Gene E.

Likens

Biogeochemistry
of a Forested
Ecosystem
Third Edition
Biogeochemistry of a Forested Ecosystem
Gene E. Likens

Biogeochemistry of a
Forested Ecosystem
Third Edition

With assistance from

Donald C. Buso
Cary Institute of Ecosystem Studies
Gene E. Likens
Cary Institute of Ecosystem Studies
Millbrook, NY, USA

ISBN 978-1-4614-7809-6 ISBN 978-1-4614-7810-2 (eBook)

DOI 10.1007/978-1-4614-7810-2
Springer New York Heidelberg Dordrecht London

Library of Congress Control Number: 2013942392

© Springer Science+Business Media New York 2013

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed. Exempted from this legal reservation are brief excerpts in connection
with reviews or scholarly analysis or material supplied specifically for the purpose of being entered and
executed on a computer system, for exclusive use by the purchaser of the work. Duplication of this
publication or parts thereof is permitted only under the provisions of the Copyright Law of the Publisher’s
location, in its current version, and permission for use must always be obtained from Springer.
Permissions for use may be obtained through RightsLink at the Copyright Clearance Center. Violations
are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

I am pleased to dedicate this THIRD
EDITION to my colleagues and friends,
F. Herbert Bormann, John S. Eaton,
Noye M. Johnson, and Robert S. Pierce.

These original coauthors made major

contributions to the First Edition, published
in 1977. Herb Bormann was a coauthor
of the Second Edition, published in 1995.
There have been many changes in the
biogeochemistry of the Hubbard Brook
Valley, several quite unexpected, since 1977.
Unfortunately, Herb, John, Noye, and Bob
are now deceased, but they are fondly
remembered and acknowledged for their
numerous scholarly and stimulating
contributions and especially for their
collegial interactions and friendship.
Special Dedication

A special dedication to my wife, Phyllis who,

despite struggling with a serious illness, did
all of the word processing and reference
checking and much of the reference
searching for this edition. Her dedication,
enthusiastic support and love throughout this
project are acknowledged with utmost
appreciation and respect. Her insightful
inputs and constant encouragement were
vital to the completion of this book. I am
most pleased to recognize her unique and
profound contributions to this 3rd Edition
and to the Hubbard Brook Ecosystem Study.
Preface to the Third Edition

As we approach 50 years of research in the Hubbard Brook Experimental Forest in

the White Mountains of New Hampshire, it is a time again for analysis and synthe-
sis, as well as for reflection. This is a momentous event for the Hubbard Brook
Ecosystem Study (HBES), which officially began on 1 June 1963. These continu-
ous, long-term data on precipitation and streamwater chemistry are, to my knowl-
edge, the longest in existence. Remarkably, funding from the National Science
Foundation for the HBES has been continuous and generous since 1963. And other
sources of funding including The Andrew W. Mellon Foundation, the Environmental
Protection Agency, and support from the home institutions of the numerous princi-
pal investigators working in the HBES, particularly Dartmouth College, Cornell
University, Yale University, and the Institute of Ecosystem Studies during the early
years of the project, are gratefully acknowledged for support of the research under-
pinning this Third Edition. This long-term work would not have been possible with-
out the cooperation and support of Forest Service colleagues at the Durham, NH
office of the Northern Research Station; U.S. Forest Service in Newtown Square,
PA; and especially those Forest Service personnel stationed at the Hubbard Brook
Experimental Forest.
This is the Third Edition of Biogeochemistry of a Forested Ecosystem, first pub-
lished in 1977. At that time, there were five authors. Coauthors F. Herbert Bormann,
Robert S. Pierce, John S. Eaton, and Noye M. Johnson are now deceased. The wise
counsel of all four of these colleagues is greatly missed as I prepare this Third
Edition, but more, I miss their close friendship, lively encouragement, and construc-
tive generation and questioning of ideas and concepts.
The goal of this Third Edition is to update long-term data presented in earlier
editions and to generate new syntheses and conclusions about the biogeochemistry
of the Hubbard Brook Valley based on these longer-term data. There have been
many changes, revelations, and exciting new insights generated from the longer-
term data records. For example, the impact of acid rain peaked during the period of
the HBES and is now declining. The longer-term data also posed challenges in that
very marked changes in fluxes occurred in some components, such as hydrogen ion
and sulfate deposition, calcium and nitrate export in stream water, and biomass

ix
x Preface to the Third Edition

accumulation, during the almost 50 years of record. Thus, presenting “mean” or

“average” conditions for many components for such a long period, when change
was so prominent, does not make sense. In some cases, pentads or decades of time
are compared to show these changes in a more smoothed and rational way for this
long period. In some cases, a single period, often during periods of rapid change,
such as acidification, is used to illustrate the main point(s). And, for some elements,
a unique mass-balance approach, allowing the calculation of the net ecosystem flux
(NEF), is shown on an annual basis throughout the study.
Because of new data and longer trends since the previous editions, the text of this
Third Edition has been largely rewritten, figures added and redrawn, and tables
reformulated, but the relatively open format of the First Edition has been retained in
response to positive feedback received over the years. I believe the concepts and
approaches originally laid out in the First Edition in 1977 and updated and enlarged
here are still compelling for use in courses such as biogeochemistry, limnology,
ecology, forestry, hydrology, and environmental science.
Throughout this third edition, “Hubbard Brook,” as used in earlier editions, has
been replaced with HBEF (Hubbard Brook Experimental Forest) to avoid confusion
between the river and the place that it drains.
Most of the routine long-term data are now accessible on the website for the
HBES (http://www.hubbardbrook.org) in an unsynthesized form. These data have
been accessed and used by numerous scholars, teachers, and students interested in
long-term biogeochemical data. We have welcomed and encouraged this use but
have urged users to be guided by ethical considerations, especially regarding attri-
bution and partnerships. The HBES was one of the first long-term ecological proj-
ects to make data freely available by initiating a “Source of the Brook” effort to
share our data with colleagues, scholars, and the public as this option became pos-
sible through computer technology (Anonymous 1990).
From the beginning, field notes were routinely and meticulously prepared at the
time samples were collected to document and clarify environmental conditions and
data relative to each sample collected. These notes have been of enormous value in
providing an historical perspective and a context for critical conditions that other-
wise may have become clouded in our memories with time.
In addition, samples of all kinds, but focused on precipitation and stream water
collected throughout the period of the HBES, are archived in a special building con-
structed for this purpose. Robert Pierce’s leadership and vision was instrumental in the
construction of this “Archive Building” at the Hubbard Brook Experimental Forest.
As mentioned in the Prefaces to previous editions, biogeochemical data from
experimentally manipulated watershed-ecosystems are not featured in this book.
Those data are presented and considered elsewhere, including our book, Pattern and
Process in a Forested Ecosystem (Bormann and Likens 1979).
I am greatly indebted to Phyllis C. Likens for word processing, literature
searches, and data management for this manuscript and to Donald C. Buso for data
management, including a massive effort with figure and table preparation, and
numerous helpful and stimulating discussions. Technical support with field collec-
tions and data compilation were provided most recently by Tammy Wooster and
Preface to the Third Edition xi

Brenda Minicucci. The majority of the solute analyses conducted since the previous
edition were done under the supervision of Kathleen Weathers and Denise Schmidt
at the Institute of Ecosystem Studies (now the Cary Institute of Ecosystem Studies).
Thanks to James LaBaugh for stimulating and helpful discussions. A special thanks
to Scott Bailey for reviewing Chap. 5, to Mark Green and Kevin McGuire for
reviewing Chap. 2, to Thomas J. Butler for reviewing Chap. 3, and to Gary Lovett
for reviewing Chap. 6. Their helpful comments and insights are greatly valued and
have contributed to a much improved final product. Acknowledgments to previous
technicians, students, and postdocs contributing to these long-term data are given in
the Prefaces to earlier editions. Significant writing was done while in residence at
the University of Connecticut, Storrs and Uppsala University, Uppsala, Sweden.
Generous support from these universities is much appreciated. I obtained a special
OPUS grant from the National Science Foundation to help support the data manage-
ment and writing of this Third Edition. This support is gratefully acknowledged.

Millbrook, NY, USA Gene E. Likens

Preface to the Second Edition

When we originally published Biogeochemistry of a Forested Ecosystem in 1977,

the Hubbard Brook Ecosystem Study (HBES) had been in existence for 14 years,
and we included data through 1974, or a biogeochemical record of 11 years. Now
our continuous, long-term biogeochemical records cover more than 31 years, and
there have been many changes. The most notable change, however, is that three of
our coauthors on the original volume are now deceased. They are deeply missed in
so many ways.
In spite of the longer records, different trends, and new insights, we believe that
the basic concepts and approaches we presented in 1977 represent the most valuable
contribution of the original edition. They are still valid and useful, particularly for
an introductory study of, or course in, biogeochemistry. Our goal in this revision is
to preserve these features, correct errors, and revise or eliminate misleading or
ambiguous short-term data (11 years!), while maintaining approximately the origi-
nal length and the modest cost.
We have left the data of the original text largely unchanged; a few figures and
tables have been replaced, revised, or added. We provide selected references to
guide the interested reader to some of the longer-term data and newer research of
the HBES. To accomplish these objectives, each chapter has a brief addendum of
especially interesting new approaches, important long-term trends, and new insights
gained from the longer-term perspective (31 years). We specifically draw attention
to a few of the HBES articles published since 1977 for a more detailed description
and analysis. A complete listing of the publications from the HBES can be found in
“Publications of the Hubbard Brook Ecosystem Study” (Likens, 1994). Data from
the HBES are available on a publicly accessible system (computer modem tele-
phone number 603-868-1006) maintained by the U.S.D.A. Forest Service and the
HBES.
As in the original version of this book, we do not include or consider here data
from the experimentally manipulated (e.g., by clearcutting) watershed-ecosystems
at Hubbard Brook, which is covered in our book, Patterns and Processes in a
Forested Ecosystem (Bormann and Likens, 1979).

xiii
xiv Preface to the Second Edition

We thank Phyllis C. Likens for help in preparing the manuscript for this revision.
We also acknowledge the help of Donald C. Buso with data analysis. Major finan-
cial support for the HBES since 1977 has been provided by the National Science
Foundation and the Andrew W. Mellon Foundation.

September 1994 Gene E. Likens

F. Herbert Bormann
Preface to the First Edition

About 15 years ago we began the Hubbard Brook Ecosystem Study with the devel-
opment of an ecosystem model and the conception of a method whereby some
major parameters of the model could be directly measured under field conditions.
The method, called “the small watershed technique,” allowed measurement of input
and output of chemicals and the construction of ecosystem nutrient budgets.
Although the Hubbard Brook study of nutrient cycling originated with ideas devel-
oped by F.H. Bormann and G.E. Likens, its early growth was aided by Robert
Pierce, forest hydrologist; Noye Johnson, geochemist; and John Eaton, forest ecolo-
gist. Donald W. Fisher of the U.S. Geological Survey also cooperated in the early
phases of the project and provided numerous data on the chemistry of precipitation
and stream water. Particular credit is due the U.S. Forest Service, whose scientists
chose the Hubbard Brook Valley as a hydrologic study site, selected particular
watersheds for intensive measurement, carried out a variety of basic hydrologic
studies, and in general cooperated with us in many ways to make the Hubbard
Brook Ecosystem Study a reality.
The initial part of the ecosystem study was concerned primarily with nutrient
flux and cycling, and it was done slowly and deliberately. The entire effort during
the first few years of study was carried forward by three of us at Dartmouth College,
with the cooperation of the U.S.D.A. Forest Service. We had no precedents to fol-
low, because similar, comprehensive studies of natural ecosystems had not been
done. We reasoned that it would be best to first construct a solid base of studies on
nutrient–hydrologic interactions upon which subsequent studies could be built. In
this regard, we were fortunate to rather quickly determine quantitative nutrient bud-
gets for replicated ecosystems. These results gave us guiding insight into the func-
tion of natural ecosystems and into the development of future lines of research.
Nevertheless, we felt that slower growth would be more manageable and would
allow for substantial interaction among all senior investigators to ensure proper
coordination and development of the overall study. Generally, the method we used
to guide the growth of the Hubbard Brook research proposals was as follows: First,

xv
xvi Preface to the First Edition

based on our own perceptions and feedback from ongoing studies, from cooperating
scientists, and from outside advisors, we identified research problems that were
timely and particularly pertinent to our overall goals. Some of the studies were
launched under our direction; others were brought to the attention of an established
investigator working in that area. We then sought a mutually satisfactory arrange-
ment to allow the investigator to work at Hubbard Brook. From the beginning, we
have called attention to the kinds of information a cooperative study might produce,
but we have always encouraged individuality in the design and execution of
research. We deem this individual research freedom one of the greatest assets of the
Hubbard Brook Study. Individuality in selection of problems and conduct of
research is also encouraged among graduate students at Hubbard Brook. Not only
does this contribute to the intellectual ferment and sound growth of the Hubbard
Brook Study, but as educators, we feel this approach absolutely necessary if the
Hubbard Brook Study is to contribute to graduate education.
Over the last 15 years the study has grown to include not only hydrology and
input–output chemistry, but also aspects of the structure, function, and dynamics of
the forest ecosystem itself. In addition to studies of undisturbed ecosystems, entire
ecosystems have been experimentally manipulated to allow comparative study of
undisturbed and perturbed ecosystems. Stream and lake ecosystems within the
Hubbard Brook Valley have been studied in detail, and the connection between the
forest ecosystem and these interlinked aquatic ecosystems is now the focus of con-
siderable attention.
Approximately 50 senior scientists and scores of graduate students have carried
out studies at Hubbard Brook during the last 15 years. Over 200 publications have
resulted from this work.
We now plan to organize some of this information in ways we consider useful
both to the scientists concerned with the theory of biogeochemical cycles and the
structure, function, and development of forested ecosystems, and to the land-use
specialists who are concerned with the production of a variety of benefits, goods,
and services from northern hardwood ecosystems.
We have developed two volumes so far, and a third is planned for the future. This,
the first volume, presents a detailed examination of the biogeochemistry of an
undisturbed, aggrading, second-growth northern hardwood forest at Hubbard
Brook. Ecosystems similar to this cover a large area of northern New England and
New York. Major emphasis in this book is on the physical aspects of nutrient and
hydrologic flow through the ecosystem and nutrient budgetary. Not only are we
concerned in this book with the presentation of data and conclusions that may be
drawn from them, but we also discuss methodology as it influences results, and we
try to share some of our experience at Hubbard Brook in the hope that other workers
can avoid pitfalls where we have not been so fortunate. Moreover, we hope that the
data presented in this volume can serve as a basis for decision-making and manage-
ment schemes, including modeling within the northern hardwood forest landscape.
The second volume has as its primary concern the structure, function, and devel-
opment through time of the northern hardwood ecosystem and is designed for the
advanced student in ecology as well as for the ecosystem specialist. It concentrates
Preface to the First Edition xvii

on the interrelationships among biogeochemical processes, structure, and species

behavior within the ecosystem and how these change with time following perturba-
tion. We plan for the information in these two volumes to interrelate and thus draw
support for each other. Biogeochemistry is presented more fully in this first volume,
whereas biological considerations are dealt with more fully in the second volume.
By no means do these two volumes summarize the Hubbard Brook Study. In the
future we hope to put together a more comprehensive volume that involves contri-
butions from the many cooperating scientists who have brought their particular spe-
cialties to bear on the Hubbard Brook Ecosystem.
There are many persons to acknowledge and thank for contributions to such a
multidisciplinary study conducted over so many years. Many data were collected
during rainstorms, blizzards, intense cold, darkness, and numerous other types of
northern New Hampshire weather. Long, often tedious hours in the laboratory and
at the computer center were also required. The ideas, devotion, loyalty, and encour-
agement and genuine enthusiasm of field technicians, laboratory technicians, stu-
dents, and various other colleagues made the overall Hubbard Brook Ecosystem
Study a reality and is beyond a simple “thank you.”
For this little book, we owe debts great and small to many people—to those
whose ideas and encouragement provided an intellectual thrust to move ahead; to
those without whose devotion, skills, and endurance there would have been no prog-
ress at all; and to those mostly good-natured colleagues whose shafts of humor and
criticisms have helped us all to keep a realistic perspective. For all of these things
and more, we thank:
J. Aber, B. Ackerman, M. Alexander, T. Allison, H. Art, P. Art, S. Bennett,
R. Bilby, J. Bishee, C. Black, A. Bock, H. Bond, B. Bormann, R. Bormann, D.
Botkin, E. Brastrup, J. Brink, S. Brown, A. Brush, L. Burckes, T. Burton, D. Buso,
P. Cantino, J. Chamberlain, S. Chisholm, B. Coffin, C. Cogbill, B. Corson, W.
Covington, M. Cushman, H. Davis, M. Davis, F. deNoyelles, G. deSylva, A.
Dominski, J. Duggin, F. Edington, Y. Falusi, A. Federer, N. Fernando, S. Fiance,
A. Field, D. Fisher, S. Fisher, L. Forcier, M. Fox, J. Fryer, G. Furnival, J. Galloway,
A. Gambell, D. Gerhart, J. Gilbert, W. Glanz, J. Gosz, C. Grant, M. Gross, R. Hame,
G. Hart, B. Haywood, J. Hobbie, R. Holmes, S. Holmgren, J. Hornbeck, S. Howe,
J. Janak, P. Johnson, M. Jordan, F. Juang, H. Karsten, M. Keller, W. Kent, L. Kivell,
J. Koonce, S. Koppes, D. Korbobo, M. Koterba, B. Kuehn, R. Lavigne, T. Ledig, J.
Lehman, R. Leonard, V. Levasseur, E. Loker, B. Mahall, J. Makarewicz, P. Marks,
W. Martin, D. Masza, K. McConnochie, O. MacGregor, D. McNaught, A. McPhail,
J. Melillo, R. Mendelsohn, R. Miller, J. Meyer, R. Moeller, M. Mollony, D. Mulcahy,
R. Muller, J. Murphy, P. Murtaugh, D. Nelson, J. Nobel, R. Patrick, P. Pelton, B.
Peterson, N. (Crane) Poole, G. Potter, R. Reifsnider, W. Reiners, R. Reynolds, L.
Riggs, J. Roskoski, D. Ryan, M. Schmale, P. Sargent, E. Sayetta, V. Scarlet, W.
Schlesinger, J. Schorske, L. Schwartz, M. Sebring, T. Sherry, T. Siccama, D. Smith,
W. Smith, P. Sprague, D. Sprugel, D. Stout, S. Strom, F. Sturges, J. Tinker, P.
(Fruend) Tinker, J. Thompson, G. Thornton, D. Tiernan, P. Toyryla, K. Turekian, G.
Voigt, J. Wallis, B. Walsh, R. Walter, P. Wicklund, R. Withington, G. Whitney, R.
Whittaker, T. Wood, G. Woodwell, J. York, and C. Zoltov.
xviii Preface to the First Edition

We all acknowledge continued indulgence from our and our coauthors’ wives
(Kay, Chris, Carol, Diane, and Delores) and families for time spent away from
them.
Financial support for the field studies was provided by the National Science
Foundation and the United States Forest Service. This work was done through the
cooperation of the Northeastern Forest Experiment Station, Forest Service, United
States Department of Agriculture, Upper Darby, Pennsylvania.
The manuscript was initially formulated while Gene E. Likens was a John Simon
Guggenheim Fellow at the Australia National University. This support is gratefully
acknowledged.

January 1977 Gene E. Likens

F. Herbert Bormann
Contents

Preface to the Third Edition .......................................................................... iv

Preface to Second Edition............................................................................... xi
Preface to First Edition................................................................................... xiii

1 Ecosystem Analysis ................................................................................... 1

Small Watershed Approach......................................................................... 4
The Hubbard Brook Ecosystem .................................................................. 5
Location .................................................................................................. 5
Climate .................................................................................................... 5
Area, Topography, and Aspect ................................................................ 7
Geology ................................................................................................... 7
Soils......................................................................................................... 8
Vegetation and Fauna .............................................................................. 8
Drainage Streams .................................................................................... 13
Concluding Thoughts .................................................................................. 13
2 Hydrology .................................................................................................. 17
The Water-Year ........................................................................................... 17
Precipitation ................................................................................................ 18
Streamflow and Evapotranspiration ............................................................ 23
Deep Seepage .............................................................................................. 29
Representativeness of the Hydrology of the Hubbard Brook
Experimental Forest for Northern New England .................................... 32
3 Chemistry................................................................................................... 35
Precipitation Chemistry .............................................................................. 38
Annual Volume-Weighted Concentrations.............................................. 38
Origin of Ions in Bulk Precipitation........................................................ 41
Dry Deposition ........................................................................................ 42

xix
xx Contents

Elevational Effects .................................................................................. 43

Acid Precipitation ................................................................................... 44
Throughfall and Stemflow .......................................................................... 66
Streamwater Chemistry ............................................................................... 68
Concentrations of Dissolved Substances ................................................ 69
Valleywide Streamwater Chemistry: A Synoptic View .............................. 74
Seasonal Variations in Bulk Precipitation
and Streamwater Concentrations ............................................................ 77
Long-Term Trends in Bulk Precipitation
and Streamwater Chemistry .................................................................... 85
4 Input–Output Budgets .............................................................................. 89
Mass Meteorologic Input of Dissolved Substances .................................... 90
Annual Variation in Net Hydrologic Budgets ............................................. 90
Some Examples of Important Components of Input–Output Budgets ....... 101
Nitrogen .................................................................................................. 101
Chlorine................................................................................................... 103
Sulfur....................................................................................................... 103
Dissolved Silica ...................................................................................... 104
Input–Output Budgets: Summary ............................................................... 105
Monthly Variations...................................................................................... 109
Output Consistently Greater Than Input ................................................. 109
Input Consistently Greater Than Output ................................................. 110
Crossover Patterns................................................................................... 113
Relationship of Annual Mass Output of Dissolved Substances
to Annual Streamflow ............................................................................. 116
Annual Variation in Mass Output of Dissolved Substances ................... 117
Particulate Matter .................................................................................... 120
Particulate Matter Versus Dissolved Substance Export .......................... 121
The Role of Debris Avalanches in Landscape Denudation......................... 124
Long-Term Changes in Input–Output Budgets ........................................... 126
5 Weathering................................................................................................. 127
Sources of Hydrogen Ion in the Weathering Reaction................................ 130
Carbon ..................................................................................................... 131
Nitrogen .................................................................................................. 132
Sulfur....................................................................................................... 133
Estimates of Ecosystem Weathering Flux................................................... 133
Net Soil Release .......................................................................................... 134
6 Nutrient Cycles and Mass Balances ........................................................ 139
The Calcium Cycle ..................................................................................... 139
The Potassium Cycle................................................................................... 141
The Sulfur Cycle ......................................................................................... 142
Nutrient Cycle Relationships at the HBEF ................................................. 146
Contents xxi

Annual Watershed-Ecosystem Mass Balances (“Budgets”) ....................... 147

Calcium ................................................................................................... 149
Sodium .................................................................................................... 151
Potassium ................................................................................................ 152
Sulfur....................................................................................................... 153
Nitrogen .................................................................................................. 153
Nitrogen Retention in Forested Watershed-Ecosystems ............................. 155
Chlorine................................................................................................... 155
Figures 53, 54, and 55. Ecosystem Diagrams: Further Explanations ......... 158
7 The Northern Hardwood Ecosystem in the Hubbard Brook Valley
in Relation to Other Forest Ecosystems .................................................. 163
8 Summary Discussion and Conclusions.................................................... 171
Epilog .......................................................................................................... 180
Odyssey ................................................................................................... 180

Erratum .......................................................................................................... E1
References ........................................................................................................ 183
Index ................................................................................................................. 201
Chapter 1
Ecosystem Analysis

An ecological system has a richly detailed series of inputs and outputs of energy and
matter. Because of the lack of precise information about these relationships and the
internal functions that maintain the ecosystem, it is often difficult to assess the
impact of human activities on the biosphere. As a result, land-use managers and
planners often cannot take into account or even foresee the full range of conse-
quences a project may have. Without full information, the traditional practice in the
management of land resources has been to emphasize strategies that maximize the
output of some desirable product or service and give little or no thought to the long-
term secondary effects. As a result often there is much conflict between conserva-
tion and management goals. For example, in an all-out effort to increase food
production, drainage waters carry large burdens of pollutants from fertilizers, pesti-
cides, and farm wastes. Forests may be cut with inadequate perception of the effects
on regional water supplies, wildlife, recreation, and aesthetic values. And despite
recent regulations to protect wetlands, they have been converted worldwide to com-
mercial use with little concern over important hydrologic, biodiversity, esthetic, and
commercial values lost in the conversion.
For some time, it has been evident that a new conceptual approach to the man-
agement of natural resources is desirable. One approach that has received consider-
able attention recently is to consider entire ecological systems as integrated
interacting units, instead of limiting studies to only a few components of the system.
In an experimental forest in New Hampshire, we have conducted a large-scale and
long-term investigation aimed at supplying the kind of information that is usually
lacking about ecosystems.
A vast number of variables, including biologic structure and diversity, geologic
heterogeneity, climate, and season, control the flux and cycling of both water and
chemicals through and within ecosystems. Clearly, both living and nonliving com-
ponents of ecosystems are important in defining and regulating the flow of matter
within and between ecosystems. Because chemicals tend to circulate from nonliving
components to living organisms and back to the nonliving environment, the path-
ways have been termed biogeochemical cycles. Ecosystems continually exchange

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2_1, 1

© Springer Science+Business Media New York 2013
2 1 Ecosystem Analysis

Fig. 1 Conceptual model depicting nutrient relationships in a terrestrial ecosystem. Inputs and
outputs to the ecosystem are moved by meteorologic, geologic, and biologic vectors (Bormann and
Likens 1967; Likens and Bormann 1972). Major sites of accumulation and major exchange path-
ways within the ecosystem are shown. Nutrients that, because they have no prominent gaseous
phase, continually cycle within the boundaries of the ecosystem between the available nutrient,
organic matter, and primary and secondary mineral components tend to form an intrasystem cycle.
Fluxes across the boundaries of an ecosystem link individual ecosystems with the remainder of the
biosphere. Modified from Likens (1992)

(gain or lose) matter and energy with other ecosystems and the biosphere as a whole.
These biogeochemical fluxes and the internal cycles are vital to the functioning and
continuance of the system. Comparisons of biogeochemical data from natural eco-
systems with those that have been manipulated by humans or otherwise disturbed
then provide important information about the functional efficiency or “health” of an
ecosystem.
For almost five decades, the biogeochemistry of the Hubbard Brook Experimental
Forest, hereafter referred to as HBEF (Fig. 1), has been the subject of continuous
and detailed study, referred to as the Hubbard Brook Ecosystem Study (HBES).
More than 35,000 water samples, including rain, snow, stream water, ground water,
throughfall, and stemflow, have been collected and chemically analyzed. In addi-
tion, thousands of samples of air, soil, rocks, and biota have been collected and
chemically analyzed. These extensive data now provide the means and the perspec-
tive to identify, isolate, and quantify some of the complex and interlocking biogeo-
chemical processes of a forest ecosystem. As this study shows, seasonal and
year-to-year variations in flux rates and associated biologic responses often require
long-term analysis before reliable generalizations can be drawn about such complex
natural systems. This book summarizes the understanding, as derived from almost
1 Ecosystem Analysis 3

50 years of data, of the biogeochemical flux and cycling of water and nutrients for
a northern hardwood forest ecosystem in New England, refers to the timescale at
which the various biogeochemical processes operate, and relates these data to other
ecosystems with long-term data.
Studies began in 1963 to determine the magnitude of the biogeochemical flux
and cycling of water and nutrients in northern hardwood forest ecosystems in the
White Mountains of New Hampshire. The ecosystem unit we used is a watershed or
drainage area, with vertical and horizontal boundaries defined functionally by bio-
logic activity and the drainage of water (Bormann and Likens 1967). Six adjacent,
small south-facing and three north-facing watersheds within the HBEF, with roughly
similar vegetation, geology, and climate, have been used for these studies (Bormann
and Likens 1967; Likens and Bormann 1972). Differences across the valley and on
south-facing and north-facing slopes gave opportunities for other comparisons and
analyses (e.g., Likens and Buso 2006). This series of watershed-ecosystems pro-
vided the possibility for replication, as well as a design for experimental manipula-
tion of entire natural ecosystems (e.g., Likens et al. 1970). Nevertheless, this book
will focus largely on the long-term biogeochemical results of Watershed 6 (W6), the
biogeochemical reference watershed for the HBEF.
A conceptual model of these deciduous forest ecosystems was formulated at the
outset to facilitate the development of key questions and to guide the quantitative
evaluation of the input–output flux and cycling of water and chemicals (Bormann
and Likens 1967; Likens and Bormann 1972). The continuous flow of energy, water,
nutrients, and other materials across the ecosystem’s boundaries is considered to be
inputs and outputs, which are transported by meteorologic, geologic, and biologic
vectors (Fig. 1). Meteorologic inputs and outputs consist of wind-borne particulate
matter, dissolved substances in rain and snow, aerosols, and gases (e.g., CO2).
Geologic flux includes dissolved and particulate matter transported by surface and
subsurface drainage and the mass movement of colluvial materials by gravity.
Biologic flux results when chemicals or energy gathered by animals in one ecosys-
tem are deposited in another (e.g., local exchange of nitrogen or phosphorus in fecal
matter, seed dispersal by seasonal bird movements, or transport of calcium or phos-
phorus in moose antlers). These input–output categories are therefore defined as
vectors or “vehicles” for transport of nutrients, matter, or energy rather than sources;
i.e., a leaf blown into an ecosystem represents meteorologic input rather than bio-
logic input.
Within the ecosystem, the nutrients may be thought of as occurring in any one of
the four major compartments: (1) atmosphere, (2) living and dead organic matter,
(3) available nutrients, and (4) primary and secondary minerals (soil and rock). The
atmospheric compartment includes all elements in the form of gases or aerosols
both above- and belowground. Available nutrients are ions that are absorbed on or
in the soil complex or dissolved in soil water. The organic compartment includes all
nutrients incorporated in living and dead biomass. The woody tissues of living veg-
etation are considered a part of the living biomass. The primary and secondary
minerals contain nutrients that comprise the inorganic soil and rock portions of the
ecosystem.
4 1 Ecosystem Analysis

The biogeochemical cycling of elements involves an exchange between these

various compartments within the boundaries of the ecosystem. Available nutrients
and gaseous nutrients may be taken up and assimilated by the vegetation and micro-
organisms; some may be passed on to heterotrophic consumers and then made avail-
able again through respiration, biologic decomposition, and/or leaching from living
and dead organic matter. Insoluble primary and secondary minerals may be con-
verted to soluble available nutrients through the general process of weathering;
soluble nutrients may be redeposited as secondary minerals.

Small Watershed Approach

In humid regions, chemical flux and cycling are intimately linked to the hydrologic
cycle. Hence, it is impossible to measure the input and output of nutrients without
simultaneously measuring the input and output of water. The problem usually is that
subsurface flows of water, which can be a significant fraction of the hydrologic
cycle, are very difficult to measure quantitatively.
About 50 years ago, it occurred to us that under certain circumstances, the inter-
action of the nutrient cycle and the hydrologic cycle could be turned to good advan-
tage in the quantitative study of an ecosystem. The requirements were that the
ecosystem be a watershed underlain by watertight bedrock or some other relatively
impermeable base. In that case, the only inputs to the ecosystem would be meteoro-
logic and biologic; geologic input need not be considered because there is no trans-
fer by gravity between adjacent watersheds on an ecological timescale. In humid
areas where surface wind abrasion is a minor factor, losses from the system are only
geologic and biologic. Given an impermeable base, all the geologic output would
inevitably turn up in the streams draining the watershed where it could be measured
quantitatively. When the watershed is part of a larger and fairly homogenous biotic
unit, the biologic output tends to balance the biologic input because most animals
move indiscriminately into and out of the watershed, randomly acquiring or releas-
ing nutrients. Therefore, it is necessary to measure only the meteorologic input and
the geologic output (net hydrologic flux, NHF) of nutrients to arrive at the net gain
or loss of a given nutrient in the ecosystem. Watershed-ecosystems in the HBEF are
small enough to allow quantitative evaluation of ecosystem complexity and permit
whole-watershed manipulations, but large enough to be relevant to management
considerations.
Using this approach, it has been possible to obtain quantitative annual NHF bud-
gets for most of the major chemical nutrients. The flux of some nutrients with a
gaseous phase (e.g., N and S) is more difficult to evaluate, but the ecosystem
approach allowed realistic assumptions and estimates to be made about some of
these unmeasured gaseous components. More recently, the use of isotopes, “flux
towers,” and other techniques has provided more quantitative estimates of these
fluxes. Incorporating other components, such as weathering, net soil release, dry
deposition, and long-term biomass storage, allowed calculations of net ecosystem
The Hubbard Brook Ecosystem 5

flux (NEF) for watershed-ecosystems. Despite its apparent simplicity, the NEF is a
powerful tool for understanding the biogeochemistry of watershed-ecosystems,
particularly when done for long periods.
We have used these approaches in an intensive study of the biogeochemistry of
six contiguous south-facing watersheds and three contiguous north-facing water-
sheds within the HBEF (Fig. 2). These watershed-ecosystems are all tributary to
Hubbard Brook. As mentioned above, this book will focus on W6 of the HBEF.

The Hubbard Brook Ecosystem

The HBEF was established in 1955 by the U.S. Forest Service as the principal
research area for forest hydrology and the management of watersheds in New
England (Hornbeck 2001; McGuire and Likens 2011; Likens and Bailey 2013). It is
operated and managed by the Northeastern Research Station of the U.S. Forest
Service in cooperation with the White Mountain National Forest. The name of the
area is derived from the major drainage stream in the valley, Hubbard Brook.
Hubbard Brook flows generally from west to east for about 13 km (Fig. 2) until it
joins with the Pemigewasset River, which ultimately forms the Merrimack River
and discharges into the Atlantic Ocean in Massachusetts. Water from more than 20
tributaries enters Hubbard Brook along its course. Mirror Lake, a small oligotrophic
lake, also discharges into Hubbard Brook at the lower end of the valley. Hubbard
Brook is a fifth-order river at its outlet from the HBEF.

Location

The HBEF is located within the townships of North Woodstock (~60 %), Ellsworth
(~40 %), and Warren (<1 %), NH, within the White Mountain National Forest of
north central New Hampshire. Coordinates of 43°56′N, 71°45′W bisect the area.
The Atlantic Ocean is about 116 km to the southeast. The nearest town is West
Thornton, NH.

Climate

Although the climate varies with elevation, it is classified as humid continental with
short, cool summers and long, cold winters (Trewartha 1954). The climate may be
characterized by (1) changeability of the weather, (2) a large range in both daily
and annual temperatures, and (3) equable distribution of precipitation on a monthly
basis. The HBEF lies in the heart of the middle latitudes, and the majority of the
air masses therefore flow from west to east. During the winter months, these are
 6
1

Fig. 2 Hubbard Brook Experimental Forest, located in the townships of North Woodstock, Ellsworth, and Warren, New Hampshire, USA. Modified from
Ecosystem Analysis

J. Campbell, USDA Forest Service, Northern Research Station - 7

The Hubbard Brook Ecosystem 7

northwesterlies, and during the summer, the air generally flows from the southwest.
Therefore, the air affecting HBEF is predominantly continental. However, during
the autumn and winter, cyclonic disturbances periodically move up the east coast of the
USA providing an occasional source of maritime air. The mean air temperature in
July is 18.7 °C and in January is −8.3 °C (Federer et al. 1990; A. Bailey et al. 2003).
A continuous snowpack normally develops each winter to a depth currently of
about 1 m. Occasionally, mild temperatures in midwinter partly or wholly melt the
snowpack. A significant microclimatologic feature of this area is that even the
uppermost layer of the forest soils usually remains unfrozen during the coldest
months because of the thick humus layer and a deep snow cover (Hart et al. 1962).
The depth of snow cover and its duration, the occurrence of soil frost, and the date
of spring snowmelt have been decreasing/advancing during the past 50 years or so
(e.g., Campbell et al. 2007; Huntington et al. 2009; Vadeboncoeur et al. 2009; Judd
et al. 2011).

Area, Topography, and Aspect

The HBEF covers an area of about 32 km2 and ranges in altitude from 222 to
1,015 m. The experimental watershed-ecosystems (Watersheds 1–6, 7–9; Fig. 2)
range in size from 12 to 77 ha and in altitude from 500 to 910 m. These headwater
watersheds are all steep (average slope of 20–30 %), with well-incised channels and
relatively distinct topographic divides. The height of the land surrounding each
watershed-ecosystems, the area, and the topography all have been determined from
ground surveys and aerial photography and most recently augmented by Laser
Imaging Detection and Ranging (LIDAR). Experimental Watersheds 1–6 face S to
SE; Watersheds 7–9 face N to NE.

Geology

The geologic substrate, outcrops of bedrock and stony till, in the Hubbard Brook
Valley was exposed some 14,000 years ago when the glacial ice sheet retreated
northward (Likens and Davis 1975; Davis and Ford 1985). The eastern portion of
the Hubbard Brook Valley is underlain by a complex assemblage of metasedimen-
tary and igneous rocks (Barton et al. 1997). The major component is the Silurian
Rangeley Formation, consisting of quartz–mica schists and quartzite, metamor-
phosed to sillimanite grade. These metaphoric rocks also were intruded by a variety
of igneous rocks, including the Concord granite, quartz diorite, pegmatite, and dia-
base and lamprophyre dikes. The western portion of the valley is underlain by
Devonian Kinsman Granodiorite, a foliated, granitic rock distinguished by large
phenocrysts of potassium feldspar (Barton et al. 1997). Much of the area of the
experimental watersheds is covered with glacial till derived from local bedrock,
8 1 Ecosystem Analysis

primarily granites, schists, and quartzites; materials from other lithologies N-NW of
HBEF are common in small quantities (Bailey and Hornbeck 1992). Movement of
surface water by deep seepage through the compacted till and crystalline bedrock is
considered minimal (Bormann and Likens 1967; Likens et al. 1967; see Chap. 2).

Soils

Soils are diverse, but well-drained spodosols (Haplorthods) of sandy loam texture,
with a thick organic layer at the surface, are most common. Most liquid precipita-
tion infiltrates into the soil at all times, and there is very little overland flow because
the surface soil is very porous, the surface topography is very rough (pit and mound,
mostly from windthrown trees), and normally there is little soil frost (Pierce 1967).
Soils in the Hubbard Brook Valley were formed in glacial till of variable depth,
underlain by metamorphic rock of igneous and sedimentary origin (schists and
quartz monzonite) (Johnson et al. 1981).
Soil depths are highly variable but average about 0.5 m from surface to bedrock
or compacted till in these headwater watershed-ecosystems. In general, they are
shallow at high elevations and increase in depth with decreasing elevation (Lawrence
et al. 1986). The average soil depth to the C horizon within the experimental water-
sheds is ~57 cm (Johnson et al. 1991a). Recent measurements using ground-
penetrating radar show scattered pockets of deeper (2–3 m) till at all elevations
within the gauged watersheds (Scott W. Bailey, personal communication). A well-
developed and well-drained surface organic layer (O horizon or forest floor) is
3–15 cm in thickness. Soil on the ridges may consist of a thin accumulation of
organic matter resting directly on the bedrock. In some places and at varying depths,
impermeable pan layers restrict vertical water movement and root development.
Soils are acidic and generally infertile (e.g., Johnson et al. 1991b). The average
pH of soil is difficult to measure, particularly at the watershed scale. Early results
indicated that HBEF soils had a pH ≥ 4.5. Currently, soils appear to be significantly
more acidic, pH 3.5 for W7 (Ross et al. 2009) and pH 3.5–3.7 for W1 prior to
Wollastonite treatment in 1998 (Peter M. Groffman, personal communication).

Vegetation and Fauna

The vegetation of the HBEF is part of the northern hardwood forest ecosystem, an
extensive forest type that extends with variations from Nova Scotia to the western
Lake Superior region and southward along the Blue Ridge Mountains (Braun 1950;
Küchler 1964; Oosting 1956). Classification of mature forest stands as northern
hardwood ecosystems rests on a loosely defined combination of deciduous and
The Hubbard Brook Ecosystem 9

coniferous species that may occur as deciduous or mixed deciduous–evergreen

stands (Photograph 1).
The forest is unevenly aged and well stocked, primarily with sugar maple (Acer
saccharum), American beech (Fagus grandifolia), and yellow birch (Betula
alleghaniensis). Red spruce (Picea rubens), balsam fir (Abies balsamea), white
birch (B. papyrifera var. cordifolia), and hemlock (Tsuga canadensis) are promi-
nent on north-facing slopes, on ridge tops, on rock outcrops, and along the main
channel of Hubbard Brook. Logging in the valley began in the 1800s, with heavy
cutting between 1906 and 1919; the forest was damaged by a hurricane in 1938 and
a large ice storm in 1998, but there is no evidence of major fire in the history of these
stands (Bormann et al. 1970; Likens 1972b; Davis 1985; Merrens 1988; Merrens
and Peart 1992; Peart et al. 1992; Rhoads et al. 2002; Likens et al. 2004b; van Doorn
et al. 2011; Charles V. Cogbill, personal communication).
The forest in W6 had a basal area of about 23 m2/ha in 1965 (Bormann et al.
1970) and 28 m2/ha in 2007 (Siccama et al. 2007). Total (above- and belowground)
live-tree biomass (~245 mg/ha for trees ≥10 cm dbh) throughout the Hubbard Brook
Valley did not change significantly between 1995 and 2005 (van Doorn et al. 2011).
Aboveground biomass accumulation for the forest on W6 averaged about 4.85 Mg/
ha-year during 1965–1977 but declined to 0.89 Mg/ha-year during 1987–1992
(Likens et al. 1994) and recently has declined even further (Lindenmayer and Likens
2010, p. 123).
Other regional components of the northern hardwood ecosystem, such as white
ash (Fraxinus americana L.) and red maple (Acer rubrum), are common in the
experimental watersheds; white pine (Pinus strobus) is rare except in the valley
floor, and red oak (Quercus rubra), basswood (Tilia americana), and elm (Ulmus
americana) are absent from the HBEF. About 26 species of trees and some 100 spe-
cies of shrubs and herbs have been identified in the Hubbard Brook Valley (Holmes
and Likens 1999).
Fauna common to northern hardwood forests occurs within the Hubbard Brook
Valley. More than 125 species of birds have been observed (Likens 1973; Holmes
and Likens 1999), and snowshoe hare (Lepus americanus), beaver (Castor
canadensis), red fox (Vulpes fulva), black bear (Ursus americanus), moose (Alces
americana), and white-tailed deer (Odocoileus virginianus) are present. The
White Mountains generally have a low population density of deer because of
severe winters and hunting pressure (Siegler 1968; Pletscher 1982; Pletscher et al.
1989). The northern, red-backed salamander (Plethodon cinereus) is a dominant
vertebrate in these forests with regard to numbers and biomass (Burton and Likens
1975). Currently, total bird abundances in the HBEF are about half of what they
were in 1970 (Holmes 2011), and terrestrial salamander abundances are about
one-third. Native (unstocked) brook trout (Salvelinus fontinalis) populate Hubbard
Brook and the reaches of some of its major tributaries (Warren et al. 2008)
(Photograph 2).
10 1 Ecosystem Analysis
Photograph 1 Aerial view of the Hubbard Brook Experimental Forest, North Woodstock, New Hampshire, during winter—1973 (area within dash dotted
lines, approximately 32 km2. Experimental watersheds shown by dotted boundaries (see Fig. 2). Watershed 2 was cleared of all woody vegetation in 1965 and
sprayed with herbicides for three successive summers to prevent regrowth. No forest products were removed. Watershed 4 was cut in 25-m-wide alternating
strips in 1970, 1972, and 1974. Picture shows Watershed 4 when two-thirds of the strips were cut and harvested. Watershed 101 was cleared and products were
removed in one operation in 1970. Watershed 5 was cut (whole-tree harvest) in 1983–1984 after this photograph was taken. Watershed 1 was treated with
Wollastonite (CaSiO3) in 1999 to test the effects of acid rain on soil calcium pools. All other watersheds are untreated as yet and serve as references to these
watershed manipulations. Dark areas on the ridges and peaks generally are conifers, and the lighter areas are deciduous northern hardwoods
The Hubbard Brook Ecosystem
11
12 1 Ecosystem Analysis

Photograph 2 A typical spodosol soil developed on till at the HBEF. The uppermost dark layer,
an organic layer called the forest floor, rests on the mineral soil. Immediately below the forest floor
is a bleached and chemically leached horizon, which grades into zones of accumulated organic
matter and iron and aluminum oxides. At still lower depths, organic matter declines and the soil
varies from partially weathered to unweathered mineral components. The presence of an obvious
leached horizon is intermittent within the HBEF
Concluding Thoughts 13

Drainage Streams

Headwater streams draining the watershed-ecosystems are small and perennial.

Flow ranges from zero during infrequent summer droughts to hundreds of m3/ha-
day during snowmelt and storm events. Five of the watersheds—3, 4, 7, 8, and 9—
are larger and typically maintain flow during summer (Fig. 2).
From June to October, the streams are heavily shaded by the forest vegetation.
As a result, daily water temperatures do not vary more than a degree or two Celsius
during that time of year. Maximum streamwater temperatures (~21 °C) occur just
before leaf out of the forest canopy. The annual temperature range in these streams
is about 0 °C to about 18 °C, with a mean of about 7 °C. The mean annual temperatures
in two of the south-facing streams (W3 and W6) declined ~2 °C from 1966 to 1983
and then increased by ~1 °C and became more variable thereafter (Likens 2011).
The stream water is normally saturated or supersaturated with dissolved oxygen,
and the pH is less than 6 in the headwaters (Likens and Buso 2006).
The streambed is covered by organic debris, fine sand, gravel, cobbles, and small
boulders. There are occasional bedrock exposures in the steam channel, and these
are preferred sites for gauging weirs. Numerous dams of organic debris occur
throughout the length of the stream; the headwater streams therefore have a stair-
step appearance of alternating “waterfalls” and pools. The stream channel com-
prises about 1–2 % of the area of each watershed-ecosystem. Areas of localized
seepage are common, but not numerous (Photographs 3 and 4).

Concluding Thoughts

Additional details concerning the topography, climate, geology, and biology of the
HBEF are given elsewhere (e.g., Likens et al. 1967; Johnson et al. 1968; Bormann
et al. 1970; Holmes and Sturges 1973; Likens 1973, 1985; Bormann and Likens
1979; Siccama et al. 1970, 2007; Whittaker et al. 1974; Holmes and Likens 1999;
Holmes 2011; http://www.hubbardbrook.org). A large number of papers have been
published on the biogeochemistry of the Hubbard Brook Valley including eight
monographs (Bormann et al. 1974; Whittaker et al. 1974; McDowell and Likens
1988; Likens et al. 1994, 1998, 2002a; Fahey et al. 2005; Lovett et al. 2005). By
coupling results from long-term biogeochemical monitoring with process-level
research, detailed mass balances have been developed for many elements
(e.g., Likens et al. 1994, 1998, 2002a, b; Likens 2004; Lovett et al. 2005; Fahey
et al. 2005; Likens and Buso 2010a; Bernal et al. 2012). Element mass balances in
combination with the small, watershed-ecosystem approach provide an excellent
conceptual framework to investigate the ecology, hydrology, and biogeochemistry
of forest watershed-ecosystems and associated aquatic ecosystems, as well as to
evaluate the response of ecosystems to disturbance. However, assessing the response
of complicated and interrelated ecosystem processes to environmental change often
requires more than monitoring and mass balances. Long-term monitoring has been
14 1 Ecosystem Analysis

Photograph 3 Hubbard Brook at an elevation of about 215 m. The characteristic boulder

substrate and “stair-step” nature of this fifth-order stream are evident

Photograph 4 Hubbard Brook at elevation of about 425 m. Photo by D.C. Buso

Concluding Thoughts 15

supplemented with plot and whole-ecosystem manipulations, long-term synthesis,

as well as modeling efforts in the HBES to generate and test hypotheses concerning
change in ecosystem structure and function and response to disturbance (some
recent HBES examples include Aber et al. 1997, 2002; Berger et al. 1997; Driscoll
et al. 1998, 2001; Christ et al. 1999; Groffman et al. 1999, 2001, 2006; Rosenberry
et al. 1999; Hall et al. 2001; Gbondo-Tugbawa et al. 2001, 2002; Peters et al. 2004;
Fitzhugh et al. 2001, 2003; Likens et al. 2004b; Likens and Buso 2006,
2010a, 2012; Bernhardt et al. 2005; Blum et al. 2002; Kobe et al. 2002; Dittman
et al. 2007; McLauchlan et al. 2007; Judd et al. 2007, 2011; Warren et al. 2007;
Conley et al. 2008; Bernal et al. 2012).
In addition to providing fundamental data on ecosystem function and the eco-
logical relationships between structure and function, these biogeochemical data
provide a baseline for judging the effects of human manipulations and disturbance
of forest landscapes as well as information useful to managers wrestling with these
problems (e.g., Likens et al. 1970; Pierce et al. 1972; Bormann and Likens 1979;
Likens and Franklin 2009; Lindenmayer and Likens 2010; Bernal et al. 2012;
Likens and Buso 2012).
The next chapters, Hydrology through Nutrient Cycles and Mass Balances, pres-
ent detailed information about the long-term biogeochemistry of the HBEF. These
sections summarize our current understanding of the flux and cycling of chemicals
in the watershed-ecosystems at the HBEF in the context of the long-term changes
that have occurred. An attempt is made to synthesize results and to interpret and
comment on outstanding problems, solved and unsolved.
Chapter 2
Hydrology

Because of the vital role of water as a transporting agent, chemical solvent, weathering
agent, and catalyst, quantitative data on hydrology are of paramount importance in
understanding the biogeochemistry of a forest ecosystem. The U.S. Forest Service
has monitored and maintained accurate records of precipitation and streamflow for
numerous gauged watersheds of the HBEF since 1956 (Federer et al. 1990;
A. Bailey et al. 2003; http://www.hubbardbrook.org). The HBEF was established in
1955 to study forest hydrology in New England utilizing a paired watershed
approach (see McGuire and Likens 2011).

The Water-Year

Selection of a suitable water-year is a primary consideration for analysis of interan-

nual hydrology. The ideal water-year begins and ends for a period that shows little
interannual change in watershed water storage, often a point in the year of lowest
water storage. One way to identify a water-year is to find that successive 12-month
period that most consistently, year after year, gives the highest correlation between
precipitation and streamflow and negligible changes in storage (i.e., soil water and
snow). In watersheds of the HBEF, streamflow is largely dependent on (1) precipita-
tion, (2) the capacity for water storage in the soil and thereby the amount of water
stored in the soil at any time, and (3) the amount of water stored in the snowpack.
After a large array of linear regressions for successive 12-month periods of pre-
cipitation and streamflow had been examined, the water-year beginning on 1 June
and ending on 31 May was chosen, with correlation coefficients of about 0.99
(see Federer et al. 1990). We repeated Federer’s calculations for the period 1963–
2010 and concluded that a water-year beginning 1 June was still valid based on the
highest regression coefficient (r2 0.95) and a regression slope closest to one.
The high correlation coefficients indicate that water storage in the watershed is con-
sistently at its seasonal low, which can be explained by seasonal water storage

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2_2, 17

© Springer Science+Business Media New York 2013
18 2 Hydrology

dynamics: (1) Evapotranspiration each summer leaves the soil with relatively
depleted water content; (2) autumn rains usually completely replenish water
depleted during the summer; and (3) the addition of between 20 and 30 cm of water
as snowpack, followed by spring soil recharge during melting of the snowpack. The
selection of a water-year starting on 1 June also is advantageous because this cor-
responds to the period when foliage is getting established in the HBEF and allows
for a separation of the water-year into periods essentially coincident with growing
and dormant seasons for the vegetation. The effects of climate change on phenology
and components of the water cycle, such as snowmelt, have the potential to change
the water-year in the future.

Precipitation

Because precipitation serves as a major vehicle of nutrient input into an ecosystem,

its accurate measurement is of prime importance in an evaluation of any biogeo-
chemical cycle. Precipitation amount is measured at the HBEF by a fixed network
of standard U.S. Weather Bureau volumetric precipitation gauges sited in cleared
areas, approximately one for every 13 ha within the areas of the experimental water-
sheds. The Thiessen polygon method (Thiessen 1911) is used to calculate amounts
of areal precipitation for discrete watersheds (see Federer et al. 1990).
On average, 143.4 cm of water falls on each unit area of Watershed 6 (W6) annu-
ally within the HBEF (1963–2009). Of this amount 65 %, or 92.6 cm, becomes
streamflow, and the remainder, about 51 cm, is lost as vapor through evapotranspira-
tion (Table 1, Fig. 3). The evapotranspiration is estimated as the difference between
precipitation and streamflow for the water-year because the geologic base of the
watershed is essentially watertight (see the section on “Deep Seepage” below).
There are on average 170 precipitation days (>trace) per year at HBEF or about
three per week (Table 2). Dividing the average annual precipitation of 143.4 cm by
170 days gives an average intensity of 0.8 cm per precipitation day.
The seasonal effects of temperature on precipitation (rain or snow), as well as the
structure and function of the biomass (e.g., interception of raindrops by the forest
canopy and litter on the forest floor), are important in lowering the erosive potential

Table 1 The average annual hydrologic budget for Watershed 6 of the Hubbard Brook
Experimental Forest
W6, 1963–2009 (period of biogeochemical
study) cm/unit area ± SD Percent of input
Input (precipitation) 143.4 ± 20.0 100
Output (streamflow) 92.6 ± 20.6 64.6
Output (evaporation plus transpiration)a 50.8 ± 5.0 35.4
a
Calculated (Input – Output)
Precipitation 19

Fig. 3 Average hydrologic 1800

Average Water-year Amount (mm/yr)

budget for Watershed 6 of the
Hubbard Brook Experimental 1600 Precipitation ±1 SD
Stream Flow ±1 SD
Forest with standard
1400 Calculated ET ±1 SD
deviations (SD) during
1963–2009 1200

1000

800

600

400

200

Table 2 Annual input of water as snow, rain, or mixed snow and rain for the Hubbard Brook
Experimental Forest during 1974–2010a
Average number Percent of average annual water input
Precipitation type (days/year) (if applied to W6, 143.4 cm)
Snow 42 12 (17.2 cm)
Rain 106 71 (101.8 cm)
Mixed 20 17 (24.4 cm)
a
Data are from calendar-year summations of USFS Daily Precipitation volume (mm) at the Robert
S. Pierce Ecosystem Laboratory (RSPEL) and matched with daily field notes. Precipitation is
recorded about 170 days per year on average or about 47 % of the time. There was no significant
trend in the percent of precipitation volume as rain

of precipitation and to biogeochemical flux and cycling. Precipitation that falls as

snow has a much smaller potential energy of impact, and once a snowpack is estab-
lished, it acts as a “sponge” for any nonfrozen precipitation.
The form of precipitation at HBEF is being altered by climate change. Some
29 % of the annual precipitation amount falls as snow or mixed snow and rain at
HBEF (Table 2). Decreasing proportions of snow and increasing amounts of mixed
snow and rain have been occurring during the winter months since 1974 at the
HBEF (Fig. 4). During the period (1974–2010), there was a 7 % decline (p < 0.01)
in amount of precipitation that fell as snow (42 days/year and 12 % of annual vol-
ume) and an 8 % increase (p < 0.01) in precipitation that fell as a mixture of rain and
snow (~20 days/year and 17 % of annual volume). These changes in type of precipi-
tation track well with other measures of climate change at HBEF (e.g., Hamburg
and Cogbill 1988; Campbell et al. 2007; Huntington et al. 2009; Likens 2011;
Hamburg et al. 2013; Groffman et al. 2012). Climate change in the future will
undoubtedly change various components of the hydrologic budget, including effects
on chemical budgets at the HBEF, and may have wide-ranging ecological and
20 2 Hydrology

100
Percentage of Annual Precipitation Volume

90
80
70
60
50 Percent RAIN
Percent MIXED
40 Percent SNOW

30
20
10
0
1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
Calendar Year

Fig. 4 Percent of total annual precipitation volume as rain (▿), snow (☆), or mixed (●) in
Watershed 6 of the Hubbard Brook Experimental Forest during 1963–2009. Rain trend (ns), snow
decline (p < 0.001), mixed increase (p < 0.004)

biogeochemical consequences. Documenting and predicting these hydrologic

and concurrent biogeochemical changes are difficult but ongoing at the HBEF
(e.g., Campbell et al. 2011; Judd et al. 2011).
The concept of a normal or average water-year is of limited utility because
interannual variation of precipitation is high. Although the long-term average was
143.4 ± 20.0 cm/yearss ( x ± SD) during 47 years of record (1963–2009), in only 2
year has precipitation been within ±10 cm of this value. The mean deviation from
the average annual precipitation is 13.9 % at HBEF (W6) and is less than the
expected mean variability, about 16 %, for similar rainfall areas of the world (Conrad
1941). In the driest water-year on record, 1964-1965 at HBEF, precipitation was
97.5 cm (streamflow, 49.6 cm), whereas in the wettest year, 1973–1974, it was
188.8 cm (streamflow, 140.1 cm). These values represented the extremes of wet and
dry conditions relative to precipitation; they were 34 % less and 32 % greater,
respectively, than the long-term average at the HBEF (W6). This range of 91.3 cm
emphasizes the need for a long-term record to understand the ecosystem’s hydro-
logic flux (Fig. 5). If the biogeochemistry for these ecosystems were based upon
either of these years alone, estimates of meteorologic input and streamwater output
of nutrients would have been quite misleading. Despite the occurrence of both
“extreme” wet and dry years during the period of our study (1963–2009), the range
of annual precipitation varied by less than twofold. This relative “stability” in the
amount of precipitation for the HBEF imparts stability to the chemical budgets as
well because the two are interdependent.
Precipitation 21

Fig. 5 Annual precipitation (P), streamflow (RO), and evapotranspiration (ET ) for Watershed 6 of
the Hubbard Brook Experimental Forest. The regression lines indicate significant trends during
1963–2009. ET = P – RO. Each water-year is from 1 June to 31 May

Based on linear regression, amount of annual precipitation in the HBEF (W6)

has increased significantly during 1963–2009 (p = 0.043), but the variability is large
(r2 = 0.09). Likewise, streamflow has also increased during this period (p = 0.015;
r2 = 0.12) (Fig. 5; see below). Campbell et al. (2011) found essentially the same
result using nonparametric, Mann–Kendall statistics, but Hamburg et al. (2013)
reported nonsignificant trends for a shorter period (i.e., through 2005). Precipitation
amount and streamflow during 1963–2009 in W6 are distributed normally
(Kolmogorov-Smirnov, p > 0.20).
Historically, quantitative measurements of precipitation at the HBEF began in
1958. Nevertheless, records can be extended back in time by comparison with
nearby weather stations. A comparison of monthly records of precipitation from
W6 with monthly data from nearby Bethlehem, NH, where records exist since 1895,
gave an r2 of 0.55 (Fig. 6a). Using this relationship, the precipitation record for W6
at HBEF was extended back to 1895 (Fig. 7). Based on this extrapolation, several
features of the precipitation amount over this 115-year period (1895–2010) become
clear:
(a) On average, the HBEF (W6) has become significantly wetter since 1964
(~64-mm increase between 1964 and 2009 and ~167-mm increase between
1985 and 2009).
22 2 Hydrology

Fig. 6 Correlation between (a) average monthly precipitation for Watershed 6 of the Hubbard
Brook Experimental Forest and precipitation at nearby Bethlehem, NH (NOAA Historical Climate
Network) during 1963–2009 (r2 0.55), (b) correlation between annual streamflow for Watershed 6
and the Pemigewasset River USGS gauging station in nearby Plymouth, NH (r2 0.86)

Fig. 7 Average annual precipitation for Watershed 6 of the Hubbard Brook Experimental Forest
extended to 1895 based on a regression with nearby Bethlehem, NH. Dashed lines represent the
mean ± 1 SD

(b) Year-to-year variability in amount of precipitation has increased since ~1964

(the standard deviation for the annual mean value during 1964–2010 was more
than two times higher than during 1895–1963; 1 standard deviation of the mean
was exceeded in 22 years out of 48—5 below the mean and 17 above the mean).
(c) The wettest and driest years occurred during the period of the HBES
(1963–2010).
Streamflow and Evapotranspiration 23

300

Average Water (mm/mo) ± 1 SD

Precipitation
250 Stream Flow

200

150

100

50

0
Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May
Month

Fig. 8 Average monthly precipitation (open circle) and streamflow (filled circle) for Watershed 6
of the Hubbard Brook Experimental Forest during 1963–2009. The vertical bars are ±1 standard
deviation of the mean

In any particular year, monthly amounts of precipitation may show random

variation, but for the longer term, the monthly pattern becomes quite regular (Fig. 8).
The average amount of monthly precipitation shows very little seasonal variation.
On average, November (13.7 ± 3.9 cm/month) is the wettest month and February
(9.3 ± 4.7 cm/month) is the driest, but no month deviates widely from the overall
monthly average of 11.9 ± 1.2 cm. Nevertheless, long-term data show that both pre-
cipitation and streamflow have significantly increased during the summer period,
June through September (Photograph 5).

Streamflow and Evapotranspiration

Streamflow is measured continuously throughout the year by automatic recorders at

stream-gauging stations, which include either a V-notch weir or a combination of
V-notch and San Dimas flume (see Federer et al. 1990). The stream-gauging sta-
tions are anchored to the bedrock at the base of each watershed. Heating units in the
ponding basin above the V-notch or in the flumes permit accurate measurements of
streamflow during the coldest weather. In contrast to the monthly uniformity of
precipitation, nearly half of the streamflow at the HBEF occurs in spring when the
accumulated snowpack melts (Fig. 8). Some 47 % of the annual streamflow from
W6 occurs during March, April, and May, with more than 23 % in April alone.
During the summer months, streamflow is very low, particularly during July, August,
and September (Fig. 8). During summer, water loss from the ecosystem occurs pri-
marily through evapotranspiration. A secondary, minor peak in streamflow occurs
in November and December when transpiration has virtually ceased after the loss of
24 2 Hydrology
Streamflow and Evapotranspiration 25

the deciduous leaves and with somewhat greater autumn rainfall. At that time,
soil-water content increases and most of the rain moves through the soil and into the
streams (Federer et al. 1990). Winter streamflow is diminished as precipitation
accumulates as a snowpack of approximately 10 % density (water weight) with a
capacity for storage of rain to a maximum of ~50 % density depending on snowpack
conditions, such as temperature. Thus, when the snowpack finally melts, the runoff
is a mixture of both melted snow and stored rain (Photograph 6).
On an annual basis, evapotranspiration in W6 has remained relatively constant
(~51 cm/year) despite major fluctuations and a long-term increase of ~4.5 %
(p ≤ 0.05) in precipitation (Fig. 5). This result is in contrast to annual streamflow,
which is highly correlated with the amount of annual precipitation (Fig. 9). New
precipitation is generally subjected to evapotranspiration first, and once evapotrans-
piration demand is satisfied, the remainder of the precipitation leaves the watershed
as streamflow. In a general way, this appears to be true, but a more detailed analysis
suggests a more complex situation.
Streamflow during the months of June through September is linearly correlated
with amount of precipitation (r2 = 0.75). Calculating evapotranspiration during the
growing season, June through September, shows that evapotranspiration is rela-
tively constant (35–45 cm) until precipitation falls below ~50 cm. Below 50 cm, the
seasonal evapotranspiration can fall as low as 20 cm as trees shut down their
transpiration demand. Calculated evapotranspiration cannot match potential
evapotranspiration when precipitation falls below ~40–50 cm (water limited), and
evaporation is energy limited (sensu Budyko 1974) when precipitation is above
~45 cm (Fig. 10). Interestingly, evapotranspiration seems to peak at a maximum
of ~45 cm during the growing season, and streamflow is zero at precipitation of
~30 cm (Fig. 10). Although these data are approximate, because we do not account

Photograph 5 (a–d) Precipitation collectors. Various types of precipitation collectors used at the
HBEF circa 1960s and 1970s. Collectors are located in clearings within the forest to avoid contami-
nation. No overhead obstruction is allowed above an angle of 45° from the opening of the collector.
Collectors are positioned at least 2 m above the ground. (a) Gauges for quantitatively monitoring
amounts of precipitation (rain and snow). Left: Standard United States Weather Bureau nonrecord-
ing gauge with 20.3-cm orifice. Right: Weight-activated recording gauge. Alter-type windshields
reduce wind velocity around gauge opening to provide a sample equivalent to deposition on
ground. A mixture of ethylene glycol and methanol is added to gauges in winter to convert snow
to liquid state. (b) A polyethylene apparatus used for collecting bulk precipitation for chemical
analysis consisting of a 28-cm diameter funnel, tubing, 2-L reservoir, and vapor barrier. Samples
are retrieved and a clean apparatus installed at 1-week intervals. (c) An automatic sampler for the
collection and separation of wet and dry deposition for chemical analysis. Water on heated sensor
activates the mechanism to move the lid from one bucket to the other at the beginning of wet and
dry periods. (d) Plastic garbage can used to collect snow for chemical analysis. Samples are
retrieved and a clean collector installed at 1-week intervals. (e–h) Precipitation collectors used at
HBEF currently. (e) A PVC pipe shielded bulk precipitation collector in fenced clearing at the
Robert S. Pierce Ecosystem Laboratory, (f) a 21-L polyethylene bulk precipitation collector for use
during winter, (g, h) winter scenes showing weather stations in the Hubbard Brook Experimental
Forest (g–h, photos by D.C. Buso)
26 2 Hydrology

Photograph 6 Stream-gauging weir. Tandem arrangement of modified San Dimas flume and
V-notch weir for monitoring streamflow on Watershed 6. The metal flume on the left measures high
flows; the 90° sharp-edged, V-notch weir on the right measures low flows. The gauging station is
built on bedrock so that all streamflow is channeled first through the flume and then into the weir.
Recording instruments in shelters for both flume and weir monitor streamflow continuously.
Propane burners are used during the winter to prevent freezing at the notch. Large particulate mat-
ter transported by the stream during high flows is caught either in the screen as the flow cascades
from the flume over the trough, bypassing the V-notch, or in the ponding basin behind the V-notch.
This particulate matter is removed annually and weighed. Gauging stations similar to this one or
utilizing a V-notch weir only have been established on nine experimental watersheds in the HBEF
(see Fig. 2)

for changes in stored soil water, they suggest a certain reasonable logic about the
hydrologic behavior of these watershed-ecosystems.
Obviously, the vast majority of the transpiration for this northern hardwood for-
est ecosystem occurs during the growing season from about 15 May to 15 September.
The seasonal importance of transpiration was demonstrated experimentally by a
deforestation experiment on Watershed 2 of the HBEF (see Hornbeck et al. 1970;
Likens et al. 1970; Bormann and Likens 1979). Transpiration may account for about
60–70 % of the evapotranspirational water loss on an annual basis at HBEF
(C.A. Federer, personal communication), but transpiration is difficult to quantify at
large scales like a watershed. There is a major need for additional quantitative mea-
surements to partition evapotranspiration into evaporation from the forest canopy
and the ground and transpiration. The relationship between annual precipitation
input and streamflow continues to be quite robust in the long-term record (r2 = 0.94;
p < 0.001) even though there has been a statistically significant increase in both
Streamflow and Evapotranspiration 27

1600

Annual Streamflow or ET (mm/yr)

1400

1200

1000

800
1964
600

400

200
600 800 1000 1200 1400 1600 1800 2000
Annual Precipitation (mm/yr)

Fig. 9 Relationship between annual precipitation and annual streamflow (filled circle) and annual
evapotranspiration (open circle) for Watershed 6 of the Hubbard Brook Experimental Forest dur-
ing 1963–2009. Calculated evapotranspiration is determined as the difference between the amount
of precipitation and streamflow. The relation between streamflow and precipitation has a slope of
0.99, r2 0.94, p < 0.001

Fig. 10 Relationship
between precipitation and
streamflow (filled circle) and
evapotranspiration (open
circle) during June through
September for Watershed 6 of
the Hubbard Brook
Experimental Forest during
1963–2009

precipitation (p = 0.043) and streamflow (p = 0.015) during the period 1963–2009

(Fig. 9). The relationship of precipitation to annual evapotranspiration appears to be
changing during the wetter years of 2003 through 2009, with calculated evapotrans-
piration (336 mm) in 2009–2010 being the lowest on record (Fig. 5).
The relationship between evapotranspiration and precipitation underscores the
powerful regulating role that the living components of the ecosystem play in the
hydrologic cycle. During the active growing season, evapotranspiration (mostly
28 2 Hydrology

Table 3 Annual hydrologic budget for four watershed-ecosystems of the Hubbard Brook
Experimental Forest during 1979–1980a
watershed-
ecosystem Area (ha) Precipitation (cm) Streamflow (cm) Evapotranspirationb (cm)
W1 11.8 113.7 71.4 42.4
W3 42.4 113.9 69.3 44.6
W5 21.9 115.5 70.2 45.3
W6 13.2 119.3 70.8 48.5
Mean 115.63 70.41 45.21
Standard error 1.30 0.45 1.27
of the mean
% of the total 100 61 39
Data prior to watershed manipulations applied to W1 (1999) and W5 (1983–1984)
a
Modified from Likens et al. (1985)
b
Precipitation minus streamflow; such estimates of annual evapotranspiration do not consider dif-
ferences in soil–water storage from year to year. Estimates of such storage differences between
years, based on the BROOK model (Federer and Lash 1978), rarely exceed 4 cm and commonly
are less than 1 cm. Neglecting changes in soil-water storage could lead to errors of up to 7 % in
estimating annual evapotranspiration

transpiration) is an important factor in regulating the amount of streamflow.

Evapotranspiration then not only acts as a regulator for the hydrologic cycle but also
contributes to tighter nutrient cycles by (1) diminishing streamflow that can carry
nutrients out of the ecosystem, (2) by effectively concentrating the inputs from pre-
cipitation, and (3) increasing nutrient uptake as a function of transpiration.
Seasonal streamflow varies by orders of magnitude, whereas interannual stream-
flow varies by no more than twofold and is usually less. Because the chemical flux is
so intimately tied to the hydrologic cycle, seasonal hydrologic events have correspond-
ing effects on the seasonal chemical budgets (see Chap. 4). In contrast, the annual
hydrologic budget, with its relatively small variability, imparts long-term stability to
the associated annual chemical budgets, as is shown in subsequent chapters.
Some differences in hydrology exist between the experimental watersheds
(W1–W6) largely related to east-west orientation. On average, about 78 mm/year or
5.4 % more precipitation falls on the westernmost experimental watershed, W6, in
comparison with the most easterly south-facing watershed, W3 (see Fig. 2; Federer
et al. 1990). Overall, however, differences between these south-facing watersheds of
the HBEF are small (Table 3), which further demonstrates the utility of the paired
small watershed approach at the HBEF.
In general, erosion rates are low and relative soil stability is high at the HBEF
since glacial retreat from this area (M. Davis et al. 1985; Bormann and Likens 1979;
Kaste et al. 2007). A factor of prime importance in regulating erosion and stream-
water chemistry is that the HBEF soils have a high infiltration capacity. Because the
surface soil is very porous, virtually all incident precipitation, rain or snowmelt,
percolates through the soil even during the largest storms. Overland flow, where
infiltration rate is exceeded, is not manifested at HBEF except in a few saturated
boggy areas and in channels (Pierce 1967) or temporarily during extreme rainfall
Deep Seepage 29

events or extreme soil frost events. The downhill flow of water toward stream beds
is by lateral subsurface interflow (Detty and McGuire 2010b). The soil usually
remains unfrozen during the coldest months because of the thick humus layer and a
deep snow cover (Hart et al. 1962). There is long-term potential, however, for
increasing soil frost because of changing climate (e.g., Groffman et al. 2001, 2006;
Judd et al. 2011). The surface topography is rough (pit and mound), which also
facilitates infiltration rather than overland flow. Because the geologic substrate is
essentially watertight, losses of water by deep seepage are minimal, and areas of
non-vegetated, exposed bedrock ledge are minimal, almost all water has intimate
contact with the biologic and inorganic components of the soil and subsoil as it
moves downhill toward the stream channel. Frozen conditions in the stream chan-
nels add still further to the control of particulate matter losses in stream water as
erodibility in winter months normally is less than that during the growing season
(Bormann et al. 1974). There have been two incidents of midwinter channel erosion,
in 1980 and 2011, when rain on snow caused “slush” avalanches in several streams.
In these exceptional situations, these streams had a thick ice layer that permitted a
snow/ice slurry to build up momentum and scour the channel.
The actual path that water follows as it flows through the soil and/or downslope
is of great importance biogeochemically. Much effort has been expended at HBEF
to quantify these relationships. For example, Stresky (1991) found that macropores,
or “pipes,” were common in the soils of the experimental watershed-ecosystems at
HBEF. He found that the average macropore was 2.5 cm in diameter (0.5–7.5 cm)
and was 16 cm deep in the soil (2–130 cm) but comprised only about 0.2 % of the
area of soil profiles examined. Patterns of hillslope hydrology have been studied at
HBEF by Shattuck (1991), Mau (1993), and Detty and McGuire (2010a, b), which
have provided new insight about streamflow and solute generation. Lawrence and
Driscoll (1990) suggested that a combination of factors, including spatial variability
in soil type and depth, modification of soil profiles adjacent to stream channels
because of lateral flow patterns, macropore flow, and vegetation type, determines
the spatial and temporal patterns of streamwater chemistry observed at HBEF.
The work of Detty and McGuire (2010a, b) demonstrates the presence of shallow
ground water at HBEF, and the importance of soil moisture and groundwater thresh-
olds that must be met before hillslopes can contribute substantially to streamflow.
Recent work by Zimmer et al. (2012) has shown that lateral flow paths can have a
major affect on soil development and streamwater chemistry, particularly in the
headwaters.

Deep Seepage

One of the key features of the HBEF that enables a close accounting of water bud-
gets is the relatively watertight bedrock. Precipitation entering a watershed must
leave by three avenues: (1) evaporation from leaf and soil surfaces, or both, to the
atmosphere; (2) deep seepage, including lateral transfer through soil and bedrock
30 2 Hydrology

across watershed boundaries and vertical movement through deep geologic strata;
and (3) streamflow. Direct measurement and quantification of evapotranspiration
via eddy covariance methodology in steep terrain like HBEF and deep seepage and
lateral transfer loss of water are extremely difficult to accomplish. Streamflow,
however, can be measured readily and with a fair degree of accuracy in many
watersheds.
When there are strong indications that a given watershed has relatively imperme-
able bedrock such that deep seepage is minimal and streamflow can be measured
quantitatively, then the other major component of the water budget, evapotranspira-
tion, can be calculated as the difference between streamflow and precipitation.
Reasonable determinations of these components of the water budget for an entire
water-year are feasible provided the water-year begins when storage of soil water is
consistent each year (see above).
There is substantial evidence suggesting that the south-facing experimental water-
sheds at the HBEF have reasonable watertight bedrock and therefore have negligible
water loss through deep seepage. Findings from a report of a geologic survey con-
ducted at the HBEF by Bradley and Cushman (1956) point out that piezometric sur-
face divides tend to coincide with the shape of the topography. Unconsolidated
deposits lie entirely within valleys, which are rimmed by walls of bedrock.
The Littleton formation (an intensely metamorphosed rock, currently mapped as
the Rangeley Formation) is crossed by several sets of joints although openings are
not large. These may extend one to a few meters below the surface, but their size
and number decrease with depth. Only relatively small quantities of water occur in
and move through these cracks. Based on these findings, Bradley and Cushman
concluded that it is improbable that measurable quantities of ground water move
from one watershed to another or that measurable quantities of ground water escape
from the watershed through major fault zones or solution openings.
Further evidence for minimal deep seepage comes from comparison of the
experimental watersheds. Assuming that two small adjacent watersheds have rea-
sonably similar precipitation, vegetative cover, soil, and evapotranspiration, then
streamflow on a unit area basis should also be similar. If there were considerable
deep seepage, however, then streamflow from the two watersheds would likely be
widely dissimilar. A comparison of streamflow among Watersheds 1, 3, 5, and 6
(11.8, 42.4, 21.9, and 13.2 ha, respectively) for 12 water-years was made using an
analysis of variance. The forest on Watersheds 2 and 4 was cut in 1965–1966 and
1970–1974, respectively, and therefore these two watersheds cannot be used in the
long-term simultaneous comparisons of streamflow with Watersheds 1, 3, 5, and 6
(Watershed 5 was whole-tree harvested in 1983–1985). In spite of wide variability
in yearly streamflow during the period, e.g., 496 mm to 1,401 mm, there was no
significant difference in streamflow among Watersheds 1, 3, 5, and 6. Because
streamflow was virtually the same for each watershed, we conclude that the chances
of deep seepage are negligible.
As stated previously, if water loss through deep seepage were negligible, then
evapotranspiration should approximate precipitation minus streamflow (runoff) for
the common water balance equation: ET = P − RO ± ΔS, where ET is evapotranspira-
tion, P is precipitation, RO is streamflow, and ΔS is the change in storage of soil
Deep Seepage 31

water. Over the course of an entire year, we assume that the change in storage of soil
water is zero as discussed above. Hart (1966) compared P − RO values for Watersheds
1 and 3 (HBEF) for 5 years with calculated values of potential evapotranspiration
based on the method by Thornthwaite (1948). Potential evapotranspiration, PE, is
defined as the evaporation that would occur from a large fully wetted surface, that
is, one with an unlimited water supply (Sellers 1965). The average water-year values
for Watersheds 1 and 3 for a 5-year period (1958–1963) were 56.2 cm for PE and
53.2 cm for P − RO. There are several empirical methods for estimating PE
(Thornthwaite 1948; Penman 1956; Hamon 1961), and PE calculations by each of
these vary slightly; in general, however, such values of PE for HBEF are reasonably
close to 56 cm. Actual evapotranspiration should be slightly less than PE except in
rare cases. For practical purposes, the values of P − RO at the HBEF can be consid-
ered as close approximations of actual evapotranspiration. The remarkable close-
ness of the values of P − RO and PE indicates that water loss from deep seepage is
not a significant factor in the water balance at the HBEF (see Fig. 9).
Chloride has not accumulated in sizeable quantities within these watershed-
ecosystems in aggrading living or dead biomass since 1982 (~32 mol Cl/ha-year;
Lovett et al. 2005). Moreover, chloride-bearing rocks in the HBEF are minor and
release of chloride via weathering is thought to be small (~2.5 mol Cl/ha-year;
Lovett et al. 2005; Chap. 5). Therefore, if losses by deep seepage were negligible,
then inputs in wet and dry deposition should be observed quantitatively in the
streamwater outputs from the ecosystem. In fact, annual mass balances, particu-
larly since 1982 show small net ecosystem losses, and analyses do not suggest
significant “hidden” losses (see Chap. 4 and 6).
Finally, if there were significant inputs or outputs of groundwater, then stream-
water chemistry should change during low-flow periods. Such changes are not
observed in the long-term record (see Chap. 3). Also, spatially dispersed chemical
hot spots associated with groundwater seeps have been observed and measured for
a long time at the HBEF (e.g., Dominski 1971; Mulder 1980; Likens and Buso
2006). Recently, Zimmer et al. (2012) looked more closely at ephemeral and inter-
mittent channels, and groundwater seeps in headwaters of W3 and suggested that
the high levels of solutes found there might indicate water moving through deep
glacial till or fractured bedrock. Pockets of glacial till can be relatively thick in these
watersheds and where compacted has lower permeability to water. Undoubtedly
there is some deep seepage (in and out) in these watersheds, but the robust trend
toward extremely dilute conditions in stream water observed currently (Chap. 3;
Likens and Buso 2012) would be very difficult to explain at the watershed scale if
deep seepage were sizeable in W6.
Thus, based on the long-term record, we have no basis to reject our assumption
about the relatively watertight nature of the experimental watershed-ecosystems of
the HBEF (Federer et al. 1990). This feature of these watershed-ecosystems is of
great importance for calculating quantitative mass balances of water and chemicals,
as will be done in subsequent chapters.
Certainly some small amount of deep seepage, in or out, must occur but is negli-
gible relative to the errors involved (Verry 2003) and to the watershed mass balances
obtained for W6 of the HBEF.
32 2 Hydrology

Quantifying uncertainty of hydrologic measurements at HBEF is currently being

reevaluated by computer simulations and sensitivity analyses of rain gauge net-
works and stream calibration data (e.g., Yanai et al. 2012). Here, the uncertainty for
precipitation volume and streamwater discharge is estimated at ±5 % on an annual
basis (Hornbeck et al. 1970; Hornbeck 1973; Winter 1981; Bailey et al. 2003a, b).
Any deep seepage gains or losses are likely to be much smaller.
Detailed and extensive studies of groundwater flow have been done in the water-
shed for Mirror Lake (e.g., Winter 1985; Paillet et al. 1987; Winter et al. 1989;
Winter and Likens 2009). This large watershed, at lower elevation in the Hubbard
Brook Valley and with deeper till and fracture-flow systems (Powers et al. 1999),
has different hydrologic characteristics than those of the experimental watersheds.
Nevertheless, groundwater inputs to Mirror Lake comprised only ~16 % of the
annual water inputs to the lake (Rosenberry and Winter 2009).

Representativeness of the Hydrology of the Hubbard Brook

Experimental Forest for Northern New England

One of the main concerns about small experimental watersheds is whether the
research results obtained can be interpreted and extended to other, larger water-
sheds. Climate, soil, geologic formation, topography, vegetative cover, and past
land use all vary from place to place, yet they play important roles in the behavior
of water fluxes through a watershed. In light of the influences of these watershed
features on water flow, the question arises: How representative is the hydrology of
the HBEF to other areas of northern New England? Fortunately, there is one water-
shed characteristic that integrates the influences of most landscape features into a
single measurable characteristic—streamflow. Considerable insight into the inner
workings of a watershed-ecosystem can be obtained by continuously monitoring the
flow from that watershed over long periods of time.
In a comprehensive study of streamflow characteristics of watersheds in northern
New England, Sopper and Lull (1965, 1970) addressed the question of representa-
tiveness of small experimental watersheds. They compared streamflow from eight
watersheds in Maine, New Hampshire, Vermont, and Massachusetts with Watershed
1 (W1) at the HBEF.
Data adapted from Sopper and Lull are presented in Table 4. The eight water-
sheds were chosen on the basis of being less than 259 km2 (100 mi2) in size, being
90 % or greater forest covered, having streams that were undisturbed by diversion
or regulation, and having relatively complete records of precipitation and stream-
flow—the latter having simultaneous periods of records from 1 December 1956 to
30 November 1962. Annual and seasonal streamflow at the HBEF were very close
to the average of the eight regional watersheds and fell within the maximum and
minimum values for all periods.
Direct streamflow (that portion of the annual streamflow that rapidly leaves the
watershed during and after storms, as opposed to prolonged base flow) for W1 of
Table 4 Hydrologic characteristics for selected watersheds in northern New Englanda
Average annual and seasonal water yield (cm) Direct
Forest Growing Dormant runoff, %
Stream and location Area (km2) cover (%) Precipitation (cm) Annual Spring Summer Fall Winter season season of annual
Austin Stream, ME 236 95 120.7 72.3 36.0 9.4 14.8 12.1 28.5 43.8 16.4
Swift River, ME 249 97 126.0 72.0 37.3 9.2 15.9 9.6 33.1 38.9 22.5
Ashuelot River, NH 184 94 113.2 58.4 32.7 3.6 10.5 11.6 15.7 42.7 11.0
N. Br. Contoocook R., NH 142 95 129.8 66.8 34.0 3.9 13.8 15.1 18.7 48.1 8.5
Otter Brook, NH 109 93 114.2 57.7 31.5 3.9 10.5 11.8 15.6 42.1 8.4
Smith River, NH 223 95 119.0 56.3 32.1 4.6 9.1 10.5 16.6 39.7 11.8
Saxtons River, VT 186 92 118.1 52.2 30.5 3.4 8.2 10.1 13.2 39.0 14.4
N. Br. Hoosic River, MA 104 90 155.5 74.8 38.9 6.3 13.2 16.4 20.6 54.2 23.8
Regional average 179 94 124.5 63.8 34.1 5.5 12.0 12.2 20.2 43.6 14.6
HBEF, W1 0.12 100 124.6 70.4 36.9 3.5 15.3 14.7 16.7 53.7 21.3
a
After Sopper and Lull (1965, 1970)
Representativeness of the Hydrology of the Hubbard Brook Experimental Forest…
33
34 2 Hydrology

the HBEF was high, as expected, but was less than direct streamflow from two other
watersheds, one of which was much larger in size, 249 km2.
Additional comparisons made by Sopper and Lull (Table 4) indicated close simi-
larities in streamflow characteristics between experimental and regional watersheds,
such as (1) number and amount of mean daily discharge above a given volume of
flow and (2) flow duration (percent of time flow was equaled or exceeded).
In general, streamflow in small watersheds tends to have sharper peaks, higher
storm levels, and shorter periods of sustained flow than it does in large water-
sheds. Such streamflow characteristics in small watersheds may be accentuated in
steep, mountainous terrain with shallow soils. Watersheds at the HBEF have these
characteristics, but it is remarkable and reassuring that streamflow at HBEF, e.g.,
in a watershed (W1) of 0.12 km2, closely approximates flow from watersheds that
are more than three orders of magnitude greater in size! Further, the HBEF water-
sheds show behavior that is similar to the river they feed. Comparing the flow in
the Pemigewasset River into which Hubbard Brook drains, with flow from W6 on
a unit area basis (Fig. 6b), showed a strong correlation during 1963–2006
(p ≤ 0.001, r2 = 0.86). The much larger Pemigewasset River drains an area of
1,611 km2 at the U.S. Geological Survey gauging station in Plymouth, NH (Slack
et al. 1997). This strong relationship provides an excellent proxy for long-term
streamflow from HBEF, as the Bethlehem precipitation station provides a proxy
for precipitation prior to 1958, because the Pemigewasset River gauge has been
operating since 1886.
The close relationships between streamwater quantity for streams in the HBEF
and larger streams throughout northern New England add considerable support to
the contention that HBEF streams are representative in many ways and that research
results from the HBEF can be useful in extension and application to other areas for
management purposes and for increasing ecological understanding at larger spatial
scales.
Chapter 3
Chemistry

Samples of precipitation and stream water are obtained from the experimental
watersheds for chemical analysis. Rain and snow are sampled with continuously
open plastic collectors, i.e., bulk precipitation samples (Likens et al. 1967; Buso
et al. 2000) collected at weekly intervals. Samples of stream water are collected by
hand each week and during high-flow events approximately 10 m upstream from
each weir in clean polyethylene bottles. This location is necessary because water
collected from the ponding basin above the weir could be contaminated by calcium
and carbonate from the cement in the weir itself. The concentrations of dissolved
chemicals characterizing a period of time are reported as volume-weighted aver-
ages. These averages are computed by summing the amount of chemical from indi-
vidual samples of precipitation or stream water during the period and then dividing
this value by the total amount of water during the period. Nutrient flux (see Chap. 4)
across ecosystem boundaries is determined by two methods. Precipitation inputs are
calculated by multiplying the measured concentration of dissolved chemicals in the
accumulated composite sample of precipitation by the daily amounts of precipita-
tion during the interval; the daily values (g/ha-day) are summed to provide mass
inputs by week, month, or year (e.g., g/ha-month). Streamwater outputs are calcu-
lated by multiplying the measured concentrations of dissolved chemicals in stream-
water samples taken at the beginning and at the end of the interval by the streamflow
on those dates. For dates between samples, the average of the beginning and ending
concentrations is multiplied by the daily flows. We have tested other approaches for
determining streamwater export, but the approach described here gives the most
accurate and reproducible results for the HBEF. Daily values (g/ha-day) are summed
on a weekly, monthly, or yearly basis to derive mass streamwater outputs (e.g., g/
ha-month).
It has been shown that sampling according to a standard time series may seri-
ously underestimate or overestimate a highly variable parameter (e.g., Claridge
1970); however, most of the dissolved chemicals in stream water at the HBEF fluc-
tuate relatively little in concentration so that serious errors are not produced by
weekly sampling in most cases (Fig. 11; Likens et al. 1967; Johnson et al. 1969;

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2_3, 35

© Springer Science+Business Media New York 2013
36 3 Chemistry

2.0
Calcium Stream Water
1.5 Precipitation
mg/L

1.0

0.5

0.0
12
Sulfate
mg/L

0
16
Nitrate
12
mg/L

8 Dissolved Silica
6
mg/L

0
1996 1998 2000 2002 2004 2006 2008
Sample Date

Fig. 11 Measured weekly, measured concentrations in bulk precipitation (gray) and stream water
(black) for Watershed 6 of the Hubbard Brook Experimental Forest during 1996–2009

Buso et al. 2000). Concentrations of dissolved organic carbon, nitrate, hydrogen

ion, and dissolved silica are more variable with instantaneous streamflow, but still
show only several-fold change in concentration over >5 orders of magnitude change
in instantaneous streamflow (Fig. 12).
As standard practice, we do not include data from bulk precipitation samples that
contain fresh leaves, bird feces, bud scales, insects, or any other significant foreign
particles. Such contamination can represent a serious problem, because a variety of
substances can leach from these materials as the water comes in contact with them
in the collector. Our intent is to have enough collectors in position to provide at least
one “clean” sample per interval. This procedure normally works very well at our site.
3 Chemistry 37

From June 1963 to May 2010, 4 % of samples collected were rejected because of
obvious contamination and 9 % failed QA/QC protocols. Considering the problems
of contamination and biogeochemical transformations that can occur in the reser-
voir of a bulk precipitation collector, experience has shown that sampling intervals
of not longer than a week are highly desirable, if not necessary, to obtain quantita-
tive data on atmospheric deposition.
Details concerning the routine methods used in collecting samples of precipita-
tion and stream water and analytical and data-management procedures have been
given by Bormann and Likens (1967), Likens et al. (1967), Fisher et al. (1968),
Bormann et al. (1969), and Buso et al. (2000).

a Calcium Magnesium Potassium

3.0 1.0 1.0
Concentration (mg/L)

Concentration (mg/L)

Concentration (mg/L)
2.5
0.8 0.8
2.0
0.6 0.6
1.5
0.4 0.4
1.0

0.5 0.2 0.2

0.0 0.0
-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3
0.0 -3 -2 -1 0 1 2 3
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec)

Sodium Ammonium Total Aluminum

3.0 0.5 1.0
Concentration (mg/L)

Concentration (mg/L)

Concentration (mg/L)

2.5 0.4 0.8

2.0
0.3 0.6
1.5
0.2 0.4
1.0

0.5 0.1 0.2

0.0 0.0 0.0

-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec) Instantaneous Flow Log 10(L/sec)

Hydrogen Ion Sum Base Cations Sum all Cations

100 300 300
Concentration (µEq/L)

Concentration (µEq/L)

Concentration (µEq/L)

80 250 250

200 200
60
150 150
40
100 100
20
50 50
0 0 0
10-3 10-2 10-1 100 101 102 103 10-3 10-2 10-1 100 101 102 103 10-3 10-2 10-1 100 101 102 103
Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec)

Fig. 12 Relation between (a) cation concentration in stream water and instantaneous streamflow
during 1963–2010 and (b) anion concentrations in stream water and instantaneous streamflow dur-
ing 1964–2010 for Watershed 6 of the Hubbard Brook Experimental Forest
38 3 Chemistry

b Sulfate Nitrate Chloride

10 10 1.0
Concentration (mg/L)

Concentration (mg/L)

Concentration (mg/L)
8 8 0.8

6 6 0.6

4 4 0.4

2 2 0.2

0 0 0.0
-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec)

Phosphate Sum (SO4+ NO3) Sum all Anions

0.10 300 300
Concentration (µEq/L)

Concentration (µEq/L)
Concentration (mg/L)

0.08 250 250

200 200
0.06
150 150
0.04
100 100
0.02 50 50

0.00 0 0
-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec)

Dissolved Organic Carbon Dissolved Silica

10 12
Concentration (mg/L)
Concentration (mg/L)

10
8
8
6
6
4
4
2 2

0 0
-3 -2 -1 0 1 2 3
10
-3
10
-2
10
-1
10
0
10
1
10
2
10
3 10 10 10 10 10 10 10

Instantaneous Flow Log10(L/sec) Instantaneous Flow Log10(L/sec)

Fig. 12 (continued)

Precipitation Chemistry

Precipitation chemistry has been monitored at the HBEF since 1963. To our knowl-
edge this unbroken series is the longest comprehensive record of precipitation
chemistry in the world.

Annual Volume-Weighted Concentrations

The chemical concentration of precipitation has been dominated by hydrogen ions

since 1963, constituting 67 and 48 % of the total cationic strength in 1965–1966 and
2009–2010, respectively. The remaining cationic charge is shared among
Precipitation Chemistry 39

Fig. 13 Percent of total ionic charge in bulk precipitation for Watershed 6 of the Hubbard Brook
Experimental Forest during 1963–2009. The values for 1955–1956 were extrapolated from iso-
pleth maps produced by Junge (1958) and Junge and Werby (1958) (modified and extended from
Likens and Bormann 1974b)

ammonium, calcium, sodium, magnesium, potassium, and aluminum, in that order

(Fig. 13; Tables 5 and 6). Aluminum is found only in trace amounts. Bulk precipita-
tion has become much more dilute since 1963, but the relative contribution of the
various cations has remained the same (Likens and Buso 2012).
Although decreasing with time, sulfate and hydrogen ion are the most prevalent
inorganic ions in bulk precipitation at the HBEF on an equivalent basis (Fig. 13;
Tables 5 and 6). The relative importance of nitrate in precipitation on an equivalence
basis has been generally increasing since 1964 (Fig. 13). Essentially, then, the inci-
dent precipitation at HBEF may be characterized as a solution of sulfuric and nitric
acids at a pH < 4.6. Small errors (~0.05 pH unit) in the determination of pH, which
has been used to estimate the concentration of hydrogen ion, are sufficient to explain
the small discrepancy in the cation–anion balance of bulk precipitation (Tables 5
and 6), particularly in the early years of the study. Considering that these long-term
averages include the various sampling and analytical errors over a 49-year span, the
agreement in cation–anion equivalents is quite good.
Surprising concentrations of dissolved organic carbon are found in rain and snow
at HBEF (Tables 5 and 6). Precipitation falling on the watersheds of the HBEF con-
tains an average total organic carbon concentration of about 1 mg/L of which 84 %
is dissolved (Likens et al. 1983). Only a small fraction of this organic matter is dis-
sociated organic acids (Likens et al. 1976, 1983; Galloway et al. 1976) and does not
Table 5 Arithmetic means of volume-weighted average annual concentrations for Watershed 6 of the Hubbard Brook Experimental Forest during
1963–2009a
A Bulk precipitation B Stream water
Mean n (obs) Mean n (obs) C of
Substance (µEq/L) ±SE (years) C of E (%) Substance (µEq/L) ±SE (years) E (%)
Calcium 4.8 0.38 47 7.9 Calcium 48.4 2.15 47 4.4
Magnesium 2.3 0.21 47 9.1 Magnesium 21.5 0.70 47 3.3
Potassium 1.3 0.11 47 8.5 Potassium 5.2 0.15 47 2.9
Sodium 4.2 0.22 47 5.3 Sodium 31.5 0.41 47 1.3
Aluminumb – – – – Aluminum 33.7 1.50 41 4.5
Ammonium 10.6 0.38 46 3.6 Ammonium 1.1 0.02 46 1.5
Hydrogen Ion 51.5 2.56 46 5.0 Hydrogen Ion 11.7 0.40 47 3.4
pH 4.29 pH 4.93
Sulfate 42.3 1.98 46 4.7 Sulfate 102.1 2.83 46 2.8
Nitrate 22.8 0.77 46 3.4 Nitrate 14.4 1.94 46 13.5
Chloride 6.8 0.44 46 6.5 Chloride 12.8 0.39 46 3.0
Phosphate 0.4 0.04 38 11.5 Phosphate 0.1 0.01 38 12.0
Bicarbonatec – – – – ANCd −10.6 1.7 46 15.7
Dissolved organic carbon – – – – Dissolved organic carbon 12.4 3.2 15 25.8
Dissolved silica (µmol/L) – – – – Dissolved silica (µmol/L) 63.9 0.66 46 1.0
Total cations 74.7 Total cations 153.0
Total anions 72.2 Total anions 141.8
Organic carbon has a charge of µEq/L per mg C in stream water Aluminum assumed to have +3 valence state
a
Means in µEq/L (or µmol/L) ± 1 Standard Error of the Mean and the percent Coefficient of Error of the Mean (±SE/Mean)
b
Only trace aluminum or dissolved silicon in clean precipitation
c
No Bicarbonate at pH < 5 (–ANC)
d
ANC calculated from difference of sum (CB) and sum (AA)
Precipitation Chemistry 41

Table 6 Volume-weighted mean concentrations of dissolved substances in bulk precipitation for

Watershed 6 at the Hubbard Brook Experimental Forest
Bulk Decade Decade
precipitation 1964–1973 Bulk precipitation 2000–2009
Conc Conc Conc Conc
Substance (mg/L) (µEq/L) Substance (mg/L) (µEq/L)
Water (mm/ 1,388 Water (mm/ 1,488
ha-year) ha-year)
Ca 0.15 7.5 Ca 0.07 3.5
Mg 0.04 3.5 Mg 0.02 1.5
K 0.06 1.5 K 0.04 1.0
Na 0.12 5.1 Na 0.08 3.5
NH4 0.22 12.1 NH4 0.16 8.9
Ala 0 0 Ala 0 0
+
H 0.073 73.1 H+ 0.028 27.6
SO4 2.85 59.3 SO4 1.19 24.8
NO3 1.47 23.8 NO3 1.07 17.2
Cl 0.38 10.7 Cl 0.15 4.4
PO4b 0.009 0.3 PO4 0.003 0.1
ANC 0.0 ANC 0.0
(bicarbonate)c (bicarbonate)c
Cations 102.7 Cations 46.0
Anions 94.1 Anions 46.4
DOCd 1 91.6 =µmol/L DOC 0.8 68.4 =µmol/L
Silicona 0 0 =µmol/L Silicona 0 0 =µmol/L
pH 4.14 pH 4.56
a
Al and Si found in trace amounts in clean precipitation
b
PO4 based on 1971–1973
c
Bicarbonate = 0 at pH < 5
d
DOC based on 1976–1977 (Likens et al. 1983)

contribute significantly to the ionic balance. The composition of the dissolved

organic carbon in precipitation is undoubtedly a complex mixture of many organic
species, most of which are present in trace amounts. During 1976–1977 particulate
plus dissolved macromolecular (>1,000 MW) organics accounted for 51 %, and
carboxylic acids represented 14 %, aldehydes 11 %, carbohydrates 8 %, and tannin/
lignin 8 % of the total organic carbon in precipitation (Likens et al. 1983). Dissolved
organic carbon in precipitation represents an additional energy source for the eco-
system and thus may be important ecologically and biogeochemically.

Origin of Ions in Bulk Precipitation

Cations and anions in precipitation originate from a variety of sources, including

oceanic spray, terrestrial dust, gaseous and particulate pollutants, and volcanic
emissions. How ionic materials find their way into a precipitation collector is an
42 3 Chemistry

important enigma. The solubility of atmospheric gases, such as CO2, from either
natural or anthropogenic origin, in atmospheric water is obvious, but the origin of
nonvolatile constituents, such as sodium, magnesium, calcium, and potassium in
precipitation is not so obvious. The dispersion of sea salts, such as NaCl and MgSO4,
in atmospheric aerosols is a well-known process (Junge 1963; MacIntyre 1974).
Nevertheless, only a part of the calcium and potassium in precipitation can be rea-
sonably attributed to marine sources, and the “excess” is customarily assigned to
unknown “continental sources” (Junge 1963; Granat 1972; Cogbill and Likens
1974). At the HBEF the molar ratio for Cl:Na in bulk precipitation averaged ~2.3
prior to 1975 (peaking at 3.5 in 1976) or roughly double that anticipated from trans-
ported sea salt (1.2), suggesting transport of chloride to the HBEF from distant
anthropogenic sources (Gorham 1958; Junge 1963; Lovett et al. 2005). Since ~1980
the Cl:Na ratio has approximated or been slightly higher than the sea-salt ratio (cf.
Table 6; Junge 1963; Juang and Johnson 1967; Lovett et al. 2005). In contrast, the
ionic Ca:Mg ratio for these periods is representative of continental precipitation
(Eriksson 1952; Eaton et al. 1973).
One explanation for the excess cations in precipitation is that soil particles and
industrially generated aerosols in the atmosphere may be chemically altered to pro-
duce cations. The ubiquitous hydrogen ions in precipitation may react with these
entrained particles and release cations into solution (see subsequent sections).
Therefore, some of the cations found in solution in a precipitation collector may
have been derived from an “artificial weathering” reaction within the continuously
open (bulk) collector such as used at HBEF. Such “weathering reactions” that occur
within the precipitation collection reservoirs would not change the total amount of
cations input to the ecosystem but might alter their chemical form or underestimate
the input of hydrogen ion in precipitation.
It is important relative to quantitative meteorologic inputs to determine whether
the dissolved and particulate materials in a bulk precipitation sample have origi-
nated outside the ecosystem’s boundaries or whether they have merely been circu-
lating within the boundaries. Because the HBEF watersheds and the surrounding
region (1) are almost entirely forested or have other vegetational ground cover, (2)
have a long period of snow cover, and (3) have generally humid conditions during
other seasons, we conclude that relatively little dust arises from within the ecosys-
tem and that our bulk precipitation collectors measure dissolved and particulate
matter that has largely originated outside the ecosystem’s boundaries. Consistent
and rigorous application of our QA/QC protocols is an important constraint on this
assumption.

Dry Deposition

Early in our studies, we were aware of the dry deposition of particles and gases from
the atmosphere, but had little quantitative information relative to such fluxes for the
ecosystems at HBEF (e.g., Eaton et al. 1976). Because dry deposition represents a
potentially important input of nutrients and pollutants to forest ecosystems, much
Precipitation Chemistry 43

effort has been expended to quantify this flux at HBEF. Bulk precipitation collectors
may be very inefficient in collecting dry deposition, particularly aerosols smaller
than 1 µm. Estimates, based on measurements of aerosol concentrations and a rea-
sonable deposition velocity, suggest that the dry deposition of calcium, magnesium,
sodium, and potassium, as measured in bulk precipitation, probably contributes a
relatively small fraction (<~20 %) of total deposition inputs at HBEF (e.g., Likens
et al. 1998). This is not necessarily the case for nitrogen and sulfur. Small aerosol
particles and gases of nitrogen and sulfur may be generated from biogenic activity
and from the combustion of fossil fuels. Both gases and small aerosols of these ele-
ments may be transported long distances in the atmosphere. These chemicals there-
fore exist in the atmosphere at HBEF in much higher concentrations than that
expected from the local environment, and their dry deposition is seriously underes-
timated by bulk precipitation collectors. Using the long-term record and a mass-
balance approach, Likens et al. (1990a) initially estimated that for sulfur, some
37 % of the total (wet plus dry) deposition annually occurred as dry deposition.
Later work using stable isotopes (Alewell et al. 1999, 2000) showed that the mass-
balance approach was not directly suitable for estimating dry deposition of sulfur
for the HBEF.
Lovett et al. (1992) compared the results of three independent methods for esti-
mating dry deposition of sulfur at HBEF: the mass-balance approach, a throughfall
approach, and an inferential method using measured atmospheric concentrations
and assumed deposition velocities. The results of these approaches give annual val-
ues that differed by about fourfold, with the mass-balance approach producing the
largest value (Lovett et al. 1992). Likens et al. (2002a) considered various factors
affecting the dry deposition estimates of sulfur at HBEF and the effect of uncertain-
ties in these estimates on mass balances. They estimated dry deposition as 21 % of
bulk deposition. More recently, the dry deposition of sulfur was estimated by using
four different models based on regional values of SO2 emissions. This approach
gave dry deposition values for 2002 ranging from ~0.5 to 2.5 kg S/ha-year at the
HBEF (Mitchell et al. 2011). Each of these approaches has assumptions and errors
associated with it, and dry deposition remains an extremely complicated flux to
measure quantitatively in forest ecosystems, particularly in mountainous terrain.
Currently, we operate an EPA-funded air quality and dry deposition monitoring
site as part of the Clean Air Status and Trends Network (CASTNET). This site has
been operational (at 250-m MSL, mean sea level) at HBEF since 1988. It uses an
inferential method for estimating dry deposition, i.e., dry deposition fluxes are cal-
culated as the product of measured ambient concentration and modeled deposition
velocity (http://www.epa.gov/castnet/epa_jsp/sites.jsp).

Elevational Effects

There is no significant difference in the concentration of calcium, magnesium,

sodium, or potassium in bulk precipitation at different elevations (from 545- to
758-m MSL) within the south-facing experimental watersheds of the HBEF (Likens
44 3 Chemistry

et al. 1967). Furthermore, a study done in 1971–1972 indicated no significant

difference between the concentration of calcium, magnesium, sodium, potassium,
ammonium, or chloride in bulk precipitation samples from the south-facing experi-
mental watersheds at 610-m MSL and those from the U.S. Forest Service’s Robert
S. Pierce Ecosystem Laboratory at 252-m MSL. However, based upon an analysis
of paired samples from these two locations, the concentrations of sulfate and nitrate
were significantly higher (p < 0.01 and p < 0.05, respectively) at the lower elevation.
In a later, more comprehensive 10-year study done between 1978 and 1989, Martin
et al. (2000) found no statistical difference for any ions collected with bulk precipi-
tation collectors at the two elevations. The source of the higher values in the earlier
study may have been related to human activity at the lower elevations, but this was
not confirmed.
A recent analysis (2000–2009) of volume-weighted average monthly concen-
trations in three bulk precipitation collectors located at 252-m MSL, 550-m MSL
(W6), and 640-m MSL (W9) showed no significant (p > 0.05) differences in any
solute measured. Nevertheless, the largest differences observed were between
252-m MSL and the higher elevations demonstrating the importance of using
bulk precipitation collectors co-located with the watershed ecosystems being
studied.

Acid Precipitation

One of the more interesting and environmentally important discoveries from the
Hubbard Brook Ecosystem Study was finding high acidity in rain and snow at
HBEF (Likens et al. 1972; Likens 2010). The first sample of rain collected at the
HBEF in July 1963 had a pH value of 3.7. The average annual weighted pH from
1964 to 1965 was 4.03 (Fig. 14); the lowest value recorded for a storm at HBEF
during 1963–2009 was pH 2.85, and the highest value was 5.85 during 1963–2009.
During the early years of the study (1963–1974), no value for weekly precipitation
exceeded a pH of 5.0. Such acidic precipitation is decidedly abnormal chemically
because pure water in equilibrium with atmospheric concentrations of CO2 should
have a pH of not less than about 5.6 (Barrett and Brodin 1955). In other words, the
precipitation at HBEF had a hydrogen ion concentration 50–500 times greater than
expected during this early period.
The increased acidity of precipitation is the result of anthropogenic emissions of
SO2 and NOX from the combustion of fossil fuels, which are hydrolyzed and oxi-
dized to strong acids (H2SO4, HNO3) in the atmosphere (Bolin 1971; Likens et al.
1972; Likens and Bormann 1974a; Likens 2010). The occurrence of acid precipita-
tion at HBEF, which is >100 km distant from any large urban-industrial area,
emphasizes that one of the major ways humans can influence natural ecosystems is
through pollution of the atmosphere and that these effects are not limited to areas
adjacent to sources of pollution (Likens et al. 1979; Likens 1984).
Precipitation Chemistry 45

Fig. 14 Volume-weighted annual concentration of hydrogen ion in bulk precipitation for

Watershed 6 of the Hubbard Brook Experimental Forest from 1963 to 2009. The linear regression
is highly significant (p < 0.01, r2 0.84). The shorter dashed lines are fitted by eye

During the 1970s and 1980s, the acid rain debate in the USA and Europe was in
full swing (Likens 2010). The First International Symposium on Acid Precipitation
and the Forest Ecosystem was held in 1975 at Ohio State University. We helped to
organize this Symposium and participated in it. Numerous studies of the ecological
impact of acid rain were underway at the HBEF by 1975, but the backbone of our
research on this important environmental problem was the developing long-term
record of precipitation and streamwater chemistry. Partly on the basis of this long-
term record of precipitation chemistry at HBEF, the U.S. Congress enacted amend-
ments in 1990 to the 1970 Clean Air Act, which for the first time explicitly included
provisions for regulating acid rain. Both of these regulations reduced emissions of
SO2 and particulates and impacted precipitation concentrations at HBEF. The 1990
Amendments to the Clean Air Act called for a reduction of 10 million tons (9.1 mil-
lion metric tons) in SO2 emissions/year and reductions of 1.2 million tons (1.1 mil-
lion metric tons) in NOX emissions/year below 1980 levels by the year 2000.
Additional reductions in NOX emissions were required after 2000 (see http://www.
epa.gov/airmarkets/progsregs/arp/nox.html and http://www.epa.gov/airmarkets/
progsregs/arp/reductions.html). Acidity of bulk precipitation decreased during this
period, but some 18 years of continuous monitoring were required before we could
report that the acidity at the Hubbard Brook Experimental Forest was significantly
reduced (Fig. 14; Likens 1989). During the past 45 years, hydrogen ion has decreased
54 µEq/L (r2 0.84; Fig. 14).
46 3 Chemistry

30

25
Total US Emissions (Mmt/yr)

20

15 SO2 Emissions
NOx Emissions

10

0
1960 1970 1980 1990 2000 2010 2020 2030
Calendar year

Fig. 15 Total US emissions of SO2 and NOX for calendar years from 1970 to 2010. The third-order
polynomial regressions have r2 of 0.99. Data from U.S. Environmental Protection Agency

The sources, distribution, and ecological effects of acid rain have been studied in
North America during the past four decades, and much has been learned (e.g.,
Likens et al. 1972, 1979, 2001, 2012; Cogbill and Likens 1974; Likens and Bormann
1974a; Schindler et al. 1985; Charles 1991; Likens 1998, 2010; Driscoll et al. 2001;
Weathers et al. 2007). The HBEF is an important site for monitoring atmospheric
pollutants in the northeastern USA because of the long record, location “down-
wind” of major sources, and lack of local, major pollution sources. An extremely
important finding from the long-term data is that bulk precipitation concentrations
and both streamwater concentrations and output of sulfate at HBEF are significantly
correlated with emissions of SO2 from the source area for HBEF (Likens et al. 2001,
2002a, 2005). For example, US total emissions of SO2 and NOX have markedly
declined since 1970 (Fig. 15). Correlated with this decrease, long-term data from
HBEF showed a striking relationship (p < 0.01, r2 0.92) between emissions of SO2
from the source area for HBEF (24-h back trajectory) (Fig. 16) and sulfate concen-
trations in bulk precipitation at HBEF (see also Likens et al. 2005). A similar rela-
tionship was observed for national SO2 emissions and sulfate concentrations in bulk
precipitation at the HBEF (Fig. 16; r2 0.93).
Like sulfate, atmospheric deposition of nitrate is directly correlated with NOX
emissions (both p < 0.001; Butler et al. 2003, 2005), but ammonium deposition is
not correlated with ammonia emissions (p > 0.10) as changes in both emissions and
deposition have been small. Annual deposition of sulfate has been declining rapidly
(−0.5 kg/ha-year, p < 0.001) for several decades, and recently declines in nitrate, as
Precipitation Chemistry 47

Fig. 16 Annual average bulk

precipitation sulfate
concentrations for Watershed
6 of the Hubbard Brook
Experimental Forest related
to annual SO2 emissions from
the 24-h back trajectory,
source area for the Hubbard
Brook Experimental Forest
(r2 0.92) and to annual US
SO2 emissions, r2 0.93 (also
see Likens et al. 2005)

well as DIN (NH4+ + NO3−) deposition have declined in concordance with the most
recent NOX emissions declines. The EPA estimated that US emissions of NOX
declined from 24.2 million tons in 1998 to 11.3 million tons in 2012 (Fig. 15; http://
www.epa.gov/ttnchiel/trends/). Unfortunately, the US EPA has changed their meth-
odology and timeframe for calculating state and national emissions repeatedly,
which has led to much frustration and difficulty in attempts to establish these impor-
tant relationships, particularly over the long term (see Likens et al. 2005).
There is a strong correlation (r2 = 0.81) between the decrease in streamwater sul-
fate and sum of base cation (CB) concentrations at HBEF, but the nitrate concentra-
tion drives the major annual fluctuations in acid-neutralizing capacity (ANC) (as
determined by the difference between CB and [SO42− + NO3−]) (Fig. 17). It is expected
that the ANC of stream water in W6 will continue to increase as SO2 and NOX emis-
sions decrease and the ecosystems recover from acidification (Fig. 17). Also, it is
expected that as sulfur deposition declines, sulfate concentration in stream water
will decline less in W6 because of the release of stored sulfur in soil (Kerr et al.
2012).
In 2005–2009, the estimated sulfur loading from the atmosphere at HBEF wet
plus dry (~7 kg/year) was still much higher than values recommended for protec-
tion of sensitive forest and associated aquatic ecosystems like those found at HBEF
(see Likens 1992, 2010). Moreover, declining concentrations and inputs in atmo-
spheric deposition of base cations, particularly calcium and magnesium at HBEF
(Fig. 18b; Driscoll et al. 1989a; Hedin et al. 1994; Likens et al. 1996, 1998) and loss
of base cations from the soil (Chaps. 4 and 5), are causing forest and associated
aquatic ecosystems at HBEF to become even more sensitive to atmospheric inputs
of acidic substances (Driscoll et al. 1989a; Likens et al. 1996, 1998; Likens 2010;
Likens and Buso 2012).
48 3 Chemistry

Fig. 17 Acidification and recovery of acid-neutralizing capacity in stream water of Watershed 6

in the Hubbard Brook Experimental Forest. PIR preindustrial revolution, PAD post-acid deposition
(modified from Likens and Buso 2012)

Acid rain continues to have major environmental impact on terrestrial and aquatic
ecosystems of the HBEF despite large decreases in emissions of SO2 and NOX from
the source area (Likens et al. 2002a, 2005; Likens 2010) and declines in bulk pre-
cipitation acidity of 80 % since 1963. Undoubtedly, emissions will continue to
change in the future as a result of economic recovery, changing energy demand, and
federal and state actions, both within the USA and internationally.

Acid Rain and Base Cation Depletion

Long-term mass-balance analyses (see Chap. 4) led to the discovery that calcium
and other base cations had been depleted from watershed-ecosystems of the HBEF
by decades of acid deposition (Likens et al. 1996, 1998). It had been generally
assumed that annual changes in the storage of base cations within available nutrient
pools of forest soils like those at HBEF were negligible (see Likens and Bailey
2013; Chap. 5), but Reuss and Johnson (1986) on theoretical grounds, and detailed
studies of terrestrial ecosystems at various locations (e.g., Likens et al. 1996, 1998;
S. Bailey et al. 2003, 2005; Fernandez et al. 2003, 2010; Long et al. 2009; Warby
Precipitation Chemistry 49

30
a Precipitation

Concentration
25

(µmol/L)
20
15 Magnesium
Calcium
10
5
0

90 b Stream Water
80
70
Concentration

60
(µmol/L)

50
40
30
20
10
0
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Year

Fig. 18 Annual volume-weighted mean concentrations of calcium and magnesium in (a) bulk
precipitation and (b) stream water for Watershed 6 of the Hubbard Brook Experimental Forest
during 1963–2009

et al. 2009; Watmough and Dillon 2001, 2004) and in aquatic ecosystems (e.g.,
Buso et al. 2009; Jeziorski et al. 2008; Korosi et al. 2012), surprisingly showed
significant, widespread calcium depletion resulting from acid rain. Moreover, long-
term data from W6 of the HBEF showed sizeable declines in calcium (~50 %) and
magnesium (~30 %) concentrations in stream water (Fig. 18b) and even larger rela-
tive declines in precipitation concentrations (~75 % and ~80 %, respectively)
(Fig. 18a; see “Dilution” section below; Likens and Buso 2012; Likens and Bailey
2013).
Johnson et al. (1981) proposed a 2-step process in the neutralization of acid rain
where atmospheric inputs of hydrogen ion are first neutralized by the dissolution of
reactive alumina in the soil zone, and then both hydrogen ion and aluminum acidity
are neutralized by chemical weathering of primary silicate minerals. Increased loss
of base cations results in reduced aluminum neutralization. This study was impor-
tant for demonstrating the neutralization of acid rain and for showing the role of this
process in generating toxic dissolved aluminum in the rooting zone of the soil and
in stream water. Thus, mineral weathering is the dominant process that neutralizes
acidity, and its rate is governed largely by (1) presence of soil water, (2) types of
50 3 Chemistry

minerals present, (3) distribution of soil particle sizes, and (4) depths of soils and
unconsolidated deposits (Likens and Bailey 2013; see Chap. 5). Because upland
forests in the northern Appalachian Mts., like HBEF, have silicate minerals that
decompose slowly, have relatively coarse soils with high sand and low clay content
(small surface area), and have shallow depths to bedrock (Bailey 2000), neutraliza-
tion of incoming acids is limited and the impact of acid rain on these sensitive sys-
tems is relatively large.
Because the podzol soils at HBEF were generally quite acidic (pH < 4.7), the
geochemical effects of acid deposition were thought early in the HBES to be mini-
mal (Johnson et al. 1972). Krug and Frink (1983) suggested that acid deposition was
not responsible for the observed acidification of surface fresh waters, rather sug-
gesting it was due to changes in land use, and that anthropogenic sulfate deposition
merely replaced the organic acidity in fresh waters with little or no change in pH.
Other scientists concluded that Krug and Frink were partly correct (e.g., R. Davis
et al. 1985; Driscoll et al. 1989b) but not offsetting. The ecosystem response is
much more complicated (see below and Chap. 5), and there are harmful ecological
effects from the anthropogenic acidification and mobilization of dissolved alumi-
num. Over the long term, acid deposition has had several important effects on a
variety of soil processes, e.g., reduction in nitrogen fixation, loss of base cations,
and loss of buffering capacity. Theoretically, 1 m of rain at pH 4 could leach 20 kg/
ha of CaCO3 from the system in drainage water.
Acid precipitation apparently has been falling on much of the eastern USA since
the early 1950s (Cogbill and Likens 1974; Likens and Bormann 1974a; Butler et al.
1984; Cogbill et al. 1984). There have been many upward and downward short-term
trends in annual volume-weighted average concentrations of hydrogen ion between
1964–1965 and 2009–2010; however, no long-term trend in hydrogen ion concen-
tration was statistically significant until 18 years of continuous data had been
obtained (Likens 1989; Fig. 14). Such findings auger strongly for the value of long-
term studies.
Sulfuric acid is the dominant acid in precipitation at HBEF. Based on a stoichio-
metric formation process in which a sea-salt anionic component is subtracted from
the total anions (Cogbill and Likens 1974), sulfate contributed more than 65 % of
the acidity in precipitation during the early years of our study. Although sulfuric
acid dominated bulk precipitation at HBEF and accounts for most of the hydrogen
ion in precipitation, the increase in input of hydrogen ion during the first decade of
our study was caused by an increase in the annual nitric acid content of precipitation
falling on these watershed-ecosystems.
Precipitation chemistry has changed both qualitatively and quantitatively at
HBEF during the past five decades, including absolute concentrations (Fig. 19a)
and relative proportions of the component chemicals (Fig. 20; Tables 5 and 6).
Based upon the acid anion composition, the sulfate contribution to acidity dropped
from 75 to 50 % and nitrate increased from <15 to 35 % from 1964–1965 to 2009–
2010. Annual inputs reflect these changes in complex ways. In a further attempt to
resolve the relative importance of the various factors controlling annual hydrogen
ion inputs during the early years when acidity was high, a stepwise multiple
Precipitation Chemistry 51

regression analysis was done to relate the annual hydrogen ion input to a variety of
independent variables. An analysis of five independent variables indicated that
86 % of the variability in annual hydrogen ion input during the decade 1964–1974
was related to annual nitrate input. Six percent of the variability was caused by
annual sulfate input, 5 % by the input of the sum of all cations minus hydrogen ion,
2 % by year, and <0.01 % by the annual amount of precipitation. A multiple linear
regression of annual volume-weighted average concentrations during 1964–2009
showed that hydrogen ion concentrations could be predicted from sulfate concentra-
tions (p < 0.001) and chloride concentrations (p < 0.01) but not by sum of base cat-
ions, ammonium, nitrate concentrations, or precipitation volume (p > 0.19).
The sources and changing concentrations of many of these ions in precipitation
illustrate how economic decisions and activities far removed from this region can

a
100
Precipitation
80 Stream Water
Calcium

60
µEq/L

40

20

0
30
µEq/L

Magnesium
20
10
0
µEq/L

10 Potassium
5
0
Sodium
µEq/L

40

20

0
2000 Annual Water
mm/yr

1600
1200
800
400
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Year

Fig. 19 Annual volume-weighted mean concentrations in (a) bulk precipitation and (b) stream
water for Watershed 6 of the Hubbard Brook Experimental Forest during 1963–2009. Linear
regressions shown when correlations are significant (p < 0.05)
52 3 Chemistry

b
100
Hydrogen Ion
80
µEq/L

60
40
20
0
Ammonium
µEq/L

20
10
0

60
Nitrate
µEq/L

40

20

Sulfate
120
µEq/L

80

40

Precipitation
0 Stream Water

2000 Annual Water

mm/yr

1600
1200
800
400
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Year
Fig. 19 (continued)

influence the dynamics of the northern hardwood forest and associated aquatic eco-
systems. Several decades ago, changes in atmospheric chemistry resulting from
changes in fossil fuel consumption, disposal of gaseous effluents into the atmo-
sphere, and reduction of alkaline particulates in the atmosphere resulted in decreased
pH of precipitation over the northeastern USA. Initially, increases in hydrogen ion
concentration were related to increases in sulfur emissions (Likens et al. 1972,
2005). However, later in the long-term record of precipitation chemistry at HBEF,
significant changes in the concentration of hydrogen ion were caused by changes in
emissions of both nitrogen oxides and sulfur dioxide (Figs. 13 and 15).
Precipitation Chemistry 53

Because precipitation is weakly buffered, the presence of small amounts of strong

mineral acids, such as H2SO4 and HNO3, can greatly depress the pH value. Even
though only relatively small amounts of acid are involved, the potential effect of
increased amounts of hydrogen ions on biologic and chemical systems can be large.
Fish kills of major proportions and diminished reproduction of salmonoid-type
fish have been observed and attributed to acid precipitation in areas of similar geol-
ogy in the Adirondack Mountains of New York (Schofield 1976), in Canada
(Beamish 1976), and in Scandinavia (Wright et al. 1976). No fish occur in the small
headwater streams of the HBEF, but Eastern brook trout (Salvelinus fontinalis)
occur in the streams at lower elevations in the HBEF (Warren et al. 2008). This pat-
tern of distribution is thought to be the result of extreme acidification of headwaters
and from physical barriers within the drainage network.
Some charge equivalence changes and biotic responses: Appreciable amounts of
dissolved organic matter are found in the undisturbed tributaries to Hubbard Brook.
Fisher (1970) reported an average dissolved organic matter concentration of
2.34 mg/L during 1968–1969 in Bear Brook (downstream of W6). This amount
would be 1.05 mg of dissolved organic carbon per liter if the dissolved organic mat-
ter were 45 % carbon. Hobbie and Likens (1973) found an average annual weighted
concentration of 1.0-mg dissolved organic carbon (DOC) per liter in stream water
from W6. Most of the streamwater carbon values ranged between 0.3 and 2.0 mg/L,
but a maximum carbon value of 4.8 mg/L was observed during an extremely high
streamflow event (29 July 1969, 34 × 104 L/ha-day). McDowell and Likens (1988)
reported an average DOC concentration of 1.8 mg/L in Bear Brook stream water
during 1976–1977 and 3.1 mg/L in 1978–1979. Lawrence et al. (1986) measured an
average DOC concentration of 1.8 mg/L in 1984. Systematic weekly measurement
of DOC began in W6 in 1995. During 1995–2010, the mean DOC concentration
was 2.2 ± 0.30 SD mg/L (range 0.18–8.31 mg DOC/L) with no temporal trend.
Values during the fall tend to be higher than at other times of the year, probably
because of rewetting of the stream channel. Moreover, there is an inverse relation
between DOC concentrations and pH of stream water (greater in W9, pH 3.9–5.7)
and a direct relation with instantaneous flow.
Several authors have reported recent increases in DOC concentrations in surface
waters of North America and Europe. Suggested causes for this increase include
reduction in acid deposition and recovery from acidification, land-use change, cli-
mate change, and changes in atmospheric deposition, e.g., nitrogen (e.g., Findlay
2005; Roulet and Moore 2006; Montieth et al. 2007; Erlandsson et al. 2011; Evans
et al. 2006, 2012). Unfortunately, the record of measurement in the HBEF is not
continuous from the early 1960s and analytical methods have changed, but available
data do not provide a convincing case for an increase in DOC concentration in
stream water during the past 3+ decades.
Watershed 9 is the most acidic (current average pH ~4.5) and humic (average
DOC ~10 mg/L) watershed in HBEF (Likens and Buso 2006). It is fishless, likely
due to low pH and high dissolved aluminum concentrations (Driscoll et al. 1980),
but brook trout (Salvelinus fontinalis) are present in Hubbard Brook at the conflu-
ence with W9 (Warren et al. 2008). Watershed 9 is typical of coniferous-dominated
54 3 Chemistry

(spruce/fir) headwater streams in New England. There has been a 20 % decline in

ionic strength (based on total dissolved cations and anions) in W9 since 1995—a
decrease of ~5 µS/cm in average electrical conductivity. Average sulfate concentra-
tion has declined by ~37 µEq/L (87–50 µEq/L) (Fig. 21) during this period or ~58 %
lower than measured in 1989 (Hedin et al. 1990). The DOC concentrations and
exports from W9 have not increased as expected (e.g., Erlandsson et al. 2011) dur-
ing this time (p > 0.10), but may in the future if the current sulfate decline were to
level off because of the release of sulfate from large reservoirs of absorbed sulfate
in the humic soils and wetlands of W9. Changes in DOC export have important
implications for watershed carbon budgets and stream biota. For example, with
decreasing ionic strength and lower competition from sulfate (De Wit et al. 2007;
Montieth et al. 2007), the charge density of DOC probably will continue to increase,
surpassing sulfate contributions to charge density in the next decade or so. As a
result of these changes, brook trout may return to W9 in the foreseeable future.

a
100
90
Precipitation Calcium
Percent of Cation Charge (%)

CATIONS Ammonium
Hydrogen Ion
80 Sodium

70
60
50
40
30
20
10
0
90 Precipitation
Percent of Anionic Charge (%)

Sulfate

80 ANIONS Nitrate
Chloride

70
60
50
40
30
20
10
0
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Year

Fig. 20 Percent of ionic charge for cations and anions in bulk precipitation (a) and stream water
(b) for Watershed 6 of the Hubbard Brook Experimental Forest during 1963–2009
Precipitation Chemistry 55

b
50
Stream Water
Percent of Cation Charge (%) CATIONS
40

30

Calcium
20 Ammonium
Hydrogen Ion
Sodium

10

0
90
Percent of Anionic Charge (%)

80
70
60 Stream Water
ANIONS
50
Sulfate
40 Nitrate
Chloride
30
20
10
0
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
Water Year

Fig. 20 (continued)

Headwater results scaled to Hubbard Brook and beyond: Watershed 6, a headwater

stream, has an area of 13.2 ha and drains into Hubbard Brook, a fifth-order river,
with a drainage area of ~32 km2. Because of the 2-order magnitude difference in
size of these drainage areas, discharge differs widely: high flows in W6 approach
0.1 m3/s, whereas high flows in Hubbard Brook can reach an estimated 200 m3/s.
Hubbard Brook is a tributary to the Pemigewasset River (2,644 km2 drainage area)
that, in turn, drains into the Merrimack River (12,200 km2 catchment area), a major
source of drinking water in an area of New England with nearly 200 communities
and over two million inhabitants. The Merrimack River watershed has also been
designated for anadromous fish reintroduction, including Atlantic salmon (Salmo
salar), a species requiring high-quality headwater streams, such as in the upper
Pemigewasset River system, for spawning. Atlantic salmon fry are found in the
lower part of Hubbard Brook.
56 3 Chemistry

60
a

µEq/L 40

20 Hydrogen Ion
Aluminum

0
b
80
2021
µEq/L

60
Sulfate
40 DOC Charge

20

0
1990 1995 2000 2005 2010 2015 2020 2025
Water Year

Fig. 21 Trends in hydrogen ion (r2 0.48, p = 0.003), monomeric aluminum (r2 0.40; p = 0.009),
sulfate (r2 0.79; p = 0.01), and DOC charge (r2 = 0.27; p = 0.039) in W9 stream water of the Hubbard
Brook Experimental Forest. Anionic charge for DOC is projected to exceed sulfate in 2021

Both W6 and Hubbard Brook represent the headwaters of this important

Merrimack River watershed, and research into the impacts of disturbances on these
smaller catchments is essential to understand and manage the larger system.
Maintaining high water quality with water processing plants downstream is prohibi-
tively expensive compared to maintaining pristine natural catchments upstream.
Yet, because they are small in area and dilute, headwater streams are among the
most sensitive of aquatic resources to pollutants (e.g., Lowe and Likens 2005).
Unfortunately, acidic deposition has thoroughly permeated W6 and Hubbard
Brook (and all of New England). Concentrations of strong acid anions (AA: sulfate
and nitrate) are very similar in these different-sized drainages (Fig. 22). The slopes
of the time-series trends are slightly different (−1.93 ± 0.08 se for Hubbard Brook;
−2.30 ± 0.12 se for W6), although the median AA values are not statistically differ-
ent. Evidently, W6 is releasing AA about 19 % faster than Hubbard Brook. Both
these streams are recovering from acidic deposition and should achieve similar, low
AA values in the near future (11 µEq/L is the target baseline within three decades if
deposition rates were to continue to fall) (see “Dilution” below and Likens and Buso
2012). The full impact of strong acid anions depends on the availability of neutral-
izing base cations in the soil and till and on the hydrologic pathways where the
dissolved ions are generated. The sensitivity of these two watersheds to acid deposi-
tion is high, but differs. Hubbard Brook has on average 14 µEq/L more base cations
Precipitation Chemistry 57

200
a W6 Base Cations
180
Hubbard Brook Base Cations
160
Concentration (µEq/L)

140 Base Cations

120
100
2039
2032
80
60
66 µEq/L
40 39 µEq/L
20
0
b W6 Acid Anions
180
Hubbard Brook Acid Anions
160
Concentration (µEq/L)

140 Sulfate + Nitrate

120
100
80
60 2032 2039
40
20
11 µEq/L
0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Water Year

Fig. 22 (a) Trends in the sum of base cation concentrations for Watershed 6 (W6) and Hubbard
Brook (HB) stream water from 1963 to 2009. Slopes significant at p < 0.01; W6 slope −1.48 µEq/
L-year; HB −1.15 µEq/L-year. (b) Strong acid anion concentration (sulfate + nitrate) trends are
highly significant (p < 0.001) and rigorous, r2 > 0.90 for both W6 and HB. Baseline target values
explained in text

than W6, and the medians are statistically different. However, annual base cation
values fluctuate more in Hubbard Brook than in W6 (r2 = 0.56 in Hubbard Brook;
r2 = 0.87 in W6 for linear regressions shown in Fig. 22), reflecting a broader range
of water pathways from long-contact times at low flows (high base cations) to
“short-circuiting” at high flows (low base cations) and more variable geologic sub-
strates in the much larger Hubbard Brook drainage. Larger, more complex water-
sheds probably will be slower to release acid anions, even with reduced acid
deposition (Mitchell and Likens 2011).
These findings have important ramifications for setting and maintaining water
quality standards downstream. Moreover, the requirement of young salmonids to
live in relatively uniform chemical conditions is compromised in Hubbard Brook,
where shifts in pH can range from pH 6.5 to pH 5.0, mobilizing toxic aluminum at
58 3 Chemistry

Photograph 7 Winter conditions at the HBEF ca. 1969. Note that the snowpack has mostly bur-
ied the instrument shelter for the gauging weir on Watershed 6 (see page 26)

the lower pH values (Johnson et al. 1981). Any changes (±) in acidic deposition will
be realized sooner in W6 than in Hubbard Brook, but both will react relatively
quickly because they have been sensitized by decades of acidic inputs and conse-
quent depletion of base cations.
The biogeochemistry in the Hubbard Brook Valley is poised for large and signifi-
cant biogeochemical changes in the upcoming decade or two—this situation pro-
vides compelling motivation for the maintenance of a continuing and adaptive
monitoring program (sensu Lindenmayer and Likens 2009, 2010).

Snow and Snowpack Chemistry

The chemistry of the considerable amount of snow that frequently accumulates as

snowpack during the winter at HBEF (see Photograph 7) may have little similarity
to the chemistry of the snow that falls (Hornbeck and Likens 1974) but can signifi-
cantly affect the chemistry of stream water during snowmelt (e.g., Johnson et al.
1981; Hooper and Shoemaker 1985; Hooper 1986). For example, potassium con-
centrations tend to be up to two orders of magnitude higher in the snowpack than in
ambient snowfall. In addition to concentration effects due to sublimation, much of
this increase comes from the leaching of potassium from plant parts also incorpo-
rated in the snowpack (Hornbeck and Likens 1974). In sharp contrast, amounts of
nitrate and ammonium in the snowpack tend to be lower than amounts input in
ambient snowfall.
Precipitation Chemistry 59

Fig. 23 General relationships for two seasonal periods in Watershed 6 at Hubbard Brook
Experimental Forest. The values are hydrogen ion concentrations in µEq/L (modified from
Hornbeck et al. 1976)

The first portions of meltwater (~20 %) from the snowpack may be significantly
more acidic than subsequent portions (Hornbeck et al. 1976; Johannessen et al.
1976). This phenomenon may cause a sharp drop in the pH of drainage streams and
lakes during the spring and has been blamed for fish kills in Scandinavia (e.g.,
Wright et al. 1976) and eastern North America. This seasonal change in pH also has
been observed in streams and in Mirror Lake within the Hubbard Brook Valley, but
it is a very temporal condition of variable intensity. During the winter and with
intermittent melting, the snowpack becomes less acidic than the incident precipita-
tion (Hornbeck et al. 1976). When the snow melts rapidly, there may be a flush of
highly acidic water into the stream channels; however, when the snowpack melts
more slowly, the reduction of pH in stream water is much less pronounced.
Despite seasonal and long-term differences in acidity of precipitation, effects of
snow accumulation and melt, and variations in residence time of soil water, the acid-
ity of stream water at the HBEF at the level of the gauging weirs is remarkably
constant throughout the year (Fig. 23). Even though the average monthly concentra-
tion for hydrogen ion in bulk precipitation for W6 ranged from ~40 to ~70 µEq/L,
average monthly concentrations in stream water from W6 only changed from ~8 to
~12 µEq/L during 1964–2009 (Fig. 24). The uniformity of streamwater acidity indi-
cates a strong buffering action of the terrestrial ecosystem (Fig. 23). The terrestrial
ecosystem is therefore very effective in ameliorating the short-term impact of acid
precipitation on the associated aquatic ecosystems. An exception is the acid pulse
60 3 Chemistry

3.9
120
Hydrogen Ion Concentration (µEq/L)

Precipitation
Stream Water
100 4.0

80 4.1

4.2

pH
60
4.3
40 4.4
4.5
20 4.7
5.0

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Month of Year

Fig. 24 Volume-weighted mean monthly hydrogen ion concentration and pH for bulk precipita-
tion and stream water for Watershed 6 of the Hubbard Brook Experimental Forest during
1965–2009. The vertical lines for each month represent 1 standard deviation from the mean

during snowmelt in the upper headwaters. Because of the long-term nature of this
acid deposition problem, the buffering capacity of the ecosystem has been slowly,
but relentlessly, depleted (Likens et al. 1998; Likens 2010; Likens and Buso 2012).

Dilution and Effect of Acid Rain on Baselines

Long-term records (>10 years) provide a unique opportunity for new insights and
for developing realistic temporal trends. As shown repeatedly in this little book,
short-term records, e.g., <6 years, can be quite misleading regarding biogeochemi-
cal patterns and trends (e.g., Fig. 14).
Using our long-term data now approaching five decades, it is apparent that the
chemistry of both precipitation and stream water in the Hubbard Brook Valley has
moved from what we may have thought to be “normal” conditions when we initi-
ated the Hubbard Brook Ecosystem Study in 1963 to very dilute conditions today.
In fact, based on electrical conductivity, concentrations in W6 are projected to
approach demineralized water in 2015 (bulk precipitation) and 2025 (stream water)
(Fig. 25; Likens and Buso 2012). Because such dilute chemistry is unrealistic for
natural waters, theoretical baseline values of 3 µS/cm and 5 µS/cm were calculated
for bulk precipitation and stream water, respectively.
Precipitation Chemistry 61

40
VWA Electrical Conductivity (µS/cm)

Precipitation

30 Precipitation EC
Streamwater EC

20

Stream Water
2015
2025
10
2025
5 µS/cm
3 µS/cm
0
1960 1970 1980 1990 2000 2010 2020 2030 2040
Water Year
Fig. 25 Volume-weighted annual electrical conductivity in bulk precipitation and stream water for
Watershed 6 of the Hubbard Brook Experimental Forest during 1964–2009 (modified from Likens
and Buso 2012)

The projected baseline values for sum of base cations (calcium, magnesium, potas-
sium, sodium) and acid anions (sulfate and nitrate) in bulk precipitation and stream
water are lower (Table 7; post-acid deposition, PAD) than predicted earlier (Likens
et al. 1996) for preindustrial revolution (PIR) conditions and the result of federal air
pollution regulations reducing inputs and of excessive leaching of base cations from
the watershed-ecosystems by decades of acid rain (Likens and Buso 2012). Rapid
dilution of surface waters to very low values has important ecological, biogeochemi-
cal, and natural resource management implications (Likens and Buso 2012).
Projected PAD baseline concentrations have been calculated for precipitation
and stream water at electrical conductivity of 3 and 5 µS/cm, calcium of 2 and
11 µEq/L, magnesium of 1 and 6 µEq/L, potassium of 1 and 3 µEq/L, sodium of 3
and 19 µEq/L, sulfate of 8 and 10 µEq/L, nitrate of 2 and 1 µEq/L, chloride of 3
and 8 µEq/L, and pH 5.4 and 5.5, respectively (Likens and Buso 2012). We assume
that stream water in PIR was dominated by Ca(HCO3)2, but now is moving toward
NaHCO3 because of preferential calcium depletion from soils due to acid rain
(Likens and Buso 2012).
Although such long-term data are indispensible in order to think about baseline
conditions, there are many pitfalls, including the following: (1) Are the records long
enough to make accurate projections and (2) can simulation models be used to
improve these extrapolations? Given the difficulty in describing baseline condi-
tions, the effort is important because of the cost of implementing federal regulations
and the biological imperative to manage and protect sensitive ecosystems.
62 3 Chemistry

Table 7 Post acidic deposition (PAD) models for bulk precipitation and streamwater
concentrations in Watershed 6 of the Hubbard Brook Experimental Forest (modified from Likens
and Buso 2012)
Precipitation Stream water
PAD Model for 2015 PAD Model for 2025
Conc Charge Conc Charge
Ion (mg/L) (μEq/L) Ion (mg/L) (μEq/L)
Calcium 0.04 2.0 Calcium 0.22 11.0
Magnesium 0.01 1.0 Magnesium 0.07 6.0
Potassium 0.04 1.0 Potassium 0.12 3.0
Sodium 0.07 3.0 Sodium 0.44 19.0
Aluminum 0.000 0.0 Aluminum 0.001 0.1
Ammonium 0.036 2.0 Ammonium 0.002 0.1
H+ (pH) 5.40 4.0 H+ (pH) 5.52 3.0
sum CB 7.0 sum CB 39.0
CATION sum 13.0 CATION sum 42.2
Sulfate 0.38 8.0 Sulfate 0.48 10.0
Nitrate 0.12 2.0 Nitrate 0.06 1.0
Chloride 0.11 3.0 Chloride 0.28 8.0
Phosphate 0.001 0.1 Phosphate 0.001 0.1
Bicarbonate (ANC) 0 0.1 Bicarbonate (ANC) NA 15.0
DOC 1.0 0.0 DOC 1.5 8.0
sum AA 10.0 sum AA 11.0
ANION sum NA 13.2 ANION sum NA 42.1
IMBALANCE −0.2 IMBALANCE 0.1
Theoretical 3.0 Theoretical 5.4
Electrical Electrical
conductance (μS/cm) conductance (μS/cm)

Impact of Acid Rain on Accumulation of Forest Biomass

Our model of the forest ecosystem (Fig. 1) suggests a number of functions internal
to the forested ecosystem that may be affected by acid precipitation. For example,
leaching of substances from the canopy can be accelerated and, in fact, that seems
to be the case. Eaton et al. (1973) found that 90 % of the hydrogen ions striking the
summer canopy of the forest were consumed within the canopy, presumably releas-
ing an equivalent amount of base cations. Laboratory studies of seedlings of major
tree species at HBEF (Wood and Bormann 1974, 1975) indicated increased leaching
of cations as the hydrogen ion content of artificially applied mist increased.
The forest ecosystem is regulated by a large number of environmental factors
such as the amount and distribution of precipitation; temperature; the length of
growing season; unusual meteorologic events, such as ice and wind storms; the
availability of nutrients; insect and disease outbreaks; and pollutants such as acid
rain and ozone. Significant effects of acid rain on forest growth are difficult to
detect. Moreover, it is extremely difficult to separate the effects of acid rain from
other variables. It is entirely possible, however, that acid rain is reducing forest
growth or tree survival over wide areas by small, but persistent impacts, or by
Precipitation Chemistry 63

Biomass for live and dead trees on W6 over time

240 50

Live trees > 10 cm dbh 45

220
40

200 35
Mg/ha 30 Mg/ha
180
(live trees
> 10 cm dbh) 25 (dead trees
and live trees
160
20 2-9 cm dbh)
Dead trees > 10 cm dbh
140 15

10
120
5
live trees 2-9 cm dbh
100 0
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

Year
Fig. 26 Biomass accumulation for live and dead trees, above- and belowground in Watershed 6 of
the Hubbard Brook Experimental Forest from 1965 to 2007 (Data from Whittaker et al. 1974;
Likens et al. 1994; Siccama et al. 2007; van Doorn et al. 2011)

upsetting ecological conditions so as to reduce the capacity of certain species to

resist insect or disease attacks or by reducing the ecosystem’s buffering capacity
through excessive leaching of base cations. Slow but cumulative effects over many
years may result in serious deterioration of the forest ecosystem, particularly as acid
deposition is but one of an array of pollutants affecting the landscape (e.g., Bormann
1974). We (Likens et al. 1996) suggested that some 840 kg Ca/ha had been depleted
from W6 soil pools during 1940–1995 by acid rain, but our recent models of long-
term streamwater losses from the ecosystem suggest a value closer to 300 kg Ca/ha
during this period (see Chap. 6). Nevertheless, 300 kg Ca/ha per 55 years is still a
large loss from the ecosystem.
How has the forest ecosystem in W6, the reference watershed, responded to acid
rain? Whittaker et al. (1974) and Cogbill (1976) reported a decline in forest growth
on south-facing slopes of the HBEF during the 1960s and 1970s, but total forest
biomass in W6 increased rapidly until ~1982 and then leveled off (Likens et al.
1994; Fahey et al. 2005; Siccama et al. 2007; Lindenmayer and Likens 2010, p. 123;
Fig. 26). Recent measurements, 2007, showed that the forest is now declining in
biomass (Lindenmayer and Likens 2010, p. 123; van Doorn et al. 2011). Soil carbon
is a very large pool at the HBEF; annual changes are small and difficult to measure
(Huntington et al. 1988, 1989; S. P. Hamburg, personal communication).
Nevertheless, the forest in W6 is probably no longer accumulating biomass, but now
may be emitting carbon to the atmosphere rather than sequestering it (Fig. 26;
Likens and Franklin 2009; Lindenmayer and Likens 2010; Likens and Bailey 2013).
There are several possible explanations for this decrease in living forest biomass
at the HBEF. It could be a natural pattern; a response to climate change (e.g.,
64 3 Chemistry

Campbell et al. 2007; Likens 2010); an effect of air pollution, such as ozone (e.g.,
Ollinger et al. 1997) or a direct effect of acid rain; or possibly disease was a major
causal factor (e.g., Lovett et al. 2006); perhaps it is nutrient limitation such as by
nitrogen (e.g., LeBauer and Treseder 2008) or calcium. Applying a hallmark HBES
approach to this complicated question, a watershed-scale manipulation was initiated
in 1999 whereby Wollastonite (CaSiO3) was added experimentally to W1 of the
HBEF. The calcium added (1,189 kg calcium/ha) was roughly equivalent to the
amount calculated to have been leached from the ecosystem by acid rain in the pre-
vious 50 years or so (Likens et al. 1996). The early results of the experiment are
encouraging regarding the important role of calcium in this ecosystem (e.g., Juice
et al. 2006; Halman et al. 2008). The seedlings, germinants, and canopy of sugar
maple (Acer saccharum) are responding positively to the addition of the wollaston-
ite, in comparison with sugar maple in adjacent, untreated areas. Likewise, red
spruce (Picea rubens) showed less winter injury in the treated area in comparison to
the reference area (Hawley et al. 2006).
A number of researchers have linked the decline in sugar maple in Pennsylvania,
New York and New England to changes in base cations and aluminum availability
in the soil, e.g., Horsley et al. (2002), Bailey et al. (2004, 2005), Hallett et al. (2006),
Long et al. (2009), Likens and Bailey (2013). In the Duchesnay Experimental Forest
near Quebec City, Quebec, Moore et al. (2012) evaluated the effect of a dolomitic
lime addition on sugar maple 15 years after the application. They found that a single
addition had long-term beneficial effects on soil chemistry and sugar maple nutri-
tion, vigor, growth, and regeneration. To obtain maximum information useful to
policy makers and those responsible for the management of the northern hardwood
forest, efforts such as this must be integrated and interpreted fully within an ecosys-
tem framework. A program mounted in this way is the highest socially responsible
investment of research funds for helping to resolve a potentially critical environ-
mental problem.

Trace Metals

Another component of regional air pollution impact on the HBEF is from atmo-
spheric input of various trace metals, including copper, lead, cadmium, mercury,
and zinc, to these watershed-ecosystems (Siccama and Smith 1978; Siccama et al.
1980; Smith and Siccama 1981; Johnson et al. 1995; Driscoll et al. 2007). Here we
will briefly report on the biogeochemistry of lead in the HBEF. Sizeable concentra-
tions of lead were found in bulk precipitation input to W6 during 1975–1985, lead-
ing to significant inputs and accumulation of this toxic metal within the ecosystem
(Fig. 27). Concentrations in bulk precipitation during this period often exceeded the
US drinking water standard of 25 µg/L (Johnson et al. 1995).
Much of the lead in rain and snow came from the combustion of leaded gasoline
by automobiles and buses. In the 1970s, legislation was passed to restrict the sale of
leaded gasoline (gasoline with alkyl-lead additives) in the USA. As a result of
 a Gasoline Use
250

200

Pb amount
(106 kg)
150

100

50

b Precipitation
80 400

300

(g+ha−t • yr−1)
60

Pb input
Pb (µg/L)

200
40
100
20
0

c Forest Floor
1200
Pb (mg/m2)

1000

800

600

d Stream water
10

8
Pb (µg/L)

0
1975 1980 1985 1990

Fig. 27 Lead (Pb) use in gasoline and lead in precipitation, forest floor, and stream water at
Watershed 6 of the Hubbard Brook Experimental Forest. Data for Pb consumed in gasoline (a) are
from Nriago (1990). Squares (b) are estimates of Pb input to the HBEF, based on volume-weighted
average Pb concentrations in bulk precipitation. Forest floor Pb pool sizes (c) are computed as the
average Pb content in 15 × 15-cm block samples; sample size varied from 58 to 119. Error bars
show 1 standard error. Streamwater concentrations (d) from weekly samples of stream water in
Watershed 6. (Redrawn from Johnson et al. 1995)
66 3 Chemistry

Table 8 Ecosystem budgets of lead (Pb) for two time periods: high atmospheric deposition of Pb
(1975–1977) and low Pb deposition (1985–1987) (adopted from Johnson et al. 1995)
1975–1977 (g/ha-year) 1985–1987 (g/ha-year)
Bulk precipitation input 325 29
Streamwater output 6 4
Net plant uptake 33 1
Weathering release 7 7

decreased use of leaded gasoline, concentrations and inputs of lead in bulk precipi-
tation, accumulation in the forest floor, and outputs in stream water of W6 declined
precipitously (Fig. 27). Johnson et al. (1995) determined some ecosystem fluxes for
lead during a period of high atmospheric deposition (1975–1977) and during a
period of low atmospheric deposition (1985–1987) (Table 8).
Similar to the acid rain situation, pollutants generated some distance from the
HBEF were transported through the atmosphere to be deposited at HBEF. Also as
with acid rain, legislation restricting the release of lead to the atmosphere was
reflected in improving environmental conditions relative to this toxic metal in areas
like HBEF remote from the source(s).

Throughfall and Stemflow

The water that reaches the forest floor is of vastly different chemical content than
the incident precipitation. Only about 85 % of the incident precipitation actually
reaches the ground of the northern hardwood forest ecosystem during the growing
season. The remainder is evaporated (or absorbed) directly from the surface of
leaves, branches, and stems (Leonard 1961; Lovett et al. 1996). As the precipitation
passes through the forest canopy, its chemistry is altered significantly (Table 9).
Throughfall and stemflow (about 5 % runs down the stems of deciduous trees) dur-
ing the growing season are greatly enriched in potassium (about 91 times), phos-
phorus (18 times), magnesium (15 times), and calcium (10 times). All other
dissolved substances also increase, with the exception of hydrogen ion, which is
retained within the canopy by cation exchange reactions.
Based on an average DOC concentration of bulk precipitation of 1.1 mg/L
(1973–1974; Likens et al. 1983) and an average, long-term precipitation of
1,440 mm, some 15.8 kg/ha-year of dissolved organic carbon would be added to the
ecosystems of the HBEF each year in direct precipitation. As the precipitation
passes through the forest canopy, it is enriched in organic carbon (Table 10;
McDowell and Likens 1988). Throughfall and stemflow during the growing season
contain some 52 kg carbon per hectare (Eaton et al. 1973), so an additional 36 kg
carbon/ha must be leached from the forest canopy and added to that already present
in the precipitation. These relationships probably are much more complicated than
Throughfall and Stemflow 67

Table 9 Weighted average chemical composition of incident precipitation and throughfall under
different size classes of sugar maple, yellow birch, and American beech trees of Watershed 6
during the growing season (values in mg/L, except H)
Precipitation under canopy
Precipitation above Throughfall and
canopy stemflow Throughfall Net change
Solutes (a) (b) (a) (b) (a) (b)
Ca 0.16 0.11 1.59 0.72 +1.43 +0.61
Mg 0.03 0.02 0.45 0.20 +0.42 +0.18
K 0.07 0.06 6.37 1.98 +6.30 +1.92
Na 0.06 0.07 0.14 0.10 +0.08 +0.03
NO3–N 0.22 0.32 0.67 0.41 +0.45 +0.09
NH4–N 0.21 0.23 1.21 0.225 +1.00 ‒0.005
Total N 0.44 – 2.44 – +2.00 –
PO4–P 0.0026c – 0.15 – +0.147 –
SO4–S 0.9 0.84 5.4 1.15 +4.5 +0.31
Cl 0.45 0.15 1.46 0.28 +1.01 +0.13
H (µEq) 86.3 74.6 0.010 30.4 ‒0.077 ‒44.2
Organic C 2.4d 3.41 12 8.75 +10 +5.34
(a)
Modified from Eaton et al. (1973) for 1969
(b)
Modified from Lovett et al. (1996) for 1989–1992
c
Weighted average for 1972–1974
d
Weighted average for 1973–1974

shown by this simple calculation, because the microflora of the leaf and bark sur-
faces of the vegetation may metabolize the more labile, dissolve organic substances
in precipitation, and in turn secrete other organic substances.
There are various sources for the dissolved substances in throughfall and stem-
flow, but their relative importance is not entirely clear (e.g., Eaton et al. 1973): (1)
Some nutrients are contained in the incident precipitation and (2) some of the nutri-
ents may have been impacted aerosols that are washed off by the incident precipita-
tion. These two portions should be considered as a part of the chemical flux from
outside the ecosystem. Similarly, (3) nutrients with a normal gaseous phase, which
were incorporated directly on or into the plant (e.g., fixation of carbon by photosyn-
thesis, absorption, and reaction with SO2 or NH3) and then removed from tissues by
incident precipitation, should be considered as meteorologic inputs in the watershed-
ecosystem balance. The remaining nutrients in throughfall and stemflow, which
have been leached from the vegetational tissues or associated microflora, are nutri-
ents cycling primarily within the intrasystem cycle of the ecosystem (Fig. 1) and
should not be considered part of the meteorologic flux. We have not been able to
quantify the amounts of chemicals transported along each of these pathways long
term for the ecosystems at HBEF. Some specific data are presented later in the dis-
cussion of nutrient cycles at HBEF (Chap. 6).
68 3 Chemistry

Table 10 Volume-weighted mean concentrations of dissolved substances in stream water of

Watershed 6 of the Hubbard Brook Experimental Forest
Decade Decade
Stream water 1964–1973 Stream water 2000–2009
Conc Conc Conc Conc
Substance (mg/L) (µEq/L) Substance (mg/L) (µEq/L)
Water (mm/ 882 Water (mm/ 1,006
ha-year) ha-year)
Ca 1.36 67.8 Ca 0.62 30.8
Mg 0.34 28.0 Mg 0.19 15.8
K 0.23 5.8 K 0.16 4.2
Na 0.78 33.9 Na 0.69 29.8
NH4 0.04 2.1 NH4 0.01 0.3
Al 0.32 35.7 (@+3) Ale 0.23 25.0 (@+3)
H+ 0.015 15.0 H+ 0.008 7.9
SO4 6.0 124.6 SO4 3.60 74.9
NO3 1.8 28.7 NO3 0.16 2.6
Cl 0.5 15.0 Cl 0.38 10.7
PO4b 0.002 0.1 PO4b 0.001 0.0
ANC 0 0.0 ANC 8.0
(bicarbon- (bicarbon-
ate)a ate)a
DOCc 1.8 10.8 DOCc 2.1 12.6
Cations 188.3 Cations 113.7
Anions 179.2 Anions 108.9
DOCc, d 1.8 149.9 =µmol/L DOCd 2.1 174.9 =µmol/L
Dissolved silica 4.1 144.5 =µmol/L Dissolved silica 3.6 128.9 =µmol/L
pH 4.82 pH 5.10
a
Bicarbonate = 0 at pH < 5
b
PO4 based on n = 3 years (1971–1973)
c
DOC based on 1976–1977 (Likens et al. 1983)
d
DOC charge is assumed at −6 µEq/L per mg/L C
e
Aluminum is assumed at +3 valence

Streamwater Chemistry

The chemistry of stream water has been monitored on a weekly basis in Watersheds
1–6 of the HBEF since 1963 and in Watershed 7–9 since 1995. Base cations were
measured during 1963–1964 in W7 and a full suite of chemistry in W8 during
1974–1978. Experiments that involved cutting of the forest vegetation were under-
taken in Watershed 2, starting in November of 1965; in Watershed 4, starting in
October of 1970; and in Watershed 5, starting in 1983. An experimental addition of
Wollastonite was added to Watershed 1 in 1999. Only long-term data on streamwa-
ter chemistry from W6, the biogeochemical reference watershed at the HBEF, are
presented here.
Streamwater Chemistry 69

Concentrations of Dissolved Substances

The HBEF forest ecosystem brings about both qualitative and quantitative chemical
changes in rain and snowmelt water as it passes through the system (Tables 5, 6, and 10).
Water enters the system as a dilute solution of sulfuric and nitric acid (pH < 5) but
leaves the system containing primarily neutral sulfates (pH ~5). Calcium and sulfate
dominate the streamwater chemistry, but sodium, magnesium, and aluminum also
are relatively important cations on an equivalent basis. Sulfate is seven times more
abundant than the next most abundant anion, currently chloride, in stream water.
The balance between cations and anions in stream water is good over a 47-year span
(Tables 5 and 10). The ionic charge assigned to aluminum and to DOC is problem-
atic. The charge for aluminum can be modeled using pH, but the DOC charge is the
residual, although the anionic charge per mg carbon is typical of values in the litera-
ture (Buso et al. 2000).
Although there have been 40–50 % declines in ionic strength of both precipita-
tion and stream water (Tables 5, 6, and 10), stream water ionic strength is still
about twice that of the incoming bulk precipitation. A principal factor here is the
concentration effect (distillation) of evapotranspiration. As water is lost from the
system by evapotranspiration, chemicals in solution tend to be concentrated into a
smaller volume of water. Based upon an average annual evapotranspiration loss of
36 % (Table 3), the concentration factor would be 1.6, which would fall short of the
observed concentration factor for chemicals (2.2 times during 1964–1973; 1.7
times during 2000–2009; Table 11). However, the concentration factor is not so
straightforward as it appears, for not only do concentrations in stream water change
but the proportions of ionic or dissolved species change as well, indicating impor-
tant internal chemical and biologic reactions (Fig. 20b). Concentration factors
(stream water out ÷ precipitation in) vary from ~2 times for DOC and 7–10 times
for calcium, magnesium, and sodium to 0.3 times for hydrogen ion and 0.2 times for
nitrate and phosphate (comparing early and recent decades; Table 11). These fac-
tors are relatively stable for individual solutes, but over the long term the concentra-
tion factor is decreasing for the combined sums of all ions (Table 11). In any case,
an additional quantity of dissolved salts is acquired by the water from within the
ecosystem’s boundaries. This process is related to chemical weathering reactions
(Johnson et al. 1968; Likens et al. 1994, 1998, 2002a; S. Bailey et al. 2003) and is
discussed separately in Chap. 5.
In general, the measured streamwater concentrations of most dissolved sub-
stances vary within a narrow range (less than a factor of 2–4), even though discharge
of water may fluctuate over five orders of magnitude during an annual cycle at the
HBEF (Figs. 12 and 48 below). This relation is particularly true for calcium, mag-
nesium, and sulfate, which decrease with increasing streamflow. Sodium and dis-
solved silica concentrations may be diluted up to fourfold during periods of high
streamflow, whereas aluminum, hydrogen ion, dissolved organic carbon, nitrate,
and potassium concentrations are increased with increased discharge. Biotic activ-
ity within the ecosystem plays an important role in determining these relationships
70 3 Chemistry

Table 11 Concentration factors for all solutes across two decadal periods in Watershed 6 of the
Hubbard Brook Experimental Forest
Decade 1964–1973 Decade 2000–2009
Stream Conc Stream
Precipitation water fac- Precipitation water Conc
(µEq/L) (µEq/L) tor = S/P (µEq/L) (µEq/L) factor = S/P

Ca 7.5 67.8 9.1 3.5 30.8 8.9

Mg 3.5 28.0 8.0 1.5 15.8 10.3
K 1.5 5.8 4.0 1.0 4.2 4.1
Na 5.1 33.9 6.6 3.5 29.8 8.5
NH4 12.1 2.1 0.2 8.9 0.3 0.0
Al 0.0 35.7 NA 0.0 25.0 NA
H+ 73.1 15.0 0.2 27.6 7.9 0.3
SO4 59.3 124.6 2.1 24.8 74.9 3.0
NO3 23.8 28.7 1.2 17.2 2.6 0.2
Cl 10.7 15.0 1.4 4.4 10.7 2.5
PO4 0.27 0.06 0.2 0.09 0.03 0.3
DOC 91.6 166.6 1.8 68.4 174.9 2.6
dSi 0.0 65.9 NA 0.0 61.1 NA
dSi is dissolved silica in µmoles/L

for potassium and nitrate in stream water. Nitrate and potassium are quite sensitive
indicators of biologic activity; therefore, streamwater concentrations for these two
nutrients are markedly reduced during periods of plant growth (growing season) and
increased during periods of vegetation dormancy. Ammonium and phosphate show
no relationship with discharge.
To describe these variations in streamwater chemistry empirically, we developed
a model that predicts streamwater concentration in relation to stream discharge
(Johnson et al. 1969). The model is based on some rather simple and intelligible
assumptions. First of all, we assume that two discrete water types exist in our water-
shed-ecosystem: one is represented by relatively dilute water recently added to the
system in rain or snowmelt, and the other is older, more concentrated, subsurface
water (mostly soil water). Second, we assume that stream water at a given point at
any time is some mixture of these two water types. Last, we assume that subsurface
water is added to the stream at a constant rate, whereas rainwater or snowmelt water
is added on a variable, day-to-day basis. Essentially, then, stream water becomes a
variable mixture between pure rainwater or snowmelt water and pure subsurface
water, the proportions of which vary from day to day and season to season. During
flood periods, say during the spring snowmelt, stream water is composed primarily
of snowmelt water. During drought periods, say in midsummer, stream water is
composed of subsurface drainage water. In spite of these tightly prescribed and
perhaps simplistic assumptions, the model does in fact work and serves to describe
both reliably and accurately quantitative fluctuations in streamwater chemistry on a
Streamwater Chemistry 71

day-to-day or season-to-season basis (Johnson et al. 1969). Within an undisturbed

watershed, the model predicts quite precisely the changes in streamwater chemistry
elicited by a given storm or a given dry period. The simple premises of the model
testify to the remarkable chemical stability and chemical controls inherent in these
forested ecosystems. This model was developed when concentrations were signifi-
cantly higher (1964–1973), but still applies today despite major dilution of precipi-
tation and streamwater chemistry. It also is significant that the model completely
breaks down when the watershed is disturbed, such as by clear-cutting (e.g., Likens
et al. 1970).
The maintenance of relatively constant streamwater chemistry is in part a result
of the yearlong high infiltration capacity and permeability of the forest soils at the
HBEF and in part of a high rate constant for the chemical reactions manifested in
the soil zone. Apparently, chemical equilibria between aqueous and solid phases are
achieved rapidly through various geologic and biologic reactions in the soil. As a
consequence, streamwater chemistry is essentially established in the soil zone or the
interflow drainage passages within the upper reaches of each watershed-ecosystem
(Johnson et al. 1981; Likens and Buso 2006; Zimmer et al. 2012). This is an excel-
lent example of the way in which both the form and rate of chemical flux are modi-
fied and tightly regulated by a forest ecosystem. Similar constancy of streamwater
chemistry also has been observed for the major cations in the Coweeta Experimental
Forest of North Carolina (Johnson and Swank 1973).
These examples suggest that the relative independence of streamwater chemistry
from streamwater discharge is a rather general feature of eastern deciduous forest
ecosystems on weathering resistant granitic and schistic substrates. The importance
of this result should not be underestimated, for it demonstrates that the chemistry of
regional headwater streams is the product of the natural ecosystem, the control of
which may be altered by disturbance. For example, in disturbed forested areas and
agricultural lands, concentrations of chemicals in stream water usually increase but
vary markedly with discharge (e.g., Powell 1964, pp. 18–19; Likens et al. 1970;
Likens and Bormann 1974b).
Despite long-term trends, the concentration of most ions in stream water varies
little from year to year (Fig. 19), because gradual change is a characteristic of undis-
turbed watershed-ecosystems of the HBEF. In contrast, the annual weighted con-
centration of nitrate, generally increasing from 1964 to 1970, suddenly doubled
from 1969 to 1976 (water-years) and then sharply declined (Fig. 28). For example,
the average weighted concentration during 1964–1967 was 0.92 mg/L; during
1967–1969 it was 1.40 mg/L, in 1969–1972 it was 3.14 mg/L, and in 1992–1993 it
was 0.23 mg/L. An explanation for this initial rise in the nitrate concentration is
primarily circumstantial, and there are several possibilities. A severe drought in
1963 and 1964 may have “primed” the ecosystem for a nitrate release triggered by
rewetting and soil frost (see Bernal et al. 2012). The snow cover formed very late in
the winter of 1969–1970, and widespread freezing of the soils was observed. Annual
streamwater concentration of nitrate rose markedly that year (Fig. 28). Then, in
1973–1974, widespread freezing of the forest soils occurred again, and again nitrate
concentration increased. Various workers have reported that the freezing and
72 3 Chemistry

60

Nitrate
50
Concentration (µEq/L)

Precipitation
40 Stream Water

30

20

10

0
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Year

Fig. 28 Volume-weighted annual concentration of nitrate in bulk precipitation and stream water
for Watershed 6 at the Hubbard Brook Experimental Forest during 1964–2009

thawing of soils promotes nitrification and the mobilization of nitrate to drainage

waters (Arefýeva and Kolesnikof 1964; references cited in McGarity and Rajaratnam
1973). Therefore, freezing and thawing of the HBEF soils may have induced nitri-
fication and the subsequent loss of relatively large amounts of nitrate in stream
water. However, we have no explanation for the very slow rate of recovery (residual
effect) observed during the 2 years after the freezing occurred in 1969–1970. It
might have been expected that if soil frost enhanced nitrification, the nitrate would
have been flushed rapidly from the system and in subsequent years nitrate concen-
trations in stream water would have returned to “normal.” Then, after ~1978 nitrate
concentrations dropped to low values, barely above detection limits during summer
after 1980. Soil freezing in December 1989 and 1997 may have increased nitrate
concentration again in stream water of W6 (Fig. 28; see Mitchell et al. 1996).
However, following a severe soil freezing event during winter of 2006, Judd et al.
(2011) observed only minimal export of nitrate in stream water. Because it is diffi-
cult to differentiate among the many factors that could cause such different ecosys-
tem responses, Judd et al. (2011) suggested that other changes over the past decades
(e.g., acid rain impacts) may have combined to reduce the effects of frost events on
nitrogen fluxes (see also Bernal et al. 2012).
Initially (1964–1974), an inverse relationship between nitrate and sulfate con-
centration in stream water was observed, but the long-term record shows the oppo-
site (Fig. 29a; Likens et al. 1970). The relationship between nitrate and sulfate in
precipitation also has changed with time (Fig. 29b). Streamwater concentrations of
nitrate generally reach a maximum during the late winter (Fig. 30), whereas concen-
trations of sulfate at that time are at a minimum.
Streamwater Chemistry 73

a 70

1:1
Stream Water
60

1969
50
Nitrate (µEq/L)

40

30

20

10

0
0 20 40 60 80 100 120 140
Sulfate (µEq/L)
b 50
Precipitation
1:1

1965-1992 Data
40
1993-2009 Data
All Data 1965-2009
Nitrate (µEq/L)

30

20

X
10 1983
1964
2009
0
0 20 40 60 80 100
Sulfate (µEq/L)

Fig. 29 Relationship between annual volume-weighted nitrate and sulfate concentrations in

(a) stream water and (b) bulk precipitation of Watershed 6 of the Hubbard Brook Experimental
Forest during 1964–2009. The second-order regression for (a) is significant (p < 0.001; r2 0.48) and
for (b) the dashed line is significant (p<0.001; r2 0.59); regression for 1965–1992 is not significant,
but regression for 1993–2009 is significant (p<0.001; r2 0.89)

There is very poor correlation between the pattern for annual values of ammo-
nium and/or nitrate in precipitation and in stream water (Fig. 19b) range in
seasonal.
A final complicating factor is that some larger (untreated) watersheds show a
larger range in seasonal nitrate concentrations in stream water than smaller water-
sheds (see below; Likens and Buso 2006), but there are differences in flow paths
74 3 Chemistry

Fig. 30 Monthly volume-weighted concentrations of nitrate in bulk precipitation (a) and stream
water (b) for Watershed 6 of the Hubbard Brook Experimental Forest during the period
1964–2012

(seeps) and geology; even so, a causal relationship is not clear. In particular, W1
(the 1999 Wollastonite-treated watershed) has recently begun to export significantly
more nitrate, even during the dormant season, but an explanation for this response
is elusive. Surface seeps tend to have a much higher nitrate concentration than
nearby streams (Likens and Buso 2006).

Valleywide Streamwater Chemistry: A Synoptic View

In spring and fall of 2001, streamwater chemistry was measured at 100-m intervals
in all streams of the Hubbard Brook Valley (Fig. 31; Likens and Buso 2006). The
chemistry of streams varied throughout this large drainage network according to
elevation, channel length, size of area drained, and type of drainage. Relatively
large longitudinal differences were observed for elements like hydrogen ion, dis-
solved aluminum, and dissolved organic carbon throughout the drainage network,
whereas other elements like sodium, chloride, and dissolved silica changed little.
The chemistry of Hubbard Brook, a fifth-order river draining the valley, was
 2001 Valley-wide Survey of HBEF 2001 Valley-wide Survey of HBEF
Mt. Cushman 3
+ Mt. Cushman 3
12 +
2
6 5 4 5 41
6

Mirror
Lake Mirror
Lake

pH Key
Mt. Kineo + pH < 4.5 Calcium Key
pH 4.5 – 5.0 Mt. Kineo +
8 [Ca] < 0.5 mg/L
pH 5.0 – 5.5 8
7 0.5 – 1.0 mg/L
9 pH 5.5 – 6.0 7
9 1.0 – 2.0 mg/L
pH 6.0 – 6.5
2.0 – 3.0 mg/L
pH 6.5 – 7.0
Stream water pH Calcium [Ca] > 3.0 mg/L
pH > 7.0

2001 Valley-wide Survey of HBEF 2001 Valley-wide Survey of HBEF

3
Valleywide Streamwater Chemistry: A Synoptic View

3 Mt. Cushman
Mt. Cushman 2
2 + 6
+ 41
6 5 41 5

Mirror
Mirror
Lake
Lake

Sulfate Key Nitrate Key

[NO3] < 0.01 mg/L
[SO4] < 4.0 mg/L Mt. Kineo +
Mt. Kineo + 0.01 – 0.10 mg/L
4.0 – 5.0 mg/L
8 0.10 – 0.50 mg/L
8 5.0 – 6.0 mg/L 7
7 9 [NO3] > 0.50 mg/L
9 [SO4] > 6.0 mg/L
Sulfate Nitrate
75

Fig. 31 Streamwater concentrations for pH, calcium, sulfate, and nitrate throughout the Hubbard Brook Valley (Adapted from Likens and Buso 2006)
76 3 Chemistry

Fig. 32 Distribution of pH 7.0

concentrations along the WEIR 8
longitudinal distance for Bear
Brook (W6), Canyon Brook 6.5
(W8), Paradise Brook (W3),
and Cascade Brook (W9) of
6.0 WEIR 6
the Hubbard Brook
Experimental Forest.
Gauging stations are marked
with arrows. The third-order 5.5

pH
polynomial regressions have WEIR 3
r2 of 0.97 for Bear Brook
5.0
(W6), 0.97 for Paradise
Brook, 0.94 for Canyon
Brook, and 0.93 for Cascade 4.5
Brook (modified from Likens
and Buso 2006)
WEIR 9
4.0
Confluence First
with main Running
Hubbard Brook Water
3.5
100 80 60 40 20 0
Percent of Wetted Channel above Sample Site (%)

remarkably constant throughout its length in spite of different chemistry of inflow-

ing tributaries (Likens and Buso 2006). Such synoptic view of streamwater chemis-
try at close intervals throughout and entire, large drainage system is unusual, but
quite informative, as the larger scale adds new insights about patterns and dynamics
of streamwater chemistry. For example, using streamwater pH as a critical diagnos-
tic biogeochemical variable in the Hubbard Brook Valley, we learned that it was
highly fortunate that the gauging weirs in the experimental watersheds had been
sited at or near the pH inflection points along the longitudinal distance of the head-
water streams being studied (Fig. 32).
As described elsewhere in this book, streamwater samples are collected routinely
for chemical analysis from fixed locations above the gauging weirs. This approach
is based on the assumption that the chemistry from these sites integrates the diverse
biogeochemical processes occurring throughout the watershed-ecosystem above the
sampling point (Bormann and Likens 1967; Likens and Buso 2006; Chap. 1). This
synoptic study helped to clarify this assumption. Moreover, this study helped us to
understand the temporal and spatial dynamics within the entire drainage network
(Likens and Buso 2006). Likens and Buso (2006) review several previous, smaller-
scale attempts to evaluate longitudinal/elevational changes in streamwater chemis-
try, but the effort in 2001 was the first, high-density, valleywide, synoptic study at
HBEF.
The gauging stations had been situated in the HBEF on the basis of hydrologic,
not biogeochemical, considerations, primarily whether bedrock was exposed for
Seasonal Variations in Bulk Precipitation and Streamwater Concentrations 77

attachment of the weir to prevent significant leakage of water around or under the
gauging station. As it turns out, these locations are close to optimal for measuring
stream chemistry because change is most obvious and pronounced at the inflection
points along the stream channel in the longitudinal distribution of data (Fig. 32).
A variety of factors influence streamwater chemistry including soils, till, bed-
rock, vegetation, and hydrologic flow paths. For example, almost all tributaries had
obvious pH trends along their longitudinal channels until they became fourth- to
fifth-order streams. In contrast, the pattern of pH in Hubbard Brook was subdued,
rather constant and quite different than the headwater streams (Fig. 31). As such, it
is much more interesting and informative to study the biogeochemistry of the head-
water streams. Moreover, the unprotected nature of Hubbard Brook in the vicinity
of Mirror Lake leads to higher concentrations of various pollutants, e.g., road salt
(Likens and Buso 2010b).

Seasonal Variations in Bulk Precipitation

and Streamwater Concentrations

As noted elsewhere, climate and biologic activity at HBEF are distinctly seasonal,
and much of the biogeochemistry reflects these climatic patterns. Normally the
snowpack persists from about mid-November until mid-April (Table 2). The trees
start to produce leaves during May, with full canopy closure by mid- to late May,
and major leaf fall occurs in mid-October. These phenological events are chang-
ing, however, because of climate change (Campbell et al. 2007; Richardson et al.
2006; Hayhoe et al. 2007; Huntington et al. 2009; Vadeboncoeur et al. 2009;
Groffman et al. 2012; Hamburg et al. 2013). Some organisms of the ecosystem,
such as stream algae, have adapted to utilize the favorable environmental condi-
tions of the transition periods between seasons (see Muller and Bormann 1976;
Bernhardt et al. 2005).
The year has been divided into four seasons to investigate the short-term varia-
tions in chemistry of bulk precipitation and stream water: summer from 1 June to 30
September, autumn from 1 October to 30 November, winter from 1 December to 28
February, and spring from 1 March to 31 May. Although these periods are arbitrary,
they correspond well with actual observed conditions and biologic activity at HBEF.
Given the strong temporal trends in bulk precipitation and streamwater chemistry
that have emerged during this long-term study, the seasonal data are presented here
as means for both the entire period (1963–2009; Table 12) and for two decades
(based on water-years): one early in the study 1964–1973, and one late, 2000–2009
(Tables 13–16).
Fall precipitation (+34 %) and especially streamflow (+71 %) have increased
between the decades (Tables 13 and 15). It is unknown how ecosystem water stor-
age has changed during this period. The weighted concentrations of various ions in
bulk precipitation are highly variable on a storm or weekly basis, but several
78 3 Chemistry

Table 12 Seasonal volume-weighted average (a) bulk precipitation and (b) streamwater
concentrations for entire period (1963–2009) for Watershed 6 of the Hubbard Brook Experimental
Forest
Summer Fall Winter Spring
June– October– December–
Obs September November February March–May
Years mg/L µEq/L mg/L µEq/L mg/L µEq/L mg/L µEq/L
(a) Bulk precipitation concentration
Calcium 47 0.08 4.0 0.08 4.0 0.06 3.0 0.14 7.0
Magnesium 47 0.02 1.6 0.03 2.5 0.02 1.6 0.04 3.3
Potassium 47 0.05 1.3 0.07 1.8 0.03 0.8 0.06 1.5
Sodium 47 0.05 2.2 0.14 6.1 0.11 4.8 0.11 4.8
Aluminuma 34 0.00 0.0 0.00 0.0 0.00 0.0 0.00 0.0
Ammonium 46 0.217 12.0 0.151 8.4 0.119 6.6 0.244 13.5
Hydrogen ionb 47 0.061 60.5 0.045 44.6 0.039 38.7 0.054 53.6
pH 4.22 4.35 4.41 4.27
Sulfate 46 2.62 54.5 1.59 33.1 1.20 25.0 2.19 45.6
Nitrate 46 1.27 20.5 1.33 21.4 1.44 23.3 1.60 25.8
Chloride 46 0.18 5.1 0.30 8.5 0.25 7.1 0.26 7.3
Phosphate 39 0.011 0.3 0.007 0.2 0.007 0.2 0.021 0.7
DOCc 15 1.3 0.0 0.6 0.0 0.4 0.0 0.8 0.0
Dissolved 38 0.00 NA 0.00 NA 0.00 NA 0.00 NA
silicaa
Totals 5.81 (mg/L) 4.36 (mg/L) 3.70 (mg/L) 5.54 (mg/L)
Cations 81.6 (µEq/L) 67.3 (µEq/L) 55.5 (µEq/L) 83.7 (µEq/L)
Anions 80.4 (µEq/L) 63.1 (µEq/L) 55.5 (µEq/L) 79.4 (µEq/L)
(b) Streamwater concentration
Calcium 47 0.91 45.4 0.94 46.9 0.98 48.9 0.96 47.9
Magnesium 47 0.24 19.8 0.26 21.4 0.27 22.2 0.25 20.6
Potassium 47 0.11 2.8 0.21 5.4 0.19 4.9 0.23 5.9
Sodium 47 0.76 33.1 0.77 33.5 0.77 33.5 0.66 28.7
Aluminumd 40 0.27 20.3 0.29 21.8 0.30 22.5 0.33 24.8
Ammonium 46 0.020 1.1 0.023 1.3 0.017 0.9 0.016 0.9
Hydrogen ionb 47 0.011 11.3 0.010 10.0 0.011 10.8 0.013 12.4
pH 4.95 5.00 4.97 4.91
Sulfate 46 4.94 102.8 5.01 104.2 4.82 100.3 4.75 98.8
Nitrate 46 0.17 2.7 0.27 4.4 1.08 17.5 1.25 20.2
Chloride 46 0.39 11.0 0.50 14.1 0.48 13.5 0.44 12.4
Phosphate 38 0.003 0.1 0.002 0.1 0.003 0.1 0.003 0.1
DOCe 15 2.8 16.9 2.5 14.9 1.9 11.6 2.0 12.1
Dissolved 46 1.94 NA 1.95 NA 1.88 NA 1.63 NA
silica
Totals 12.57 (mg/L) 12.72 (mg/L) 12.74 (mg/L) 12.55 (mg/L)
Cations 133.7 (µEq/L) 140.2 (µEq/L) 143.7 (µEq/L) 141.1 (µEq/L)
Anions 133.4 (µEq/L) 137.7 (µEq/L) 143.0 (µEq/L) 143.6 (µEq/L)
a
Aluminum and dissolved silica are trace in precipitation
b
Hydrogen ion in µg/L VWA-concentration
c
DOC has no charge in precipitation
d
Aluminum valence assumed to be +2 (Buso et al. 2000)
e
Dissolved organic carbon at 6 µEq/mg-C
Table 13 Seasonal volume-weighted average concentrations in bulk precipitation for Watershed 6 of the Hubbard Brook Experimental Forest for decades 1964–1973
and 2000–2009
Decade
(10-year Precip Ca Mg K Na Al NH4 H+ SO4 NO3 Cl PO4 dSi
Season block) (mm/ha) (mg/L) (mg/L) (mg/L) (mg/L) (mg/L)a (mg/L) (µEq/L) (mg/L) (mg/L) (mg/L) (mg/L)b (mg/L)a
Summer 1964–1973 120.5 0.14 0.04 0.06 0.08 0.00 0.25 0.1 3.75 1.30 0.38 0.013 0.00
2000–2009 128.3 0.06 0.02 0.04 0.05 0.00 0.19 0.0 1.55 0.92 0.10 0.005 0.00
Fall 1964–1973 114.9 0.16 0.07 0.08 0.21 0.00 0.20 0.1 2.43 1.51 0.50 0.009 0.00
2000–2009 153.5 0.05 0.02 0.04 0.09 0.00 0.12 0.0 0.95 0.93 0.19 0.002 0.00
Winter 1964–1973 114.9 0.10 0.03 0.03 0.11 0.00 0.15 0.0 1.64 1.44 0.32 0.010 0.00
2000–2009 108.2 0.05 0.02 0.02 0.11 0.00 0.10 0.0 0.80 1.17 0.21 0.002 0.00
Spring 1964–1973 110.5 0.21 0.05 0.06 0.12 0.00 0.26 0.1 3.08 1.74 0.36 0.034 0.00
2000–2009 114.4 0.12 0.03 0.06 0.06 0.00 0.23 0.0 1.37 1.21 0.16 0.009 0.00
Years of observations (n) 47 47 47 47 47 34 46 47 46 46 46 39 38
Years are water-years: 1 June–31 May
a
Aluminum and dissolved silica (dSi) are trace only in precipitation
b
Phosphate data N/A in 1964–1968; early data based on 1971–1980 decade
80 3 Chemistry

Table 14 Seasonal water and volume-weighted average concentration changes in bulk

precipitation for Watershed 6 of the Hubbard Brook Experimental Forest between decades 1964–
1973 and 2000–2009
Past to
present Precip Ca Mg K Na Ala NH4 H+ SO4 NO3 Cl PO4 dSia
Summer %Change 6 −54 −57 −29 −43 NA −25 −65 −59 −29 −74 −59 NA
Fall %Change 34 −71 −74 −48 −54 NA −37 −65 −61 −38 −62 −78 NA
Winter %Change −6 −46 −46 −39 3 NA −33 −46 −51 −18 −35 −83 NA
Spring %Change 4 −43 −39 −9 −54 NA −12 −63 −56 −30 −55 −75 NA
a
Aluminum and dissolved silica (dSi) are trace only in precipitation

seasonal trends have emerged. In general, average cation concentrations are mostly
lower during winter and somewhat higher during spring; sulfate concentrations are
lowest in winter, but nitrate and chloride are lowest during summer. Highest average
sulfate concentrations occur during summer, nitrate in spring, and chloride in fall
(Table 12a; Fig. 33).
Seasonal bulk precipitation chemistry concentrations are much lower in the
recent pentad (Fig. 33) reflecting the long-term trends in annual bulk precipitation
chemistry (see Fig. 19). The total ionic concentration of bulk precipitation during
the winter at HBEF is about 68 % that of the spring season of the year (Table 12a).
Currently, this trend is reflected in the long-term concentrations, particularly of cal-
cium, magnesium, hydrogen ion, chloride, and sulfate. Early in the record, a peak in
concentration for sodium, potassium, and magnesium occurred in the autumn, but
this increase has disappeared in the more recent pentad (Fig. 33; Table 13). As
expected, all base cations in bulk precipitation have decreased in all seasons, except
for sodium in the winter (+3 % increase). Inexplicitly, there is no similar increase in
winter chloride. Annual patterns for many of the elements may result in part from
the origin of air masses and in part from lower aerosol scavenging efficiency by
raindrops as compared to snowflakes. Herman and Gorham (1957) suggested from
studies in Nova Scotia that snowflakes may be less efficient than raindrops in remov-
ing materials from the atmosphere. Nitrate concentrations in bulk precipitation tend
to increase in winter and spring, whereas sulfate and ammonium tend to increase in
spring and summer (Fig. 33; Tables 12a and 13). Higher sulfate concentrations in
the spring and summer are usually attributed to more rapid oxidation of SO2 to sul-
fate in the warmer months. Likewise, SO2 concentrations in the atmosphere are
usually higher in the winter because SO2 is not being converted to sulfate as rapidly.
Ammonium is higher in the spring and summer because there are greater NH3 emis-
sions from increased volatilization (followed by conversion to ammonium) from the
soil during these warmer months.
Other unexplained variations include that calcium equivalent concentrations in
bulk precipitation were much greater than (almost double) sodium values during the
spring and summer but are less than sodium during the autumn and winter during
1964–1973 (Table 13). The relatively high concentrations of sodium and chloride in
autumn rains relative to other seasons may reflect a larger influence from marine
aerosols from coastal storms at this time in the eastern USA (e.g., Likens et al. 1990b).
Table 15 Seasonal volume-weighted average concentrations in stream water for Watershed 6 of the Hubbard Brook Experimental Forest between decades
1964–1973 and 2000–2009
Decade
(10-year Flow Ca Mg K Na Al NH4 H+ SO4 NO3 Cl PO4 dSi
Season block) (mm/ha) (mg/L) (mg/L) (mg/L) (mg/L) (mg/L)a (mg/L) (µEq/L) (mg/L) (mg/L) (mg/L) (mg/L)b (mg/L)
Summer 1964– 31.5 1.26 0.31 0.15 0.82 0.19 0.045 0.0 6.37 0.42 0.47 0.004 4.45
1973
2000– 42.6 0.85 0.26 0.12 0.98 0.26 0.007 0.0 4.88 0.02 0.45 0.002 5.28
2009
Fall 1964– 63.6 1.37 0.37 0.23 0.90 0.34 0.074 0.0 6.50 0.70 0.60 0.004 4.77
1973
2000– 108.9 0.62 0.20 0.19 0.71 0.24 0.005 0.0 3.65 0.04 0.43 0.001 3.74
2009
Winter 1964– 52.3 1.43 0.37 0.22 0.85 0.27 0.037 0.0 5.92 2.21 0.56 0.002 4.29
1973
2000– 67.7 0.63 0.20 0.14 0.72 0.21 0.005 0.0 3.69 0.21 0.41 0.001 3.87
2009
Spring 1964– 157.2 1.36 0.33 0.25 0.71 0.35 0.026 0.0 5.76 2.29 0.52 0.003 3.50
1973
2000– 138.5 0.61 0.19 0.19 0.19 0.24 0.005 0.0 3.52 0.26 0.36 0.001 3.44
2009
Years of observations (n) 47 47 47 47 47 40 46 47 46 46 46 38 46
Seasonal Variations in Bulk Precipitation and Streamwater Concentrations

Years are water-years: 1 June–31 May

dSi dissolved silicate
a
Aluminum data N/A 1971–1975; data from 1964–1970 (7-years) and 1976–1978 (3-years) used for early decade
b
Phosphate data N/A in 1964–1971: early data based on 1972–1983 decade
81
82 3 Chemistry

Table 16 Seasonal water and volume-weighted average concentration changes in stream water for
Watershed 6 of the Hubbard Brook Experimental Forest between decades 1964–1973 and
2000–2009
Past to
present Flow Ca Mg K Na Al NH4 H+ SO4 NO3 Cl PO4 dSi
Summer %Change 35 −32 −18 −19 19 36 −84 −27 −23 −95 −5 −54 19
Fall %Change 71 −55 −48 −18 −22 −31 −93 −36 −44 −95 −29 −78 −22
Winter %Change 29 −56 −45 −34 −15 −22 −87 −51 −38 −91 −26 −69 −10
Spring %Change −12 −55 −43 −26 −74 −31 −80 −48 −39 −89 −31 −66 −2
dSi dissolved silica

Overall, large declines in average concentrations occurred for almost all solutes
between 1964–1973 and 2000–2009 decades, especially for calcium, hydrogen ion,
sulfate, nitrate, and chloride. The least amount of change occurred between winter
and the largest change occurred between fall decades (Tables 13 and 14). The fall
precipitation volume increased by 63 % between these decades.
Concentrations of most ions in stream water are much less variable seasonally
than in bulk precipitation (Figs. 34 and 35; Table 12b). Like seasonal precipitation
values, seasonal streamwater concentrations are consistently lower in the recent
pentad than the earlier one (Figs. 34 and 35; Table 15). Concentrations in stream
water are slightly lower in the summer, in part because of the intense biologic utili-
zation (uptake) of potassium and nitrate during the growing season (Fig. 34). Both
dissolved silica and sodium concentrations in stream water have increased in sum-
mer between decades (Tables 15 and 16). Is this because of deeper flow paths?
In general, nitrate concentrations in stream water begin to rise in the autumn
when there is a reduction in biologic activity and attain a maximum concentration
in the winter or early spring (Figs. 30 and 34; Likens 2004). As streamwater nitrate
concentrations have declined with time (Fig. 28), the seasonal pattern also has
become more subdued (Figs. 30 and 34). US Geological Survey Water Resources
data show a similar pattern for rivers and streams throughout New England and New
York State. This pattern also has been observed in the surface waters of Mirror Lake
(Buso et al. 2009, p. 122) and in other freshwater lakes where nitrate concentrations
reach maximum levels in winter and early spring. The increase in non-anthropogenic
nitrate concentrations in lakes has been attributed to increased nitrification during
the winter (Hutchinson 1957, p. 871), but at HBEF the early spring maximum may
be the result of decreased plant uptake by terrestrial deciduous vegetation and
increased soil production (and retention) during winter (e.g., Judd et al. 2007).
At HBEF, we initially observed that nitrate concentrations increased markedly
with increased streamflow (Johnson et al. 1969). However, this may have been a
fortuitous relationship since the late winter–early spring period of maximum
streamflow is potentially also a time for relatively high nitrification in the soil (see
Judd et al. 2007). With time and changing conditions, e.g., reduced winter snow-
pack, however, this relationship has all but disappeared. Even though there may be
no causal relationship between high nitrate concentration and increased streamflow,
the fact that the late winter–early spring period of high nitrate concentrations also is
Seasonal Variations in Bulk Precipitation and Streamwater Concentrations 83

0.6
Calcium 1964-68 Pentad
0.4
mg/L

2005-09 Pentad

0.2

0.0
mg/L

0.2 Magnesium
0.1
0.0
0.2 Potassium
mg/L

0.1
0.0
0.6 Sodium
mg/L

0.4
0.2
0.0
mg/L

0.4 Ammonium
0.2
0.0
150
µEq/L

Hydrogen Ion
100
50
0
5
4 Sulfate
mg/L

3
2
1
0
2.0 Nitrate
mg/L

1.5
1.0
0.5
0.0
0.6 Chloride
mg/L

0.4
0.2
0.0
Summer Fall Winter Spring
Season
Fig. 33 Mean seasonal bulk precipitation concentrations ±1 SD for calcium, magnesium, potas-
sium, sodium; ammonium, hydrogen ion, sulfate, nitrate, and chloride during two pentads, 1964–
1968 and 2005–2009, in Watershed 6 of the Hubbard Brook Experimental Forest
84 3 Chemistry

2.0
1.5 Calcium
mg/L

1.0
0.5
0.0
0.4
mg/L

0.2
0.0
Magnesium
Potassium
mg/L

0.4
0.2
0.0
1.5
mg/L

1.0
0.5
0.0
Sodium 1964-68 Pentad
0.4 2005-09 Pentad
Aluminum
mg/L

0.3
0.2
0.1
0.0
20 Hydrogen Ion
µEq/L

15
10
5
0
6
mg/L

4
2
Sulfate
0 1964-68 Pentad
4 2005-09 Pentad
Nitrate
mg/L

3 1969-73 Pentad
2
1
0
0.6
mg/L

0.4
0.2 Chloride
0.0
Summer Fall Winter Spring
Season
Fig. 34 Mean seasonal streamwater concentrations ±1 SD for calcium, magnesium, potassium,
sodium; aluminum, hydrogen ion, sulfate, nitrate, and chloride during two pentads, 1964–1968 and
2005–2009, in Watershed 6 of the Hubbard Brook Experimental Forest
Long-Term Trends in Bulk Precipitation and Streamwater Chemistry 85

0.25
0.20 Ammonium
1964-69 Pentad
0.15
mg/L
2005-09 Pentad
0.10
0.05
MDL
0.00

0.003
Phosphate
mg/L

0.002
MDL
0.001
0.000

6
mg/L

2 Dissolved Silicate
0
Summer Fall Winter Spring
Season

Fig. 35 Mean seasonal streamwater concentrations ±1 SD for dissolved silica, ammonium, and
phosphate during two pentads, 1964–1968 and 2005–2009, in Watershed 6 of the Hubbard Brook
Experimental Forest

characterized by large discharge remains of prime importance to nutrient budget

calculations.
The seasonal pattern for ammonium concentrations in stream water is quite inter-
esting. During the early pentad (1964–1968; Fig. 35; Table 15), there was a clear
peak in ammonium concentration during the autumn, which probably reflects
increased decomposition of organic matter in the forest floor after leaf fall. In the
current period (Table 15), however, streamwater concentrations of ammonium con-
sistently are very low throughout the year and show no distinct seasonal pattern
(Fig. 35). Most solutes have an individual pattern of seasonal change between the
two decades (Tables 15 and 16). Patterns are not necessarily in synchrony for bulk
precipitation and stream water, raising many intriguing biogeochemical questions.

Long-Term Trends in Bulk Precipitation

and Streamwater Chemistry

Longer-term biogeochemical records, now spanning some 49 years, have revealed

some remarkable trends in the chemistry of bulk precipitation and stream water at
HBEF. Notably, annual volume-weighted concentrations of hydrogen ion, sulfate,
and calcium in bulk precipitation decreased by 80 %, 75 %, and 66 %, respectively,
86 3 Chemistry

since 1963 or 1964 (Fig. 19). Since the amount of annual precipitation increased
only 14 % during this period (Fig. 19), the annual inputs of these ions also
decreased significantly (see Chap. 4; Likens 1992; Likens and Buso 2012). The
decline of acid anions and base cations in surface waters recorded at HBEF since
1963–1964 has also been found more widely in the eastern USA and in northwest-
ern Europe, although the length of record at these other locations is shorter (e.g.,
Hedin et al. 1994; Watmough et al. 2005; Clair et al. 2010; Navrátil et al. 2010).
Most of the decline in base cations at HBEF was due to a decline in calcium
concentrations, and most of this decline occurred prior to about 1975 (Figs. 18
and 19a). By 1969 it appeared that the annual nitrate concentration in bulk precipi-
tation was on an upward trend (Fig. 19b), but the longer-term record indicates that
between about 1972 and 2002 nitrate concentrations were relatively flat, and now
concentrations have declined and are similar or lower than values in the mid-1960s
(Figs. 19b, 28, and 30).
The long-term record for concentration of hydrogen ion in bulk precipitation is
particularly interesting (Fig. 14). Clearly the average precipitation at HBEF is not as
acidic today as it was in the 1960s, but it required 18 years of measurement to fit a
statistically significant linear regression (p < 0.05) to these temporal data (Likens
1989). A close examination of this long-term record, however, shows periods when
several years of data give contrasting trends (Fig. 14). Obviously, short-term
(3–6 years) records can be misleading relative to the overall trend. This is an impor-
tant finding not only for interpreting results from the HBES, but also for general
ecological studies and state and national monitoring programs of management rel-
evance (see Lindenmayer and Likens 2010).
Coincident with the declines in annual volume-weighted concentrations in bulk
precipitation were declines in the chemistry of stream water at HBEF (Table 10;
Likens and Buso 2012). Average calcium, sulfate, and hydrogen ion concentrations
declined by 64 %, 52 %, and 59 %, respectively, during 1963 or 1964–2009 (Fig. 19).
Calcium concentrations in stream water actually increased from 1963–1964 to
1969–1970 and then decreased steadily thereafter. The decline in annual sum of
base cations and sulfate concentrations in stream water was closely correlated with
the decline in atmospheric inputs of gaseous and particulate pollutants (e.g., Driscoll
et al. 1989a, b; Likens 1992; Likens et al. 1998, 2002a, 2005). The decrease in
annual concentration of hydrogen ion in stream water was relatively small in the
long-term record (Fig. 19b), primarily because the decline in sum of base cations
was paralleled by the decline in sulfate concentration (Fig. 19a; Driscoll et al.
1989a, b).
It has been suggested that a few streamwater samples can serve as a biogeo-
chemical baseline for a terrestrial ecosystem. This assumption is tenuous at best
given the long-term data presented in this chapter. For example, from a monthly
sample effort over 18 years, Palmer et al. (2004) did not detect the long-term trends
in streamwater chemistry we observe from our systematic, weekly sampling extend-
ing over 48 years in the same watershed. Moreover, the marked and often unex-
plainable changes in the intensively sampled stream water of the HBEF suggest
caution in attributing baseline characteristics to a few samples that unwittingly may
Long-Term Trends in Bulk Precipitation and Streamwater Chemistry 87

characterize a high or low period in the history of the site. It would be misleading to
generate different conclusions about the biogeochemistry of W6 from two different
decadal-long data sets for precipitation and streamwater chemistry separated by
nearly three decades. Clearly, the most useful biogeochemical baseline is one of
sufficient length to allow trend analyses, and even then the baseline may be elusive
(Likens and Buso 2012).
Chapter 4
Input–Output Budgets

Because our ecosystems are watersheds, the flux of chemicals into the system may
be simplified to that in atmospheric deposition (meteorologic vector) and the output
to that in drainage waters (geologic vector), at least for those nutrients without a
prominent gaseous phase (Bormann and Likens 1967; Likens and Bormann 1972).
Theoretically, then, the difference between annual hydrologic input (I) and output
(O), the Net Hydrologic Flux (NHF) for a given chemical constituent, shows whether
that constituent is being accumulated within the ecosystem, I > O; is being lost from
the system, I < O; or is quantitatively passing through the system, I = O. By this defi-
nition, the release of weathering products does not represent an input to the ecosys-
tem per se; instead, weathering is considered as a process internal to the ecosystem
and a part of the intrasystem cycle (Fig. 1; Bormann and Likens 1967; Likens and
Bormann 1972). Likewise, uptake and long-term storage of nutrients by plants are
internal ecosystem processes and not outputs, per se. On the other hand, complete
mass-balance calculations or NEF (Net Ecosystem Flux), on the time frame of the
HBES (about 50 years), considers annual weathering release or net soil release as an
input and annual biomass storage as an output (see below and Chap. 6).
Explicit boundaries for an ecosystem (a watershed in our case) are necessary to
make quantitative measures of flux when determining input–output budgets.
Because ecosystem boundaries usually are determined by the investigator to make
these quantitative measures of flux, rather than on the basis of some known or obvi-
ous functional discontinuity, between ecosystems, such as the shoreline of a lake,
has raised criticism about this ecosystem approach (e.g., see Likens 1998;
Fitzsimmons 1996; Currie 2011). Nevertheless, the theoretical and methodological
components and constraints for establishing ecosystem boundaries are given in
Bormann and Likens (1967, 1979); Likens and Bormann (1972); Likens (1975,
1992, 2001); Likens and Bormann (1985); and Wiens et al. (1985). Using watershed
boundaries, when topographic and phreatic boundaries are the same and when the
geologic substrate is relatively watertight, as at HBEF, alleviates this problem as
water is the primary transport vector in humid areas (Bormann and Likens 1967;
Likens and Bormann 1972).

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2_4, 89

© Springer Science+Business Media New York 2013
90 4 Input–Output Budgets

Mass Meteorologic Input of Dissolved Substances

The long-term average annual input of dissolved inorganic substances in rain and
snow was calculated to be 23 kg/ha or 2,081 Eq/ha during 1963–2009 (see Tables 17
and 19). Characterizing the “dilution effect” (see above), the atmospheric input of
total solutes during an early wet year (1973–1974) was 3,377 Eq/ha-year, whereas
the input during a similar wet year in 1995–1996 (6 % less water input) was
2,390 Eq/ha (29 % less than the solute input in 1973–1974). Nevertheless, even with
this gradual “dilution effect,” the input of chemicals in bulk precipitation to these
forested ecosystems is substantial, both as nutrients and as weathering agents.

Annual Variation in Net Hydrologic Budgets

There now are, depending on the chemical constituent, almost 50 years of quantita-
tive annual net hydrologic budgets (NHF) for precipitation (P) input and streamwa-
ter (S) output at the HBEF. Over the long term, there has been considerable change
but relatively small uncertainty in mean annual budgets for most of the major ions
(Tables 17 and 18). Comparing medians, there has been hydrologic net loss (S > P)
of calcium (6.9 times), magnesium (6.5 times), potassium (2.9 times), sodium (4.9
times), sulfate (1.6 times), chloride (1.3 times), DOC (1.9 times), aluminum,
and dissolved silica and net gain (P > S) of ammonium (21.6 times), nitrate (4.3
times), hydrogen ion (7.1 times), and phosphate (6.6 times) from these watershed-
ecosystems (Tables 17 and 18; Likens 2004). The long-term averages show a small
net gain in phosphate and a small net loss of chloride each year, but there is no sta-
tistical difference (p > 0.05) between mean annual inputs in bulk precipitation and
streamwater outputs (Table 18).
There are two separate considerations in evaluating annual hydrologic input
minus output budgets: (1) the direction of the net change, whether input is greater
or less than output, and (2) the magnitude of the net change, i.e., the difference in
amount between input and output. These parameters may vary from year to year for
the individual chemical elements (Figs. 36a–c and 37). For some elements, data
from a single annual cycle (n = 1) are sufficient to establish the direction of net
change; for others, data from several years (n > 1) are necessary. Similarly, a vari-
able number of years are required to characterize the magnitude of net change with
meaningful confidence limits. Based on almost 50 years of study at the HBEF, it is
possible to make some generalizations about the annual budgets for the various
ions. Generally, the direction of net change for budgets of bulk precipitation input
minus streamwater output for calcium, magnesium, sodium, aluminum, ammo-
nium, hydrogen ion, sulfate, and dissolved silica was predictable from only a few
years (1–3) of data (Table 17; Fig. 36a–c). For these ions, the direction of the budget
difference is not in doubt, but the magnitude may be. For potassium, nitrate, phos-
phate, and chloride, one or a few years were likely to give unreliable results for both
Annual Variation in Net Hydrologic Budgets 91

Table 17 Inputs in bulk precipitation (IN) and outputs in stream water (OUT) for Watershed 6 of
the Hubbard Brook Experimental Forest during 1963–2009
Ca Ca Ca Mg Mg Mg K K K
Water-year IN OUT NET IN OUT NET IN OUT NET
1963 3.0 10.7 −7.7 0.7 2.4 −1.6 2.3 1.8 0.6
1964 3.1 4.8 −1.7 1.2 1.8 −0.6 1.4 1.1 0.3
1965 2.6 10.4 −7.7 0.6 2.7 −2.1 0.6 1.3 −0.8
1966 2.1 10.8 −8.7 0.4 2.9 −2.5 0.6 1.7 −1.1
1967 2.8 12.2 −9.4 0.9 3.5 −2.6 1.0 2.4 −1.4
1968 1.6 11.7 −10.1 0.4 3.0 −2.6 0.7 2.3 −1.6
1969 2.2 14.1 −12.0 0.4 3.2 −2.8 0.8 2.3 −1.5
1970 1.9 12.1 −10.1 0.6 2.8 −2.3 1.2 1.7 −0.6
1971 1.2 10.7 −9.4 0.4 2.5 −2.1 0.3 1.7 −1.4
1972 1.1 14.0 −12.8 0.5 3.2 −2.7 0.6 2.4 −1.8
1973 2.0 19.0 −17.0 0.4 4.3 −3.9 0.8 3.1 −2.3
1974 1.2 11.2 −10.0 0.4 2.8 −2.4 0.5 2.1 −1.6
1975 1.2 15.2 −14.0 0.3 3.6 −3.3 0.5 2.3 −1.8
1976 1.3 10.1 −8.8 0.3 2.5 −2.2 0.4 1.9 −1.5
1977 1.1 10.9 −9.8 0.4 2.6 −2.2 0.7 1.8 −1.2
1978 0.8 8.9 −8.1 0.3 2.2 −1.8 0.6 1.8 −1.2
1979 1.1 8.2 −7.0 0.4 2.0 −1.7 0.9 2.1 −1.2
1980 1.3 8.7 −7.4 0.5 2.3 −1.8 0.8 2.0 −1.2
1981 1.1 10.0 −8.9 0.3 2.5 −2.3 0.6 1.8 −1.2
1982 0.9 7.7 −6.8 0.3 2.1 −1.7 0.7 1.7 −1.0
1983 1.1 8.7 −7.6 0.3 2.2 −1.9 0.7 2.3 −1.6
1984 1.0 6.2 −5.2 0.3 1.7 −1.4 0.6 1.5 −0.9
1985 0.9 8.0 −7.1 0.3 2.2 −1.9 0.4 1.9 −1.5
1986 0.6 7.6 −7.1 0.2 2.1 −1.9 0.3 2.1 −1.8
1987 1.1 7.1 −6.0 0.3 2.1 −1.8 0.7 1.4 −0.7
1988 1.1 6.9 −5.7 0.3 2.0 −1.7 0.5 1.9 −1.4
1989 1.4 10.1 −8.7 0.4 2.7 −2.3 0.9 2.4 −1.5
1990 1.2 9.8 −8.5 0.4 2.7 −2.3 0.8 1.6 −0.8
1991 1.1 7.4 −6.3 0.3 2.1 −1.7 0.6 1.4 −0.8
1992 0.9 7.0 −6.1 0.3 2.0 −1.7 0.6 1.5 −0.9
1993 1.2 7.3 −6.0 0.3 2.0 −1.7 0.6 1.4 −0.8
1994 1.0 5.4 −4.4 0.2 1.6 −1.3 0.6 1.0 −0.4
1995 1.3 9.7 −8.4 0.5 2.7 −2.2 0.7 2.6 −1.9
1996 1.1 7.8 −6.7 0.4 2.2 −1.9 0.6 1.9 −1.3
1997 1.0 5.6 −4.5 0.3 1.7 −1.4 0.8 1.5 −0.7
1998 1.5 8.3 −6.7 0.3 2.4 −2.0 0.7 2.0 −1.4
1999 1.7 8.1 −6.4 0.5 2.3 −1.8 0.9 2.9 −1.9
2000 0.8 4.9 −4.1 0.2 1.5 −1.3 0.5 1.6 −1.1
2001 1.2 5.3 −4.1 0.3 1.6 −1.3 0.6 1.4 −0.8
2002 1.1 4.4 −3.4 0.2 1.4 −1.1 0.6 1.2 −0.6
2003 1.2 7.7 −6.5 0.4 2.2 −1.8 0.6 1.9 −1.3
2004 0.8 6.1 −5.3 0.3 1.9 −1.6 0.5 1.6 −1.1
2005 1.2 7.6 −6.5 0.3 2.3 −2.0 0.8 1.8 −1.0
(continued)
92 4 Input–Output Budgets

Table 17 (continued)
Ca Ca Ca Mg Mg Mg K K K
Water-year IN OUT NET IN OUT NET IN OUT NET
2006 0.8 6.2 −5.4 0.2 2.0 −1.7 0.7 2.0 −1.3
2007 1.3 6.5 −5.2 0.3 2.1 −1.8 0.5 1.8 −1.3
2008 1.0 6.6 −5.6 0.3 2.1 −1.8 0.6 1.4 −0.8
2009 1.0 6.8 −5.8 0.3 2.2 −1.9 0.5 1.7 −1.2
SUM 63.4 414.5 −351.1 18.2 111.0 −92.8 32.9 87.1 −54.2
AVG 1.3 8.8 −7.5 0.4 2.4 −2.0 0.7 1.9 −1.2
±SD 0.6 3.0 2.8 0.2 0.6 0.5 0.3 0.4 0.5
Na Na Na Al Al Al NH4 NH4 NH4
Water-year IN OUT NET IN OUT NET IN OUT NET
1963 1.0 5.4 −4.3 0.0 0.9 −0.9 na na na
1964 2.2 4.1 −1.9 0.0 1.4 −1.4 2.0 0.29 1.8
1965 1.9 6.3 −4.4 0.0 2.4 −2.4 2.7 0.94 1.7
1966 1.2 6.7 −5.4 0.0 2.8 −2.8 2.4 0.37 2.0
1967 2.2 8.7 −6.5 0.0 3.0 −3.0 3.1 0.24 2.8
1968 1.4 6.9 −5.6 0.0 2.7 −2.7 3.3 0.13 3.2
1969 1.8 7.1 −5.3 0.0 3.2 −3.2 3.2 0.39 2.8
1970 1.4 6.3 −4.9 0.0 na na 4.6 0.23 4.4
1971 1.4 5.6 −4.2 0.0 na na 2.9 0.07 2.8
1972 1.5 7.9 −6.5 0.0 na na 2.3 0.21 2.1
1973 1.4 9.1 −7.7 0.0 na na 3.8 0.44 3.4
1974 1.8 5.9 −4.2 0.0 na na 2.3 0.16 2.1
1975 1.0 8.0 −7.0 0.0 na na 2.4 0.19 2.2
1976 0.9 5.9 −5.0 0.0 2.9 −2.9 2.4 0.07 2.3
1977 1.3 6.3 −5.0 0.0 3.8 −3.8 2.1 0.10 2.0
1978 1.5 5.6 −4.1 0.0 3.7 −3.7 2.3 0.10 2.2
1979 1.3 5.3 −4.0 0.0 2.5 −2.5 2.1 0.13 2.0
1980 2.2 6.4 −4.2 0.0 3.5 −3.5 2.8 0.16 2.6
1981 0.6 7.1 −6.4 0.0 4.2 −4.2 2.9 0.16 2.7
1982 1.3 6.0 −4.7 0.0 3.1 −3.1 2.9 0.12 2.7
1983 1.1 6.5 −5.4 0.0 3.6 −3.6 1.5 0.17 1.3
1984 0.8 5.1 −4.3 0.0 2.1 −2.1 2.5 0.10 2.4
1985 0.8 6.0 −5.2 0.0 2.9 −2.9 2.0 0.19 1.9
1986 0.7 5.8 −5.1 0.0 2.7 −2.7 2.1 0.23 1.8
1987 1.3 6.0 −4.7 0.0 2.3 −2.3 2.7 0.17 2.6
1988 1.1 5.6 −4.6 0.0 2.2 −2.2 2.5 0.15 2.4
1989 1.3 8.2 −6.9 0.0 4.2 −4.2 3.5 0.25 3.2
1990 1.2 7.6 −6.4 0.0 4.4 −4.4 3.8 0.30 3.5
1991 1.3 5.7 −4.5 0.0 1.7 −1.7 2.9 0.21 2.7
1992 1.5 6.0 −4.5 0.0 2.6 −2.6 3.0 0.31 2.7
1993 1.8 6.4 −4.6 0.0 3.0 −3.0 3.4 0.11 3.3
1994 1.0 5.2 −4.2 0.0 1.6 −1.6 2.8 0.08 2.7
1995 3.0 9.9 −6.9 0.0 4.8 −4.8 3.3 0.13 3.2
1996 1.9 8.2 −6.3 0.0 3.9 −3.9 3.1 0.08 3.0
(continued)
Annual Variation in Net Hydrologic Budgets 93

Table 17 (continued)
Na Na Na Al Al Al NH4 NH4 NH4
Water-year IN OUT NET IN OUT NET IN OUT NET
1997 0.9 5.3 −4.4 0.0 2.2 −2.2 3.0 0.11 2.9
1998 1.2 7.4 −6.2 0.0 3.6 −3.6 3.1 0.06 3.0
1999 1.7 7.5 −5.8 0.0 3.8 −3.8 3.6 0.06 3.5
2000 0.7 4.8 −4.1 0.0 1.8 −1.8 2.3 0.04 2.2
2001 1.3 5.5 −4.2 0.0 1.9 −1.9 2.9 0.04 2.9
2002 0.9 4.8 −3.9 0.0 1.6 −1.6 2.7 0.03 2.6
2003 1.9 8.6 −6.7 0.0 4.1 −4.1 3.2 0.06 3.1
2004 1.4 7.1 −5.7 0.0 2.7 −2.7 2.0 0.05 1.9
2005 1.6 9.0 −7.4 0.0 3.3 −3.3 2.3 0.06 2.3
2006 0.9 6.9 −6.0 0.0 2.7 −2.7 1.8 0.05 1.7
2007 1.1 7.3 −6.2 0.0 2.6 −2.6 3.6 0.05 3.5
2008 1.1 7.2 −6.0 0.0 1.0 −1.0 1.9 0.06 1.9
2009 1.1 7.5 −6.5 0.0 1.0 −1.0 1.3 0.06 1.3
SUM 64.0 312.1 −248.0 0.0 114.4 −114.4 125.2 7.7 117.5
AVG 1.4 6.6 −5.3 0.0 2.8 −2.8 2.7 0.2 2.6
±SD 0.5 1.3 1.2 0.0 1.0 1.0 0.7 0.2 0.6
H+ H+ H+ SO4 SO4 SO4 NO3 NO3 NO3
Water-year IN OUT NET IN OUT NET IN OUT NET
1963 0.86 0.10 0.76 na na na na na na
1964 0.74 0.07 0.67 30.4 29.7 0.7 6.9 4.8 2.1
1965 0.89 0.09 0.80 42.7 47.1 −4.5 17.9 6.5 11.4
1966 1.06 0.11 0.95 42.1 51.4 −9.3 19.8 6.1 13.7
1967 1.08 0.11 0.96 46.0 58.2 −12.2 23.6 12.4 11.2
1968 0.93 0.14 0.79 34.1 51.5 −17.4 16.8 11.6 5.2
1969 0.99 0.14 0.85 31.5 47.1 −15.6 18.4 28.2 −9.8
1970 1.28 0.14 1.14 41.7 52.6 −11.0 25.7 19.3 6.3
1971 0.92 0.12 0.80 34.0 46.0 −12.0 22.1 16.8 5.2
1972 1.05 0.17 0.88 39.1 62.4 −23.3 22.4 18.6 3.8
1973 1.20 0.22 0.97 54.0 81.5 −27.6 31.1 32.7 −1.6
1974 0.87 0.10 0.77 35.7 48.7 −13.0 20.8 19.8 1.0
1975 1.03 0.15 0.89 40.8 70.7 −29.9 24.9 26.3 −1.4
1976 0.85 0.10 0.75 32.8 46.4 −13.6 25.2 20.1 5.1
1977 1.05 0.12 0.93 39.5 54.8 −15.3 24.7 7.5 17.2
1978 0.94 0.12 0.81 34.5 46.9 −12.4 23.1 6.4 16.6
1979 0.72 0.08 0.64 27.1 38.2 −11.1 20.2 10.9 9.3
1980 0.79 0.11 0.69 31.6 43.2 −11.6 22.4 11.3 11.1
1981 0.80 0.14 0.66 34.3 55.7 −21.4 22.5 6.9 15.6
1982 0.74 0.09 0.65 31.2 44.1 −13.0 19.4 4.9 14.5
1983 0.56 0.12 0.44 20.6 50.5 −29.9 15.0 2.2 12.9
1984 0.66 0.08 0.58 28.1 34.8 −6.7 21.0 4.1 16.9
1985 0.61 0.12 0.48 24.1 44.8 −20.7 19.1 4.0 15.0
1986 0.62 0.10 0.52 25.1 39.5 −14.4 16.3 7.1 9.2
1987 0.68 0.07 0.61 25.7 38.0 −12.3 20.9 3.7 17.3
(continued)
94 4 Input–Output Budgets

Table 17 (continued)
H+ H+ H+ SO4 SO4 SO4 NO3 NO3 NO3
Water-year IN OUT NET IN OUT NET IN OUT NET
1988 0.66 0.07 0.58 25.4 35.7 −10.3 22.0 6.7 15.3
1989 0.79 0.12 0.67 30.6 50.4 −19.8 24.3 16.8 7.5
1990 0.82 0.15 0.67 34.7 53.7 −19.0 24.7 8.2 16.5
1991 0.70 0.10 0.60 26.9 42.4 −15.4 21.5 2.8 18.7
1992 0.71 0.11 0.61 26.7 41.5 −14.8 22.2 2.0 20.2
1993 0.78 0.09 0.69 30.4 43.7 −13.3 25.9 1.2 24.7
1994 0.63 0.06 0.57 23.3 30.0 −6.7 21.3 2.0 19.3
1995 0.79 0.18 0.61 27.3 59.8 −32.5 25.8 3.4 22.4
1996 0.70 0.14 0.56 26.5 48.8 −22.3 23.3 1.9 21.4
1997 0.50 0.07 0.42 20.3 32.0 −11.7 19.7 1.6 18.1
1998 0.66 0.11 0.54 24.4 39.3 −14.9 22.7 12.9 9.8
1999 0.70 0.14 0.55 25.5 44.1 −18.5 22.1 6.8 15.3
2000 0.45 0.06 0.39 16.7 27.9 −11.2 17.4 3.4 14.1
2001 0.47 0.07 0.40 18.6 30.8 −12.2 19.4 1.2 18.2
2002 0.40 0.06 0.34 16.3 26.1 −9.8 15.5 1.1 14.5
2003 0.54 0.13 0.41 22.3 46.6 −24.4 19.5 2.4 17.1
2004 0.44 0.08 0.35 18.3 37.1 −18.8 14.7 1.7 13.0
2005 0.45 0.10 0.34 20.5 45.9 −25.3 14.8 0.9 13.8
2006 0.34 0.08 0.26 15.9 35.9 −20.0 12.0 2.9 9.1
2007 0.48 0.07 0.41 22.3 37.7 −15.4 17.3 1.3 16.0
2008 0.36 0.08 0.29 16.4 37.7 −21.3 12.9 0.8 12.1
2009 0.19 0.06 0.13 9.8 36.5 −26.7 8.8 0.4 8.4
SUM 34.44 5.05 29.39 1,325.6 2,067.4 −741.8 927.9 374.8 553.1
AVG 0.73 0.11 0.63 28.8 44.9 −16.1 20.2 8.1 12.0
±SD 0.24 0.03 0.22 9.1 11.0 7.0 4.7 8.0 7.0
Cl Cl Cl PO4 PO4 PO4 DOC DOC DOC
Water-year IN OUT NET IN OUT NET IN OUT NET
1963 na na na na na na na na na
1964 3.8 2.9 1.0 na na na na na na
1965 6.4 4.3 2.1 na na na na na na
1966 5.4 4.6 0.8 na na na na na na
1967 5.5 5.3 0.1 na na na na na na
1968 5.5 5.1 0.5 na na na na na na
1969 4.7 4.1 0.6 na na na na na na
1970 5.3 4.6 0.6 na na na na na na
1971 5.1 3.5 1.6 0.082 na na na na na
1972 6.6 6.6 −0.1 0.195 0.018 0.177 na na na
1973 4.5 5.9 −1.4 0.131 0.032 0.100 na na na
1974 5.2 4.2 1.0 0.097 0.017 0.081 na na na
1975 3.6 4.6 −1.0 0.148 0.031 0.117 na na na
1976 4.9 3.3 1.6 0.139 0.023 0.116 na na na
1977 4.4 3.5 0.9 0.125 0.045 0.080 na na na
1978 4.2 4.1 0.1 0.225 0.037 0.188 na na na
(continued)
Annual Variation in Net Hydrologic Budgets 95

Table 17 (continued)
Cl Cl Cl PO4 PO4 PO4 DOC DOC DOC
Water-year IN OUT NET IN OUT NET IN OUT NET
1979 3.2 3.9 −0.7 0.209 0.033 0.176 na na na
1980 4.3 5.6 −1.3 0.561 0.017 0.545 na na na
1981 1.7 4.5 −2.7 0.470 0.033 0.437 na na na
1982 2.4 3.7 −1.3 0.246 0.025 0.221 na na na
1983 2.3 4.6 −2.3 0.237 0.042 0.195 na na na
1984 1.9 3.3 −1.4 0.174 0.027 0.147 na na na
1985 1.9 3.5 −1.6 0.198 0.065 0.133 na na na
1986 1.8 3.1 −1.3 0.256 0.069 0.187 na na na
1987 2.9 3.4 −0.6 0.412 0.067 0.345 na na na
1988 2.2 3.5 −1.3 0.166 0.020 0.145 na na na
1989 3.0 4.8 −1.8 0.343 0.050 0.293 na na na
1990 2.4 4.6 −2.2 0.385 0.080 0.305 na na na
1991 2.4 3.2 −0.8 0.165 0.036 0.129 na na na
1992 2.8 3.3 −0.5 0.101 0.023 0.077 na na na
1993 3.5 4.3 −0.7 0.161 0.012 0.149 na na na
1994 2.0 2.8 −0.8 0.201 0.009 0.191 na na na
1995 6.0 8.4 −2.4 0.100 0.029 0.071 15.6 32.2 −16.6
1996 3.6 5.4 −1.8 0.065 0.011 0.054 15.3 26.5 −11.3
1997 2.0 2.7 −0.7 0.153 0.012 0.141 11.7 16.4 −4.7
1998 2.5 4.0 −1.5 0.127 0.023 0.104 14.8 25.7 −10.8
1999 3.4 4.9 −1.5 0.188 0.044 0.144 12.4 33.0 −20.6
2000 1.6 2.4 −0.7 0.060 0.015 0.045 13.1 15.6 −2.5
2001 2.3 3.5 −1.2 0.140 0.012 0.128 13.3 16.5 −3.2
2002 1.7 2.8 −1.1 0.089 0.011 0.078 11.8 10.2 1.6
2003 3.5 5.9 −2.4 0.162 0.016 0.146 15.4 33.4 −18.0
2004 2.6 4.2 −1.6 0.068 0.007 0.061 11.3 22.4 −11.1
2005 3.1 5.2 −2.1 0.099 0.006 0.092 12.5 25.2 −12.7
2006 1.7 3.8 −2.1 0.008 0.005 0.003 8.4 22.0 −13.6
2007 2.1 3.5 −1.4 0.011 0.006 0.006 13.1 20.3 −7.3
2008 2.2 3.3 −1.1 0.023 0.006 0.017 14.7 24.0 −9.3
2009 2.2 3.7 −1.5 0.008 0.006 0.002 8.7 24.9 −16.1
SUM 156.3 192.5 −36.2 6.728 1.020 5.625 192.0 348.1 −156.1
AVG 3.4 4.2 −0.8 0.173 0.027 0.148 12.8 23.2 −10.4
±SD 1.4 1.1 1.2 0.123 0.019 0.114 2.2 6.7 6.3
di Silica di Silica di Silica
Water-year IN OUT NET
1963 na na na
1964 0 20.9 −20.9
1965 0 31.2 −31.2
1966 0 36.3 −36.3
1967 0 36.4 −36.4
1968 0 29.2 −29.2
1969 0 35.3 −35.3
(continued)
96 4 Input–Output Budgets

Table 17 (continued)
di Silica di Silica di Silica
Water-year IN OUT NET
1970 0 34.7 −34.7
1971 0 31.6 −31.6
1972 0 42.8 −42.8
1973 0 50.5 −50.5
1974 0 32.3 −32.3
1975 0 47.3 −47.3
1976 0 35.3 −35.3
1977 0 38.6 −38.6
1978 0 30.1 −30.1
1979 0 29.3 −29.3
1980 0 32.6 −32.6
1981 0 43.3 −43.3
1982 0 32.3 −32.3
1983 0 35.3 −35.3
1984 0 27.0 −27.0
1985 0 33.3 −33.3
1986 0 31.8 −31.8
1987 0 30.8 −30.8
1988 0 26.1 −26.1
1989 0 40.9 −40.9
1990 0 42.2 −42.2
1991 0 32.3 −32.3
1992 0 32.7 −32.7
1993 0 32.8 −32.8
1994 0 26.5 −26.5
1995 0 47.7 −47.7
1996 0 39.8 −39.8
1997 0 29.5 −29.5
1998 0 40.4 −40.4
1999 0 42.1 −42.1
2000 0 27.2 −27.2
2001 0 28.6 −28.6
2002 0 24.3 −24.3
2003 0 43.2 −43.2
2004 0 35.3 −35.3
2005 0 44.2 −44.2
2006 0 36.7 −36.7
2007 0 40.5 −40.5
2008 0 43.2 −43.2
2009 0 46.1 −46.1
SUM 0.0 1,630.6 −1,630.6
AVG 0.0 35.4 −35.4
±SD 0.0 6.8 6.8
Table 18 Mean annual (arithmetic average of annual values) bulk precipitation input, streamwater output and net loss or gain, and standard deviation of the mean for dissolved
substances, during 1963–2009 for Watershed 6 of the Hubbard Brook Experimental Foresta
Inputs Outputs Net loss or gainb
% of Mean % of % of % of % of
Obs Mean (kg/ total (Eq/ total Obs Mean (kg/ total Mean (Eq/ total Mean (kg/ total Mean (Eq/
Substance (years) ha-year) ±1 SD (kg) ha-year) ±1 SD (Eq) (years) ha-year) ±1 SD (kg) ha-year) ±1 SD (Eq) ha-year) ±1 SD (kg) ha-year) ±1 SD
Calcium 47 1.35 0.58 2 67 29 3 47 8.82 2.96 7 440 148 17 −7.47 2.78 16 −373 139
Magnesium 47 0.39 0.18 1 32 15 2 47 2.36 0.57 2 194 47 7 −1.97 0.55 4 −162 45
Potassium 47 0.70 0.32 1 18 8 1 47 1.85 0.44 2 47 11 2 −1.15 0.53 2 −29 14
Sodium 47 1.36 0.47 2 59 20 3 47 6.64 1.28 6 289 56 11 −5.28 1.16 11 −230 50
Aluminum 34 0.00 0.00 0 0 0 0 40 2.79 0.98 2 209 73 8 −2.79 0.98 6 −209 73
Ammonium 46 2.72 0.65 4 151 36 7 46 0.17 0.15 <1 9 9 <1 2.55 0.65 −5 142 36
Hydrogen ion 47 0.733 0.237 1 727 235 35 47 0.107 0.035 <1 107 34 4 0.625 0.216 −1 620 214
Sulfate 46 28.82 9.06 40 600 189 29 46 44.94 10.96 38 936 228 36 −16.13 7.01 34 −336 146
Nitrate 46 20.17 4.66 28 325 75 16 46 8.15 8.01 7 131 129 5 12.02 6.98 −25 194 113
Chloride 46 3.40 3.40 5 96 96 5 46 4.18 1.14 3 118 32 5 −0.79 1.19 2 −22 33
Phosphate 46 0.173 0.123 <1 5.4 3.9 <1 38 0.027 0.019 <1 0.8 0.6 <1 0.146 0.114 <1 4.6 3.6
DOC 15 12.80 2.23 18 na na na 15 23.20 6.74 19 139 40 5 −10.41 6.28 22 na na
Dissolved 38 0.00 0.00 0 na na na 46 16.59 3.19 14 na na na −16.59 3.19 35 na na
silica
Total 72.61 kg 100 2,081 Eq 100 119.83 kg 100 2,621 Eq 100 −47.22 kg 100 −401 Eq
Cations 1,054 1,296 −241
Anions 1,027 1,325 −159
a
These data emphasize that for some elements, data from one or a few annual cycles can give grossly aberrant estimates of long-term mean values of both direction and magnitude
of change
b
There are small rounding errors associated with some of these mean values and their calculation from gross output minus input
 98

Table 19 Summary of average input–output (NHF) budgets for Watershed 6 of the Hubbard Brook Experimental Forest during 1963–2009a
Inputs Outputs
BP DD SW M BP − SW (BP + DD) − SW (BP + DD) − (SW + M)
Years Bulk precip Dry dep (kg/ Stream (kg/ Particulate NHF (kg/ NHF + DD (kg/
Element of data (kg/ha-year) ha-year) ha-year) (kg/ha-year) ha-year) ha-year) ΣWS I–O (kg/ha-year)
Calcium 47 1.35 Included in BP ‒8.82 ‒0.21 ‒7.47 ‒7.47 ‒7.68
Magnesium 47 0.39 Included in BP ‒2.36 ‒0.19 ‒1.97 ‒1.97 ‒2.16
Potassium 47 0.70 Included in BP ‒1.85 ‒0.52 ‒1.15 ‒1.15 ‒1.67
Sodium 47 1.36 Included in BP ‒6.64 ‒0.25 ‒5.28 ‒5.28 ‒5.53
Aluminum 42 <0.01 Included in BP ‒2.79 ‒1.38 ‒2.79 ‒2.79 ‒4.17
Ironb na <0.01 Included in BP ‒0.05 ‒0.64 ‒0.05 ‒0.05 ‒0.69
Hydrogen 47 0.73 Included in BP ‒0.11 0.00 0.63 0.63 0.63
Sulfur 46 9.60 2.05 ‒14.97 ‒0.03 ‒5.37 ‒3.32 ‒3.35
Nitrogen 46 6.76 1.40 ‒2.00 ‒0.11 4.76 6.16 6.05
Chlorine 46 3.40 0.68 ‒4.18 0.68 ‒0.78 ‒0.10 na
Phosphorus 39 0.056 Included in BP ‒0.009 ‒0.012 0.047 0.047 0.035
Carbon 18 12.8 Included in BP ‒23.20 ‒3.98 ‒10.40 ‒10.40 ‒14.38
Dissolved 46 <0.01 Included in BP ‒35.45 ‒6.19 ‒35.45 ‒35.45 ‒41.64
silicac
4

Dry deposition (DD) included with BP for elements without gaseous/aerosol phase, except carbon. Particulate losses (M) modified from Bormann et al. (1974);
particulate matter losses are from water-years 1966 through 1969. S, N, and P converted from solutes measured as SO4, NO3 + NH4, and PO4. Average value
from Mitchell et al. (2011). Estimated as 1.4 kg/ha N (Lovett et al. 1997). Not estimated, but long-term balance would be equal to 0.7 kg/ha Cl (Lovett et al.
2005)
a
Not all years have data: these are from the years of continuous, systematic measurements
b
Iron streamwater outputs estimated from average trace concentrations in stream water [0.005 mg/L] times average flow (see Fuss et al. 2011)
c
Dissolved silica measured and expressed as SiO2
Input–Output Budgets
Annual Variation in Net Hydrologic Budgets 99

a 2
Gains
0

-2
NHF= (I-O) (kg/ha-yr)

-4

-6

-8

-10
Calcium
Magnesium
-12
Potassium
-14 Sodium

-16
Losses
-18
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
Water Year
b
30 Sulfate Gains
Ammonium
Nitrate
20
NHF= (I-O) (kg/ha-yr)

10

-10

-20

-30
Losses
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
Water Year
Fig. 36 Net hydrologic flux (bulk precipitation input–streamwater output) for (a) calcium ( filled
circle), magnesium (open circle), potassium ( filled triangle), and sodium (square); (b) sulfate
( filled circle), ammonium ( filled triangle), and nitrate ( square); and (c) dissolved silica ( filled
circle), in Watershed 6 of the Hubbard Brook Experimental Forest from 1963 to 2009
100 4 Input–Output Budgets

c
Gains
0

-10
NHF= (I-O) (kg/ha-yr)

Dissolved Silica
-20

-30

-40

-50

Losses
-60
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
Water Year
Fig. 36 (continued)
Annual Inputs, Outputs and NHF (kg/ha-yr)

10

Chloride Outputs
8 Net Hydrologic Flux
Chloride Inputs

-2

-4
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
Water Year
Fig. 37 Annual bulk precipitation inputs (open circle), streamwater outputs (filled circle), and Net
Hydrologic Flux (filled triangle) for chloride in Watershed 6 of the Hubbard Brook Experimental
Forest from 1964 to 2009. Chloride inputs (r2 0.51), outputs (ns), NHF (r2 0.47, p < 0.001)
Some Examples of Important Comonents of Input–Output Budgets 101

net change and magnitude. With the standard deviation of the mean and Table 18
and Fig. 36a–c used as a guide, a few annual measurements would yield reliable
estimates of magnitude of loss or gain for calcium, magnesium, sodium, aluminum,
ammonium, hydrogen ion, and dissolved Si. Nitrate, chloride, and potassium also
have a large range in annual magnitude of net change, but they are much more
unpredictable relative to direction of net change than the other ions (Tables 17 and
18; Figs. 36b and 37). Annual bulk precipitation inputs of phosphate and chloride
have been particularly variable (Table 18, Fig. 37). The standard deviation of the
mean for long-term potassium inputs is 44 %; for phosphate inputs, it is 72 %; and
for chloride, it is 100 % of the mean (equivalence basis). These highly variable
inputs are therefore the major reason for the uncertainty in the average annual bud-
gets for these elements.
Following an initial decline in NHF (increased losses) for calcium since ~1973,
there was a significant (r2 = 0.66) upward trend of diminished losses (Fig. 36a) as
available weatherable calcium was depleted and precipitation inputs decreased (see
Chaps. 3 and 5). The NHF of sodium tended (insignificantly) toward increased
losses, whereas the NHF for magnesium and potassium showed no trend (Fig. 36a).
Sulfate showed highly variable but a decreasing trend toward increasing losses
(r2 = 0.24; Fig. 36b), whereas chloride hovered at NHF of zero, mostly positive prior
to ~1977 and negative thereafter (Fig. 37). Nitrate, ammonium, and dissolved inor-
ganic nitrogen were mostly positive (net gain) throughout the 47-year period
(Fig. 36b). The NHF for dissolved silica was highly variable and consistently nega-
tive (Fig. 36c).
These data emphasize that for some elements data from one or a few annual
cycles may give grossly aberrant estimates of long-term mean values of both direc-
tion and magnitude of change.

Some Examples of Important Components

of Input–Output Budgets

Nitrogen

During the 46-year period, 1964–2009, 927.8 kg/ha of nitrate (209.7 kg/ha NO3–N)
were added in bulk precipitation and 374.8 kg of nitrate per hectare (84.7 kg/ha
NO3–N) were lost in stream water from forested, watershed-ecosystems at HBEF.
Likewise, 125.2 kg of ammonium per hectare (97.4 kg/ha NH4–N) have been added
in precipitation, and 7.72 kg of ammonium per hectare (6.0 kg/ha NH4–N) have
been lost in stream water during 1964–2009. These amounts give a total dissolved
inorganic nitrogen input of 307.1 kg N/ha and a total dissolved inorganic nitrogen
output of 90.7 kg N/ha during this 46-year period. These long-term data show
clearly that >3 times more dissolved inorganic nitrogen has been added to W6 in
bulk precipitation than has been lost in stream water. This amount although
102 4 Input–Output Budgets

substantial did not account for the nitrogen accumulating within the ecosystem in
living and dead biomass during the early years of the study, 1963–1982 (Bormann
et al. 1977; Covington 1975, 1976; Whittaker et al. 1974; see Chap. 6). It was esti-
mated that nitrogen was incorporated and stored in biomass during this early period
at a rate of about 16.7 kg N/ha-year in these forest ecosystems (about 9 kg N/ha-
year in living biomass and about 8 kg N/ha-year in soil organic matter; Bormann
et al. 1977). Because nitrogen is essentially absent from the bedrock and till at
HBEF, it can be assumed that the release of nitrogenous substances by weathering
and the formation of secondary nitrogenous minerals are negligible. Therefore, the
net gaseous and aerosol exchange of nitrogen across the ecosystem’s boundaries
was estimated (Fig. 1) from net biomass accumulation minus net hydrologic flux
(bulk precipitation input minus streamwater output). The net gain in dissolved inor-
ganic nitrogen during 1963–1974 was 2.5 kg/ha-year (bulk precipitation input
minus loss of dissolved and particulate nitrogen compounds in streamflow; Bormann
et al. 1977). Therefore, applying an ecosystem analysis to the small watershed
approach and based on net biomass uptake (16.7 kg N/ha-year) and NHF (2.5 kg N/
ha-year), it was estimated that on average about 14.2 kg of nitrogen/ha-year entered
the ecosystem in a gaseous form or from impacted nitrogenous aerosols (Bormann
et al. 1977). This value is almost certainly too large (see Chap. 6) as inputs of gases
and particles (dry deposition) were not fully evaluated for the ecosystems at HBEF
at that time. More recently, the dry deposition of N to these forest watersheds has
been estimated at ~1.4 kg N/ha-year (Lovett et al. 1997). Field data indicated that
microbial fixation of nitrogen occurs within these forested ecosystems, but the val-
ues are relatively small (~2 kg N/ha-year; Roskoski 1980). Denitrification, particu-
larly chemical denitrification, may also occur, but recent estimates indicate that it
also is small. Annual fluxes (losses) of N2O ranged between 0.27 and 1.4 kg N/ha-
year between 1986 and 2004 (Bowden and Bormann 1986; Groffman et al. 2006,
2009). In situ, temporal measurements of denitrification are difficult and frequently
contentious, particularly at the watershed scale. Using stable isotopes, Bernal et al.
(2012) found no evidence for long-term changes in denitrification rates at the HBEF.
Long-term measurements of the forest floor and mineral soil have not detected any
significant changes in N content, in large part because of the difficulty in detecting
rather small changes within large N pools in the soil.
Inputs and outputs and the NHF for inorganic nitrogen have changed dramatically
with time at HBEF (Figs. 36b and 45), and extensive work now has been done on
many aspects of the nitrogen budget (see also Chap. 6). For example, Bernal et al.
(2012) have shown the legacies of previous disturbance, the critical value of long-
term data, and, based on an analysis of nitrogen isotopes in archived water samples,
that denitrification appears to be a negligible contributor to this change. Nitrogen
isotopes (15N) in soil organic matter show little trend over time, whereas this isotopic
signature should have increased if denitrification were to have increased (see
Chap. 6). Our long-term biogeochemical record and climate dynamics showed that
climate alone cannot explain the occurrence of the dramatic >90 % drop in watershed
nitrate concentration and export in stream water (Fig. 19b) over the past 46 years,
despite longer growing seasons and higher soil temperatures (Bernal et al. 2012).
Some Examples of Important Components of Input–Output Budgets 103

Bernal et al. (2012) evaluated eight potential drivers/vectors of this streamwater

decline in nitrate export (decreased N deposition, increased losses of DON, longer
growing season, increased plant accumulation, increased denitrification, changes in
soil-water flow paths/temperature, sugar maple decline, historical disturbances) and
identified two capable of causing the observed change in the N cycle: altered paths of
soil-water flow and historic effects of forest cutting (Bernal et al. 2012).

Chlorine

The long-term annual hydrologic budget for chloride has changed dramatically
since 1964 (Table 17; Fig. 37). Initially, the chloride hydrologic budget at HBEF
was thought to be in balance or net retention occurred (Juang and Johnson 1967;
Lovett et al. 2005). Prior to 1981–1982, the annual hydrologic chloride budget
ranged from +2.1 (1965–1966) and −2.7 (1981–1982) to −0.6 kg Cl/ha-year in
1987–1988, averaging −0.8 kg Cl/ha-year during 1964–2009 (Table 17).
Chlorine, like nitrogen, may have a gaseous phase at normal biologic tempera-
tures; however, chlorine gas is very reactive and thus is not common in long-range
atmospheric transport. Significant amounts of chloride in throughfall and stemflow
(Eaton et al. 1973; Lovett et al. 1996; Berger et al. 2001) suggested that small
amounts of impacted aerosols containing chloride were washed off vegetation sur-
faces or that chloride was leached from vegetation by incident precipitation.
Experimental additions of chloride to the forest floor showed that biologic uptake
and accumulation of chloride can occur in foliage (Berger et al. 2001). Also, sub-
stantial releases of chloride occurred in stream water from some internal reservoir,
possibly biomass decay, after experimental disturbance by deforestation (Likens
et al. 1970). Somewhat surprisingly, long-term data, experimental watershed manip-
ulation, and experimental plot studies (Berger et al. 2001) suggest an important role
for forest vegetation in the chloride cycle, which raises questions about the value of
chloride as a conservative tracer in various ecosystem studies.
Lovett et al. (2005) suggested that the dominant source of chloride in the atmo-
sphere during the 1960s and 1970s was from regional coal burning, whereas in more
recent decades, chloride deposition originated primarily from marine sources. Dry
deposition of chloride is difficult to evaluate but may contribute ~20 % of bulk pre-
cipitation inputs (Lovett et al. 2005). Mineralization from soil organic matter also
may be an important source of chloride in stream water (Lovett et al. 2005).

Sulfur

Except for 1964, the NHF for sulfate has been consistently negative (Fig. 36b), sug-
gesting some “missing” source. Storage (net uptake) of sulfur in the annual biomass
increment occurred (above- and belowground) at a rate of about ~37 mol/ha-year
104 4 Input–Output Budgets

Photograph 8 Collecting a streamwater sample from Watershed 4 at the Hubbard Brook

Experimental Forest (Photo by D.C. Buso)

(3.6 kg/ha-year) in 1964–1969 and 3.1 mol/ha-year (0.3 kg/ha-year) in 1993–1998

(Likens et al. 2002a). But, chemical weathering releases ~50 mol/ha-year (4.8 kg S/
ha-year) within the ecosystems of the HBEF and probably more than balances the
net biomass storage (Likens et al. 2002a). Sulfur also may exist as a gas or as an
aerosol in the atmosphere. Unfortunately, the amount of S input from atmospheric
gaseous and aerosol sources annually has high uncertainty—estimated in 1964–
1969 and 1993–1998 as 85 (8.1 kg/ha-year) and 57 mol/ha-year (5.5 kg/ha-year),
respectively (see Chap. 6). A new approach in 2011 for evaluating dry deposition of
S gave a range of 0.5–2.5 kg S/ha-year (Mitchell et al. 2011). This amount should
be added to the bulk precipitation input value in calculating the net ecosystem flux
for these ecosystems (see Chap. 6). The origin, quantification, and ecological effects
of gaseous and/or aerosol inputs of sulfur require detailed studies to elucidate the
role of sulfur in forested ecosystems. Unfortunately, these fluxes frequently are not
quantitatively or carefully evaluated in biogeochemical studies.

Dissolved Silica

Negligible amounts of dissolved silica are input in bulk precipitation to watershed-

ecosystems of the Hubbard Brook Experimental Forest, so the NHF is strongly and
consistently negative (Fig. 36c). These outputs in stream water are made up from
weathering of silicate minerals within the ecosystem (Photograph 8).
Input–Output Budgets: Summary 105

Input–Output Budgets: Summary

Overall, during the period 1963–2009, there was an average annual net loss of total
dissolved inorganic substances from W6 amounting to 47.2 kg/ha-year (Table 18).
The net loss averaged 241 Eq/ha-year for cations and 159 Eq/ha-year for anions
(Table 18). Organic anions make up the imbalance. The average annual net output
of dissolved inorganic substances minus dissolved silica (1963–2009) was 30.6 kg/
ha-year (Table 18). However, sizeable yearly variations and long-term trends have
occurred during the study (Figs. 19, 27, 36, and 37; Table 17). For example, the
smallest net loss of dissolved inorganic substances (27.8 kg/ha or 7.0 kg/ha for total
dissolved material minus dissolved silica) occurred during 1964–1965, the driest
year of our study. The largest net losses of dissolved inorganic substances occurred
during the wettest year, 1973–1974 (139.7 kg/ha and 986 cationic equivalents per
hectare). Hydrogen ion dominates equivalent fluxes due to its high reactivity, but
because it is the lightest element, it does not significantly affect mass.
Budgets adjusted to all major sources of input and output for the entire period are
presented in Table 19 and in Chap. 6. Table 19 is somewhat misleading because of
strong temporal trends in inputs and outputs such as for sulfate, nitrate, and calcium.
The explanation for the overall net losses lies in the biogeochemical reactions that
occur within the ecosystem, notably chemical weathering, biotic uptake and net
storage of nutrients, and related phenomena (see Chap. 6).
We show average inputs in bulk precipitation (Table 20) and outputs in stream
water (Table 21) for two decades: one at the beginning of the long-term study
(1964–1973) and one currently (2000–2009).
The pattern of average bulk precipitation solute input and streamwater solute
output during the seasonal periods for the entire study (Tables 22 and 23) and for
two decades (Tables 24 and 25) is similar to and driven by the monthly flux (Figs. 38,
39, 40, and 41), but the resolution is not as great. Input (on a weight basis) using the
most recent decade (2000–2009) is greatest during the summer and lowest during
the winter by ~41 % (Table 24). The equivalent input is lowest during the winter,
largely because of the sharp decline in sulfate, ammonium, and DOC inputs. On a
weight basis, sulfate, nitrate, and dissolved organic carbon provide about 80–88 %
of the dissolved inorganic substances in precipitation. On an equivalent basis,
hydrogen ion and sulfate dominate, as expected. In general the input of nitrate
reflects more closely than any other ion the amount of water added as precipitation
because nitrate concentration has remained high in precipitation until recently
(Table 17, Fig. 28). Inputs of magnesium, potassium, and to a lesser extent calcium
and sodium are relatively greater than expected on the basis of amount of precipita-
tion during the autumn (enriched precipitation relative to the annual average) and
less during the winter. All ionic inputs during the winter season were less than
expected on the basis of amount of precipitation; in general, however, inputs of
individual dissolved substances were highly variable on a seasonal basis (Table 22).
In the early decade (1964–1973), more than 55 % of the annual output of dis-
solved substances in stream water occurred during the spring. More recently (2000–
2009), however, only 36 % of dissolved substances in stream water output occur in
Table 20 Average decadal bulk precipitation inputs for two periods (1964–1973 and 2000–2009) for Watershed 6 of the Hubbard Brook Experimental Forest
Past precipitation inputs 1964–1973 decade Present precipitation inputs 2000–2009 decade
Water INPUTS Total (mm) Average (mm/ Water Total (mm) Average (mm/
ha-year) INPUTS ha-year)
Precipitation 13,880 1,388 Precipitation 14,877 1,488
Solute INPUTS 10-year total 10-year avg VWA conc % Total cations Solute 10-year total 10-year avg VWA conc % Total cations
(Eq) (Eq/ha-year) (µEq/L) or anions INPUTS (Eq) (Eq/ha-year) (µEq/L) or anions
Calcium 1,035 103.5 7.5 7.3 Calcium 514.8 51.5 3.5 7.5
Magnesium 482.0 48.2 3.5 3.4 Magnesium 229.3 22.9 1.5 3.3
Potassium 203.1 20.3 1.5 1.4 Potassium 151.8 15.2 1.0 2.2
Sodium 714.2 71.4 5.1 5.0 Sodium 523.0 52.3 3.5 7.6
Aluminuma 1.0 0.1 0.01 0.0 Aluminuma 1.0 0.1 0.01 0.0
Ammonium 1,677 167.7 12.1 11.8 Ammonium 1,326 132.6 8.9 19.4
H+ ion 10,142 1,014 73.1 71.2 H+ ion 4,104 410.4 27.6 59.9
Average pH 4.14 Average pH 4.56
Sulfate 8,230 823.0 59.3 62.9 Sulfate 3,689 368.9 24.8 54.1
Nitrate 3,299 329.9 23.8 25.2 Nitrate 2,459 245.9 16.5 36.1
Chloride 1,489 148.9 10.7 11.4 Chloride 649.2 64.9 4.4 9.5
Phosphateb 60.4 6.0 0.4 0.5 Phosphate 21.1 2.1 0.1 0.3
Bicarbonatec 0.0 0.0 0.0 0.0 Bicarbonatec 0.0 0.0 0.0 0.0
DOCd 0.0 0.0 0.0 0.0 DOCd 0.0 0.0 0.0 0.0
Total cations 1,425 102.7 100.0 Total cations 685.0 46.0 100.0
Total anions 1,308 94.2 100.0 Total anions 681.8 45.8 100.0
Imbalance 117.6 8.5 Imbalance 3.2 0.2
Dissolved silica 0.7 (kg) 0.1 (kg/ha-year) 0.005 na Dissolved 0.7 (kg) 0.1 (kg/ha-year) 0.005 na
(mg/L) silica (mg/L)
a
Aluminum and dissolved silica found only in trace amounts in precipitation (Buso et al. 2000)
b
Phosphate analysis not available for all of 1964–1973: used 1971–1980 input data
c
Bicarbonate (as ANC) not measureable below pH 5
d
DOC = dissolved organic carbon not a charged species in precipitation
Table 21 Average decadal streamwater outputs for two periods (1964–1973 and 2000–2009) for Watershed 6 of the Hubbard Brook Experimental Forest
Past streamwater outputs 1964–1973 decade Present streamwater outputs 2000–2009 decade
Water OUTPUTS Total (mm) Average (mm/ Water Total (mm) Average
ha-year) OUTPUTS (mm/ha-year)
Streamflow 8,817 881.7 Streamflow 10,064 1,006
Solute OUTPUTS 10-year 10-year avg VWA conc % Total cations Solute 10-year 10-year avg VWA conc % Total cations
total (Eq) (Eq/ha-year) (µEq/L) or anions OUTPUTS total (Eq) (Eq/ha-year) (µEq/L) or anions
Calcium 5,975 597.5 67.8 36.3 Calcium 3,104 310.4 30.8 29.3
Magnesium 2,465 246.5 28.0 15.0 Magnesium 1,590 159.0 15.8 15.0
Potassium 513.9 51.4 5.8 3.1 Potassium 417.7 41.8 4.2 3.9
Sodium 2,989 298.9 33.9 18.2 Sodium 2,999 299.9 29.8 28.3
Aluminuma 3,010 301.0 34.1 18.3 Aluminuma 1,677 167.7 16.7 15.8
Ammonium 183.6 18.4 2.1 1.1 Ammonium 28.3 2.8 0.3 0.3
H+ ion (and pH) 1,324 132.4 15.0 8.0 H+ ion (and pH) 792.7 79.3 7.9 7.5
Average pH 4.82 Average pH 5.10
Sulfate 10,984 1,098 124.6 70.4 Sulfate 7,542 754.2 74.9 71.7
Nitrate 2,533 253.3 28.7 16.2 Nitrate 260.7 26.1 2.6 2.5
Chloride 1,326 132.6 15.0 8.5 Chloride 1,078 107.8 10.7 10.2
Phosphateb 9.0 0.9 0.1 0.1 Phosphate 2.9 0.3 0.0 0.0
Bicarbonatec 0.0 0.0 0.0 0.0 Bicarbonatec 350.0 35.0 3.5 3.3
DOCd 740.0 74.0 8.4 4.7 DOCd 1,290 129.0 12.8 12.3
Total cations 1,646 186.7 100.0 Total cations 1,061 105.4 100.0
Total anions 1,559 176.8 100.0 Total anions 1,052 104.6 100.0
Imbalance 86.9 9.9 Imbalance 8.5 0.8
Dissolved silica 348.9 (kg) 34.9 4.0 na Dissolved silica 98.6 (kg) 9.9 (kg/ha-year) 1.0 (mg/L) na
(kg/ha-year) (mg/L)
a
Aluminum valence estimated +2.5 for pH 4.8 (1964–1973) and +2.0 for pH 5.1 (2000–2009) (Buso et al. 2000)
b
Phosphate analysis not available in 1964–1973; used 1973–1982 export data
c
Bicarbonate (as ANC) not measureable below pH 5; ANC measured routinely since 1990 (Buso et al. 2000)
d
DOC assigned a charge of −6 µEq/mgC (Buso et al. 2000); DOC mass for 1964–1973 based on sporadic measurements and may be higher
108 4 Input–Output Budgets

Table 22 Seasonal volume-weighted average bulk precipitation values for Watershed 6 of the
Hubbard Brook Experimental Forest during 1963–2009
Summer Fall Winter Spring
October– December–
Obs June–September November February March–May
Solute Years mg/L µEq/L mg/L µEq/L mg/L µEq/L mg/L µEq/L
Calcium 47 0.08 4.0 0.08 4.0 0.06 3.0 0.14 7.0
Magnesium 47 0.02 1.6 0.03 2.5 0.02 1.6 0.04 3.3
Potassium 47 0.05 1.3 0.07 1.8 0.03 0.8 0.06 1.5
Sodium 47 0.05 2.2 0.14 6.1 0.11 4.8 0.11 4.8
Aluminuma 34 0.00 0.0 0.00 0.0 0.00 0.0 0.00 0.0
Ammonium 46 0.217 12.0 0.151 8.4 0.119 6.6 0.244 13.5
Hydrogen ionb 47 0.061 60.5 0.045 44.6 0.039 38.7 0.054 53.6
pH 4.22 4.35 4.41 4.27
Sulfate 46 2.62 54.5 1.59 33.1 1.20 25.0 2.19 45.6
Nitrate 46 1.27 20.5 1.33 21.4 1.44 23.3 1.60 25.8
Chloride 46 0.18 5.1 0.30 8.5 0.25 7.1 0.26 7.3
Phosphate 39 0.011 0.3 0.007 0.2 0.007 0.2 0.021 0.7
DOCc 15 1.3 0.0 0.6 0.0 0.4 0.0 0.8 0.0
Dissolved 38 0.00 na 0.00 na 0.00 na 0.00 na
silicaa
Totals 5.81 (mg/L) 4.36 (mg/L) 3.70 (mg/L) 5.54 (mg/L)
Cations 81.6 (µEq/L) 67.3 (µEq/L) 55.5 (µEq/L) 83.7 (µEq/L)
Anions 80.4 (µEq/L) 63.1 (µEq/L) 55.5 (µEq/L) 79.4 (µEq/L)
a
Aluminum and dissolved silica are trace in precipitation
b
Hydrogen ion in µg/L VWA-concentration
c
DOC has no charge in precipitation

spring, followed closely by 34 % during autumn (Table 25). This dramatic change
is due to a sharp drop in sulfate and nitrate export and an overall dilution of stream
water with time (Likens and Buso 2012). In contrast, only 11 % and 12 % of the
output occurs during summer in the early and recent decades, respectively (Table 25).
Sulfate clearly dominates the output of dissolved substances on a weight and equiv-
alent basis. Calcium, magnesium, sodium, sulfate, chloride, and hydrogen ion losses
in stream water generally reflect the amount of water lost. Phosphate losses during
the early decades bore little relationship to the amount of streamflow; summer and
winter losses are appreciably greater than expected, and autumn and spring losses
are less than expected from the amount of streamflow. The relation to streamflow
was much stronger in the recent decade. Likewise, output of nitrate during summer
and autumn is less than expected and during winter and spring is more than expected
on the basis of water loss. Outputs of potassium and ammonium on a seasonal basis
were more variable with regard to streamflow during the early decade. Ammonium
outputs are less so currently; the reduction of potassium concentration during the
summer (Fig. 34) produces an obvious drop in streamwater output assumed to be
from biologic utilization and storage within the forested ecosystem (Table 25).
Monthly Variations 109

Table 23 Seasonal volume-weighted average streamwater values for Watershed 6 of the Hubbard
Brook Experimental Forest during 1963–2009
Summer Fall Winter Spring
October– December–
Obs June–September November February March–May
Solute Years mg/L µEq/L mg/L µEq/L mg/L µEq/L mg/L µEq/L
Calcium 47 0.91 45.4 0.94 46.9 0.98 48.9 0.96 47.9
Magnesium 47 0.24 19.8 0.26 21.4 0.27 22.2 0.25 20.6
Potassium 47 0.11 2.8 0.21 5.4 0.19 4.9 0.23 5.9
Sodium 47 0.76 33.1 0.77 33.5 0.77 33.5 0.66 28.7
Aluminuma 40 0.27 20.3 0.29 21.8 0.30 22.5 0.33 24.8
Ammonium 46 0.020 1.1 0.023 1.3 0.017 0.9 0.016 0.9
Hydrogen ionb 47 0.011 11.3 0.010 10.0 0.011 10.8 0.013 12.4
pH 4.95 5.00 4.97 4.91
Sulfate 46 4.94 102.8 5.01 104.2 4.82 100.3 4.75 98.8
Nitrate 46 0.17 2.7 0.27 4.4 1.08 17.5 1.25 20.2
Chloride 46 0.39 11.0 0.50 14.1 0.48 13.5 0.44 12.4
Phosphate 38 0.003 0.1 0.002 0.1 0.003 0.1 0.003 0.1
DOCc 15 2.8 16.9 2.5 14.9 1.9 11.6 2.0 12.1
Dissolved 46 4.15 na 4.17 na 4.02 na 3.49 na
silica
Totals 14.78 (mg/L) 14.94 (mg/L) 14.88 (mg/L) 14.41 (mg/L)
Cations 133.7 (µEq/L) 140.2 (µEq/L) 143.7 (µEq/L) 141.1 (µEq/L)
Anions 133.4 (µEq/L) 137.7 (µEq/L) 143.0 (µEq/L) 143.6 (µEq/L)
a
Aluminum valence assumed to be +2 (Buso et al. 2000)
b
Hydrogen ion in µg/L VWA-concentration
c
DOC = dissolved organic carbon at 6 µEq/mg-C

Monthly Variations

A greater insight into the biogeochemistry of these forest ecosystems can be gained
by examining the nutrient budgets on a monthly scale. Three patterns of input–output
emerge: (1) those in which input exceeds output in every month, (2) those in which
output exceeds input in every month, and (3) those in which there is a crossover
from input dominance to output dominance on a monthly basis. These patterns are
not related to whether the elements have a gaseous component or are entirely sedi-
mentary. For example, calcium, phosphorus, and potassium have sedimentary
cycles, yet they exhibit different patterns; sulfur and nitrogen, however, both with
prominent gaseous phases, exhibit two patterns. Instead of this simple assumption,
a complex interaction involving such factors as input, effect of biologic activity, and
annual variations in weather are determining factors.

Output Consistently Greater Than Input

Calcium exemplifies a nutrient in which the output exceeds the input each month of
the year on a long-term basis (Fig. 38). The effects of increased streamflow are
110 4 Input–Output Budgets

Table 24 Seasonal average bulk precipitation inputs for Watershed 6 of the Hubbard Brook
Experimental Forest for two decades (1964–1973 and 2000–2009)
Summer Fall Winter Spring
October– December–
June–September November February March–May
1964– 2000– 1964– 2000– 1964– 2000– 1964– 2000–
Substance 1973 2009 1973 2009 1973 2009 1973 2009
Precipitation 120.5 128.3 114.9 153.5 114.9 108.2 110.5 114.4
(mm/ha)
g/ha g/ha g/ha g/ha
Calcium 166.9 77.1 188.6 72.0 109.4 55.8 233.7 137.3
Magnesium 43.5 18.9 80.1 27.9 33.0 16.8 50.8 32.3
Potassium 75.1 53.4 89.0 62.2 38.1 22.0 67.2 63.3
Sodium 99.0 56.8 238.1 145.3 124.9 121.3 131.8 106.9
Aluminuma 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Ammonium 304.6 229.6 226.7 190.2 167.3 105.8 284.2 258.9
Hydrogen ionb 107.7 37.9 83.1 39.2 54.5 27.9 84.6 32.3
Sulfate 4,519.8 1,873.8 2,795.4 1,465.7 1,882.3 866.0 3,404.9 1,564.4
Nitrate 1,566.4 1,104.7 1,731.4 1,434.5 1,653.5 1,270.2 1,920.4 1,381.7
Chloride 457.1 120.8 571.6 292.1 365.0 223.8 402.8 187.1
Phosphatec 15.8 6.5 10.5 3.2 11.3 1.9 37.2 9.7
DOCd na 1,513.6 na 908.7 na 487.0 na 919.1
Dissolved 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
silicaa
a
Aluminum and dissolved silica are trace in precipitation
b
Hydrogen ion is in µg/ha
c
Phosphate data not available prior to 1971: used 1972–1981 data for early decade
d
Seasonal DOC flux not available prior to 1990

clearly shown by the peak chemical outputs in April and a smaller peak in
November–December. During a specific year, the monthly input and output may be
quite variable, and monthly outputs may not exceed inputs, particularly during sum-
mer. Magnesium, sodium, aluminum, and dissolved silica all have patterns similar
to calcium; i.e., outputs consistently exceed inputs on a monthly basis (Fig. 38).

Input Consistently Greater Than Output

In sharp contrast to calcium, ammonium is characterized by monthly inputs that

consistently exceed outputs on a long-term basis (Fig. 39). Monthly streamwater
outputs of ammonium are relatively constant and low, with a small peak in April,
driven by large water outputs. The pattern of ammonium inputs may be highly vari-
able for a specific year, but it becomes much clearer and distinctly sinusoidal with
long-term data. The monthly input during late autumn and winter is relatively low.
Monthly Variations 111

Table 25 Seasonal average streamwater exports for Watershed 6 of the HBEF for two decades
(1964–1973 and 2000–2009)
Summer Fall Winter Spring
October– December–
June–September November February March–May
1964– 2000– 1964– 2000– 1964– 2000– 1964– 2000–
Substance 1973 2009 1973 2009 1973 2009 1973 2009
Streamflow 31.5 42.6 63.6 108.9 52.3 67.7 157.2 138.5
(mm/ha)
g/ha g/ha g/ha g/ha
Calcium 398.0 268.9 872.4 672.3 748.1 426.7 2,130.8 840.4
Magnesium 97.6 80.5 238.1 213.3 192.1 136.3 517.4 258.5
Potassium 46.5 37.6 147.6 207.1 113.1 97.0 396.5 259.4
Sodium 259.4 309.0 572.8 767.8 444.5 489.5 1,118.4 885.1
Aluminuma 59.3 80.7 216.9 256.6 140.3 142.0 546.6 332.8
Ammonium 14.1 2.2 47.0 5.5 19.4 3.4 40.9 7.1
Hydrogen 4.7 3.4 7.6 8.3 7.7 4.9 25.2 11.4
ionb
Sulfate 2,003.3 1,534.7 4,130.3 3,974.4 3,096.6 2,499.2 9,064.0 4,880.2
Nitrate 132.5 7.1 445.5 40.8 1,156.9 140.2 3,603.4 362.0
Chloride 149.3 141.8 382.7 463.7 291.4 278.1 820.4 496.9
Phosphatec 1.1 0.5 2.5 0.9 1.1 0.5 4.4 1.3
DOCd na 1,116.3 na 2,700.4 na 1,221.6 na 2,634.4
Dissolved 1,399.4 1,663.2 3,030.4 4,070.8 2,246.0 2,615.5 5,498.5 4,766.6
silica
a
Aluminum data from 1964–1970 (7-years) and 1976–1978 (3-years) used for early decade
b
Hydrogen ion in µg/ha
c
Phosphate data not available prior to 1971: used 1973–1982 data for early decade
d
Seasonal DOC flux not available prior to 1990

This pattern may reflect a reduced rate of ammonia input into the atmosphere from
decay in the soil because of the lower temperatures, or the transfer of ammonia to
the atmosphere may be reduced by the presence of a snowpack. Likewise, higher
inputs during the spring and summer may reflect the general increase in biologic
activity of the soil and the elevated generation of ammonia. However, because
ammonia may be converted rapidly to nitrate, N2O, N2, etc., it is difficult to inter-
pret these monthly fluxes. The overall pattern of monthly fluxes for hydrogen and
phosphate are similar to ammonium (input exceeds output) (Fig. 39). Calculated
output values for ammonium and phosphate are compromised somewhat because
many streamwater values are at or below method detection levels (Buso et al.
2000).
A maximum phosphate concentration in bulk precipitation often occurs in May
(Fig. 39) due to pollen contamination. If pollen contamination were obvious, the
sample would be rejected, but small amounts of pollen may not be visible and may
decompose contributing to elevated concentrations of phosphate (and potassium).
112 4 Input–Output Budgets

Fig. 38 Average volume-weighted monthly bulk precipitation inputs and streamwater outputs for
calcium, magnesium, sodium, aluminum, and dissolved silica for Watershed 6 of the Hubbard
Brook Experimental Forest during 1963–2009
Monthly Variations 113

140
120
(mol/ha-mo) ±1 SD

100 Hydrogen Ion

Average Flux

80
60
40
20
0

500
Ammonium
(g/ha-mo) ±1 SD
Average Flux

400
300
200
100
0
(g/ha-mo) ±1 SD

120
Average Flux

100 Phosphate
Stream Output
80
Precipitation Input
60
40
20
0
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Month of Year
Fig. 39 Average volume-weighted monthly bulk precipitation inputs and streamwater outputs for
hydrogen ion, ammonium, and phosphate for Watershed 6 of the Hubbard Brook Experimental
Forest during 1963–2009

Crossover Patterns

Long-term potassium inputs barely exceed outputs from mid-June through mid-
September, but during the remainder of the year, monthly outputs are larger than
inputs, with a large peak in April (Fig. 40). This pattern is of interest, for as sug-
gested earlier the potassium budget (particularly potassium in stream water) is par-
ticularly sensitive to seasonal biologic control within the ecosystem (Likens et al.
1994; Table 25; Fig. 40). As mentioned above, the forest vegetation of these water-
sheds normally breaks dormancy in May, and the deciduous leaves are shed about
mid-October. Apparently the biotic portion of the ecosystem actively extracts potas-
sium from the soil and drainage waters during this growing season (cf. Johnson
et al. 1969). The “crossover times” separating the NHF values (Fig. 40) were not
114 4 Input–Output Budgets

(g/ha-mo) ±1 SD
800 Potassium
Average Flux

Stream Output
600 Precipitation Input
400
200
0
16000
14000 Sulfate
(g/ha-mo) ±1 SD

12000
Average Flux

10000
8000
6000
4000
2000
0
1400
1200 Chloride
(g/ha-mo) ±1 SD

1000
Average Flux

800
600
400
200
0
Dissolved
(g/ha-mo) ±1 SD

6000
Average Flux

Organic
4000 Carbon

2000

0
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Month of Year
Fig. 40 Average volume-weighted monthly bulk precipitation inputs and streamwater outputs for
potassium, sulfate, chloride, and dissolved organic carbon for Watershed 6 of the Hubbard Brook
Experimental Forest during 1963–2009

identifiable in the “long-term” monthly budgets by 1969 (1963–1969), which

presumably was a residual effect of the drought years (1963–1965) on the cumula-
tive average (see Table 17). Longer-term data show that the large monthly export of
potassium in stream water during March through May of each year generally sur-
passes the average annual input to these forest ecosystems (Fig. 40; Tables 18, 24,
and 25). Nevertheless, the monthly pattern of potassium inputs and outputs for any
Monthly Variations 115

7000
Average Flux (g/ha-mo) ±1 SD

6000 Nitrate

5000 Stream Output

Precipitation Input

4000

3000

2000

1000

0
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Month of Year

Fig. 41 Average volume-weighted monthly bulk precipitation inputs and streamwater outputs for
nitrate in Watershed 6 of the Hubbard Brook Experimental Forest during 1964–2009

specific year may be highly variable. For example, as mentioned above, inputs
exceeded outputs during the drought years of 1963–1965, but less dramatic,
although complicated, patterns occurred in other years. Potassium has a relatively
stable annual NHF pattern despite its crossover pattern on a monthly basis.
The monthly pattern for sulfate, chloride, and dissolved organic carbon is gener-
ally similar to that of potassium (Fig. 40) except that the long-term results show that
the crossover from input to output dominance in sulfate, chloride, and dissolved
organic carbon is more pronounced than for potassium. The monthly input of sulfate
and dissolved organic carbon bulk precipitation have a dampened symmetrical, sinu-
soidal pattern, with a peak in June–July–August and a low in January–February.
Monthly nitrate flux requires special consideration (Fig. 41). The monthly pat-
tern is generally consistent but distinctly different than the pattern for ammonium
(Fig. 39), although the annual patterns for these two nitrogenous ions are similar
(i.e., inputs exceed outputs; Table 18). The pattern for nitrate after 8 years (1964–
1972) showed that inputs exceeded outputs during every month except April, even
though after 8 years the average output for April was 55 % higher than it was in
1968–1969. However, during 1969–1970, 1970–1971, and 1973–1974, a distinct
change in the annual pattern occurred (Table 17). During these 3 years, the long-
term annual relationship, where input exceeds output, was reversed. The input dur-
ing August, September, October, November, and January 1969–1970 was less than
the long-term average, but it was the relatively large outputs in April and to a lesser
116 4 Input–Output Budgets

extent in December and February that overwhelmed the long-term pattern of the
annual budget. As mentioned previously, this change in pattern during 1969–1970
and 1973–1974 may have been related to the unusual occurrence of soil frost
increasing streamwater losses. Other disturbances have occurred in the long-term
record (e.g., Bernal et al. 2012). The long-term pattern (1964–2009) shows that
inputs of nitrate exceed output during every month except March and April, although
monthly values tend to be highly variable (Fig. 41).
Marked changes in concentrations of nitrate in precipitation and stream water
during the period of study (Fig. 28; Tables 6 and 10) also have affected the long-
term monthly and seasonal flux (e.g., Fig. 41). As a result of these perturbations in
precipitation and streamwater flux, the resolution of the “average” long-term
monthly pattern for nitrate is not particularly useful. Nevertheless, based on 47 years
of data, monthly inputs of nitrate exceed outputs during all months except April.
Inputs and outputs are essentially balanced during March (Fig. 41).

Relationship of Annual Mass Output of Dissolved Substances

to Annual Streamflow

One of the major findings from the long-term Hubbard Brook Ecosystem Study is
the highly significant direct relationship of annual gross output of cations to annual
streamflow (Figs. 42 and 43). In retrospect, this finding should not have been too
surprising since chemical concentrations are quite constant with stream discharge
(Johnson et al. 1969, Fig. 48). Nevertheless, the direct relation of gross annual
export of most ions to the annual streamflow is highly predictable (r2 = >0.95), but
because concentrations have changed dramatically with time, there are interesting
but important subtleties. For example, the relation for gross output of calcium (Eq/
ha-year) to annual discharge (mm) is strongly positive and highly significant
(p = <0.01). When this relation is partitioned into shorter temporal periods reflecting
the change in streamwater chemistry with time, an interesting pattern emerges
(Fig. 42a). The slope of the relations has flattened with time, and linear regressions
for the three periods after 1976 no longer go through the origin (Fig. 42a). For
example, based on the linear regressions for 1964–1975 and 2000–2009, more than
2 times as much calcium would have been lost during the earlier period than now at
a flow of ~1,500 mm/year. This result is due primarily to the depletion of calcium
from the watershed-ecosystem by decades of acid rain (see Chap. 3).
Because of the regulation effect of the forest ecosystem on chemical concentra-
tions in stream water as it emerges from watersheds of the HBEF, even after the
sizeable impact of a pollutant, such as acid rain, mass output can be predicted
rather reliably simply from a knowledge of annual hydrologic output. This rela-
tionship is useful for (1) characterizing conditions in an undisturbed forest and (2)
providing baseline information for evaluating biogeochemical conditions in a for-
est disturbed by acid rain (but, see Likens and Buso 2012). These important rela-
tions also provide critical information to land-use, hydrologic, and water-quality
planners.
Relationship of Annual Mass Output of Dissolved Substances… 117

Fig. 42 Relationship between annual streamflow and annual gross outputs of (a) calcium and (b)
magnesium for different periods in Watershed 6 of the Hubbard Brook Experimental Forest during
1963–2009

Annual Variation in Mass Output of Dissolved Substances

On a mass basis (kg/ha), sulfate, dissolved silica, and DOC are the dominant sub-
stances exported in stream water, whereas on an ionic basis, sulfate and calcium
dominate, although both of these ions have declined markedly (see section on
118 4 Input–Output Budgets

Fig. 43 Relationship between annual streamflow and annual gross output of sum of base cations
during different periods in Watershed 6 of the Hubbard Brook Experimental Forest during
1963–2009

Dilution in Chap. 3). The annual gross output of total dissolved substance (excluding
dissolved organic carbon) during 1963–2009 averaged 96.6 kg/ha; excluding
dissolved silica, it was 80.0 kg/ha (Table 18). With time, sulfate has declined, but
its predominance is still striking. In 2009–2010, the annual gross output of sulfate
is 36.5 kg/ha or 759 Eq/ha, which represents 26 % of the total dissolved substances
(including DOC) or 87 % of the anionic equivalency (not accounting for DOC).
The annual gross output of total inorganic dissolved substances varies relatively
little from year to year. Annual mass export varied by threefold and equivalency by
2.9-fold, which could be largely explained by the range (2.7-fold) in annual stream-
flow (Fig. 5). The largest output of dissolved solids occurred during the 1973–1974
water-year (119 % greater than the long-term average), and the drought year of
1964–1965 had the smallest output (26 % less than the long-term average). The
average annual output of cations during the period 1963–2009 was calculated to be
1,296 Eq/ha, and the average annual output of anions during 1963–2009 was
1,325 Eq/ha. The poorest agreement in equivalents, 7 and 8 % (difference based on
the smaller value), between cations and anions, respectively, occurred in 1964 and
1975, and the best agreement, <1 %, was in 1990 and 2008.
Hydrogen ion and nitrate inputs in bulk precipitation were not as closely corre-
lated as hydrogen ion outputs in stream water and nitrate inputs in bulk precipitation
(Fig. 44). Annual inputs of dissolved inorganic nitrogen (DIN) were more closely
correlated to annual outputs of DIN in stream water (Fig. 45). The explanation for
Relationship of Annual Mass Output of Dissolved Substances… 119

1400

Hydrogen Ion Flux (eq/ha-yr)

1200

1000

800
1:1
600

400

200

0 100 200 300 400 500 600 700

Nitrate Inputs (eq/ha-yr)

Fig. 44 Relationship between hydrogen ion inputs and nitrate inputs in bulk precipitation (open
circle) and hydrogen ion inputs and nitrate outputs in stream water (filled circle) in Watershed 6 of
the Hubbard Brook Experimental Forest during 1964–2009

800
Dissolved Inorganic Nitrogen
Annual Inputs or Outputs (eq/ha-yr)

600

400

200

Precipitation
Stream Water
0

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Year

Fig. 45 Annual bulk precipitation inputs (open circle) and streamwater outputs (filled circle) of
dissolved inorganic nitrogen in Watershed 6 of the Hubbard Brook Experimental Forest during
1964–2009
120 4 Input–Output Budgets

Fig. 46 Sampling procedure employed in the measurement of total nutrient losses from a
watershed-ecosystem. A = water sample for dissolved substances, B = suspended and bed load
dropped in basin, C = water sample filtered through a fine-meshed net, D = water sample, after fil-
tered as C, through the finer pore of membrane filter. Total losses = dissolved substances (A) + par-
ticulate matter (B + C + D). Modified from Bormann et al. (1969)

the marked decline in streamwater concentrations and outputs of DIN during the
long-term record (Fig. 28) currently is the subject of much research at HBEF
(Likens 2004; Bernhardt et al. 2005; Judd et al. 2011; Bernal et al. 2012).

Particulate Matter

Measurement of Mass Output

In addition to dissolved chemical substances, both organic and inorganic particulate

matter may be exported from ecosystems in stream water. Particulate matter is
removed from the watershed as suspended load carried by turbulent water and as
bed load rolled, slid, or bounced along the stream bed. Heavier particulate matter
collects in the ponding or stilling basin behind the weir where, over an 8-year
period, it was periodically dug out, weighted, proportionally sampled, oven dried,
and analyzed for dry weight and organic and inorganic content (Bormann et al.
1969, 1974). Suspended particulate matter that passed over the weir was periodi-
cally sampled using a 1-mm mesh net and by passing a sample of water through the
net and then filtering it through a 0.45-µm pore Millipore filter at 40 psi. Samples
were then analyzed for dry weight and organic and inorganic content and expressed
as concentrations per unit volume of water. These concentrations were used in com-
bination with the hydrologic discharge record to determine the export of materials
over the weir. These three components, basin, netted, and filtered, were combined to
give total particulate export from the ecosystem (Fig. 46).
It is noteworthy that this estimate of total particulate export includes both sus-
pended particulate matter and bed load, whereas most reports of particulate export
from ecosystems are concerned with suspended load alone. If the heavier inorganic
Relationship of Annual Mass Output of Dissolved Substances… 121

material that collects behind the weir were considered as a measure of bed load,
then bed load constitutes about 55 % of the total particulate export.

Annual Loss

Erosion and transport of particulate matter from these forested watersheds are rela-
tively low even though the watersheds are on steep slopes (12–13°) and are subject
to large amounts of precipitation (1,434 cm/year). Although the glacial till in these
watersheds is relatively resistant to erosion (Hunt 1967), it is primarily the living
and dead biomass that minimizes the loss of particulate matter by regulating the
amount, timing, and effect of moving water within the undisturbed ecosystem.
Organic debris dams in the stream channels of these headwater ecosystems appar-
ently play a major role in minimizing the export of particulate matter (Bilby and
Likens 1980).
The undisturbed ecosystems tightly regulate erosion and transport and lose only
about 33 ± 13.4 kg/ha of particulate matter with annual runoff (Table 26).
This material averages about 33 ± 4 % (standard error) organic, but the organic–
inorganic proportion is strongly influenced by flow rate of the stream, with higher
organic proportions associated with higher flow rates (Fig. 47; Bormann et al. 1974).

Seasonal Variation in Erodibility

Although the maximum output of particulate matter occurs during high-discharge

periods, summer flows are more effective in exporting particulate matter than are
flows of about the same velocity when vegetation is dormant (Bormann et al. 1974).
The 30 % greater erodibility during the summer is probably related to increased
biologic activity and decomposition within the ecosystem. In winter, the stream
channels are often stabilized by ice and snow, and in addition, precipitation gener-
ally falls as snow, with low potential energy of impact. Rarely (twice in 50 years),
there have been ice flows that scour the stream channel and move large amounts of
organic and inorganic bed load materials.

Particulate Matter Versus Dissolved Substance Export

There is a sharp contrast in the response of dissolved substance and particulate mat-
ter export to discharge rate of the stream. Concentrations of dissolved substances
are relatively little affected by flow rates, whereas particulate matter concentrations
are directly and exponentially related to stream discharge (Figs. 47 and 48). As a
consequence, the bulk of particulate matter is moved during storms. Output of dis-
solved substances is therefore closely related to annual output of water, whereas
removal of particulate matter is more of a stochastic process, strongly related to the
occurrence of random storms. This pattern is reflected in the fact that during the
122 4 Input–Output Budgets

Table 26 Annual particulate matter output in kilograms of oven-dry weight organic and inorganic
materials per hectare for Watershed 6a circa 1970
Source of output Organic Inorganic Total
1965–1966 Ponding basin 2.12 1.77 3.89
Net 0.34 0.01 0.35
Filter 1.37 1.28 2.65
Total 3.83 3.06 6.89
1966–1967 Ponding basin 13.41 17.07 30.48
Net 0.39 0.01 0.40
Filter 2.72 2.95 5.67
Total 16.52 20.03 36.55
1967–1968 Ponding basin 3.83 5.93 9.76
Net 0.43 0.01 0.44
Filter 2.61 2.82 5.43
Total 6.87 8.76 15.63
1968–1969 Ponding basin 4.61 8.31 12.92
Net 0.42 0.01 0.43
Filter 2.57 2.81 5.38
Total 7.60 11.13 18.73
1969–1970 Ponding basin 11.28 30.67 41.90
Net 0.40 0.01 0.41
Filter 3.30 3.69 6.99
Total 14.98 34.37 49.30
1970–1971 Ponding basin 2.70 2.75 5.45
Net 0.39 0.01 0.40
Filter 1.77 1.78 3.55
Total 4.86 4.54 9.40
1971–1972 Ponding basin 3.66 2.52 6.18
Net 0.37 0.01 0.38
Filter 1.76 1.80 3.56
Total 5.79 4.33 10.12
1972–1973 Ponding basin 17.88b 77.40 95.28
Net 0.60 0.01 0.61
Filter 10.09 13.61 23.70
Total 28.57 91.02 119.59
8-Year average Ponding basin 7.44 18.30 25.74
Net 0.42 0.01 0.43
Filter 3.27 3.84 7.11
Total 11.13 22.15 33.28
a
Ponding basin output = particulate matter collected in a ponding basin of the weir upstream from
the V-notch; net output = suspended particulate, >1 mm, that passes over the V-notch of the weir;
filter output = suspended particulate matter, >0.45 µm <1 mm, that passes over the V-notch of the
weir
b
Underestimate

period 1965–1973, particulate matter output ranged from 7 to 120 kg/ha-year and
was highly correlated with the occurrence of individual storms in a particular year.
An analysis of more than 4 years of data indicated that 86 % of the total particulate
matter was exported in 1.6 % of the total time in the period and with 23 % of the
Relationship of Annual Mass Output of Dissolved Substances… 123

Fig. 47 Percent organic material in basin collections from Watershed 6 plotted against the highest
flow recorded during the period. 1 ft3/s = 28.3 L/s. After Bormann et al. (1974). Circa 1974

Fig. 48 General
relationships between the
concentration of dissolved
substances and particulate
matter and stream discharge
in the Hubbard Brook
Experimental Forest. After
Bormann et al. (1969)

water, whereas 16 % was exported in 0.0025 % of the total time and with 0.2 % of
the water (Bormann et al. 1974).
These relationships have important implications for the export of certain ele-
ments. Most of the chemical elements with a sedimentary cycle are exported from
124 4 Input–Output Budgets

Table 27 Average annual gross output of some elements as dissolved (D) and particulate (P)
substances from Watershed 6 of the Hubbard Brook Experimental Forest during 1966–1967 to
1969–1970a
Particulate Dissolved
Particulate + dissolved % of P + D % of P + D
Element element total, kg/ha kg/ha element total kg/ha element total
Aluminum 3.37 1.38 40.9 1.99 59.1
Calcium 13.93 0.21 1.7 13.7 98.3
Chloride 4.58 – 0 4.58 100
Iron 0.89 0.64 72 0.25b 28
Magnesium 3.34 0.19 5.7 3.15 94.3
Nitrogen 4.01 0.11 2.7 3.90 97.3
Phosphorus 0.019 0.012 63.2 0.007 36.8
Potassium 2.40 0.52 21.7 1.88 78.3
Silicon 23.8 6.19 26.0 17.6c 74.0
Sodium 7.48 0.25 3.3 7.23 96.7
Sulfur 17.63 0.03 0.2 17.6 99.8
Carbond 12.3 3.98e 32.4 8.35f 67.5
a
Particulate matter losses during 1966–1967 to 1969–1970 are modified from Bormann et al.
(1974). Dissolved substance losses are averages during the period 1963–1974
b
Average of Smith et al. (1986) and Fuss et al. (2011)
c
Assuming that dissolved silica is in the form of SiO2
d
Organic only
e
Assuming organic matter is 40 % C
f
Assuming an average carbon concentration of 1 mg/L (Hobbie and Likens 1973)

the watershed-ecosystem primarily in the dissolved form, as shown above; however,

the majority of the iron and phosphorus is lost in the form of particulate matter
(Table 27). In contrast to other elements, therefore, the total output of iron and phos-
phorus is more directly related to stream discharge rate and is less predictable from
annual streamflow. Moreover, phosphorus in particulate form is mostly unavailable
to aquatic microorganisms and plants within the stream ecosystem unless it first
undergoes decomposition or mineralization.
Our data show that in the forest ecosystems at HBEF, solution losses are the
major force in the geologic process of fluvial denudation. In terms of gross export
of mass, dissolved substances—about 119 kg/ha, including dissolved organic mat-
ter during 1964–2009—are about eight times greater than particulate matter losses
(Tables 26 and 27).

The Role of Debris Avalanches in Landscape Denudation

Our data suggest that headwater watersheds, with their well-developed forests, are
gradually lowered in place by the action of solution coupled with slow mass move-
ments. Both of these actions deliver material to the stream, where it is removed by
The Role of Debris Avalanches in Landscape Denudation 125

erosion and transportation. The possibility that infrequent debris avalanches play an
equal or greater role in moving material downslope in headwater watersheds should
also be considered. For example, Hack and Goodlett (1960) have shown in the
Appalachian Mountains of Virginia and West Virginia that infrequent but cata-
strophic landslides play a major role in removing material from low-order water-
sheds and that this activity is independent of any restraining influence by the biota.
Flaccus (1958a, b) has shown essentially the same thing for the White Mountains of
New Hampshire. Our data on stream losses allow a rough comparison of these very
different denudational processes.
Flaccus (1958a, b) mapped 543 debris avalanches on aerial photos of the White
Mountains. His study area, most of which is heavily forested, is adjacent to the
Hubbard Brook Valley and contains extensive stretches of the Littleton formation
(now mapped as the Rangeley Formation in our watershed). The mapped avalanches
occur in about equal proportions on slopes of the Littleton formation and slopes
underlain by plutonic rocks. Flaccus examined in detail a range of avalanches,
thought to be a fair sample of the total, and calculated the average weight of material
moved downslope as 21,800 metric tons. In those slides fresh enough to permit
location of the topmost elevation, it was found that 100 % occurred above the 610-m
elevation contour. Within his study, we have calculated an area of about 128,000 ha
above the 610-m elevation. Using Flaccus’ data for numbers and dates of avalanches
and average amounts of material moved in each event, we have estimated that
462 kg/ha-year are moved downslope by debris avalanches occurring above 610 m
in the White Mountains (Bormann et al. 1969). This estimate is about triple our
estimate of 200 kg/ha-year for dissolved plus particulate export from headwater
watersheds of the HBEF. This comparison suggests that debris avalanches are at
least equivalent to the combined action of solution and slower mass movements in
lowering the surface of low-order watersheds in the White Mountains.
However, two additional factors must be considered: (1) Disturbances that
destroy the biotic stability of the ecosystem, such as fire, can lead to greatly increased
losses by solution, mass wasting, and erosion; and (2) the incidence of avalanches
per unit area increases sharply with increasing elevation and increasing slope
(Flaccus 1958a, b). In the White Mountains, only 18 % of the recorded avalanches
occur wholly below 910 m.
In the Hubbard Brook Valley, it seems unlikely that debris avalanches have
occurred within recent times in the area currently occupied by northern hardwood
forest. With minor exceptions, elevations within the Hubbard Brook Valley are
below 910 m, and except for a few sites of very restricted area, our slopes do not
meet Flaccus’ minimum slope requirement of 25° or is there any evidence of recent
or old debris avalanches. It seems safe to conclude that debris avalanches have not
played a significant role in the denudation of the relatively gentle slopes of the
HBEF, at least during the last millenium.
Apparently, denudation in the area of the Hubbard Brook Valley is primarily a
result of the less dramatic action of solution and erosion in combination with slower
mass movements, such as creep, which may deliver materials to the streambeds.
This result, in turn, suggests that the weathering rind in our watersheds is fairly old.
126 4 Input–Output Budgets

Locally, this is shown by well-developed soil profiles; however, in some places, the
surface is subject to shallow stirring by burrowing animals and from trees uprooted
by the wind or during ice storms.

Long-Term Changes in Input–Output Budgets

Because of long-term changes in the chemistry of precipitation and stream water,

there have been associated changes in the input–output budgets for the watershed-
ecosystems at HBEF. In general, the overall patterns for the input–output relations
have remained remarkably the same with time, but there are some important differ-
ences in detail. The long-term data for watershed-ecosystems at HBES show signifi-
cant net gains in NHF (precipitation input minus streamflow output) for dissolved
inorganic nitrogen, hydrogen ion, and phosphate. The discrepancy between the sum
of anions and cations is small for bulk precipitation and streamwater fluxes in two
decades spanning the long-term study (Tables 20 and 21).
Importantly, current streamwater concentrations of nitrate in Watershed 6 are the
lowest on record (Fig. 28), even though forest biomass accumulation is very low or
negative (Fig. 26; Likens 2004; Lindenmayer and Likens 2010; Bernal et al. 2012).
Significant net losses of calcium, magnesium, sodium, potassium and sulfate, alu-
minum, organic carbon, and dissolved silica occur, but streamwater outputs of cal-
cium, magnesium, and sodium are appreciably smaller than during 1964–1973 and,
as a percentage of total cation loss, have become smaller (Tables 18 and 21).
Chloride shows a very small long-term net loss, but annual values are highly vari-
able as discussed above (Tables 18 and 19).
The balance between the sum of cations and the sum of anions for streamwater
outputs in the earlier decade (Table 21) is not good, but assigning a charge for the
organic anion and dissolved aluminum simultaneously was problematic. Making
such adjustments, if data were available, probably would reduce the discrepancy to
a smaller error.
Chapter 5
Weathering

Both the qualitative and the quantitative changes in water chemistry elicited during
the passage of water through a forest ecosystem are related in part to the process of
chemical weathering. The bulk ionic composition of water entering the ecosystems
of the HBEF in precipitation is characterized by acid salts, primarily H2SO4 and
HNO3. In contrast, water leaving these systems is characterized mainly by neutral
salts, such as CaSO4 and Na2SO4, and to a lesser extent by nitrates and chlorides.
This qualitative change in chemistry exemplifies the general chemical weathering
reactions. This reaction and some of the chemical pathways in an open system, such
as the watershed-ecosystems of the HBEF, are shown in Fig. 49. As described in
Chap. 1, chemicals generated from weathering may not quickly cross the ecosystem
boundary as they may be taken up by vegetation, participate in cation exchange
reactions in the soil, etc., and, therefore, are referred to as “weathering release.”
Nevertheless, the release of ionic substances from primary minerals by chemical
weathering is an important source of nutrients in soils, soil water, and output flux in
stream water from forest ecosystems (e.g., Likens et al. 1967, 1994, 1998, 2002a;
Mast and Drever 1990; Williams et al. 1986; Yuretich and Batchelder 1988; Johnson
et al. 1994; S. Bailey et al. 2003; Houle et al. 2012). These chemicals also are
important contributors to the ecosystem mass balance, but often these contributions
are difficult to quantify, especially since the amounts may be “masked” by biotic
uptake and long-term storage (see below). Mineral weathering is a one-way reac-
tion, and HBEF soils are composed of minerals derived from igneous and metamor-
phic bedrock. The soil environment is much different, however, than the igneous
and metamorphic environments where these minerals formed. Hence, the minerals
are inherently unstable tending to decompose on reaction with water and enhanced
by the action of acids, both mineral and organic.
As H+ is consumed within the system, base cations (M+) may be leached from
various components of the system, and primary minerals (M+X) are converted into
soluble ions and left behind as secondary minerals or coatings on soil surfaces
(Fig. 49). When mechanical erosion or other disturbance is sufficiently slow, the gla-
cial drift and to a lesser extent bedrock interact with organic substances to form soil.

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2_5, 127

© Springer Science+Business Media New York 2013
128 5 Weathering

Fig. 49 M+ represents a base cation, such as calcium, and X is an ionic exchange substrate, usually
a primary silicate mineral but also includes such derivatives as soil minerals or soil humates. The
above diagram illustrates that as H+ is added to the system from external and internal sources, base
cations (M+) are leached from the system, and primary silicate minerals (M+X) are transformed
into secondary minerals (H+X)

The rate at which the chemical weathering reaction proceeds may be operationally
defined as the rate at which hydrogen ion is supplied to the system (see Figs. 49 and 50;
Driscoll and Likens 1982).
Following the final retreat of the Pleistocene continental glacier some 14,000
years ago in the Hubbard Brook Valley, chemical and physical weathering of freshly
exposed rock and mineral surfaces was thought to be relatively high (Likens and
Moeller 1985, p. 397). With time, however, chemical weathering flux declined as
labile, relatively easily weatherable minerals were depleted or reduced, and the cli-
mate/soil/vegetation complex became more stable. Currently, weathering fluxes are
relatively quite low, even with recent increased acidity of atmospheric deposition.
These fluxes are difficult to measure quantitatively, particularly at the watershed
scale. Efforts to extrapolate results of laboratory studies of weathering to a water-
shed scale have not been successful (e.g., Nezat et al. 2004). Initially, one of the
ideas of the Hubbard Brook Ecosystem Study was that very difficult to measure,
large-scale ecosystem processes, such as weathering, could be estimated from mass
balances generated by the small watershed-ecosystem approach (Bormann and
Likens 1967). This approach proved to be difficult, however, for weathering at the
watershed scale because of uncertainty in other components of the mass balance and
gradually morphed into a somewhat broader concept for ecosystem analysis, net
soil release (see Likens et al. 1994, p. 79). Some estimates of weathering and net
soil release are compared in Table 28.
5 Weathering 129

Meteorologic Inputs
Bulk Precipitation +960
Dry Deposition +362

Weathering Reactions Net Forest Accumulation

Forest Biomass Forest Floor Total

Ca –1055 Ca + 405 + 70 + 475

Mg – 288 Mg + 57 + 17 + 74
Na – 252 Na + 7 < 1 + 7
K – 182 K + 147 + 9 + 156
S + 25 S – 75 – 50 – 125
Al – 211 Fe + 40 + 63 + 103
Fe – 78 NO3- ------ ------ – 47
P + 83 NH4+ ------ ------ + 144
P – 74 – 16 – 90

TOTAL –1957 TOTAL + 697

Stream Exports
Stream pH (H+) – 100
Stream alkalinity (HCO3-) + 126
Discrepancy in charge balance + 26
(organic anions, hydroxide ligands)
SUMMARY
Hydrogen Ion Sources + 2541
Hydrogen Ion Sinks – 2428
Budget Discrepancy + 113

Fig. 50  Hydrogen ion budget for Watershed 6 of the Hubbard Brook Experimental Forest circa
1982 (modified from Driscoll and Likens 1982). Hydrogen ion flux is in Eq H+/ha-year. Positive
values indicate hydrogen ion generation (source); negative values indicate consumption (sink).
It was not possible to distinguish dissolved inorganic nitrogen accumulation in the forest floor and
forest biomass

Initially at HBEF, we attempted to estimate weathering as the long-term cat-

ionic denudation rate, i.e., the rate at which cations are exported from the system
(Reynolds and Johnson 1972). A major insight gained from the ecosystem
approach relative to this estimate was that net sequestration of cations by the bio-
mass must be included in the mass-balance calculation. Moreover, the role of cat-
ion exchange and secondary mineral formation was identified, but could not be
quantified or differentiated. Recent data on the biogeochemistry of potassium
130 5 Weathering

Table 28  Some estimates of chemical weathering fluxes and net soil release rates in Watershed 6
of the Hubbard Brook Experimental Forest (mol/ha-year)
Element Weathering Net soil release References
K – 179 in 1964–1969; 60 Likens et al. (1994)
in 1987–1992
Na 250–370 207–241 Likens et al. (1998),
S. Bailey et al. (2003)
Ca 50–80; 54–132 175–701 Likens et al. (1998),
S. Bailey et al. (2003)
S 50 (max) 44 in 1964–1969; 128 Likens (2002a)
in 1993–1998
Cl 1.7–3.4 – Lovett et al. (2005)

(Likens et al. 1994), calcium (Likens et al. 1998), and sulfur (Likens et al. 2002a)
suggested that secondary minerals (particularly vermiculite for potassium), the
cation exchange complex for potassium and calcium, soil adsorption for sulfate,
and the biota play important roles in regulating the loss of these elements in stream
water on an annual basis at HBEF.
The full, mass balance for the calculation of cationic denudation on an annual
basis is

Pi + Wi = Si + ∆Bi ± ∆Oi ± ∆Xi ± ∆Mi (1)

where Pi is atmospheric input of element i, W is the weathering release from p­ rimary
minerals, S is dissolved loss in stream water, ΔB is net storage in biomass, ΔO is net
long-term storage in the soil organic matter pool, ΔX is the net change in the
exchangeable pool, and ΔM is the net change in the secondary mineral pool.
Bailey et al. (1996) used a detailed analysis of the isotopic composition of stron-
tium, coupled with a mass-balance analysis of both calcium and strontium at the
Cone Pond watershed, some 7 km from the HBEF, in an attempt to distinguish
between weathering release and net exchange with various ecosystem pools (e.g.,
cation exchange complex, forest floor). They estimated the weathering flux for cal-
cium to be only about 30 % of the value estimated by a standard mass balance.
Hyman et al. (1998) found a similar value for weathering using a different technique
involving the estimation of mineral weathering rinds on rock fragments.

Sources of Hydrogen Ion in the Weathering Reaction

Hydrogen ion, which is a major driver of the chemical weathering reaction, is sup-
plied from two sources at HBEF: one external to the ecosystem’s watershed bound-
aries and the other internal (Fig. 50). The external source is from acids supplied in
Sources of Hydrogen Ion in the Weathering Reaction 131

bulk precipitation and dry deposition (meteorologic input); the internal source is
from various biologic and chemical processes and transformations occurring within
the ecosystem. Currently, the meteorologic input of hydrogen ion at HBEF is mainly
in the form of H2SO4 and HNO3 (Tables 18 and 19). This external input of hydrogen
ion at HBEF has been decreasing over the past decades, but the long-term average
in bulk precipitation is 0.727 ± 0.24 (SD) × 103 Eq/ha-year (Table 18). If this were
the only source of hydrogen ions at HBEF (and the ecosystem were in a steady state,
which it is not), we might expect this value for consumption of hydrogen ion to be
more or less balanced by the mean net rate at which ionic calcium, magnesium,
potassium, sodium, and aluminum have been leached from the system in stream
water. In fact, more of these cations are removed from the ecosystem each year on
average (1.19 × 103 Eq/ha) in stream water than there are external hydrogen ions to
replace them (Table 18). The difference (0.462 × 103 Eq/ha-year) implies the action
of internally generated hydrogen ion. By this definition “internal” is taken to mean
everything that is not sensibly added by meteorologic input.
Under prevailing biologic and chemical conditions circa 1982, external and
internal generation of hydrogen ions play about equal roles in driving the weather-
ing reaction at HBEF (Fig. 50). This conclusion is based on the assumption that
most of the hydrogen ions entering the ecosystem from both external and internal
sources are consumed in the weathering reactions. This hydrogen ion balance was
done in 1982, but biotic and chemical conditions have changed significantly during
the period 1963–2009. For example, net forest biomass accumulation since 1982
has been small or negative (Siccama et al. 2007; Lindenmayer and Likens 2010; van
Doorn et al. 2011), and both precipitation inputs and streamwater outputs of several
dissolved solutes have declined significantly since 1982 (Likens and Buso 2012).
Some of the major internal sources of hydrogen ion at HBEF are as follows.

Carbon

Reduced carbon in the form of CHO compounds is continually oxidized in the soil
zone with the production of CO2 at partial pressures normally greatly exceeding
(thousands of μL/L) those found in the atmosphere (Fernandez et al. 1993; Brady
and Weil 2008). Some CO2 reacts with the soil water to form carbonic acid, which
has a potential to generate hydrogen ions. However, the soils at HBEF, at least in the
upper soil horizons, (Johnson et al. 1991b) are quite acidic (pH < 4.7). Under these
conditions H2CO3 dissociates only very slightly, and hence carbonation weathering
reactions (Carroll 1970) are inhibited. An additional potential source of hydrogen
ion at HBEF is from the production of organic acids by biologic activity within the
soil zone, such as citric, tartaric, tannic, and oxalic acids (McDowell and Likens
1988). The role of such organic acids in chemical weathering has undoubtedly been
important over the long term in ecosystems of the HBEF. However, the effects of
132 5 Weathering

Fig. 51  Long-term relationship between sum of base cations (calcium, magnesium, sodium, and
potassium) output and sum of acid anions (sulfate, nitrate) input in stream water of Watershed 6 of
the Hubbard Brook Experimental Forest (r2 0.93)

these organic acids have not been quantitatively assessed at the HBEF. There is no
obvious signature of organic acids in the charge balance between streamwater acid
anion and base cation exports from W6 (Figs. 17 and 51).

Nitrogen

Another aspect of the chemical weathering reaction at HBEF is the role of the
ammonium ion. It is well known to agriculturalists that additions of ammonium
have an acidifying effect on soils (e.g., Reuss 1976; Brady and Weil 2008). When
ammonium ion is fully oxidized to nitrate ion, two hydrogen ions are produced.
Ammonium is added to the ecosystem both by meteorologic input and by biological
decomposition within the system. The actual acidifying effect of ammonium in the
HBEF soils is unknown; however, it is estimated that only about 10–20 % of the soil
ammonium is oxidized (Melillo 1977). The rest is apparently taken up directly by
plants, which may release additional hydrogen ions (Reuss 1976). In the long term,
NH4–N contributes some 32 % (weight basis) of dissolved inorganic nitrogen
(DIN) in bulk precipitation inputs, but only 7 % of the DIN in streamwater outputs
(Table 18) reflecting the biogeochemical transformations occurring within the eco-
system. Although nitrification represents a substantial source of hydrogen ion, it is
not sufficient by itself to make up the observed deficit between cationic output and
external, meteorological hydrogen ion input (Fig. 50; Groffman et al. 1999, 2001;
Fitzhugh et al. 2003; Judd et al. 2007; Ross et al. 2012).
Estimates of Ecosystem Weathering Flux 133

Sulfur

Although sulfur concentrations in bedrock at HBEF can be high, even much higher
concentrations than average crustal rocks (Bailey et al. 2004), due to small amounts
of sulfide minerals, such as pyritite, Bailey et al. (2004) show that soil C horizon
concentrations of sulfur are low. This result is probably due to both low sulfur-
bearing rocks in the glacial drift and early postglacial depletion of iron sulfides.
These minerals, which are extremely reactive, are subject to oxidation in the soil
zone with the concomitant production of sulfuric acid (Carroll 1970; Likens et al.
2002a). In addition, microbially liable organic sulfur compounds, such as carbon-
bonded sulfur and ester sulfates, are subject to oxidation with the production of
sulfuric acid or sulfates (Mitchell et al. 1992; Zhang et al. 1999). Oxidation may
occur by strict chemical reactions, but most sulfur oxidation occurring in the soil
profile is thought to be biochemical in nature (e.g., Brady and Weil 2008).

Estimates of Ecosystem Weathering Flux

As mentioned above, the cationic denudation during 1963–1974 in the HBEF eco-
system was estimated at 2.0 × 103 Eq/ha-year, comparable to that calculated for New
England (2.2 × 103 Eq/ha), but substantially less than the average (3.8 × 103 Eq/ha)
for the North American continent (Reynolds and Johnson 1972). This low value is
reasonable given the relatively resistant nature of the silicate-based material at
HBEF to weathering. Estimating weathering flux is more complicated, however,
because we know that both living and dead biomass have been accumulating and
soils have been developing within the forest ecosystems at HBEF for the past
14,000 years (Likens and Davis 1975; M. Davis et al. 1985). In effect, this net accre-
tion of biomass represents a long-term sink for some of the nutrients supplied from
weathering release, and for an ecosystem mass balance, estimates of weathering
must be added to that actually removed from the system as dissolved and particulate
matter export. This disposition of cations within the ecosystem allows some impor-
tant conclusions: (a) Cations stored within the biomass and soil cation exchange
pools must be included in calculations of contemporary weathering; (b) the rate of
storage is a consequence of the ecosystem’s current state of forest succession and,
therefore, it changes with time; and (c) existence of the forest and its state of devel-
opment must be considered in mass-balance estimates of weathering. Thus, cationic
denudation values during the first two decades of the HBES, including biomass
accumulation, were about twice those reported previously (Johnson et al. 1972).
Currently, annual biomass storage is minimal or declining (Chap. 3). Thus, the bio-
logical demand for weathering products is currently greatly reduced.
Weathering flux is not uniform across the watershed. Nezat et al. (2004) exam-
ined the effect of elevation and vegetation on long-term weathering flux in Watershed
1 of the HBEF. Coniferous vegetation is much more abundant in the upper third
of south-facing watersheds, and soils and glacial till are thinner (see Chap. 1).
134 5 Weathering

Using titanium mineral depletion in the soil profile, they found that the weathering
flux of base cations decreased by twofold from this coniferous zone (a decrease in
elevation of 260 m) to lower elevation hardwood dominated zones in this watershed
over the period since glacial retreat.

Net Soil Release

Since direct estimates of Wi, ΔXi, ΔMi, and ΔOi currently are quantitatively unknown
or poorly known at HBEF, we recently proposed the use of “net soil release” rather
than “weathering” or “cationic denudation” in our ecosystem analysis of these
­systems (see Likens et al. 1994, 1998, 2002a), and it is calculated as

Net soil release = Wi ± Mi ± ∆Xi ± ∆Oi = Si + ∆Bi − Pi (2)

[symbols as in Eq. (1)].
Net soil release includes mobilization of an element, e.g., calcium, via chemical
weathering, cation exchange, and mineralization, in response to watershed inputs,
exports, and net biomass uptake (Likens et al. 1994, p. 79).
The long-term comprehensive data collected in the HBES allow us to compen-
sate for the changing effects of precipitation chemistry, biologic activity, hydrologic
dilution, and seasonal variation in characterizing streamwater chemistry. Our expe-
rience suggests that the most important factor in this regard is the growth status of
the forest and how that affects net storage in biomass. This factor may mask the
actual rate at which rocks are chemically decomposing within the system. Changing
precipitation inputs also are critical to these reactions and mass balances, for exam-
ple, weathering flux is likely to increase if climate change were to produce warmer
and wetter conditions at the HBEF (Chap. 2).
An example is provided by the case for calcium weathering. Several Ca-bearing
minerals at the HBEF like plagioclase, biotite, garnet, and hornblende could release
calcium to soil water through chemical weathering (Likens et al. 1998). Plagioclase,
a calcium–sodium silicate mineral, is relatively abundant and thought to be a major
source of weatherable calcium at HBEF, but inputs from bulk precipitation and
estimates of weathering release are less than calcium outputs in stream water plus
biomass storage over the long term (see Chap. 6). Thus, release from soil exchange
sites and organically bound calcium also must be occurring (Likens et al. 1998).
Blum et al. (2002) suggested that another source of calcium for the ecosystem is
from fungal mycorrhizal weathering (“mining” sensu van Breeman et al. 2000) of
apatite inclusions in silicate minerals in deeper soil horizons. Release of calcium
from calcium oxalate crystals may be an unaccounted source of enhanced leaching
of calcium in the HBEF, driven by hydrogen ion release from root uptake processes
(S. Bailey et al. 2003).
S. Bailey et al. (2003) used net sodium output (streamwater loss minus bulk pre-
cipitation inputs) as an indicator of weathering fluxes at HBEF because sodium is
Net Soil Release 135

1.4
Ecosystem Ca:Na Molar Ratio
1.2

1.0

0.8

0.6 0.59

0.4
range of calcium weathering release based on reactants

0.2 0.24

0.0
1950 1960 1970 1980 1990 2000 2010 2020
Water Years

Fig. 52  Net ecosystem Ca:Na ratios (streamwater outputs minus bulk precipitation inputs) for
Watershed 6 of the Hubbard Brook Experimental Forest (modified and extended from S. Bailey
et al. 2003)

negligible as a component of the soil exchange complex (Johnson et al. 1991b), in

soil organic matter and in secondary minerals (Likens et al. 1998), and long-term
storage is minimal in biomass of the HBEF. Thus, weathering release of sodium
may be estimated simply as the difference between streamwater loss and bulk pre-
cipitation inputs (Eq. 1). Assuming that weathering release of calcium and sodium
is stoichiometrically related, weathering release of calcium can be estimated by
multiplying the weathering release of sodium by the Ca:Na ratio of the weathering
reactants. At HBEF these range from 0.24 for available soil plagioclase to 19.0 for
calcic plagioclase in the Rangeley Formation (Likens et al. 1998; S. Bailey et al.
2003), to even higher ratios for trace minerals such as epidote and apatite.
S. Bailey et al. (2003) suggested that the streamwater–bulk precipitation ratio for
Ca:Na in W6 would stabilize between 0.24 and 0.59 based on the release from the
principal weathering reactants. In fact, this ratio now may have stabilized as it has
hovered near 0.50 since 2000 (Fig. 52). S. Bailey et al. (2003) further argued that the
sodic plagioclase Ca:Na ratio of 0.24 gives the lower bound to weathering release of
calcium, whereas the minimal ecosystem Ca:Na ratio they observed (0.59) provided
an upper bound. These values are equivalent to a calcium weathering rate of 54 and
132 mol/ha-year, respectively, thereby constraining this release of calcium to the
ecosystem. Higher ratios prior to ~1975 reflect much larger amounts of calcium
input to the HBEF in bulk precipitation (Likens et al. 1998, p. 97) and depletion of
labile soil exchange pools (S. Bailey et al. 2003). The current ratio also shows the
depletion of calcium in proportion to the availability of weatherable minerals, such
as plagioclase, hornblende, and apatite, and that major ecosystem pools (exchange,
136 5 Weathering

vegetation) may be relatively stable, where weathering release is dominated by

sodic plagioclase, with minor contributions from calcic plagioclase, hornblende,
and apatite.
The Ca:Na ratio also can be used to assess various ecosystem dynamics. For
example, evaluating retention and leaching of calcium within and from the ecosys-
tem suggest that (1) depletion of exchangeable calcium pools, (2) decomposition in
the forest floor, and (3) changes in mineral weathering do not appear to be important
sources of calcium in streamwater outputs following major disturbance, such as
clear-cutting (S. Bailey et al. 2003). As mentioned elsewhere, net release of hydro-
gen ion from various internal ecosystem processes (Fig. 49) may contribute signifi-
cantly to increased calcium leaching.
There are a number of calcium-bearing minerals in the bedrock of the HBEF,
including slowly weathering plagioclase and biotite and intermediate-weathering
minerals like amphibole (hornblende and actinolite), pyroxenes (mostly diopside),
and clinozoisite. These are also present in the glacial till of the valley, which is the
parent material for soil in the watershed-ecosystems (Likens et al. 1998). It is
believed, however, that most calcium comes from weathering of the relatively abun-
dant plagioclase in the HBEF (Likens et al. 1998; Nezat et al. 2004). Finally, cal-
cium is also released from soil exchange sites and mineralization of soil organic
matter.
Weathering generally accounts for ≪50 % of the net soil release value for cal-
cium in the HBEF (Likens et al. 1998). Likens et al. (1998) estimated that total
calcium lost from exchange sites and soil organic matter between 1965 and 1992
was 9.9–11.5 kmol/ha. This amount is greater than the entire exchangeable pool
measured in 1983, is 53–61 % of the living biomass pool in 1992, and is 2.0–2.4
times greater than the forest pool of calcium in 1969–1970 (Likens et al. 1998).
The calcium minerals of the bedrock obviously are the most prone to weathering,
and calcium compromises some 1.4 % abundance in bedrock (Table 27; Johnson
et al. 1968). If this were the only source, then some 1.5 metric tons per hectare of
bedrock must undergo weathering every year to provide the requisite net soil release
of calcium to the system (Table 28). If we were to assume that this calcium-based
value (1.5 tons of rock per hectare year) was the lower limit for the rate of rock
weathering, we could then calculate the fraction of each element extracted by the
weathering process (Table 29). These estimates seem to be compatible with what is
known about the vulnerability of various silicate minerals to chemical attack
(Goldich 1938; Marshall 1964; White and Brantley 1995); i.e., calcium-rich miner-
als are the least stable and potassium-rich minerals are the most stable in the bed-
rock of the HBEF exposed to an acidic chemical regime.
Secondary minerals (especially vermiculite) and exchangeable pools probably
play an important role in regulating streamwater losses of potassium from watershed-
ecosystems of the HBEF (Likens et al. 1994). Biotite and muscovite are primary
potassium-bearing minerals in bedrock, till, and soil with some potassium feldspar;
mixed-layer biotite/vermiculite and illite are the dominant secondary minerals at the
HBEF (Likens et al. 1994). Biotite is probably the most weatherable of this group
(Sverdrup and Warfvinge 1988). In the acidic soils of the HBEF, weathering of
Net Soil Release 137

Table 29  Chemical composition of bedrock and glacial till at the Hubbard Brook Experimental
Forest
Abundance in Amount contained in Mean composition
Element bedrocka, % 1,500 kg of bedrock, kg of glacial tillb
Ca2+ 1.4 21.1 8.7
Na+ 1.6 24.1 14.7
Mg2+ 1.1 16.5 3.9
K+ 2.9 43.6 21.6
Al3+ 8.3 124.8 57.5
Sc 0.2‒0.8 7.5 –
Si4+ 30.7 461.7 394.2
a
Taken from Johnson et al. (1968)
b
Unpublished data from S. Bailey
c
Bailey et al. (2004)

potassium feldspar, muscovite, and biotite to koalinite and ­hydroxyl-­interlayered

vermiculite are ultimate sources of dissolved potassium, but on short timescales,
potassium may be sequestered in interlayers providing a sink (Douglas 1989; Likens
et al. 1994). With declining inputs of mineral acid deposition from the atmosphere
(Chap. 3), is weathering now rebuilding exchange pools in HBEF soils?
A Final Point: When compared with the usual published chemical analyses for
large, high-order streams in New England (Johnson et al. 1972), the headwater
streams of the HBEF appear to be rather unusual in chemical makeup; i.e., they are
dominated by strong acids (sulfuric and nitric). Significantly, the sulfate content of
New England stream water is fixed at ~4–6 mg/L, and it stays essentially at this
concentration during the remainder of its transit to the sea (Johnson et al. 1972;
Pearson and Fisher 1971). Because of its physiographic location, the HBEF (Fig. 2)
represents the incipient stages of drainage water, i.e., precipitation that recently per-
colated through the soil zone and is appearing in variable source areas or first-order
stream channels. These acids, however, are in the process of being neutralized by
chemical weathering reactions. The rate of neutralization varies somewhat as a
function of local streamflow rate and local bedrock chemistry, but our experience at
HBEF suggests that the major chemical composition of stream water is determined
within a few meters of the initiation of soil water flow and at most the first few hun-
dred meters of streamflow (e.g., Hall and Likens 1980; Likens and Buso 2006;
Zimmer et al. 2012).
The relatively low pH typical of headwater streams at HBEF (pH 4–5) precludes
the presence of significant quantities of ionized carbonic acid (bicarbonate) (Stumm
and Morgan 1996). As the water moves downstream and through longer flow paths
from the headwaters, it is progressively neutralized by weathering reactions (e.g.,
Johnson et al. 1981). As the pH rises above 5, carbonic acid can dissociate and car-
bonation weathering reactions may proceed. Clear evidence for this is that bicar-
bonate alkalinity increases downstream as shown by chemical analyses in the
mainstream of Hubbard Brook itself.
Chapter 6
Nutrient Cycles and Mass Balances

Chemical elements without a prominent gaseous phase at normal biologic

temperature, such as calcium, magnesium, and potassium, have what is referred to
as a sedimentary biogeochemical cycle (Odum 1959). That is, flux and cycling are
affected primarily by hydrologic factors (including dissolution, erosion, and sedi-
mentation), landslides, vulcanism, and biologic agents. At some places, for example,
arid regions, calcium, magnesium, sodium, potassium, and phosphorus have
relatively large airborne fluxes of particulates, which complicate quantitative mea-
surement of their fluxes and cycles. The presence of a prominent gaseous compo-
nent in the biogeochemical cycle (e.g., C, N, and S) greatly complicates measurement
and analysis at the ecosystem scale, as discussed previously. To illustrate a sedimen-
tary cycle, the average biogeochemical relationships for calcium and potassium at
HBEF will be summarized for two pentads: one near the beginning of the long-term
record and one nearer the end. The cycle for sulfur will be used to show relation-
ships for an element with a major meteorologic component. “Average” is somewhat
misleading in this regard because of the large temporal changes that have occurred
since 1963, as discussed in previous chapters. Nevertheless, summarizing and syn-
thesizing long-term data into a conceptual diagram for various nutrients provides
new understanding and insights.

The Calcium Cycle

This synthesis for calcium will be done for two 5-year periods, 1964–1969 and
1987–1992 (Fig. 53). In addition, mass-balance diagrams will be presented below
showing annual values for bulk precipitation input, net soil release, streamwater
output, and biomass storage to allow the calculation of a net ecosystem flux (NEF)
annually throughout the period 1963–2009.
Bulk precipitation contributed 1.5 kg/ha (61 mol/ha) of calcium to the ecosystem
each year during the initial pentad (1964–1969), but less than half that amount,

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2_6, 139

© Springer Science+Business Media New York 2013
 a CALCIUM
1964–69
Northern Hardwood Forest Ecosystem
Wet deposition
48 Biological
Dry deposition vectors
18
Above ground
living biomass Standing
9,590 dead biomass 61
Resorption INPUT
Litter
plus dry
Throughfall
and Wet
135 0
Stemflow
1,015 Deposition
220 boles
Below ground Dead biomass
living biomass
67 Forest floor BIOSPHERE
Net uptake 2,470

Available Pool Total pool 9,230 Ga )

root litter seou na
12,700 440 s lo s s (
St re p o r t
FOREST FLOOR inorganic root 90 organic organic a m ex
Exchange fraction exudates fraction
and
inorganic OUTPUT
390 fraction
E + Bh Net soil release (particulate)
Exchange 5 .7
Mineral soil
Bs Total pool < 2mm Units
Exchange 250,0002 Pools: mol/ha
Fluxes: mol/ha-yr
Dissolved Fraction 249

b CALCIUM
1987–92
Northern Hardwood Forest Ecosystem
Wet deposition
23 Biological
Dry deposition vectors
6
Above ground
living biomass Standing
14,970 dead biomass
815 29
Resorption INPUT
Litter
plus dry
Throughfall
and Wet
43 0
Stemflow
1,015 Deposition
110 boles
Below ground Dead biomass
11 living biomass 122 265
Net uptake 3,480 Forest floor BIOSPHERE
1
Available Total pool 7,930 Ga )
root litter seou na
Pool 440 s lo s s (
St re p o r t
FOREST FLOOR a m ex
1,850 organic
inorganic root 90? organic and
OUTPUT
376 inorganic
fraction exudates fraction fraction
E + Bh (particulate)
5 .7?
1,460 230
279 Net soil release
Units
Bs 175 Mineral soil
3,390 Pools: mol/ha
Total pool < 2mm Fluxes: mol/ha-yr
165 250,0002
Dissolved Fraction 205

Fig. 53 Ecosystem pools (boxes) and fluxes (arrows) for calcium for Watershed 6 at the Hubbard
Brook Experimental Forest. Average values in mol/ha or mol/ha-year for periods specified. Values
for 1964–1969 (a) are modified from Likens et al. (1977); and values for 1987–1992 (b): above-
plus belowground biomass is average of 1987 and 1992 values, belowground assumed to be 20 %
of aboveground; net uptake values are based on the difference in biomass storage between 1982
and 1992; (1) root and aboveground litter, root exudates, and mineral soil assumed to be same as
1964–1969; total forest floor pool from Johnson (1989); (2) total mineral soil pool from Federer
et al. (1989) (from Likens et al. 1998). Reprinted with permission of Kluwer Publishers
The Potassium Cycle 141

0.7 kg/ha (29 mol/ha), during the later one (1987–1992). Estimated dry deposition
contributed 0.3 and 0.1 kg Ca/ha-year, respectively, during these two pentads. In
addition, some 9.7 and 5.7 kg of Ca/ha-year, respectively, were generated within the
ecosystem by net soil release. Initially (1964–1969), the 55-year-old forest was
strongly accumulating biomass, so 3.4 kg of Ca/ha-year was stored aboveground
and 1.7 kg/ha-year belowground in the annual vegetation growth increment. Since
1982, however, above- and belowground biomass accumulation has been small with
only 1.3 kg Ca/ha-year stored in total (Fig. 53b). Approximately 25.3 kg Ca/ha-year
is returned to the soil via aboveground litter, 11.0 kg/ha-year in belowground litter,
and 6.7 kg/ha-year in fine root turnover; throughfall and stemflow decreased by half
from the earlier to later pentad. The remainder is lost from the system in drainage
water and decreased from 6.2 to 5.1 kg Ca/ha-year from 1964–1969 to 1987–1992.
Most (97 %) of this loss is in the dissolved form. The more recent (1987–1992;
Fig. 53b) annual net loss from the ecosystem (5.2 kg/ha) in streamwater is only 3 %
of the exchangeable pool of calcium (~161 kg/ha) in the soil and about 3.8 times
greater than the amount currently stored annually by the biomass (1.3 kg/ha). The
exchangeable pool of calcium has been depleted markedly during the past 50 year
or so, primarily through mobilization by acid rain (Likens et al. 1996, 1998; Likens
and Buso 2012). Annual gross uptake of calcium by trees is about 26–30 % of the
exchangeable pool, and some 7–8 times more calcium is cycled through the vegeta-
tion than is lost in streamwater annually (Likens et al. 1998).
Most of the calcium in the ecosystem, or more than 93 %, is present in the soil
complex, whereas only about 7 % (460 kg/ha) is bound in the vegetation.
Nevertheless, currently <1 % of the net biomass storage is provided by the available
nutrient pool each year. Meteorologic input supplies an amount of calcium equiva-
lent to about 50 % of all calcium taken up (net) by vegetation annually, whereas net
soil release generates about four times more than the annual net vegetation uptake.
Annually, calcium stored by vegetation is returned to the forest floor by litter, stem-
flow, throughfall, and root exudates. More recent (1987–1992) inputs to the avail-
able nutrient pool come from net soil release, 50 %; leaching from the canopy,
24 %; root exudates, 20 %; and atmospheric deposition, 6 % (Fig. 53b; Likens et al.
1998). Interestingly, root exudates from living vegetation release ~20 % as much
calcium as is lost as root litter. Gosz et al. (1973) calculated that some 17.3 kg Ca/
ha was released by decomposition and leaching of current litter each year at HBEF.
Comparing calcium dynamics from 1964–1969 to 1987–1992, net annual bio-
mass storage decreased by ~75 %, atmospheric deposition by >50 %, throughfall
and stemflow by ~50 %, stream output of dissolved calcium by ~20 %, and net soil
release by ~40 % (Fig. 53).

The Potassium Cycle

Potassium is another element with a sedimentary cycle. Potassium is an important

nutrient in forest ecosystems (Tripler et al. 2006), and although cycled relatively
rapidly in the forest floor, it is strongly retained within the ecosystem against losses
142 6 Nutrient Cycles and Mass Balances

in streamwater (Likens et al. 1994). Comparing an ecosystem analysis in 1964–

1969 (Fig. 54a) with one done more recently in 1987–1992 (Fig. 54b) showed the
same loss in streamwater and similar inputs in bulk precipitation for these different
periods. There were decreases in biomass storage, throughfall flux, and net soil
release with time, correlating with decreases in biomass accretion during this period.
Potassium in precipitation inputs, in streamwater outputs, and in the exchange
complex of the mineral soil is relatively small compared to rock and biomass pools
(Fig. 54). Annual net hydrologic flux (NHF) was negative (streamwater output > pre-
cipitation input) for all years except 1963–1965, drought years at the HBEF (Likens
et al. 1994). Following these drought years, annual NHF values varied, but tended
to fluctuate between ~ −0.4 and −1.4 kg/ha-year.
Because of changes in the long-term chemistry of bulk precipitation and particu-
larly because of the decline in biomass accumulation at HBEF, significant changes
in the flux and cycling of potassium occurred between the early (1964–1969) period
of study and the current (1987–1992) period (Likens et al. 1994). For example, stor-
age of potassium in biomass, net soil release, and throughfall all decreased, whereas
resorption increased from the earlier pentad to the later one (Fig. 54).
An elevational gradient of biotic and abiotic conditions in W6 provided an
opportunity to assess effects on the biogeochemistry of the watershed-ecosystem.
Different patterns in pools and fluxes of potassium were evident along this gradient.
There were higher concentrations and fluxes of potassium in vegetation, aboveg-
round litter, throughfall, forest floor, and soil water in the lower, more deciduous
zones than in the higher, more spruce–fir dominated zone. Abiotic patterns included
elevational decreases in soil cation exchange capacity (as a function of reduced soil
organic matter) and soil/till depth at higher elevations (Likens et al. 1994). These
abiotic patterns tended to increase the relative impact from strong acid anions like
sulfate and nitrate in atmospheric deposition.

The Sulfur Cycle

Elements like carbon, sulfur, and nitrogen have a prominent gaseous phase at nor-
mal biologic temperatures, and this greatly complicates quantitative measurements
of the biogeochemical flux and cycling because the flux of gases and airborne par-
ticles are difficult to measure quantitatively, particularly at the watershed scale. To
provide an example of this type of biogeochemical cycle, sulfur will be used
(Fig. 55).
Sulfur plays a key role in the biogeochemical and ecological function of for-
ested watershed-ecosystems at HBEF. Even though it has been declining since
1964, sulfate is the dominant anion in both precipitation and streamwater
(see Table 5, Chap. 3). Sulfate represented 71 % and 53 % of anions in bulk precipi-
tation and 80 % and 87 % of anions in streamwater in 1964–1965 and 1997–1998,
respectively. More than 95 % of the sulfur in the ecosystem at any one moment is
found in the soil complex, and the remaining 5 % occurs in living biomass (Fig. 55).
The Sulfur Cycle 143

Fig. 54 Ecosystem pools (boxes) and fluxes (arrows) for potassium for Watershed 6 at the
Hubbard Brook Experimental Forest. Average values in mol/ha or mol/ha-year for periods speci-
fied. Values for 1964–1969 (a) and for 1987–1992 (b). From Likens et al. (1994). Reprinted with
permission of Kluwer Academic Publishers
 a SULFUR
1964–69
Northern Hardwood Forest Ecosystem
Bulk deposition 406
Biological
Dry deposition vectors
85
Above ground
living biomass Standing
1,300 dead biomass
nd 491
Resorption INPUT
Litter
plus dry
Throughfall Wet
and
25 40
Stemflow
180 Deposition
514 boles
Below ground Dead biomass
12 living biomass nd
Net uptake 530 Forest floor BIOSPHERE

Available root litter Total pool 4,696 Ga )

Pool 19 s e o u s l o s s (?
St re t
FOREST FLOOR inorganic root 90 organic organic 498 a m expo r
Exchange fraction exudates fraction and
59 59 <3 inorganic
fraction OUTPUT
E + Bh (particulate)
Net soil release <3
Exchange
44 Mineral soil
Bs weathering Total pool Units
Exchange 50 53033 Pools: mol/ha
Fluxes: mol/ha-yr
Dissolved Fraction 495

b SULFUR
1993–98
Northern Hardwood Forest Ecosystem
Bulk deposition 266
Biological
Dry deposition vectors
57
Above ground
living biomass Standing
2100 dead biomass
100 323
Translocation INPUT
Litter
plus dry
Throughfall Wet
and
2.3 40
Stemflow
180 Deposition
346 boles
Below ground Dead biomass
living biomass 30
0.8
Net uptake 790 Forest Floor BIOSPHERE
Oie 1370
Oa 3326
Available Pool root litter Ga )
Total Pool 4696 s e o u s l o s s (?
FOREST FLOOR 20 St re t
inorganic root organic organic 448 a m expo r
482 fraction exudates fraction and
59 60? <3 inorganic
E + BH fraction OUTPUT
526 (particulate)
Net soil release <3
B S1 128
Mineral soil
507 weathering Total pool Units
B S2 50 53033 Pools: mol/ha
Fluxes: mol/ha-yr
Dissolved Fraction 445

Fig. 55 Ecosystem pools (boxes) and fluxes (arrows) of Sulfur for Watershed 6 at the Hubbard
Brook Experimental Forest. Average values in mol/ha or mol/ha-year for periods specified. Values
for 1964–1969 (a) and for 1993–1998 (b). Above-plus belowground biomass is for 1997; root and
aboveground litter and root exudates assumed to be same as in 1964–1969; net uptake values based
on difference in biomass storage between 1982 and 1998; Forest floor pools and total mineral soil
pool for both periods from Zhang et al. (1999) as estimated for W5 in 1983. From Likens et al.
(2002a) and reprinted with permission of Kluwer Academic Publishers
The Sulfur Cycle 145

In a recent pentad (1993–1998), vegetation stores sulfur at a rate of 0.1 kg/ha-year.

Net soil release is 4.1 kg S/ha-year, and mineral weathering generates an amount of
sulfur equivalent to about 16 times the sulfur stored annually by vegetation.
Meteorologic input is approximately 100 times the annual vegetation storage
(Fig. 55). The sulfur budget of the forest ecosystems at HBEF is therefore domi-
nated by meteorologic inputs (i.e., wet and dry deposition).
In 1993–1998, gross uptake of S in trees was estimated at ~9.2 kg S/ha-year,
based on estimates of root exudates, root litter, canopy litterfall, canopy leaching,
and net biomass storage (Fig. 55; Likens et al. 2002a). Only ~1 % of gross uptake
of S is accrued, whereas 99 % is returned to the soil. Gross uptake during 1964–
1969 was ~11 % more than in 1993–1998, and net biomass storage of sulfur was
~12 % of gross uptake (Fig. 55; Likens et al. 2002a). Loss of sulfate via streamwater
decreased significantly from the earlier to the later pentad, but net biomass storage
decreased even more. Because annual net biomass storage of sulfur decreased by
more than an order of magnitude from the earlier to the later pentad, streamwater
loss of S was >13 times than net biomass storage in 1964–1969 and 145 times
greater in 1993–1998 (Fig. 55; Likens et al. 2002a).
Recent (1993–1998) inputs to the available nutrient compartment are partitioned
as follows: atmospheric bulk precipitation (50 %), net soil release (24 %), dry depo-
sition (11 %), root exudates (11 %), and canopy leaching (4 %) (Fig. 55).
It is well known that plants can utilize SO2 directly from the atmosphere (Hill
1971; Hoeft et al. 1972; Cowling et al. 1973), and various authors have suggested
input from aerosols and dust impacted on vegetation surfaces (Eriksson 1952;
Tamm and Troedsson 1955; Duvigneaud and Denaeyer-DeSmet 1964; Carlisle
et al. 1967; White and Turner 1970). Our throughfall and stemflow data (Table 9)
suggest that sulfurous aerosols may be impacted on vegetation surfaces in large
amounts during the summer. We do not know the actual proportions of these two
meteorologic sources of sulfur at HBEF, but we estimate that the net gaseous input
is about fivefold greater than the aerosol deposition on an annual basis (Lovett et al.
1997). Currently (1993–1998), the combination of net gaseous uptake and aerosol
deposition accounts for about 18 % of the total meteorologic input or 13 % of the
annual hydrologic export (Fig. 55). Obviously such inputs must be carefully
assessed in biogeochemical studies.
A summation of the values for biomass accretion, weathering release, precipita-
tion input, and streamwater output shows a sulfur imbalance, suggesting some addi-
tional source(s) of sulfur for the ecosystem. Because net hydrologic fluxes
(precipitation inputs‒streamwater outputs) for S are consistently imbalanced at the
HBEF, with streamwater outputs exceeding bulk precipitation inputs (Chap. 4;
Likens et al. 2002a), dry deposition of sulfur needs to be added to this balance. One
exception occurred during the drought year of 1964–1965, when precipitation
inputs exceeded streamwater outputs.
Dry deposition of S is difficult to measure quantitatively in forest landscapes in
mountainous terrain. During 2002, dry deposition was estimated to be 0.5–2.5 kg S/
ha-year using a new approach (Mitchell et al. 2011), as compared to an estimate of
1.8 kg/ha-year during 1993–1998 (Likens et al. 2002a). Adding dry deposition
reduced the overall imbalance. Both unmeasured desorption and mineralization of
146 6 Nutrient Cycles and Mass Balances

sulfur stored in the soil also probably contributed to this imbalance. In the long
term, annual balances were both positive and negative (Likens et al. 2002a).

Nutrient Cycle Relationships at the HBEF

The allocation of budgetary items for the major nutrients at HBEF circa ~1974 is
given in Tables 30 and 31. This was a time when forest biomass was strongly
aggrading and meteorologic inputs of acid, sulfate, and nitrate were high. These
values, along with those of Tables 17, 18, and 19, suggest several important conclu-
sions regarding the biogeochemistry of the northern hardwood ecosystem. These
are the following: (a) In relation to gross losses, nutrient inputs in bulk precipitation
represent a significant addition to the ecosystem. Such inputs can be the major
source of nutrients in terrestrial ecosystems low in weathering substrates (e.g., Art
et al. 1974); (b) for some elements, such as nitrogen, sulfur, and chloride, dry
deposition or biologic gaseous fixation can provide significant input to forest

Table 30 Standing stocks and annual biogeochemical fluxes for a 55-year-old forested ecosystem
at the Hubbard Brook Experimental Forest (circa 1974)
Chemical element
Component Ca Mg Na K N S P Cl
Standing stock (kg/ha)
a
Aboveground biomass 383 36 1.6 155 351 42 34
a
Belowground biomass 101 13 3.8 63 181 17 53
a
Forest floor 372 38 3.6 66 1256 124 78
Annual flux (kg/ha-year)
Bulk precipitation input 2.2 0.6 1.6 0.9 6.5 12.7 0.04 6.2
a a a a a
Gaseous or aerosol input 6.1 ?
Weathering releaseb 21.1 3.5 5.8 7.1 0 0.8 ? a

Streamwater output
Dissolved substances 13.7 3.1 7.2 1.9 3.9 17.6 0.01 4.6
a
Particulate matter 0.2 0.2 0.2 0.5 0.1 <0.1 0.01
Vegetation uptake 62.2 9.3 34.8 64.3 79.6c 24.5c 8.9 a

a
Litter fall 40.7 5.9 0.1 18.3 54.2 5.8 4.0
a
Root litter 3.2 0.5 0.01 2.1 6.2 0.6 1.7
Throughfall and stemflow 6.7 2.0 0.3 30.1 9.3 21.0 0.7 4.4
Root exudates 3.5 0.2 34.2 8.0 0.9 1.9 0.2 1.8
Net mineralization 42.4 6.1 0.1 20.1 69.6 5.7 ? ?
a
Aboveground biomass accretion 5.4 0.4 0.03 4.3 4.8 0.8 0.9
a
Belowground biomass accretion 2.7 0.3 0.12 1.5 4.2 0.4 1.4
a a
Forest floor accretion 1.4 0.2 0.02 0.3 0.8 0.5
a
Small, unmeasured
b
Currently calculated as net soil release; see Chap. 5
c
Root uptake
Annual Watershed-Ecosystem Mass Balances (“Budgets”) 147

Table 31 Allocation of budgetary items in percent for watershed-ecosystems of the Hubbard

Brook Experimental Forest (circa 1974)
Ca K Mg Na N S
Source
Bulk precipitation input 9 11 15 22 31 65
a
Net gas or aerosol input — — — — 31
Weathering releaseb 91 89 85 78 — 4
Storage or loss
Biomass accumulation
Vegetation 35 68 17 2 43 6
Forest floor 6 4 5 <1 37 4
Streamflow
Dissolved substances 59 22 74 95 19 90
Particulate matter <1 6 5 3 <1 <1
a
See Table 30
b
Currently calculated as net soil release; see Chap. 5

ecosystems; (c) although variable and changing with time, hydrologic-driven

budgets for the forest watershed-ecosystems currently show gains in carbon, hydro-
gen ion, nitrogen, phosphorus, and chloride and losses in sulfur, dissolved silica,
calcium, sodium, aluminum, magnesium, and potassium (Likens 2004; Chap. 4).
Losses of the latter substances from the intrasystem nutrient cycle are made up by
weathering of primary minerals/net soil release, dry deposition, and mineralization
of soil organic matter (Fig. 1); (d) net soil release is the major source of calcium,
potassium, magnesium, and sodium in the ecosystem; (e) a dramatic decrease in net
biomass uptake since 1982 has greatly decreased the NEF for calcium and magne-
sium, and a decrease in net biomass uptake and streamwater output increased the
NEF for nitrogen. The forest floor is a particularly effective storage site for nitrogen.
In contrast, <10 % of the sodium and sulfur added to the ecosystem from meteoro-
logic sources is stored in the biomass, and the remainder, more than 90 %, is lost in
streamflow annually; (f) root exudates are very important to the intrasystem cycle of
sodium, whereas stemflow and throughfall are particularly important for cycling
sulfur and potassium, and litter plays a major role in the cycling of phosphorus,
nitrogen, magnesium, and calcium.

Annual Watershed-Ecosystem Mass Balances (“Budgets”)

Another informative way to summarize and synthesize long-term biogeochemical

dynamics is to measure or estimate the full range of ecosystem fluxes and organize
this information into annual mass balances (“budgets”) in order to calculate Net
Ecosystem Flux (NEF). This is done here for calcium, sodium, potassium, sulfur,
dissolved inorganic nitrogen, and chlorine (Figs. 56, 57, 58, 59, 60, and 63), which
provides a sampling of elements with and without a prominent gaseous phase.
148 6 Nutrient Cycles and Mass Balances

300

Source / Gain
Calcium Mass Balance
200

100

0
Flux (mol/ha-yr)

Sink / Loss
-100

-200

-300

-400 Net Soil Release

Bulk Precipitation Inputs
Total Stream Outputs
-500 Net Biomass Uptake
Net Ecosystem Flux
-600

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Years

Fig. 56 Annual ecosystem mass balances for calcium in Watershed 6 of the Hubbard Brook
Experimental Forest during 1963–2009. Bulk precipitation input, including dry deposition (open
circle); filled diamond is net soil release based on midrange reaction rate (of 0.37 × NHF for
sodium; see Chap. 5); filled circle is streamwater output; filled triangle is net biomass storage; and
is net ecosystem flux obtained by difference (straight line) Net Ecosystem Flux is obtained by dif-
ference [modified and extended from Likens (1996, 1998)]

Displaying these long-term budgetary data in this way helps to put the detailed,
biogeochemical data for the watershed-ecosystem for selected pentads into perspec-
tive and provides new insights about the biogeochemical dynamics of these
elements. The inputs to the mass balance shown are bulk precipitation, dry deposi-
tion, and net soil release. The outputs are streamwater exports, net biomass storage,
and net gas flux. The difference, annually, between inputs and outputs is net ecosys-
tem flux, NEF (or NHF when only the hydrologic components are known).
These long-term data on inputs and outputs allow mass balances to be con-
structed and compared over time. The values in these mass balances are updated and
modified from those published previously (e.g., Likens et al. 1994). The uncertainty
bars on the NEF term in these mass-balance diagrams are root mean squares for
estimated errors for the other components of the mass balance: bulk precipita-
tion = measured with ±5 % uncertainty; dry deposition = included in bulk precipita-
tion for calcium, sodium, and potassium; sulfur, nitrogen, and chlorine estimated as
a % of bulk precipitation with 50 % estimated uncertainty (the uncertainty in S dry
deposition, Mitchell et al. 2011); net soil release = estimated as ±50 % uncertainty
(based on S. Bailey et al. 2003); streamwater output = measured with ±5 % uncer-
tainty; net biomass uptake = for above-plus belowground; the range was 27 to
150 %, uncertainty from high to low uptake (based on Yanai et al. 2012);
Annual Watershed-Ecosystem Mass Balances (“Budgets”) 149

700
Net Soil Release
600 Sodium Mass Balance Bulk Precipitation Inputs
Total Stream Outputs
500 Net Biomass Uptake
Net Ecosystem Flux
400

Source / Gain
300
Flux (mol/ha-yr)

200
100
0

Sink / Loss
-100
-200
-300
-400
-500
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Years

Fig. 57 Annual ecosystem mass balances for sodium in Watershed 6 of the Hubbard Brook
Experimental Forest during 1963–2009. Open circle is bulk deposition input, including dry deposi-
tion; filled diamond is net soil release (net soil release approximates NHF for sodium); filled circle
is streamwater output; filled triangle is net biomass storage; and (straight line) is Net Ecosystem
Flux obtained by difference

see element-by-element details in figure legends; mass balances—further explana-

tions [see page 160–161].
Annual budgets of this sort over almost five decades are a unique contribution
from the long-term biogeochemical studies at the HBEF. Long-term budgetary pat-
terns have changed, are often individualistic for each element in adjacent watershed-
ecosystems, but provide key insights regarding the important drivers of change,
such as hydrology, biomass aggradation, air pollution, and climate change. Examples
for six elements are given below.

Calcium

The long-term annual mass budgets for calcium in W6 are shown in Fig. 56. This
figure is extended some 15 years beyond that published in Likens et al. (1998); this
recent period occurs when biomass uptake and storage in W6 are minimal (Likens
et al. 1994; Lindenmayer and Likens 2010; van Doorn et al. 2011).
In this current analysis (Fig. 56), we adopted the conservative assumption of
Likens et al. (1998) that bulk precipitation collectors sample both wet and dry
150 6 Nutrient Cycles and Mass Balances

300
Potassium Mass Balance

Source / Gain
200

100
Flux (mol/ha-yr)

Sink / Loss
-100
Net Soil Release
Bulk Precipitation Inputs
Total Stream Outputs
-200 Net Biomass Uptake
Net Ecosystem Flux

-300
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Years

Fig. 58 Annual ecosystem mass balances for potassium in Watershed 6 of the Hubbard Brook
Experimental Forest during 1963–2009. Open circle is bulk deposition input including dry deposi-
tion; filled diamond is net soil release; (filled circle) is streamwater output; (filled triangle) is net
biomass storage; and straight line is Net Ecosystem Flux obtained by difference [modified and
extended from Likens et al. (1994)]

deposition of calcium. With annual streamwater output of calcium decreasing from

~1973 to ~1985 and then relatively constant, bulk precipitation inputs decreasing
since the mid-1950s (Fig. 19; Likens and Bailey 2013), net biomass storage mini-
mal since ~1982 and net soil release (based on sodium biogeochemistry, S. Bailey
et al. 2003) similar to bulk precipitation inputs and rather constant, net ecosystem
flux (NEF) of calcium is much smaller after ~1982 (Fig. 56). Streamwater outputs
have dominated the fluxes of calcium throughout the study. Nevertheless, the strik-
ing result shown by this long-term record is the marked loss of calcium (−NEF)
from the watershed-ecosystem, particularly prior to ~1985. It was calculated earlier
that these watershed-ecosystems lost ~840 kg Ca/ha or approximately 50 % of the
available nutrient pool during the period 1940–1995 (Likens et al. 1996, 1998; but
see p. 63). These net losses have been attributed to the impact of acid rain during
this period (Likens et al. 1996, 1998; Likens and Bailey 2013). The NEF during
1963–2009 was −8.3 kmol Ca (207 kg Ca) (Fig. 56; Table 32). Total streamwater
losses of calcium were larger than NEF during this 47-year period (Table 32).
The acid-neutralizing capacity in streamwater became positive in W6 for the first
time in our study ~2000 (Fig. 17), but acid rain continues, albeit at a smaller amount
(Fig. 14), and the NEF continues to be negative, but not significantly different than zero
(Fig. 56). An increase in acidic input in bulk precipitation and/or net biomass uptake
(forest recovery) would increase NEF losses.
Annual Watershed-Ecosystem Mass Balances (“Budgets”) 151

1000
Sulfur Mass Balance Dry Deposition

Source / Gain
800 Net Soil Release
Bulk Precipitation Inputs
Total Stream Outputs
600
Net Biomass Uptake
Net Ecosystem Flux
400
Flux (mol/ha-yr)

200

Sink / Loss
-200

-400

-600

-800

-1000
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Years

Fig. 59 Annual ecosystem mass balances for sulfur for Watershed 6 of the Hubbard Brook
Experimental Forest during 1964–2009. Open circle is bulk deposition input; filled diamond is net
soil release; filled circle is streamwater output; filled triangle is net biomass storage; and inverted
filled triangle is dry deposition based on Mitchell et al. (2011), and straight line is net ecosystem
flux obtained by difference [modified and extended from Likens et al. (2002a)]

Sodium

As is typical of an element with a sedimentary cycle, streamwater outputs greatly

exceeded bulk precipitation inputs of sodium (Fig. 57) throughout the long-term
study (1963–2009). In these mass balances, dry deposition inputs were included as
a part of bulk precipitation inputs, and net soil release was calculated as a function
of annual streamflow and a variable annual NHF for sodium. Net soil release and
streamwater exports dominate the annual, long-term mass-balance values for
sodium. The uncertainty for net soil release of sodium is estimated at ±10 %.
Because net soil release plus bulk precipitation input approximately equal stream-
water output on an annual basis, net storage of sodium in biomass is very small even
before biomass stopped accumulating in ~1982. NEF is approximately in balance
(−0.6 kmol/ha) in the long-term record, particularly given the uncertainties (Fig. 57).
Sodium concentrations are slightly decreasing in both bulk precipitation and stream-
water, so the recent increasing amount of precipitation and streamflow are the pri-
mary drivers in these fluxes (Figs. 19a and 57). Streamwater output and net soil
release/weathering are the dominant fluxes in these long-term mass balances for
sodium at the HBEF.
152 6 Nutrient Cycles and Mass Balances

200
Nitrogen Mass Balance

Source / Gain
150

100
Flux (mol/ha-yr)

50

Sink / Loss
-50

Dry Deposition
-100 Bulk Precipitation Inputs
Total Stream Outputs
Net Biomass Uptake
-150 Net Gaseous Flux
Net Ecosystem Flux

-200
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Years

Fig. 60 Annual ecosystem mass balance for dissolved inorganic nitrogen in Watershed 6 of the
Hubbard Brook Experimental Forest during 1963–2009. Open circle is bulk deposition input; filled
circle is streamwater output; filled triangle is net biomass storage; inverted filled triangle is dry
deposition; red square is gaseous losses based on N2O flux estimates; and straight line is Net
Ecosystem Flux obtained by difference

Potassium

This record of annual mass balance is extended some 17 years beyond that pub-
lished by Likens et al. (1994). As was the situation for calcium and sodium, stream-
water outputs greatly exceeded bulk precipitation inputs of potassium (Fig. 58)
throughout the long-term study (1963–2009). Also, as with calcium and sodium,
dry deposition inputs were included as part of bulk precipitation (Likens et al.
1994). Annual net soil release was calculated as a function of net biomass uptake
and a fluctuating NHF, which in turn is a function of annual streamflow (Fig. 54).
The net biomass uptake had an estimated uncertainty of ±30 to 120 %, the NHF
7 %, giving an estimated NEF uncertainty of ~30 %. The uncertainty in NEF was
large during 1963–1980 and much smaller thereafter because of the smaller effect
of reduced net biomass storage (Fig. 58).
Mirroring the annual net storage of potassium in biomass, net soil release was
large during 1963 to ~1982 (especially 1963 to 1977) and then fluctuated somewhat
above zero thereafter. The NEF value fluctuated around zero throughout the study
with periods of positive values, suggesting possible storage of potassium in the soils
of the HBEF. The aggressive demand for potassium by the biomass is obvious in
these mass balances (Fig. 58). Over the long term essentially balanced total NEF
was slightly negative, ‒0.4 kmol/ha (Fig. 58; Table 32). Nevertheless, long-term
Annual Watershed-Ecosystem Mass Balances (“Budgets”) 153

total net soil release of potassium is ~37 % that for sodium, but the NEF values are
not different (Table 32; Figs. 57 and 58). The weathering sources of potassium and
sodium are different in the HBEF (see Chap. 5).

Sulfur

In response to long-term values for atmospheric emissions of SO2 (Chap. 3; Likens

et al. 2001, 2002a), sulfate input in bulk precipitation peaked at HBEF in 1973.
Sulfate input has been declining steadily since (Table 17). Dry deposition of sulfur
is assumed to be ~21 % of bulk precipitation inputs (Mitchell et al. 2011), and net
soil release is prorated and extended based on 44 mol S/ha-year in 1964–1969 and
128 mol S/ha-year in 1993–1998 (Fig. 55). Interestingly, net soil release is equal to
or larger than the bulk precipitation inputs in 2008 and 2009, suggesting that desorp-
tion of stored sulfur in the ecosystem from past atmospheric inputs may be becom-
ing increasingly important in the mass balance of sulfur. As described above, the
NEF values are small and varied between +86.8 and −229.5 mol S/year (+0.3 and
−0.8 kg S/ha-year) and averaged −54.3 ± 72.9 (SD) mol S/year (−0.19 ± 0.25 kg S/
ha-year) during 1964–1998. The NEF is roughly balanced over the 48 years because
of the tendency for declining bulk precipitation inputs to be balanced by declining
streamwater outputs and increasing net soil release (Fig. 59). The long-term cumu-
lative NEF is small, but negative, −0.3 kmol/ha (Table 32).
Likens et al. (2002a) suggested four hypotheses to account for the small and
inconsistent imbalance in NEF: underestimates of (1) dry deposition, (2) sulfate
desorption from soils, (3) weathering release, and (4) net mineralization of organic
sulfur. In evaluating these hypotheses, they concluded that net desorption of sulfate
and primarily net mineralization of soil organic sulfur were the most likely mecha-
nisms to cause any imbalance (more likely since 1985; Fig. 59) in the NEF.

Nitrogen

Nitrogen is a very interesting, yet difficult, element to characterize quantitatively by

mass balance, particularly over a long period and at the watershed scale. It is an
extremely interesting element, primarily because it is fundamentally important to
biota and has been thought to be limiting (sensu Liebig’s Law of the minimum) to
productivity in the northern hardwood forest ecosystem. Since there are negligible
nitrogen-bearing rocks in the HBEF, all of the nitrogen in the watershed-ecosystems
of the HBEF must have been obtained from the atmosphere and accumulated by the
biota in the past 14,000 years following the glacial retreat from this area. Gaseous
fluxes into the ecosystem via fixation and out of the ecosystem via denitrification
are very difficult to measure quantitatively at the watershed scale (e.g., Groffman
and Rosi-Marshall 2013).
 154

Table 32 Cumulative Totals of all Fluxes in Mass Balance for Watershed 6 of the Hubbard Brook Experimental Foresta
Bulk precipitation Dry deposition Net soil release Stream export Net biomass Gaseous losses NEF
(kmol/ha) (kmol/ha) (kmol/ha) (kmol/ha) intake (kmol/ha) (kmol/ha) (kmol/ha) Flux
Element In In In Out Out Out Net In/Out
Calcium 1.581 In BP 3.991 –10.586 –3.268 na –8.282 Loss
%Total in or out 28.4 71.6 76.4 23.6
Potassium 0.841 In BP 3.988 –2.853 –2.395 na –0.419 Loss
%Total in or out 17.4 82.6 54.4 45.6
Sodium 2.784 In BP 10.804 –14.085 –0.105 na –0.602 Loss
%Total in or out 20.5 79.5 99.3 0.7
Sulfur 14.149 2.971 4.946 –21.963 –0.437 na –0.334 Loss
%Total in or out 64.1 13.5 22.4 98.0 2.0
6

DIN 5.087 1.068 0.000 –1.570 –2.204 –0.534 1.847 Gain

%Total in or out 82.6 17.4 0.0 36.4 51.2 12.4
Chlorine 4.411 0.882 0.129 –5.431 –0.486 na –0.495 Loss
%Total in or out 81.3 16.3 2.4 91.8 8.2
a
Calcium, potassium and sodium for 1963–2009. Sulfur, DIN and chloride for 1964–2009. For calcium, potassium and sodium dry deposition included in bulk
precipitation total. Stream export includes dissolved and particulate substances. There are no gaseous losses for calcium, potassium, sodium, sulfur and chlo-
rine. DIN is dissolved inorganic nitrogen. NEF is Net Ecosystem Flux; values may differ slightly due to rounding errors
Nutrient Cycles and Mass Balances
Nitrogen Retention in Forested Watershed-Ecosystems 155

Here (Fig. 60), only hydrologic fluxes of dissolved inorganic nitrogen (DIN),
that is, nitrate and ammonium, will be considered as dissolved organic nitrogen
(DON) in bulk precipitation averaged 1.1 ± 0.6 kg N/ha-year and DON in stream
water averaged 1.0 ± 0.4 N/ha-year for 1995–2012, representing small and essen-
tially balanced fluxes.
In Fig. 60, dry deposition is assumed to be 21 % of bulk precipitation (Lovett
et al. 1997), which agrees reasonably well with the CASTNET value for our site.
Gaseous losses (0.27 – 1.4 kg N/ha-year) are based on N2O estimates (Kulkarni
et al. 2008; Groffman et al. 2009), and N-fixation inputs (~2 kg N/ha-year) are
based on Roskoski (1980).
The long-term mass balance (NEF) for nitrogen essentially shows a sinusoidal
pattern of ecosystem release during 1964–1980 and accumulation during 1980–
2009 (Fig. 60). Overall, the NEF balance for DIN is positive (1.8 kmol N; Table 32).
This NEF pattern is driven largely by net biomass uptake, and secondarily by bulk
precipitation input, which has been declining since ~1995 and currently is lower
than in 1964 (Fig. 60). This fascinating pattern begs the question of where this nitro-
gen is being stored in the ecosystem and whether this pattern has occurred in the
past and whether it will occur in the future.

Nitrogen Retention in Forested Watershed-Ecosystems

Normally, watershed-ecosystems of the HBEF are very retentive (bulk precipitation

input minus stream water output) against loss of dissolved inorganic nitrogen (DIN)
in stream water (Fig. 61). This “action” against loss might be expected since nitro-
gen is thought to be a limiting element for biological productivity and its ultimate
source in the atmosphere. Even vernal “dams” have been proposed to function as
retention mechanisms at the HBEF (Muller and Bormann 1976).
The interannual pattern of retention is quite variable, however (Fig. 61), with
minimal % retention in 1969 and very high % retention from 1992 to 2009.
Relatively high inputs and losses of DIN in streamwater occurred during 1965–
1978 and 1969–1976, respectively (Fig. 60). Decreased retention occurred during
water-years of forest disturbance (frost 1969, 1988; ice storm 1998; Fig. 61).
Seasonally, the summer growing season is consistently, highly retentive of DIN,
and spring and fall tend to be less retentive than summer and winter (Fig. 62). The
spring season is more variable in terms of % retention of DIN than the other seasons.

Chlorine

Chloride deposition prior to about 1980 was presumably higher than currently
because of coal burning, which releases HCl into the atmosphere (Graedel and
Keene 1996; Lovett et al. 2005). Since about 1980 chloride input in bulk
 100

Annual Retention (%)

80

60

40

20

0
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Year

Fig. 61 Annual retention of dissolved inorganic nitrogen (bulk precipitation input minus stream-
water export for Watershed 6 of the Hubbard Brook Experimental Forest during 1964–2009)

100
Retention (%)

50
0
-50
-100 Winter
-150

100
Retention (%)

50
0
-50
-100
-150 Spring

100
Retention (%)

80
60
40
Summer
20
0
-20
100
Retention (%)

80
60
40 Fall
20
0
-20
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Year
Fig. 62 Average seasonal retention of dissolved inorganic nitrogen (bulk precipitation input
minus streamwater output for Watershed 6 of the Hubbard Brook Experimental Forest during
1964–2009). Winter = December, January, February; Spring = March, April, May; Summer = June,
July, August, September; Fall = October, November
Nitrogen Retention in Forested Watershed-Ecosystems 157

300
Chlorine Mass Balance Dry Deposition
Weathering
Bulk Precipitation
200 Streamwater Outputs

Source / Gain
Net Biomass Uptake
Net Ecosystem Flux

100
Flux (mol/ha-yr)

Sink / Loss
-100

-200

-300
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Water Years

Fig. 63 Annual ecosystem mass balances for chloride in Watershed 6 of the Hubbard Brook
Experimental Forest during 1964–2009. Open circle is bulk deposition inputs; inverted filled
triangle is dry deposition inputs (estimated as 20 % of bulk deposition). Dry deposition in winter
assumed negligible; filled diamond is weathering (estimated at 1.7 mol ha/year); filled triangle
is net biomass uptake; filled circle is streamflow output; and straight line is Net Ecosystem Flux
obtained by difference [modified and extended from Lovett et al. (2005)]

precipitation has been lower and variable, but averaged ~60 mol/ha-year (1.7 kg
Cl/ha-year) (Fig. 63). Wet years in 1973 and particularly 1995 produced high
inputs in bulk precipitation. Dry deposition of chloride is assumed at 20 % of bulk
precipitation values (Lovett et al. 2005). Rather than using net soil release, which
could be large, highly variable, and difficult to estimate (Svensson et al. 2012), we
used chloride weathering, which is very small (1.7–3.4 mol/ha-year) (Lovett et al.
2005). Annual values are calculated as a function of streamflow. Relatively large
amounts of chloride are lost in streamwater and wet years (1973 and 1995) pre-
dominate in this regard (Fig. 63). There is a strong correlation between inputs of
chloride in precipitation and streamwater outputs (Fig. 63). The average streamwa-
ter output from 1964 to 2009 is ~125 mol/ha-year (3.5 kg Cl/ha-year). Small
amounts of chloride (~30 mol Cl/ha-year) are accreted in biomass in an aggrading
forest, but this storage is assumed to be minimal after ~1982 (Lovett et al. 2005).
The result of this approach is that NEF becomes a “proxy” for net soil release. Net
soil release, primarily from soil organic matter, has been relatively large, but may
have decreased with time at the HBEF (Lovett et al. 2005). Thus, the NEF is
158 6 Nutrient Cycles and Mass Balances

variable but mostly negative after 1978, averaging −0.5 kmol Cl for the period
1964–2009 (Table 32; Fig. 63). Despite the decline in atmospheric inputs, the coin-
cidental decline in the net biomass sink allowed for continued streamwater export
of chloride (Fig. 63).

Figures 53, 54, and 55. Ecosystem Diagrams:

Further Explanations

General. The boxes within the ecosystem diagrams represent ecosystem standing
stocks or pools of elements. The values within these boxes are the size of the
elemental pool in mol/ha. The arrows between the various pools represent annual
ecosystem fluxes in mol/ha-year. All values are given for Watershed 6 of the Hubbard
Brook Experimental Forest.
The last cutting of the forest at Hubbard Brook took place between 1909 and
1917. Indications are that the majority of the cutting took place closer to 1917 than
to 1909. Therefore, we made the assumption that major forest regrowth was initiated
around 1915.

Pools. Aboveground Living Biomass: Living biomass was measured by Whittaker

et al. (1974). At the same time representative samples were taken and analyzed by
Likens and Bormann (1970) for various elemental constituents. Whittaker et al.
(1974) also measured the productivity of the forest for two pentads (1956–1960 and
1961–1965). The biomass accretion values in the diagrams for 1964–1969 represent
a mean of these two pentads and are multiplied by the elemental concentrations of
Likens and Bormann (1970). This approach assumes that tissue concentrations
have not changed since 1970. Biomass and chemical data in later pentads come
from Likens et al. (1994) (potassium), (1998) (calcium), and (2002a) (sulfur).

Belowground Living Biomass: Belowground biomass was measured, and elemental

composition was determined in the same manner as the aboveground living
biomass.
Forest Floor: For the early pentad (1964–1969), the biomass and the elemental
composition of the upper organic layers of the soil (to an average depth of 8.9 cm)
were measured by Gosz et al. (1976) and Dominski (1971). The biomass of larger
litter (e.g., limbs and boles of trees) on the forest floor was measured by Covington
(1976), and its elemental composition was determined from Gosz et al. (1972).
The rate of forest floor accretion was estimated by Covington (1976) by measuring
the biomass of a successional sequence of northern hardwood stands ranging in age
from 3 to 200 years. Elemental composition is from Gosz et al. (1976).
Figures 53, 54, and 55. Ecosystem Diagrams: Further Explanations 159

Biomass accretion and elemental composition in the forest floor in later pentads
come from Likens et al. (1994) (potassium), (1998) (calcium), and (2002a)
(sulfur).
Mineral Soil and Rock: Estimates of the amount of organic and inorganic matter in
the mineral soil and of rock in the soil profile to a depth of 45 cm for 1964–1969
were measured by Dominski (1971), and the elemental composition of mineral soil
is from Gosz et al. (1976), Lunt (1932), and Johnson et al. (1968); the elemental
composition of rock is from Johnson et al. (1968) assuming 50 % unweathered
Kinsman quartz monzonite and 50 % unweathered Littleton gneiss. Although an
arbitrary soil depth of 45 cm was used, we assume the majority of the active soil
processes (biogeochemical) occur within the upper 45 cm of soil.
Available Soil Nutrients: Estimated from soil depth, bulk density, and elemental
analyses (Lunt 1932).
Estimates for later pentads are as above.

Elemental Fluxes. Bulk Precipitation: Chemical inputs in rain and snow to the
ecosystem have been measured since 1963 (Buso et al. 2000). Numbers used here
are weighted mean annual inputs (see Table 11).

Litterfall: The fall of litter from forest vegetation to the forest floor was measured,
and its elemental composition was determined by Gosz et al. (1972).
Translocation: The movement of chemicals back into woody tissues at the time
of autumn senescence for leaves is calculated as translocation = (production +
bud litter + root exudates + throughfall and stemflow) – (uptake – shrub and
herb litter).
Throughfall and Stemflow: The washing and leaching of chemicals from the forest
canopy by precipitation was determined by Eaton et al. (1973).
Root Exudation: The loss of chemicals from roots by root exudates was measured
by Smith (1976).
Root Litter: The biomass of root litter was estimated by Whittaker et al. (1974).
Elemental composition of roots was taken from Likens and Bormann (1970).
Estimates for later pentads are as above.
Net Soil Release: As described in Chap. 5.
Uptake: Uptake by forest vegetation was calculated as uptake = litterfall + net through-
fall and stemflow + root exudates + root litter + living biomass accumulation.
Hydrologic Export: Export of dissolved chemicals represents means for 1964–1969
and 1987–1992 (Ca, K) and 1993–1998 (S). Particulate losses represent means for
4 years (Tables 19, 27).
160 6 Nutrient Cycles and Mass Balances

Mass Balances (NEF)– Further Explanations: P = bulk precipitation input, D = dry depo-
sition, NSR = net soil release, S = streamwater output, W = weathering, B = net
biomass uptake, G = net gaseous exchange, NHF = Net Hydrologic Flux.

Calcium
P Measured, ±5 % uncertainty
D Included in P
NSR 37 % of Na NHF based on possible reactants; ±35 % uncertainty
S Measured ±5 % plus particulate matter loss based on average annual streamflow
(increases uncertainty to 6 %)
B Maximum value from Fig. 53 = −202 mol/ha-year; ±30 to 120 % uncertainty
G No gas flux
Potassium
P Measured, ±5 % uncertainty
D Included in P
NSR Used 1964–1969 average (179) until 1982 and =to NHF for 1982–2009; ±50 %
uncertainty
S Measured ±5 % plus particulate losses based on average annual streamflow (increases
uncertainty to 6 %)
B Maximum value from Fig. 54 = −148 mol/ha-year; ±30 to 120 % uncertainty
G No gas flux
Sodium
P Measured, ±5 % uncertainty
D included in P
NSR =NHF, but modified as average NHF value per mm annual streamflow; ±10 % uncertainty
S Measured, ±5 % plus particulate matter losses based on average annual streamflow
(increases uncertainty to 6 %)
B Maximum value from −6.5 mol/ha-year from Bormann and Likens 1979; ±100 %
uncertainty
G No gas flux
Sulfur
P Measured, ±5 % uncertainty
D Used mean of Mitchell et al. (2011) estimate =21 % of P; ±50 % uncertainty based on
span of estimate
NSR Used gradually increasing linear extension of 44 mol/ha 1964–1969 through 128 mol/ha
1993–1998 from Fig. 55 to 2008–2009; ±50 % uncertainty
S Measured, ±5 % uncertainty, plus particulate matter losses based on average annual
streamflow (increases uncertainty to 6 %)
B Maximum value from HBES model = −27 mol/ha; ±50 to 120 % error
G No gas flux
DIN
P Measured, ±5 % uncertainty
D 21 % of P; ±50 % uncertainty based on span of estimate
NSR No weathering; NSR might be = NEF
S Measured, ±5 % uncertainty plus particulate matter losses based on average annual
streamflow (increases uncertainty to 7 %)
B Maximum value based on Bormann and Likens 1979 = 145 mol/ha (9 kg N) ±30 to 120 %
uncertainty
(continued)
Figures 53, 54, and 55. Ecosystem Diagrams: Further Explanations 161

G NET gas flux based on N2O release (4 to 23 mol/ha; Kulkarni et al. 2008) and fixation (0
to 32 mol/ha; Roskoski 1980); both fluxes regressed against wetness (streamwater
range); mostly cancel each other; ±100 % uncertainty on Net value
Chloride
P Measured, ±5 % uncertainty
D Estimated at 20 % based on similar D for S and DIN; ±50 % uncertainty
NSR =W used average weathering inputs in a ratio with Average annual streamflow ±30 %
uncertainty
S Measured, ±5 % uncertainty; no estimate for particulate matter losses available
B Estimated at 32 mol/ha-year max ±~100 % uncertainty
G No gas flux
Chapter 7
The Northern Hardwood Ecosystem
in the Hubbard Brook Valley in Relation
to Other Forest Ecosystems

Forests and woodlands cover some 57 × 106 km2, which is about 38 % of the total
continental area or 11 % of the Earth’s surface. Despite this relatively small area,
92 % of the Earth’s plant biomass and 46 % of its annual net primary production
come from forest (Table 33). The ~80 billion metric tons of dry plant matter pro-
duced (net) each year by forests exceeds the total net primary production of the
oceans, even though the oceans are more than six times larger in area. The animal
biomass in these forests represents some 68 % of the total found on continental land
masses [Whittaker and Likens 1973; but see Cole (2013, pp. 113–114)].
The structure and well-being of forest ecosystems to a large extent reflect the
chemical balance that is extant for the ecosystem. Previous editions of this book
compared biogeochemical data from the HBEF with data from other studies done at
various places throughout the world. In 1977 (First Edition) the Internet did not
exist to host large amounts of publically available biogeochemical data. Thus, the
attempt to capture data on streamwater chemistry from the literature for sites around
the world was interesting and informative. Now, however, massive amounts of data
from the U.S. Geological Survey (http://www.usgs.gov or http://water.usgs.gov/
data/), the U.S. Environmental Protection Agency (http://www.epa.gov), the U.S.
Forest Service Experimental Forests and Ranges, the U.S. National Science
Foundation LTREB and LTER sites, state and local municipalities, and private
groups in the USA are available on the Internet. Likewise, biogeochemical data are
available on the Internet from similar international groups worldwide. Thus, the
original table may be of some historical interest, but is neither representative nor
comprehensive given the available information today. It is possible, however, that
this little book, first published in 1977, did help to promote this proliferation of
long-term biogeochemical data and its availability.
Data in Table 34 have been revised and focused on a few long-term temperate,
forested watersheds in eastern North America for comparison with data from W6 of
the HBEF. While we emphasize in this book that long-term data are critical for
analyzing the biogeochemistry of complicated watershed-ecosystems, these “snap-
shots” focused on one or a few annual inputs, outputs, and net hydrologic fluxes for
selected watersheds and solutes allow for concise, but useful, comparisons.

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2_7, 163

© Springer Science+Business Media New York 2013
164 7 The Northern Hardwood Ecosystem in the Hubbard Brook Valley…

Table 33 Estimates of area, biomass, and productivity circa 1975 for major components of the
biospherea
Plant Net plant Animal Animal
Area, biomass, productivity, 109 biomass, productivity, 106
106 km2 109 dry tons dry tons/year 106 dry tons dry tons/year
Forest and 57 1,700 79.9 680 442
woodlands
Total continental 149 1,837 117.5 1,015 827
area
Total ocean 361 4 55.0 998 3,057
Total biosphere 510 1,841 172.5 2,013 3,884
a
It is assumed that dry matter × 0.45 = carbon. Modified from Whittaker and Likens (1973, 1975)

In general, the findings from these studies (Table 34) show many common
patterns. Several conclusions emerge:
1. Chemicals in precipitation and in dry deposition are significant additions quan-
titatively to the budgets for most of these forest ecosystems.
2. These atmospheric additions are biogeochemically and ecologically important
(see previous chapters).
3. Precipitation inputs of inorganic nitrogen exceed losses in streamwater by ~2 to
>10 times for these forest ecosystems. The net retention of dissolved inorganic
nitrogen in the Cone Pond watershed is the highest and almost complete.
Phosphorus inputs in precipitation probably exceed streamwater outputs as
well, but the values are small and difficult to quantify. The Taughannock Creek
watershed, which includes drainage from agricultural land, shows small net
losses of phosphorus. In this case, human activities or production of animal
wastes probably has overwhelmed the capacity of the natural ecosystem to con-
serve phosphorus within this watershed (e.g., Likens 1974a, b; Likens and
Bormann 1974b).
4. With the exception of potassium in Cone Pond Inlet, there are consistent net
losses of calcium, magnesium, sodium, and potassium from forest ecosystems
that must be made up from weathering. The small net gain of potassium in
Cone Pond Inlet probably reflects reduced amounts of readily weatherable min-
erals and strong biotic retention in this depleted ecosystem.
5. Geologic substrates play a major role in determining the amount and composi-
tion of chemicals lost in streamwater. Watersheds on relatively soluble bedrock
(Sleepers River, VT; Taughannock Creek, NY; Walker Branch, TN) lose large
quantities of the major cations via streamflow.
6. Forest ecosystems in coastal locations or otherwise dominated by maritime air
masses have relatively large inputs of sea salts, e.g., sodium and chloride in
precipitation (e.g., Art 1976; Svensson et al. 2012), whereas interior sites are
more dominated by inputs of calcium and magnesium salts.
7. With the exception of the Turkey Lakes Watershed, which is approximately
balanced, all of the net hydrologic flux values for chloride in these diverse
watershed-ecosystems are negative, suggesting an unmeasured input, but more
likely an internal source of chloride.
 7

Table 34 Nutrient budgets for various temperate zone terrestrial ecosystems in Eastern North America (kg/ha-year)
Atmospheric Streamwater Net gain or Annual
Locationa deposition output loss precipitation (cm) Dominant vegetationb Geologyc
Calcium
Coweeta, NC (1) 4.4 5.3 ‒0.9 215 Q, C, Ac Mg, Msh
HBEF, NH (2) 1.0d 6.6 ‒5.6 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH (3) 1.2d 2.5 ‒1.3 145 Ts, Pic, Ab Ig
Long Island, NY (3) 3.3 9.6e ‒6.3 124 Q, Pin Ss
E. Bear Brook, ME (4) 0.4f 10.1 ‒9.7 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. (5) 6 19 ‒13 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV (6) 2.2d 8.6 ‒6.4 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY (7) 11.2d (0.8g) 182 ‒171 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT (8) 1.3d 226 ‒225 149 Ac, B, Fr, F Mc
Walker Branch, TN (9) 1.3f 147 ‒146 118 Q, C Sc
Biscuit Brook, NY (10) 1.2h 16.5 ‒15.3 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON 2.7h 20.8 ‒18.1 107 Ac, B, Pic, Ts Ig
(Canada) (11)
Magnesium
Coweeta, NC 1.0 3.1 ‒2.1 215 Q, C, Ac Mg, Msh
HBEF, NH 0.3d 2.2 ‒1.9 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 0.3d 0.8 ‒0.5 145 Ts, Pic, Ab Ig
Long Island, NY 2.1 7.3e ‒5.2 124 Q, Pin Ss
The Northern Hardwood Ecosystem in the Hubbard Brook Valley…

E. Bear Brook, ME 0.4f 2.5 ‒2.1 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. 2 6 ‒4 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV 0.3d 3.3 ‒3.0 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY 1.6d (0.1g) 34.8 ‒33.2 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 0.2d 9.5 ‒9.3 149 Ac, B, Fr, F Mc
Walker Branch, TN 0.2f 77 ‒77 118 Q, C Sc
(continued)
165
Table 34 (continued)
166

Atmospheric Streamwater Net gain or Annual

Locationa deposition output loss precipitation (cm) Dominant vegetationb Geologyc
Biscuit Brook, NY 0.3h 3.5 ‒3.2 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON 0.5h 1.5 ‒1.0 107 Ac, B, Pic, Ts Ig
(Canada)
Sodium
7

Coweeta, NC 3.6 8.2 ‒4.6 215 Q, C, Ac Mg, Msh

HBEF, NH 1.1d 7.2 ‒6.0 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 1.0d 4.6 ‒3.6 145 Ts, Pic, Ab Ig
Long Island, NY 17 23e ‒6 124 Q, Pin Ss
E. Bear Brook, ME 5.2f 15.6 ‒10.4 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. 5 26 ‒21 127 Q, Pin Ss, Ssh, Mg, Msh
Taughannock Creek, NY 1.4d (0.3g) 18.9 ‒17.5 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 0.8d 6.5 ‒5.7 149 Ac, B, Fr, F Mc
Walker Branch, TN 1.0f 2.7 ‒1.7 118 Q, C Sc
Biscuit Brook, NY 1.2 2.2 ‒1.0 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON 0.7h 2.1 ‒1.4 107 Ac, B, Pic, Ts Ig
(Canada)
Potassium
Coweeta, NC 4.1 5.0 ‒0.9 215 Q, C, Ac Mg, Msh
HBEF, NH 0.6d 1.4 ‒0.8 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 0.4 0.3 +0.1 145 Ts, Pic, Ab Ig
Long Island, NY 2.4 3.9e ‒1.5 124 Q, Pin Ss
E. Bear Brook, ME 0.5f 1.8 ‒1.3 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. 1 6 ‒5 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV 1.1d 3.6 ‒2.5 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY 1.1d (0.2g) 5.6 ‒4.5 96 (111g) Ac, Ti, Ts Ssh, Sc
The Northern Hardwood Ecosystem in the Hubbard Brook Valley…
 Atmospheric Streamwater Net gain or Annual
Locationa deposition output loss precipitation (cm) Dominant vegetationb Geologyc
7

d
Sleepers River, VT 0.7 11.2 ‒10.5 149 Ac, B, Fr, F Mc
Biscuit Brook, NY 0.2h 2.7 ‒2.5 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON 0.3h 0.6 ‒0.3 107 Ac, B, Pic, Ts Ig
(Canada)
Dissolved inorganic nitrogen (NH4–N + NO3–N)
Coweeta, NC 6.2 0.66 +5.5 215 Q, C, Ac Mg, Msh
HBEF, NH 5.3j 0.23 +5.1 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 3.5d, k <0.1 +3.5 145 Ts, Pic, Ab Ig
E. Bear Brook, ME 1.7f 0.3 +1.4 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. 2k 1k +1k 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV 7.4h 4.4 +3.0 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY 10.3d (5.4g) 5.6 +4.7 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 4.7d 1.9 +2.8 149 Ac, B, Fr, F Mc
Walker Branch, TN 1.7f 0.03 +1.7 118 Q, C Sc
Biscuit Brook, NY 8.2h 3.3 +4.9 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON 25.6h 2.1 +23.5 107 Ac, B, Pic, Ts Ig
(Canada)
Phosphorus
Coweeta, NC 0.04 <0.01 +0.04 215 Q, C, Ac Mg, Msh
HBEF, NH 0.02d 0.01 +0.01 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Taughannock Creek, NY 0.186d 0.197 ‒0.011 96 (111g) Ac, Ti, Ts Ssh, Sc
(0.03g)
The Northern Hardwood Ecosystem in the Hubbard Brook Valley…

Sleepers River, VT 0.15d 0.13 +0.02 149 Ac, B, Fr, F Mc

Walker Branch, TN – 0.03 – 118 Q, C Sc
Biscuit Brook, NY – <0.001 – 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON – 0.01 – 107 Ac, B, Pic, Ts Ig
(Canada)
(continued)
167
 168

Table 34 (continued)
Atmospheric Streamwater Net gain or Annual
Locationa deposition output loss precipitation (cm) Dominant vegetationb Geologyc
Sulfate–Sulfur
Coweeta, NC 7.5 6.4 +1.1 215 Q, C, Ac Mg, Msh
7

HBEF, NH 6.7j 12.6 ‒5.9 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 6.7d 9.7 ‒3.0 145 Ts, Pic, Ab Ig
E. Bear Brook, ME 3.7f 13.2 ‒9.5 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. 8 7 +1 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV 11.6h 8.1 +3.5 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY 18.1d (4.4g) 38 ‒19.9 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 3.8d 17.6 + 149 Ac, B, Fr, F Mc
Walker Branch, TN 5.5f 5.4 +0.1 118 Q, C Sc
Biscuit Brook, NY 8.2h 9.9 ‒1.7 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON 6.1h 5.5 +0.6 107 Ac, B, Pic, Ts Ig
(Canada)
Chloride
Coweeta, NC 6.2 9.1 ‒2.9 215 Q, C, Ac Mg, Msh
HBEF, NH 2.6j 3.3 ‒0.7 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 1.9d 3.4 ‒1.5 145 Ts, Pic, Ab Ig
E. Bear Brook, ME 8.3f 18.0 ‒9.7 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
Taughannock Creek, NY 11.5d (0.7g) 41.7 ‒30.2 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 1.7d 2.4 ‒0.7 149 Ac, B, Fr, F Mc
Walker Branch, TN 2.2f 6.5 ‒4.3 118 Q, C Sc
Biscuit Brook, NY 2.5h 3.7 ‒1.2 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON 0.9h 0.6 +0.3 107 Ac, B, Pic, Ts Ig
(Canada)
The Northern Hardwood Ecosystem in the Hubbard Brook Valley…
 7
a
(1) S. Laester, personal communication for precipitation and streamwater chemistry in Watershed 27 (2008). (2) Present study; average for 2008 water-
year. (3) S. Bailey, unpublished averages for water-year 2007–2008 (4) Woodwell and Whittaker (1967). (5) L. Rustad, I. Fernandez for 2008 water-year.
(6) Gambell and Fisher (1966). (7) M.B. Adams for 2005 calendar year for Watershed 4 (Adams et al. 2006). (8) Likens (1974a, b) for 1970–1971.
(9) S. Bailey and J. Shanley, unpublished averages for water-year 2008 and 2009, Watershed 9, personal communication. (10) Lutz et al. (2012); 2008 water-
year. (11) D. Burns, personal communication, calendar-year, volume-weighted average for 2008. (12) I. Creed, personal communication, average for 2006
water-year catchment #32
b
Vegetation footnotes: Ab, Abies; Ac, Acer; B, Betula, C, Carya; F, Fagus; Fr, Fraxinus, L, Liriodendron; Pin, Pinus; Pic, Picea; Pr, Prunus; Ps, Pseudotsuga;
Q, Quercus; Ti, Tilia; Ts, Tsuga
c
Geology footnotes: Ig, igneous, granitic; Sc, sedimentary, carbonate; Ssh, sedimentary, shale; Ss, sedimentary, sandstone; Mx, Metamorphic (x refers to sub-
scripts used above
d
Bulk precipitation
e
To water table
f
Wet only
g
Wet plus dry data for 2010 from nearby Connecticut Hill CASTNET site (T.J. Butler)
h
Wet plus dry deposition
i
Mean 2007–2009
j
Bulk precipitation plus dry deposition (see Table 17 for bulk precipitation)
k
NO3–N only
The Northern Hardwood Ecosystem in the Hubbard Brook Valley…
169
170 7 The Northern Hardwood Ecosystem in the Hubbard Brook Valley…

8. A close correlation between the amount of precipitation and the loss of nutri-
ents may apply to individual watersheds, such as the HBEF, but as a general
principle, there is no close correlation between the amount of precipitation and
loss of nutrients in forest ecosystems with different climates.
9. Net loss or gain values for sulfur budgets are significantly affected by
anthropogenic contributions to wet and dry deposition. Contaminated air
masses over all of these sites in eastern North America produce relatively high
levels of sulfur and nitrogen in local precipitation, but these concentrations are
declining in recent decades because of reductions in emissions. Interestingly
and importantly, the biogeochemical control over sulfate export in stream water
in forested watershed-ecosystems in eastern North America appears to be shift-
ing from atmospheric S deposition to climate factors through effects on soil
moisture as SO2 emissions are declining and soils are becoming wetter due to
climate change (Mitchell and Likens 2011; Kerr et al. 2012).
10. These forest ecosystems have conservative losses of major nutrients, such as
nitrogen and phosphorus, relative to inputs or amounts cycled internally.
11. Streamwater losses may be very significantly changed by disturbance of the
ecosystem, such as forest cutting, ice storms, agricultural activity, or various
other land management practices, e.g., Taughannock Creek watershed. The
array of biogeochemical responses to clear-cutting and recovery of the northern
hardwood ecosystem at HBEF is discussed in Likens et al. (1970); Bormann
and Likens (1979); Likens and Bormann (1974b); Hornbeck et al. (1986, 1987,
1990); Martin et al. (1986); Likens (2004).
Adjustments for net gaseous input and dry deposition of aerosols were not incor-
porated systematically in Table 34, but much more information is available than 35
years ago. Nevertheless, considerable effort should be devoted to the problem of
measuring dry deposition and gaseous exchange in forested ecosystems in the future
in order to reflect quantitatively these important fluxes in biogeochemical cycling
and fluxes. Recently, Mitchell and colleagues (2011) used a new approach to predict
SO2 concentrations as a function of SO2 emissions, latitude and longitude, and then
estimated dry deposition from relations between concentrations and deposition flux
from the U.S. CASTNET and Canadian CAPMoN networks at 15 sites in northeast-
ern USA and southeastern Canada. Although dry deposition estimates are still rela-
tively uncertain within the mass balance for these systems, this approach seems
promising and allowed the prediction of dry deposition values for the HBEF of 0.5
to 2.5 kg S/ha-year.
Chapter 8
Summary Discussion and Conclusions

This book summarizes current understanding of the biogeochemistry of a northern

hardwood forest ecosystem at the HBEF in the White Mountains of New Hampshire.
It emphasizes the usefulness of the small watershed technique for biogeochemical
measurement in shedding light on ecosystem function and change. The characteris-
tics of the HBEF watersheds revolve around the forest that covers these hillslopes
and the water that passes through this thin veneer of biological organisms.
Trees are long-lived and over time accumulate large amounts of biomass as wood
and nutrients in that wood; it is therefore not surprising that trees dominate the for-
est landscape and are used to name the ecosystem types (plant associations or
biomes). Humans have long looked to forest ecosystems for firewood and timber, as
well as for clean water, wildlife, and recreation, but their demands often have been
at variance with the structure of the forest ecosystem, and large areas have been (and
are being) cleared of forest vegetation. This clearing is currently proceeding at an
unprecedented rate in many forest ecosystems worldwide for human purposes, such
as agriculture, housing, and reservoirs above dams (e.g., Croat 1972; Gómez-Pompa
et al. 1972; Ågren and Andersson 2012, p. 280; Lindenmayer et al. 2013).
Water in some form is also a conspicuous feature of the humid forested land-
scape. Rain and snow not only supply the water that replenishes groundwater
reserves and fills stream and river channels, ponds, lakes, and wetlands but also
provide the water to facilitate weathering and the moisture and many of the nutrients
for forest growth. Chemicals, particularly those without a prominent gaseous phase,
such as calcium, magnesium, or potassium, are transported largely into and out of
ecosystems by moving water.
In forested landscapes much of the water that falls to the land’s surface as pre-
cipitation is first intercepted by the leaves and branches of the trees and herbaceous
layer and then by litter on the ground. Therefore, the kinetic energy and chemical
composition of the water may be appreciably altered before the water comes into
contact with the mineral soil. The diminishment of kinetic energy both minimizes
erosion and regulates nutrient flow. This ostensibly simple and obvious interaction
between falling precipitation and vegetation structure emphasizes that the

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2_8, 171

© Springer Science+Business Media New York 2013
172 8 Summary Discussion and Conclusions

regulation of ecosystem processes, such as biogeochemical flux, depends upon the

structural and functional integrity of the whole ecosystem.
Water that has fallen to the surface may run off overland (minimal in most forests
except during extreme events), infiltrate into the soil, or evaporate. Again, the biotic
portion of the forest ecosystem may alter both the distribution of potential and
kinetic energy and the chemistry of water circulating within, as well as moving
through, an ecosystem.
There are several ways, then, in which water may affect the biogeochemistry of
a forest ecosystem. It is helpful for quantitative studies to select ecosystems that (1)
are watersheds, (2) have a relatively watertight bedrock or other substrate, and (3)
have a relative uniform biogeochemical environment (Bormann and Likens 1967;
Likens and Bormann 1972).
The flux of water and nutrients across an ecosystem’s boundaries, as well as
internal cycling, are vital for the maintenance of a natural ecosystem. Environmental
stress, such as pollution, deforestation, or road building, will alter these conditions
of nutrient flux and cycling and thereby disrupt the function of the ecosystem. Too
little is known about the long-term legacy effects of such alterations or about the
ability of an ecosystem to respond, recover, or improvise following disturbance
(Franklin et al. 2000; Likens and Franklin 2009; Likens and Buso 2012).
The yield of liquid water in streamflow is an important consideration to water-
shed managers. For example, Douglass and Swank (1972) state that “Because the
difference between precipitation input and vapor loss represents the quantity of
water available for man’s use, the watershed manager seeks to reduce the total vapor
loss from the forest vegetation in order to increase the flow of streams” (p. 1).
However, because liquid water that flows out of an ecosystem contains nutrients
(whether at a constant concentration or not), it may be argued that transpiration is a
nutrient conservation mechanism for the ecosystem. Moreover, as precipitation
comes in contact with the vegetation, its chemistry is altered significantly. The
kinetic energy of the falling raindrop is largely absorbed by the vegetation canopy,
and the potential for erosion is reduced. Litter and roots in the forest soil also reduce
erosion. The change of liquid water to vapor by the ecosystem also may have great
impact in reducing the potential for erosion from liquid water. Therefore, the biotic
structure of these forest ecosystems may alter both the potential energy and the
chemistry of water as it passes through the ecosystem. Inadequate consideration has
been given to the interplay of such changes in the hydrologic cycle on nutrient flux
for forest ecosystems. This interplay is of particular concern when the basic struc-
ture of the forest ecosystem is altered by vegetation disturbance or removal or by
climate change (e.g., Likens et al. 1970; Houlton et al. 2003; Groffman et al. 2012).
There is currently considerable effort being made to quantify biogeochemical
relationships for various kinds of terrestrial ecosystems. Based on long-term studies
in the HBEF, it can be emphasized that:
1. Considerable thought be given (a) to selection of the ecosystem, including
boundaries, for study—millions of dollars and thousands of hours of time may
be wasted by a poor initial choice; (b) to methodology, such as frequency and
8 Summary Discussion and Conclusions 173

location of sampling, criteria for acceptance or rejection of particular samples,

and types and reliability of chemical analyses; (c) to measures of uncertainty;
and (d) to the development of a conceptual framework or model for the ecosys-
tem. This methodology is important because small analytical mistakes can eas-
ily be blown out of proportion in the multiplication procedures that are inevitable
in the construction of ecosystem budgets.
2. Ecosystems are open. Water and nutrients continually flux across the boundar-
ies and cycle internally between the various components of the ecosystem.
Studies at HBEF have demonstrated the importance of the abiotic factors and
quantitative mass-balance studies for understanding the structure, function, and
temporal change of forest ecosystems.
3. Long-term records are important in identifying patterns of nutrient flux through
ecosystems. Flux is greatly influenced by hydrology and, during 49- year pre-
cipitation, has ranged from 95 to 186 cm/year in W6 at the HBEF. Measurements
only in these extreme years would have given poor estimates of average condi-
tions. Over the initial 10-year period of the HBES, moreover, annual nitrate and
hydrogen ion inputs in precipitation increased 2.3 and 1.4 times, respectively.
The output of NO3 in streamwater also increased during this period and seemed
to be strongly influenced by soil freezing during two of the 11 winters.
Subsequently, however, the pattern changed dramatically. Since 1980, average
annual bulk precipitation concentrations of hydrogen ion, sulfate, and nitrate
have decreased by ~60, ~40, and ~65 %, respectively. Streamwater concentra-
tions and output of nitrate since ~1990 are consistently the lowest on record;
higher values occur only following major disturbances.
4. Although the chemistry of bulk precipitation is highly variable and dependent
on complex emission patterns and meteorologic conditions, meteorologic input
is an important source of chemicals for the ecosystem. For sulfur, calcium, and
potassium, meteorologic input supplies about 104 times (during 1993–1998),
54 % (during 1987–1992), and 55 % (during 1987–1992), respectively, of the
annual net uptake of these nutrients by green plant biomass.
5. Meteorologic input may be strongly influenced by advertent and inadvertent
manipulation by humans. Acid precipitation is a good example: most acidity in
the rain and snow at HBEF is traceable to sulfur pollution emanating primarily
from fossil fuel burning by electrical utilities in the midwestern USA (Likens
et al. 1972; Likens and Bormann 1974a; Likens 2010), but increasingly with
time and because of federal regulation of SO2 emissions, the input of acid pre-
cipitation became increasingly related to NOX emissions resulting from internal
combustion engines.
6. Acid rain was discovered in North America from samples of rain first collected
at the HBEF in 1963 (Likens et al. 1972; Likens 2010). Long-term research and
monitoring at the HBEF and elsewhere built an evidence-based framework
needed for developing policy and federal and state regulation required to reduce
this form of air pollution and mitigate environmental impacts (e.g., Likens et al.
1996, 1998, 2001, 2002a, 2005; Likens and Lambert 1998; Butler et al. 2001;
Driscoll et al. 2001; Likens 2004; Siccama et al. 2007; Likens and Franklin
174 8 Summary Discussion and Conclusions

2009; Likens and Buso 2010a). Several findings from the long-term HBES
were important to the understanding, policy development, and management of
this major environmental problem, including:
• Changes in emissions of sulfur and nitrogen dioxides, SO2 and NOX (major
precursors to acid rain), as a result of federal legislation, are strongly corre-
lated with changes in sulfate and nitrate concentrations in precipitation and
stream and lake water at HBEF (Likens et al. 2001, 2002a, 2005; Butler
et al. 2001; Likens and Buso 2010a).
• Nitric acid is increasing in importance in precipitation at HBEF and is pre-
dicted to be the dominant acid in precipitation in the future without further
controls on emissions of SO2 and NOX (Likens and Lambert 1998; Likens
and Buso 2010a).
• Calcium and other base cations have been markedly depleted in the soils of
the HBEF as a result of acid rain inputs (Likens et al. 1996, 1998).
– As much as one-half of the pool of exchangeable calcium in the soil has
been depleted during the past 50 years by acid rain (Likens et al. 1998).
– As a result of losses in soil buffering, the forest ecosystem is currently
more sensitive to acid rain impacts than previously thought (Likens et al.
1996, 1998; Likens 2010).
• Enigmatically, rapid forest biomass accumulation has ceased since ~1982 at
HBEF due to increased tree mortality, related in part to the effects of acid
rain (Likens et al. 1998, 2002a; Siccama et al. 2007). This important finding
is the subject of intense, ongoing investigation (Likens and Franklin 2009;
Likens 2010).
7. The ecosystem acts as a “filter” for atmospheric contaminants, such as hydro-
gen ion, nitrogen, sulfur, phosphorus, and certain heavy metals, such as lead,
which accumulate within the ecosystem. The forest ecosystem, therefore, has
been acting as a pollution buffer for society; the ultimate cost of this activity is
unknown, but if continued has the obvious potential for long-term ecosystem
degradation.
8. As a general rule, annual evapotranspiration (mostly transpiration from vegeta-
tion) at HBEF is essentially constant over a wide range of precipitation and
environmental conditions. Water converted to vapor by evapotranspiration can-
not erode the system or transport dissolved substances out of the ecosystem.
Thus, evapotranspiration not only serves as a regulator for the hydrologic cycle,
but also affects nutrient flux as well. Moreover, the forest ecosystem exercises
a strong regulatory function over the chemistry of liquid water passing through
it. This regulation is achieved because the structure of the ecosystem causes all
precipitation to come into intimate contact with, first, the mantle of living and
dead biomass of the ecosystem and then the organic matter and minerals in the
soil profile. Rapid exchange reactions in each of these compartments govern the
chemical composition of liquid water leaving the ecosystem.
8 Summary Discussion and Conclusions 175

9. Evapotranspiration reduces the volume of liquid water in the ecosystem.

Streamwater concentrations in W6 at the HBEF could be increased by 1.5 times
in this way. If evapotranspiration were the only operational factor, the average
streamwater output (average concentration × amount of water) would equal the
average input. Instead, the ecosystem is much more complex. Nutrients are
released internally by weathering and net soil release, accumulated and stored
by vegetation, and returned to the soil in throughfall, stemflow, root exudates,
and litter. The result at HBEF is that on average (1963–2009) the ionic strength
of streamwater (147 µEq/L) is twice that of bulk precipitation (73.5 µEq/L) and
the average streamwater output is 2.6 × 103 Eq/ha-year, whereas the average
bulk precipitation input is 2.1 × 103 Eq/ha-year. In addition, the proportions of
dissolved species change from precipitation to streamflow, further illustrating
the importance of biologic and chemical reactions occurring within the ecosys-
tem. Nitrate input in bulk precipitation is about 2:1 on an equivalent basis with
ammonium, but in streamwater output, the proportion increases to 15:1. For
most of the five decades of study, hydrogen ion and sulfate have dominated
precipitation (pH ~4 to 4.5), whereas calcium and sulfate dominated streamwater
(pH ~5 to 5.5) chemistry.
10. Streamwater chemistry of the forest ecosystem is highly predictable.
Concentrations of sodium and dissolved silica are predictably diluted as stream-
flow rates increase, whereas aluminum, nitrate, hydrogen ion, and dissolved
organic carbon increase as streamflow increases (Johnson et al. 1969). However,
changes in concentrations are relatively small (two to fourfold) in relation to the
great range in flow rates (up to five orders of magnitude, annually). It is well
known that some fixed time series for sampling may seriously underestimate or
overestimate a highly variable parameter (e.g., Claridge 1970); however, most
of the dissolved chemicals in stream water at the HBEF have such constant
concentrations that serious errors are not produced by weekly sampling in most
cases (Johnson et al. 1969; Likens et al. 1967). Concentrations of both nitrate
and potassium in streamwater are highly dependent on seasonal biologic activ-
ity within the ecosystem. The biogeochemical reaction rates within these eco-
systems are rapid; therefore, streamwater chemistry quickly and clearly reflects
environmental conditions within the forest ecosystem.
11. As a consequence of generally weak solute, concentration–flow rate relations in
streams in the HBEF, the output of most individual nutrients, and the total
export of dissolved substances can be closely predicted from the annual output
of streamwater alone.
12. The aggrading northern hardwood forest ecosystem is very resistant to particu-
late matter erosion. This resistance occurs because of its geology (stony till),
high infiltration capacity, and highly efficient mechanisms of controlling water
flow associated with living and dead biomass. In contrast to dissolved sub-
stance removal, which is essentially independent of flow rate, particulate matter
erosion and export are highly dependent on flow rate. Occasional, individual
storms with high discharges do most of the erosion and transport of particulate
176 8 Summary Discussion and Conclusions

matter. In contrast to all of the other nutrients, the bulk of iron and phosphorus
is lost from watershed-ecosystems at HBEF in the particulate form. The stream-
water output for these two elements therefore may reflect more closely the rate
of stream discharge than the amount of annual streamflow. Samples of stream-
water obtained according to a standard time series therefore may be less accu-
rate in estimating streamwater concentrations of iron and phosphorus than for
substances occurring primarily in the dissolved form. Because discharge affects
the concentration of dissolved substances and particulate matter differently, any
management of water quality or total export of these materials requires differ-
ent approaches at different levels of discharge.
13. In particulate matter removal, bed load accounts for slightly more than 50 % of
the material removed. Many studies can only measure suspended matter and
hence tend to underestimate the total output of particulate matter from an
ecosystem.
14. Long-term dissolved substance losses (1963–2009) from the ecosystem (80 kg
of ionic substances, 17 kg of dissolved silica, and 23 kg of dissolved organic
matter per hectare-year) are about four times greater than particulate matter
losses (11 kg of organic matter plus 22 kg of inorganic particles). Total losses
of dissolved substances and particulate matter (198 kg/ha-year) in stream water,
circa 1970, are about 1.5 times greater than total inputs in bulk precipitation
(134 kg/ha-year) in that same period.
15. The long-term lowering of the landscape in the HBEF, which is dominated by
forests, is primarily caused by solution losses and, to a lesser extent, by particu-
late matter losses. Slow mass movements, such as soil creep, may deliver mate-
rials to stream beds, where they are removed. At higher elevations and with
steeper slopes, debris avalanches also may play a major role in lowering the
landscape. Another factor that may be of considerable importance is distur-
bances that temporarily destroy the forest canopy and make the ecosystem
more vulnerable to both solution and particulate matter losses (e.g., Likens
et al. 1970; Bormann and Likens 1979; Houlton et al. 2003).
16. In establishing input–output budgets for individual nutrients, both direction of
change, i.e., whether there is a net loss or gain, and magnitude of change are
important. Budgets have been done in two ways in our studies, by NHF (stream-
water output minus bulk precipitation input) and by NEF where all of the major
ecosystem components are measured or estimated. For some ions, the NHF
budget parameters may be established with one or a few years of data, for
example, direction: Ca2+(−), Mg2+(−), Na+(−), Al3+(−), NH4+ (+), H+(+), SO42–
(−), PO43– (+), dissolved silica (−), HCO3− (−). For other ions, K+, NO3−, and
Cl−, one or a few years gave unreliable results for direction and magnitude. For
some budgets, therefore, longer term studies are necessary to obtain valid bud-
getary data.
17. Not only must net losses of ions from the ecosystem be included in an estimate
of weathering flux, but those elements built into any net accumulation of bio-
mass must also be included. At HBEF failure to include annual biomass accre-
tion of nutrients results in an ecosystem weathering estimate that is 50 % too low.
8 Summary Discussion and Conclusions 177

18. Three input–output patterns emerge when small watershed data are examined
on a monthly basis: output exceeds input in every month for magnesium, cal-
cium, sodium, aluminum, and dissolved silica; output is less than input for
ammonium, hydrogen ion, and phosphate; and crossover patterns, where for a
time output exceeds input and then output is less than input, occur for potas-
sium, sulfate, chloride, dissolved organic carbon, and nitrate. These patterns are
distributed among elements with both sedimentary and atmospheric cycles and
reflect a complex interaction involving such factors as bulk precipitation input,
biologic activity, and climatic variations.
19. Chemical weathering of minerals within the ecosystem is driven by water and
by protons derived from hydrogen ions in precipitation and those generated
internally within the system by biogeochemical reactions involving carbon,
nitrogen, and sulfur. About 50 % of the hydrogen ions generated during the
weathering reaction is derived from external sources, i.e., acid precipitation,
and 50 % is from internal biogeochemical sources. However, this estimate is
based on the assumption that all hydrogen ions, external plus internal, are con-
sumed in exchange reactions.
20. Complete mass balances (NEF) for the ecosystems at HBEF reveal certain
fundamental features of nutrient flux and cycling not otherwise apparent. These
are the following: (a) wet and dry deposition are major ecosystem sources of
nutrients, particularly for sulfur, nitrogen, chloride, and phosphorus; (b) weath-
ering/net soil release is a major source for calcium, magnesium, potassium, and
sodium; (c) biologic activity and photosynthesis play significant roles in the
input and storage of carbon and nitrogen and are important (impaction surfaces
plus gaseous uptake) for sulfur; (d) the northern hardwood forest shows abso-
lute gains for hydrogen ion, dissolved inorganic nitrogen, potassium, and phos-
phorus and losses for dissolved silica, calcium, sodium, sulfur, aluminum,
magnesium, chlorine, and iron. The losses of dissolved silica, calcium, sodium,
magnesium, potassium, and iron are made up by weathering of primary miner-
als within the ecosystem. Available weathering substrate may be decreasing
with time, particularly for elements like calcium.
21. Calcium, an example of a sedimentary cycle, tends to be held within the sys-
tem, with net losses in 1987–1992 being but a small proportion (2.7 %) of the
calcium in the available pool (Fig. 53; Likens et al. 1998). However, net losses
are 3.4 larger than annual net uptake by green plants. Although sulfur has a
gaseous cycle, some 95 % of the sulfur in the ecosystem is found in the soil
(Fig. 55; Likens et al. 2002a). Net gaseous uptake and aerosol deposition of
sulfur accounted for about 18 % of the total meteorologic input for the system
during 1993–1998. Meteorologic inputs dominate, in general, and release by
weathering is small, but increasing in relative importance as meteorologic
inputs decrease (Likens et al. 2002a).
22. Because the HBEF has negligible amounts of nitrogen-bearing rocks, aggrad-
ing forest ecosystems have accumulated significant amounts of nitrogen from
atmospheric inputs during the past 14,000 years since glacial retreat.
178 8 Summary Discussion and Conclusions

23. Many of the important biogeochemical relationships and transformations within

the forest ecosystem are regulated by microorganisms, such as nitrogen fixa-
tion, nitrification, and denitrification or sulfur oxidation and reduction, at HBEF
(e.g., Zhang et al. 1999; Findlay 2013; Groffman and Rosi-Marshall 2013).
24. Streamwater chemistry collected and measured at high density in all tributaries
throughout the Hubbard Brook Valley during spring (May–July) and fall
(October–December) of 2001 gave a remarkable new landscape view of stream-
water chemistry in the valley (Likens and Buso 2006). Stream water chemistry
varied throughout the valley in relation to changes in elevation, distance from
origin of streamflow, drainage area, and type of drainage. Concentrations of
hydrogen ion, aluminum, and dissolved organic carbon varied by two to tenfold
in relation to these factors, but other elements, e.g., chloride and dissolved sil-
ica, changed relatively little throughout the valley. The main Hubbard Brook, a
fifth-order river, changed concentration very little throughout its 13-km length
despite inputs from some 20 major tributaries.
25. Based on pH as a critical diagnostic variable in the Hubbard Brook Valley, the
gauging stations in the experimental watersheds were sited near the pH inflec-
tion points of the longitudinal distance of these headwater streams. Thus, these
locations are nearly optimal for measuring change in acid–base chemistry of
stream water.
26. Stream water was acidified from atmospheric inputs of H2SO4 and HNO3 dur-
ing 1963 to 1969 and then is recovering from this acidification (r2 = 0.96) from
1970 to the present (Likens et al. 1996; Likens and Buso 2012). The projected
preindustrial revolution/post-acidic deposition value is 39 µeq/L for the sum of
base cations and 11 µeq/L for the sum of acid anions (sulfate plus nitrate)
(Likens and Buso 2012).
27. Long-term data show a dramatic and relentless dilution of both precipitation
and streamwater concentrations at various watershed scales of the HBEF
(Likens and Buso 2012). Concentrations project to approximately demineral-
ized water in the near future. Since demineralized water would be unrealistic
for surface waters, theoretical baselines of 3 and 5 µS/cm (electrical conductiv-
ity), 7 and 39 µeq/L (sum of base cations), <1 and 14 µeq/L (acid-neutralizing
capacity), and 5.5 (pH) are estimated for precipitation and stream water, respec-
tively (Likens and Buso 2012).
28. Input–output budgets for a diverse series of relatively undisturbed vegetated
watersheds throughout eastern North America have many patterns in common
with HBEF. Although the magnitude of precipitation, geologic substrate, veg-
etation type, and proximity to marine or anthropogenic emissions varies for
these diverse watersheds, the constant pattern that emerges from these data is
remarkable. Inputs of phosphorus and nitrogen in bulk precipitation are gener-
ally greater than losses in drainage waters on an annual basis. Conversely, out-
puts of calcium, magnesium, sodium, and potassium are generally greater than
precipitation inputs. Net losses of potassium are small compared to inputs. The
magnitude and direction for sulfate budgets are variable and depend largely on
geologic substrate and intensity of air pollution.
8 Summary Discussion and Conclusions 179

29. Ecosystem studies are now at the stage to provide quantitative answers for
important ecological questions and management interventions. The small
watershed technique is powerful in providing an experimental framework for
these studies, which must consider geologic and meteorologic aspects as well
as biologic features. Such questions include: (a) How do ecosystems change
with time? Do they become more or less efficient in cycling and storing nutri-
ents? (b) What are the long-term effects of disturbance on the biogeochemical
flux for ecosystems (see, e.g., Bernal et al. 2012)? (c) How do the biogeochemi-
cal dynamics of individual elements interact (e.g., Likens 1981; Schlesinger
et al. 2011; Schlesinger and Bernhardt 2013)? (d) How would weathering rates
change in relation to changes in rates of biologic activity and/or meteorologic
input? Are there conditions under which weathering does not occur? (e) Of
critical importance is the finding that some of the ions produced by weathering
do not move downhill (sensu Aldo Leopold), but are instead accumulated in the
growing vegetation. If the forest were to reach some kind of steady state or
rapid decline of growth, what would happen to the biogeochemistry of a com-
plex ecosystem? This latter question is of utmost importance and urgency at the
HBEF (e.g., Likens and Franklin 2009; Lindenmayer and Likens 2010, p. 123).
30. There are many major unanswered questions emanating from the long-term
research of the HBES:
– how can biogeochemical results from small plot scales be extrapolated
quantitatively to the watershed scale (e.g., denitrification, soil acidity, pri-
mary and secondary weathering, soil frost, hydro-pedology) when the
watershed-ecosystem approach is not applicable or difficult?
– is biological control becoming more important in the biogeochemistry of the
Hubbard Brook Valley as the input of acid rain decreases? Will organic
acids play an increasing role in chemical reactions in the future?
– how, or at what rate, are soils recovering from the impact of decades of acid
rain? How will base cations be replenished in the rooting zone? How long
will it take to deplete stored sulfur? Will nitrogen ‘saturation’ occur before
recovery is complete?
– what will be the ecological and biogeochemical effects of relentless dilution
of precipitation and stream water in the Hubbard Brook Valley?
– how can important ecosystem relationships be predicted with reliability in a
changing future, based on past relationships under very different
conditions?
– how can the small watershed approach be applied productively in attempts
to integrate even more diverse and more complicated information generated
in the future?
31. Because so much has changed biogeochemically at the HBEF during the past
50 years, several of the fundamental analyses and syntheses need to be extended,
redone, or revised, e.g., long-term hydrological and chemical records (Chaps.
2, 3 and 4), role of hydrogen ion in weathering (Chap. 5), nutrient budget rela-
tionships (Chap. 6), quantitative elucidation of the nitrogen cycle, quantitative
180 8 Summary Discussion and Conclusions

impact of climate change on hydrology, and mass balances (Chaps. 2 and 6).
Indeed, as in all good research, there is no lack of important research questions,
even after five decades of intensive effort of the HBES.

Epilog

Aldo Leopold was a very wise and perceptive observer of nature. Some might have
called him a conservationist; others might have called him an ecologist or a biogeo-
chemist. Actually he was a brilliant, yet humble, student of nature in the purest
sense. Much of what this little book has described in quantitative terms Aldo
Leopold outlined in general terms >60 year ago in his essay Odyssey. Odyssey of
Atom X captures the complicated theory of nutrient flux and cycling clearly and
efficiently. Resources (nutrients) “enter” an ecosystem via atmospheric deposition
and weathering release (chemical, physical, and biological); do “work” within the
ecosystem by supporting productivity, food webs, decomposition, etc.; and then are
lost from the ecosystem via drainage water or volatile or particulate loss to the
atmosphere across the ecosystem’s boundary and by biotic uptake and long-term
storage. Leopold has said this more effectively and beautifully in the Odyssey of
Atom X than anyone!
His insight and prose have been an inspiration to me for a long time. Therefore,
it seems appropriate for Aldo Leopold to have the last word in this book and so,
from A Sand County Almanac (pp. 104–108) (1949), the following:

Odyssey

X had marked time in the limestone ledge since the Paleozoic seas covered the land.
Time, to an atom locked in a rock, does not pass.
The break came when a bur-oak root nosed down a crack and began prying and
sucking. In the flush of a century the rock decayed, and X was pulled out and up into
the world of living things. He helped build a flower, which became an acorn, which
fattened a deer, which fed an Indian, all in a single year.
From his berth in the Indian’s bones, X joined again in chase and flight, feast and
famine, hope and fear. He felt these things as changes in the little chemical pushes
and pulls that tug timelessly at every atom. When the Indian took his leave of the
prairie, X moldered briefly underground, only to embark on a second trip through
the bloodstream of the land.
This time it was a rootlet of bluestem that sucked him up and lodged him in a leaf
that rode the green billows of the prairie June, sharing the common task of hoarding
sunlight. To this leaf also fell an uncommon task: flicking shadows across a plover’s
eggs. The ecstatic plover, hovering overhead, poured praises on something perfect:
perhaps the eggs, perhaps the shadows, or perhaps the haze of pink phlox that lay on
the prairie.
Epilog 181

When the departing plovers set wing for the Argentine, all the bluestems waved
farewell with tall new tassels. When the first geese came out of the north and all the
bluestems glowed wine-red, a forehanded deermouse cut the leaf in which X lay,
and buried it in an underground nest, as if to hide a bit of Indian summer from the
thieving frosts. But a fox detained the mouse, molds and fungi took the nest apart,
and X lay in the soil again, foot-loose and fancy-free.
Next he entered a tuft of side-oats grama, a buffalo, a buffalo chip, and again the
soil. Next a spiderwort, a rabbit, and an owl. Thence a tuft of sporobolus.
All routines come to an end. This one ended with a prairie fire, which reduced the
prairie plants to smoke, gas, and ashes. Phosphorus and potash atoms stayed in the
ash, but the nitrogen atoms were gone with the wind. A spectator might, at this
point, have predicted an early end of the biotic drama, for with fires exhausting the
nitrogen, the soil might well have lost its plants and blown away.
But the prairie had two strings to its bow. Fires thinned its grasses, but they thick-
ened its stand of leguminous herbs; prairie clover, bush clover, wild bean, vetch,
lead-plant, trefoil, and Baptisia, each carrying its own bacteria housed in nodules on
its rootlets. Each nodule pumped nitrogen out of the air into the plant, and then
ultimately into the soil. Thus the prairie savings bank took in more nitrogen from its
legumes than it paid out to its fires. That the prairie is rich is known to the humblest
deermouse; why the prairie is rich is a question seldom asked in all the still lapse of
ages.
Between each of his excursions through the biota, X lay in the soil and was car-
ried by the rains, inch by inch, downhill. Living plants retarded the wash by
impounding atoms; dead plants by locking them to their decayed tissues. Animals
ate the plants and carried them briefly uphill or downhill, depending on whether
they died or defecated higher or lower than they fed. No animal was aware that the
altitude of his death was more important than his manner of dying. Thus a fox
caught a gopher in a meadow, carrying X uphill to his bed on the brow of a ledge,
where an eagle laid him low. The dying fox sensed the end of his chapter in foxdom,
but not the new beginning in the odyssey of an atom.
An Indian eventually inherited the eagle’s plumes, and with them propitiated the
Fates, whom he assumed had a special interest in Indians. It did not occur to him
that they might be busy casting dice against gravity; that mice and men, soils and
songs, might be merely ways to retard the march of atoms to the sea.
One year, while X lay in a cottonwood by the river, he was eaten by a beaver, an
animal that always feeds higher than he dies. The beaver starved when his pond
dried up during a bitter frost. X rode the carcass down the spring freshet, losing
more altitude each hour than heretofore in a century. He ended up in the silt of a
backwater bayou, where he fed a crayfish, a coon, and then an Indian, who laid him
down to his last sleep in a mound on the riverbank. One spring an oxbow caved the
bank, and after one short week of freshet X lay again in his ancient prison, the sea.
An atom at large in the biota is too free to know freedom; an atom back in the sea
has forgotten it. For every atom lost to the sea, the prairie pulls another out of the
decaying rocks. The only certain truth is that its creatures must suck hard, live fast,
and die often, lest its losses exceed its gains.
***
182 8 Summary Discussion and Conclusions

It is the nature of roots to nose into cracks. When Y was thus released from the
parent ledge, a new animal had arrived and begun redding up the prairie to fit his
own notions of law and order. An oxteam turned the prairie sod, and Y began a suc-
cession of dizzy annual trips through a new grass called wheat.
The old prairie lived by the diversity of its plants and animals, all of which were
useful because the sum total of their co-operations and competitions achieved con-
tinuity. But the wheat farmer was a builder of categories; to him only wheat and
oxen were useful. He saw the useless pigeons settle in clouds upon his wheat, and
shortly cleared the skies of them. He saw the chinch bugs take over the stealing job,
and fumed because here was a useless thing too small to kill. He failed to see the
downward wash of over-wheated loam, laid bare in spring against the pelting rains.
When soil-wash and chinch bugs finally put an end to wheat farming, Y and his like
had already traveled far down the watershed.
When the empire of wheat collapsed, the settler took a leaf from the old prairie
book: he impounded his fertility in livestock, he augmented it with nitrogen-
pumping alfalfa, and he tapped the lower layers of the loam with deep-rooted corn.
But he used his alfalfa, and every other new weapon against wash, not only to
hold his old plowings, but also to exploit new ones which, in turn, needed holding.
So, despite alfalfa, the black loam grew gradually thinner. Erosion engineers
built dams and terraces to hold it. Army engineers built levees and wing-dams to
flush it from the rivers. The rivers would not flush, but raised their beds instead, thus
choking navigation. So the engineers built pools like gigantic beaver ponds, and Y
landed in one of these, his trip from rock to river completed in one short century.
On first reaching the pool, Y made several trips through water plants, fish, and
waterfowl. But engineers build sewers as well as dams, and down them comes the
loot of all the far hills and the sea. The atoms that once grew pasque-flowers to greet
the returning plovers now lie inert, confused, imprisoned in oily sludge.
Roots still nose among the rocks. Rains still pelt the fields. Deermice still hide
their souvenirs of Indian summer. Old men who helped destroy the pigeons still
recount the glory of the fluttering hosts. Black and white buffalo pass in and out of
red barns, offering free rides to itinerant atoms (“Odyssey” was first published in the
May-June 1942 issue of Audubon).
 ERRATUM

Biogeochemistry of a Forested Ecosystem

Gene E. Likens

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2,

© Springer Science+Business Media New York 2013

DOI 10.1007/978-1-4614-7810-2_9

The publisher regrets that the online and print versions of this book were published
with some errors on the below pages of this book. The correct datas are listed below:

On page xx, in the “Contents” a new entry “Trace Metals” has been added
before “Throughfall and Stemflow”. The revised contents is given below:

Elevational Effects .................................................................................. 43

Acid Precipitation ................................................................................... 44
Trace Metals................................................................................................ 64
Throughfall and Stemflow .......................................................................... 66
Streamwater Chemistry............................................................................... 68

The online version of the original book can be found at E1

1
http://dx.doi.org/10.1007/978-1-4614-7810-2
E2 Erratum

On page 14, Photograph 3 has been replaced and the new photograph is given
below:

Photograph 3 Hubbard Brook at an elevation of about 215 m. The characteristic boulder

substrate and “stair-step” nature of this fifth-order stream are evident
Erratum E3

On page 104, Photograph 8 has been replaced and the new photograph is given
below:

Photograph 8 Collecting a streamwater sample from Watershed 4 at the Hubbard Brook

Experimental Forest (Photo by D.C. Buso)

On page 111, the term “HBEF” in the caption of Table 25 has been expanded to
“Hubbard Brook Experimental Forest”.

Table 25 Seasonal average streamwater exports for Watershed 6 of the Hubbard Brook
Experimental Forest for two decades (1964–1973 and 2000–2009)

Table 34 on page 165 has been modified and the new version is given below:
Table 34 Nutrient budgets for various temperate zone terrestrial ecosystems in Eastern North America (kg/ha-year)
E4

Atmospheric Streamwater Net gain Annual

Locationa deposition output or loss precipitation (cm) Dominant vegetationb Geologyc
Calcium
Coweeta, NC (1) 4.4 5.3 ‒0.9 215 Q, C, Ac Mg, Msh
HBEF, NH (2) 1.0d 6.6 ‒5.6 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH (3) 1.2d 2.5 ‒1.3 145 Ts, Pic, Ab Ig
Long Island, NY (4) 3.3 9.6e ‒6.3 124 Q, Pin Ss
E. Bear Brook, ME (5) 0.4f 10.1 ‒9.7 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. (6) 6 19 ‒13 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV (7) 2.2d 8.6 ‒6.4 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY (8) 11.2d (0.8g) 182 ‒171 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT (9) 1.3d 226 ‒225 149 Ac, B, Fr, F Mc
Walker Branch, TN (10) 1.3f 147 ‒146 118 Q, C Sc
Biscuit Brook, NY (11) 1.2h 16.5 ‒15.3 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON (Canada) (12) 2.7h 20.8 ‒18.1 107 Ac, B, Pic, Ts Ig
Magnesium
Coweeta, NC 1.0 3.1 ‒2.1 215 Q, C, Ac Mg, Msh
HBEF, NH 0.3d 2.2 ‒1.9 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 0.3d 0.8 ‒0.5 145 Ts, Pic, Ab Ig
Long Island, NY 2.1 7.3e ‒5.2 124 Q, Pin Ss
E. Bear Brook, ME 0.4f 2.5 ‒2.1 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. 2 6 ‒4 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV 0.3d 3.3 ‒3.0 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY 1.6d (0.1g) 34.8 ‒33.2 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 0.2d 9.5 ‒9.3 149 Ac, B, Fr, F Mc
Walker Branch, TN 0.2f 77 ‒77 118 Q, C Sc
Biscuit Brook, NY 0.3h 3.5 ‒3.2 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON (Canada) 0.5h 1.5 ‒1.0 107 Ac, B, Pic, Ts Ig
Erratum
Sodium
Coweeta, NC 3.6 8.2 ‒4.6 215 Q, C, Ac Mg, Msh
Erratum

HBEF, NH 1.1d 7.2 ‒6.0 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 1.0d 4.6 ‒3.6 145 Ts, Pic, Ab Ig
Long Island, NY 17 23e ‒6 124 Q, Pin Ss
E. Bear Brook, ME 5.2f 15.6 ‒10.4 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. 5 26 ‒21 127 Q, Pin Ss, Ssh, Mg, Msh
Taughannock Creek, NY 1.4d (0.3g) 18.9 ‒17.5 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 0.8d 6.5 ‒5.7 149 Ac, B, Fr, F Mc
Walker Branch, TN 1.0f 2.7 ‒1.7 118 Q, C Sc
Biscuit Brook, NY 1.2 2.2 ‒1.0 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON (Canada) 0.7h 2.1 ‒1.4 107 Ac, B, Pic, Ts Ig
Potassium
Coweeta, NC 4.1 5.0 ‒0.9 215 Q, C, Ac Mg, Msh
HBEF, NH 0.6d 1.4 ‒0.8 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 0.4 0.3 +0.1 145 Ts, Pic, Ab Ig
Long Island, NY 2.4 3.9e ‒1.5 124 Q, Pin Ss
E. Bear Brook, ME 0.5f 1.8 ‒1.3 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. 1 6 ‒5 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV 1.1d 3.6 ‒2.5 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY 1.1d (0.2g) 5.6 ‒4.5 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 0.7d 11.2 ‒10.5 149 Ac, B, Fr, F Mc
Biscuit Brook, NY 0.2h 2.7 ‒2.5 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON (Canada) 0.3h 0.6 ‒0.3 107 Ac, B, Pic, Ts Ig
Dissolved inorganic nitrogen (NH4–N + NO3–N)
Coweeta, NC 6.2 0.66 +5.5 215 Q, C, Ac Mg, Msh
HBEF, NH 5.3j 0.23 +5.1 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 3.5d, k <0.1 +3.5 145 Ts, Pic, Ab Ig
E. Bear Brook, ME 1.7f 0.3 +1.4 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
E5

(continued)
Table 34 (continued)
E6

Atmospheric Streamwater Net gain Annual

Locationa deposition output or loss precipitation (cm) Dominant vegetationb Geologyc
S.E. U.S. 2k 1k +1k 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV 7.4h 4.4 +3.0 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY 10.3d (5.4g) 5.6 +4.7 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 4.7d 1.9 +2.8 149 Ac, B, Fr, F Mc
Walker Branch, TN 1.7f 0.03 +1.7 118 Q, C Sc
Biscuit Brook, NY 8.2h 3.3 +4.9 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON (Canada) 25.6h 2.1 +23.5 107 Ac, B, Pic, Ts Ig
Phosphorus
Coweeta, NC 0.04 <0.01 +0.04 215 Q, C, Ac Mg, Msh
HBEF, NH 0.02d 0.01 +0.01 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Taughannock Creek, NY 0.186d (0.03g) 0.197 ‒0.011 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 0.15d 0.13 +0.02 149 Ac, B, Fr, F Mc
Walker Branch, TN – 0.03 – 118 Q, C Sc
Biscuit Brook, NY – <0.001 – 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON (Canada) – 0.01 – 107 Ac, B, Pic, Ts Ig
Sulfate–Sulfur
Coweeta, NC 7.5 6.4 +1.1 215 Q, C, Ac Mg, Msh
HBEF, NH 6.7j 12.6 ‒5.9 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 6.7d 9.7 ‒3.0 145 Ts, Pic, Ab Ig
E. Bear Brook, ME 3.7f 13.2 ‒9.5 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
S.E. U.S. 8 7 +1 127 Q, Pin Ss, Ssh, Mg, Msh
Fernow, WV 11.6h 8.1 +3.5 125 Q, L, Ac, B, F Ss, Ssh, Ms, Msh
Taughannock Creek, NY 18.1d (4.4g) 38 ‒19.9 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 3.8d 17.6 + 149 Ac, B, Fr, F Mc
Walker Branch, TN 5.5f 5.4 +0.1 118 Q, C Sc
Biscuit Brook, NY 8.2h 9.9 ‒1.7 165i Ab, Ac, B, F, Pic, Ts Ss
Erratum

Turkey Lakes Watershed, ON (Canada) 6.1h 5.5 +0.6 107 Ac, B, Pic, Ts Ig
Chloride
Coweeta, NC 6.2 9.1 ‒2.9 215 Q, C, Ac Mg, Msh
Erratum

HBEF, NH 2.6j 3.3 ‒0.7 165 Ac, F, B, Pic, Ab, Ts Ig, Msh
Cone Pond Inlet, NH 1.9d 3.4 ‒1.5 145 Ts, Pic, Ab Ig
E. Bear Brook, ME 8.3f 18.0 ‒9.7 159 F, Ac, B, Pic, Ab, Ts Ig, Msh
Taughannock Creek, NY 11.5d (0.7g) 41.7 ‒30.2 96 (111g) Ac, Ti, Ts Ssh, Sc
Sleepers River, VT 1.7d 2.4 ‒0.7 149 Ac, B, Fr, F Mc
Walker Branch, TN 2.2f 6.5 ‒4.3 118 Q, C Sc
Biscuit Brook, NY 2.5h 3.7 ‒1.2 165i Ab, Ac, B, F, Pic, Ts Ss
Turkey Lakes Watershed, ON (Canada) 0.9h 0.6 +0.3 107 Ac, B, Pic, Ts Ig
a
(1) S. Laester, personal communication for precipitation and streamwater chemistry in Watershed 27 (2008). (2) Present study; average for 2008 water-
year. (3) S. Bailey, unpublished averages for water-year 2007–2008 (4) Woodwell and Whittaker (1967). (5) L. Rustad, I. Fernandez for 2008 water-year.
(6) Gambell and Fisher (1966). (7) M.B. Adams for 2005 calendar year for Watershed 4 (Adams et al. 2006). (8) Likens (1974a, b) for 1970–1971.
(9) S. Bailey and J. Shanley, unpublished averages for water-year 2008 and 2009, Watershed 9, personal communication. (10) Lutz et al. (2012); 2008
water-year. (11) D. Burns, personal communication, calendar-year, volume-weighted average for 2008. (12) I. Creed, personal communication, average for
2006 water-year catchment #32
b
Vegetation footnotes: Ab, Abies; Ac, Acer; B, Betula, C, Carya; F, Fagus; Fr, Fraxinus, L, Liriodendron; Pin, Pinus; Pic, Picea; Pr, Prunus; Ps, Pseudotsuga;
Q, Quercus; Ti, Tilia; Ts, Tsuga
c
Geology footnotes: Ig, igneous, granitic; Sc, sedimentary, carbonate; Ssh, sedimentary, shale; Ss, sedimentary, sandstone; Mx, Metamorphic (x refers to
subscripts used above
d
Bulk precipitation
e
To water table
f
Wet only
g
Wet plus dry data for 2010 from nearby Connecticut Hill CASTNET site (T.J. Butler)
h
Wet plus dry deposition
i
Mean 2007–2009
j
Bulk precipitation plus dry deposition (see Table 17 for bulk precipitation)
k
NO3–N only
E7
E8 Erratum

On page 170, in the paragraph beginning “Adjustments for net gaseous

input....” the sentence “Nevertheless, considerable effort...” has been moved to
the end of the paragraph. The revised page is listed below:

Adjustments for net gaseous input and dry deposition of aerosols were not incorpo-
rated systematically in Table 34, but much more information is available than 35
years ago. Recently, Mitchell and colleagues (2011) used a new approach to predict
SO2 concentrations as a function of SO2 emissions, latitude and longitude, and then
estimated dry deposition from relations between concentrations and deposition flux
from the U.S. CASTNET and Canadian CAPMoN networks at 15 sites in northeast-
ern USA and southeastern Canada. Although dry deposition estimates are still rela-
tively uncertain within the mass balance for these systems, this approach seems
promising and allowed the prediction of dry deposition values for the HBEF of 0.5
to 2.5 kg S/ha-year. Nevertheless, considerable effort should be devoted to the prob-
lem of measuring dry deposition and gaseous exchange in forested ecosystems in
the future in order to reflect quantitatively these important fluxes in biogeochemical
cycling and fluxes.

On page 173, in the list entry 3, the sentence beginning “Flux is greatly influ-
enced...” has been revised and the corrected sentence is given below:

Flux is greatly influenced by hydrology and, during 49 years, precipitation has

ranged from 95 to 186 cm/year in W6 at the HBEF.

On page 174, a reference citation “van Doorn et al. 2011” has been added
after “Likens 2010” at the end of list entry 6, in the paragraph beginning
“Enigmatically, rapid forest...”.

Enigmatically, rapid forest biomass accumulation has ceased since ~1982 at HBEF
due to increased tree mortality, related in part to the effects of acid rain (Likens et
al. 1998, 2002a; Siccama et al. 2007). This important finding is the subject of
intense, ongoing investigation (Likens and Franklin 2009; Likens 2010; van Doorn
et al. 2011).

On page 178, in list entry 23, the comma after “reduction” has been removed
and the corrected entry is given below:

23. Many of the important biogeochemical relationships and transformations within

the forest ecosystem are regulated by microorganisms, such as nitrogen fixation,
nitrification, and denitrification or sulfur oxidation and reduction at HBEF
(e.g., Zhang et al. 1999; Findlay 2013; Groffman and Rosi-Marshall 2013).
Erratum E9

On page 178, in list entry 24, the word “Stream water” has been closed up to
read “Streamwater” in the sentence “Streamwater chemistry varied....”.

24. Streamwater chemistry collected and measured at high density in all tributaries
throughout the Hubbard Brook Valley during spring (May–July) and fall
(October–December) of 2001 gave a remarkable new landscape view of stream-
water chemistry in the valley (Likens and Buso 2006 ). Streamwater chemistry
varied throughout the valley in relation to changes in elevation, distance from
origin of streamflow, drainage area, and type of drainage. Concentrations of
hydrogen ion, aluminum, and dissolved organic carbon varied by two to tenfold
in relation to these factors, but other elements, e.g., chloride and dissolved
silica, changed relatively little throughout the valley. The main Hubbard Brook,
a fifth-order river, changed concentration very little throughout its 13-km length
despite inputs from some 20 major tributaries.

In References, on page 193, the references “Likens GE, Bormann FH (1970);

Likens GE, Bormann FH (1985); Likens GE, Bormann FH (1995)” have been
rearranged and the revised order is given below:

Likens GE, Bormann FH (1970) Chemical analyses of plant tissues from the Hubbard Brook
Ecosystem in New Hampshire. Bulletin 79, Yale University School of Forestry, New Haven,
CT, 25 pp
Likens GE, Bormann FH (1972) Nutrient cycling in ecosystems. In: Wiens J (ed) Ecosystem
structure and function. Oregon State University Press, Corvallis, OR, pp 25–67
Likens GE, Bormann FH (1974a) Acid rain: a serious regional environmental problem. Science
184(4143):1176–1179
Likens GE, Bormann FH (1974b) Linkages between terrestrial and aquatic ecosystems.
Bioscience 24(8):447–456
Likens GE, Bormann FH (1985) An ecosystem approach. In: Likens GE (ed) An ecosystem approach
to aquatic ecology: Mirror Lake and its environment. Springer-Verlag, New York, pp 1–8
Likens GE, Bormann FH (1995) Biogeochemistry of a forested ecosystem, 2nd edn. Springer-
Verlag, New York, 159 pp
Likens GE, Buso DC (2006) Variation in streamwater chemistry throughout the Hubbard Brook
Valley. Biogeochemistry 78:1–30
Likens GE, Buso DC (2010a) Long-term changes in streamwater chemistry following disturbance
in the Hubbard Brook Experimental Forest, USA. Verh Int Verein Limnol 30(10):1577–1581
Likens GE, Buso DC (2010b) Salinization of Mirror Lake by road salt. Water Air Soil Pollut
205:205–214
Likens GE, Buso DC (2012) Dilution and the elusive baseline. Environ Sci Technol 46(8):
4382–4387. doi:10.1021/es3000189
Likens GE, Davis MB (1975) Post-glacial history of Mirror Lake and its watershed in
New Hampshire, U.S.A.: an initial report. Verh Int Verein Limnol 19:982–993
Likens GE, Fallon Lambert K (1998) The importance of long-term data in addressing regional
environmental issues. Northeast Nat 5(2):127–136
Likens GE, Franklin JF (2009) Ecosystem thinking in the Northern Forest – and beyond.
Bioscience 59(6):511–513
Likens GE, Moeller RE (1985) Chemistry. In: Likens GE (ed) An ecosystem approach to aquatic
ecology: Mirror Lake and its environment. Springer-Verlag, New York, pp 392–410
Likens GE, Bormann FH, Johnson NM, Pierce RS (1967) The calcium, magnesium, potassium
and sodium budgets for a small forested ecosystem. Ecology 48(5):772–785
Likens GE, Bormann FH, Johnson NM, Fisher DW, Pierce RS (1970) The effect of forest cutting
and herbicide treatment on nutrient budgets in the Hubbard Brook watershed-ecosystem. Ecol
Monogr 40(1):23–47
References

Aber JD, Ollinger SV, Driscoll CT (1997) Modeling nitrogen saturation in forest ecosystems in
response to land use and atmospheric deposition. Ecol Model 101:61–78
Aber JD, Ollinger SV, Driscoll CT, Likens GE, Holmes RT, Freuder RJ, Goodale CL (2002)
Inorganic nitrogen losses from a forested ecosystem in response to physical, chemical, biotic,
and climatic perturbations. Ecosystems 5(7):648–658
Adams MB, DeWalle DR, Horn JL (eds) (2006) The Fernow watershed acidification study.
Springer, 279 pp
Ågren GI, Andersson FO (2012) Terrestrial ecosystem ecology – principles and applications.
Cambridge University Press. ISBN 978-1-107-01107-6
Alewell C, Mitchell MJ, Likens GE, Krouse HR (1999) Sources of stream sulfate at the Hubbard
Brook Experimental Forest: long-term analyses using stable isotopes. Biogeochemistry
44:281–299
Alewell C, Mitchell MJ, Likens GE, Krouse HR (2000) Assessing the origin of sulfate- deposition
at the Hubbard Brook Experimental Forest. J Environ Qual 29(3):759–767
Anonymous (1990) The Source of the Brook (brochure). USDA Forest Service NE-INF-89-90
Arefýeva ZN, Kolesnikof BD (1964) Chemistry and biochemistry dynamics of ammonia and
nitrate nitrogen in forest soils of the Transurals at high and low temperatures. Sov Soil Sci
3:246–260
Art HW (1976) Ecological studies of the Sunken Forest Island national seashore, New York.
National Park Service. Scientific monograph series, No. 7
Art HW, Bormann FH, Voigt GK, Woodwell GM (1974) Barrier Island forest ecosystem: the role
of meteorological nutrient inputs. Science 184:60–62
Bailey SW (2000) Ecological setting of the northern forest – geologic and edaphic factors influenc-
ing susceptibilty to environmental change. In: Birdsey R, Hom J, Mickler R (eds) Responses of
Northern US forests to environmental change. Ecological studies series, vol 139. Springer-
Verlag, pp 27–49
Bailey SW, Hornbeck JW (1992) Lithologic composition and rock weathering potential for for-
ested, glacial-till soils. US Forest Service Research Paper NE-662
Bailey SW, Hornbeck JW, Driscoll CT, Gaudette HE (1996) Calcium inputs and transport in a
base-poor forest ecosystem as interpreted by Sr isotopes. Water Resour Res 32(3):707–719
Bailey AS, Hornbeck JW, Campbell JL, Eagar C (2003) Hydrometeorological database for
Hubbard Brook Experimental Forest: 1955–2000. USDA Forest Service, Northeastern
Research Station, General Technical Report NE-305, 36 pp
Bailey SW, Buso DC, Likens GE (2003) Implications of sodium mass balance for interpreting the
calcium cycle of a forested ecosystem. Ecology 84(2):471–484

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2, 183

© Springer Science+Business Media New York 2013
184 References

Bailey SW, Mayer B, Mitchell MJ (2004) Evidence for influence of mineral weathering on stream
water sulphate in Vermont and New Hampshire (USA). Hydrol Process 18:1639–1653
Bailey SW, Horsley SB, Long RP (2005) Thirty years of change in forest soils of the Allegheny
Plateau, Pennsylvania. Soil Sci Soc Am J 69:681–690
Barrett E, Brodin G (1955) The acidity of Scandinavian precipitation. Tellus 7:251–257
Barton CC, Camerlo RH, Bailey SW (1997) Bedrock geologic map of Hubbard Brook Experimental
Forest and maps of fractures and geology in roadcuts along Interstate-93, Grafton County, New
Hampshire. Sheet 1, Scale 1:12,000; Sheet 2, Scale 1:200. US Geological Survey, Miscellaneous
Investigations Series, Map I-2562
Beamish RJ (1976) Effects of precipitation on Canadian lakes. In: Dochinger LS, Seliga TA (eds)
Proceedings of the first international symposium on acid precipitation and the forest ecosys-
tem. USDA Forest Service General Technical Report NE-23, pp 479–498
Berger TW, Tartowski SL, Likens GE (1997) Trifluoroacetate retention in a northern hardwood
forest soil. Environ Sci Technol 31(7):1916–1921
Berger TW, Eagar C, Likens GE, Stingeder G (2001) Effects of calcium and aluminum chloride
additions on foliar and throughfall chemistry in sugar maples. For Ecol Manage 149:75–90
Bernal S, Hedin LO, Likens GE, Gerber S, Buso DC (2012) Complex response of the forest nitro-
gen cycle to climate change. Proc Natl Acad Sci 109(9):3406–3411. doi:10.1073/
pnas.1121448109
Bernhardt ES, Likens GE, Hall RO Jr, Buso DC, Fisher SG, Burton TM, Meyer JL, McDowell
WH, Mayer MS, Bowden WB, Findlay SEG, Macneale KH, Stelzer RS, Lowe WH (2005)
Can’t see the forest for the stream? In-stream processing and terrestrial nitrogen exports.
Bioscience 55(3):219–230
Bilby RE, Likens GE (1980) Importance of organic debris dams in the structure and function of
stream ecosystems. Ecology 61(5):1107–1113
Blum JD, Klaue A, Nezat CA, Driscoll CT, Johnson CE, Siccama TG, Eagar C, Fahey TJ, Likens
GE (2002) Mycorrhizal weathering of apatite as an important calcium source in base-poor for-
est ecosystems. Nature 417:729–731
Bolin B (ed) (1971) Report of the Swedish Preparatory Committee for the United Nations
Conference on human environment. Norstedt and Söner, Stockholm
Bormann FH (1974) Acid rain and the environmental future. Environ Cons, p 270
Bormann FH, Likens GE (1967) Nutrient cycling. Science 155:424–429
Bormann FH, Likens GE (1979) Pattern and process in a forested ecosystem. Springer-Verlag,
New York, NY, 253 pp
Bormann FH, Likens GE, Eaton JS (1969) Biotic regulation of particulate and solution losses from
a forested ecosystem. Bioscience 19(7):600–610
Bormann FH, Siccama TG, Likens GE, Whittaker RH (1970) The Hubbard Brook Ecosystem
Study: composition and dynamics of the tree stratum. Ecol Monogr 40(4):373–388
Bormann FH, Likens GE, Siccama TG, Pierce RS, Eaton JS (1974) The effect of deforestation on
ecosystem export and the steady-state condition at Hubbard Brook. Ecol Monogr
44(3):255–277
Bormann FH, Likens GE, Melillo J (1977) Nitrogen budget for an aggrading northern hardwood
forest ecosystem. Science 196(4293):981–983
Bowden WB, Bormann FH (1986) Transport and loss of nitrous oxide in soil water after forest
clear-cutting. Science 233:867–869
Bradley E, Cushman RV (1956) Memorandum report on geologic and ground-water conditions in
the Hubbard Brook watershed, New Hampshire. On file at Northeastern Forest Experiment
Station, Durham, NH, 15 pp
Brady NC, Weil R (2008) The nature and properties of soils, 14th edn. Prentice Hall
Braun EL (1950) Deciduous forests of Eastern North America. The Blakiston, Philadelphia, PA,
594 pp
Budyko MI (1974) Climate and life. Academic, New York
References 185

Burton TM, Likens GE (1975) Salamander populations and biomass in the Hubbard Brook
Experimental Forest, New Hampshire. Copeia 3:541–546
Buso DC, Likens GE, Eaton JS (2000) Chemistry of precipitation, streamwater and lakewater from
the Hubbard Brook Ecosystem Study: a record of sampling protocols and analytical proce-
dures. USDA Forest Service, Northeastern Research Station, Newtown Square, PA, General
Technical Report NE-275, 52 pp
Buso DC, Likens GE, LaBaugh JW, Bade D (2009) Nutrient dynamics. In: Winter TC, Likens GE
(eds) Mirror Lake: interactions among air, land and water. University of California Press, pp
69–203
Butler TJ, Cogbill CV, Likens GE (1984) Effect of climatology on precipitation chemistry. Bull
Am Meteorol Soc 65(6):639–640
Butler TJ, Likens GE, Stunder BJB (2001) Regional-scale impacts of Phase I of the Clean Air Act
Amendments in the USA: the relation between emissions and concentrations, both wet and dry.
Atmos Environ 35(6):1015–1028
Butler TJ, Likens GE, Vermeylen FM, Stunder BJB (2003) The relation between NO× emissions
and precipitation NO3− in the eastern USA. Atmos Environ 37(15):2093–2104
Butler TJ, Likens GE, Vermeylen FM, Stunder BJB (2005) The impact of changing nitrogen oxide
emissions on wet and dry nitrogen deposition in the northeastern USA. Atmos Environ
39:4851–4862
Campbell JL, Driscoll CT, Eagar C, Likens GE, Siccama TG, Johnson CE, Fahey TJ, Hamburg SP,
Holmes RT, Bailey AS, Buso DC (2007) Long-term trends from ecosystem research at the
Hubbard Brook Experimental Forest. US Department of Agriculture, Forest Service, Northern
Research Station, Newtown Square, PA, General Technical Report NRS-17, 41 pp
Campbell JL, Driscoll CT, Pourmokhtarian A, Hayhoe K (2011) Streamflow responses to past and
projected future changes in climate at the Hubbard Brook Experimental Forest, New Hampshire,
United States. Water Resour Res 47, W02514. doi:10.1029/2010WR009438
Carlisle A, Brown AHF, White EJ (1967) The nutrient content of tree stem flow and ground flora
litter and leachates in a sessile oak (Quercus petraea) woodland. J Ecol 55:615–627
Carroll D (1970) Rock weathering. Plenum, New York, NY, 203 pp
Charles DF (ed) (1991) Acidic deposition and aquatic ecosystems: regional case studies. Springer-
Verlag, New York
Christ MJ, Driscoll CT, Likens GE (1999) Watershed- and plot-scale tests of the mobile anion
concept. Biogeochemistry 47(3):335–353
Clair TA, Dennis IF, Vet R (2010) Water chemistry and dissolved organic carbon trends in lakes
from Canada’s Atlantic Provinces: no recovery from acidification measured after 25 years of
lake monitoring. Can J Fish Aquat Sci 68:663–674
Claridge GGC (1970) Studies in element balances in a small catchment at Taita, New Zealand. In:
Proceedings of the international association of scientific hydrology, UNESCO symposium, on
results of research on representative and experimental basins, Wellington, New Zealand,
December 1970, pp 523–540
Cogbill CV (1976) The effect of acid precipitation on tree growth in eastern North America. In:
Dochinger LS, Seliga TA (eds) Proceedings of the first international symposium on acid pre-
cipitation and the forest ecosystem. USDA Forest Service General Technical Report NE-23, pp
1027–1032
Cogbill CV, Likens GE (1974) Acid precipitation in the northeastern United States. Water Resour
Res 10:1133–1137
Cogbill CV, Likens GE, Butler TJ (1984) Uncertainties in historical aspects of acid precipitation:
getting it straight. Atmos Environ 18(10):2261–2270
Cole JJ (2013) The carbon cycle. In: Weathers KC, Strayer DL, Likens GE (eds) Fundamentals of
ecosystem science. Academic, New York, pp 109–135
Conley DJ, Likens GE, Buso DC, Saccone L, Bailey SW, Johnson CE (2008) Deforestation causes
increased dissolved silicate losses in the Hubbard Brook Experimental Forest. Glob Chang
Biol 14:2548–2554
186 References

Conrad V (1941) The variability of precipitation. Mon Weather Rev Wash 69:5
Covington WW (1975) Successional dynamics of organic matter and nutrient content of forest
floor and leaf fall in northern hardwoods. Bull Ecol Soc Am 56(2):12
Covington WW (1976) Secondary succession in northern hardwoods: forest floor organic matter
and nutrients and leaf fall. Ph.D. Thesis, Yale University, New Haven, CT, 117 pp
Cowling DW, Jones LHP, Lockyer DR (1973) Increased yield through correction of sulphur defi-
ciency in ryegrass exposed to sulphur dioxide. Nature 243(5407):479–480
Croat TB (1972) The role of overpopulation and agricultural methods in the destruction of tropical
ecosystems. Bioscience 22(8):465–467
Currie WS (2011) Units of nature or processes across scales? The ecosystem concept at age 75.
New Phytol 190:21–34
Davis MB (1985) History of the vegetation on the Mirror Lake watershed. In: Likens GE (ed) An
ecosystem approach to aquatic ecology: Mirror Lake and its environment. Springer-Verlag,
New York, pp 53–65
Davis MB, Ford J (1985) Late-glacial and Holocene sedimentation. In: Likens GE (ed) An ecosys-
tem approach to aquatic ecology: Mirror Lake and its environment. Springer-Verlag, New
York, pp 345–355
Davis MB, Moeller RE, Likens GE, Ford J, Sherman J, Goulden C (1985) Paleoecology of Mirror
Lake and its watershed. In: Likens GE (ed) An ecosystem approach to aquatic ecology: mirror
lake and its environment. Springer-Verlag, New York, pp 410–429
Davis RB, Anderson DS, Berge F (1985) Paleolimnological evidence that lake acidification is
accompanied by loss of organic matter. Nature 316:436–438
De Wit HA, Mulder J, Hindar A, Hole L (2007) Long-term increase in dissolved organic carbon in
streamwaters in Norway in response to reduced acid precipitation. Environ Sci Technol
41:7706–7713
Detty JM, McGuire KJ (2010a) Threshold changes in storm runoff generation at a till-mantled
headwater catchment. Water Resour Res 46, W07525. doi:10.1029/2009WR008102
Detty JM, McGuire KJ (2010b) Topographic controls on shallow groundwater dynamics: implica-
tions of hydrologic connectivity between hillslopes and riparian zones in a till mantled catch-
ment. Hydrol Process 24:2222–2236
Dittman JA, Driscoll CT, Groffman PM, Fahey TJ (2007) Dynamics of nitrogen and dissolved
organic carbon at the Hubbard Brook Experimental Forest. Ecology 88(5):1153–1166
Dominski AS (1971) Accelerated nitrate production and loss in the northern hardwood forest eco-
system underlain by podzol soils following clear cutting and addition to herbicides. Ph.D.
Thesis, Yale University, 157 pp
Douglas LA (1989) Vermiculites. In: Dixon JB, Weed SB (eds) Minerals in soil environments, 2nd
edn. Soil Science Society of America, Madison, WI, pp 635–674
Douglass JE, Swank WT (1972) Streamflow modification through management of Eastern forests.
USDA Forest Service Research Paper SE-94. Southeast Forest Experiment Station, Washington,
DC. 15 pp
Driscoll CT, Likens GE (1982) Hydrogen ion budget of an aggrading forested ecosystem. Tellus
34:283–292
Driscoll CT, Baker JP, Bisogni JJ, Schofield CL (1980) Effect of aluminum speciation on fish in
dilute acidified waters. Nature 284:161–164
Driscoll CT, van Breemen N, Mulder J (1985) Aluminum chemistry in a forested spodosol. Soil
Sci Soc Am J 49(2):437–444
Driscoll CT, Likens GE, Hedin LO, Eaton JS, Bormann FH (1989a) Changes in the chemistry of
surface waters: 25-year results at the Hubbard Brook Experimental Forest, NH. Environ Sci
Technol 23(2):137–143
Driscoll CT, Fuller RD, Schecher WD (1989b) The role of organic acids in the acidification of
surface waters in the eastern US. Water Air Soil Pollut 43:21–40
Driscoll CT, Likens GE, Church MR (1998) Recovery of surface waters in the northeastern U.S.
from decreases in atmospheric deposition of sulfur. Water Air Soil Pollut 105:319–329
References 187

Driscoll CT, Lawrence GB, Bulger AJ, Butler TJ, Cronan CS, Eagar C, Fallon Lambert K, Likens
GE, Stoddard JL, Weathers KC (2001) Acidic deposition in the northeastern United States:
sources and inputs, ecosystem effects, and management strategies. Bioscience 51(3):180–198
Driscoll CT, Han Y-J, Chen C, Evers DC, Fallon Lambert K, Holsen TM, Kamman NC, Munson
RK (2007) Mercury contamination in forest and freshwater ecosystems in the northeastern
United States. BioScience 57(1):17–28
Duvigneaud P, Denaeyer-DeSmet S (1964) Le cycle des éléments biogènes dans l'écosystème forêt
(Forêts tempèrées caducifoliées). Lejeunia 28:1–148
Eaton JS, Likens GE, Bormann FH (1973) Throughfall and stemflow chemistry in a northern hard-
wood forest. J Ecol 61:495–508
Eaton JS, Likens GE, Bormann FH (1976) The biogeochemistry of sulfur in a northeastern forest
ecosystem. In: The tenth Middle Atlantic regional meeting abstracts, American Chemical
Society, 23–26 February, Philadelphia, PA, p 43
Eriksson E (1952) Composition of atmospheric precipitation. II. Sulfur, chloride, iodine com-
pounds. Bibliography. Tellus 4:280–303
Erlandsson M, Cory N, Fölster J, Köhler S, Laudon JM, Weyhenmeyer GA, Bishop K (2011)
Increasing dissolved organic carbon redefines the extent of surface water acidification and
helps resolve a classic controversy. Bioscience 61(8):614–618
Evans CD, Chapman PJ, Clark JM, Monteith DT, Cresser MS (2006) Alternative explanations for
rising dissolved organic carbon export from organic soils. Glob Chang Biol
12(11):2044–2053
Evans CD, Jones TG, Burden A, Ostle N, Zielinski P, Cooper MDA, Peacock M, Clark JM,
Oulehle F, Cooper D, Freeman C (2012) Acidity controls on dissolved organic carbon mobility
in organic soils. Glob Chang Biol 18:3317–3331
Fahey TJ, Siccama TG, Driscoll CT, Likens GE, Campbell J, Johnson CE, Aber JD, Cole JJ, Fisk
MC, Groffman PM, Holmes RT, Schwarz PA, Yanai RD (2005) The biogeochemistry of carbon
at Hubbard Brook. Biogeochemistry 75(1):109–176
Federer CA, Lash D (1978) BROOK: a hydrologic simulation model for eastern forests. Water
Resources Research Center, University of New Hampshire, Durham. Research Report No. 19,
84 pp
Federer CA, Hornbeck JW, Tritton LM, Martin CW, Pierce RS, Smith CT (1989) Long-term deple-
tion of calcium and other nutrients in eastern U.S. forests. Environ Manag 13(5):593–601
Federer CA, Flynn LD, Martin CW, Hornbeck JW, Pierce RS (1990) Thirty years of hydrometeo-
rologic data at the Hubbard Brook Experimental Forest, New Hampshire. USDA Forest Service
General Technical Report NE-141, 44 pp
Fernandez IJ, Son Y, Kraske CR, Rustad LE, David MB (1993) Soil carbon dioxide characteristics
under different forest types and after harvest. Soil Sci Soc Am J 57(4):1115–1121
Fernandez IJ, Rustad LE, Norton SA, Kahl JS, Cosby BJ (2003) Experimental acidification causes
soil base-cation depletion at The Bear Brook Watershed in Maine. Soil Sci Soc Am J
67:1909–1919
Fernandez IJ, Adams MB, SanClements MD, Norton SA (2010) Comparing decadal responses of
whole-watershed manipulations at the Bear Brook and Fernow experiments. Environ Monit
Assess 171:149–161
Findlay SEG (2005) Increased carbon transport in the Hudson River: unexpected consequence of
nitrogen deposition? Front Ecol Environ 3:133–137
Findlay SEG (2013) Organic matter decomposition. In: Weathers KC, Strayer DL, Likens GE
(eds) Fundamentals of ecosystem science. Academic, New York, pp 75–94
Fisher SG (1970) Annual energy budget of a small forest stream ecosystem: Bear Brook, West
Thornton, New Hampshire. Ph.D. Thesis, Dartmouth College, Hanover, NH, 97 pp
Fisher DW, Gambell AW, Likens GE, Bormann FH (1968) Atmospheric contributions to water
quality of streams in the Hubbard Brook Experimental Forest, New Hampshire. Water Resour
Res 4(5):1115–1126
188 References

Fitzhugh RD, Driscoll CT, Groffman PM, Tierney GL, Fahey TJ, Hardy JP (2001) Effects of soil
freezing disturbance on soil solution nitrogen, phosphorus, and carbon chemistry in a northern
hardwood ecosystem. Biogeochemistry 56(2):215–238
Fitzhugh RD, Likens GE, Driscoll CT, Mitchell MJ, Groffman PM, Fahey TJ, Hardy JP (2003)
Role of soil freezing events in interannual patterns of stream chemistry at the Hubbard Brook
Experimental Forest, New Hampshire. Environ Sci Technol 37:1575–1580
Fitzsimmons AK (1996) Sound policy or smoke and mirrors: does ecosystem management make
sense? Water Resour Bull 32(2):217–227
Flaccus E (1958a) Landslides and their revegetation in the White Mountains of New Hampshire.
Ph.D. dissertation, Duke University, Durham, NC. 186 pp
Flaccus E (1958b) White Mountain landslides. Appalachia 32:175–191
Fowler-Billings K, Page LR (1942) The geology of the Cardigan and Rumney quadrangles, New
Hampshire. New Hampshire Planning and Development Commission, Concord, NH, 31 pp
Franklin JF, Lindenmayer DB, MacMahon JA, McKee A, Magnuson J, Perry DA, Waide R, Foster
DR (2000) Threads of continuity: ecosystem disturbances, biological legacies and ecosystem
recovery. Conserv Biol Pract 1(1):8–16
Fuss CB, Driscoll CT, Johnson CE, Petras RJ, Fahey TJ (2011) Dynamics of oxidized and reduced
iron in a northern hardwood forest. Biogeochemistry 104:103–119
Galloway JN, Likens GE, Edgerton ES (1976) Acid precipitation in the northeastern United States:
pH and acidity. Science 194:722–724
Gambell AW, Fisher DW (1966) Chemical composition of rainfall of eastern North Carolina and
southeastern Virginia. United States Geological Survey Water Supply Paper 1535-K,
Washington, DC, pp K1–K41
Gbondo-Tugbawa SS, Driscoll CT, Aber JD, Likens GE (2001) Evaluation of an integrated bio-
geochemical model (PnET-BGC) at a northern hardwood forest ecosystem. Water Resour Res
37(4):1057–1070
Gbondo-Tugbawa SS, Driscoll CT, Mitchell MJ, Aber JD, Likens GE (2002) A model to simulate
the response of a northern hardwood forest ecosystem to changes in S deposition. Ecol Appl
12(1):8–23
Goldich SS (1938) A study of rock weathering. J Geol 46:17–58
Gómez-Pompa A, Vázquez-Yanes C, Guevara S (1972) The tropical rainforest: a nonrenewable
resource. Science 177:762–765
Gorham E (1958) Atmospheric pollution by hydrochloric acid. Quart J Roy Meteorol Soc
84(361):274–276
Gosz JR, Likens GE, Bormann FH (1972) Nutrient content of litter fall on the Hubbard Brook
Experimental Forest, New Hampshire. Ecology 53(5):769–784
Gosz JR, Likens GE, Bormann FH (1973) Nutrient release from decomposing leaf and branch lit-
ter in the Hubbard Brook Forest, New Hampshire. Ecol Monogr 43(2):173–191
Gosz JR, Likens GE, Bormann FH (1976) Organic matter and nutrient dynamics of the forest and
forest floor in the Hubbard Brook Forest. Oecologia (Berlin) 22:305–320
Graedel TE, Keene WC (1996) The budget and cycle of Earth’s natural chlorine. Pure Appl Chem
68:1689–1697
Granat L (1972) On the relation between pH and the chemical composition in atmospheric precipi-
tation. Tellus 24:550–560
Groffman PM, Rosi-Marshall E (2013) The nitrogen cycle. In: Weathers KC, Strayer DL, Likens
GE (eds) Fundamentals of ecosystem science. Academic, New York, pp 137–158
Groffman PM, Hardy JP, Nolan S, Fitzhugh RD, Driscoll CT, Fahey TJ (1999) Snow depth, soil
frost and nutrient loss in a northern hardwood forest. Hydrol Process 13:2275–2286
Groffman PM, Driscoll CT, Fahey TJ, Hardy JP, Fitzhugh RD, Tierney GL (2001) Colder soils in
a warmer world: a snow manipulation study in a northern hardwood forest ecosystem.
Biogeochemistry 56(2):135–150
Groffman PM, Fisk MC, Driscoll CT, Likens GE, Fahey TJ, Eagar C, Pardo LH (2006) Calcium
additions and microbial nitrogen cycle processes in a northern hardwood forest. Ecosystems
9:1289–1305
References 189

Groffman PM, Hardy JP, Fisk MC, Fahey TJ, Driscoll CT (2009) Climate variation and soil carbon
and nitrogen cycling processes in a northern hardwood forest. Ecosystems 12(6):927–943
Groffman PM, Rustad LE, Templer PH, Campbell JL, Christenson LM, Lany NK, Socci AM,
Vadeboncoeur MA, Schaberg PG, Wilson GF, Driscoll CT, Fahey TJ, Fisk MC, Goodale CL,
Green MB, Hamburg SP, Johnson CE, Mitchell MJ, Morse JL, Pardo LH, Rodenhouse NL
(2012) Long-term integrated studies show complex and surprising effects of climate change in
the northern hardwood forest. Bioscience 62(12):1056–1066
Hack JT, Goodlett JC (1960) Geomorphology and forest ecology of a mountain region in the
Central Appalachians. United States Geological Survey Professional Paper 347, Washington,
DC, 66 pp
Hall RJ, Likens GE (1980) Ecological effects of experimental acidification on a stream ecosystem.
In: Drabløs D, Tollan A (eds) Ecological impact of acid precipitation. Proceedings of the inter-
national conference, Sandefjord, Norway, March 1980, pp 375–376
Hall RO Jr, Macneale KH, Bernhardt ES, Field M, Likens GE (2001) Biogeochemical responses
of two forest streams to a two-month calcium addition. Freshw Biol 46(3):291–302
Hallett RA, Bailey SW, Horsley SB, Long RP (2006) Influence of nutrition and stress on sugar
maple at a regional scale. Can J Forest Res 36:2235–2246
Halman JM, Schaberg PG, Hawley GJ, Eagar C (2008) Calcium addition at the Hubbard Brook
Experimental Forest increases sugar storage, antioxidant activity and cold tolerance in native
red spruce (Picea rubens). Tree Physiol 28(6):855–862
Hamburg SP, Cogbill CV (1988) Historical decline of red spruce populations and climatic warm-
ing. Nature 331(6155):428–431
Hamburg SP, Vadeboncoeur MA, Richardson AD, Bailey AS (2013) Climate change at the ecosys-
tem scale: a 50-year record in New Hampshire. Clim Change 116:457–477. doi:10.1007/
s10584-012-0517-2
Hamon WR (1961) Estimating potential evapotranspiration. Am Soc Civ Eng Proc
87(HY3):107–120
Hart GE Jr (1966) Streamflow characteristics of small forested watersheds in the White Mountains
of New Hampshire. Ph.D. Thesis, University of Michigan, Ann Arbor, MI, 141 pp
Hart GE Jr, Leonard RE, Pierce RS (1962) Leaf fall, humus depth, and soil frost in a northern
hardwood forest. Research Note No. 131, Northeastern Forest Experiment Station, Upper
Darby, PA, 3 pp
Hawley GJ, Schaberg PG, Eagar C, Borer CH (2006) Calcium addition at the Hubbard Brook
Experimental Forest reduced winter injury to red spruce in a high-injury year. Can J For Res
36(10):2544–2549
Hayhoe K, Wake CP, Huntington TG, Luo L, Schwartz MD, Sheffield J, Wood E, Anderson B,
Bradbury J, DeGaetano A, Troy T, Wolfe D (2007) Past and future changes in climate and
hydrological indicators in the U.S. Northeast. Clim Dyn 28:381–407
Hedin LO, Granat L, Likens GE, Rodhe H (1990) Strong similarities in seasonal concentration
ratios of SO42-, NO3- and NH4+ in precipitation between Sweden and the northeastern US. Tellus
42B:454–462
Hedin LO, Granat L, Likens GE, Buishand TA, Galloway JN, Butler TJ, Rodhe H (1994) Steep
declines in atmospheric base cations in regions of Europe and North America. Nature
367:351–354
Herman FA, Gorham E (1957) Total mineral material, acidity, sulfur, and nitrogen in rain and snow
at Kentville, Nova Scotia. Tellus 9:180–183
Hill AC (1971) Vegetation: a sink for atmospheric pollutants. J Air Poll Contr Assoc 21:341–346
Hobbie JE, Likens GE (1973) The output of phosphorus dissolved organic carbon and fine particulate
carbon from Hubbard Brook watersheds. Limmol Oceanogr 18(5):734–742
Hoeft RG, Keeney DR, Walsh LM (1972) Nitrogen and sulfur in precipitation and sulfur dioxide
in the atmosphere in Wisconsin. J Environ Qual 1:203–208
190 References

Holmes RT (2011) Avian population and community processes in forest ecosystems: long-term
research in the Hubbard Brook Experimental Forest. For Ecol Manage 262:20–32
Holmes RT, Likens GE (1999) Organisms of Hubbard Brook Valley, New Hampshire. USDA
Forest Service, Northeastern Research Station, General Technical Report NE-257, 32 pp
Holmes RT, Sturges FW (1973) Annual energy expenditure by the avifauna of a northern hard-
woods ecosystem. Oikos 24:24–29
Hooper RS (1986) The chemical response of an acid-sensitive headwater stream to snowmelt and
storm events: a field study and simulation model. Ph.D. Thesis, Cornell University, Ithaca, NY,
279 pp
Hooper RP, Shoemaker C (1985) Aluminum mobilization in an acidic headwater stream: temporal
variation and mineral dissolution disequilibria. Science 229:464–465
Hornbeck JW (1973) Storm flow from hardwood-forested and cleared watersheds in New
Hampshire. Water Resour Res 9(2):346–354
Hornbeck JW (2001) Events leading to establishment of the Hubbard Brook Experimental Forest.
http://www.hubbardbrook.org/overview/HBEF_establishment.html
Hornbeck JW, Likens GE (1974) The ecosystem concept for determining the importance of chemi-
cal composition of snow. In: Advanced concepts and techniques in the study of snow and ice
resources. National Academy of Sciences, Washington, DC, pp 139–151; Also, 31st annual
eastern snow conference proceedings, pp 145–155
Hornbeck JW, Pierce RS, Federer CA (1970) Streamflow changes after forest clearing in New
England. Water Resour Res 6(4):1124–1132
Hornbeck JW, Likens GE, Eaton JS (1976) Seasonal variation in acidity of precipitation and the
implications for forest-stream ecosystems. In: Dochinger LS, Seliga TA (eds) Proceedings of
the first international symposium on acid precipitation and the forest ecosystem. USDA Forest
Service General Technical Report NE-23, pp 597–609
Hornbeck JW, Martin CW, Pierce RS, Bormann FH, Likens GE, Eaton JS (1986) Clearcutting
northern hardwoods: effects on hydrologic and nutrient ion budgets. For Sci 32(3):667–686
Hornbeck JW, Martin CW, Pierce RS, Bormann FH, Likens GE, Eaton JS (1987) The northern
hardwood forest ecosystem: ten years of recovery from clearcutting. USDA Forest Service,
Northeastern Forest Experiment Station, NE-RP-596, 30 pp
Hornbeck JW, Smith CT, Martin CW, Tritton LM, Pierce RS (1990) Effects of intensive harvesting
on nutrient capitals of three forest types in New England. For Ecol Manage 30:55–64
Horsley SB, Long RP, Bailey SW, Hallett RA, Wargo PM (2002) Health of eastern North American
sugar maple forests and factors affecting decline. North J Appl For 19:34–44
Houle D, Lamoureux P, Bélanger N, Bouchard M, Gagnon C, Couture S, Bouffard A (2012) Soil
weathering rates in 21 catchments of the Canadian Shield. Hydrol Earth Syst Sci 16:685–697
Houlton BZ, Driscoll CT, Fahey TJ, Likens GE, Groffman PM, Bernhardt ES, Buso DC (2003)
Nitrogen dynamics in ice storm-damaged forest ecosystems: implications for nitrogen limita-
tion theory. Ecosystems 6(5):431–443
Hunt CB (1967) Physiography of the United States. Freeman, San Francisco, CA
Huntington TG, Ryan DF, Hamburg SP (1988) Estimating soil nitrogen and carbon pools in a
northern hardwood forest ecosystem. Soil Sci Soc Am J 52:1162–1167
Huntington TG, Johnson CE, Johnson AH, Siccama TG, Ryan DF (1989) Carbon, organic matter
and bulk density relationships in a forested spodosol. Soil Sci 148(5):380–386
Huntington TG, Richardson AD, McGuire KJ, Hayhoe K (2009) Climate and hydrological changes
in the northeastern United States: recent trends and implications for forested and aquatic eco-
systems. Can J Forest Res 39:199–212
Hutchinson GE (1957) A treatise on limnology, vol 1. Wiley, New York, NY, 1015 pp
Hyman ME, Johnson CE, Bailey SW, April RH, Hornbeck JW (1998) Chemical weathering and
cation loss in a base poor watershed. Geol Soc Am Bull 110:85–95
Jeziorski A, Yan ND, Paterson AM, DeSellas AM, Turner MA, Jeffries DS, Keller B, Weeber RC,
McNicol DK, Palmer ME, McIver K, Arseneau K, Ginn BK, Cumming BF, Smol JP (2008)
The widespread threat of calcium decline in fresh waters. Science 322:1374–1377
References 191

Johannessen M, Dale T, Gjessing ET, Henriksen A, Wright RF (1976) Proceedings of the

international symposium on isotopes and impurities in snow and ice. International association
of hydrological science, Grenoble, France, 28–30 August 1975. International Association of
Hydrological Science Publication, 118
Johnson CE (1989) The chemical and physical properties of a northern hardwood forest soil: har-
vesting effects, soil-tree relations and sample size determination. Ph.D. Thesis, University of
Pennsylvania, Philadelphia, PA, 221 pp
Johnson PL, Swank WT (1973) Studies of cation budgets in the southern Appalachians on four
experimental watersheds with contrasting vegetation. Ecology 54(1):70–80
Johnson NM, Likens GE, Bormann FH, Pierce RS (1968) Rate of chemical weathering of silicate
minerals in New Hampshire. Geochim Cosmochim Acta 32:531–545
Johnson NM, Likens GE, Bormann FH, Fisher DW, Pierce RS (1969) A working model for the
variation in streamwater chemistry at the Hubbard Brook Experimental Forest, New Hampshire.
Water Resour Res 5(6):1353–1363
Johnson NM, Reynolds RC, Likens GE (1972) Atmospheric sulfur: its effect on the chemical
weathering of New England. Science 177:514–516
Johnson NM, Driscoll CT, Eaton JS, Likens GE, McDowell WH (1981) “Acid rain”, dissolved
aluminum and chemical weathering at the Hubbard Brook Experimental Forest, New
Hampshire. Geochim Cosmochim Acta 45(9):1421–1437
Johnson CE, Johnson AH, Huntington TG, Siccama TG (1991a) Whole-tree clear-cutting effects
on soil horizons and organic-matter pools. Soil Sci Soc Am J 55:497–502
Johnson CE, Johnson AH, Siccama TG (1991b) Whole-tree clear-cutting effects on exchangeable
cations and soil acidity. Soil Sci Soc Am J 55:502–508
Johnson CE, Litaor MI, Bittlett MF, Bricker OP (1994) Chemical weathering in small catchments:
climatic and anthropogenic influences. In: Moldan B, Cerny J (eds) Biogeochemistry of small
catchments: a tool for environmental research. SCOPE 51. Wiley, Chichester, pp 323–341
Johnson CE, Siccama TG, Driscoll CT, Likens GE, Moeller RE (1995) Changes in lead biogeo-
chemistry in response to decreasing atmospheric inputs. Ecol Appl 5(3):813–822
Juang FHT, Johnson NM (1967) Cycling of chlorine through a forested catchment in New England.
J Geophys Res 72:5641–5647
Judd KE, Likens GE, Groffman PM (2007) High nitrate retention during winter in soils of the
Hubbard Brook Experimental Forest. Ecosystems 10:217–225
Judd KE, Likens GE, Buso DC, Bailey AS (2011) Minimal response in watershed nitrate export to
severe soil frost raises questions about nutrient dynamics in the Hubbard Brook Experimental
Forest. Biogeochemistry 106:443–459. doi:10.1007/s10533-010-9524-4
Juice SM, Fahey TJ, Siccama TG, Driscoll CT, Denny EG, Eagar C, Cleavitt NL, Minocha R,
Richardson AD (2006) Response of sugar maple to calcium addition to northern hardwood
forest. Ecology 87(5):1267–1280
Junge CE (1958) The distribution of ammonia and nitrate in rain water over the United States.
Trans Am Geophys Union 39:241–248
Junge CE (1963) Air chemistry and radioactivity. Academic, New York, NY, 382 pp
Junge CE, Werby RT (1958) The concentration of chloride, sodium, potassium, calcium and sul-
fate in rain water over the United States. J Meteorol 15:417–425
Kaste JM, Heimsath AM, Bostick BC (2007) Short-term soil mixing quantified with fallout radio-
nuclides. Geology 35(3):243–246
Kerr JG, Eimers MC, Creed IF, Adams MB, Beall F, Burns D, Campbell JL, Christopher SF, Clair
TA, Courchesne F, Duchesne L, Fernandez I, Houle D, Jeffries DS, Likens GE, Mitchell MJ,
Shanley J, Yao H (2012) The effect of seasonal drying on sulphate dynamics in streams across
southeastern Canada and the northeastern USA. Biogeochemistry 111:393–409. doi:10.1007/
s10533-011-9664-1
Kobe RK, Likens GE, Eagar C (2002) Tree seedling growth and mortality responses to manipula-
tions of calcium and aluminum in a northern hardwood forest. Can J Forest Res 32:954–966
Korosi JB, Burke SM, Thienpont JR, Smol JP (2012) Anomolous rise in algal production linked to
lakewater calcium decline through food web interactions. Proceedings of the Royal Society B
279:1210–1217. doi:10.1098/rspb.2011.1411
192 References

Krug EC, Frink CR (1983) Acid rain on acid soil: a new perspective. Science 221:520–521
Küchler AW (1964) Potential natural vegetation of the conterminous United States. American
Geographical Society Special Publication No. 36. New York, NY, 116 pp
Kulkarni MV, Groffman PM, Yavitt JB (2008) Solving the global nitrogen problem: it’s a gas!
Front Ecol Environ 6:199–206
Lawrence GB, Driscoll CT (1990) Longitudinal patterns of concentration-discharge relationships
in streamwater draining the Hubbard Brook Experimental Forest, New Hampshire. J Hydrol
116:147–165
Lawrence GB, Fuller RD, Driscoll CT (1986) Spatial relationships of aluminum chemistry in the
streams of the Hubbard Brook Experimental Forest, New Hampshire. Biogeochemistry
2(2):115–135
LeBauer DS, Treseder KK (2008) Nitrogen limitation of net primary productivity in terrestrial
ecosystems is globally distributed. Ecology 89(2):371–379
Leonard RE (1961) Interception of precipitation by northern Hardwoods. Station Paper No. 159,
Northeastern Forest Experiment Station, USDA Forest Service, Upper Darby, PA, 16 pp
Leopold A (1949) A Sand County Almanac. Oxford University Press, Oxford, 226 pp
Likens GE (1972a) The chemistry of precipitation in the Central Finger Lakes Region. Technical
Publication No. 50, Cornell University Water Resources and Marine Sciences Center, Ithaca,
New York, 47 pp and 14 Figs
Likens GE (1972b) Mirror Lake: its past, present and future? Appalachia 39(2):23–41
Likens GE (1973) A checklist of organisms for the Hubbard Brook ecosystems. Section of Ecology
and Systematics, Cornell University, Ithaca, NY (Mimeo), 54 pp
Likens GE (1974a) The runoff of water and nutrients from watersheds tributary to Cayuga Lake,
New York. Water Resources and Marine Sciences Center Publication 81. Cornell University,
Ithaca, New York, 124 pp
Likens GE (1974b) Water and nutrient budgets for Cayuga Lake, New York. Water Resources and
Marine Sciences Center Publication 82. Cornell University, Ithaca, New York, 91 pp
Likens GE (1975) Nutrient flux and cycling in freshwater ecosystems. In: Howell FG, Gentry JB,
Smith MH (eds) Mineral cycling in southeastern ecosystems. ERDA symposium series CONF-
740513, Augusta, Georgia, May 1974, pp 314–348
Likens GE (ed) (1981) Some perspectives of the major biogeochemical cycles, vol 17. SCOPE
IVth general assembly, Stockholm, Sweden. Wiley, Chichester, 170 pp
Likens GE (1984) Acid rain: the smokestack is the “smoking gun”. Garden 8(4):12–18
Likens GE (ed) (1985) An ecosystem approach to aquatic ecology: Mirror Lake and its environ-
ment. Springer-Verlag, New York, 516 pp
Likens GE (1989) Some aspects of air pollution effects on terrestrial ecosystems and prospects for
the future. Ambio 18(3):172–178
Likens GE (1992) The ecosystem approach: its use and abuse. Excellence in ecology, book 3. The
Ecology Institute, Oldendorf-Luhe, Germany, 167 pp
Likens GE (1998) Limitations to intellectual progress in ecosystem science. In: Pace ML, Groffman
PM (eds) Successes, limitations and frontiers in ecosystem science. 7th Cary Conference,
Institute of Ecosystem Studies, Millbrook, New York. Springer-Verlag, New York,
pp 247–271
Likens GE (2001) Ecosystems: energetics and biogeochemistry. In: Kress WJ, Barrett G (eds) A
new century of biology. Smithsonian Institution Press, Washington, pp 53–88
Likens GE (2004) Some perspectives on long-term biogeochemical research from the Hubbard
Brook Ecosystem Study. Ecology 85(9):2355–2362
Likens GE (2010) The role of science in decision making: does evidence-based science drive envi-
ronmental policy? Front Ecol Environ 8(6):e1–e8. doi:10.1890/090132
Likens GE (2011) Limnological measures related to climate change in the Hubbard Brook Valley,
USA. Inland Waters 1(2):93–99
Likens GE, Bailey SW (2013) The discovery of acid rain at the Hubbard Brook Experimental
Forest – a story of collaboration and long-term research. In: Experimental forest and range
(in press)
References 193

Likens GE, Bormann FH (1972) Nutrient cycling in ecosystems. In: Wiens J (ed) Ecosystem
structure and function. Oregon State University Press, Corvallis, OR, pp 25–67
Likens GE, Bormann FH (1974a) Acid rain: a serious regional environmental problem. Science
184(4143):1176–1179
Likens GE, Bormann FH (1974b) Linkages between terrestrial and aquatic ecosystems. Bioscience
24(8):447–456
Likens GE, Buso DC (2006) Variation in streamwater chemistry throughout the Hubbard Brook
Valley. Biogeochemistry 78:1–30
Likens GE, Buso DC (2010a) Long-term changes in streamwater chemistry following disturbance
in the Hubbard Brook Experimental Forest, USA. Verh Int Verein Limnol 30(10):1577–1581
Likens GE, Buso DC (2010b) Salinization of Mirror Lake by road salt. Water Air Soil Pollut
205:205–214
Likens GE, Buso DC (2012) Dilution and the elusive baseline. Environ Sci Technol 46(8):
4382–4387. doi:10.1021/es3000189
Likens GE, Davis MB (1975) Post-glacial history of Mirror Lake and its watershed in
New Hampshire, U.S.A.: an initial report. Verh Int Verein Limnol 19:982–993
Likens GE, Bormann FH (1970) Chemical analyses of plant tissues from the Hubbard Brook
Ecosystem in New Hampshire. Bulletin 79, Yale University School of Forestry, New Haven,
CT, 25 pp
Likens GE, Bormann FH (1985) An ecosystem approach. In: Likens GE (ed) An ecosystem approach
to aquatic ecology: Mirror Lake and its environment. Springer-Verlag, New York, pp 1–8
Likens GE, Bormann FH (1995) Biogeochemistry of a forested ecosystem, 2nd edn. Springer-
Verlag, New York, 159 pp
Likens GE, Fallon Lambert K (1998) The importance of long-term data in addressing regional
environmental issues. Northeast Nat 5(2):127–136
Likens GE, Franklin JF (2009) Ecosystem thinking in the Northern Forest – and beyond. Bioscience
59(6):511–513
Likens GE, Moeller RE (1985) Chemistry. In: Likens GE (ed) An ecosystem approach to aquatic
ecology: Mirror Lake and its environment. Springer-Verlag, New York, pp 392–410
Likens GE, Bormann FH, Johnson NM, Pierce RS (1967) The calcium, magnesium, potassium and
sodium budgets for a small forested ecosystem. Ecology 48(5):772–785
Likens GE, Bormann FH, Johnson NM, Fisher DW, Pierce RS (1970) The effect of forest cutting
and herbicide treatment on nutrient budgets in the Hubbard Brook watershed-ecosystem. Ecol
Monogr 40(1):23–47
Likens GE, Bormann FH, Johnson NM (1972) Acid rain. Environment 14(2):33–40
Likens GE, Johnson NM, Galloway JN, Bormann FH (1976) Acid precipitation: strong and weak
acids. Science 194(4265):643–645
Likens GE, Bormann FH, Pierce RS, Eaton JS, Johnson NM (1977) Biogeochemistry of a forested
ecosystem. Springer-Verlag, New York, 146 pp
Likens GE, Wright RF, Galloway JN, Butler TJ (1979) Acid rain. Sci Am 241(4):43–51
Likens GE, Edgerton ES, Galloway JN (1983) The composition and deposition of organic carbon
in precipitation. Tellus 35B:16–24
Likens GE, Bormann FH, Pierce RS, Eaton JS (1985) The Hubbard Brook valley. In: Likens GE
(ed) An ecosystem approach to aquatic ecology: Mirror Lake and its environment. Springer-
Verlag, New York, pp 9–39
Likens GE, Bormann FH, Hedin LO, Driscoll CT, Eaton JS (1990a) Dry deposition of sulfur:
a 23-yr record for the Hubbard Brook Forest ecosystem. Tellus 42B:319–329
Likens GE, Hedin LO, Butler TJ (1990b) Some long-term precipitation chemistry patterns at the
Hubbard Brook Experimental Forest: extremes and averages. Verh Int Verein Limnol
24(1):128–135
Likens GE, Driscoll CT, Buso DC, Siccama TG, Johnson CE, Lovett GM, Ryan DF, Fahey T,
Reiners WA (1994) The biogeochemistry of potassium at Hubbard Brook. Biogeochemistry
25:61–125
194 References

Likens GE, Driscoll CT, Buso DC (1996) Long-term effects of acid rain: response and recovery of
a forest ecosystem. Science 272:244–246
Likens GE, Driscoll CT, Buso DC, Siccama TG, Johnson CE, Lovett GM, Fahey TJ, Reiners WA,
Ryan DF, Martin CW, Bailey SW (1998) The biogeochemistry of calcium at Hubbard Brook.
Biogeochemistry 41(2):89–173
Likens GE, Butler TJ, Buso DC (2001) Long- and short-term changes in sulfate deposition: effects
of The 1990 Clean Air Act Amendments. Biogeochemistry 52(1):1–11
Likens GE, Driscoll CT, Buso DC, Mitchell MJ, Lovett GM, Bailey SW, Siccama TG, Reiners
WA, Alewell C (2002a) The biogeochemistry of sulfur at Hubbard Brook. Biogeochemistry
60(3):235–316
Likens GE, Buso DC, Hornbeck JW (2002b) Variation in chemistry of stream water and bulk
deposition across the Hubbard Brook Valley, New Hampshire, USA. Verh Int Verein Limnol
28(1):402–409
Likens GE, Buso DC, Dresser BK, Bernhardt ES, Hall RO Jr, Macneale KH, Bailey SW (2004a)
Buffering an acidic stream in New Hampshire with a silicate mineral. Restor Ecol
12(3):419–428
Likens GE, Dresser BK, Buso DC (2004b) Short-term, temperature response in forest floor and
soil to ice storm disturbance in a northern hardwood forest. North J Appl For 21(4):209–219
Likens GE, Buso DC, Butler TJ (2005) Long-term relationships between SO2 and NOX emissions
and SO42- and NO3- concentration in bulk deposition at the Hubbard Brook Experimental
Forest, New Hampshire. J Environ Monit 7(10):964–968
Likens GE, Butler T, Rury M (2012) Acid rain. In: Anheier H, Juergensmeyer M (eds) Encyclopedia
of global studies. SAGE, Thousand Oaks, CA, pp 18–20. doi:10.4135/9781452218557.n6
Lindenmayer D, Likens GE (2009) Adaptive monitoring: a new paradigm for long-term research
and monitoring. Trends Ecol Evol (TREE) 24(9):482–486
Lindenmayer DB, Likens GE (2010) Effective ecological monitoring. CSIRO Publishing and
Earthscan, 170 pp
Lindenmayer DB, Banks SC, Blanchard W, Blair D, Franklin JF, Gibbons P, Ikin K, Laurance WF,
Likens GE, Manning AD, McBurney L, Stein JAR (2013) New policies for old trees: averting
a global crisis in populations of a keystone ecological structure. Biol Conserv
Long RP, Horsley SB, Hallett RA, Bailey SW (2009) Sugar maple growth in relation to nutrition
and stress in the northeastern United States. Ecol Appl 19:1454–1466
Lovett GM, Likens GE, Nolan SS (1992) Dry deposition of sulfur to the Hubbard Brook
Experimental Forest: a preliminary comparison of methods. In: Schwartz SE, Slinn WGN
(coordinators) Fifth international conference on precipitation scavenging and atmosphere-
surface exchange, vol 3 – the summers volume: applications and appraisals. Hemisphere,
pp 1391–1401
Lovett GM, Nolan SS, Driscoll CT, Fahey TJ (1996) Factors regulating throughfall flux in a
New Hampshire forested landscape. Can J For Res 26:2134–2144
Lovett GM, Bowser JJ, Edgerton ES (1997) Atmospheric deposition to watersheds in complex
terrain. Hydrol Process 11:645–654
Lovett GM, Likens GE, Buso DC, Driscoll CT, Bailey SW (2005) The biogeochemistry of chlorine
at Hubbard Brook, New Hampshire, USA. Biogeochemistry 72:191–232
Lovett GM, Canham CD, Arthur MA, Weathers KC, Fitzhugh RD (2006) Forest ecosystem
responses to exotic pests and pathogens in eastern North America. Bioscience 56:395–405
Lowe W, Likens GE (2005) Moving headwater streams to the head of the class. Bioscience
55(3):196–197
Lunt HA (1932) Profile characteristics of New England forest soils. Conn Agric Exp Sta Bull
342:743–836
Lutz BD, Mulholland PJ, Bernhardt ES (2012) Long-term data reveal patterns and controls on
stream water chemistry in a forested stream: Walker Branch, Tennessee. Ecol Monogr
82(3):367–387
MacIntyre F (1974) The top millimeter of the ocean. Sci Am 230:67–77
References 195

Marshall CE (1964) The physical chemistry and mineralogy of soils, vol 1, Soil materials. Wiley,
New York, NY, 388 pp
Martin CW, Pierce RS, Likens GE, Bormann FH (1986) Clearcutting affects stream chemistry in
the White Mountains of New Hampshire. USDA Forest Service Research Paper NE-579, 12 pp
Martin CW, Likens GE, Buso DC (2000) Comparison of long-term precipitation chemistry mea-
surements at the Hubbard Brook Experimental Forest, New Hampshire. Water Air Soil Pollut
120(3–4):359–379
Mast MA, Drever JI (1990) Chemical weathering in the Loch Vale watershed, Rocky Mountain
National Park, Colorado. Water Resour Res 26:2971–2978
Mau DP (1993) Estimating ground water recharge and baseflow from streamflow hydrographs for
a small Appalachian Mountain basin. M.S. Thesis, University of Colorado, Denver, CO, 130 pp
McDowell WH, Likens GE (1988) Origin, composition, and flux of dissolved organic carbon in
the Hubbard Brook Valley. Ecol Monogr 58(3):177–195
McGarity JW, Rajaratnam JA (1973) Apparatus for the measurement of losses of nitrogen as gas
from the field and simulated field environments. Soil Biol Biochem 5:121–131
McGuire KJ, Likens GE (2011) Historical roots of forest hydrology and biogeochemistry (Chapter
1). In: Levia DF, Carlyle-Moses D, Tanaka T (eds) Forest hydrology and biogeochemistry:
synthesis of past research and future directions. Ecological studies series, vol 216. Springer-
Verlag, Heidelberg, pp 3–26
McLauchlan KK, Craine JM, Oswald WW, Leavitt PR, Likens GE (2007) Changes in nitrogen
cycling during the past century in a northern hardwood forest. Proc Natl Acad Sci
104(18):7466–7470
Melillo JM (1977) Nitrogen dynamics in an aggrading northern hardwood forest ecosystem. Ph.D.
Thesis, Yale University, New Haven, CT
Merrens EJ (1988) Effects of the 1938 hurricane on forest structure and growth rates in a northern
hardwoods forest at Hubbard Brook, New Hampshire. Undergraduate Honors Thesis,
Department of Biological Sciences, Dartmouth College, 37 pp
Merrens EJ, Peart DR (1992) Effects of hurricane damage on individual growth and stand structure
in a hardwood forest in New Hampshire, USA. J Ecol 80(4):787–795
Mitchell MJ, Likens GE (2011) Watershed sulfur biogeochemistry: shift from atmospheric deposi-
tion dominance to climatic regulation. Environ Sci Technol 45:5267–5271. doi:10.1021/
es200844n
Mitchell MJ, David MB, Harrison RB (1992) Sulphur dynamics of forest ecosystems. In: Howarth
RW, Stewart JWB, Ivanov MV (eds) Sulphur cycling on the continents, vol 48, SCOPE. Wiley,
Chichester, pp 215–254
Mitchell MJ, Driscoll CT, Kahl JS, Likens GE, Murdoch PS, Pardo LH (1996) Climatic control of
nitrate loss from forested watersheds in the Northeast United States. Environ Sci Technol
30(8):2609–2612
Mitchell MJ, Lovett G, Bailey S, Beall F, Burns D, Buso D, Clair TA, Courchesne F, Duchesne L,
Eimers C, Fernandez I, Houle D, Jeffries DS, Likens GE, Moran MD, Rogers C, Schwede D,
Shanley J, Weathers KC, Vet R (2011) Comparisons of watershed sulfur budgets in southeast
Canada and northeast US: new approaches and implications. Biogeochemistry 103:181–207.
doi:10.1007/s10533-010-9455-0
Montieth DT, Stoddard JL, Evans CD, De Wit HA, Forcius M, Hogasen T, Wilander A, Skjelkvale
BL, Jefferies DS, Vuorenmaa J, Keller B, Kopacek J, Vesely J (2007) Dissolved organic carbon
trends resulting from changes in atmospheric deposition chemistry. Nature 450:537–541
Moore J-D, Ouimet R, Duchesne L (2012) Soil and sugar maple response 15 years after dolomitic
lime application. For Ecol Manage 281:130–139
Mulder J (1980) The neutralization of acid rain in the Hubbard Brook Experimental Forest, New
Hampshire, U.S.A. Department of Chemistry, University of Wageningen, The Netherlands.
Unpublished Report. 44 pp
Muller RN, Bormann FH (1976) The role of Erythronium americanium Ker. in energy flow and
nutrient dynamics of a northern hardwood forest ecosystem. Science 193:1126–1128
196 References

Navrátil T, Norton SA, Fernandez IJ, Nelson SJ (2010) Twenty-year inter-annual trends and
seasonal variations in precipitation and stream water chemistry at the Bear Brook Watershed in
Maine, USA. Environ Monit Assess 171:23–45. doi:10.1007/s10661-010-1527-z
Nezat CA, Blum JD, Klaue A, Johnson CE, Siccama TG (2004) Influence of landscape position
and vegetation on long-term weathering rates at the Hubbard Brook Experimental Forest,
New Hampshire, USA. Geochim Cosmochim Acta 68(14):3065–3078
Nriago JO (1990) The rise and fall of leaded gasoline. Sci Total Environ 92:13–28
Odum EP (1959) Fundamentals of ecology, 2nd edn. W.B. Saunders, Philadelphia, PA, 546 pp
Ollinger SV, Aber JD, Reich PB (1997) Simulating ozone effects on forest productivity. Ecol Appl
7:1237–1251
Oosting HJ (1956) The study of plant communities. W.H. Freeman, San Francisco, CA, 440 pp
Paillet FL, Hess AE, Cheng CH, Hardin E (1987) Characterization of fracture permeability with
high-resolution vertical flow measurements during bolehole pumping. J Ground Water
25(1):28–40
Palmer SM, Driscoll CT, Johnson CE (2004) Long-term trends in soil solution and stream water
chemistry at the Hubbard Brook Experimental Forest: relationship with landscape position.
Biogeochemistry 68(1):51–70
Pearson FJ, Fisher DW (1971) Chemical composition of atmospheric precipitation in the
Northeastern United States. United States Geological Survey Water Supply Paper 1535-P,
Washington, DC, pp P1–P23
Peart DR, Cogbill CV, Palmiotto PA (1992) Effects of logging history and hurricane damage on
canopy structure in a northern hardwoods forest. Bull Torrey Bot Club 119(1):29–38
Penman HL (1956) Estimating evapotranspiration. Trans Am Geophys Union 37:43–50
Peters SC, Blum JD, Driscoll CT, Likens GE (2004) Dissolution of wollastonite during the experi-
mental manipulation of Hubbard Brook Watershed 1. Biogeochemistry 67:309–329
Pierce RS (1967) Evidence of overland flow on forest watersheds. In: Sopper WE, Lull HW (eds)
International symposium on forest hydrology. Pergamon, New York, NY, pp 247–252
Pierce RS, Martin CW, Reeves CC, Likens GE, Bormann FH (1972) Nutrient loss from clearcut-
tings in New Hampshire. Symposium on watersheds in transition, Ft. Collins, Colorado.
American Water Resources Association and Colorado State University, pp 285–295
Pletscher DH (1982) White-tailed deer and nutrient cycling in the Hubbard Brook Experimental
Forest, New Hampshire. Ph.D. Thesis, Yale University, 145 pp
Pletscher DH, Bormann FH, Miller RS (1989) Importance of deer compared to other vertebrates
in nutrient cycling and energy flow in a northern hardwood ecosystem. Am Midl Nat
121:302–311
Powell ST (1964) Quality of water. In: Chow VT (ed) Handbook of applied hydrology. McGraw-
Hill, New York, NY, pp 19-1–19-37
Powers CJ, Singha K, Haeni FP (1999) Integration of surface geophysical methods for fracture
detection in bedrock at Mirror Lake, New Hampshire. In: US geological survey toxic substance
hydrology program. Proceedings of the technical meeting, Charleston, South Carolina, 8–12
March 1999. US Geological Survey Water-Resources Investigations Report 99-4018C, vol 3,
pp 757–768
Reuss JO (1976) Chemical and biological relationships relevant to the ecological effects of acid
rainfall. In: Dochinger LS, Seliga TA (eds) Proceedings of the first international symposium on
acid precipitation and the forest ecosystem. USDA Forest Service General Technical Report
NE-23, pp 791–813
Reuss JO, Johnson DW (1986) Acid deposition and acidification of soils and waters. Ecol.
Studies 59. Springer-Verlag NY
Reynolds RC, Johnson NM (1972) Chemical weathering in the temperate glacial environment of
the northern Cascade Mountains. Geochim Cosmochim Acta 36:537–554
Rhoads AG, Hamburg S, Fahey TJ, Siccama TG, Hane EN, Battles J, Cogbill C, Randall J, Wilson
G (2002) Effects of an intense ice storm on the structure of a northern hardwood forest. Can J
For Res 32:1763–1775
References 197

Richardson AD, Bailey AS, Denny EG, Martin CW, O’Keefe J (2006) Phenology of a northern
hardwood forest canopy. Glob Chang Biol 12(7):1174–1188
Rosenberry DO, Winter TC (2009) Hydrologic processes and the water budget. In: Winter TC,
Likens GE (eds) Mirror Lake: interactions among air, land and water. University of California
Press, pp 23–68
Rosenberry DO, Bukaveckas PA, Buso DC, Likens GE, Shapiro AM, Winter TC (1999) Movement
of road salt to a small New Hampshire lake. Water Air Soil Pollut 109:179–206
Roskoski JP (1980) Nitrogen fixation in a northern hardwood forest in the northeastern United
States. Plant Soil 54:33–44
Ross DS, Wemple BC, Jamison AE, Fredriksen G, Shanley JB, Lawrence GB, Bailey SW,
Campbell JL (2009) A cross-site comparison of factors influencing soil nitrification rates in
northeastern USA forested watersheds. Ecosystems 12:158–178
Ross DS, Shanley JB, Campbell JL, Lawrence GB, Bailey SW, Likens GE, Wemple BC, Fredriksen
G, Jamison AE (2012) Spatial patterns of soil nitrification and nitrate export from forested
headwaters in the northeastern United States. J Geophys Res 117, G01009. doi:10.1029/200
9JG001740
Roulet N, Moore TR (2006) Browning the waters. Nature 444:283–284
Schindler DW, Mills KH, Malley DF, Findlay DL, Shearer JA, Davies IJ, Turner MA, Linsey GA,
Cruikshank DR (1985) Long-term ecosystem stress: the effects of years of experimental acidi-
fication on a small lake. Science 228:1395–1401
Schlesinger WH, Bernhardt ES (2013) Biogeochemistry: an analysis of global change, 3rd edn.
Academic, San Diego
Schlesinger WH, Cole JJ, Finzi AC, Holland EA (2011) Introduction to coupled biogeochemical
cycles. Front Ecol Environ 9:5–8
Schofield CL (1976) Lake acidification in the Adirondack Mountains of New York: causes and
consequences. In: Dochinger LS, Seliga TA (eds) Proceedings of the first international sympo-
sium on acid precipitation and the forest ecosystem. USDA Forest Service General Technical
Report NE-23, p 477
Sellers WD (1965) Physical climatology. The University of Chicago Press, Chicago, IL, 272 pp
Shattuck PC (1991) Shallow water-table response to precipitation and evapotranspiration in an
ephemeral stream valley, Woodstock, New Hampshire. M.S. Thesis, University of New
Hampshire, Durham, NH, 113 pp
Siccama TG, Smith WH (1978) Lead accumulation in a northern hardwood forest. Environ Sci
Technol 12:593–594
Siccama TG, Bormann FH, Likens GE (1970) The Hubbard Brook Ecosystem Study: productivity,
nutrients and phytosociology of the herbaceous layer. Ecol Monogr 40(4):389–402
Siccama TG, Smith WH, Mader DL (1980) Changes in lead, zinc, copper, dry weight and organic
matter content of the forest floor of white pine stands in central Massachusetts over 16 years.
Environ Sci Technol 14:54–56
Siccama TG, Fahey TJ, Johnson CE, Sherry T, Denny EG, Girdler EB, Likens GE, Schwarz P
(2007) Population and biomass dynamics of trees in a northern hardwood forest at Hubbard
Brook. Can J For Res 37:737–749
Siegler HR (ed) (1968) The white-tailed deer of New Hampshire. New Hampshire Fish and Game
Department, Concord, NH, 256 pp
Slack JR, Lumb AM, Landwehr JM (1997) Hydro-climatic data network: Pemigewasset River,
Plymouth, NH Data Set 1874-1988. USGS WRI Report 93-4076
Smith WH (1976) Character and significance of forest tree root exudates. Ecology 57:324–331
Smith WH, Siccama TG (1981) The Hubbard Brook Ecosystem Study: biogeochemistry of lead in
the northern hardwood forest. J Environ Qual 10(3):323–333
Smith WH, Siccama TG, Clark S (1986) Atmospheric deposition of heavy metals and forest
health: an overview including a ten year budget for the input/output of seven heavy metals to a
northern hardwood forest, vol FWS-87-02. Virginia Polytechnic Institute and State University,
Blacksburg, VA, p 27
198 References

Sopper WE, Lull HW (1965) The representativeness of small forested experimental watersheds in
northeastern United States. Int Assoc Sci Hydrol 66(2):441–456
Sopper WE, Lull HW (1970) Streamflow characteristics of the northeastern United States. Bulletin
766. The Pennsylvania State University, University Park, PA, 129 pp
Stresky SJ (1991) Morphology and flow characteristics of pipes in a forested New England
hillslope. M.S. Thesis, University of New Hampshire, Durham, NH, 131 pp
Stumm W, Morgan JJ (1996) Aquatic chemistry, chemical equilibria and rates in natural waters,
3rd edn. Wiley, New York, NY, 1022 pp
Svensson T, Lovett GM, Likens GE (2012) Is chloride a conservative ion in forest ecosystems?
Biogeochemistry 107:125–134. doi:10.1007/s10533-010-9538-y
Sverdrup H, Warfvinge P (1988) Weathering of primary silicate minerals in the natural soil envi-
ronment in relation to a chemical weathering model. Water Air Soil Pollut 38:387–408
Tamm CO, Troedsson T (1955) An example of the amounts of plant nutrients supplied to the
ground in road dust. Oikos 6:61–70
Thiessen AH (1911) Precipitation for large areas. Mon Weather Rev 39:1082–1089
Thornthwaite CW (1948) An approach toward a rational classification of climate. Geogr Rev
38:55–94
Trewartha GT (1954) Introduction to climate. McGraw-Hill, New York, NY, 402 pp
Tripler CE, Kaushal SS, Likens GE, Walter MT (2006) Patterns in potassium dynamics in forest
ecosystems. Ecol Lett 9:451–466
Vadeboncoeur M, Hamburg S, Richardson A, Bailey A (2009) Examining climate change at the
ecosystem level: a 50-year record from the Hubbard Brook Experimental Forest. http://www.
hubbardbrook.org/research/climate/vadeboncoeur06.html
van Breeman N, Finlay R, Lundstrom U, Jongmans AG, Giesler R, Olsson M (2000) Mycorrhizal
weathering: a true case of mineral plant nutrition? Biogeochemistry 49:53–67
van Doorn NS, Battles JJ, Fahey TJ, Siccama TG, Schwarz PA (2011) Links between biomass and
tree demography in a northern hardwood forest: a decade of stability and change in Hubbard
Brook Valley, New Hampshire. Can J For Res 41:1369–1379
Verry ES (2003) Estimating groundwater yield in small research basins. Ground Water
41(7):1001–1004
Warby RAF, Johnson CE, Driscoll CT (2009) Continuing acidification of organic soils across the
northeastern USA: 1984–2001. Soil Sci Soc Am J 73:274–284
Warren DR, Bernhardt ES, Hall RO Jr, Likens GE (2007) Forest age, wood and nutrient dynamics
in headwater streams of the Hubbard Brook Experimental Forest, NH. Earth Surf Process
Landf 32:1154–1163
Warren DR, Likens GE, Buso DC, Kraft CE (2008) Status and distribution of fish in an acid-
impacted watershed of the Northeastern United States (Hubbard Brook, NH). Northeast Nat
15(3):375–390
Watmough SA, Dillon PJ (2001) Base cation losses from a coniferous catchment in central Ontario,
Canada. Water Air Soil Pollut Focus 1:507–524
Watmough SA, Dillon PJ (2004) Major element fluxes from a coniferous catchment in central
Ontario, 1983–1999. Biogeochemistry 67:369–398
Watmough SA, Aherne J, Alewell C, Bailey SW, Clair MOM, Dillon P (2005) Sulphate, nitrogen
and base cation budgets at 21 forested catchments in Canada, the United States and Europe.
Environ Monit Assess 109:1–36. doi:10.1007/s10661-005-4336-z
Weathers KC, Likens GE, Butler TJ (2007) Acid rain. In: Rom WN (ed) Environmental and occu-
pational medicine, 4th edn. Lippincott Williams & Wilkins, a Wolters Kluwer business,
Philadelphia, PA, pp 1507–1520
White AF, Brantley SL (1995) Chemical weathering rates of silicate minerals. Mineralogical
Society of America 31, Washington, DC, 583 pp. ISBN 0939950383
White EJ, Turner F (1970) A method of estimating income of nutrients in catch of airborne parti-
cles by a woodland canopy. J Appl Ecol 7:441–461
References 199

Whittaker RH, Likens GE (1973) Carbon in the biota. In: Woodwell GM, Pecan EV (eds) Carbon
and the biosphere. CONF-720510, United States Atomic Energy Commission, Springfield, VA,
pp 281–302
Whittaker RH, Likens GE (1975) The biosphere and man. In: Lieth H, Whittaker RH (eds) Primary
productivity of the biosphere. Springer-Verlag, NY, pp 305–328
Whittaker RH, Bormann FH, Likens GE, Siccama TG (1974) The Hubbard Brook Ecosystem
Study: forest biomass and production. Ecol Monogr 44(2):233–254
Wiens JA, Crawford CS, Gosz JR (1985) Boundary dynamics: a conceptual framework for study-
ing landscape ecosystems. Oikos 45:421–427
Williams AF, Ternan L, Kent M (1986) Some observations on the chemical weathering of the
Dartmoor granite. Earth Surf Process Landf 11:557–574
Winter TC (1981) Uncertainties in estimating the water balance of lakes. Water Resour Bull
17(1):82–115
Winter TC (1985) Approaches to the study of lake hydrology. In: Likens GE (ed) An ecosystem
approach to aquatic ecology: Mirror Lake and its environment. Springer-Verlag, New York,
NY, pp 128–135
Winter TC, Likens GE (eds) (2009) Mirror Lake: interactions among air, land and water. University
of California Press, 361 pp
Winter TC, Eaton JS, Likens GE (1989) Evaluation of inflow to Mirror Lake, New Hampshire.
Water Resour Bull 25(5):991–1008
Wood T, Bormann FH (1974) The effects of an artificial acid mist upon the growth of Betula
alleghaniensis Britt. Environ Pollut 7:259–268
Wood T, Bormann FH (1975) Increases in foliar leaching caused by acidification of an artificial
mist. Ambio 4(4):169–171
Woodwell GM, Whittaker RH (1967) Primary production and the cation budget of the Brookhaven
Forest. In: Young HE (ed) Symposium on primary productivity and mineral cycling in natural
ecosystems. University of Maine Press, Orono, ME, pp 151–166
Wright RF, Dale T, Gjessing ET, Hendrey GR, Henriksen A, Johannessen M, Muniz IP (1976)
Impact of acid precipitation on fresh-water ecosystems in Norway. In: Dochinger LS, Seliga
TA (eds) Proceedings of the first international symposium on acid precipitation and the forest
ecosystem. USDA Forest Service General Technical Report NE-23, pp 459–476
Yanai RD, Levine C, Green M, Campbell J (2012) Quantifying uncertainty in forest nutrient bud-
gets. J For 110(8):448–456
Yuretich R, Batchelder G (1988) Hydrogeochemical cycling and chemical denudation in the Fort
River watershed, central Massachusetts: an appraisal of mass balance studies. Water Resour
Res 24:105–114
Zhang Y, Mitchell MJ, Driscoll CT, Likens GE (1999) Changes in soil sulfur constituents in a
forested watershed 8 years after whole-tree harvesting. Can J For Res 29:356–364
Zimmer M, Bailey SW, McGuire KJ, Bullen TD (2012) Fine scale variations of surface water
chemistry in an ephemeral to perennial drainage network. Hydrol Process. doi:10.1002/
hyp.9449
Index

A snow and snowpack chemistry, 58–60

Acer saccharum, 61 sulfuric acid, 50
Acid winter conditions, 58
carbonic, 131, 137 Aerosols, 3, 42, 43, 80, 102–104,
in Hubbard Brook streams, 14 145, 170, 177
hydrochloric, 155 Aluminum
nitric, 44–48, 174 input-output budget, 97, 98, 103, 126
organic, 41–42 monthly concentration in precipitation and
sulfuric, 39, 50, 69, 133, 137 stream water, 78
Acid precipitation, 44–48, 173 output in stream water, 103, 107
acidification and recovery, 48 in particulate matter, 124
in Adirondack Mountain, 53 in precipitation, 62
ammonium deposition, 46 in stream water, 62, 69
anthropogenic emissions, 44 weathering release, 131
base cation, 47 Ammonium
in Canada, 53 in bulk precipitation, 38–50
effects on fish, 53 input-output budgets, 108
effects on forest, 45 monthly concentration in precipitation, 62
effects on soil, 47 monthly concentration in stream water, 62
in Scandinavia, 53 role in weathering, 132
SO2 and NOx emissions, 45, 46 seasonal concentration in precipitation, 103
sulfate concentrations, 47 seasonal concentration in stream
volume-weighted annual concentration, 45 water, 78, 85, 109
Acid rain, 48–68, 173, 174 Ashuelot River, 33
acid precipitation, 48 Atmospheric transport, 103
(see also Acid precipitation) Austin Stream, ME, US, 33
aluminum neutralization, 49 Avalanches, role in landscape denudation,
calcium and magnesium in, 49 124–126
dilution and effect of, 60–62
DOC concentration, 53, 54
effects on fish, 53 B
forest ecosystem, 62–68 Base cation depletion. See Acid rain
Hubbard Brook, 55, 56 Bedrock, 8
hydrogen ion concentrations, 50 Kinsman Granodiorite, 7
ionic charge for, 54–55 Littleton formation, 30
lead, 66 net soil release, 137
mineral acids, 53 Rangeley formation, 7

G.E. Likens, Biogeochemistry of a Forested Ecosystem, DOI 10.1007/978-1-4614-7810-2, 201

© Springer Science+Business Media New York 2013
202 Index

Bicarbonate Chemistry. See also Precipitation chemistry;

in bulk precipitation, 38–41 Streamwater chemistry
input-output budget, 103, 107 cation and anion concentrations, 37–38
in stream water, 137 precipitation inputs, 35
Biomass, accumulation of nutrients, 103, 126 streamwater outputs, 35
Biotite, 136, 137 weekly sampling, 35, 36
Biscuit Brook, NY. See Nutrient budgets Chloride
Budgets. See Mass balances annual bulk precipitation inputs of, 97,
Bulk precipitation 100, 101
annual inputs, 51, 91–97, 100, 101, 119 input in bulk precipitation, 44, 94–95, 103,
average inputs, 103, 104, 112–114 114
calcium, 49, 148 input-output budget, 31
in cations and anions, 54 monthly concentration in precipitation, 114
dissolved substances, 41 output in stream water, 94–95, 97, 100,
in electrical conductivity, 61 107, 114, 157
hydrogen ion concentration, 45, 118, 119 in particulate matter, 146
input–output budgets, 103, 104 in precipitation, 40, 51, 62
ionic charge, 39 in root exudates, 146
ions origin, 41–42 seasonal concentration in precipitation, 78
in long-term trends, 85–87 seasonal concentration in stream water, 78,
magnesium concentration, 49 82, 84
nitrate concentration, 72, 74, 118, 119 in stream water, 62, 69, 111
pH for, 60 in through fall and stemflow, 146
phosphate concentration in, 111 Chlorine
potassium, 150, 152 input–output budgets, 103
seasonal inputs and variations, 77–85, mass balance,,154–159
108, 110 nutrient budgets, 168
sodium, 149, 151 Coweeta Experimental Forest, NC, US.
sulfate concentrations, 47 See Nutrient budgets
sulfur, 151, 153
watershed, 72
D
Dams, organic debris, 13, 121
C Decomposition, 85, 121, 132, 136, 141
Calcium Deep seepage, 8, 29–32
acid rain, 174 Denitrification, 102, 103, 153
mass balance, 148–150, 154, 160 Denudation
net hydrologic flux, 99 cationic, 129, 130, 133, 134
net soil release, 134–136 fluvial, 124
nutrient budgets, 163 landscape, 124–126
nutrient cycle Dissolved inorganic nitrogen (DIN)
biomass accumulation, 141 input–output budgets, 118
bulk precipitation, 139 mass balance, 152, 154–156, 160–161
dry deposition, 141 nutrient budgets, 166
ecosystem pools and fluxes, 140 Drought, 13, 70, 71, 113, 114, 118, 142, 145
meteorologic inputs, 141 Dry deposition, 31, 42–43, 98, 102–104, 141,
sedimentary cycle, 177 145, 148–153, 157, 160, 170
streamwater chemistry, 69
streamwater concentrations for, 75
Carbon, organic, 131–132 E
dissolved, 36, 39–41, 53, 66, 74, 94–95, E. Bear Brook, ME. See Nutrient budgets
104, 113, 118, 175, 178 Ecosystem components
input-output budget, 40 atmosphere, 2, 3
Chemical budget, 18, 19 available nutrients, 2–4
Index 203

Hubbard Brook ecosystem seasonal concentration in stream water,

(see Hubbard Brook) 78, 84, 109
intrasystem cycle, 2, 3 sources for weathering, 130–133
mass balance, 15 in stream water, 60, 86
soil and rock, 2–3 in throughfall and stemflow, 66
Erosion, 127, 175 Hydrology, 17–34
Evapotranspiration, 18, 21, 23–29 deep seepage, 29–32
annual, 21, 23–29, 174 evapotranspiration (see Evapotranspiration)
bulk precipitation, 175 precipitation (see Precipitation)
potential, 25, 31 representativeness, 32–34
seasonal, 25–28 streamflow, 23–29
watertight substrate, 4, 30–31

F
Fernow, WV. See Nutrient budgets I
Forest floor, 8, 12, 65, 103, 147, 158 Inorganic dissolved substances, 90, 105, 118
Fossil fuels, combustion of, 43, 44, 52 Input–output budgets, 89–126, 176–178
ammonium, 108
annual variation in
G arithmetic average of annual values, 97
Gases, 2–4 average, 98
Geology, 7–8, 29–31. See also Bedrock; bulk precipitation and stream water,
Nutrient budgets 91–96, 100
Glacial till, net soil release, 137 direction of net change, 90
magnitude of net change, 90
net hydrologic flux, 99–100
H bulk precipitation inputs for, 103, 104
Hubbard Brook, 5–15 chlorine, 103
area and aspect, 7 debris avalanches, 124–126
climate, 5, 7 dissolved silica, 104
drainage streams, 13 gross output of dissolved substances
geology, 7–8 and annual streamflow, 117–118
location, 5 DIN, 118
representativeness, 32–34 hydrogen ion and nitrate, 118, 119
soil, 8 particulate matter (see Particulate
topography, 7 matter export)
vegetation and fauna, 8–13 pollutant, 114
Human activities, 1, 15 long-term changes in, 126
agriculture, 171 monthly variations in
fire, 171 ammonium, 110, 111
logging, 11, 12 calcium, 109, 110
Hydrogen ion. See also Weathering crossover patterns, 110, 111, 113, 115
budget, 129 phosphate concentration, 111
in bulk precipitation, 86 volume-weighted bulk precipitation,
increased input in precipitation, 112–113
42, 50, 173 nitrogen (see Nitrogen)
input in bulk precipitation, 39, 59, streamwater outputs, 107
90, 93–94, 97, 110, 113, sulfur, 103–104
118, 119 volume-weighted average, 108–109
monthly concentration in precipitation, 113 Iron
monthly concentration in stream water, 113 dissolved in stream water, 176, 177
output in stream water, 90–94, 97, 113, output in stream water, 98
118, 119 in particulate matter, 124, 176
204 Index

K potassium, 150, 152–154, 160

Kinetic energy, 171, 172 sodium, 149, 151, 154, 160
sulfur, 151, 153–154, 160
Merrimack River, NH, US, 5
L Mirror Lake, NH, US, 5, 8, 11
Land use management, 1 Model
Lead streamwater chemistry, 62, 70, 71
acid rain, 66 watershed-ecosystem, 28, 62, 71
ecosystem budgets of, 66
gasoline, 65
Litter. See Forest floor N
Long Island, NY, US. See Nutrient budgets N. Branch Contoocook River, NH, US, 33
N. Branch Hoosic River, MA, US, 33
Net hydrologic flux (NHF), 142, 147, 148, 176
M dissolved silica, 101
Maesnant Catchment, Mid-Wales. inorganic nitrogen, 102
See Nutrient budgets input–output budgets, 99–100
Magnesium magnesium and potassium, 99
in biomass, 129 Net soil release
biomass accumulation, 9 bedrock, 137
cycle, 139, 147 biotite, 136, 137
in forest floor, 129, 146 calcium, 134–136
input in bulk precipitation, 91–92, 97, elements mobilization, 134
99, 103, 110 forest growth status, 134
input-output budget, 90, 98, 101, 177 glacial till, 137
in litter, 146 neutralization rate, 137
monthly concentration in precipitation, 112 pH, 137
monthly concentration in stream water, 112 plagioclase, 134
net mineralization, 146 potassium, 136, 137
nutrient budgets, 165–166 precipitation inputs, 134
output in stream water, 91–92, 97, 99, rates of, 128, 130
107, 111 sodium, 134, 135
in particulate matter, 146 streamwater–bulk precipitation ratio, 135
in precipitation, 39, 40, 49, 67 vermiculite, 136, 137
in root exudates, 146 New England Rivers, US, 82
seasonal concentration in precipitation, 104 New York State Rivers, US, 82
seasonal concentration in stream water, 78, Nitrate. See also Nitrogen
80–82, 84, 109 biologic utilization, 82, 84
in soil, 177 in bulk precipitation, 36
in stream water, 40, 49, 61, 62, 68 increase in stream water, 82, 126
in throughfall and stemflow, 66, 67, 146 input in precipitation, 103
weathering release, 146 input-output annual budget, 97, 101,
Maine, US. See Watersheds 102, 105
Massachusetts, US. See Watersheds monthly concentration in precipitation, 116
Mass balances, 147–161 monthly concentration in stream water, 116
annual, 148, 149 monthly volume-weighted
biogeochemical data, 148 concentrations, 74
calcium, 148–150, 154, 160 output in stream water, 97, 99, 102,
chlorine, 151, 154–158, 161 107, 116
components of, 148, 149 seasonal concentration in stream water,
dissolved inorganic nitrogen, 150, 78, 80, 82, 84, 109, 111
153–155, 160–161 in stream water, 70–73, 82
NEF, 147, 148 and sulfate concentrations in, 73
nitrogen, 152–155 in throughfall and stemflow, 67
Index 205

Nitrification, 72, 82, 132, 178 potassium, 166–167

Nitrogen. See also Ammonium; Nitrate S.E. U.S., 165–168
cycle, 4 Sleepers River, VT, 164–168
dioxides, acid rain, 174 sodium, 166
hydrogen ion, 132 sulfur, 168
input–output budgets Taughannock Creek, NY, US, 164–168
denitrification, 102 Turkey Lakes Watershed, ON (Canada),
isotopes, 102 164–168
stream water, 101 Walker Branch, TN, US, 164–168
mass balance, 152–155 Nutrient cycles, 139–147, 158–161
net mineralization, 146 calcium, 139–141
vs. N-fixation, 155 features, 177
output in drainage water, 50, 178 flux, 173
output in stream water, 119, 146 gaseous phase, 139
plant uptake, 146 HBEF, 146–147
role in weathering, 132 potassium, 141–143
in throughfall and stemflow, 146 sulfur, 142, 144–146
volatilization, 80
weathering release, 146, 147
Nitrogen fixation, 50, 102, 155, 178 O
Northern Hardwood ecosystem, 163–170, Otter Brook, NH, US, 33
175, 176
biogeochemical data, 163
geologic substrates, 164 P
net primary production, 163 Particulate matter export, 120–124
nutrient budgets, 165–169 annual loss, 121, 123
calcium, 165 vs. dissolved substance export, 121–124
chloride, 166 measurement of, 120–121
dissolved inorganic nitrogen, 167 seasonal variation in erodibility, 121
magnesium, 165–166 Pemigewasset River, NH, US, 5, 8, 10
phosphorus, 167 pH
potassium, 168 distribution of, 76
sodium, 168 net soil release, 137
sulfur, 168 of precipitation, 44–48
precipitation inputs, 164 streamwater concentrations for, 13, 75
streamwater losses, 164 watersheds, 178
Northern Hardwood Forest, 3, 8, 11 Phosphate, 97, 101
basal area, 9 Phosphorus
biomass, 9 in biomass, 129, 146
net primary productivity, 8 cycle, 147
species composition, 9 in forest floor, 129, 146
Nova Scotia. See Precipitation chemistry input in bulk precipitation, 146
Nutrient budgets, 165–169 input-output budget, 98, 109, 167
Biscuit Brook, NY, 165–168 in litter, 146
calcium, 165 nutrient budgets, 167
chloride, 168 output in stream water, 146
Coweeta Experimental Forest, NC, US, in particulate matter, 124
71, 165–168 plant uptake, 146
dissolved inorganic nitrogen, 167 in precipitation, 67, 164
E. Bear Brook, ME, 165–168 in root exudates, 146
Fernow, WV, 165–168 in throughfall and stemflow, 66, 67, 146
Long Island, NY, US, 165–168 Piezometric divide, 30
magnesium, 165–167 Plagioclase, net soil release, 134
phosphorus, 167 Post acidic deposition (PAD) models, 61, 62
206 Index

Potassium lead, 64–66 (see also Lead)

in biomass, 129, 146 live and dead trees, biomass, 63
biomass accumulation sugar maple, 64
cycle, 141–142 throughfall and stemflow, 66–68
in forest floor, 129, 146 ionic ratio, 41–42
input in bulk precipitation, 91–92, 97, 99, Nova Scotia, 80
103, 114 volume-weighted annual concentrations, 38
input-output budget, 98 arithmetic means of, 40
in litter, 146 ionic charge, 39
mass balance, 150, 152, 154, 157 mean concentrations, 41
net mineralization, 146 sulfate and hydrogen ion, 39
net soil release, 130, 136, 137 Precipitation collection, 24
nutrient budgets, 166–167 collectors, 43, 44
nutrient cycle, 141–143
output in stream water, 91–92, 97, 99, 107,
111, 114, 146 R
in particulate matter, 124 Root exudates, 140, 141, 144–147, 159
plant uptake, 146 Runoff. See also Watersheds
in precipitation, 61, 67 experimental watersheds, 32–34
in root exudates, 146 infiltration, 8, 28
seasonal concentration in stream water, overland flow, 8
84, 109
in stream water, 40, 51, 61, 62, 68, 78, 82
streamwater chemistry, 70 S
in throughfall and stemflow, 66, 146 Salt
weathering release, 146 excess, 42
Precipitation, 18–23, 30 sea, 42, 50, 162
annual, 18–21 Saxtons River, VT, US, 33
average, 18, 20, 22, 23 S.E. U.S. See Nutrient budgets
bulk, 36, 39, 41–44, 47, 49, 51, 54, 61, Silica, 99
72–74 dissolved in stream water, 68
linear regression, 21 input-output budget, 98, 104
maximum, 18 output in stream water, 95–96, 107, 112
minimum, 18 Sleepers River, VT. See Nutrient budgets
monthly, 21–23 Small watershed technique, 4–5, 128, 171, 177
quantitative measurements, 21 Smith River, NH, US, 33
seasonal, 18, 19 Snow, 18, 19
Precipitation chemistry, 38–48 Snowpack, 7, 17–19, 23
acid precipitation Sodium
acidification and recovery, 48 in biomass, 129
ammonium deposition, 46 in bulk precipitation, 41, 43, 44, 51, 62, 79,
anthropogenic emissions, 44 80, 146
base cation, 47 in forest floor, 146
SO2 and NOx emissions, 45, 46 input in bulk precipitation, 92–93, 97, 99,
sulfate concentrations, 47 103, 110, 112
volume-weighted annual input-output budget, 98
concentration, 45 in litter, 146
acid rain and base cation depletion, 48–68 mass balance, 149, 151, 154, 160
(see also Acid rain) net mineralization, 146
dry deposition, 42–43 net soil release, 134, 135
elevational effects, 43–44 nutrient budgets, 166
forest biomass output in stream water, 92–93, 97, 99, 107,
canopy, 62 112, 146
environmental factors, 62 in particulate matter, 146
Index 207

plant uptake, 146 in stream water, 40, 47, 52, 62, 75, 84, 109
in root exudates, 146 streamwater chemistry, 69, 73, 86
seasonal concentration in precipitation Sulfur. See also Sulfate
seasonal concentration in stream water, 78, acid rain, 174
81, 82, 84 in biomass, 129, 146
in stream water, 40, 62, 68, 109 in forest floor, 146, 147
in through fall and stemflow, 67, 146 gaseous cycle, 177
weathering release, 135, 146 hydrogen ion, 133
Soil input in bulk precipitation, 146, 147
depth, 8 input–output budgets, 103–104
frost, 7, 8, 29 in litter, 146
humus, 29 mass balance, 151, 152, 154, 157
pH, 8 net mineralization, 146
water storage, 17, 28 nutrient budgets, 168
Streamflow, 13, 17, 18, 21, 23–29 nutrient cycle
annual, 21, 23, 25, 27 dry deposition, 145, 146
Stream-gauging weir, 23 ecosystem pools and fluxes, 144
Streamwater chemistry, 68–87, 178 gross uptake of, 145
acidification and recovery, 48 net biomass storage, 145
calcium concentrations, 69, 86 ouccrrence, 142
cation and anion, 37–38 streamwater, 145
chemistry of, 74, 76 vegetation surface, 145
collection of samples, 35 output in stream water, 146
concentration factors, 70 in particulate matter, 124, 146, 147
evapotranspiration, 69 plant uptake, 146, 147
nitrate, 70–73 in root exudates, 146
pH concentrations, 76 in throughfall and stemflow, 67, 146
potassium, 70 weathering release, 133, 146, 147
sample collection, 76
seasonal variations in
ammonium concentration, 85 T
bulk precipitation concentrations, 83 Taughannock Creek, NY, US. See Nutrient
mean seasonal streamwater budgets
concentrations, 84–85 Temperature
nitrate and sulfate concentrations, mean air, 7
80, 82 stream, 13
SO2 concentrations, 80 Thiessen polygon method, 18
stream algae, 77 Throughfall, 66–68, 103
streamflow, 82 Till, glacial, 7, 8, 31, 121, 133, 136, 137
volume-weighted average, 78–82 Topography, 7
sulfate concentrations, 69, 73, 86 pit and mound, 29
Sulfate. See also Sulfur Transpiration, 23, 25, 26, 28
input-output budget, 90, 99, 101, 103, 105 Turkey Lakes Watershed, ON (Canada).
net hydrologic flux, 99 See Nutrient budgets
in New England streams, 137
nutrient budgets, 168
output in stream water, 93–94, 97, V
99, 107, 114 Vectors
relative to nitrate concentration in stream biologic, 2, 3
water, 72, 73 geologic, 2, 3, 89
seasonal concentration in stream water, 78 meteorologic, 2, 3, 89
208 Index

Vegetation type, 8, 13 Water-year, 17–18

Vermiculite, net soil release, 136, 137 Weathering, 127–137
cation exchange complex, 129, 130
chemical weathering flux, 128, 130, 177
W flux estimation, 128, 133–134
Walker Branch, TN, US. See Nutrient budgets hydrogen ion
Watershed-ecosystem approach, 4–5 budget, 127–129
Watersheds, 4–5, 32–34, 172 carbon, 131–132
annual budgets (see Mass balances) external source, 130, 131
average annual and monthly precipitation, internal source, 131
22, 23 long-term relationship, 132
calcium, 140 nitrogen, 132
experimental, 32–34 stream water, 131
hydrogen ion budget, 129 sulfur, 133
hydrologic budget, 18, 19, 28 ionic composition, 127
Maine, 32, 33 mass balance, 128, 130
Maine, US, 32 mineral, 127
management, 172 net soil release (see Net soil release)
Massachusetts, 32, 33 qualitative and the quantitative changes, 127
New Hampshire, 32, 33 release, 127
potassium, 143 Weir, 13, 23, 35, 58, 59, 76, 77, 120–122
sulfur, 144 West Thornton, NH, US, 5, 8–9
Vermont, 32, 33 White Mountains, NH, US, 3, 5