* 1

GEODESY
The History and Concepts of Modern Geodesy

Introduction to

i-

Smith
C o ntro l, Roy M in n ick , Series Editor

INTRODUCTION TO GEODESY

This text is printed on acid-free paper. Copyright 1997 by John W iley & Sons, Inc. A ll rights reserved. Published simultaneously in Canada. Reproduction or translation o f any part o f this work beyond that perm itted by Section 107 o r 108 o f the 1976 United States Copyright A ct without the permission o f the copyright owner is unlawful. Requests for permission o r further inform ation should be addressed to the Permissions Department, John W iley & Sons, Inc., 605 T hird Avenue, N ew York, N Y 10158-0012. This publication is designed to provide accurate and authoritative inform ation in regard to the subject matter covered. It is sold with the understanding that the publisher is not engaged in rendering legal, accounting, o r other professional services. I f legal advice o r other expert assistance is required, the services o f a com petent professional person should be sought. Library o f Congress Cataloging in Publication Data: Smith, James R. (James R aym ond), 1935Introduction to geodesy : the history and concepts o f modern geodesy / J.R. Smith p. cm. (W iley series in surveying and boundary control) Wiley-Interscience publication. Includes bibliographical references and index. ISBN 0-471-16660-X (alk. paper) 1. Geodesy. I. Title. QB281.S55 1996 526'. 1 dc20 II. Series. 96-33301

Printed in the U nited States o f Am erica 10 9 8 7 6 5 4

Dedicated to Eleanor and Peter Cursed be he who moves his neighbor's boundary stone.

IF _ S 5

& 6

Contents

P re fa c e Introduction 1 I listory o f G e o d e s y
Pythagoras, 7 Aristotle, 7 Archimedes, 7 Eratosthenes, 7 Poseidonius, 11 I-Hsing, 13 Al-Mamun, 15 Columbus, 15 Frisius and Snellius, 17 Fernel, 17 Picard, 17 The Cassinis, 19 Mason and Dixon, 25

XIII

2
5

2 E a rth s S h a p e
Earth Parameters, 31 Spheroid and Ellipsoid, 33 Geoid, 35

27

3 U nits o f M e a s u re
Distance Measure, 37 Foot and Toise, 37 Meter, 39 Field Measurements, 41 Definitive Value, 41 The Legal Metre, 42 A Natural Standard, 44 Angle Measure, 44

37

4 T raditional S u rvey P ositioning T e c h n iq u e s

Horizontal Positioning, 49 Astronomical Techniques, 49 Triangiilntion, 63

47

Traversing, 69 Networks, 71 Vertical Positioning, 73 Precise or Geodetic Spirit Leveling, 73 Trigonom etric Heigh ting, 79 Barometric Heigh ting (Leveling), 79 Echo Sounding, 79

5 G e o d e tic S y s te m s
Horizontal Geodetic Datums, 83 Orientation o f Ellipsoid to Geoid, 85 Satellite Datums, 91 Discrepancies between Datums, 91 Datum Connection, 93 Datums before W orld War II, 93 Preferred Datums, 95 Other Arrangements, 97 Vertical Datums, 99 Tides, 101 Coriolis Effect, 107 Satellite Altimetry, 107 Transformations, 108

6 Physical G e o d e s y
Isostasy, 113 Gravity Anomalies, 115 Free-Air Anomaly, 119 Bouguer Anomaly, 119 Isostatic Anomalies, 121 Gravimetric Method, 121

7 W o rld G e o d e tic S ys te m
Derivation o f W orld Geodetic Systems, 125 Intercontinental Links, 126 Collection o f Gravity Data, 127 Gravity Values in Blank Areas, 127 Derivation o f Ellipsoid, 129 Gravimetric Orientation, 131 Army Map Service W orld Geodetic System, 135 U.S. Department o f Defense W orld Geodetic System 1960 (D O D W G S 6 O), 135 Later W orld Geodetic Systems, 135 Other Notable Systems, 136 EUREF, 136 ETRF 89, 136 ITRS and ITRF, 137

8 S atellite G e o d e s y
Early Satellites, 138 Interferometry, 141 Doppler, 141 Point Positioning, 145 Translocation, 145 Short-Arc Translocation, 147 Transit Doppler, 149 Observational Systems, 151

138

9 V ery Long B aseline In terfero m etry 10 G lo b al P ositioning S y s te m

Basic Description, 159 Other Positioning Systems, 161 GPS Operating Modes and Methods, 163 A Stationary (at Rest) Positioning, 163 B Mobile or Kinematic (Dynamic) Positioning, 163 A l, B1 Point (Stand-Alone) Positioning, 163 A2, B2 Relative (Differential) Positioning, 163 A l . l , A2.1 Static Positioning, 164 A2.2 Stop-Go Kinematic Positioning, 164 A2.3 Pseudo-Kinematic Positioning, 164 A2.4 Rapid Static Positioning, 164 A2.5, B2.1 Nonstop Kinematic Positioning, 164 B2.2 Kinematic On-the-Fly Positioning, 165 Seeded Rapid Static Positioning, 165 Ambiguities and Their Resolution, 165 Codes, 166 P Code, 166 C/A Code, 166 Y Code, 166 Range Values, 166 Real-Time GPS, 167 Sources o f Error, 169 Satellite, 169 Signal and Atmosphere, 170 Receiver, 170 Selective Availability, 170 Datum, 170 The Future, 170

155 159

11 G ravity
Absolute Measurements, 171 Relative Measurements, 172 Measurement Use in Geodesy, Geology, and Geophysics, 173

171

1 2 G y ro s c o p e

174

13 Inertial S urveying
Accelerometers, 176 Zero Velocity Update, 177

176

14 V elo city o f Light, E D M , a n d L aser R an g in g

Velocity o f Light, 179 Basic Relationships, 179 Early History, 180 Electromagnetic Distance Measurement (ED M ), 181 Laser Ranging, 185

179

15 P ro jectio n s
W orld Projection, 191 Mercator Projection, 192 Transverse Mercator Projection, 193 Lambert Conical Orthomorphic Projection, 195 State Plane Coordinates, 197

187
K

16 E x a m p le s o f M o d e rn P ro jec ts
Superconducting Super Collider (U SA), 198 Gotthard Tunnel (Swiss Alps), 199 Mineral Exploration in Kazakhstan, 200 Heighting by GPS in Lesotho, 200 Plate Tectonics, 202 Determining H eight o f Mount Everest, 202 Changing National Grid and Geodetic Datum in Great Britain, 205

198

B iblio graphy A b o u t th e A u th o r In d ex

207 213 214

List of Figures

Realizing the Earth Is not Flat, 4 Principle o f Determining Size o f Earth as a Sphere, 6 2. Method o f Eratosthenes for Determining Size o f Earth, 230 b .c ., 8 3. Method o f Poseidonius for Determining Size o f Earth, 100 b .c ., 10 4. Method o f I-Hsing for Determining Size o f Earth, a .d . 724, 12 5. Method o f Al-Mamun for Determining Size o f Earth, a .d . 820, 14 6. Use o f Triangulation by Frisius (1533), Snellius (1620s), 16 7. 8 . Oblate and Prolate Spheroids, 18 9. Spread o f 18th-Century Arcs, 20 10 . Relation o f Oblate Spheroid to Fitting Sphere, 22 11 . Mason and Dixons (1766) Method o f Arc Measure, 24 1 2 . Curved Surface and Plane Equivalent, 26 13. Effect o f Elevation on Arc Length, 28 14. Elements o f an Ellipse, 30 15. Spheroids and Ellipsoids Defined, 32 16. Relation o f Geoid to Ellipsoid, 34 17. Units o f Length, 36 18. Questions about the Earth, 46 19. Celestial Coordinate Systems, 48 2 0 . Azimuth, Bearing, and Direction, 50 2 1 . Deflection o f Vertical at Point P, 52 2 2 . Different Concepts o f Year, 54 23. Motions o f Earths Axis, 58 24. Precession, 60 25. Triangulation Network, 62 26. Equipment for Distance Measurement, 64 27. A Trilatration Network, 6 6 28. Traverses, 68 29. Different North Lines, 70 30. Leveling, 72 31. Orthometric and Dynamic Heights, 74 32. Ellipsoidal and Geoidal Heights, 76 33. Effect o f Curvature and Refraction, 78 34. Horizontal Datum, 82 35. Datum Orientation for Single Astronomical Station, 84 36. Orientation for Asli'ogcodetic Datum, 8 6
1.

37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78.

Deflection o f the Vertical, 8 8 Astrogeodetic Deflections Are Relative, 90 Example o f Many Datums in Southeast Asia Area, 02 Preferred Datums, 94 Fitting Different Ellipsoids to Different Areas, 98 Tide Generation, 100 Selection o f Ocean Levels, 102 Sea-level and Depth Measurement, 104 Satellite Altimetry, 106 Deviation Due to Mountain Mass, 110 Isostatic Hypotheses, 112 Effect o f Mass Anomalies, 114 Factors Involved in Reducing Observed Gravity, 116 Gravity Anomalies, 118 Products o f Gravimetric Method,, 120 Geoid-Ellipsoid Relationships, 122 Shoran Connection between Scotland and Norway, 124 Arc Method o f Determining Earths Size, 128 Preferred Datums are Relative Datums, 130 Gravimetric Datum Orientation, 132 Deflections o f Vertical Computed from Earths Gravity Can Determine a W orld Geodetic System, 134 Principle o f Radio Interferometry, 140 Doppler Principle, 142 Doppler Positioning Methods, 144 Simultaneous and Orbital Methods First Used with SECOR, 146 Short-Arc Method, 147 Single Difference, 148 Double and Triple Difference, 150 VLBI Baseline Determination, 154 Global Positioning Systems Orbital Configuration, 158 GPS Control Stations, 160 GPS Operating Modes and Techniques, 162 Position Circles, 168 Reduction o f EDM Measurements, 182 EDM Waves, 184 Map Projections, 186 The Effect o f Using Different Spheroids, 188 Transverse Mercator Projection, 189 U TM Zone 36, 191 Lambert Grid Typical State or Country Greater East-West than North-South, 194 Transverse Mercator Grid Typical State or Country Greater NorthSouth than East-West, 176 Heighting Mount Everest, 204

Preface

I o many who study land surveying, whether in its own right or as part o f another discipline such as civil engineering, geography, or topographic sci ence, the word geodesy can strike an element o f fear. This is probably because it is associated with higher mathematics, complicated formulae, and the prob lems o f thinking in three dimensions. T o become fully conversant with the subject a high level o f mathematical ability is an essential prerequisite. It should be obvious that to teach such a subject, as with any other branch o f surveying, one would have to become involved in all its mathematical intricacies. Here, however, the idea is simply to give the flavor o f geodesy and to develop an appreciation o f the range o f topics encompassed and the many facets o f science that it impinges upon. I set out with the idea o f trying not to include any formulae or equations and have nearly succeeded. N ot until I recently came across an old copy o f E. B. Denisons Astronomy without Mathe matics London 1867 would I have considered it as even a remote possibility. For those who wish to follow this branch o f surveying as a career the daunting elements have to be firmly grasped, but on the other hand it is quite feasible to get an overall feel for the discipline without becoming involved in the intricacies. It is the aim here, as it was in Geodesy for the Layman and Basic Geodesy, which this replaces, to put across the elements o f geodesy in an elementary way, using simple language and with the sid o f numerous diagrams. The pure geodesist may well have criticisms, since in order to achieve simplicity I have made generalizations that in the strict sense may not be correct but which are sufficiently near the truth for the purposes o f this volume. Certainly any reader with more than a passing interest must refer to specialized up-to-date texts. Good examples are Geodesy by Guy Bomford and GPS Satellite Surveying by Alfred Leick. Geodesy is a very complicated, highly mathematical science, and in no way should readers assume that this text is any more than an outline introduction. It covers the scope o f geodesy and its capabilities but does not attempt to touch many o f the intricate areas that have to be accounted for in rigorous applications. For example, some explanation is given o f ocean tides but not o f earth tides, and o f equipotential but not o f the spherical harmonics that can be used to express it. On the other hand, this new edition does include brief mention o f some o f the more abstruse quantities, such as nutation and aberration.

ProtlK :o

Many instruments are mentioned, but without details as to the necessary adjustment and observational procedures and subsequent data analysis through statistical techniques. Variations in atmospheric conditions play a considerable role and are still far from adequately understood. Particularly as we enter the satellite-reliant age, a wider range o f profession als will come into contact with geodesy and require an outline knowledge o f its capabilities and fields o f application. The area o f satellite research and the general application o f electronics to surveying is changing so rapidly that it is difficult to keep pace with develop ments. H ere no attempt has been made to be up to date immediately with technological developments. Rather, an outline is given o f basic techniques and principles. In many instances these survive equipment changes with only slight modification. Much o f the material in this book is based on the two very popular predeces sor volumes mentioned above, with updating especially in areas o f rapid change, such as those involving satellites. However, in the short 8 years since the first edition o f Basic Geodesy, the use o f satellites for positioning has pro gressed so rapidly on a global scale that it now warrants a much enlarged, separate treatment. Thus, the present volume should not be taken as a defini tive textbook on GPS; it is an introduction intended no more than to lay the foundations. It must be said that many readers o f the first edition have sug gested that more space should be given to GPS, but this volume is really meant to focus on the basics o f geodesy. Hence a fine line is being trod between maintaining the theme and trying to satisfy potential readers. On the other hand, it can truthfully be said that with the introduction o f satellites new factors influence measurements for geodetic and these are mentioned in the appropriate sections. An additional chapter has been added in this new edition to illustrate the wide range o f modern projects where geodesy in one form or another is very pertinent. The examples are by no means exhaustive but represent a cross section o f topics that are fully written up elsewhere. The interested reader can delve deeper as he or she so wishes. Previous to the 1950s no one in geodesy was interested in more than the first few hundred meters o f the atmosphere or in the characteristics o f electromagnetic wave propagation. I f stations were not intervisible, they were not readily o f use, and thus hilltops were a priority. Now that is no longer the case. Above all, it is my wish for this volume to be a painless initiation into geodesy and a precursor to immersion in textbooks for the serious student or, alternatively, to provide a general understanding for those in cognate professions. A selected, but by no means comprehensive, bibliography for further reading is included at the end o f the book. Particular thanks must be extended to the original authors for the clarity o f both their text and diagrams. They have served the profession well for forty years, and it is to be hoped that this book, in its small way, will be able to match their success. Am ong colleagues to whom thanks must be extended

I mention in particular Alan Wright o f Global Surveys Ltd., who has had considerable input on both the first and second editions, and Carl Calvert, Adam Chrzanowski, Kevin Dixon, Dave Doyle, Kurt Egger, Keith Greggor, Alan Haugh, and Walt Robillard. Thanks must also be extended to Roy Minnick for setting the challenge o f updating such a popular works. J. R.
S m it h

Petersfield, U.K. 1997

Second edition

Introduction

W h a t is this subject o f geodesy? Is it just an esoteric area o f science with little modern practical application, or does it function quietly to the benefit o f all but receive little public attention? Is it a product o f the age o f electronics, or does it have a much longer history? In brief, what is it, how has it developed, what areas o f life does it impinge upon, and how is it put to practical use? Definitions vary from the inevitable one line science o f measuring the earth, or surveying any large part of it to detailed paragraphs setting out all the implications o f that single sentence. It would serve no useful purpose to offer yet another definition solely for this volume. Instead, Ihe following is an amalgam o f several current definitions so you can appreciate from the start the range o f topics involved. Geodesy, from the Greek, literally means dividing the earth, and as a first objective the practice o f geodesy should provide an accurate framework for the control o f national topographical surveys. Thus geodesy is the science that determines the figure o f the earth and the interrelation of selected points on its surface by either direct or indirect techniques. These cliai act eristics further makes it a branch o f applied mathematics, one that must m < hide observations that can be used to determine the size and shape o f the eai 111. 1 1id the definition o f coordinate systems for three-dimensional positioning; the variation o f phe nomena near to or on the surface, such as gravity, tides, earth rotation, crustal movement, and deflection of the plumb line; (ogrtliei with units o f measurement and methods o f representing the curved e.irlli surface on a flat sheet o f paper. Unfortunately there is a trend today for the term tfm/nv to be applied in an umbrella manner, particularly in the European <'.>1111111111iIy, to describe all activities from valuation, land management, soil icsling, cartography, setting-out, underground surveys, national mapping, Itound.iiy surveys, Land Information Systems, and in fact every activity except n<< > <Irsy in its traditional definition! In addition, we are now being pressed to us< ilie term geomatics

to cover almost as wide a selection as the above list. Maybe it will soon be possible to refer to geomatics and geodesy as covering everything that is understood to be under the authority o f a surveyor. Some 65 000 geodetic surveyors are said to work in the European Commu nity (EC) alone! A t least two noughts should be deleted from this figure to get to the traditional geodesists and those who fit the sphere o f activity o f the International Association o f Geodesy (IA G ) an organization that would recognize few, if any, o f the topics encompassed by the EC statement. The problem, unfortunately is further complicated by the wide range o f uses now being found for the global positioning system (GPS). It is an earthcentered system, relying on earth-orbiting satellites, but it has recreational uses in addition to survey applications and acceptable requirements from a few centimeters to many meters depending on the use. T o quote from the ACSM Bulletin o f March/April 1992:
Is the ability to routinely determine boundary lines accurate to 1:100 000 by GPS within the province o f the land surveyor? I think not unless that person is also trained in the geodetic sciences. GPS is a tool o f geodesy land surveyors using GPS are practicing geodesy. Th erefore they should be thoroughly trained in the geodetic science. This in cludes higher mathematics, physics, measurement theory, analysis, adjustment computations, geodetic concepts and field experience. (B loodgood, 1992)

Here it is the intention to follow more the view o f the LAG than that o f the EC, but with some editorial license. T o be o f wide topical use, this second edition has a section on GPS that is perhaps disproportionately long. N o apology is given for this, as it is a current tool o f geodesy and as such warrants fuller treatment than the more historic topics. In this last quarter o f the 20th century so much now revolves around very accurate positioning whether for oil rigs many kilometers offshore, intercontinental missile targets, shuttle missions into space, monitoring crustal movements, laser bathymetry, navigation, seismic surveys, gravity observations, deformation measurements, all forms o f military positioning, or small-scale mapping from space that the fruits o f geodetic labors have an almost limitless range o f applications. However, before venturing into the realms o f artificial satellites it will be instructive to trace the origins o f geodesy from the first few centuries B.C. up to the present day. From the times o f a flat earth concept, through the sphere and spheroid to the geoid; from the knotted rope for measurement to suspended wires, electromagnetic systems, laser ranging to the moon, and the use o f orbiting satellites.

Pole

1. Realizing the Earth Is Not Flat

CH A P T E R

1
History of Geodesy

ram earliest times man was increasing his knowledge o f the planet on which he lived, although much o f its function would have been a mystery. As he tilled the fields, he noticed the changing lengths o f daylight and the changes in the maximum height o f the sun as the year progressed. I f he happened to be a fisherman or simply a watcher o f seaborne activity, he could well have realized that when a distant vessel disappears from view the lower part goes first and the top o f the mast is the last to vanish. I f a person stands on the beach (figure la ) with his eyes 2 m (meters) above the water level, and the top o f the mast o f a retreating vessel is 30 m above water level, the vessel would first begin to disappear when 5V2 km (kilometer) from shore and finally vanish when 22 km out not that humans had the means at that time either to measure the distance or to use kilometers! As man became more adventurous and traveled farther from his home territory, he would have appreciated that if he went in an easterly or westerly direction, the star now recognized as the Pole Star, and around which the signs o f the zodiac revolved, would stay more or less at the same altitude, whereas if he went in a northerly or southerly direction, the altitude would change (figure lb ). In addition, he may well have noticed that under the heat o f the midday sun the length o f his shadow changed as he traveled north or south, whereas, for east or west travel over periods o f a few days, the shadows remained sensibly constant in length. Whether or not these phenomena gave him any early ideas as to the shape o f the earth he was standing on is unknown. One might even question whether the shape o f the earth was o f any interest at all since travel was very limited and adventurers few and far between. However, thoughts on its shape gradually progressed from flat, through a disc or short cylinder, to a number o f variations on these.
0---------------------------------------------------------------------------------------------------------------------------

niMiury ui vjwuuesy

C ircum ference = ^ ~ ' s

PYTHAGORAS Certainly by the time o f Pythagoras (c. 580-500 B.C.) the earth was considered to be spherical if for no other reason than that the sphere was thought by the philosophers o f the time to be the perfect regular solid. As might be expected, not all authorities agreed, and even as late as the sixth century a .d ., scorn was still being poured on the idea o f a round earth and queries raised as to how such a shape could retain the waters o f lakes and oceans. I f we had a table tennis ball, there would be no difficulty measuring its diameter with calipers. Enlarge this to a dome 1 km across, and, while the problem would be a little more difficult, there would still be several ways o f measuring it. W e might use techniques that were the reverse o f setting out curves for example, we could directly measure the tangent lengths or calcu late offsets from a tangent and then do the necessary calculation. But what happens when the dome has a radius o f some 7000 km? Some totally different approach is required. Granted, in the light o f methods avail able today, there would be no difficulty in using satellites, but what about 2 0 0 0 + years ago, before the developments o f the electronic age?

ARISTOTLE First attempts at putting a dimension on the sphere are credited to Aristotle (c. 384-322 B.C.) who recorded a diameter o f 400 000 stad.es. This figure could vary from 84 000 to 63 000 km, depending on our choice o f conversion factor. But how he arrived at his value is unknown.

ARCHIMEDES A century later, Archimedes quoted 300 000 stades (63 000 to 47 000 km). It might even be that each used a different length o f stade and that both used the figures equivalent to 63 000 km possibly from the same initial guess.

ERATOSTHENES For an approach o f any scientific significance we turn to Egypt, where the Greek philosopher Eratosthenes (276-195 B.C.) was librarian at the famous library in Alexandria. He adopted a principle, shown in figure 2, that is still acceptable today. I f for the moment we assume the earth is a sphere, then its

3. Method of Eratosthenes for D eterm ining Size of the Earth, 2 3 0 B.C.

size can be found if two quantities are known: the distance .s between the two points that lie on the same meridian (a line joinin g the two poles), and the angle a subtended by those two points at the center o f the earth. Thus

circumference = 360 .

s a

On the face o f it the determination o f a might appear particularly difficult, but luckily there is no need to get to the center o f the earth to observe its value. This can be achieved on the surface in various ways, as illustrated in the following examples. In his position as librarian, Eratosthenes had access to a vast range o f material and ideas. Thus, whether he actually made the measure with which he is credited or it was a combination from other reports is open to question. Nevertheless, as the story goes (figure 3), Eratosthenes noticed that at the summer solstice the sun shone directly into a deep well at Syene (Aswan). This spot would have to be on the tropic o f Cancer for this observation to apply. A t the same time at Alexandria he found (probably by using a scaphe a hemispherical bowl with a central gnom on) that the sun cast a shadow equiva lent to Vso o f a circle about Vi o f a sign o f the zodiac or 712' in present terminology. This angle is the same as the one subtended at the center o f the earth by the two terminal points on the earths surface, Syene and Alexan dria. H e combined this measure with a value o f 5000 stades for the distance between the two cities. Whether this was measured by the royal road surveyors, estimated in terms o f camel-days journey, or an educated guess is unknown. The observations and assumptions are full o f possible errors. The modern equivalent o f 5000 X 50 = 250 000 stades (52 500 to 39 400 km) is close to the present-day accepted distance o f around 40 000 km, but this can only be considered fortuitous, for several reasons. Syene is not on the tropic o f Cancer but some 60 km north o f it; Syene and Alexandria are several degrees away from being on the same meridian, and the measure o f their distance apart was some 10% in error. Then when we realize that Eratosthenes apparently added 2 0 0 0 stades simply to make the result divisible by 60, the uncertainty o f the whole calculation becomes apparent. Aside from errors o f position, varying interpretations o f the stade, and other such problems, what about basic inaccuracies in measuring the distance and the angle? Let us assume for the moment that the earth is a perfect sphere (it makes the arithmetic far simpler), an arc o f 1 is measured, and the distance between the ends is found to be X meters (about 110 000 m ). Then the circumference is 360X meters. I f the angle was incorrectly measured by 1' o f arc, the resulting circumference would be incorrect by 6 X me ters (approximately 660 km). Conversely, if the distance were incorrect by 1000 m, then the circumference would be incorrect by 360 000 m (360 km). Neither o f these figures is insignificant, and yet both would have been possible with the equipment available at the time.

I llstory ol ( too* limy

C ircum ference = 48 x 5000 = 2 4 0 0 0 0 stades

4. Method of Poseidonius for D eterm ining Size of Earth, 100 B.C.

History ot Ueodesy

POSEI DON I US A century later, Poseidonius used a different technique (figure 4) to obtain a comparable result o f 240 000 stades. H e noticed that the star Canopus was on the horizon when viewed from Rhodes while at Alexandria its elevation was V48 o f a circle (730'). The distance between the two cities was estimated as 5000 stades, but since they are separated by water, this could only have resulted from mariners estimates. As with Eratosthenes, the angle used by Poseidonius was considerably in error, in fact by 21 /4, and the distance was o ff by some 30% to 40%. The errors were compensatory, however: the too-small angle and too-long distance combined to give a fortuitously acceptable result.

Difference in shadow lengths is a fu n ctio n of a n g le a

n isio ry or ueoaesy

l-H S IN G
A not Iter variation on the same theme was used in China in the eighth century \ i> (ligure 5) when I-Hsing, a Tantric Buddhist monk who was also a famous m.ii lumatician and astronomer, was charged with organizing observations nvci an arc o f some 11 400 li (5000 km) on the 114 east meridian. This ic(|uiicment arose from the long-held belief that the shadow length o f an 8 lool (1.96-m) gnom on changed by 1 Chinese inch (0.02 to 0.03 m) for every 1000 li traveled along a meridian. In the event, the rate o f change was more in .ii ly 4 inches (0.1 m ). The same data, however, gave a value for the earths i in umference o f 128 300 li (56 700 km).

6. Method of Al -M amun for D eterm ining Size of Earth, a.d. 8 2 0

AL-MAMUN Yrl another alternative was given in a .d . 820 by the Caliph Al-Mamun in Arabia (figure 6 ). H e directed his astronomers to measure several (probably 4) lines around Baghdad and A1 Raqqah. From a central starting point, parties went both north and south until the vertical angles to Polaris (the Pole Star) had ( hanged by 1. How the distances were measured is again open to doubt. Some records suggest long knotted ropes, others the distance traveled by horsemen in a given time. However measured, the distances were recorded in Arab miles with an accepted result o f 56% Arab miles for 1. Conversion is again uncertain, but an accepted equivalent o f 111.073 km gives the circum ference as 39 986 km.

COLUMBUS
1 1 was probably such uncertainties as these for the size o f the earth that led Christopher Columbus to think that Asia was only some 4000 miles (6400 km) west o f Europe. For example, the geographer to whom all seafarers o f the Middle Ages referred was Ptolemy ( a .d . 100-178), who accepted a value for the earths circumference o f 180 000 stades. This could be converted by I he wrong factor to 28 350 km. I f this were then used to determine the circumference at, say, 36N latitude, the result would be 23 200 km, as com pared to the more realistic value o f about 32 800 km. I f the earth were thought to be only some 70% o f its actual size, then the navigator/explorer had little chance o f success. Some historians give an alternative explanation that Columbus used the result o f Al-Mamun but assumed it was given in Italian (Roman) rather than Arab miles a difference o f some 25% too small. Yet another way o f accounting for the shortfall in Columbuss estimate is to consider that a degree o f latitude was taken to be 56% Roman miles instead o f the more nearly correct 75 Roman miles. This results in a ratio o f 0.76, again producing a figure far too small. It is not my intention here to examine the Columbus voyages in detail. In summary, however, one scenario might have been that Columbus took the inhabitable world to be 13 750 Roman miles and 300 extent at 36 N, whereas it should have been more like 9600 miles and 160. For the earths circumfer ence he took 16 500 instead o f 21 580 miles. Thus sailing west from Europe using his figures would require traveling only 60 longitude and 2750 miles, whereas it should have been 220 and 12 000 miles. Many other scenarios could be set up, but they would all imply that the distance going west was far shorter when in fact it was far longer. It is likely that Columbus not only underestimated the size o f the earth but also overestimated the size o f the area known as the inhabitable world.

15

7. Use of T riangulation by Frisius (1 5 3 3 ), S nellius (1 6 2 0 s )

The 16th and 17th centuries were the turning point in most aspects o f science in the Western world, and this applies also in no small measure to surveying. Developments in instrumentation including the telescope, ver nier, thermometer, and barometer and in computing techniques such as logarithmic and trigonometric tables all opened the way for the introduction and development o f triangulation. (See Chapter 4.) FRISIUS AND SNELLIUS In a publication o f 1533, Gemma Frisius put forward the principles oftriangulation, but it is uncertain whether he actually put them into practice. The scheme he illustrated was certainly not observable. The technique was further developed by Tycho Brahe in the late 16th century, but the real landmark was the technique used by W illibrord Snellius in the 1620s. Between Bergen-op-Zoom and Alkmaar in the Netherlands, Snellius ob served a triangulation scheme with five baselines in the vicinity o f Leiden. This technique deviated from all previous arc measures in that the distance between the terminal points could now be determined indirectly rather than directly. That is, instead o f measuring 100+ km by tape or similar device, all Ibat was now required was to very accurately measure at least one line (in the case o f Snellius the lines averaged 1300 m) for the calculation o f an arc o f around 130 km. (See figure 7; triangulation is explained in more detail in Chapter 4.) FERNEL At about 1525, Jean Fernel measured an arc o f 1 between Paris and Amiens using the revolutions o f a carriage wheel to determine the distance. For his angle he first measured the height o f the sun at midday. H e then calculated what the height would be at a point 1 farther north and traveled in that direction until he found a position where he was able to observe that value. (There was a small error in his calculation, but that is only to be expected.) H e found the distance between these two points to be 68.096 Italian miles. There is some doubt about the exact equivalent o f the Italian mile in other units such as the toise, although it is generally taken as 56 746 toise. PICARD I n 1669,1 AbbeJean Picard introduced the telescope for observing star altitudes and angles o f triangulation. A t that point the observation work became similar in principle to that subsequently used until the last quarter o f the 2 0 th century, with the advent o f the artificial satellite and the global positioning system (G P S ). With wooden rods, Picard measured two baselines o f 5663 and 3902 toises (1 1 038 and 7605 m) on his triangulation from Malvoisin near Paris to Sourdon near Amiens. He calculated that 1 equaled 57060 toises (111 210 m ).

M _____________________________________________________

m sio ry ui u u o u e s y

(a) Prolate s2 > s i

*- si

(b) Oblate s j > s2 8. Oblate and Prolate Spheroids

THE CASSINIS By the end o f the 16th and early 17th centuries, several conflicting measure ments and calculations as to the shape o f the earth had been made in Europe. These in turn affected estimates o f the earths size. Contemporaneous with the founding o f the Royal Society in London and lAcadmie Royale des Sciences in Paris came European expeditions abroad. These had wide terms o f reference, such as observations o f the oscillations o f a pendulum at different elevations and in different latitudes, and o f changes in the velocity o f sound and o f atmospheric pressure with elevation. These observations, together with the theories o f Isaac Newton, suggested the earth must be flattened at the poles. Almost in parallel with these activities, several long arcs were being mea sured in France under the guidance o f the Cassini family, who occupied posts in the Paris Observatory over four generations. All the French arc measure ments suggested equatorial flattening rather than flattening at the poles. How could such overwhelming evidence be disputed by theory? As we shall see, the discrepancies resulted from the inadequacy o f the available instrumentation, which was unable to detect the small changes involved. This was a scientific controversy par excellence as it has been described by some, the battle between the pumpkin and the egg shape. Th e technical names given to the rival shapes are oblate spheroid for flattening at the poles and prolate spheroid for flattening at the equator (figure 8 ). For any given angular value, the equivalent arc length will increase toward the equator for a prolate spheroid and increase toward the poles for an oblate one. It will be noticed from figure 8 that when the earth is no longer considered to be a sphere then the varying radii o f curvature o f the surface do not meet at a single central point.

History of Geodesy

LAPLAND Pello

North Pole

Quito PERU Cuenca

9. Spread of 18th-C entury Arcs

History of Geodesy

I'he controversy over shape needed urgent resolution, since it had critical implications for the growing number o f intrepii explorers and navigators, as witnessed earlier in the case o f Columbus. T o resolve the debate, in the early 1730s it vas suggested to the Academy in Paris that an expedition be sent to as nejr the equator as possible to measure a long arc and a second expedition slould go as near to the North Pole as possible to do the same. Two such widely separated arcs (figure 9) should prove conclusive: if the length o f 1 o f arc was greater near the North Pole than it was near the equator, then Newton would be proved correct; if the opposite, then the Cassinis would be right So it was that groups o f Academicians went to Peru (now Ecuador) and to the borders o f Sweden and Finland in southern Lapland. Several famous people were among these parties: Pierre Bou^uer and Charles-Marie de La Condamine went to Peru in a party that induced two Spanish naval officers; Pierre-Louis de Maupertuis, Alexis-Claude Clairaut, and Anders Celsius went to Lapland.

North Pole

Oblate spheroid

10. Relation of Oblate Spheroid to F ittin g Sphere

I heir results convincingly proved Sir Isaac Newton to be correct and the I ,i >N i|)is wrong. As Voltaire said at the time about Maupertuis, II avait aplati hi l i n e et les Cassinis. Maupertuis in a letter to the British astronomer |inn s Bradley wrote, Thus Sir, You See the earth is Oblate, according to the \i 111.11 Measurements, as it has been already found by the Law o f Staticks: m il this flatness appears even more considerable than Sir Isaac Newton IIit > i i^l 1 1 it. Later Voltaire commented to returning members o f the Peruvian i s| < diiion, You have found by prolonged toil, what Newton had found wllliout even leaving his hom e. I hr Iwo main radii o f the earth calculated from the results o f these expedi tions were 6376.45 and 6355.88 km. In terms o f the length o f 1 o f arc, this inn's .i difference o f only 1350 m in 111 km. It is not surprising that the errors inherent in the equipment and techniques o f the Cassinis swamped the small uiioimt they were trying to resolve. Figure 10 shows more recent, but only ilit lilly different, values and indicates the relationship to a fitting sphere.

Mason and Dixons (1766) Method of Arc Measure

MASON AND DIXON I ii illustrate the complicated nature o f comparisons between arcs because o f ilillcrcnt units o f measure, let us consider the arc in the United States mea.iiird in 1768 by Charles Mason and Jeremiah Dixon. (This is notxhe famous I ii hindary line between Pennsylvania and Maryland, but the arc o f the meridi.m, which they measured at the same time.) They found 1 to be 363 763 English feet at a mean latitude o f 3912'. Now .ill other arcs for comparison were in French toises. Thus, to quote from ( Diaries Huttons Abridged Philosophical Transactions,
T o reduce the measure o f a degree to the measure o f the French toise, it must he [remised, that the measure o f the French foot was found on a very accurate comparison made by Mr Graham, o f the toise o f the Royal Academy o f Sciences it Paris, with the Royal Societys brass standard, to be to the English foot as 114 to 107. . . . , which, divided by six, the number o f feet in a toise, gives the length o f the degree as = 56 904% Paris toises. . . . Such is the length o f a degree in this latitude supposing the 5 ft brass standard made use o f in the measure to have been exactly adjusted to the length o f the Royal Society brass standard. It was really adjusted by Mr Bird by his accurate brass scale o f equal parts . . . and which is just 1/1000 th part o f an inch shorter than the Royal Societys standard. . . . H e then finds other reasons to amend the result by a further 3 feet or a total o f 13 feet to be subtracted. But because o f other considerations that would negate this correction he concludes that the best value should be 56 904% toises!!

What all this illustrates is that the smallest uncertainty in a standard o f measure, when multiplied up to the length o f a degree, can be very significant. The figure o f 56 9041 / 2 comes from 363 763 X (107/114) X V fe. But if Birds scale was shorter by 1/1000 inch over a yard, or 1/36 000, then the whole arc would be affected by about 10 ft (P /2 toises). The arc Mason and Dixon used was not o f a design comparable to other arc measures, either o f the same vintage or any other time, so it is worth a brief description here. In terms o f figure 11 it was required to determine the length and amplitude o f line NA. Mason and Dixon first traced the meridian from N t o P, a distance o f 14.8 miles (23.8 km). The point Cwas then placed in the same latitude as P at a distance o f 2.99 miles (4.8 km). The meridian was then traced from C to D to get 5.0 miles (8.0 km). Point B was then found such that BDM was a straight line with BD = 0.28 mile (0.45 km) and B was south o f D by 90 ft (27.4 m ). BA was then traced as 81.98 miles (131.9 km). This last length was then moved proportionately to arrive at AR. Thus NA = A R + gD + D C + P N = 538 067 ft = 128'44.9" and 1 = 363 763 English feet.

------- 4 0 1 - - ---------------------------------------------------------------------------------------

1 0 0 0 km

(a)

L J A Okm B
(b)

Curved Surface and Plane Equivalent

CH A P T E R

2
Earths Shape

I he previous chapter has recounted how geodesy was born, but what is all this talk about spheroids surely all can see that the surface o f the earth is highly irregular in shape? While the irregularity is an indisputable fact, there are severe difficulties in even attempting any mathematical calculations on such a surface. For practical purposes it is necessary to find a regular solid figure that most nearly matches the topographic surface o f the earth. Remember that even the most rugged topography o f the Himalayas, at some 8 km high, is only a thin veneer on an earth o f nearly 6400 km radius. I f the topography o f a globe o f 2-m radius were shown to scale, it would be no more than 2 V2 mm thick. The oblate spheroid mentioned earlier is a refinement o f the true sphere such that its N-S radii are slightly less than those in the E-W direction at the equator. The difference between the two axes is only some 22 km, so again it is difficult to show even this correct to scale. On the 2-m-radius globe the distortion would only amount to 7 mm, which would not be discernible to the naked eye. Thus it is that all diagrams (such as figure 10) have to be grossly distorted to illustrate the situation. Even so, the topography does complicate the issue because o f its irregulari ties. Those parts o f the earth that have the most regular surfaces are the oceans. I f we imagine them extending under the continents in small frictionless channels, then the surface represented might be useful. In fact such a surface would be close to the accepted oblate spheroid and will be met again. Following this idea, the mean sea level is taken as a reference surface, and any measurements on the irregular topography are reduced to their equivalents on this sea-level surface (or one near it ). But this leads to the further question o f what is meant by sea level, as we are all aware o f the rise and fall o f the sea as we sit on the beach in the summer snn (Chapter 5 will consider the sea-level problem further.)

Now with an earth o f such shape and dimensions, if we are interested Im llu relation o f widely separated points, we need to take the shape into .......nictation. On the other hand if only small areas are o f interest, it may I* i II he feasible to revert to considering the earth flat. But what is meant by ... .ill area ? This is not readily defined; it will depend upon such factors as n li , 1 1 we wish to depict and the degree o f accuracy we seek in the answer. What might be the criteria, and can they be varied at all? These questions h .ni us into the field o f map projections (Chapter 15). However, for the .... ment, to get a feel for the situation, consider two points A and B at the i xIn mes o f an area o f interest and 1000 km apart (figure 12a). T o consider |i.iil o f the earth flat we could in effect imagine a sheet o f paper touching ih< globe in the vicinity o f the area o f interest. I f it were possible to look at this situation from outer space (that is, from near infinity), the two points would appear to be 998.95 km apart, whereas their true separation would be 1000 km, a distortion o f more than 1 in 1000. I f this were taken to the i \1 1 emes o f 1. A and B only 10 km apart (figure ] 2 b ), then the difference is a few milli meters. 2. A and B as points at opposite ends o f a diameter at the equator (figure 12c), then the apparent distance ABwould be only % o f the true distance. But what o f the effect o f elevation above mean sea level? In figure 13 imagine three plateau areas each o f which subtend the same angle d at the ( enter o f the earth but which are at different elevations. One o f the many positions where the three distances are represented as being the same length (since the angles 6 are the same) is their equivalent representation at sea level. For all mapping purposes this type o f reduction is used. If the line AB is 2000 m long and 1000 m above sea level, the equivalent distance at sea level would be 1999.68 m. While the difference in this example is relatively small, but not insignificant, it can obviously become large over long distances. Although by the 1730s the earth was recognized to have a flattened form, refinements have continued to be made to get the best values for the parame ters o f the spheroid. Since the Peru and Lapland arcs there has been a steady stream o f scientists, astronomers, and surveyors who have analyzed various observations to produce new values for the parameters and so get their names into the annals o f survey history. Each successive rvaluation has been named after the relevant personage, and some, like Everest and Clarke, have sev eral claims. Now, whenever a country has decided to carry out national survey work, it has selected a particular set o f earth parameters for the computation. While usually it has chosen the best value known at the time, or maybe one dictated by a national determination, subsequent parameters are probably much better.

P' a = semi major axis b = semi m inor axis F = fla tte n in g = ^ - = r P P ' = axis of revolution of the earth e llipsoid 14. Elements of an Ellipse

30

I lo w ever, o n c e a c h o ic e has b e e n m ad e, so> m u ch is then d e p e n d e n t o n it that later ch an ges to b e tte r param eters are u n likely. F o r e x a m p le, G re a t B ritain was active in n a tion a l survey w o rk in the first h a lf o f the 19th cen tu ry an d so selected th e values o f th eir a stro n o m er royal, Sir G e o r g e A iry. S u bsequ en t w o rk has in d ic a te d that the c h o ic e was n o t a g o o d o n e; n everth eless it con tin u es to b e u sed since it is n o w the basis o f all o f n a tion a l m a p p in g.

EARTH PARAMETERS
The p aram eters by w h ich a s p h e ro id (e llip s o id ) is d e fin e d (fig u r e 14) are the values o f the tw o m ain axes, w h ere a = th e s em im a jo r axis o r e q u a to ria l radius,

b = the s e m im in o r axis o r p o la r radius.

A ltern a tively th e fig u re is fu lly d e fin e d by a and th e rela tio n sh ip {a b)/ a, w hich is ca lled the fla tten in g and d e n o te d b y / A ll m o d e rn calcu lation s p la ce the valu e o f / c l o s e to 1/300. T h e fo llo w in g list gives a selection fr o m th e n u m erou s values available and indicates som e o f the areas in w h ich they h a ve b e e n a d o p ted . Year 1830 1830 1841 Name Airy Everest Bessel a(m ) 6 377 563 6 377 276 6 377 397 b( m) 6 356 257 6 356 075 6 356 079 W here used Great Britain India, Pakistan, Burma Germany, Indonesia, Netherlands, Japan, N.E. China Australia USA Cen., S., and W. Africa; France

1// 299.325 300.802 299.153

1858 1866 1880 1907 1909

Clarke Clarke Clarke Helm ert Hayford

6 378 294 6 378 206 6 378 249 6 378 200 6 378 388

6 356 618 6 356 584 6 356 515 6 356 818 6 356 912

294.261 294.978 293.466 298.300 297.000

1927 1948 1960 1966 1967

NAD 27 Krassovsky Fischer WGS 66 IU G G

6 6 6 6 6

378 378 378 378 378

206.4 245 155 145 160

6 6 6 6

356 356 356 356

863 773 760 775

1972 1980 1983 1984

WGS 72 International NAD 83 WGS 84

6 6 6 6

378 378 378 378

135 137 137.0 137

6 356 751 6 356 752

USA; from 1924 used internationally; Europe, N. Africa 294.978 698 2 298.300 Russia, Eastern countries S. Asia 298.3 298.25 298.247 Internationally; W. Europe, S. America, Australia 298.26 From satellite geodesy 298.257 Recent best fit to geoid 298.257 222 101 298.257 223 563

31

D
(a) Oblate Spheroid B

15. Spheroids and E llipsoid Defined

It is noticeable that, whereas only 50 years ago new determinations could differ from previous ones by hundreds o f meters, today the modifications are o f the order o f 1 to 2 m only. Before accepting any o f the above values for practical application, the user should contact the appropriate documentation office o f a country so as to ascertain any local modifications to the numbers.

SPHEROID AND ELLIPSOID Two confusing terms will occur in the context o f the earth figure spheroid and ellipsoid. Is there any difference? Certainly they tend to be used inter changeably. Mathematical reference books define them separately: a spheroid is a solid generated by rotating an ellipse about either axis, while an ellipsoid is a solid for which all plane sections through one axis are ellipses and through the other are ellipses or circles. However, if any two o f the three axes o f the ellipsoid are equal, the figure becomes a spheroid, and if all three are equal, it becomes a sphere. In the case where two o f the axes are equal, the strict definition is then o f an ellipsoid of revolution. The dictionary suggests, however, that the term spheroid can also refer to any slightly nonspherical shape but not necessarily a mathematically definable one. I f an ellipse is rotated about its minor axis, it produces an oblate shape, and if rotated about its major axis, it produces a prolate shape (figure 15). Thus in figure 15, for oblate there is rotation o f BDO about BO and for prolate rotation o f ADO about AO. In the ellipsoid, when EDF moves parallel to AOB, it produces ellipses o f varying size. Though modern texts use the terms in an arbitrary manner, in geodetic circles there is a tendency toward accepting the terms ellipsoid and ellipsoid of revolution. As you can see from the table on page 31, there are many ellipsoids. The first international figure was developed in 1924, but today the advanced technology o f GPS and related systems allows the use o f various world-based figures (see Chapter 10). Even so, the most acceptable figure, DOD WGS 84, differs from the geoid (see next section) in the range o f +60 to 1 0 0 m.

Scale of variations

16. Relation of Geoid to Ellipsoid (A fte r King-Hele, 1 9 63)

GEOID
In the same way that lack o f accurate observations delayed recognition o f the earths nonsphericity, so it was not until the advent o f artificial satellites and measurements from space that the standard o f refinement o f collected data would be sufficient to allow scientists to move confidently away from a biaxial figure o f reference. Now an even more complicated figure may be the case (figure 16), one with a slight flattening in the northern middle latitudes and a bulge o f similar dimensions in the southern middle latitudes. At the time o f this discovery, some likened it to a pear shape, but there are those who vehemently oppose the use o f such a description, particularly because the deformation is so slight only 2 0 m or so in relation to the overall size. The term most appropriate to the irregular earth surface is a geoid. It does not have a mathematical representation, but it is the equipotential surface at mean sea level, which means that at any point it is perpendicular to the direction o f gravity. Since that can be affected by variations in the earth structure, the geoid, while a smooth surface, is undulatory. It is not a fully known surface because o f the sparseness o f gravity observations over large areas o f the earth, particularly the oceans. It is the dimension o f the equipotential surface at mean sea level that is referred to when quoting the values o f the semimajor and semiminor axes. Many have tried to represent the shape mathematically, but even to get somewhere near requires a series with 18th-power terms. The Defense Map ping Agency developed a mathematical model with terms to the power 180 with 32 755 coefficients. The average (or even the above-average) user can with confidence usually find an alternative way around any problem that calls for manipulating expressions o f that magnitude. In 1993 a new geoid model, Geoid93, was introduced for Puerto Rico, Hawaii, and the Virgin Island. Tests indicated that it was a noticeable improve ment on Geoid90. More such figures will no doubt follow.

35

3 .2 7 cu b its = 0 .0 1 stades

, ______d_____

d = S cu b its = t stades

2 1 6 Barley corns = 72 inches

, ______d_____

d = z barley corns = z/3 inches

2 yards = 6 ft

, ______d_____

d = q standard yards = 3 q feet

0.9387 toise = 5.632 Paris feet = 811 lignes

, ______ d______ , ^ ^

d = r toise

= 864.000r lignes

1 .8 3 0 m

|-<-------------------- d = p x

10M

quadr ant

2 8 4 2 2 8 9 .6 A of red cadm ium = 3 0 2 0 8 9 7 .6 A o f krypton-86

i ^

_ , H

d = y wavelengths

1 /1 6 3 8 2 1 0 1 5 secs = 6 . 1 x 1 0 - 9 secs

, d , r< --------------------- H

, d = x seconds

Where next? 17. U nits of Length

CHAPTER

3
Units of Measure

DISTANCE MEASURES It is one thing to mark points o f interest on the earths surface but quite another to determine how far apart they are. In what sort o f unit can the distance be recorded? We are very familiar with the foot as a unit o f measure, but in the past numerous lengths o f the foot, varying by large amounts, have been recognized in different countries. An even better example o f an inconsistent unit is the mile, where a Swedish mile was some seven times the length o f an English mile. Thus considerable care has had to be exercised over the centuries to ensure that the particular unit in which a quantity was measured was properly defined. Even so, numerous confusions have arisen. In earliest times a journey was measured by how long it took to cover the ground; this might be so many camel-days or sailing-days. By the time o f Eratosthenes, for example, the distance between Alexandria and Syene was given in terms o f the stade. But which particular form o f the stade did he use, and what would be its equivalent today? Historians quote different values varying from 157.5 to 210 m, but even these are only approximate equivalents.

F o o t and T o is e
Both the foot and the toise (the forerunner o f the meter) can be traced back to the cubit, (defined as the length o f a forearm ), some values o f which are known to have been in use in 3000 B.C. (figure 17). Chapter 1 mentioned Arabian miles, Italian miles, and Chinese li, but

---- --

---------------------------------------------------------------------------------

numerous other units were also used historically. None o f these can be given accurate modern equivalents, although attempts have been made to assign unrealistic accuracies to some relationships. O f the early units the first to be physically represented in any reasonable form were the yard in England and the toise in France. As early as a .d . 960 the offical standard yard was defined by King Edgar as the length o f an hexagonal brass rod kept at Winchester. King Edward I (1239-1307) introduced the first table o f relations between measurement units. It started with 3 grains o f barley, dry and round = 1 inch and with 3 feet = 1 ulna (represented by a bar o f iron made in 1305). The year 1490 saw a new yard standard represented in an octagonal bar introduced by King Henry VII. In 1588 a !/ 2-inch-square brass rod standard yard was approved by Queen Elizabeth I (1533-1603), based on the previous standards, and this remained standard until 1824. However, just browsing through papers on historical metrology highlights many different types o f foot. The royal foot can be found quoted as equal to 13.75 present inches, the Amsterdam foot as 11.41 inches and the natural foot as 9.8 inches. The time o f the building o f Stonehenge is probably as far back as we can go with any degree o f certainty. Sir Flinders Petrie, investigat ing in 1880, found the inner diameter o f the circle o f faces o f the sarcen stones to be 1167.9 inches, which is almost exactly 100 times the Roman foot o f 11.67 inches. This foot was descended from the royal cubit o f 20.63 inches. The French, not to be outdone, had a foot measure descended from the Assyrian cubit o f about 21.6 inches. This cubit later became the Hashimi cubit, a half measure said to have been introduced into France by Charlemagne in around a .d . 790. It was then known as the pied du Roi and survived for 1 0 0 0 years. The pied du Roi o f 12 pouces equates to 12.79 present inches. Six pied du Roi became the toise o f 6.395 British feet, a unit that was later used to measure both the northern (Lapland) and southern (Peru) arcs and other arcs o f the 17th and 18th centuries. Its most notable physical representation prior to the expeditions was as an iron rod set in the stairs o f the Chatelet in Paris. From the middle o f the 17th century there was a growing realization o f the need for a universal unit o f measurement. The problem was how to define such a unit. It needed to be in a form that could be readily reproduced, but at that time there was no obvious candidate. A pioneer in this field was Gabriel Mouton, a vicar from Lyons, who in 1670 proposed that a minute o f arc o f a great circle be adopted as a universal unit. Mouton, like other prominent scientists o f those times, was a Jesuit priest who treated astronomy as a hobby. H e used the work o f another Jesuit, John Riccioli, to illustrate his idea. In essence it consisted o f the following relations:

1 inilliar 1 1 1 1

1 centuria

= 1 0 0 0 virga = 1 0 0 virga decuria = 10 virga 1 virga = 0.001 minute o f are decima = 0.1 virga centesima = 0.01 virga millesima = 0.001 virga

The decima was to have the same length as a pendulum that made 3959.2 use illations in 30 minutes at Lyons, but nothing came o f these efforts. In England Sir Christopher Wren had similar aspirations for a universal unit and was considering the length o f a half-second pendulum in London .is a possibly appropriate unit. Over the English Channel, the toise du Chatelet had become some 5 lignes (where 1 ligne = V12 inch) too long through usage and wear. In 1671 lAbbe licard, while occupied in restoring the toise, worked along the same lines and suggested that a universal foot should be Vz o f the length o f a second pendulum (see Chapter 11) or rayon astronomique at the Paris Observatory. I his was 36 pouces, 8.5 lignes in length where the Paris pouce (inch) was equivalent to about 1.065 present inches. A similar pendulum suggestion came in 1673 from the Dutch astronomer Christiaan Huygens. Th e scent went cold for some 50 years until Jacques Cassini II in 1720 reverted to the notion o f using the length o f a minute o f arc by suggesting that a geodetic foot be 1/6000 o f a terrestrial minute o f arc. This would have been very close to the approximation used in some present-day calculations o f l"(second o f arc) = 100 ft. Meanwhile the toise was reconstructed specifi cally for the expeditions o f 1735 to Peru and Lapland as the toise du Perou and the toise du Nord, respectively. They were in the form o f a flat iron bar o f 864.000 Paris lignes (where 1 pied du Roi = 144 lignes) or 6 Paris feet.

Meter As a result o f the Peru expedition, La Condamine recommended in 1747 that 1he equatorial seconds pendulum should be used to define a standard. This he found to be 439.15 lignes. He was emphatic in saying that it should be an international unit and that perhaps a group o f academics from various coun tries should investigate the problem. In 1789 a conference o f delegates from numerous French towns petitioned Louis XVI to unify the system o f measures. Th e year 1790 saw the propo sal o f a new unit to be called the metre, based, yet again, on the seconds pen dulum. Th e impetus for a start came from Prieur Du Vernois, later known as

C.-A. Prieur de la Cte-dOr. On February 9,1790, he presented tu the National Academy a memoir on the necessity for a uniform set o f measures. This occurred during the difficult times o f the French Revolution, but he was adept at the politics necessary to promote his ideas without risking the guillotine. He proposed relating the unit to the length o f the seconds pendulum at the Paris Observatory. A platinum bar deposited at the City Hall would represent the length. One-third o f its length would represent the national foot, which in turn would divide into 10 inches, each o f 10 lignes. Ten feet would equal 1 rod. Charles Maurice de Tallyrand, bishop o f Autun and later president o f the National Assembly, put Prieurs ideas to the Assembly. In doing so he suggested using the length o f a seconds pendulum at 45 latitude so as to be independent o f any particular location. Around the same time, both the United States and Great Britain were investigating the possibilities o f a ne\V system o f measures. As a result it was suggested that the British and French parliaments, the French Academy o f Sciences, and the British Royal Society should cooperate in creating the new system, but this did not materialize. The members o f the committee o f the French Academy o f Sciences includ ing such well-known names as Jean Borda, Pierre Laplace, Joseph Lagrange, Gaspard Monge, and Marie de Condorcet. Their first report in March 1791 particularly emphasized that on no account should any existing unit o f mea sure be used as the new standard since it would only generate national pride in the lucky country but hurt in every other country. Am ong the methods they considered were:
1 . a seconds pendulum at the equator 2 . a fraction o f the circumference o f the equator

3. a fraction o f a meridian quadrant They rejected the first because o f the influence on it o f time and gravity variations and the second because o f the difficulty in measuring along the equator. That left only the meridian arc. It was decided that this would be derived from a new meridian measured from Dunkirk to Barcelona that would pass through the Paris Observatory. The National Assembly approved the Academys plan in March 1791 and as a result the Academy established five commissions to investigate triangulation, observation o f latitude, baseline measurement, length o f the seconds pendu lum, comparison o f units, and standardization o f weights. On the eve o f his flight from Paris, the king discussed the projects with the commissioners and in particular with Cassini, whose father and grandfather had previously measured a chain o f triangulation along the meridian through Paris. Cassini pointed out that the instruments avail-

u i m a u i iv it ja u ie

able to these new efforts were much improved over those used by his an cestors.

F ie ld M e a s u r e m e n t s
A year was spent on the production o f suitable equipment, among which was the celebrated Borda circle and other equipment constructed to his design. By the end o f 1792 Jean Delambre had started on the section from Dunkirk south to Rodez, and Pierre Mechain was beginning in the Pyrenees, an area he had not previously visited. As a war was still in progress, life was not easy for the surveyors. They were looked upon with suspicion by villagers and on many occasions were arrested as spies. Measurements for the 939' arc took until 1798 to complete. While the survey work was still in progress, the Third Decimal Committee was anxious to establish a provisional unit. A report from the Academy to the convention in May 1793 recommended dividing the new units decimally, giving the units simple and internationally intelligible names, and assigning the meter the provisional length o f 443.443 Paris lignes. This last was based on l Abbe Nicholas La Cailles value o f 57 027 toises for 1 o f the meridian at 45 latitude. A brass standard o f this length was produced by Jean Borda and Mathurin Brisson. However, when the arc was finally completed, the commission gave the relation as 1 meter = 443.295 936 Paris lignes at 0C, and a new scale was ordered from Lenoir. This gave 1 French foot = 0.324 839 meter. The precision with which Delambre and Mechain measured their arc was about 5 X 10~4 . The metre des Archives that resulted from the arc measures was made in 1799 to a precision o f about 5 X 10 5, or about one order higher than the actual arcs. It was around 1850 that the precision o f calibrations improved by another order o f magnitude, and the 1880s before 10 7 was reached.

D e f in it iv e V a l u e
When the results became available the assumption was made that the earth was an exact spheroid and that the ellipticity, dimensions, and quadrant length should be determined from this new arc together with that o f Bouguer and La Condamine in Peru. The result gave the ellipticity as 1/304 and the meter as 443.296 lignes. The data they used to achieve these values were:

Arc length (toises)

Amplitude

Mid. lat.

for France: for Peru:

551 584.7 17(i 873

940'25" 3 07 01

46 11'58" 1 31 00

The approximate solution can be found thus: 1. I f 2A and 2 B are the sum and difference o f the semimajor and -minoi axes, then neglecting powers o f B/A we get R = A - 3B cos 2 4 > 2. Integrating from 0 to 77/2 gives the length o f the quadrant: Q = An/2 3. From the radii at the two latitudes 4 > \ and < f>2 there are two equations from which B can be eliminated. Then A = e + at cot a cot 8 where 2 e and 2 o > are the sum and difference o f the radii, and a and h are the sum and difference o f the mean latitudes. 4. From these data the curvatures o f centers o f the two arcs are 3 266 977 and 3 251 285

Whence 8 = 3 259 131 and w = 7846. 5. From this we get A = 3 266 345 and multiplying by 77/2 and dividing by 107 gives the meter as 0.513 0766 parts o f the toise du Perou (figure 17).
6 . Since the toise = 864 lignes, the meter became 443.296 lignes.

T h e Le ga l m e tre
The metric system was legalized in France on April 7, 1795, and resulted in the expression Legal metre. Remember however, that the meter was at first defined only as a certain proportion o f the toise du Perou and did not actually materialize until Bordas metre des Archives. At 0C this defined value was 443.296 lignes o f the toise du Perou when the latter was at 16.25"( that is, 443.296/864 became the old French legal meter. This gave tlx- i l lation

1 toise = 1.949 0366 m

or

1 m = 0.513 0740 toise

Since at the outset the meter was not represented by any bar, the Toise i In IVrou remained the national material standard. That is, initially the meter w.is simply defined by law, so when countries wanted copies, they were issued 111| ii< s o f the Toise du Prou. 1 1 1 1798 Borda made the metre des Archives, which was a flat bar o f impure platinum in the form o f an end standard. When it was found that the length ni die meter was impossible to redetermine accurately in terms o f its physical li.isis, the metre des Archives became the basic standard. Actually this meter w.is one o f three platinum standards and several iron ones made at this lime. At about the same time the French Academy considered the toise du Prou unsatisfactory, so it instructed Borda to prepare new bars. He made four bimetallic rules o f strips o f copper and platinum. The length was a double ii use o f about 12.8 English feet, N o I o f which became known as the module uid was regarded as replacing the toise du Prou. On (une 22, 1799, the new meter standards were formally presented to the Corps Lgislatif and legalized by a statute that abolished the provisional standards. Hence from about 1800 there were two standards in existence in I'Vance the toise and the meter. In the 1810s there was much mistrust o f the new units o f measure. Napoleon lound the popular disregard for the new system so great that he sanctioned . 1 parallel system that applied the old names and fractions to the new units. In 1812 there came a French imperial decree for facilitating and accelerating i lie establishment o f a uniform system o f weights and measures in our Empire. I lie decimal division was abandoned in favor o f binary and duodecimal divi sions o f metric units. Soon, however, the decree o f 1812 was repealed, and on July 4, 1837, the use o f any units other than decimal ones became an offense with effect liom January 1, 1840. In 1841 Friedrich Bessel gave the earth quadrant as 10 000 856 m, so that already the indefiniteness o f the standard was be<oining evident. By 1857 Alexander Clarke was quoting a quadrant o f 10 001 983 m. Despite the introduction o f the meter, standard yards and other units per sisted. In 1864 a committee was form ed from delegates at an international exhibi tion in Paris to bring uniformity into measures. This goal was further pursued .it the Geodetic Association meeting in Berlin. In 1869 the French government invited foreign governments to attend a conference in Paris on length standardization. The year 1870 saw the first meeting o f the International Metric Commission, and in August o f that year 1 1 was resolved to establish the Bureau Internationale des Poids et Msures

(B IP M ) and to construct a number of precise line standards .is similar as possible to the metre des Archives. From here the metei gradually began to gain international acceptance, although, as we all know, there are still those who are not happy using either it or its relatives. Various intercomparisons o f standards and manufacture o f copies took place and in 1889; 30 copies o f the prototype meter were distributed to different countries. For example, numbers 11 and 28 went to Russia, number 16 to the United Kingdom, and numbers 21 and 27 to the United States.

A N a tura l S ta n d a rd
By 1890 Sir David Gill was agitating for a natural standard and suggested the wavelength o f sodium vapor as a possibility. There were other suggestions before Albert Michelson and Justin Benoit in 1893 defined the meter in terms o f a specific number o f wavelengths o f red cadmium radiation (figure 17). A comparison o f the meter with the yard in 1894-1895 remained the U.K. standard for the meter until 1964. In 1946 the Empire Scientific Conference redefined the yard as exactly 0.9144 m. Prior to this time the U.S. inch was 25.400 05 mm and the British inch was 25.399 93 mm. In 1960 the meter was redefined in terms o f a wavelength o f light as 1650 763.73 A o f krypton-86 gas in a vacuum (figure 17). A line in the orange part o f the visible spectrum corresponding to the unperturbed transition between 2Pio and 5N 5 levels. This allowed a comparison o f standard lengths that was accurate to 4 parts in 109 . In 1983 the General Conference on Weights and Measures meeting in Paris redefined the meter yet again, this time in terms o f a time standard. It is now the length o f the path traveled by light in a vacuum during a time interval o f 1/299 792 458 o f a second. Notice that this number is the reciprocal o f the accepted value o f the velocity o f light. Comparison is now possible to better than 1 part in 1 0 '. Thus the 20th century has seen several shifts in defining units. In the 1940s accurately measured distances were used to determine the velocity o f light. Later this was turned round so that the velocity could be used to determine distance. W e now have distance defined in terms o f time.

ANGLE MEASURE In parallel with the introduction o f the meter came the gon, a decimalized angular unit. A right angle contains 100 gons; hence the full circle is 400g. Instead o f the sexagesimal (60) minutes and seconds, a gon subdivides into 100c ( centesimal minutes), each o f 100 (centesimal seconds). Thus an angle will be written as 379.4582s or 379R45c 82 . Note that most modern calculators incorrectly call this unit a grad or grade.

AA

Introduction o f the gon does not mean that the sexagesimal degree system li r. gone. It is still used in parallel with the gon system. When you think about ii sexagesimal divisions affect not only angles but also time, latitude, and li tngilude, so any move to sole use o f the gon would have enormous repercusii is in measurements.

18. Questions about the Earth

CHAPTER

4
Traditional Survey Positioning Techniques

^ ) n c e knowledge is available on the size and shape o f the earth, the next question is Where are we on that earth? or Where on earth are we? (figure 18). One o f the main functions o f geodesy is to determine the exact positions o f points on or near the earths surface. This includes points below the surface for mining or tunneling, under the oceans for underwater engineering structures or oil wells, and above the surface for rocketry or radio telescope positioning. Recent years have seen the traditional techniques supplemented by new ones. It would now seem to be only a matter o f time before the older methods are fully superseded by satellite methods and the need for permanent survey marks disappears. This is not to suggest that all marks would be destroyed or let fall into disrepair since undoubtedly there would still be small operators who would require them. This chapter will concentrate on the time-honored approaches, and later chapters will cover the newer techniques. While there is interest in all three spatial dimensions, it is convenient to separate the horizontal positioning from the vertical element since in traditional survey work they tended to be separate operations. With the newer methods this is not so since the data used can normally solve for all three dimensions at the same time. Horizontal positioning methods include 1. astronomical techniques 2. triangulation 3. trilateration 4. traversing

while positioning in the vertical plane includes 1. precise or geodetic leveling, 2. trigonometric leveling 3. barometric leveling 4. echo sounding

47

Zenith

North Pole

(b) 19. Celestial Coordinate Systems

HORIZONTAL POSITIONING Astronomical "pchniques From the very early days o f civilization Pc heavens have been used to aid positioning. Until the last few centuries tie N-S component, or latitude, was far easier to find than was the E-W component, or longitude. While the latter had to await the invention o f good chrtfiometers getting an indication o f lalitude was relatively straightforward usin,' simple stellar observation. Initially astronomy was used for marine navigation later explorers found it very useful in uncharted regions. Then the geodesis used it in combination with other measurements to establish precise position on which to base national surveys. Single isolated fixes by astronomical lethods were o f little value to the surveyor, because they could not be intefelated. But taken at positions that were connected by other means, they wee o f extreme importance. From early times it was recognized thatthe stars appeared to rotate around the heavens and that there was a star near o the apparent hub o f that rotation. Even before the idea o f an Equator was filly appreciated, it was observed that by traveling north or south the elevation >f this hub star (Pole Star) changed by amounts that could be related to the listance traveled. Unfortunately for the surveyor, as well as the navigator andsxplorer, there is no equivalent star visible in the Southern Hemisphere thit gives a direct approximation to the position o f the South Pole. The newest star there is er Octantis in the constellation known as the Southern Crds, but its relationship to the pole at any time requires computation. In the early centuries after Christ, streiuous efforts were made to catalogue as many o f the heavenly bodies as possibt in terms o f their relative positions. Primative observatories came and went the importance o f detailed knowl edge o f the movements and positions o f t h e stars and planets came to be ap preciated. Am ong the methods used for describing the positions o f the stars on a celestial sphere arc those shown in fig me 19: 1. by right ascension (R A ) and declination (8) (figure 19a) 2. by hour angle (H A ) and declination (5) (figure 19a) 3. by azimuth (A z) and altitude (a ) figure 19b) where right ascension = the angle along tie celestial equator from an initial point called the firit point o f Aries (y ) to the declination circle o f the body declination = the arc between the body and the celestial equator

True north

(a) Azim uth

Grid or m agnetic north

Grid or m agnetic north

(b) Bearing

1432'29"

T heodolite circle Angle A O B = a = D irection B - D irection A = 1 0 8 2 4 '1 1 " - 1 4 3 2 '2 9 " = 9 3 5 1 ,4 2 "

rection 20. A zim uth, Bearing, and Direction

hour angle a/mmth .illitude

= the angle between the plane o f the meridian through Greenwich and the plane through the body = the horizontal angle along the horizon plane between the meridian o f the observer and that through the body = the vertical angle from the horizon to the plane o f the body

< )l the three systems, the first is used in tabulating star catalogues since it In independent o f the observers position. The other two systems are used for In Id observations in conjunction with determining point position and orienliilio n .

I he 17th century saw the formation o f important scientific bodies, such as

11 n Academy o f Sciences in France and the Royal Society in London, to

in ourage these among numerous other aspects o f science. It also saw the Inundation o f the Royal Greenwich Observatory in 1675 under warrant from l ing Charles II and with John Flamsteed as astronomical observator, and ulluT major observatories on the continent o f Europe. These were great iiiinuli to production o f accurate tables o f the stars (ephemerides) and to improvements in the methods o f recording time. Astronomical methods for positioning yield results essentially in terms o f l.iiilude and longitude, although they may at a later stage be converted into . 1 rectangular coordinate system o f some form. With these two elements we ,ii c also normally interested in the azimuth o f a particular line, which can be Iniind from similar observations. A/.imuth, bearing, and direction Three terms need to be distinguished (fig ure 20). The azimuth from point A to point B is the clockwise angle measured .it A between the plane containing the North Pole and the plane containing It. There are subtle differences between astronomical azimuth and geodetic a/.imuth. The form er relates to the geoid, while the latter relates to the adopted ellipsoid. The term bearing is normally retained to indicate the clockwise angle from id north when working in terms o f rectangular coordinates. Direction is used to denote the circle reading on an instrument, such as a theodolite, and the difference between two recorded directions gives the angle il that point. We must accept, however, that, while it is preferable for all nations to use the same definitions and terminology, it is a fact o f life that neither these nor some other survey terms are universally exclusive, and so should be interpreted in the context o f the publication in which they appear.

21. D eflection of Vertical at Point P

I rue north, grid north, and magnetic north Th e previous section made mention o f grid north. Surveying uses several different norths, so the differ ii( < \ s need to be known. /ru e north is the direction from a point toward the north end o f the axis I rotation o f the earth, and is nearly equal to the direction o f the Pole Star i ii North Pole. (hid north is the north direction o f a set o f rectangular coordinates and limy (though more likely will not) be coincident with true north. It can be a < 11111 < - arbitrary direction. Magnetic north is the direction indicated by a magnetic compass. It is not ci instant, and if shown on maps is usually accompanied by an indication o f its relation to grid north and the amount by which it is changing annually. l atitude There are many ways the surveyor can determine his latitude through astronomical observations aside from that already indicated, where the altitude o f the Pole Star closely approximates the required result. It follows from this that if the altitude (the vertical angle from the horizon to the body) is measured to a star or the sun and we know the position o f the body (because it has been catalogued) in relation to the celestial pole, then we can determine the latitude. For the best results the observations i< ( |iiire correction for refraction. I f it is the sun being observed, then allowance is needed for its semidiameter; that is, when viewed through a theodolite tele scope, the sun has a finite diameter that fills a large percentage o f the field o f view, whereas a star can be considered as a point. Am ong the alternative methods is the use o f circummeridian observations. A series o f readings are taken to either the sun or a star over a period from some 10 minutes before until 10 minutes after it has crossed the meridian. With accurate knowledge o f the time o f each observation, corrections can be applied to get the equivalent values at the meridian. With suitable observing techniques, a high degree o f precision is possible. Longitude For centuries longitude proved to be a much more difficult nut to crack. Early explorers and navigators made errors o f many degrees in their longitude position because o f the methods they had to adopt. Longitude is essentially a function o f time a quantity that could not be measured suffi ciently accurately until the second half o f the 18th century. Although the ancient Chinese had their water clocks and other cultures their sand glasses, these were not timepieces that could be transported from place to place. Positioning required knowing the difference in time between a datum position nowadays this is the Greenwich meridian, but in the past many other positions have been used, including Paris and a point in the Canary Isles and the point in question. Before the introduction o f radio signals, the only way o f getting the time from A to was to physically carry it in the form o f a timepiece. W hile early c locks had reasonable accuracies when stationary, they were completely upset

(a) Tropical and Sidereal Year Defined

(b) D ifference between Apparent Solar Time and Mean Solar Time 2 2 . D ifferent Concepts of Year

Traditional Survey Positioning Techniques

when taken on board ship. So much so that a reward was offered by the government for a solution to the longitude problem in effect for building a timepiece that would keep accurate time on a long sea voyage. Other solutions were suggested, but were not at all practicable. It was John Harrison who produced the first acceptable chronometer in 1765 and who, after a considerable struggle, managed to claim the reward. Nearly another 100 years were to elapse (1852) before it was possible to send time signals by electric telegraph, and it wasnt until the turn o f this century that wireless telegraphy was introduced. Nowadays it is possible to receive radio time signals throughout the world to a very high degree o f accuracy. When considering time and longitude it must be remembered that 1 second o f time is 15 seconds o f arc, or around 0.5 km at the equator. Th e difference in longitude between two places is then the difference between the times at those places at the same instant. This is the same as saying the difference in time between a particular star being directly overhead at the Greenwich merid ian and the same star directly overhead at the point. By using star or sun observations the local time at a particular point can be found, and the requirement is then to find the local time at the reference station at the same instant. This can be done using either time signals or a portable chronometer. Azimuth can similarly be found by several astronomical methods. Since the stars appear to circle with the Pole Star as a hub (almost), then by noting the positions o f a particular star when it is farthest east and farthest west in other words, at elongation we can find the meridian direction, which is clearly halfway in between. Next we record the angle between the star and a ground point. Finally, we find the azimuth o f the point o f observation by noting the angle between that point and the ground point. Astronomical observations made with a theodolite or similar instrument all relate to some leveling device on the instrument. Thus, when properly adjusted, the vertical axis o f the instrument coincides with the direction o f gravity and is therefore perpendicular to the geoid (figure 21). Thus astronom ical positions are referenced to the geoid. Since the geoid is an irregular, nonmathematical surface, astronomical positions are wholly independent o f each other. I f sufficient knowledge o f the geoid and ellipsoid is available, then conversions are possible.

Tim e Modern requirements for time have lai ouisii ippcd I lie* accuracy that can be obtained from a pendulum clock. As the explosion in technology has gathered pace, so better and better timepieces have become necessary to achieve finer and finer resolutions. Tim e is o f course o f particular interest because it is also a measure o f longitude: since 360 = 24 hours, 1 = 4 minutes. Year Since the sun returns to a different position after a year, there are two different concepts o f year. Th e tropical year is the time required for the sun to return to the same position as the previous year in relation to the equinoctial points (the points where the equator intersects the ecliptic). And the sidereal year is the time it takes the sun to return to the same position as before in relation to the stars. In terms o f figure 22a, this means that the sun goes from point A to A for the tropical year from A to 1 1for the sidereal year. Thus, sidereal year = 365.242 199 mean solar days tropical year = 365.256 360 mean solar days a difference o f about 20 minutes. Where a mean solar day is the interval between two successive transits o f the mean sun over a given meridian, and the mean sun is a fictitious body, describing a circuit o f the equator in the time required for the actual sun to describe the ecliptic. The difference between apparent solar and mean solar time at any instant is given by the equation o f time, which varies between 14.25 to +16.25 minutes (figure 22b). This all begins to get confused with different types o f sun, day, and year but these stem from the nonuniform motion o f the actual sun in its apparent orbit. It is much easier to deal with a hypothetical sun moving at the average velocity o f the actual sun in other words, with the mean sun. The big problem with time is that there is such a bewildering range o f forms in which it can be expressed. When writing this text, I had no difficulty amassing a list o f 25 different forms o f time that could warrant mention. To describe them all in detail here would have little effect other than to create confusion. What it does highlight is how careful we have to be to know which system we are working in and how to convert from one system to another if necessary. Day The time unit day is defined as the interval between successive transits o f a celestial body across the same part o f a meridian. There are three versions depending whether the body is the sun (solar), a star (sidereal), or the moon (lunar). All are slightly variable. The last can be ignored here. Solar time This is the natural clock in terms o f the daily motion o f the sun. The solar day is the interval between successive upper transits o f the sun, but for various reasons this is not a constant. Using the mean solar time (a function o f the average o f all apparent solar days throughout a year) overcomes this irregularity. This gives rise to an imaginary mean sun moving at a uniform rate round the equator.

Sidereal time While the sidereal time is the hour angle o f the first point < ! A i ies, it is also equal to the hour angle plus the right ascension o f a given star. ( Weenwich time As the prime meridian is taken to be that through Green-

vvh h, England, so various measures o f time are called Greenwich time. This

may be sidereal, apparent solar, or mean solar time. Th e basic relation is that ( reenwich time = local time + longitude o f the place o f observation. I Ii is leads to the concept standard time, which is the time adopted for different hands o f longitude for everyday purposes. Whence standard time = Greenwich mean time an integral number o f hours Greenwich mean time (G M T ) is now often referred to as universal time (U T ). Allhough there is still an observatory at Greenwich, its time, denoted U T 1, is now the mean value derived from some 54 observatories around the world. A modification o f this, to allow for annual variations in the earths rotation, ^ives U T 2, where the difference between U T 1 and U T 2 varies by a few hundredths o f a second o f time over a year. Greenwich apparent sidereal local apparent sidereal time (GAST) time (LA ST) A where A is the longitude o f the point o f observation. ( 1AST can be converted to universal time, as U T 0, by using a set o f tables. Then U T 1 = U T 0 + a small reduction to the conventional international origin (C IO ) U T 2 = U T 1 + a small correction related to the polar motion It is U T 1 that the surveyor or geodesist uses when reducing his astronomi cal observations. Ephemeris time T o accommodate the higher accuracies required by dynami cal astronomy, geodesists replaced mean sidereal time with ephemeris time and atomic time (A T ). Ephemeris tim.e (E T ) is related to U T 1 via a small correction found in the Astronomical Ephemeris, where ET = U T 1 + dT The correcting term was zero in 1902 and +34 seconds in 1961, but it does not increase uniformly. A

Seconds of arc (b) Polar Motion (1 9 0 0 -1 9 0 5 ) 2 3 . Motions of E arths Axis

Atomic time Modern atomic clocks are far more stable than any previous form o f timepiece. Because o f this, they are able to detect irregularities in Ilie earths rotation. Hence the introduction in 1967 o f the concept atomic time (AT, now sometimes called the SI second), which began as the weighted mean o f eight atomic standards. It was defined that year such that 1 second = 9 192 631 770 cycles o f a caesium-133 atom under defined conditions. The atomic second is very close to an ephemeris second. International atomic time (T A I) was arranged to have a distinct relation to ephemeris time asJanuary 1, 1977, and has since replaced the latter. Coordinated universal time (U T C ) is directly related to atomic time with a periodic jum p o f 1 second to keep it within that amount o f U T 2. Other forms of time As you investigate other relevant publications it will become obvious that there are yet more versions o f time. It is not appropriate to refer to them all here, but you must follow them if youre aiming to become a specialist in geodesy. Time in context of global positioning system (GPS) Each satellite contains sev eral atomic frequency standards since the highest-possible orders o f accuracy are required. These are atomic clocks based on both caesium and rubidium. The latter is the cheaper o f the two versions. Despite their high accuracy, I hey are still subject to a steady drift away from the GPS standard. This variation can be closely tracked at the master control station in the United States and its value incorporated in the broadcast signals. The magnitude o f such drift is kept to within 0.001 second. The ground receivers normally do not have atomic frequency standards but instead use quartz crystal clocks. Nevertheless, the stability is o f a high order better than 1 part in 101 0 , which means better than 1 second in over 300 years. Today this can be bettered to even 1 in 101 5 , an accuracy almost beyond comprehension. Remember also that since the velocity o f light is 300 000 km/second, 106 second ( = 1 microsec) is equivalent to 300 m. Hence, to get down to accuracies o f centimeters, resolutions to better than 10~1 0 seconds are essential. While it is not possible to sychronize the various clocks to this accuracy, there are ways to reduce the effect o f such synchronization errors. Terms Those small troublesome terms The earths axis is, unfortunately, not fixed in space, and neither are the positions o f the apparendy fixed stars. These effects result particularly from the fact that the earth is not a true sphere. As a result, small correcting terms are required, each amounting to a few seconds o f arc. T o complicate the issue still further, the corrections are not constant but vary with time. A brief description o f each appropriate one follows, although their effect will often come built into any software package the geodesist uses.

_____________________________________________________

(a) Apparent Change: M to AT

(b) Actual Changes: y to y ' 24. Precession

Aberration James Bradley spent a year starting in December 1725 observing a particular star through a fixed telescope. He found that it seemed to move with time in such a way as to complete a cycle in a year and moved as much as 20" o f arc from its original position. He tried various hypotheses including nutation (see below) to account for this change, but to no avail until 1728. It was then, after further observations o f more stars, that he hit upon the idea that the effect might result from the finite velocity o f light combined with the motion o f the earth in its orbit. The apparent oosition o f a star is determined by the movement o f the light from that star relative to the earth, which will vary from star to star. Any one star traces out an approximate ellipse around the point where it would be seen if there were no aberration. Over the years various estimates have been put on the maximum value o f aberration, all o f which fall in the vicinity o f 20.4". Working backward from his results, Bradley con cluded that the velocity o f light exceeded the velocity o f the earth in its orbit by a factor o f 10 210 to 1 and hence that light took approximately 500 seconds to travel from the sun to the earth. Since the velocity o f the earth was then known to be about 18.3 miles/sec, Bradleys values would suggest a velocity o f light o f 18.3 X 10 210 = 186 850 miles per hour not a bad estimate compared with todays knowledge. Nutation In addition to the complications caused by the continual varia tion o f the separation o f the attractive bodies (precession, see next section), there is a small oscillation o f the earths axis, called nutation. This also was discovered by Bradley from observations made over the 20-year period o f 1727-1747. H e found an irregularity in motion o f the pole that had a period o f 18.6 years and which became known as nutation. The pole was describing an ellipse with axes o f 18" and 16" (today taken as 18.42" and 13.75"), which gave the precessional movement a superimposed wavy or periodic variation. Precession Compared with observations made 150 years before his time, Hipparchus (c. 150 B.C.) found there was a small general increase in the longitude o f stars but no changes in their latitude. He estimated this change to be about 36" a year (from west to east). From this he deduced that the positions o f the equinoctial points were changing and that it must be the equator that was slowly moving. In this situation, the sun will return to a slightly different point at the completion o f each year, a position a little before that o f the previous year (figures 23a and 24). In fact, the attractive effects between the earth and other heavenly bodies, particularly the sun and moon, cause this slow movement, known as precession of the equinoxes, and which amounts to about 50.2" a year (actually 50.2564 + a small term ). Hence over the not inconsiderable period o f 25 800 years [(360 X 3600)/50.2] the celestial pole will describe a circular motion around the pole o f the ecliptic. The motion can be likened to a spinning top: while it rotates on a point, its uppermost part also rotates to sweep out a conic al shape.
___ . ------ ------------------------------------------------------------------------

Measured all angles in every triangle Lengths A B and CD

25. Triangulation Network

While Hipparchus was the first to conclude that such a change occurred, it was left to Newton to come up with the right explanation: because the sun seldom lies in the plane o f the earths equator, most o f the time its attractive force is not directed toward the center o f the earth. Polar motion Observations in Berlin 1884-1885 indicated minute variations in latitudes o f about 0.5". Dr. S. C. Chandler showed in 1891 that they could be explained if the earths axis described a circuit around its mean position in about 434 days. The displacement was never more than 10 m and has not yet been sensibly explained, but it is known as the Chandler period. It would appear that the earths axis o f rotation is moving by a very small amount in relation to the earths crust. Figure 23b shows an example over a longer period. Aberration, nutation, precession, and polar motion these are just a few o f the sometimes mystifying corrections that must be made either in astronomical observations to stars or in observations involving satellites.

Tria n g u la tio n
Since the 1620s, when Snellius developed the idea o f triangulation (figures 7 and 25), it has been used extensively throughout the world. It has form ed the skeleton upon which most national surveys have been based. Triangulation Iinks together a series o f triangles and quadrilaterals form ed between a number o f pertinent points over the area to be mapped. These points, because o f requirements o f the observational methods, tend to be prominent positions such as pillars on hilltops, church towers, and similar high points. Essentially, as the name implies, triangulation requires the measurement o f three angles in each triangle. A little more than that is needed, however, to get scale, position, and orientation into a scheme. So that everything is correct to scale, at least one accurately measured line is needed. In schemes o f any size there are often two, one towards each end o f the chain o f triangles as a check on the accumulation o f inaccuracies. Positioning the whole chain cor rectly on the surface o f the earth requires finding the coordinates (latitude, longitude, and height, or easting, northing, and height) o f at least one point and the azimuth o f at least one line. In a national survey scheme, the triangle sides are likely to be on average 50 km (30 miles) long, but they have been known to be over 100 km. On the other hand, any baseline would b<; about 15 to 20 km (10 to 12 miles), because o f the difficulty in finding suitable sites and carrying out the actual measuring procedure. The baseline would, however, be measured with every conceiv able precaution.

r~y 2,

(a) 18th and 19th Century

by wooden bars

(b) First H alf 2 0 th Century

by invar wire in catenary

(c) Today

by electrom agnetic distance measuring

26. E quipm ent for Distance Measurement

Baselines The baselines o f triangulation nets observed in the 17th to 19th centuries were measured with a wide range o f different equipment (figure 26): wooden rods placed end to end, glass rods, metal bars, steel chain, bimetallic rods, and variations on these. One problem with all these methods was accounting for the expansion o f the materials with changes in temperature. Toward the end o f the 19th century came a major breakthrough with the invention o f invar. This is a nickel-steel alloy whose low coefficient o f expansion is some 10 times less than that o f steel. This material could be used in the form o f long wires (up to 500 ft) suspended between tripods as a catenary and under known tension conditions. Accuracies o f the order o f 1 in 1 million (1 mm in 1 km, or 1 inch in 15.8 miles) proved possible, so the method was used to measure all important baselines up to the 1960s. Angles The angles in triangulation were measured with the best available theodolites often reading to 0.1". This allowed computation o f the lengths o f every other side in the scheme by simple trigonometry. From the initial orientation (probably from astronomical observations) the orientations o f every line could be calculated. By combining these data with the given initial coordinates, the coordinates o f every point could be computed. Various corrections and adjustments were applied to make whole scheme as mathematically sound as possible. After all, so much future work would depend on this skeleton that it was wise to take all feasible care and effort. Major early triangulation networks covered India, Great Britain, and large parts o f Europe. Th e United States and large parts o f Africa followed. A modern map would show most o f the world covered either by triangulation or the following alternatives o f trilateration and traverse.

Measured all sides in the network A T rilatration Network

T rila t ra tio n
As the name implies, trilatration requires the measurement o f three sides in each triangle. The idea is the same as for triangulation in that it is a method o f interconnecting, and hence positioning, a number o f points in relation to one another. With the advent o f electromagnetic methods for measuring distance it became quicker and more economic to record distances than angles (figure 27). The basic data required is either the coordinates o f A and B or the coordi nates o f one point and the azimuth to the other. In practice, a scheme is likely to involve a mixture o f both triangulation and trilatration (sometimes referred to as triangulateration) where a majority o f the sides are measured together with selected angles. The appropriate mix can often be determined through a network optimization exercise or computer simulation o f differ ent configurations. The electromagnetic distance measuring (EDM) equipment that has made trilat ration feasible did not come into wide use until the 1960s. It was first devel oped in the late 1940s as a result o f experiments to determine the velocity o f light. Such experiments required accurately measured distances against which to test the observations. Later, when the velocity became determinable to an accuracy equivalent to that o f the measured distance, the whole idea was turned around. Particularly under the initiative o f Erik Bergstrand in Sweden, the Geodimeter was developed, and the first production model appeared in 1949. It was extremely heavy and bulky, but those national survey departments around the world that tried it out achieved very encouraging results. By 1957 Dr. Trevor Wadley in South Africa had developed a similar instru ment the Tellurometer that operated on microwaves. In the beginning, the observational and reduction procedures were tedious and time-consuming, but as electronics developed, the equipment became more readily acceptable. As with triangulation, so trilatration requires at least one set o f known coordinates and an orientation on which to found the computed coordinates.

Open Traverse

Closed Traverse

Closed Traverse

2 8 . Traverses

68

The Geodimeter and the Tellurom eter were capable o producing com parable results; however, there were differences between them, the most ob vious being the use o f light waves versus radio waves. Am ong their other differences differences typical o f other instruments in these two cate gories are: Light Waves Geodimeter Control unit at one end o f the line, reflector at the other end. Range limited by visibility to 30 to 50 km. Operator must be able to see between the terminals. Light waves affected only a little by atmospheric conditions. Trained operator at the control unit, unskilled operator at the reflector. N o integral communication between the control unit and the reflector position. Radio Waves Tellur orfieter Similar control units at each end o f line, one called the master and the other the remote. Possible range in excess o f 100 kmy even in poor visibility; i.e., can measure through haze, drizzle, etc. Microwaves more susceptible to humidity changes. Trainee! operator at each end.

C om rm u n ication possible b etw een th e m a s te r an d re m o te instrurtfients o n the m easu rin g signals.-

In general, the accuracies achieved were o f the order o f 1 in 250 000 (4 ppm) o f the length o f line; for example, aril error o f only 80 mm over 20 km. It is usual, however, to quote accuracy levels in the form ( x ppm y mm) since part is attributable to the electronic configuration and part to the total length. By the second half o f the 1980s, accuracies in tthe submillimeter range were possible with the Mekometer at ranges o f several kilometers, for example, (0.2 mm + 0.2 mm/km) up to 8 km. Variations and improvements on these instruiments continue to be made. However, many o f the tasks they could be used ifor are now done by satellites in the GPS.

Tra v e rs in g
The whole o f surveying is involved with determ ining the relationships between points on the earths surface. For large areas wherre there is reasonable intervisi bility over great distances, then triangulation arP d trilatration will be used to form a skeleton o f known points. Where the tterrain is, for example, heavy jungle or bushland or unduly flat, another technique may be preferable. This

69

Central m eridian M agnetic north True north

Grid north

Magnetic north

c = convergence t -T = arc-to-chord correction 29. D iffe re n t North Lines

litclliod called traversing can be arranged in different ways according to ilu accuracy sought, the size, and the configuration o f the area. In essence (figure 28) traversing requires the measurement o f the angles ,n . 1 series o f points together with determination o f the distances between m m< cssive points. Angular measures invariably use the theodolite since these itir available with resolutions down to 0.1". Th e distances could be by invar i.tpi in catenary (figure 26), steel tape on the ground, EDM equipment, or \ ,ii iations on these.

N e tw o rks
Network order In each case, whether the surveyor uses triangulation, trilateration, or traverse the results are graded into first-order, second-order, and lliird-order (or primary, secondary, and tertiary) networks according to accui .ic y, area o f coverage, and usage. A First-order network is one for which the best available instrumentation lias been used, the utmost care taken, and the most rigorous adjustment procedure applied. It is thus the most time-consuming and expensive frame work but also the most accurate. Invariably, national surveys use first-order networks. The average length o f line could well be 50 km. Second-orderwork is the initial filling-in o f the first-order network with closer points and to a less exacting standard. A third-order network is yet a further breakdown o f the secondary network, with coordinated points every 5 to 10 km. Every national survey organization has its own criteria for each order based on local needs and circumstances. Grids For ease o f location o f points on a map the sheet is normally covered with a network o f lines in two sets at right angles to one another. These are ( ailed grid lines and usually one set o f lines is oriented more or less northsouth. However, this is not essential, and in essence the lines can be rotated into whatever orientation the user requires. In terms o f the geodesy o f this volume, it is usually the national grid system o f a country that is o f interest. Often it is stipulated that a survey schemes coordinates must relate to the national system, which will almost certainly conform to a north-south orientation. At lower levels o f surveying, quite arbi trary local grid systems can have any orientation the user thinks is appropriate. As noted elsewhere, coordinates on any one grid system can be converted to their equivalents on any other system, as long as there is a minimum number o f points common to both systems. Convergence The relation between the north line on a grid and the direction to true north (the geographic North Pole) at any point is a variable. The difference between the two can be readily calculated and can amount to several degrees o f arc (positive or negative) depending on the circumstances. This difference is known as convergence (figure 29).

14

s
Si

(a) S p irit Leveling

(b) Trigonom etric Leveling (S im p lifie d )

3 0 . Leveling

Magnetic north Yet another north to become familiar with is magnetic north. It is a slowly changing quantity and has to be calculated both as a function o f the position o f the point o f interest and o f the time (the nearest month or year) when the observation was made (see figure 29). (t T) Just to complicate matters further, another correcting term is required. This is known as the (t T ) or arc-to-chord correction. A straight line between two points on the earths surface will plot as a curve on a projection. This curve, or geodesic, diverges from the straight line by amounts that vary as a function o f the position o f the point o f interest in relation to the central meridian o f the projection. This divergence can be a handful o f seconds, positive or negative, and so on occasion it has to be taken into consideration (see figure 29). Spherical excess Triangles form ed on the earths surface by well-separated points are not the same as those drawn for typical mathematical problems. In the latter, the angular sum always totals 180, but that is not so with large triangles, which are essentially spherical triangles. The theoretical sum o f the three angles in such a triangle is 180 + e, where e can be a few seconds o f arc and varies as the area enclosed by the triangle. As a rule o f thumb, 76 square miles (195 sq km) equates to 1 second for e. Very roughly, this is an equilateral triangle with 10-mile (16-km) sides.

VERTICAL POSITIONING

P re c is e o r G e o d e tic S p irit L e v e lin g

Determination o f the position o f a point in the vertical plane above a defined datum, such as mean sea level, can take several forms. The most accurate is by spirit leveling. The instrument consists o f a telescope set so that it can be very accurately leveled up in relation to a plumb bob (and hence to the geoid) and that can be rotated through 360 horizontally. It is used in conjunction with a graduated staff to find the change in elevation between a succession o f points, as shown in figure 30a. The height o f the starting point, or benchmark, must be known. At the end o f the sequence o f observations a reading should be taken either to the same benchmark or to a different one if it is more convenient. This allows a ready check to be made for observational errors. When the instrument is set up at I x and leveled (figure 30a) a reading a is taken on the staff Si at the benchmark. This allows calculation o f the height o f the plane generated by the line o f sight ( plane of collimation) when the telescope is rotated. A forward reading b is then taken on a second staff position S^ The instrument moves on to / 2 , and the same procedure is used for readings c and d and successive positions.

QR'\ P S ' = level surfaces - lines of constant o rthom etric height

Q R ; PS = e q uipotential surfaces - lines of constant dynam ic height

3 1 . O rthom etric and Dynamic heights

With this form o f leveling, the distances S | to I b I] to S2 and so on, are kept as nearly equal as possible and limited to some 30 to 50 m each. This sequence is repeated as often as necessary to determine the heights o f the required points. At all positions the rotating telescope is parallel to the geoid. Since the force o f gravity increases as one goes from the equator to the poles, on leveling in such a direction the equipotential surfaces converge (figure 31a). Orthometric, ellipsoidal, geoidal, and dynamic heights For the survey prac titioner, orthometric heights are the bread-and-butter form. Differences in orthometric height are the quantities obtained when using a normal level instrument. These differences are applied to the known value o f a starting point to give orthometric heights. They are heights above the geoid. GPS gives height values related to an ellipsoid, which may be (but most likely is not) coincident with the geoid. Thus to turn GPS (ellipsoid) values into their equivalent orthometric values requires a correction called the geoidal separation (also known as geoidal height or undulation of the geoid). I f the ellipsoid and geoid happen to be coincident, then the separation is zero. On the other hand, if a different ellipsoid is used at the same point as before, there could be a noticeable separation (possibly many meters), either positive or negative. Th e sign indicates whether the geoid surface lies above or below the ellipsoidal surface at that point. Since the surfaces certainly are not parallel, the separation varies from place to place. Often it is more important to know the difference between the separations at given points than the absolute values at those points. There is a third form o f height value normally called dynamic height that needs to be appreciated. Referring to figure 31, note that the linear distance Qq is greater than Rr, although R is said to have the same dynamic height as Q. A line o f constant dynamic height is parallel to mean sea level along a line o f latitude, but the lines coverage when progressing toward the poles. The orthometric height o f Q above q is the length Qq. The dynamic height is measured as the work required to raise a unit mass from q to Q. Thus QR' is a line o f constant orthom etric elevation, while QR is a line o f constant dynamic elevation. Dynamic heights are thus defined in terms o f potential, which is a function o f the acceleration due to gravity and the distance o f the point in question from the center o f the earth. Equipotential lines (see figure 31) generally converge towards the poles. Dynamic height values are given in terms o f geopotential numbers, where points o f equal number lie on the same equipoten tial surface but will not have the same orthometric values. The surface o f a large lake is an equipotential surface. Leveling would show il to be (almost) level, but ellipsoid heights would generally show it to have a slope.

Deviation of the Vertical

N = height of geoid above ellipsoid h = H + N (approxim ately) 3 2 . E llip so id a l'a n d Geoidal Heights

Orthometric heights are disstances measured along the vertical perpendicu lar to mean sea level. The elliipsoid heights are measured along the normal to the ellipsoid. Thus heightss found via GPS are ellipsoid values related to the normal to the ellipsoid. F-^or comparison with spirit-level values, ellipsoid values have to relate to the nojrmal to the sea level and the difference between the two surfaces (figure 32). In the leveling process the bubble against which the instrument is leveled lies on an equipotential surface, but when the staff is read, it gives an orthomet ric value. An appropriate forrmula allows conversion from one to the other when the order o f accuracy wvarrants it. Note that while the traditiconal forms o f heighting are related to the direc tion o f gravity, the results frojm GPS are not so constrained. Geoidal separation The gee>idal separation can be determined with varying degrees o f reliability. Where observational coverage has been good, contour maps o f the values are often available, and an interpolation can be made for the required location. The accuracy may be o f the order o f 0.1 m depending on the part o f the world involved. Other authorities provide computer models from which predicted values, may be obtained. Alternatively, an indication can be obtained by relating a selection o f ellip soidal results to known orth ometric values in the same vicinity. A geodesist who does this, must take car-e in selecting suitable points. For scientific purposes it i possible to observe astronomical positions along a particular line across a courn try and from the results determine, for a particu lar ellipsoid, the deviation o f the vertical and geoidal separation. This is so time-consuming that it i:s seldom attempted, certainly not for standard survey operations.

= height of A hB = height of B r = e ffe c t of refraction c = e ffe c t of curvature dh = apparent d ifference of height, A -B d H = actual difference of height, A-B a = observed vertical angle

/ /

sx

/( / \( \ \ \

d H = dh

3 3 . E ffect of Curvature and Refraction

T r i g o n o m e t r i c I H e ig h t in g
Trigonometric heigh ting is a method that can be adopted when it is necessary to determine the difference o f height between points that are far separated or are difficult to level between, or where the required accuracy is not o f the highest order. Instead o f a telescope mounted horizontally, trigonometric heighting involves using a theodolite to measure the vertical angles and combining them with horizontal distances (figure 30b) to find the difference in height. Since the earth is not a flat surface, the solution is complicated by the effect o f the earths curvature, which can be combined with the effect o f atmospheric refraction, or bending o f the light rays (figure 33). Whil e the curvature is a definite correc tion as a function o f the length o f the lline o f sight, the refraction varies with the changing atmospheric conditions and so introduces an element o f uncertainty into the result. The combined effect o f curvature and refraction is about 67 mm at 1 km, or 6.7 m at 10 km, or about 15" o f arc fo r 1 km in angular terms. For this technique sight lengths can be as long as intervisibility allows. The height o f Mount Everest was determined in this manner from numerous observations in the Indian Plains some 150 km away. Variations on the technique are possible when both ends can be occupied simultaneously. It allows for better m odeling o f the refraction effect and improves the accuracy.

B a ro m e tric H e ig h tin g (L e ve lin g )

Barometric heighting is a method used when lower-order accuracy is accept able. It uses the principle o f the change in pressure with elevation, first investigated and used by Bouguer while working on the Peru arc in 1738. In an ideal atmosphere there would be a direct relation between pressure and elevation, but in real life the situation is more complicated. Nevertheless, it is possible to use a pressure-measuring device for determining the difference o f height. Using several such instruments in appropriate arrangements can improve the accuracy. Originally a mercury barometer was required, which was difficult to trans port and use satisfactorily. Nowadays aneroid barometers make the operation very straightforward. The aneroid consists essentially o f a small box from which air has been exhausted. T o its top surface is fixed a pointer and scale. The effect o f changes in pressure on the box are magnified and indicated by the pointer. The units o f measurement may be millibars o f pressure, millime ters o f mercury, or even direct graduation in terms o f height units. A variation on the mercury barometer no longer used is the hypsometer or boiling-point thermometer, which could be similarly related to height.

E c h o S o u n d in g
Besides methods to determine the heights o f points on visible terrain, there are also depths below sea level to consider. In previous centuries the technique

was to drop a weighted line over the side o f the ship and record the amount o f line paid out until the bottom was reached (see figure 44c). While this method was prone to various errors, it generally gave an acceptable indication o f depth for the purposes o f navigation. Nowadays far more information is required about the seabed, so sophisticated methods have to be used. In particular, the echo sounder is a device for bouncing pulsed signals o ff the seabed and recording their return at the vessel. This can provide continuous coverage along a line as opposed to the discrete measures o f a lead line. Variations on the basic principle allow recording o f not only vertical depths but also a band to either side o f the vessels track. In this way complete coverage o f selected areas is feasible. The higher the frequency o f the signal, the shorter the wavelength and the narrower the beam width. This relation would be pertinent when accurate measurements o f depth to specific points are o f more interest than general indications o f the seabed form and depth. Even so, accuracies much better than 1> in 200 are difficult to achieve. An average value for the velocity o f sound in seawater is 1500 m/sec, but this is a function o f salinity, temperature, and density all o f which change with time and conditions as well as with depth.

Lines of latitude

Lines of longitude

0 = latitude A = longitude a = initial azimuth Geoid separation at origin = often zero. a, b = semi-axes

34. Horizontal Datum

CH A P T E R

5
Geodetic Systems

r \ geodetic datum is defined as any numerical or geometrical quantity or set o f such quantities that serves as a reference or base for other quantities. In geodesy two types o f datum must be considered: a horizontal datum, which forms the basis for the computations o f horizontal control surveys that consider the curvature o f the earth, and a vertical datum, which elevations refer to. In other words, the coordinates for points in specific geodetic surveys are computed from certain initial quantities or datum parameters. Originally the horizontal and vertical datums were kept strictly separate, but today it is more usual to quote both at the same time. This join t quotation results from the fact that orbiting satellites relate to the datum both horizontally and vertically. Previously there was no link or common factor between the two. Because o f the way in which surveying has evolved, there are a multiplicity o f datums. In light o f satellite geodesy, there would now be no need for more than one, but there is no likelihood o f a full change since such a vast quantity o f work corresponds to each separate datum. T o convert all existing mapping information in just one country would take maybe a century. As with the property boundaries o f centuries ago in the United States, so in years to come map users will still want to refer to the ancient systems for particular aspects. Thus a complete change would not eliminate a problem but rather add yet another variable to an already difficult inheritance.

HORIZONTAL GEODETIC DATUMS A horizontal datum (figure 34) consists o f the latitude and longitude o f an initial point (called the origin), an azimuth for one line, the parameters for (he ellipsoid chosen for the computations (one radius and the flattening), and the geoidal separation at the origin. Defining an ellipsoid does not on its own form a datum; it is only one o f several required quantities.

---------

-----------------------------------------------------------------------------------------------------------------------------

Ellipsoid center 35. Datum Orientation for Single Astronomical Station

A change in any one o f the parameters would affect all points computed on that datum. For this reason, while positions within a system are directly and accurately relatable, data such as distance and azimuth derived from computations between the positions o f the same points on different datums will vary in proportion to the difference in the initial quantities.

O rie n ta tio n of E llip s o id to G e o id

Single astronomical position datum orientation (figure 35) Selection o f the reference ellipsoid provides the radius and flattening elements. The simplest means o f obtaining the other three factors to establish the geodetic datum is to select a first-order triangulation station, preferably one located near the center o f the triangulation network, to serve as the datum origin. Then the astronomical coordinates o f the station and the astronomical azimuth o f a line from that station to another control station are observed. An alternative to observing an azimuth would be to compute it by observing for the astronomical coordinates o f the second station. The observed astronomical coordinates and azimuth are adopted, without any correction, as the geodetic coordinates and azimuth o f the datum origin on that particular ellipsoid. In addition, it is usual to assume that the geoid and ellipsoid are coincident at the datum. This means that the deflection o f the vertical and the separation between the two surfaces are defined as zero at the origin. In this method o f orientation, the normal to the ellipsoid is arbitrarily made to coincide with the plumb line at the datum origin. Although the computed positions will be correct with respect to each other in this type o f orientation, the entire net will shift with respect to the axis o f the earth. This is not significant for local use o f the positions but may introduce large systematic errors as the survey is expanded. N ote that although the deflection and undulation are defined as zero at the origin, deflections will occur at other positions within the network. Therefore, when we compare the geodetic latitude and longitude o f any other point in the network with the corresponding astronomical latitude and longitude o f that point, discrepancies will appear between the two sets o f values. A datum oriented by a single astronomical point may produce large system atic geoidal separations. The ellipsoid is not earth-centered, and its rotational axis is not coincident with the axis o f the earth. The inconvenience o f such an orientation is that the positions derived from different, astronomically oriented datums are not directly comparable to each other in any geodetic computation.

E llipsoid center

A = A stronom ical la titu d e G = Geodetic latitude S = Astrogeodetic defle ctio n s of the vertical 3 6 . O rientation fo r Astrogeodetic Datum

Astrogeodetic orientation The deflections o f the vertical at a number o f Laplace stations (see below) can be used for a second type o f datum orientation known as the astrogeodetic orientation. This orientation (figure 36) makes a correction at the origin that in effect reduces the sum o f the squares o f the astrogeodetic deflections at all the Laplace stations to a minimum. One o f the Laplace stations in the adjustment is arbitrarily selected as the origin. A Laplace station named after the famous French mathematician PierreSimon Laplace (1749-1827) is one at which astronomical observations are made for azimuth and longitude to control the accumulation o f errors along the chain. I f possible, they are located near the junctions o f chains. The longitude observation is made so that allowance can be introduced for the angle between the local vertical and the ellipsoid normal, which affects the azimuth observations. From this comes the noted Laplace equation that is a function o f the difference between the astronomical and geodetic longi tudes with the difference between the two azimuths.

E llipsoid Geoidal height = d i + d 2 + d 3 3 7 . D eflection of the Vertical

This deflection o f die vertical (figure 37), or angle between the plumb line and the normal to the ellipsoid, is usually resolved into a north-south component (equal to the difference between astronomical and geodetic lati tudes) and an east-west component (proportional to the difference between astronomical and geodetic longitudes.) Th e Laplace equation provides a means o f reconciling the azimuth differences resulting from the use o f two separate reference surfaces. Laplace equations are introduced into triangula tion adjustments to control the azimuth and orient the ellipsoid. Therefore, instead o f a zero deflection at the origin as with a single astronomical position, there is a particular deflection o f the vertical. Similarly, the geoidal separation (undulation) can be determined at the origin and the ellipsoid reoriented provide a best fit for the ellipsoid and the geoid in the area o f the Laplace stations used. Consequently astrogeodetic-oriented datums can be applied to larger areas than those oriented by a single astronomical position. This, for example, was the arrangement applied a few years ago in Australia. Using an adopted origin with coincident astronomical and geodetic coordi nates, the survey networks were preliminarily computed. This provided a set o f deflections o f the vertical from which a mean value was obtained. Applica tion o f this mean value to the origin allowed a recomputation that gave a better overall fit for the whole continent. In general, a best-fitting ellipsoid can be chosen for a national geodetic network such that the undulations are no more than 10 m.

Deflection angle

Y Ellipsoid Geoid

Geoid

38. Astrogeodetic Deflections Are Relative

E The astrogeodetic orientation has the disadvantage that the deflections o f the vertical remain relative (figure 38). I f the ellipsoid is changed the deflec tions o f the vertical will also change. Second, it is necessary to assume a specific orientation o f the reference ellipsoid with respect to the geoid before computing the astrogeodetic deflections. The orientation is fixed by the initial values o f the datum origin from which the geodetic coordinates were com puted. Any change in these quantities changes the deflection o f the vertical at each point. Consequently, the astrogeodetic deflection o f the vertical depends upon a specific geodetic datum, and the use o f geodetic data developed by this method is limited to relatively small areas. Th e angle between the astronomical zenith (which relates to the geoid) and the geodetic zenith (which relates to the ellipsoid) equals the astrogeodetic deflection. This angle also equals the angle o f intersection between the geoid and ellipsoid surfaces, so it is possible to compute the separation in increments from the deflection values (see figure 36).

S a t e llit e D a t u m s
The datum used for satellite observations is defined in a different way. It is based on the positions given for several tracking stations, the geopotential model selected, and various constants that do not appear in traditional datums. These include the velocity o f light, the rate o f rotation o f the earth, corrections for clock and oscillator rates, the gravitational constant, and the earth mass. Such a system has as its origin the earths center o f mass and uses the same coordinate system as that used for the precise ephemerides (see page 145). It is usual also to quote an ellipsoid o f reference so that geographic coordinates can be obtained.

D is c re p a n c ie s b e tw e e n D a tu m s
In areas o f overlapping geodetic triangulation networks, each computed on a different datum, the coordinates o f the points given with respect to one datum will differ from those given with respect to the other. The differences occur because o f the different ellipsoids used and the different deflections o f the vertical and geoid separations at the datum origins. The latitude and longitude components o f the absolute deflection at the initial points result in a parallel shift between the systems. Such a shift is due to the fact that the minor axes o f the various reference ellipsoids do not coincide with the rotation axis o f the earth. In addition, deflection errors in azimuth cause a relative rotation between the systems. Finally, a difference in the scale o f the horizontal control may result in a stretch in the corresponding lines o f the geodetic net works.

39. Example of Many Datums in Southeast Asia Area

The discrepancies are generally larger between datums oriented by a single astronomical point than between those with an astrogeodetic orientation. In view o f the initial point translation, the azimuth rotation, and the hori zontal scale differences, any attempt to correlate the geodetic information from one datum to another, unconnected datum is quite impossible. Regard less o f the accuracy o f the individual datums within themselves, there is no accurate way to perform distance and azimuth computations between uncon nected geodetic systems. With the development o f both intermediate and long-range weapon sys tems, geodetic problems have become more critical than ever before. T o satisfy military requirements, it is necessary to provide detailed cartographic coverage o f areas o f strategic importance and to accomplish geodetic computa tions between these areas and launch sites, which are often on unrelated datums. Both o f these requirements necessitate unification o f major geodetic datums by one or a combination o f existing methods.

D a tu m C o n n e c tio n
There are three general methods by which horizontal datums can be con nected. The first method is restricted to surveys o f a limited scope and consists o f systematic elimination o f discrepancies between adjoining or overlapping triangulation networks. This is done by moving the origin and rotating and stretching the networks to fit each other. The method is generally used to connect local surveys for mapping purposes. Known as datum transformation it can only be used where control exists for common points in the different systems. A minimum o f two such points is necessary, but it is preferable to have a spread o f common points over the area o f interest. It is then possible to do a least-squares best fit between the two sets o f values. T o relate larger areas to one another, common points can be used i f they exist; in addition, extra common positions can be introduced through satellite observations. Using some o f the methods described in Chapter 8, the geodesist can interrelate satellite observations from critical points in each geodetic network and so calculate the necessary transformation parameters. In earlier years this same form o f connection could be achieved through solar eclipses, star occultations, and use o f the moon-position camera. The third form o f datum connection is a gravimetric method, described on pages 121 and 127. Coordinate transformations can be in either two or three dimensions de pending on the requirements and the common data available.

D a t u m s b e f o r e W o r l d W a r II
By 1940 every technically advanced nation had developed its own geodetic system to an extent governed by its military and economic requirements. Some systems developed through the expansion and unification o f existing

94 40. Preferred Datums

local surveys, and others hy new nationwide surveys replacing outdated local ones. The former o f these two approaches is not to be recommended since "working from the part to the whole can propagate and enlarge inherent errors and is contrary to the standard survey maxim o f always working from the whole to the part. Normally, neighboring countries did not use the same geodetic datum. There was no economic requirement for common geodetic information, and the use o f common datums was contrary to the military interests o f each country. Th e only surveys o f an international nature based on the same datum were the few measurements o f long arcs accomplished for the purpose o f determining the size and shape o f the earth. An example o f this was the Struve arc from the River Danube to the North Cape. This meridian arc was measured by triangulation during the period 1815 to 1855; it stretches 2520' in latitude or over 2800 kms, and it passes through ten countries roughly following the 26 E line o f longitude. International boundary surveys required coordinates o f marks to be in the systems pertinent to the countries involved, but this could be achieved without full conversion into one united system. The net result was that there were many different surveys o f varying size that differed from each other to a remarkable extent. The national maps based on the surveys also differed widely. Figure 39 illustrates the confusion o f geodetic information in Southeast Asia. As military distance requirements increased, positioning information o f local and even national scope became unsatisfactory. The capabilities o f the various weapon systems increased until datums o f at least continental limits were required.

P re ferre d D a tu m s
By 1940 the best solution was to establish for both military and scientific purposes a set o f preferred datums for selected areas and adjusting all local systems within each area to it. The North American, European, Tokyo, and Indian datums were selected for this purpose (figure 40). It did not mean that each country had to convert all o f its coordinated material to the preferred datum for everyday use. It was only information o f a primary character and o f scientific and military significance that obtained a second set o f values. North American Datum (N A D ) Before 1927 a U.S. standard datum was based upon a meager triangulation network. As new networks were added, the magnitude o f the inconsistencies became objectionable. As a result, a general readjustment was made o f the whole system, and the outcome was known as the NAD 27 datum. The origin is at Meades Ranch, Kansas, lat. 3913'26.686", long. 26127'29.494". The datum was computed on the Clarke 1866 ellipsoid, which was oriented by a modified astrogeodetic method. The system incorporates Canada, Mexico, and the West Indies with a Central and South American connection. Consolidation o f the various Cen tral and South American surveys has been carried out by the Inter-American Geodetic Survey.

95

Geodetic System:

As before, inconsistencies became apparent as new data was added and higher reliabilities w ere required between relative coordinate positions. Where an accuracy o f 1 in 100 000 was required, some parts had only 1 in 15 000. Hence errors and distortions were introduced. Many years o f effort based upon a vast increase in data, improved computa tional techniques, and the advent o f satellite positioning systems went into the redefinition. T h e result was the adoption o f the 1980 international ellip soid values (see page 31). Th e new datum became known as NAD 83 and had a geocentric origin. European Datum (E D ) Th e initial point o f European datum is located in Potsdam, Germany, at the H elm ertTower, lat. 5222'51.45" long. 1303'58.74". Numerous national systems were joined into this large datum based upon the international ellipsoid, which was oriented by the astrogeodetic method. The Army Map Service connected the European and African triangulation chains and filled the rem aining gap o f the African arc measurements from Cape Town to Cairo. Thus all o f Europe, South Africa, and North Africa were m olded into one great system. Through common survey stations it was also possible to convert data from the Russian Pulkovo 1932 system to the European datum; as a result the European datum includes triangulation as far east as the 84th meridian. Additional ties across the Middle East permitted connection o f the Indian and European datums. ED 50 covered various countries on mainland Europe, but in 1963 a connec tion was made across the English Channel to join the United Kingdom to the same datum. Initially ED 50 was used mainly for scientific purposes, but later it became useful in such operations as the oil industry in the North Sea. This stretch o f water lies between Scandinavia and the British Isles, and so a common system was o f particular advantage. In the 1980s and 1990s the Channel Tunnel connection has been an opportunity for cooperation between the national survey organizations o f France and Great Britian and for use o f the common datum. Tokyo datum Th e third preferred datum, the Tokyo datum, has its origin in Tokyo at the Observatory, lat. 3539'17.51", long. 13944'40.50". It is defined in terms o f the Bessel ellipsoid and oriented by means o f a single astronomical station. By triangulation ties through Korea, the Tokyo datum is connected with the Manchurian datum. Unfortunately, Tokyo is situated on a steep geoid slope, and the single station orientation has resulted in large systematic geoid separations as the system extends from its initial point. Indian datum The Indian datum is now accepted as the preferred datum for India and several adjacent countries in Southeast Asia. It is computed on the Everest ellipsoid with its origin at Kalianpur in central India, lat. 2407' 11.26", long. 7739'12.57". Derived in 1830, the Everest ellipsoid is the oldest o f the ellipsoids in common use and is noticeably out o f sympathy with more re m it determina tions. As a result, the datum cannot l>r extended too l.u Irom the origin since

Geodetic Systems

very large geoid separations will occur. For this reason and the fact that the lies between local triangulations in Southeast Asia are typically weak, the Indian datum is probably the least satisfactory o f the preferred datums. Thus it can be seen that 4 quite different ellipsoids are used. Referring to page 31, their parameters are: _____________________________________________ ________________________ W _
Clarke 1866 International (H ayford) Bessel Everest 6 6 6 6 378 378 377 377 206 388 397 276 294.978 297.000 299.153 300.802

As figure 41 shows, wide variations o f this magnitude can not be readily reconciled into one system without large distortions and deviations.

O th er A rra ng e m e n ts
For military distance and direction problems, limited to continents or smaller areas, the preferred datums were initially satisfactory. However, while they were improvements over the limited national datums, they had serious defi ciencies that prevented them from providing the geodetic information re quired for intercontinental ballistic missiles. At one stage a North Atlantic Hiran tie permitted connection o f the Euro pean datum and the North American datum. This however, did not completely solve the problem since both these datums are relative. W hile in each case the chosen ellipsoid gave an adequate fit in the area o f the origin, as already shown, neither provided a good fit for the entire earth. In addition, the process o f connecting various datums by means o f intervening datums or triangulation/trilateration ties allowed values to accumulate that do not always agree with newly observed data. The surveys join in g India to the European and Tokyo datums presented just such a major problem; the requirement for long-range geodetic information could not be considered solved by the preferred datums. The Hiran system (mentioned above) was one o f the early improvements on the Shoran system, which was the first application o f radar techniques to surveying. The principle used was similar to that o f modern EDM equipment. However, by mounting the equipment in an aircraft much longer lines could be measured, and some 500 km proved feasible. The Hiran system permitted the use o f even longer lines and achieved an accuracy o f 1 in 100 000 (see figure 53). Choice of ellipsoid While it might seem obvious that it would be preferable to have just one ellipsoid for worldwide use, this is not necessarily practical. As figure 41 shows, an ellipsoid that fits well in one part o f the world may diverge by a noticeable amount elsewhere. The effect o f this was illustrated in Australia, where, had they adopted WGS 72, the divergence between the ellipsoidal and geoidal^urfaces would have given unacceptable scale errors. Corrections varying

4 1 . F ittin g D iffe re n t E llipsoids to D iffe re n t Areas

In mi + 6 to 12 |)|)in would have had to be applie;d by everyone in the distance measuring business many o f whom would not hawe understood what they were doing, and hence would quite likely ignore the corrections. An African Datum Recently new work has beein d o n e to assess the options lor choosing a suitable datum for the whole o f A fr ic a . Earlier attempts, com pleted some 10 years ago, have not yet been im p le m e n ted because the many i hanging boundaries in the continent make a unified approach important. Doppler data relating to 185 stations throughout! the continent, together with other available information, have been used to> fin d the best-fit parameters under a variety o f situations. The results vary considerably, and much further work is necessary before any one arrangement (can be accepted.

VERTICAL DA TUM S |list as horizontal surveys refer to specific orig in a l conditions (datums), so vertical surveys also relate to an initial quantity o r datum. As already discussed, elevations are based on the geoid because the instruments used either for spirit leveling or trigonometric leveling are adjusted so that the vertical axis is coincident with the local vertical. As with horizontal datums, there are many small discrepancies between neighboring vertical datums. Although elevations in some areas relate to surfaces other than the geoid, there is never m ore than 2 m o f variance between leveling networks based on different m ean sea-level datums. In European areas there are fewer vertical datum problems than in Asia and Africa. Extensive leveling work has been d o n e in Europe, and practically all o f it has been referred to the same mean sea-level surface. However, in Asia and Africa the situation is entirely different. In some places there is precise leveling information based on mean sea level; in other areas the zero elevation is an assumed elevation that sometimes has no connection to any sea-level surface. China is an extreme example o f this situation, in that nearly all o f the provinces have their own independent zero reference. There are no reliable vertical data for as much as 75% o f the area o f Asia and Africa. As long ago as the Peruvian arc measure o f 1738 there have been difficulties in connecting to sea level. Only after several years o f trying was Bouguer able to tenuously connect the triangulation scheme o f the high Andes to sea level and make the appropriate corrections to the observations and calculations. The vertical reference system o f the United States was fixed in 1929 from some 100 000 km o f leveling on the North American continent. Twenty-six tide-gauge values were held fixed in the overall adjustment. A new vertical control network is expected to be completed in the near future. Th e datum in the U K was determined in 1921 at Newlyn while that in Australia was determined in 1971 on the basis o f a national levelling net adjustment that incorporated a large number o f tidal stations. A

o
(a) General situation

Moon

o
(b)

Moon

Sun

Spring tides

^" ^M oon

4 2 . Tide Generation

Tid e s
Whereas for horizontal datums there is an infinite choice o f datum positions when it comes to the datum plane for heights, the choice is simpler. Through out the world the oceans present an obvious choice as a datum, and that is what is often chosen, but the situation is not quite as simple as it would seem. Since the fourth century B.C. it has been appreciated that the sea level rises and falls and that the amount o f rise and fall varies throughout the year. (Tidal variations in the Mediterranean area were unimpressive and it was not until wider exploration that their regularity became accepted). What value then should we take as the sea-level position when it is to be a datum? What causes tides Consider first the forces that produce the tides. It was not until the time o f Newton that the forces producing tidal variations were properly understood. H e showed how the tides were in direct agreement with his law o f gravitation. While the sun revolves in relation to the earth in about 365.25 days, the moon does so in about 27.3 days. Between them these bodies exert forces upon the earth in keeping with Newtons law that every particle o f matter attracts every other particle with a force proportional to the mass o f each and inversely proportional to the square o f the distance between them. Since the sun and moon neither orbit in circles nor lie in similar orbits, any effect they have varies from day to day according to the relative position o f both bodies to the earth at any particular time. The orbits o f the sun and moon are ellipses, so there are positions o f nearest and farthest approach for each body. I f we consider each body separately and assume that the earth is completely covered with water, then the water height will rise at places on the earth approximately where the line normal to the body cuts the water surface. This will apply on both sides o f the earth (figure 42). Similarly, there will be depressions o f the surface at positions 90 from the high points. As the earth rotates, these positions will move around it. At the same time, as the body in its orbit varies in its distance from the earth, its attractive effect changes and the height to which the water surface rises varies accordingly. Since there are two positions o f both high and low water level, at any point on the earths surface there will usually be two high water levels and two low water levels every day. As stated above, these positions lie approximately on the intersection o f the normal to the earths surface and the body. They are only approximations because o f the time lag caused by the effect o f friction between the water and the ground surface over which it moves. The positions o f the sun and moon in relation to the equatorial plane also cause variations in the times o f occurrence o f high and low water levels and their magnitudes. In general, the m oons effect is more than twice that o f the sun. There are times during the year when the two bodies tend to act together and other times when they act against one another. Thus we can see that many parameters are involved in finding the position o f the water surface at any time. In fact, there are even longer-period variations, such as 18.6 years due to movements o f the position o f the moon relative to where its path crosses the ecliptic.

Highest astronom ical tid e Mean high-water springs Mean high-water neaps Mean sea level Mean low-water neaps Mean low-water springs Chart d a tu m --------Low est astronom ical tid e

4 3 . Selection of Ocean Levels

Since quarter-diurnal, semidiurnal, urnal, monthly, annual, and 18-year variations need to be combined with thf effect o f the topography o f the coasts and unpredictable fluctuations, it is no w<nder that choosing a datum is difficult. Mean sea level T o obtain the best dium position at any coastal location, continuous observations o f the sea leve should be taken over a period o f at least 18.6 years and the results average!- This would give the mean sea level at that location. Such a length o f time ould be impractical, and approxima tions o f varying accuracies are obtaimd by averaging regular observations over shorter periods such as a day, (actially the average length o f a tidal day is nearly 25 hours), a month, or a year The accepted constituents o f the variitions can range in number according to the detail and accuracy required fa the overall result. Up to 63 such components are sometimes referred to, out that would only be for exceptional investigations. Several groupings withn these interact with each other to provide repetitions over cycles rangingfrom a few hours to many years. Monitoring Recent environmental worries over gas emissions have suggested that sea level may be rising steadily. Obvously, for many reasons such a change requires careful monitoring. The trend may be only a millimeter or two per year, but aggregated over some years this can become a disastrous amount so much so that in various parts o f the werld continuous monitoring o f the sea level is an ongoing concern. For example, many o f the islands in the Pacific Ocean are low-lying, and any noticeabe sea level rise would cause not only inundation but perhaps annihilation o f some islands. Because o f this, a network o f 11 monitoring stations has been set up around the islands stretching from Papua New Guinea at about long. 148 E to the Cook Islands at 162 W, 20 S and to the Marshall Islands at 10 N or aver an area 50 E-W and 30 N-S. The system to be set at each o f thesi stations will be the latest state-of-theart multimeasurement arrangements to record all manner o f sea parameters and to transmit the results to the maste' processing station via a geostationary Japanese satellite. The system will be capable o f taking readings every 2 seconds with the results averaged in 6-minute blocks. While all these measurements may l>e good to 1 mm, remember that the earths surface itself is moving. Hence all such stations have to be located on a single reference system the (International Earth Rotation Service, Terrestrial Reference Frame (IERS T R F ) and monitored over many years before definite sea-level movement can be confirmed. For this, one o f the best methods o f relative positioning over thousands o f miles is very long baseline interferometry (VTB I). Definitions Mean sea level can be defined in two different ways. It is the equipotential surface that the oceans would have i f left undisturbed. Alterna tively, at any individual point on the se^coast it can be defined as the average o f the constantly changing water level over a period o f time. This could be better termed local mean sea level. Many alternative positions o f the water surface (figure 43) have their uses for navigation and other purposes. In particular, there are the values o f high . * _________________ _______________________________________

ueoaetic yysiimnn

Recording drum and pen

Weight

Float

"C f~ -

Damping inlet

(a) Tide Pole

(b) Recording Tide Gauge

44. Sea-Level and Depth Measurement

ueoaeuc aybiem s
and low water at spring and Neap tides. These are the times o f greatest and least range in the diurnal tide heights and occur about every 2 weeks. Their cause is the relative alignment o f the sun, moon, and earth. When all three are in line at the full or new moon the tides are highest (springs tides), and when the sun and moon are widest separated at the first and third quarters o f the moon their effects tend to cancel and give the lowest (neap) tides. O f particular importance is the value taken as datum for indicating depths on charts. Theoretically, it is a position below which the lowest water level never falls. This approximates what is called lowest astronomical tide, or the position below which predictions suggest that the water will not fall under normal conditions. Different parts o f the world have different definitions o f their vertical datum, but all approximate to that outlined above. Any one country should have one main datum, with any subsidiary ones determined in the same manner as the main one and connected to it by geodetic leveling. Otherwise if leveling operations started from one datum and closed on to another, there would be undue discrepancies if the basic criteria for their establishment had been different. Recording How can the mean sea level be recorded? The simplest method is to plant a graduated pole (figure 44a) in the water at the required location. Then at regular intervals say every hour take readings at the water level on the pole. This process has the obvious disadvantages o f being tedious, not particularly accurate, and occurring at discrete intervals o f time rather than being continuous. A t the other extreme is the automatically recording tide gauge (figure 44b). This has continuous recording by pen and graph paper affixed to a revolving drum. One revolution o f the drum usually equates to 24 hours. The pen is linked to a float so that as the float moves up and down, so does the pen. Various forms o f damping device are incorporated to smooth out the surges o f the waves and latest developments incorporate automatic recording onto a data storage and computer facility. One notable modem advance in this area is the air-acoustic sensor system connected to a microprocessor data acquisition system. Resolutions to the milli meter are readily obtained. The principle behind the system is that a transducer, set in a tube, emits a pulse from a fixed position toward the water surface. At a known position along its path, some o f the pulse hits a bar placed in the path and is reflected back. The rest o f the pulse continues to the water and is reflected from its surface. The first part o f the pulse that reflected from the bar travels over a known path length, and is used to calibrate the system for the velocity o f sound. Hence the value found for the overall reflection can be used to determine the position o f the water surface below the transducer. With either o f these forms o f tide gauge or with the variations on these themes (figure 44a,c), it is particularly important to select the site carefully.

It should be near the open sea but not unduly affected by abnormal wind and storm conditions. I f the gauge is to be monitored for many years, then it should be at a site that does not silt up quickly. Any change o f the seabed in this way can affect the readings. In all cases the tide-gauge zero mark should be accurately located in terms o f how far it is below a permanent first-order benchmark to which subsequent level measurements can be easily related.

C o rio lis E ffe ct

The rotation o f the earth causes what is known as the Coriolis effect. Besides having an influence on tidal movements, it also affects the movement o f artificial satellites. Since water is stationary in relation to the earth, as the earth rotates, so the water near the equator is rotating at some 1400 km/hr while toward the poles, the rotation slows. I f various forces then act to move a particle o f water toward the east in relation to the earth, the particle will actually begin to drift toward the south because it would be moving faster than normal for its latitude. I f we followed the particle for awhile, we would observe this tendency moving it steadily in a clockwise direction if north o f the equator and counterclockwise if to the south o f the equator. In a similar manner a satellite launched toward due north from near the equator would have both its own launch velocity and the velocity due to the rotation o f the earth. This combination o f components would cause it to drift east o f its initial due-north direction and to trace out a path over the earths surface that curved to the east. As with the water, so the satellite to the south o f the equator would appear to drift in the opposite direction.

S a t e llit e A l t i m e t r y
A few satellites are now able to measure heights above the mean wave surface o f the oceans. For this there is no possibility o f a ground transmitter or receiver at the required height position, so the measure has to be made in terms o f the time a signal takes to make the round trip from the satellite to the sea and back again. From a sequence o f such signals a mean value is obtained; however, as with echo sounding to the ocean floor, it is no good having a depth if its position is unknown. In the case o f satellite altimetry (figure 45), it is the satellites position that has to be accurately determined for the time o f measurement. This is achieved by observing the satellite by either laser ranging or Doppler techniques that allow determination o f the orbit. Then from knowledge o f the orientation o f the transmitter/receiver signals, the depth to the wave surface can be calcu lated. The value found will not be mean sea level because o f tidal, atmospheric, and other effects; nevertheless, it provides useful data for the determination o f mean sea level and, more particularly, for the geoid surface, which, strictly speaking, differs from mean sea level by small amounts. This arises because i

the acceptance o f mean sea level as an equipotential sin hu e is affected by a variety o f perturbing effects that cause a meter or so in variation.

T ra n s fo rm a tio n s
With such a multiplicity o f systems worldwide it is essential to be able to convert (or transform) values on one system to their equivalent on another system. I f one were thinking only o f a small area (few square miles) ovei which the values are spread, then a very simple twopoint transformation will suffice. This requires only two points common to the two coordinate systems in question. From such a minimum o f data it is possible to allow for differences between the systems that involve scale, unit o f measure (for example, from a system in imperial units to one in metric units), orientation, and origin. In such an arrangement, however, there are no redundant observations, no checks on gross errors. Two simple formulae are all that is required to convert the values o f any number o f points within that limited area. However, in the more appropriate geodetic setting, seven parameters are required to cover a scale factor, rotation around the three axes, and movement o f the origin in three directions. It is possible to convert between any two coordinate systems whether their positions are defined in cartesian form (that is, east, north, and height) or in geographic form (latitude, longitude, and height). References such as that by Soler and Hothem (1988) give a good lead into this subject. Typical datum shifts to WGS 84 are: From OS(SN)80 WGS 72 NAD 83 NAD 27 Indian dX +364.705 0 0.42 4 227 dY -109.420 0 0.95 166 803 dZ +429.003 + 4.5 - 0.62 183 274 Scale 1.2 ppm +0.23 -0.14 0.37 6.59

It can be seen from just these few examples that shifts can range from nothing to many hundreds o f meters. There are various methods by which the transformations can be executed. All are beyond the scope o f this volume, but Deakin et al. (1994) for example, compare the results from least-squares collocation with the three-parameter method o f translations only and the seven-parameter one o f translations, rotation, and scale factor. Th e differences between the results in their particu lar example and the published values for 11 stations were, in all except three cases out o f the 66, less than 1 m. N o particular conclusion can be drawn as to which is the preferable approach, and no doubt this is an area where discussion will continue for some time to come.

...
.

Center of Earth 4 6 . Deviation Due to M ountain Mass

CH.APTER

Physical Geodesy

T" his chapter introduces a branch o f geodetic science that is closely related to geology and geophysics. The biasic distinction between physical geodesy and geometrical geodesy is that the formeer utilizes measurements and characteristics o f the earths gravity field, in addition to theories regarding this field, to deduce the shape o f the earth and, in combination with arc measurements, the earths size; the latter, as we have seen, is concerned with the methods o f measurement and computation on a curved surface and the determination o f the parameters o f the reference ellipsoid. With sufficient information re garding the earths gravity field, it is possible to determine geoidal undulations, gravimetric deflections, and the (earths flattening. Pierre Bouguer in 1738 endeavored to quantify the deflection o f the vertical during the Peru survey expeditio n. H e made observations near Mount Chim borazo in the Andes but conceded that the inaccuracies in his method were probably equivalent in magnitude to the quantity he was trying to determine. Thus his results were o f little use other than to pave the way for future investi gators. In 1774 Nevil Maskelyne, Astronom er Royal at Greenwich, attempted to measure the effect o f attraction in relation to Mount Schehallion in Scotland, an east-west ridge standing 600 m above a fairly level area. Figure 46 shows attractions < 5 , and & at points A and B together with zenith distances a and (3 to a star. From knowledge o f R (radius o f the earth) and 9 (the angle subtended by A and B at the center o f the Earth) found by survey methods, the amount o f deflection could be calculated. Maskelyne found a residual displacement o f 11.7" o f arc as the attraction and concluded that the earth had a mean density 4.867 to 4.559 times that o f water. (Th e present-day accepted value is 5.52.)

-Topography jL ________ Ocean

-------- -----------------t ----------------------- T - Sea level

1 ----------- '" t

Depth of com pensation

Equal pressure surface (a) P ra tts Hypothesis

(b) A irys Hypothesis 4 7 . Isostatic Hypotheses

ISOSTASY Following the efforts o f Bouguer there were a number o f attempts to develop a theory for the distribution o f masses within the earth that would satisfy observations. Was there in fact a deficiency o f material below a mountain? Were there large, dense, hidden masses scattered through the earth? The theories o f the early 19th century allowed estimates to be made o f the attraction caused by the Himalaya mountains. One notable personality was Archdeacon John Pratt, who calculated theoretical values for the possible attraction. However, when the Survey o f India was observing its great triangula tion, the differences between astronomically observed and geodetically com puted locations gave the attraction to be only about one-third the theoreti cal value. T o explain this discrepancy, both Pratt and the then Astronomer Royal at Greenwich, Sir George Airy, put forward conflicting ideas (figure 47) based on their thoughts regarding the distribution o f masses within the earth. Pratt (in 1854) suggested that the crust o f the earth was uniformly thick below sea level but that the density varied to account for the topography. Airy (in 1855), on the other hand, considered that beneath mountains the earth crust dipped downwards like the root o f a tooth, so that down to some particular level in the earths crust the total mass per unit area would be the same. These theories o f the earths crustal structure now form part o f the science o f isostasy, although it was some years after the time o f Pratt and Airy when that term was introduced. Its basic concept is that the outer part o f the earth adjusts the distribution o f its component masses to reach equilibrium. Thus, to obtain reasonably accurate theoretical values o f the gravitational field throughout the earth, we must consider the effect o f isostasy. In using the earths gravitational field to determine the shape o f the earth, the gravimetrist (someone who measures gravity) measures the acceleration o f gravity at or near the surface o f the earth. It is interesting to compare the acceleration measured by the gravimetrist and the acceleration experienced in an airplane. In the latter, an acceleration is simply called a g force and is measured by a g meter. A g factor o f unity is used to indicate the acceleration due to the attraction o f the earth and is considered a neutral condition. The gravity unit used and measured in geodesy is much smaller: A g factor o f 1 is approximately equal to 1000 gal. Thus in geodesy, where the milligal (see p. 115) is the basic unit, we are dealing with variations in acceleration equal to 1 millionth o f a 1-g aircraft acceleration. In fact, modern intruments allow the ready resolution o f 0.01 mgal.

Ik

/f

Negative anomalies

\ \
\

Positive anomalies

\ /

I D eflection of the vertical

\\\\\'

4 8 . E ffect of Mass Anom alies

G R A V ITY A N O M A L IE S
By assuming the earth to be a regular surface without mountains or oceans, with no variations in rock densities or in the thickness o f the crust, a theoretical value o f gravity can be computed for any point by a simple mathematical formula. This theoretical value represents the combined effect o f the force o f the earths attraction due to gravitation and the centrifugal force due to the rotation o f the earth. The theoretical value o f gravity at a point on the ellipsoids surface depends both on the size and shape o f the ellipsoid and on a value, computed from observational data, that is accepted as the theoreti cal value o f gravity at the equator. It varies only with the latitude o f the observation point if the figure o f the earth is taken to be an ellipsoid o f revolution, although some experts have tried to introduce a longitude term as well. While there are several formulae for computing the theoretical gravity, the one most commonly used today, resulting from modern evaluations o f the earths parameters, is based on the Geodetic Reference System 1967 (GRS 67). Before then the formula used was based on the International Ellipsoid (1924). The gravity formula 1967 is given by y = 978 031.85 (1 + 0.005 278 895 sin2 c f> - 0.000 023 462 sin4 c P) where e p = geodetic latitude y = normal gravity in milligals (mgal) at the ellipsoidal surface The differences between theoretical and observed values are termed anoma lies. Used in this sense, an anomaly is a deviation from the normal and can be used either for a single point or to describe a regional effect (figure 48). T o make use o f the anomalies, the observed gravity must be reduced to a common frame o f reference the geoid or mean sea level. In an ideal situation all anomalies would be zero that is, theoretical values would equal observed values but this does not happen. The calculation o f possible causes for these

Earths

(a ) H eight h of observation point P (b ) Mass between mean sea level and height h @ Effects of terrain surplus and deficiency 4 9 . Factors Involved in Reducing Observed Gravity

anomalies can he achieved according to different assumptions. The reductions may take into account the elevation above or below sea level, account for the terrain surrounding the point, and the assumed structure o f the earths crust (figure 49). The position o f the point must be established with sufficient accuracy to meet the gravity survey requirements. There must also be accurate enough topographic information o f the distant as well as surrounding area to satisfy reduction needs. Three particular methods o f assessing for the anomalies are worthy o f mention here.

50. Gravity Anomalies

F re e -A ir A n o m a ly
I he free-air anomaly reduces the observed value to the geoidal surface on the assumption that there is no mass between the earths surface and the geoidal surface. This height correction is 0.3086 mgal for every meter above sea level. In other words, the point o f observation is considered to be in free air above the geoidal surface, and while the height is taken into account, the intervening mass is not. However, where there is a mass above the observation point that will affect the reading, (such as for point C in figure 50a), then the free-air anomaly does not account for this effect. Free-air anomalies correlate positively to the topography and are assumed to be balanced by the compensation for depth. Except in mountainous areas, this is a reasonably accurate method.

B o u g u e r A n o m a ly
This is an extension o f the free-air anomaly with an allowance for the topogra phy between the two surfaces. The Bouguer anomaly assumes a constant density value for that mass and that the area around the observation point is a level plane. Thus, if the point is on a sharp peak (such as B in figure 50b) correction is made for a nonexistent mass surrounding it, and a positive correction is needed. The converse applies to a point in a valley (such as C in 50b), where the observed value has been overdiminished and needs a positive correction. An approximate value o f this correction is 0.1119 mgal for every meter o f ele vation. This correction, together with the application o f the free-air reduction, gives the Bouguer reduction. What this does not allow for is the nearby topography. Corrections for the effect o f topography are made to a radius o f at least 166.6 km from the point o f observation. This apparently strange value arises from the sizes given to a series o f zones and compartments around a point, as devised byjohn Hayford in 1907 to aid the reduction o f gravity observations. It equates to 100 miles, the value Hayford no doubt initially used. The mean elevation o f each compartment is used in the topographic reduction. For ocean areas, any measured values have to be corrected up to sea level according to the depth at which they were recorded usually the depth at which a submarine took the instruments. These anomalies are extensively used in geophysical prospecting. Over wide areas they are closely related to height and are generally strongly negative for mountain stations.

119

G ravim etric defle ctio n of the vertical

51. Products of G ravim etric Method

Is o s ta tic A n o m a lie s These can be determined according to a particular hypothesis. In addition to the hypotheses o f Pratt and Airy is a variation o f the Pratt hypothesis developed by Hayford early in this century. Other scientists, such as William Bowie, Friedrich Helmert, and Weiko Heiskanen, contributed similar ones. The theoretical value at sea level is corrected for elevation, for the total global terrain effect o f topography, and for compensation according to the chosen hypothesis. Application o f the total global terrain effect to the Bouguer anomaly gives the isostatic anomaly according to the chosen hypothesis. The basic variation between the various hypotheses is the chosen depth o f compen sation, usually taken to be either 96 km (60 miles) or 115.7 km (about 70 miles). Best results are obtained by varying the compensation depths re gionally. There are two schools o f thought on the reduction o f gravity observations. One approach adjusts the observed values and compares the results with the theoretical values (usually by countries outside the United States), while the other does the reverse, adjusting the theoretical values for comparison with the observed values (particularly in the United States). Obviously both meth ods should produce identical results.

G R A V IM E TR IC M E T H O D
The method providing the basis for determining the undulations o f the geoid from gravity data was published in 1849 by the British scientist Sir George Stokes (1819-1903). However, the lack o f sufficient observed gravity data prevented its application for 100 years. In 1928 the Dutch scientist Felix Vening Meinesz (1887-1966) developed the work o f Stokes into formulae by which the gravimetric deflection o f the verdcal could be computed (figure 51). Computing the undulations o f the geoid and the deflections o f the vertical requires extensive gravity observations. The areas immediately surrounding the computation point require a dense coverage o f gravity observations, and detailed data must be obtained out to a distance o f about 800 km (500 miles). A less dense network is required for the remaining portion o f the earth. While the observational requirements for these computations appear enormous, computer facilities now available relieve much o f the tedium, and the results well justify the necessary survey work. Th e observed elements o f the gravimetric method are the gravity anomalies the differences between the observed gravity value properly reduced to sea level, and the theoretical gravity value obtained from the international gravity formula. The gravity anomalies are caused by either local or regional mass variations, which may be visible or invisible, topographic or subsurface.

52. Geoid-EIIipsoid Relationships

Figure 52 shows how the mass surplus of the mountains and the mass deficiency o f the oceans cause the deflection o f the vertical and the undula tions o f the geoid. A mountain mass appears to pull the plumb line from the normal o f the ellipsoid. In a similar way, the mass deficiency o f the oceans appear to push the plumb line. These effects o f the mass anomalies contribute to the deflection o f the vertical. However, both deflections o f the vertical and undulation values result from density variations throughout the earth. In the area o f mass surplus, the observed gravity (reduced to sea level by the elevation correction only) is generally greater than the theoretical values, so the anomalies are positive. In the areas o f mass deficiency, the observed gravity, reduced in the same way, is generally smaller than the theoret ical value, so the anomalies are negative. The deflections and undulations computed with sufficient gravity informa tion are considered absolute values referred to an earth-centered ellipsoid. In other words, the axis o f rotation for the ellipsoid coincides with the earths center o f gravity. Effective use o f the gravimetric method is dependent only on the availability o f anomalies in sufficient quantity to achieve the accuracy desired. Successful use o f Stokess integral and Vening Meineszs formulae depends on a good knowledge o f gravity anomalies in the immediate vicinity o f the point under consideration and a general knowledge o f anomalies for the entire earth.

53. Shoran Connection between Scotland and Norway

CHAPTER

7
World Geodetic System

B e c a u s e o f the failure o f local or even preferred systems to provide adequate intercontinental geodetic information, a unified world system became essen tial. T o develop such a system, all o f the military services actively participated in a program to develop a U.S. Department o f Defense W orld Geodetic System (D O D WGS 60). T o establish a world system, it was necessary to consider all the available observed data to determine the absolute reference system that best provided a good fit for the entire earth. Once such a system was estab lished, all the major geodetic networks o f the world could be unified and the coordinates o f points anywhere on the earth would be compatible.

D E R IV A TIO N O F W O R L D G E O D E T IC S Y S T E M S
The development o f a world geodetic system, however, would require enor mous effort, particularly in the following areas: 1. establishing intercontinental links 2. measuring as much gravity data as possible 3. deriving approximate gravity values in blank areas
4. defining a suitable ellipsoid

5. deriving the gravimetric orientation Let us consider each o f these areas in turn.

In te rc o n tin e n ta l L in k s
Because o f its intercontinental-range missile capability, the U.S. A ir Force played an active role with other agencies over a 10-year period to gather sufficient data to formulate a unified world geodetic system. O f particular use in this work was applying the results from Shoran and Hiran. Shoran (Short Range Navigation) was a system for measuring distances to the required target from either a vehicle or fixed ground stations. Other types o f system had the receiver at the required point and range information dictated from fixed ground stations. Shoran was developed in the United States for use in W orld War II, but in 1944 was experimented with for civilian survey use under the guidance o f Carl Aslakson at the USCGS (U.S. Coast and Geodetic Survey). Experiments consisted o f flying at high-altitude-3000 + m (10 000 + ft ) across the line joinin g two stations. Ranges recorded to each station during repeated crossings gave results with a standard error for the total range equivalent to about 1 in 100 000, although in absolute terms for the distance, this was considerably downgraded by inaccuracies in the adopted velocity o f light. (At that time Bergstrand was only just beginning his work on the velocity o f light.) The technique showed promise, particularly in the development o f large trilateration networks in remote areas and o f long-distance ties such as between Norway and Scotland (figure 53) and from Crete to North Africa. Hiran ( high-precision Shoran ) was developed from Shoran in 1949 to give higher precision. The basic mode o f operation was similar to Shoran, although Hiran incorporated improvements in the electronics. Tests over similar long lines using a better value for the velocity o f light gave standard errors in the distance o f up to 1 in 100 000. The Hiran tie from Canada to Norway permitted connection o f the North American and European datums. However, since both o f these datums are based on ellipsoids, which provide a good fit only in the areas o f their origins, connecting the two datums through the Hiran tie overextended the usefulness o f either ellipsoid. Nevertheless, ties made with Hiran provided an effective check on the two major datums oriented to a common ellipsoid. Other Hiran trilateration loops were valuable in the evaluation and extension o f the DOD WGS 60. As illustrated in figure 53, while Shoran and Hiran were forms o f trilatera tion, the geometry o f the networks was necessarily different from a normal land-based trilateration (see figure 27).

C o l l e c t i o n o f G ra vity D a t a A major area in which preparatory work wa. done in anticipation o f a world geodetic system was in the collection and analysis o f gravity observations. Through an extensive program several thoisand gravity base-reference sta tions were interconnected throughout the vorld. The connection o f these base stations by he easily transported gravimeter permitted the reduction o f numerous gravity ibservations to a common usable system. In effect, it established a world gravty system by including data col lected for oil prospecting and other geophyiical purposes. The collection o f extensive gravity data Facilitated preparation o f mean gravity anomalies for 1 X 1 and 5 X 5 giographic sectors as needed for further gravimetric computations. It also provided data for the preparation o f the gravimetric geoid used for orienting the ellipsoid for the Army Map Service World Geodetic System (see page 135). Much o f the work in the collection and analysis o f the gravity data was cone by contract with universities.

G r a v i t y V a l u e s in B la n k A r e a s
An additional phase o f gravity work involvtd computing gravity anomalies for places where there were no observations. As stated before, gravimetric computations require a knowledge o f gravity anomalies over the entire earth. Some regions o f the earth were completely void o f observation points. For these the best substitute was found by considering the topography or bathyme try (underwater topography) near the required positions and making certain applications o f knowledge o f the earths crust. Although the computations for each position were tedious and involved consideration o f the entire earth, the approximate anomalies were better than blank observation squares on the gravity anomaly map. However, filling blank areas with approximate anomalies was not a substi tute for actual observations when the latter were possible. Thus work continues gathering data by land, oceanic, and airborne (including satellite) gravity surveys.

54. Arc Method of D eterm ining Earth's Size

D e r i v a t i o n o f E l l ip s o i d
Determination o f the size and shape o f the ellipsoid o f revolution that provides a close approximation to the true figure o f the earth is essential to the establish ment o f a world geodetic system. There are several ways in which this informa tion can be obtained. The historical synopsis in Chapter 1 described the methods by which the ancient Greeks and others obtained their ideas o f the size o f the earth. The method o f arc measurement has not been changed in principle to the present day, although certain important refinements have been made. Basically, a long arc was measured on the earths surface, and by means o f astronomical observations the angle subtended by the measured arc was determined from the difference in latitude or longitude. (Although most arcs lay along a meridian, some were deliberately set along a line o f latitude). By using appropriate formulae it was possible to compute the value o f the semima jo r axis and the flattening. From this procedure it was but a simple step to combine all available arcs into one calculation so that the result represented a wider area o f the earth. It has already been indicated that astronomical observations are based on the geoid. This causes errors in arc measurements that must be eliminated in order to obtain accurate results. Figure 54 illustrates how the deflection o f the vertical can cause an error in computing the radius o f the arc. The arc AB, where the geoidal surface lies below the reference ellipsoid, gives too large a radius Ru The arc BC, where the geoidal surface is above the ellipsoid, gives too small a radius R2 . This error can be reduced by considering the effect o f topography and the structure o f the earths crust surrounding the points where the astronomical observations are made. Still better results can be obtained if the deflections are computed by the gravimetric method. Re moving the deflection o f the vertical at points A, B, and C eliminates the error, so that the two arcs AB and BC give the real radius o f the ellipsoid. The ellipsoid used for the world geodetic system was based on all available information. The data reduced from arc measurements in the Eastern and Western Hemispheres as well as the connecting tie resulting from the North Atlantic Hiran trilateration arc were all considered. The arcs provided informa tion for determining the size o f the ellipsoid. Artificial earth satellites furnished observational data for determining the earths flattening. For example, the 1958 Vanguard, satellite had been tracked several thousands o f times, and data reduced from observations indicated a probable flattening factor o f 1/298. This value had been suggested earlier by Friedrich Helm ert and also by Theodosy Krassovsky, who used it as an element o f the 1942 Pulkovo datum. More recently, satellites designed specifically to collect relevant data have m odified the accepted value just slightly, to 1/298.26.

129

Axis of

55. Preferred Datums Are Relative Datums

G ra v im e tric O rie n ta tio n

The preferred datums (figure 55) were given a relative orientation with respect to a certain portion o f the geoid by the methods already outlined (see figure 54). T o convert each o f them into a single earth-centered world system, it was necessary to change the size and shape o f each ellipsoid and reorient them so that the center o f the new ellipsoid would coincide with the center o f the earth and the semimajor axis would coincide with the earths axis o f rotation. This reorientation was achieved by the gravimetric method. T o obtain an absolute gravimetric orientation for the preferred datums (he following procedure was necessary: Compute the gravimetric geoidal height and deflection of the vertical for the initial point of the geodetic system Theoretically the absolute undulation and deflection components for one astrogeodetic point in each geodetic datum would be sufficient to reduce the datums to a world system. However, by using the mean o f the gravimetric quantities o f several stations around a selected point the inaccuracies likely to result from error in the astronomical coordinates or gravity information around a single point are minimized. For this reason, the gravimetric components for deflection o f the vertical and geoidal undulation were computed for a group o f astrogeodetic stations in the vicinity o f the origin o f the datum, and the astrogeodetic minus gravimetric differences were transferred to the origin by transformation formulae. The results were further improved by repeating the procedure for a second area within the same datum and taking a weighted mean o f the results for the two areas.

Change in de flection

56. G ravim etric Datum O rientation

As a correction, apply the mean of the differences between gravimetric and astrogeodetic deflection components and geoidal undulation at the initial point See figure 56. Use transformation formulae such as those devised by Vening Meinesz or J. de Graaff-Hunter to compute the correction for each station in the preferred datum This procedure was applied to the North American, European and Tokyo datums to establish the USAF World Geodetic System 1958. Later revised to incorporate additional survey and satellite information, it became the USAF W orld Geodetic System 1959. Deflection o f the vertical components were computed for each o f some 30 points by the Vening Meinesz procedure and geoidal heights were determined according to Stokess formulae.

57. D eflections of Vertical Computed from E arths Gravity Can Determ ine a World Geodetic System

A R M Y M A P S E R V IC E W O R L D G E O D E T IC S Y S T E M
Shortly after the USAF W orld Geodetic System o f 1958 became operational, the U.S. Army Map Service (AMS) published its results o f work on the develop ment o f a world system. Its fundamental concepts were consistent with those o f the Air Force. Th e Army studies relied largely on geoidal heights for determination o f the best-fitting ellipsoid and for obtaining the necessary earth-centered orien tation o f the ellipsoid. Using deflections at numerous astrogeodetic stations, astrogeodetic geoidal undulations were computed for the North American datum in North and South America and for the European datum in Africa and the Eurasian continent (figure 57). These were compared with gravimetrically computed undulations, and an adjustment was used to minimize the differ ences between the astrogeodetic and gravimetric geoids. By matching the relative astrogeodetic geoids o f the preferred datums with an earth-centered gravimetric geoid, the preferred datums were reduced to an earth-centered orientation, thereby producing the U.S. Arm ys tentative world datum o f 1958. In 1959 after additional information Hiran data and a better flattening factor was incorporated, the improved system became the U.S. Army W orld Geodetic System 1959.

U .S . D E P A R T M E N T O F D E F E N S E W O R L D G E O D E T I C S Y S T E M 1960 (D O D W G S 60)
The Aeronautical Chart and Information Center (ACIC ) and AMS had each developed a world geodetic system by different procedures. ACIC derived the USAF W orld Geodetic System 1959 using differences between the astrogeo detic and gravimetric deflections at specifically selected stations in the areas o f the major datums. AMS derived the U.S. Army W orld Geodetic System 1959 using a comparison o f astrogeodetic and gravimetric geoidal heights in the areas o f the major triangulation systems. Both systems incorporated satel lite observational data and intercontinental Hiran trilatration information. Since the two systems agreed remarkably well for the North American, European, and Tokyo datum areas, they were consolidated into a single system called the Department o f Defense World Geodetic System 1960 (WGS 60). The WGS 60 provided a single, unified, earth-centered system through which accurate distance and direction information could be determined for missiles o f all ranges.

L A TE R W O R L D G E O D E T IC S Y S TE M S
WGS 60 served well for some 12 years or so, but the increasing use o f satellites drove the need for an even better system.

------------------------------------------------------------------------------------------

A ll processing o f the global positioning system (( IPS) observations was then perform ed on WGS 72, which was an earth-centered and earth-fixed system. The final WGS 72 latitudes and longitudes could be converted back to a specified datum, but there were problems with heights unless the geoidellipsoid separations were known for that area. There is also a WGS 72 ellipsoid, where the semimajor axis a = 6 378 135 m and 1//= 298.26. Yet another new system WGS 84 has now been adopted and is taking over the previous role o f WGS 72. WGS 84 is almost identical with NAD 83. Although they are both defined as geocentric, they are, according to latest information, a meter or two away from being exactly so.

O TH ER

N O TAB LE SYSTEM S

New systems and new abbreviations appear almost overnight, and it is quite impossible, and even unnecessary in some instances, to keep up with all o f them. Several pages could be devoted to listing and explaining the abbrevia tions, but just a few o f the more important ones will suffice here.

EUREF
A subcommission o f the International Association o f Geodesy (IA G ) was re sponsible for defining the European Reference Frame (EUREF), a true tridi mensional control network covering Europe. Th e frame was defined using very high accuracy techniques such as very long baseline interferometry (V L B I) where a few centimeters can be achieved over thousands o f kilometers.

E T R F 89
Th e European Terrestrial Reference Frame 1989. (ETRF 89) resulted from EUREF and aimed to give more precise coordinates throughout Europe. It is compatible with WGS 84 to within 1 m and is the fundamental geod etic reference frame in Europe. It is not directly comparable to either the OSGB 36 (Ordnance Survey Great Britain 1936) or its equivalents in other countries, and the differences are not uniform. ETRF 89 was used to transform the GPS coordinates in terms o f WGS 84 to their equivalents on the Ordnance Survey Geodetic Reference System, 1980 (OSGRS 80) and thence to the national grid, and vice versa. The procedure, for which tables have been derived, is to convert the GPS values to plane coordinates (or OSGRS 80) on the Transverse Mercator projection. These values are then subject to a shift to get them to the national grid. ETRF 89 is based on the GRS 80 spheroid. W hile it is now reasonably in sympathy with ITRF (see below) and WGS 84, tectonic movements over the continent are o f the order o f 2.5 mm yr, so that as time goes by, ETRF 89 will differ increasingly from ITRF.

ITR S and ITR F

The International Terrestrial Reference System (ITR S) has been adopted by the International Union for Geodesy and Geophysics (IU G G ) for geodetic and geodynamic purposes. ITRS is maintained by the International Earth Rotation Service (IERS), which was founded in 1987 to embrace the work previously carried out by several other bodies. IERS has the task o f defining and maintaining the ITRS through high-precision techniques such as VLBI, lunar laser ranging (L L R ), and satellite laser ranging (SLR). The origin o f the system is considered to be within 0.1 m o f the earths center o f mass. In addition, IERS is active in relating tide gauges to International Terrestrial Reference Frame (IT R F ) and carrying out a long-term monitoring exercise in three dimensions. The aim is to determine the variations in sea level without some o f the disturbing parameters.

CHAPTER

8
Satellite Geodesy

W i t h the rapid increase in industrial technology after W orld War II came the requirement for improved geodetic information. Major surveys and posi tioning techniques were still laborious and difficult, as well as dangerous, in the more inaccessible parts o f the world. There was a need to integrate the surveys on a worldwide basis to relate all the major datums o f different continents. Traditional survey techniques were not sufficient without the injec tion o f a new approach. Rocket developments during the war provided the required stimulus. In theory, had it been possible to position objects in predictable orbits around the earth, a considerable step would have been made in pushing the frontiers o f geodesy forward. The breakthrough came in 1957 with the launch o f the first artificial satellites, but hand in hand with this was the need for improved tracking methods and the development o f high-speed computers. Without the tracking facilities, prediction o f the orbital changes would be impossible yet if the changes were observable the amount o f calculation re quired in short intervals o f time would be unachievable. Thus it was essential for all three satellites, tracking, and computing to progress in parallel.

E A R L Y S A T E L L IT E S
The first Sputnik was launched from Russia on October 4, 1957. It weighed 184 lb (83 k g ), was powered by ordinary chemical batteries, and emitted radio signals for just 3 weeks. With an average orbital height o f 360 miles (576 km), it orbited in 96.2 minutes. The orbit was elliptical, varying from 136 miles (216 km) to 585 miles (1136 km), but air drag and gravity rapidly decreased

these values so that it completed only 1400 orbits in a life span o f 92 days. Although it had no experimental scientific instrumentation on board, it repre sented the forward step that science had been waiting for. The second Sputnik, launched a month later, lasted for nearly 2400 orbits and had basic instrumentation for data collection. America entered the space race in February 1958 with Explorer I, which was in a higher orbit than the Sputniks, varying from 222 miles (355 km) to 1593 miles (2551 km), and as a result had a life o f 4 years. From the geodetic point o f view it was Americas Vanguard I launched in March 1958, with an even higher orbit, varying from 410 miles (656 km) to 2468 miles (3950 km), and estimated lifetime o f 100 years that made the most significant contribution. It is interesting to note that this breakthrough for surveying was very much contemporary with another major advance EDM. Although first used in the Geodimeter in the 1940s, toward the end o f the 1950s such equipment began to be accepted as an everyday tool and additional models became available (see Chapter 14). With the early satellites there were three main tracking systems optical, radio, and radar. Optical techniques were affected by the size o f the satellite as well as by adverse weather conditions. Radio methods had fewer problems but could not produce given accuracies as readily as the optical methods. Radar similarly had few problems, except that a satellite is a very small target at a considerable distance. The simplest optical method o f using the naked eye was not at all feasible and better results were possible with either an ordinary theodolite or a kinetheodolite (a theodolite with camera facilities). When a satellite was photo graphed against a background o f stars it was theoretically possible to resolve the angular position to around 3" o f arc. This presented some difficulties, however, when it came to higher accuracies because o f the size o f the required camera, such as the Baker-Nunn camera, and the visibility problems o f such an arrangement. Satellites designed specifically for optical observation con tained a light source which could be activated electronically for photographic purposes. A large satellite with no independent source o f illumination could be photographed at dawn or dusk when reflected sunlight shone on it. Addi tionally, the satellite needed to be bright enough both to allow its location, and to leave an image on the film. Radio methods could use either interferometry or Doppler techniques (see below ).

P rinciple of Radio Interferom etry

IN T E R F E R O M E T R Y
Interferometry (figure 58) required that two aerials be positioned at the same height and set at a distance apart equal to an integer number n o f wavelength A o f a particular signal. When the two sets o f received signals were in phase, then the difference in distance from the satellite to each aerial would also be an integer number o f wavelengths. I f the signals completely interfered, then the difference in distance was a multiple o f half the wavelength. At long wavelengths there were problems with bending o f the radio waves.

DOPPLER
The Doppler technique was first investigated by Christian Doppler in 1842. One o f its early survey applications was its use in Antarctica during 1972; at that time it was particularly bulky, although accuracies o f a handful o f meters were possible. The Doppler phenomenon is usually explained in terms o f the movement o f a noise source in relation to a stationary observer (figure 59). I f the source moves in such a way that it remains at a constant distance from the observer (that is, in a circle), then the pitch will be constant because the wave always travels the same distance. I f the source travels in a straight line, it must move toward and then away from the observer. As it does this, the waves have decreasing distances to travel; as they appear to pile up, the detected pitch or frequency increases. As the source moves away from the observer, the reverse effect becomes evident. I f a second observer is farther away from the source, then the change in pitch would be more gradual. I f the sounds at the observer position were recorded and analyzed, it would be possible to determine how far the source was from the observer. Note that an example like a train involves sound waves, while a satellite involves radio waves. The effect is the same, but radio waves travel at about 300 000 km/sec, while sound travels at only 330 m/sec. The basic relationship is that change in frequency = frequency X V/C where V = velocity o f source C = velocity o f radio waves Radar methods sent short wavelength pulses (1 mm to 10 m) to the satellite and endeavored to monitor the amount o f reflected energy. From this, use o f signal travel time and the velocity o f light gave the distance. For

S 3 = point of nearest approach 5 2 - S 3 = signal increases pitch 5 3 - S 4 = signal decreases pitch

S atellite o rb it

59. Doppler p rin cip le

accurate results the beam had to be narrow, which, in its turn, made it more difficult to hit the target. Note that the velocity o f light and the velocity o f radio waves are the same when in a vacuum. In the atmosphere they vary slightly due to different effects o f pressure, temperature, and humidity, but for the purposes o f this text, we can consider them equal at about 300 000 km/sec.

S atellite orbits

S atellite orbits

6 0 . Doppler P ositioning Methods

The Doji/iln count is (lie number o f cycles o f the beat frequency (the difference between the frequency generated at the receiver and that received there from the satellite). There are three ways o f determining the Doppler count: over discrete time intervals, as a cumulative total over a full passage, or in terms o f the time taken to receive a given number o f cycles. This last method is not suitable for field survey work. I f the count is combined with knowledge o f the satellite orbit parameters, then the position o f the receiver can be found with respect to the earths center o f mass or geocentric coordinates. Doppler observations can be reduced by at least three different methods: (1) point positioning; (2) translocation, and (3) short-arc translocation.

P o in t P o s itio n in g
The process point positioning (figure 60a) assumes that the ephemerides (set o f orbital parameters) o f each pass are correct. Data is collected for a number o f passes at a single station. The ephemerides may be either precise or broadcast. Precise ephemerides are derived from observations over a 48hour period from a worldwide network o f over 20 monitoring stations. Accu racy o f the data is 2 to 3 m. Broadcast ephemerides are encoded onto the satellite signal so that they are immediately accessible. The data come from tracking results at four stations only, and hence accuracy is only some 20 to 30 m. Broadcast ephemerides are essentially a forecast o f the satellite position, whereas the precise ephemerides are derived from observations on the actual orbit used. For point positioning the data reduction is relatively simple but is o f a low order o f accuracy.

T ra n s lo c a tio n
Translocation (figure 60b) is a modification o f point positioning, in which two points are used. Passes are recorded simultaneously at each station. I f the points are not too far apart, some sources o f error, particularly in the ephemeri des, can be considered equal and canceling out. This allows a far better accuracy in the relative positioning. An extension o f this is to keep one receiver stationary while a second one occupies, in succession, a number o f required points rather along the lines o f the single-base method o f barometric heighting.

O rb it 1

Orbit 2

(a) O rbital Method

6 1 . S im ultaneous and O rbital Methods First Used w ith SECOR

14R

S h o rt -A r c T ra n s lo c a tio n
Short-arc translocation is a variation o f the orbital method (figures 61a and 62). It uses a portion o f the satellite arc. The network allows a rigorous adjustment procedure to be applied, but the computational requirement is considerable since it allows adjustment o f the orbit as well as the ground coordinates. It can be modified in various ways to simplify the procedure. By extrapolation from the results o f the short-arc observations, the orbit is defined for a far distant station that observes the same satellite pass. From a set o f at least three known positions, simultaneous observations (figure 61b) to the satellite allow definition o f the orbit. Stations at unknown points, distant from each other so as to be unable to observe the satellites at the same time, make separate observations. The desired ground positions are then determined by a form o f three-dimensional resection from the satellite posi tions.

= Known positions = Required position 6 2 . Short-Arc Method

147

S a te llite

(a) One S ate llite , Two Receivers: Cancels S a te llite Clock Offset.

(b) Two S atellites, One Receiver: Cancels Receiver Clock Offset. 6 3 . Single D ifference

TR A N S IT D O PPLER
Transit Doppler is a U.S. Navy Navigation Satellite System. Under development since 1958, it has been in existence since 1964 and available for civilian use since 1967. It is expected to continue at least until the late 1990s when GPS should be fully phased in. By the end o f 1981 there were six satellites in orbit for use by this method and 12 more in store. They weigh only 61 kg, are in near-circular polar orbits just over 1000 km high, and orbit about the earth every 107 minutes. From a sequence o f range measures during 10 minutes a good fix can be obtained in various configura tions (figures 63 and 64) for either a moving vessel or a stationary ground station. Point positioning is good to 1 to 5 m depending on whether precise or broadcast ephemerides are used. It has been proved to be a very reliable system; users o f this and later satellite systems appear to be growing by some 50% each year and is approaching the tens o f thousands. With the small number o f satellites there could be long (6- to 12-hour) waiting periods between available fix times. Today there are only four satellites transmitting. When there were seven in operation, an acceptable three-dimensional pass could be obtained in Great Britain almost every hour, but with the reduction to four that is not now possible. The satellites transmit at 150 and 400 MHz with very stable frequencies. The phase-modulated signals include time information as well as orbital data. The position o f the receiver in relation to the satellite is found by the Doppler shift described on page 141. Each unit o f the count is less than 1 m, so it is a very sensitive system.

S a te llite o rb it

(a) Double D ifference: Both clock offsets cancel.

(b) T riple d ifference: Both clock offsets cancel; phase am biguity is e lim inated. 6 4 . Double and T riple Difference

Any convenient satellite pass is suitable for observation if it is above the horizon in a more or less overhead orbit during some 18 minutes, which is time enough for up to 40 counts. Refraction due to the ionosphere causes the satellite to appear to be on a path o f greater than actual curvature, which tends to reduce the Doppler shift by moving the position o f the receiver up to 0.5 km. The troposphere also produces refraction errors by affecting the propagation velocity o f the signal. It is not nearly so easily detected as the ionospheric effect. The dual frequency helps to minimize errors due to refraction. Results o f transit Doppler observations are on the world geodetic system (WGS 72) and can be subsequently transformed to any required national datum. This can change the geographic positions o f a station by the equiva lent o f several hundred meters. For a relatively small outlay on a receiver, transit Doppler provides an all-weather, high-accuracy positioning system. It has found a wide range ol applications including pipeline positioning, oil-well location, determining boundaries o f all types as well as isolated points throughout the world as pai ls o f national survey schemes. Accuracies with this system can be down to 0.25 m with observing periods from a half day up to 2 days. The system has four transmitting satellites all in polar orbits o f about 1100 km, which gives them orbit periods o f about 107 minutes.

O B S E R V A TIO N A L S Y S TE M S
Initially two basic systems were used for obtaining geodetic information from artificial earth satellites through three principal observational methods. Opti cal and electronic facilities made it possible to perform various geodetic measurements by simultaneous, orbital, or short-arc methods to determine geocentric position, relate unknown positions to established points, and ex tend existing triangulation networks into unsurveyed areas. Am ong the early developments in satellite observation wasSECOR (SEquential Collation O f Range). This consisted o f a transponder that received, ampli fied, and retransmitted high-frequency signals from ground stations. While it was an all-weather system and was described as mobile, it did need a large truck to transport it by todays standards it was very large, cumbersome, and heavy. Yet it allowed determination o f the coordinates o f positions up to 2500 km (1500 miles) from known points. SECOR was developed by a contractor to the U.S. Army Corp o f Engineers. The first satellite used with it was launched in 1964. T o avoid uncertainties in the earths gravity field, SECOR was used in a geometric mode only. This mode used a total o f four ground positions one o f which had to be designated as master from which simultaneous range observations were made to a tran sponder in the satellite.

O f the four positions, only one could be unknown, then; rather along the lines o f resection, a set o f range values allowed computation o f the unknown position. In practice, o f course, many sets o f observations were taken to allow a least-squares solution. Such an operation required the three known stations to provide a well-conditioned triangle. The observations were best when the satellite was above the center o f that triangle. Th e two main sources o f error were in the system calibration and in the effects o f refraction in both the troposphere and ionosphere. Tests suggested that the achievable accuracy was around 3 ppm probable error with improve ments from night observations. In practice, the sequence o f operations required to complete a positioning was repeated every 50 milliseconds. It required a minimum o f two satellite passes. For better results, data from positions on both passes (or more if possible) were combined to give improved geometry. With four ground sta tions for SECOR it was possible to measure some 29 000 ranges during a normal 6-minute satellite pass a position fix every 1/20 second. All data were recorded on seven-channel magnetic tape and then sent to a central processing agency. In fact, the fixes did not occur exactly every 1/20 second but only on average: when the satellite was close to the stations, it could be less than 1/20; conversely, when farther away, the time could be slightly longer. The coordinates o f the unknown point will be in the same coordinate system as the known points. Such a system o f control points can be readily extended.

0
S i

S i and S2 = d ista n t galactic sources

a,b,a,p

T i and T 2 = telescopes = components used to calculate T i T2

6 5 . VLBI Baseline D eterm ination

Very Long Baseline Interferometry

In the second half o f the 1960s astronomers developed a technique, called very long baseline interferometry to improve the resolution o f their radio telescopes based on extraterrestrial objects such as quasars. Geodesists thought the idea might be applicable to geodesy and geophysics as well as astronomy. By the early 1970s geodetic experts in the United Slates were experimenting with it. The concept is that a radio signal from a quasar or similar source is received at two or more distantly separated parabolic radio telescopes that form an interferometer. The separation o f the telescopes means the signal will have different distances to travel. The accurate times o f arrival o f the signal can be processed, in conjunction with the geometry of the situation, into information relating to the relative positions o f the telescopes. The difference in recorded time o f arrival o f a signal at the two tele scopes or relative phase delay is directly "elated to the component o f the baseline in the direction o f the source (figure 65). Thus rotating the telescopes and using a series o f sources more or less paired at right angles to each other will theoretically allow calculation o f both :he length and direction o f the baseline. The signals are recorded on tape and processed at a central location rather than locally.

This description somewhat oversimplifies the process since the earth is moving in relation to the distant sources. Allowance has to be made for this effect together with other factors such as refraction, precession, nutation, and offset o f the clocks. Obviously as time differences are used it is essential to know the relationship between the two clocks. The time standard can be based on rubidium or cesium at stabilities o f at least 1 in 10", or on the hydrogen maser at 1 in 101 3 over a period o f 1000 seconds (16 to 17 minutes). The refraction effect can be around 10~8 seconds, or 2.5 m in distance. Accuracy for the resulting separation o f the telescopes is more or less independent o f length and can fall in the centimeter region even over ranges reckoned in thousands o f kilometers. Orientation is in parts o f a second o f arc. As with so many other aspects o f surveying, it is the atmosphere that is the limiting factor. In addition to determining the relative location o f the telescopes, VLBI is also the best technique for determining the earths rotation parameters, plate motion, and crustal movements. VLBI also finds application in linking survey networks o f far-flung countries in different continents, and in establishing a global geodetic reference system. The terminals o f such an arrangement are usually the electrical centers o f the antennae, so for survey purposes, local connections by traditional methods are required to connect to the nearest part o f the national survey network. VLBI and laser ranging are closely related. They can supplement one another when there is a requirement for a high-precision goedetic network as for some monitoring schemes. The most meaningful use o f VLBI, in the geodetic sense, is when at least four receiving stations are observing simultane ously. This allows the separation o f some o f the parameters and the solution o f others.

6 6 . Global P ositioning System s O rbital C onfiguration

CH A P T E R

10
Global Positioning System

BASIC DESCRIPTION I he modern approach to the use o f satellites for positioning is the NAVSTAR (Navigation System using Tim e And Ranging), more usually know as the global positioning system ( GPS). Such are the expectations o f this system that within just a few years it has replaced much o f the traditional survey work in both the horizontal and vertical planes. The prime obstacle to using it has been the cost, which started at many thousands o f dollars or pounds per unit, but which is now around 10 000 ($16 000) for two single-frequency receivers. T o make it really effective and widely used, expenditure for the equipment should decrease to no more than that for a good theodolite. In addition, the processing must all be on the spot and not require dispatch to any central location. Its aim is to allow virtually instantaneous determination o f position to a relatively low accuracy or, by occupying more observing time, a highaccuracy positioning in three dimensions. GPS is designed in such a way that, now it is fully operational, it allows continuous provision o f information worldwide. It is an American develop ment conceived in the 1970s and for which 11 prototype satellites were put into orbit between 1978 and 1985. Two generations o f GPS satellites have been developed, and a further two are scheduled for launching by the year 2000. The satellites are generally launched from a space shuttle, and the Challenger disaster o f 1986 was a serious setback to GPS. It had been hoped that the full requirement would be completed by 1989, but this was delayed by several years. The full constellation o f 24 satellites was in operation by 1994. The full system consists o f 21 active satellites and three spare ones in six different orbits (figure 66) at 55 to the equator, at heights o f 20 200 km (12 620 m l), so that orbital times are 12 sidereal hours. Power comes from solar batteries. When in high orbits in relation to the observer (that is, passing nearly overhead), the satellites will be available for 4 to 5 hours before disap pearing over the horizon. Th e arrangement o f the orbits and positions o f the

159

67. G PS Control Stations

satellites is such that at any time o f the day or night at least four should be observable on an almost worldwide basis, although there are scattered areas in the latitude regions o f 30 to 50 N and 30 to 50 S that have degraded coverage once or twice a day. Th e control o f the system is maintained from five stations spaced around the world. The master station is at Colorado Springs (39 N 105 W ) , although it was formerly at Vandenberg, California, the others are at Ascension (8 S 14 W ); Diego Garcia (10 S 75 E), Hawaii (20 N 155 W ), and Kwajalein (9 N 167 E) (figure 67). These five stations lie on four different major tectonic plates; the two major plates not represented are those for Africa and Eurasia. Essential requirements for GPS are, as for VLBI, accurate frequency stan dards and clocks. These can be rubidium, cesium, quartz crystal, or hydrogen maser. In addition, good ephemerides catalogues o f predicted positions o f satellites against time coded into satellite signals are necessary. Just as errors in the clock or in the range can affect the result, so can inaccuracies in the orbital predictions. Any ephemeris error assumes the satellite to be at an erroneous position, and this affects the range. A satellite moves some 3 km/ sec when in a 12-hour orbit, so even a small time error or prediction error can mean appreciable ground position error.

OTHER POSITIONING SYSTEMS A USSR system similar to GPS is Glonass, which also has 24 satellites but in three orbits. Glonass started in 1982 and reached full operation on January 18, 1996. Unlike GPS, this is a purely civilian system, and there are receivers available capable o f accepting both GPS and Glonass. When these dual receiv ers become fully reliable, they will have the advantage o f allowing observations to more satellites at any one time than each system does on its own. Although both systems have the same number o f operational satellites, the sets are at different orbital inclinations and heights. Glonass satellites are at a 64.8 inclination, a height o f 19 133 km (11 960 miles), and orbit some 13 minutes quicker than the GPS ones. Glonass transmits in the L band frequen cies o f 1597 to 1617 KHz and 1240 to 1260 MHz and uses either Moscow time or UTC. Its coordinates are based on the Soviet Geocentric Coordinate System 1985. Other systems should be mentioned as an indication o f the multiplicity o f orbiting bodies that are in operation. Mention has already been made on page 149 o f the Transit network. A USSR system similar to this is called Tsicada. The European Space Agency (ESA) has developed Navsat (Navigation Satel lite). Starfix is a geostationary complex o f four satellites over the Gulf o f Mexico operating in a geosponder mode only. That is, they receive signals from earth-based stations and retransmit them to other users. The United States has a proposed geostationary system, Geostar, which has yet to get past its development stage. A

GPS

I------At rest (A) (stationary) Dynamic (B) (m obile or kinem atic)

____ I____
Stand-alone ( A l)

------- 1
D ifferential (A2)

I -------Stand-alone ( B l)

------- 1
D ifferential (B 2)

I
S tatic ( A l. 1)

____ I____
Nonstop kinem atic (B 2 .1 ) Kinem atic on the fly (B 2 .2 )

I
S tatic (A 2 .1)

I----------Stop-Go kinem atic (A 2 .2 ) Pseudokinem atic (A 2.3) Rapid sta tic (A 2 .4 ) Nonstop kinem atic (A 2.5)

68. GPS Operating Modes and Techniques

GPS OPERATING MODES AND METHODS (IPS is a passive, all-weather system, available 24 hours a day. It can be used in either a stationary (at rest) or mobile (dynamic) mode, and used for either point positioning (stand alone) or relative positioning (differential) (figure 68).

S t a t i o n a r y (at R e s t ) P o s i t i o n i n g

In this mode, observations are made using a stationary receiver. Successive observations at a point can be to different satellites and at different times. The period during which the receiver has to be stationary can range from a few seconds to many hours, or in exceptional circumstances, a day or two.

B M o b ile o r K in e m a tic (D y n a m ic ) P o s itio n in g

In this m ode the receiver is in a moving vehicle. The results are used, for example, as a navigation aid at a low level o f accuracy o f 5 to BOO m. However, if this mode is used in conjunction with a second, stationary unit (that is, in a differential m ode), accuracies from 5 m down to 1 cm are possible. With a mobile receiver there need to be at least three simultaneous observations. The receiver is positioned continuously, and a full route description is obtained.

A 1 , B1

P o in t (S t a n d -A lo n e ) P o s itio n in g

This method locates individual points three-dimensionally within an estab lished coordinate system (see figure 60a). The system may be local, or it may be worldwide, based on geocentric coordinates so as to give absolute positioning. The accuracy achieved in such a situation, using only a single receiver, is 300 m or better, although the actual figure in any instance is not readily deduced.

A 2 , B 2 R e l a t i v e ( D if f e r e n tia l) P o s i t i o n i n g
This mode involves locating one point in relation to another point (known as translocation) (see figure 60b). Because o f this it is more likely to use a local coordinate system, although it may be appropriate to later transform the values into an alternative system. Transformation would also be necessary if the two points in question were themselves on different coordinate systems. Relative positioning is simpler than absolute positioning, but it requires simul taneous observations from each point. Either o f the two arrangements can be extended into a network o f points, and then more sophisticated adjustment is possible with redundant observations. Often known as Differential GPS (DGPS), accuracies o f a few millimeters are possible.

A 1 .1 , A 2 .1

S ta tic P o s itio n in g

In this technique, the receiver is stationary for around an hour, but the time will vary according to various factors and whether it is being used in a stand alone or differential mode. As with EDM, some receiver models have different ranges from others. Single-frequency receivers usually have an upper limit in this mode o f 15 km while dual-frequency models can cover 15 km and more, with no particular limit. O f all the possible variations, static positioning gives the highest accuracies and allows observation o f the longest lines, although these can take several days to achieve.

A 2 .2 S t o p -G o K in e m a tic P o sitio n in g
This is a differential or relative technique. It is a system akin to barometric heighting with one field and one stationary instrument or the single-base method. One receiver stays at a known point recording continuously. The second receiver starts at another known position and then moves successively around the unknown points, keeping lock on the satellite signals as it goes.

A 2 .3 P s e u d o -K in e m a t ic P o s itio n in g
This technique is a modification o f A2.2. As before, there is a stationary receiver at one known point, and the moving receiver starts at a second known point. However, the moving receiver switches o ff before going to the first unknown point and also when moving between other unknown points. The drawback to this method is that each point has to be reoccupied after about an hour.

A 2 .4 R a p id S ta tic P o s itio n in g
This is the latest technique and improves on A2.3 in that no reoccupation is necessary. The time spent at each unknown point is 5 to 10 minutes. Rather akin to single-base barometric heighting, a single receiver stays at a reference position and tracks continuously. A second receiver visits all the required field positions, with the receiver switched o ff between positions, and each point is treated individually in the calculations. This approach is particularly useful over short (5 km) baselines.

A 2 . 5 , B 2.1

N o n s to p K in e m a tic P o s itio n in g

In this method, despite its name, the mobile antenna has to be stationary for about 5 minutes to determine the starting point. Then it can be moved, on any sort o f vehicle or just manually, along the required route to obtain a continuous record. It is essential to have four or five satellites visible; any less and a new start has to be made.

B 2 .2 K in e m a t ic O n -t h e -F ly P o s itio n in g
This is the same as the previous method without the need for the initial stationary period. Th e initialization can be ichieved on the move and the ambiquities resolved if five or more satellites are visible.

S e e d e d R a p i d S t a t i c P o s i t i o n in g
This is a method pioneered by the University af Pretoria for use in subsidence monitoring where a good approximation to the position is known prior to the observation. I f the position is known to within one or two times the 19-cm wavelength, then it allows a resolution of the whole number ambiguities (see below).

AMBIGUITIES AND THEIR RESOLUTION Satellites are able to provide either pseudo-range measurements or carrier beat phase measurements. In pseudo-range measurement the information required is the signal frequency, the exact time o f signal transmission, and the exact time o f signal receipt. Then, in effect, one applies the relation distance = velocity X time. The prime difficulty with the pseudo-range method is synchronizing the clocks. In carrier beat phase measurement the phase o f the received signal is compared with the phase when it was transmitted. The difference between these gives only the fine part o f the reading. The integral number o f wave lengths has to be determined by other techniques. The integral number is known as the initial phase ambiguity. GPS is a one-way ranging system where the controlling transmitters and user receivers have separate clocks and hence the need to know their interrela tion or offset. Any discrepancy in the synchronization will directly affect the range by an amount related to the velocity of light (that is, 300 000 km/sec). Thus 1 km is equivalent to 3.3 microseconds (3.3 X 10 6 seconds). The accumulation o f errors from such sources as these can amount to a centimeter or two on short lines, or about 1 in 106 (1 ppm ) o f the distance as the line becomes more than a few kilometers. Some experts use the rule o f thumb that uncertainty in the baseline is given by the expression ephemeris uncertainty satellite altitude
------ :----- L-

base length X

A particular advantage o f the relative positioning approach is that some errors, common to both stations or systems, can be eliminated, or considerably reduced, with resulting improved accuracy (see figures 63 and 64). As with ground survey techniques, the configuration o f the observations can have a bearing on the accuracy o f the results. Poor strength o f figure can be detrimental.

v a iu u a i r u o m u i m l y o y s i t J i n

The operation o f NAVSTAR is based on signals continuously transmitted on two carrier frequencies o f 1575.42 M Hz (equivalent to a wavelength o f 19 cm) and 1227.60 MHz (equivalent to a wavelength o f 24 cm) multiplied up from the high-stability oscillator on the satellite o f 10.23 MHz.

CODES GPS uses two code arrangements: P (Y ), or precision code; and C/A, or coarse acquisition code, sometimes referred to as S or standard code. Th e frequency o f 1575.42 M Hz carries both the P and C /A codes, that o f 1227.60 MHz only the P code. From these two codes stem two modes o f use: one uses the codes; the other is referred to as codeless which means it uses the noise o f the system. The disadvantages o f codeless are that the ephemerides and clock synchronization data have to be obtained externally and needs both pre- and postsynchronization with an accurate source clock.

P Code
The P code, with its dual frequency, is the basis o f PPS (Precise Positioning Service) for high-precision positioning measurements and can give accuracies o f 10 to 20 m. Originally it had the drawback o f making information available only at discrete intervals rather than in a continuous manner, but this no longer applies. The P code is at present accessible to anyone who has permis sion to acquire the appropriate receiver, but ultimately this could be restricted, for reasons o f national defense, to use by the U.S. military authorities or subject to only limited civilian access.

C/A Code
The C/A code is the basis o f SPS (Standard Positioning Service) for initial signal acquisition and course position determinations, which can give an accuracy o f 20-30m.

Y Code
When the P code is encrypted to protect it against hostile imitation, it becomes known as the P (Y ) code. This technique is called antispoofing (AS), and a special module is required for the decryption o f the Y code. This form o f tranmission is restricted to military use.

RANGE VALUES W hile it would appear that there are three basic variables to be found for each three-dimensional point position (E, N, and H or longitude, latitude,

V . J IW U U I r u o i l i u i I l y

-y tn w iii

and height), there is a fourth unknown h the form o f the clock bias or offset. Thus, to solve for four variables it is recessary to have at least four range values to satellites. In fact, each range from a satellite deines a sphere on which the required point lies, and it is the intersection o f such spheres from different satellite positions that gives the result (figure 69. W e can visualize this if we imagine a bite taken out o f an apple. The outline o f the bite is the line o f intersection, and if the cut is made by a spherical obpct, then the outline is a circle. Thus if a third such sphere cuts the other two, it will produce a trisection point at a position on the circle. The uncertainies caused by the clock offset can be resolved by introducing the sphere for i fourth range. This approach is only capable o f a few meters accuracy, so surveying generally requires alternative methods. These can be differential (relative) positioning techniques that use phase neasurement, which is similar to the translocation method as used in the transit system. T o determine the range from ground station to satellite, the receiver pro duces a code similar to that sent from the satellite. Then by comparison o f the actual and duplicate codes it is possible to find the transit time from satellite to ground station. In effect, tie measured delay between the two signals is the time o f travel from satelliie to receiver.

REAL-TIME GPS This term implies that the GPS data is being processed at the same time as it is being collected, albeit a fraction of time after the event. For this to take place, some reliable method o f communication is required for transmitting and receiving the data between the control (reference) GPS receiver and the remote (roving) unit. Initially, real time was used for lower-accuracy applica tions o f GPS. The technology has advarced so that it is now used for applica tions requiring accuracies o f a centimeter or better, particularly some civil engineering setting-out applications. Application o f the velocity o f light will give an approximation to the range. It is approximate because at that stage it contains the effects o f several errors and so is called the pseudo-range. N ot until the errors have been eliminated can the ranges be considered true and used for coordinate determination. For relative positioning there is a requirement for one station to be known (the master) so that the point in question (rem ote) can be coordinated in relation to it. Relative positioning also requires simultaneous observations at each point from the same satellites. This allows several error sources to be reduced or to cancel out. The same differential measurement technique has been used to improve the results from EDM and allow millimeter accuracy where normally only centimeters are possible. The fractional part o f a wavelength or phase shift can be determined by ;i phase meter, but the integral number o f wavelengths, called the 2 7 7

Four sate llite s = allows three-dim ensional positioning 6 9. Position circles

ambiguity, can only be resolved by comparisons o f signals received at each ground point. The phase measurement technique is again comparable to that in EDM since the distance is a multiple o f the wavelength plus a fraction o f a wave length. EDM, however, records the double distance to and from the target which does not apply to GPS. As in so many other aspects o f surveying, extra observations, if in the right configuration, can improve the results. So with the differential technique, increasing the number o f pairs o f ranges from the same two ground stations can make noticeable improvements. Over short baselines those o f a few kilometers errors due to the tropo sphere and ionosphere can be assumed to be similar at both terminals and will cancel out with appropriate station configuration. For longer separations such errors cannot be considered equal and must be measured and allowed for in some way. With pairs o f ranges to two satellite positions, additional error sources cancel out (see figure 64). Such a quadrilateral configuration could be repeated several times during the pass o f a single satellite or repeated with other satellites, to strengthen the calculated relative position. Accuracies in the 10-mm region are feasible with this technique. Data from the satellites gives geographic positions in terms o f a particular figure o f the earth (WGS 72) or rectangular coordinates on the GPS datum. Hence there may be a need to transform the results into a local system, and for this some knowledge o f that system is required. Th e heights have no direct relationship to mean sea level but can be transformed to ellipsoidal heights and then to orthometric values.

SOURCES OF ERROR As in all aspects o f surveying, GPS is subject to a host o f possible errors. While this text is not the place to go into them in depth, some at least require mention. T o begin with there are five elements to the whole system, each o f which presents its own problems. These are the satellite, the signal and atmosphere, the receiver, selective availability, and the datum.

S a t e llit e
The orbit in which the satellite moves is subject to a range o f influences, the predicted effects o f which are broadcast by the monitoring stations. In total, their effect is usually less than 100 m or 5 ppm on a baseline vector. The latest satellites are better known than the earlier ones, so this figure should show a decreasing tendency. Am ong the factors are the variability o f the earths gravitational attraction, and the effects o f both the ocean tides and earth tides. These last two are predicted from global models that are still sketchy in parts.

169

S ig n a l a n d A t m o s p h e re Atmospheric effects are well known in various aspects o f survey operation. They are similarly an incompletely understood problem with regard to GPS. While the transmission errors are modeled in the software, there are delays that could amount to many meters, due to both the ionosphere and the troposphere. This applies not only to the horizontal component but also to the vertical component o f the results. Even with modeling, the height can be in error by several centimeters due to this cause.

R e ce ive r
The receiver is subject to both clock errors and atmospheric effects. Again, m odeling o f the local situation happens within the software, but codeless-type receivers present more difficulty than coded versions.

S e le c tiv e A va ila b ility

Selective availability (SA) refers to the ability o f the U.S. Department o f Defense to deliberately degrade the signal for security purposes. The effect can be up to BOO m. It is achieved by manipulating either the navigational message or the satellite clock frequency. Its name arises because authorized (or selected) operators can get access to the unadulterated signals and hence obtain higher accuracies. For those operating in the differential mode, SA does not affect the result. Without SA a position could be determined to within 15 m after just a few minutes o f C/A code observations. When SA is turned on, the accuracy drops to the order o f 150 m. T o achieve the higher figure requires several hours o f observation.

D a tum
Particularly in relation to height values, there is the uncertainty o f the relation ship at any point between the geoid and the ellipsoid. Th e separation o f these two needs to be well known if a reliable result is to be obtained. Unfortunately, as with models o f other factors, the world is still insufficiently covered with the necessary gravity data to always produce good results for the separation o f the surfaces.

THE FUTURE As the applications for GPS increase, so the inconveniences o f selective avail ability multiply. It seems highly likely that if some o f the major operators, particularly in the area o f navigation, are not allowed full, unhindered access, then other independent systems will be developed.

CHAPTER

Gravity

T h e effect o f the intensity o f gravity impinges on many aspects o f geodesy. Its most obvious effect is to make a plumb bob hang in a particular manner often loosely referred to as vertical, although this needs qualification by saying to what it is perpendicular. As has been seen earlier (figure 46), the plumb line can be deflected from the vertical by large mountain masses, and this effect has to be accounted for. Measurement o f the intensity o f gravity has been possible for over 300 years and can be either relative or absolute. As with some other aspects o f surveying, the relative approach can produce acceptable results far more quickly than absolute measurements. It is because the earth is both nonspherical and nonhomogeneous that the acceleration due to gravity varies from point to point. Thus it varies with latitude, elevation, the distribution o f visible masses and mass deficiencies (such as ocean areas), the distribution o f invisible masses, and density varia tions through the earth. The standard unit o f measure is the gal named after Galileo Galilei which is equivalent to 10 mm/sec-2, and the magnitude at the earths surface is in the vicinity o f 980 gal. In terms o f the usual unit o f measure, the milligal (mgal or 0.01 mm/sec"2) , variations range over some 5000 mgal due to latitude and 30 mgal per 100 m (9.4 mgal per 100 ft) for changes in height.

ABSOLUTE MEASUREMENTS The absolute value is normally determined by timing a freely falling body, but high degrees o f accuracy were not possible until the advent o f electronic timing devices. Prior to this, the tedious but straightforward approach was to time oscillations o f a pendulum. The duration o f an oscillation directly relates to the length o f the pendulum and the acceleration due to gravity, g. A length

---

---------------------------------------------------------------------

o f 25 cm (10 inches) gives a time o f 1 second, and 1 m (40 inches) a time o f 2 seconds. The latter relation is normally referred to as a seconds pendulum, and the form er as a half-second pendulum since the time in the basic pendulum formula refers to a swing from one limit to the other and back again. It was Jean Richers observation o f a seconds pendulum in Cayenne in 1672 that first indicated that the earth might not be spherical. H e found that the length required to beat single seconds was 1V4 lignes (2.5 mm) shorter in Cayenne than in Paris suggesting an oblate earth. T o determine the length accurately, oscillations were observed over many hours so that the deduced time for a single swing was o f the order o f 1 ppm. This method persisted until this century with refinements in the structure o f the pendulum; an example is Henry Kater (1777-1835), who in 1818 developed a reversible pendulum with two knife edges. Absolute values were increasing in importance and various methods were devised to measure falling bodies in such a way as to obtain accurate results. One such invention involved photographing the lines on a free-falling scale. Other systems used lenses and mirrors, a falling glass sphere, falling corner cube reflectors, and symmetrical free motion o f an object as it was projected up and allowed to fall again.

RELATIVE MEASUREMENTS The procedure for relative measurements o f gravity is again rather like the single-base barometric heighting method. When pendulums were used, mea surements would commence at a point o f known absolute gravity value. Then the pendulum would be taken to each required position in turn and the period o f swing compared with that at the known station to obtain the gravity value at the new point. Today the gravimeter has replaced the pendulum. A gravimeter (or gravity meter) is any instrument in which the acceleration due to gravity is determined by measuring the force necessary to support a mass in the earths gravity field. The principle o f a gravity meter is based on the fact that when a weight hangs from a spring, the spring changes length with changes in gravity. For example, a nearly horizontal beam can be maintained in position by a spring under tension. As gravity changes, the beam tends to move, so the restraint exerted by the spring can be recorded. In operation a dial reading representing the force necessary to support the test mass is made first at a base station and again at the desired location. The difference in dial readings between the two stations represents the change in gravity. Changes as small as 0.01 mgal can be readily detected, and even 0.001 mgal is possible. Gravity meters can be used at sea or in aircraft as well as on land. Using them in vehicles, however, creates added problems. It is necessary to eliminate the effect o f the moving vehicle since the value required is a stationary one for a particular coordinate position. The effect o f motion is called the Eotvos effect after the Hungarian scientist Baron von Eotvos (1848-1919), who from about 1886 wrote extensively on gravity. The speed o f the vehicle, its direction o f travel, and the latitude are all needed to evaluate the correction.

I
MEASUREMENT USE IN GEODESY, GEOLOGY, AND GEOPHYSICS Since observed variations in the gravity field are caused by variations in the density o f the underlying rock, geologists and geophysicists frequently use gravity information in the analysis o f crustal structure. Just the opposite process is used by the geodesist, who uses the crustal structure to indicate the gravity for the area. For example, correlations between known structural and gravity variations can be used to estimate gravity anomalies for areas where no ob served values o f gravity are available but the crustal conditions are known to be similar. Correlations can also be used to evaluate the reliability o f observed gravity anomalies. The use o f geological information to estimate anomalous gravity is probably just a prelude to a much wider use o f both geological and geophysical data for geodetic purposes. Discoveries in the last 30 years indicate possible modifi cations in classical geodetic concepts regarding the relationships between topography and crustal structure. The use o f such data may also result in more precise methods o f reducing observed gravity to sea level.

CH A P T E R

12
The Gyroscope

f \ s a survey tool the gyroscope is probably best known as a unit that can be attached to a theodolite in such a way as to allow determination o f the meridian direction. In this form it is, for example, ideal for carrying orientation through a tunnel, although it can be just as usefully applied to orientation problems on surface surveys. For some years the same type o f unit has also played a prominent role in, and developed from, navigation for both missiles and aircraft, so brief details o f its operation will be relevant prior to describing its use in inertial surveying. The essential parts o f a gyroscope are as follows: 1. A heavy balance wheel in which most o f the mass is concentrated toward the rim. It is made to spin at very high speeds. 2. An axle to support the balance wheel, where the axles bearings are as nearly frictionless as possible. 3. A gimbal system in which the axle is mounted so as to allow the wheel to rotate freely and take up the appropriate orientation. As the wheel spins, it is constrained so that its spin axis remains horizontal and rotation occurs only around the vertical axis. The torque exerted upon it makes the spin axis seek true north, and it will oscillate about this direction in a simple harmonic mode. When a gyroscope is attached to a theodolite (in which form it is known as a gyrotheodolite) , it is possible to record the angular position o f the system in relation to a reference object some distance away. That is, the direction from the instrument to the reference object can be related to true north. Although the axes in a theodolite system are restrained, even when the gyroscope is free, it will maintain its plane o f rotation irrespective o f how the

174

supports move. Willi good conditiors and sufficient time it is possible to record with accuracies o f a few seconds o f arc. While the use o f a gyrotheodolite is somewhat limited it is the application o f a gyroscope in an inertial surveying system (Chapter 13) that is o f grow ing importance.

CH A P T E R

13
Inertial Surveying

I he use o f the gyroscope as an aid to navigation can be traced to the early years o f the century, but the most notable military application was for missile guidance in the V2 rockets o f W orld War II. As a tool for the surveyor it is o f much more recent date. Two firms were given U.S. government contracts in the early 1960s for military positioning systems, and by 1967 one o f these firms was asked to develop the equipment for survey purposes. N ot until the mid-1970s was a high degree o f accuracy achievable, as a development from guidance equipment needed for the space race. Inertial surveying is a completely new method o f positioning for the surveyor. In essence it consists o f a unit (albeit a rather bulky one at present) that is transported by truck or helicopter between a series o f points whose positions are required, rather along the lines o f a conventional theodolite traverse, although successive points do not need to be intervisible. Thus, by starting at a known position and closing at the same, or other known position, the intermediate stopping points can be coordinated. Repetitions o f a circuit can improve the results appreciably. The coordination is found in terms o f the amount by which the unit has moved from point to point in three dimensions. T o achieve this requires three accelerometers (see below) and three gyro scopes mounted in gimbals on a stable base. They must be rigid in space that is, maintain their orientations against attempts to move the spin axes from the original positions.

ACCELEROMETERS An accelerometer is basically a sprung mass arranged in such a way that it moves as the vehicle moves and has to be counteracted by a restoring force to return it to its original position. The amount o f restoring force is a measure o f the acceleration, which can be converted into the equivalent distance traveled.

17fi

An acceleroinetei is capable o f recording both acceleration and deceleration against time. A single unit is capable o f detecting angular movement in relation to true north. In addition, the equipment can give information on the deviation o f the vertical and acceleration due to graviy. In essence one can use the basic formulae for motion in a straight line to illustrate the operation o f inertial equipment. A single accelerometer can detect the acceleration imposed upon it at discrete intervals o f time (fractions o f a second). Starting from a stationary oosition, the velocity at the end o f an interval can be calculated as the second integral o f the acceleration, and from this the distance traveled during the interval is determined. In its turn this positions the instrument with relation to its starting point. I f instead o f one accelerometer there are three mounted orthogonally, then the three-dimensional (rather than simply the one-dimensional straight line) movement can be determined. This description is somewhat o f an over simplification since, as with most survey operations, there are side effects to consider. Examples o f these are the Coriolis effect, variations in the gravita tional field, and deviations o f the vertical. Several methods are available to keep the axes o f the accelerometers main tained in a coordinate framework as defined in terms o f a geodetic refer ence system. In addition to a vehicle moving with respect to the reference system, the earth itself is also in motion. This effect is overcome by the continuous applica tion o f torque on the gyroscopes. When an outside force tries to move the axis from its alignment, the effect will be to make the wheel precess in a direction at 90 to that o f the applied force. Since it maintains its orientation, it is said to be rigid in space. This retention o f orientation in space continues as the earth rotates, such that the axis o f the gyroscope tilts by measurable amounts through a full circle in 24 hours. I f this idea is extended to three gyroscopes mounted at right angles to one another, then it follows that move ments can be detected in all three direction. The combination o f three gyroscopes and three accelerometers allows measurement o f the directions and magnitudes o f movement as well as stabili zation o f the whole system.

ZERO VELOCITY UPDATE One particular problem is that an accelerometer cannot separate accelerations due to movement from the acceleration due to the earths gravitational field. This can only be overcome by periodic stops (every 3 to 5 minutes) for about 20 seconds to achieve zero velocity update (Z U P T ). This eliminates accumulated errors and allows the velocities o f each accelerometer to be recorded. The Z accelerometer value will then be almost all due to gravitational effects. With a land-borne system, regular ZU PT stops should present no problem, but with a helicopter system it may be necessary to use a hover period if landing is not feasible.

I 7 - 7

Prior to being moved from the starting station ihe equipment is set running for up to an hour. This allows time to calibrate the system; it can sense the vertical direction at that station, orient to astronomical north, and perform other calibrating exercises. Because the instrument is transported, there is an update o f acceleration data every 0.016 sec that, when doubly integrated into E, N, and H changes, enable the axes to be reoriented. When movement ceases, to allow for ZU PT the vertical direction at that point can be located as before, and from this can be found a measure o f the deflection o f the vertical. At the same time various calibration parameters can be updated. The data acquired and computations carried out during a traverse are recorded on tape to allow subsequent additional adjustments and conversion to any required coordinate system. The precision o f a gyroscopic system is, unlike that in a traditional traverse, not so much a function o f the distance covered but more an inverse function o f the time taken. Errors can arise from th,e mechanics o f the system, external effects such as gravity, and from the operator.

CHA P T E R

14
Velocity of Light, EDM, and Laser Ranging

VELOCITY OF LIGHT The basis o f much o f the application o f electronics to surveying is a good knowledge o f the velocity o f light or radio waves. When taken in a vacuum the two are the same, but when used in the free atmosphere, they are affected slightly differently. In particular, humidity has a greater affect on radio waves than on light waves.

B a s ic R e la tio n s h ip s
I f such waves are used to measure distances on the earth or ranges o f such bodies as artificial satellites, the basic relationship is that distance = time X velocity Before such an approach could be used at levels o f accuracy required for surveying, the velocity had to be known to better than 1 km/sec. Until the 20th century such accuracies were impossible to achieve, so before any serious attempt could be made to use the relation as it stands, it was first investigated the other way around, as velocity = distance/time and extensive experiments were carried out to obtain a reliable value for the velocity o f light waves in a vacuum.
170

E a rly H is to ry The fact that light waves travel at a finite velocity was appreciated some three centuries ago. During 1676 O laf Roemer, the Danish Astronomer Royal, was observing the eclipses o f Jupiters satellites and determined that the time taken for light to travel a distance equal to the diameter o f the earths orbit equated to a velocity o f light equivalent to 214 000 km/sec. Fifty years later James Bradley, Third Astronomer Royal at Greenwich, deduced a value equiva lent to 301 000 km/sec. The 19th century saw determined experiments to improve the value by optical-mechanical methods. D. F. Arago, director o f the Paris Observatory, experimented in 1820 with a rotating mirror. This technique was modified in 1849 by H. L. Fizeau. He transmitted a pulse o f light to a distant mirror, and on its return it was interrupted by a rotating cog wheel. At a particular velocity o f the wheel the returning ray would be intercepted by the cogs and not be visible to an observer near the source. Using a cog wheel o f 720 teeth the first eclipse was found at an angular velocity o f 12.6 rps (revolutions per second) equivalent to 313 300 km/sec for the velocity o f light. In 1862 Jean-Bernard Foucault used a mirror rotating at 500 rps on a 20-m baseline to obtain a velocity o f 298 000 km/sec. He was followed by Albert Michelson, who made many experiments over 40 years from 1879, when he obtained the value 299 910 50 km/sec; to 1926, when he experi mented over a 35-km baseline in the United States to get a value o f 299 798 4 km/sec. Electrooptical methods first developed around 1925 when the cog wheel was replaced by a Kerr cell between two Nicol prisms. The result was a value for the speed o f light o f 299 778 20 km/sec. The year 1941 was a turning point when Erik Bergstrand o f the Geographi cal Survey o f Sweden conceived a blinking light system. Here the cog wheel was replaced by light pulses o f known frequency and variable intensity pro jected over the line and returned to a receiver near the transmitter. In 1947 his tests over 7734 m gave a velocity o f 299 793.9 2.7 km/sec. At about the same time L. Essen and C. Aslakson were achieving comparable results, but the firm o f A G A was so interested in Bergstrand that they aided his research. In 1948 the first tests, made with a prototype Geodimeter, derived a velocity value o f 299 793.1 0.26 km/sec. Th e results were now so encourag ing that the whole concept could be turned around: the Geodimeter was then used to measure distance with the known velocity o f light. Subsequent years have seen various refinements, all o f which give results in the region o f 299 792.458 0.001 km/sec. Th e surveyor was now confidendy able to use velocity and time to determine distance. (The accuracy with which time could be recorded had long been in advance o f requirements for routine survey operations).

ELECTRO M AGNETIC DISTANCE M E A S U R E M E N T (EDM)

From the late 1940s it has been possible to measure distances from a few tens o f meters to over 100 km, quickly and accurately at the press o f a few buttons and twiddles o f a few knobs on a black box. The principle o f EDM, as stated above, is that distance = velocity X time. Thus, from the known velocity o f electromagnetic waves in a vacuum (about 300 000 km/sec), suitable corrections for the fact that it is not actually in a vacuum, and a measured value for the time to send a signal from one end o f a line to the other and back again, it is possible to determine the distance. In other words, the measurement is actually one o f time rather than distance. Almost all EDM instruments operate on phase difference techniques. Any signal radiated to a far point and back again will exhibit a difference in phase angle on its return. This difference can be measured, but to go with that is a need for knowledge o f the integral number o f wavelengths traveled. For microwave systems the number o f wavelengths is usually found by increasing the measuring wavelengths by factors o f 10. An alternative is to use several specific wavelengths. For light-wave instruments two frequencies suffice because o f the shorter maximum range to be resolved.

E a rly H is to ry The fact that light waves travel at a finite velocity was appreciated some three centuries ago. During 1676 O la f Roemer, the Danish Astronomer Royal, was observing the eclipses o f Jupiter s satellites and determined that the time taken for light to travel a distance equal to the diameter o f the earths orbit equated to a velocity o f light equivalent to 214 000 km/sec. Fifty years later (ames Bradley, Third Astronomer Royal at Greenwich, deduced a value equiva lent to 301 000 km/sec. Th e 19th century saw determined experiments to improve the value by optical-mechanical methods. D. F. Arago, director o f the Paris Observatory, experimented in 1820 with a rotating mirror. This technique was modified in 1849 by H. L. Fizeau. H e transmitted a pulse o f light to a distant mirror, and on its return it was interrupted by a rotating cog wheel. At a particular velocity o f the wheel the returning ray would be intercepted by the cogs and not be visible to an observer near the source. Using a cog wheel o f 720 teeth the first eclipse was found at an angular velocity o f 12.6 rps (revolutions per second) equivalent to 313 300 km/sec for the velocity o f light. In 1862 Jean-Bernard Foucault used a mirror rotating at 500 rps on a 20-m baseline to obtain a velocity o f 298 000 km/sec. H e was followed by Albert Michelson, who made many experiments over 40 years from 1879, when he obtained the value 299 910 50 km/sec; to 1926, when he experi mented over a 35-km baseline in the United States to get a value o f 299 798 4 km/sec. Electrooptical methods first developed around 1925 when the cog wheel was replaced by a Kerr cell between two Nicol prisms. The result was a value for the speed o f light o f 299 778 20 km/sec. The year 1941 was a turning point when Erik Bergstrand o f the Geographi cal Survey o f Sweden conceived a blinking light system. Here the cog wheel was replaced by light pulses o f known frequency and variable intensity pro jected over the line and returned to a receiver near the transmitter. In 1947 his tests over 7734 m gave a velocity o f 299 793.9 2.7 km/sec. At about the same time L. Essen and C. Aslakson were achieving comparable results, but the firm o f A G A was so interested in Bergstrand that they aided his research. In 1948 the first tests, made with a prototype Geodimeter, derived a velocity value o f 299 793.1 0.26 km/sec. The results were now so encourag ing that the whole concept could be turned around: the Geodimeter was then used to measure distance with the known velocity o f light. Subsequent years have seen various refinements, all o f which give results in the region o f 299 792.458 0.001 km/sec. The surveyor was now confi dently able to use velocity and time to determine distance. (The accuracy with which time could be recorded had long been in advance o f requirements for routine survey operations).

ELECTR O M AG NETIC DISTANCE M E A S U R E M E N T (EDM)

From the late 1940s it has been possible to measure distances from a few tens o f meters to over 100 km, quickly and accurately at the press o f a few buttons and twiddles o f a few knobs on a black box. The principle o f EDM, as stated above, is that distance = velocity X time. Thus, from the known velocity o f electromagnetic waves in a vacuum (about 300 000 km/sec), suitable corrections for the fact that it is not actually in a vacuum, and a measured value for the time to send a signal from one end o f a line to the other and back again, it is possible to determine the distance. In other words, the measurement is actually one o f time rather than distance. Almost all EDM instruments operate on phase difference techniques. Any signal radiated to a far point and back again will exhibit a difference in phase angle on its return. This difference can be measured, but to go with that is a need for knowledge o f the integral number o f wavelengths traveled. For microwave systems the number o f wavelengths is usually found by increasing the measuring wavelengths by factors o f 10. An alternative is to use several specific wavelengths. For light-wave instruments two frequencies suffice because o f the shorter maximum range to be resolved.

D' Ds D s' D

= = = =

recorded value corrected slope value reduced to sea level chord ground equivalent at sea level

70. Reduction of EDM Measurements

V t 7iw/U<liy V _/l l _ i y I l l ,

l_ l_/ IV I, U ' IV-* I-U W W I

IVM i y n i y

All measured distances are inclined lines between elevated terminals and require correction for this as well as for their height above sea level or equiva lent reference surface (figure 70). In modern instruments the inclination can now be allowed for automatically. Instruments that use visible light operate on wavelengths as short as 0.56 X 10 6 m, whereas those that use microwaves are on wavelengths ranging from a few millimeters to 100 or 200 mm. For light-wave instruments the light source can vary from an ordinary 6volt bulb to a laser; as a result, the range o f operation can vary from a few kilometers to some 50 km. For microwave instruments the signal is generated by an oscillator which can be much more directional than a light source and has the added advantage o f operation in poor weather conditions.

w . u ..y i il, U U IV I, Q IIU

I I l o t >1 M I 1M I|IIH |

Wavelength

h------A -------H

A m plitude

(a) Basic Wave

(b) Frequency Modulated

71. EDM Waves

Microwave systems use a frequency modulation (figure 71b), particularly as an aid to resolving ambiguities in th; recorded distance. On the other hand, the light-wave systems normally use amplitude modulation (figure 71c). Both forms o f instrument are particuhrly affected by changes in pressure, temperature, and humidity, but corrections are possible as long as m eteorolog ical observations are recorded at the sane time as the distance. Accuracies o f 1 in 100 000 and better are possible cepending on conditions. Instruments o f much longer wavelength are used for marine navigation. They are o f lower positional accuracy but sufficient for the purposes they are put to. I f better accuracy is required offshore, nowadays a satellite fix would be used instead. The arrangement for long-wavelength positioning is akin to the resection problem. Three fixed land stations are used in adjacent pairs to generate two families o f hyperbolae. Th e signals received indicate differences in distance from the master station and each slave station. Then from knowledge o f the number and separation o f successive hyperbolae it is possible to determine a position for the receiver.

LASER RANGING As long ago as the 1950s it was thought feasible to use lasers to determine the distance from the earth to the moon. T o put this idea into practice required a reflector system on the moon and this was achieved initially during the Apollo program. Ranges to the moon were then taken from various purpose-built stations around the globe. The separation o f these had little effect on the accuracy o f the range measurements which were considered to be in the centimeter region. However, a reflector at such a range is an extremely small target and the procedure was difficult. Modifications o f the same idea are used now on the orbiting satellites. Several passive satellites such as LAGEOS (launched into a 6000-km-high orbit in 1976) and STARLETTE were specifically equipped with reflectors to act as laser targets. They make an easier target than a moon reflector. Even a single-range measure can be in the order o f a few centimeters for accuracy, although generally it still requires several weeks o f observations to get to this figure. When a pulse o f intense light is aimed at the satellite, some o f its photons return. As with EDM, use o f the velocity o f light together with the transit time gives the distance. I f several ground stations are used, it is possible to determine both the orbital parameters o f the satellite and the separation o f the stations. Such systems are proving to be of particular use to the geologist and geophysicist for monitoring crustal movements, continental drift, and changes in the earths parameters. As with all other systems, a relative mode can be more accurate than an absolute mode, and this is o f particular significance in geotectonics. Accuracies in the order o f 1 X 10 7 o f the separation are possible.

r iu jo i- . u v . ji ia

Plane

Cone (a) Projection Surfaces

Cylinder

(b) Conic with Two Lines of contact 72. Map Projections

C H A P ' ER

15
Projections

O n c e points have been coordinated in an ellipsoid o f known size and in relation to an acceptable datum, therf remains the problem o f depicting those positions in a readily interpreted graphical form, which can be much more meaningful than sheets o f numbers. All through this text, reference has bien made to a nearly spherical surface on which all points are positioned. The problem then is to graphically repre sent a spherical surface on a flat sheet o f paper an exercise referred to as map projections. Despite centuries o f effort (Ptolemy, a .d . 100-178, produced one o f the earliest) no method has yet t>een found to create map projections without some forms o f distortion inherent in the results. These will vary in form and magnitude according to the size o f the area involved, the scale at which it is to be represented, and the projection method adopted. It is a vast subject that can only be touched upor here. Although there are only three basic projection surfaces, by mathematical manipulations it is possible to obtain some 200 variations. Consider an orange where the peel represents the topography that is to become the map. I f a small piece o f peel is taken off, say a cm square, and laid on the table, it can be flattened out with very little (distortion. If, however, a piece several centimeters across is taken off, when pressed flat it will split as well as not lay properly flat. This is an example o f the distortion problems found in map projections. The distortions are o f three basic forms in area, shape, and scale (or distance) and the respective projection groups are equal area {equivalent) ^onform al (orthomorphic), and equidistant. The group most useful for the surveyor is the conformal, in which the scale o f small areas is sensibly the same in all directions. The three basic projection surfaces re the plane, the cone, and the cylinder. The latter two are said to be developable since they can be cut and unrolled lo lay flat. Figure 72a shows each in contact with a globe. Consider the simplest *

__________________________________

7 3 . The E ffect of Using D ifferent Spheroids

(a) Single point of contact

7 4. Transverse M ercator Projection

Projections

80N

75. UTM Zone 36

case o f the plane. Since it touches the globe at a single point, everything on the surface around that position could be transferred to the plane with negligible distortion. The farther one progresses from the point, the greater the distor tion. Th e use o f different spheroids can also be a problem (figure 73). Imagine that the globe is transparent with a point source o f light at the center (figure 74). Then points such as A, B, C, D, and E will be projected by the light source onto the paper plane at A', B ', C , D ', and E !. This is the essence o f all map projections. T o get different projections, the position o f the supposed light source can be varied; the position where the plane, cone, or cylinder touches the globe can be varied; and these surfaces can be made to cut, rather than touch the globe. Add in projections that are mathematical contortions and entail choices as to which distortions to limit and which to accept and one has an almost infinite selection o f projections.

W ORLD

P R O JE C T IO N

W e have only to look at an atlas to see that there are many ways o f representing the whole world on one projection. However, these mostly contain quite large distortions o f one form or another. As the need for a world projection for geodetic work became significant during the 1940s, a specification was drawn up by the U.S. Army to cater to military global requirements. The need was not for an atlas map, which would be o f little advantage to soldiers, but for a system on which the whole world could be represented on sheets at a large scale. There was a need to minimize erros in azimuth, to keep scale errors within prescribed limits, to use as few slices or zones as feasible, and with straightfor ward conversions o f coordinate values from one part o f an overlap to the other. From such requirements grew the Universal Transverse Mercator (U T M ) projec tion (sometimes called the Gauss-Kriiger projection), which, as the name implies, is a modification o f the Transverse Mercator. It was found that the optimal width o f a zone was 6 o f longitude we can liken this situation to an orange with 60 segments. The U TM is a cylindrical projection in which the cylinder can be imagined to have a radius less than that o f the globe. Hence there are two lines around the globe that coincide with the cylinder. This particularly keeps the scale error across a zone within the acceptable limits o f 1/2500. Thus sections AB and DE (figure 74b) have to be stretched by up to +1/2500 to M B ' and D 'E on the plane, while BCD has to be squashed by up to 1/2500 to B C D ' . Every zone (figure 75) is treated identically so that there is a consistent referencing system, and metric units are throughout. In addition, each zone has the same coordinate values on the central meridian an Easting value o f 500 000 m to ensure that all coordinates remained positive, a Northing value o f 0 m for working north o f the equator, and 10 000 000 m for zones south o f the equator.

Th e zones extend north-south from HO" N to HO" S. /one 1 is centered at 177 W, so it extends from 180 W to 174 W, and the zone numbers increase to the east. Various ellipsoids were recommended for different parts o f the world: Clarke 1880 for Africa; Clarke 1886 for North America; Bessel for the USSR, Japan, and parts o f Southeast Asia; Everest for India and neighboring parts o f Southeast Asia; and the International 1924 for the remainder. These will be seen to agree closely with the preferred datums (page 95).

M ER CATO R

P R O JE C T IO N

The Mercator projection is based on a cylinder with its axis parallel to that o f the earths rotation. Variations on this can be obtained by having the two axes either oblique to one another or at right angles to one another (see Transverse Mercator Projection below). I f they are oblique to one another, the line o f contact is neither a parallel nor a meridian. The scale varies with distance from the parallel o f contact between the cylinder and the earth sphere which is the equator. Hence it is best for areas near the equator since distortions increase rapidly the farther one gets north or south. N o other line o f contact can be made between a sphere and cylinder with parallel axes. Where a country such as Malaya is greater in extent in the north-south direction than east-west but its orientation departs noticeably from the merid ian, then an oblique (or rotated) Transverse Mercator is possible.

TR AN SVER SE M ER C A TO R

P R O JE C T IO N

In the Transverse Mercator projection the scale varies with distance from the central meridian; hence it is suitable for areas o f limited east-west extent or for areas that can be readily subdivided into separate systems (or zones) in that direction. This is based on the cylinder with its axis at right angles to the earths axis o f rotation. However, the cylinder can be rotated such that the two axes are still retained at right angles to one another but the line o f contact is a different meridian line. With just one line o f contact, the scale error will often increase too rapidly for the required area to be covered satisfactorily. By modifying the system to have two lines o f contact achieved by imagining the enclosing cylinder to have a smaller radius than the enclosed sphere the increase in scale error can be kept within acceptable limits over a far wider extent (see figure 74). This is the system used in Great Britain. The scale error varies between +1 in 2500 and 1 in 2500, with two meridians where the scale is exact. In Georgia (U .S.A.), where there is a similar system, the range o f errors is between 4-1 in 10 000 and - 1 in 10 000.

___________Scale too large by m axim um am ount

Scale exact

Scale t oo sm all by m axim um am ount

___ _______ Scale too large by m axim um am ount 76. Lam bert G rid -T y p ic a l State or Country Greater East-West than North-South

LA M B E R T C O N IC A L O R TH O M O R P H IC

P R O JE C TIO N

In this projection the scale varies with distance from the central parallel. Hence it is suitable for areas o f limited north-south extent or for areas that can be readily subdivided into separate systems in that direction. As the name implies, the Lambert Conical Orthomorphic projection is based on a cone fitting over the sphere or cutting into it (figure 72b). In the form er there is contact along one parallel, in the latter along two. Along these standard parallels the scale is correct, and as one moves north or south from these parallels, the error varies. In the case o f two standard parallels, the scale is too small between the parallels, and too large outside o f them (figure 76).

77. Transverse Mercator Grid Typical State or Country Greater N orth-South than East-West

Scale too large b y maximum

amount

For example, in Connecticut the scale varies from being too small by 1 in 59 000 to being too large by 1 in 34 000, whereas in North Carolina the range is from 1 in 7900 too small to 1 in 5950 too large (see figure 77).

STA TE

P LAN E C O O R D IN A TE S

In a country as vast in extent as the United States, each state has its own system. Because o f the variety o f shapes o f the states, some 30 have adopted the Lambert system and 19 the Transverse Mercator. It is fortuitous that virtually all the states are oriented more or less either east-west or north-south, and as a result only two different projections are required. By using a conformal projection as the basis and limiting one direction o f a single zone, these maps are able to preserve correctly angles defined by lines o f less than about 10 miles (16 km). Moreover, the scale factor, or deviation o f the grid length (the distance on a plane) from the geodetic length (the surface distance on the spheroid), can be limited to within accept able bounds. One particular problem raised in addition to the scale factor is known as the convergence of the meridians: except at a few specific positions, there is a difference between grid azimuth o f a line and its geodetic value, and this difference increases the farther the line in question is from the central merid ian o f the system. The problem arises because, as is well known, the geodetic meridians converge toward the poles whereas the grid equivalents do not. In the 1930s an engineer submitted to the USCGS a proposal for an arrange ment whereby geodetic data could be used over an entire state by applying only the formulae for plane surveying. This resulted in 1933, in the establishment o f a coordinate system in North Carolina in which latitudes and longitudes o f points throughout the state could be transformed into plane rectangular coordinates on a single grid. Then any surveys within the states could be defined in terms o f their coordinates on a common origin.

CHAPTER

16
Examples of Modem Projects

S U P E R C O N D U C TIN G

S U P ER C O LLID E R

(U S A )

A recent example o f the requirement for high-accuracy geodetic control was the Superconducting Super Collider (SSC) in Texas. Although the project terminated because o f financial constraints, enough o f the work was completed to illustrate the care and accuracy required for such a structure. The main ring o f the structure was to be a tunnel 4 m diameter and 87 km long. This tunnel would contain 12 000 magnets, each o f which had to be aligned to better than 1 ppm in a perfect geometric plane, not a level surface. The survey tolerance for the excavation, both horizontal and vertical, was 76 mm. The added stricture here was that excavation had to meet the specification precisely at all times and with all parts; no realignment for corrections would be possible. Th e design coordinates o f the magnet positions were given in an arbitrary site reference system as cartesian coordinates. T o transform these into geodetic coordinates required an accurate model o f the relative geoidal undulations in the area. Transformed ellipsoidal heights had to be corrected by the geoidal separation to obtain the orthometric heights. It was expected that the relative geoidal heights could be determined with an average standard deviation o f 5 mm and the components o f the relative deflections o f the vertical to 0.1" o f arc. In addition, a special conformal map projection was defined for the area o f the project. Surface control was designed for high-precision GPS aimed at achieving 3 mm + 1 ppm for the baseline components. The relative height difference across the ring was to be 10 mm. T o achieve this, scientists determined the effect o f tidal accelerations on the earths equipotential surfac e, b in found it

II
to be less than 1 mm. Additionally, account was required o f the nonparallelism o f orthometric surfaces to the reference surface, and this was modeled. The possible effect o f lateral refraction in the tunneling survey was of concern since even a small temperature gradient o f 0.2 C/m over 500 m would deviate the line by 23 mm. It was possible to reduce this effect by using gyrotheodolites instead o f normal traverse angle measure. The actual achieved accuracy o f the GPS control was 3m m 10 7S. The tunnel was to be connected to the surface by a number o f vertical shafts spaced about 4 km apart. A t the time the project closed, 24 km o f tunneling was completed. The breakthrough o f individual 4 km segments between the shafts was accurate to within a few millimeters. For further details on the SSC, see Chrzanowski et al. (1995).

G O T T H A R D T U N N E L (S W IS S A L P S )
Long, deep tunnels present many problems that relate to geodesy. One o f the latest such tunnels is the proposed Gotthard base tunnel in the Alps. At 56.9 km long, it will pass between Erstfeld and Biasca. Although a similar length to the Channel Tunnel between England and France, the problems in construction are quite different. While the Channel Tunnel is only 100 m below sea level, the Gotthard tunnel will have up to 2300 m o f overburden, which is expected to create tunnel temperatures o f 55C. From the geodetic point o f view, one particular problem is the geoidalellipsoidal separation. This will be critical to achieving an accurate break through. The whole proposed route o f the tunnel has been given a geodetic control network using both traditional (triangulation and traverse) and mod ern (GPS) methods. A good mathematical model based on the ellipsoid and geoid, with devia tions o f the vertical and geoidal undulations, was essential for transforming coordinates from the GPS geocentric version to the grid systems; and for correcting both the gyro-azimuths and the astronomical observations. While the geoidal model is well known on the ground surface, geodesists can only assume that the model for the depths under consideration takes a similar form an assumption that could introduce errors. In addition, the change in gravitational acceleration affects the height difference that is used for the depth between the surface and a shaft bottom some 800 m deep. During survey o f the tunnel operations, refraction will be a particular problem because o f the expected large temperature gradients across the tunnel. In the less rigorous conditions o f the Channel Tunnel this was found to have a noticeable affect, so it is likely to be a much greater distraction in the Gotthard tunnel. As far as angular measures are concerned, normal traverse-type angles relate to the geoid while angles from the gyrotheodolite refer to the ellipsoid.

The difference between them is the deviation o f the vertical; if this correction is neglected in this situation a large systematic error could result. Despite all the likely difficulties the expected breakthrough on the most difficult section is 10 cm. For further details on the Gotthard Tunnel, see Egger (1995).

M IN ER A L E X P L O R A TIO N

IN K A Z A K H S T A N

Th e mineral wealth o f Kazakhstan and neighboring areas o f the Caspian Sea is in the process o f development and expansion. So vast is the potential that one o f the first requirements is up-to-date mapping. The local geodetic system used a Pulkovo 1942 datum. One o f the prime mapping needs was to transform coordinates from the WGS 84, which was done using a network o f seven geodetic monuments around the north end o f the Caspian Sea. These were located in absolute position to better than 0.5 m. An intricate manipulation o f various datum shifts was then required to relate Pulkovo 1942 to WGS 84. It was then possible to calculate the geoidal undulations for the area. While these agreed well in one part o f the country, there was an unexplained degrada tion in other areas. Further gravitational data and a new geoidal model could well iron out the differences. High accuracy is required in this project because o f the requirements o f the seismic operations. For further details on mineral exploration in Kazakhstan, see Nash et al. (1995).

H EIG H TIN G

B Y G P S IN L E S O T H O

In traditional survey work, heighting over long distances has always been a problem because o f the refraction effect on the grazing rays as they pass through the lower layers o f the atmosphere. Even over the steepest o f sights by trigonometric heighting, the vertical angle is seldom more than a few degrees. GPS, on the other hand, concentrates on elevations as near to the vertical as possible and so greatly reduces the refraction problem. In fact, only data from satellites above a given angle o f elevation (such as 10) are accepted in the solution. For heighting only, it is feasible to concentrate on those satellites orbiting closest to the vertical. Since usually it is differences in height that are required, the geodesist can assume that nearly vertical observations through the atmosphere from points not too far apart contain comparable refractive errors that cancel out a situation that cannot be so readily accepted in traditional trigonometric heighting. A fine example o f the advantages o f GPS is instanced in a paper by K. N. Greggor relating to the Lesotho Highlands Water Scheme. It graphically illustrates a number o f the topics discussed elsewhere in this text and demon strates the need for their recognition and application.

In this scheme there are to be five dams, two power stations, and over 200 km o f tunnel. Th e required control points are located in deep valleys since the tunnels on the project are some 1000 m below the highest summits o f the Highlands. Using traditional methods to fix such points numerous mountaintops would be occupied for the control before coordinating the valley points a process that would have taken several months. By using GPS this has all changed and the actual valley points were occupied and the 70 km o f check survey completed in only two weeks. Height differences measured from satellites differ noticeably from those sensed by flowing water or spirit leveling. In fact, over the 70 km length o f the Lesotho section o f tunnel, the height differences measured by spirit level ing and those by GPS differed by as much as 2.5 m. The cause could be attributed to the variations in gravity resulting from the variations in internal density o f the earth. A further complication arose in that the national height network refers to the Clarke 1880 ellipsoid, fitted in such a position and orientation as to form the Cape datum. However, this does not coincide with the GPS ellipsoid o f WGS 84. Besides differing in dimensions they depart from one another by some 300 m. In addition, there appears to be a small tilt o f the order o f 1" in alignment between the Cape and WGS 84 datums and there may also be a very small scale difference. In traditional precise leveling it is neccessary to account for the corrections between orthometric and measured heights which are normally very small indeed. In extreme conditions, as in Lesotho, this can amount to several centimeters over 100 km, and with tight fluid-flow parameters this could be a critical amount. On the other hand, although the effect o f the geoidal undulation is reck oned in meters, it is usually adequate to assume that in a local region both the ellipsoid and the geoid are more or less equivalently ellipsoidal in shape although slightly tilted with respect to one another. This tilt is the deviation o f the vertical, and in extreme cases can amount to 30". Some might ask why we should not just use GPS heighting and ignore the geoid. Unfortunately, particularly in projects such as the Highlands Water Scheme in Lesotho, there are difficulties with this approach, not least o f which is that it is not possible to use satellite observations in tunnels. Thus the GPS height values can only be used on the surface to check the precise spirit leveling as it emerges from underground. A further problem is that fluid-flow calculations would be o ff by several seconds o f arc, even up to 5" in Lesotho. The slope o f the tunnel has to be kept within close limits such that the water neither flows too fast nor too slow, so a few seconds can be o f considerable importance. In a trial on the edge o f an anomalous gravitational section o f the scheme, GPS measures on a dam deformation scheme were compared with traditionally obtained results. Although the GPS values were accurate to better than half a centimeter, the tilt was found from the comparison to be 11".

201

. .r .w w <_ i v iv / n <i 11

rivjjouiB

It cannot be stressed too strongly that the casual user o f GPS software should never simply assume that the global geoidal data provided is o f adequate resolution for particular heighting uses. It must also be remembered that even in areas well covered with gravity measurements the latent unknown relationship between GPS and geoidal height differences is o f the order o f I in 100 000, whereas GPS heighting alone tends toward an order o f magnitude better than this. For further details, see Greggor (1993).

PLATE TE C TO N IC S
In several parts o f the world geodetic methods are being used to determine plate tectonic movements o f a few centimeters per year. Both satellite dopplei and GPS have been used in Papua New Guinea, Greece, and New Zealand. Geodetically Papua New Guinea is an interesting area because o f its large gravity and magnetic anomalies as well as the plate movements. McClusky (1994) quotes Bouguer anomalies ranging from 180 to 200 mgal and a minimum free-air anomaly o f 300 mgal. Obviously a research gold mine for the geophysicist and volcanologist such areas also provide the surveyor with opportunities to use the latest technology to quantify the rate and direction o f movement o f the earths crust. In order to do this sort o f analysis it is first necessary to provide over the area o f interest a network o f control points positioned with the highest possible accuracy. Ideally the whole network is reobserved on a regular basis, which may be annually or every 2 or 3 years depending on the availability o f staff, equipment, and funds. Further observation points can then be related to the basic control point. In the Papua New Guinea work comparisons are possible between 1981 data and various observations in the 1990s. Movements are in the range from + 1.51 to 1.93 m over lines o f several hundred kilometers. It is also possible to compare results by Doppler and by GPS. In general, the techniques employed have proved to be appropriate to the requirements o f the geologists and geophysicists and to reliably record changes o f a few centimeters per year. For further details see McClusky et al. (1994) and Stolz et al. (1982).

D E TE R M IN IN G

H E IG H T O F M O U N T E V E R E S T

A good example o f the complications o f the geoid is the determination o f the height o f Mount Everest. This mountain is some 400 miles (645 km) from the nearest sea, so where would sea level be if extended under the mountain? For a long while the height was quoted as 29 002 ft, and the question was raised as to whether or not the 2 at the end should be omitled. II left it

might imply an enhanced accuracy, but if it were dropped it would imply that tile value was only known to the nearest 1000 ft. Is 29 002 ft an exact height? In the 1920s De Graaff-Hunter used an interest ing analogy. Th e height o f the Eiffel Tower is stated to be 984 ft. But what about the legs o f the tower and foundations should not they be included? l hen beneath the tower is a complex o f radio installations. Th e point is that, not only the top, but the bottom also needs defining before quoting a height. In the case o f an ordinary tower it would be possible to decide what the bottom was and get access to it. With the Eiffel Tower, however, we would need to remember that it changes by some 3 inches (8 cm) seasonally with temperature. With mountains the base is certainly not visible, and because o f tectonic movements the height is also changing slightly. Some o f the possible causes o f errors in any height measure would likely be o f opposite sign and so cancel out; aside from these there was estimated to be an uncertainty o f about 27 ft (8.2 m ). In general, there are three main factors affecting the accuracy o f the height determination o f a mountain: geoidal separation atmospheric refraction deviation o f the vertical In the case o f Mount Everest, however, these are compounded by: vast lengths o f sights great differences in height from observing stations to peak the huge mass o f the mountain range the distance o f the peak from the sea the impossibility o f occupying both ends o f the line

S tations added

1880-1902

0 ----- 1

30 ---1

Miles (Approx.)

T o summarize the story, Figure 78 shows that because o f the political situation in the period 1847-1850 the nearest that surveyors could get to make observations was the Plains o f northern India, about 110 miles (160 km) from the peak. Although six observing stations were used, they were far from ideally placed since the maximum intersection angle was 36. The six results varied by only 36 ft (11m), and the mean value was 29 002 ft. (8840 m ). The refraction effects were found to be up to 1375 ft, and an angular variation between morning and afternoon o f up to 200". Such agree ment under the circumstances was phenomenal. The biggest sources o f error were considered to be: the effect o f refraction with an uncertainty o f perhaps 120 ft (36.6 m) uncertainty in the geoidal separation o f perhaps 115 ft (35.1 m) In September 1992 an international Italian-Chinese expedition remeasured the height. This they did using all the modern techniques o f GPS, EDM, and electronic theodolites and related the result to sea level in the Bay o f Bengal. Simultaneous observations were made from both Tibet and China with a reflector and satellite receiver on the summit. The value quoted for the orthometric elevation was 8846.10 m. The previous Chinese value o f 1975 had been 8848.13 m. The agreement both between the two Chinese values and with the original value o f 8840 m is extremely good. It shows that the original work was far better than to the nearest 1000 ft. It also shows that original estimates for the geoidal separation were good. For further detail see Smith (1997).

C H A N G IN G N A T IO N A L G R ID A N D G E O D E T IC D A T U M IN G R E A T B R I T A I N
Why change to a new datum? There are various reasons, among which Wilson and Christie (1992) give the following: GPS measurements can now detect any shortcomings o f an existing map ping datum. Existing datums are unique to Great Britain and cannot be extended to cover the rest o f Europe. The need for more o f an international datum is overwhelming. GPS has removed the need for any intervisibility between stations; hence there is now no need for expensive hilltop monuments. With the rapidly increasing popularity o f GPS the datum and related topics have to keep in step. New applications are arising regularly and require greater accuracies than can be supplied via the existing system.

The first national mapping o f Great Britain was based on 13 different meridians on the Cassini projection. Later this was changed to the National Grid based on a m odified Transverse Mercator projection. This was latei supplemented with a Europe-wide system called the European datum 1950 (ED 50). Th e National Grid is based on the Ordnance Survey Great Britain 1936 (OSGB 36) datum, whose origin is at Herstmonceux where the orientation and position o f the Airy spheroid is defined. Since 1936 the primary triangulation has been readjusted twice. The first o f these resulted in the OSGB 70(SN) datum, in which, the coordinates o f Herstmonceux were held fixed at their previous values. Th e outcome o f this readjustment indicated that OSGB 36 was not only too large overall, but that it varied considerably in scale from one part o f the country to another. The next readjustment indicated short comings in the OSGB 70(SN), and so the process continues. With the advent o f satellite geodesy, GPS positional values were given in terms o f WGS 84. However, in Europe there were efforts to improve on this, and a continent-wide campaign o f observations took place in 1989. Some 60 receivers were used on 93 stations throughout Europe, and the results were subjected to a rigorous adjustment. Th e result was the European Terrestrial Reference Frame 1989 (ETRF 89), which is now the fundamental frame throughout Europe. In this, the standard errors attributed to the observing stations are around the 10-cm level and expected to be improved to 1 or 2 cm. What it has shown, however, is that the original OSGB 36 triangulation and its equivalents in other European countries are not compatible with ETRF 89 and that unfortunately the differences are not uniform. As a result coordinates in WGS 84 will differ from coordinates in OSGB 36 by hundreds o f meters due solely to the difference o f datums and not to error. This discrepancy obviously becomes a problem for a national mapping organization. After national consultation a policy was devised to support trans formations between the national grid (OSGB 36) and WGS 84 (or ETRF 89). T o this end, conversion tables have been compiled and made freely available that will have sufficient (2-m) accuracy for mapping purposes. A more precise transformation (0.2 m) is also available as a service. In tandem with this is a transformation service for situations requiring higher accuracies. For the time being, mapping will be retained on OSGB 36. A parallel development has been a national geoidal m odel showing the geoidal-ellipsoidal separations with a claimed accuracy o f 5 to 7 cm. For further details, see Calvert (1994).

Bibliography

a or readers interested in pursuing the subject in more depth the following is a selection o f further reading. Numerous other papers are available and in particular it is recommended to obtain the proceedings o f relevent confer ences, seminars and symposia for details o f the latest state o f this fast chang ing science.

Abbreviations ACSM American Congress o f Surveying & Mapping AFT Association Franaise de Topographie Ansi. Sur. Bull. Geod. Can. Sur. Australian Surveyor Bulletin Godsique Canadian Surveyor N.Z. Sur. P.O.B. New Zealand Surveyor

Point of Beginning

RAS Royal Astronomical Society RICS Royal Institution o f Chartered Surveyors S. & M. Sur. Rev. S. W. Surveying and Mapping Survey Review

C. E. Sur. Civil Engineering Surveyor FIG Fdration Internationale des Gomtres G.l.M. Geodetic Info Magazine

Surveying World

Aardoom , L. 1972. On a geodetic application o f multiple-station very long base interferometry. Allman, J.S. & Veenstra, C. 1984. Division o f National Mapping, Australia.

Netherlands Geodetic Commission, new series, 5, No. 2. Geodetic model of Australia 1982. Bull. Geod.

Tech. Report 33.

Anderle, R.J. 1980. Accuracy o f mean earth ellipsoid based on doppler, laser and altimeter observations. No. 54:521-527. Anderle, R.J. 1981. Geodetic control from electronic satellite observations. FIG, Montreux. Paper 506.1. Ashgaee, J . 1986. IC2 technique fo r GPS signal processing in survey applications. FIG, Toronto. Paper 509.1. Ashkenazi, V. 1985a. Positioning by satellites: trends and prospects. RAS Mason Confer ence. Glasgow. Ashkenazi, V. 1985b. Positioning by GPS and NAVSAT. Will it be the end of geodetic networks? Paper H3. Survey and Mapping 85 (con feren ce). Reading.

_*i,yiiv /v ji ri| 1 iy

Ashkenazi, V. 1986. Satellite geodesy its impact on engineering survey. FIG, Toronto. Paper 502.1. Ashkenazi, V. 1994a. T h e impact o f GPS on civil engineering projects: a blessing or a curse? Proceedings o f 3rd C ITO P Conference, London, Dec. 9, 1994. Published by A F T and RIGS. Ashkenazi, V. 1994b. GPS and maps. FIG Congress, Melbourne. Paper TS 505.1. Ashkenazi, V. & Gough, R.J. 1975. Determination o f position by satellite-doppler techniques. Univ. o f Nottingham. Ashkenazi, V. & McLintock, D. 1982. Very long baseline interferometry: an introduction and geodetic applications. Sur. Rev. 26, No. 204:279-288. Barr, J.R. & Carriere, R.J. 1981. Accuracy criteria for inertial surveys. FIG, Montreux. Paper 512.2. Bevin, A.J. & Hall, J. 1994. The review and development o f a modern geodetic datum. FIG Congress, Melbourne. Paper TS 506.3. Billaris, H. et al. 1991. Geodetic determination o f tectonic deformation in central Greece from 1900 to 1988 Nature 350, No. 6314:124-129. Bloodgood J.F. 1992. Jumping on the GPS Bandwagon. ACSM Bulletin No. 136 (M arch/April):32-34. Bomford, G. 1971. Geodesy, 3rd ed. O xford Univ. Press, Oxford. Bossier, J.D. & Hanson, R.H. 1984. The impact o f VLB1 and GPS on geodesy. S. & M. 44, No. 2:105-113. Bossier, J.D. et al. 1980. Using the global positioning system (GPS) for geodetic psoitioning. Bull. Geod. No. 54:553-563. Brown D.C. 1976. Doppler surveying with theJMR-1 receiver. Bull. Geod. No. 50:9-25. Butelli, G. et al. 1994. GPS measurements for ground subsidence m onitoring and geoid computation in the Ravenna area. FIG Congress, Melbourne. Paper TS 501.3 Calvert, C. 1994. Great Britain-changing the national grid and geodetic datum. 5. W., Sept., pp. 23-24. Carter, W.E. 1981. An elementary introduction to radio interferometric surveying. Sur. Rev. 26, No. 199:17-31. Cavero, A.P. & Czarnecki, K. 1994. GPS-related reference system as applied to engineer ing purposes. FIG Congress, Melbourne. Paper SGS 653.3. Chen, J. Y. 1994. Crustal movements, gravity field and atmospheric refraction in the Mt Everest area. Zeitschrift fu r Vermessungswesen No. 8, pp. 389-400. Christie, R.R. 1992. The establishment o f a new 3D control framework for GB based on modern space techniques. 5. W., Nov., pp. 41-43. Chrzanowski, A. et al. 1995. US Supercollider-geodetic control. 5. W., March, pp. 22-26. Clarke, F.L. 1980. Determination o f a local geodetic datum from astrogeodesy. Sur. Rev. XXV, No. 196:268-279. Cleasby, C. 1992. The determination o f height using GPS. C. E. Sur., Oct., pp. 25-27. Collier, P.A. 1994. Static and kinematic GPS for deformation monitoring. FIG Congress, Melbourne. Paper SGS 552.3. Collier, P.A., Armstrong, A.P. & Leahy, F.f. 1994. GPS heighting by least squares collocation-initial results and experiences. FIG Congress, Melbourne. Paper T A

501.2.

biDiiograpny

Cook, A .II. 1969.

(Wavity and the earth.

Wykeham Publications.

Cross, P.A. 1986. Prospects for satellite and inertial positioning methods in land survey ing. Land and Minerals Surveying, April, pp. 196-203. Cross, P.A. 1991. Position: just what does it mean?

Jnl. Inst. Navigation,

pp. 246-262.

Cross, P.A. & W ood, J.P. 1980. Instrumentation and methods fo r inertial surveying. Autumn.

Chartered Land & Minerals Surveyor,

Danson, E.F.S. 1991. Introduction to plane and geodetic coordinate systems. C. July/Aug., pp 28-29; Sept., pp. 11-12.

Sur.,

E.

Deakin, R.E., Collier, P.A. & Leahy, F.J. 1994. Transformation o f coordinates using least squares collocation. March, pp. 6-20.

Aust. Sur.,

Dewhurst, W.T. 1984. Local gravity estimates available from National Geodetic Survey. June, pp. 13-16.

ACSM Bull.,

Dodson, A.H. 1994. The status o f GPS for height determination. FIG Congress, Mel bourne, Paper TS 501.1. Doyle, D. & Dragoo, A. 1994. W here freedom stands. pp. 4-9. Egger, K. 1995. The survey o f the Gotthard base tunnel. 5. W., March, pp. 18-21. Ezeigbo, C.U. 1994. O n the choice o f a suitable datum for a unified geodetic network o f Africa. & Dec., pp. 385-392.

Professional Surveyor, Jan./Feb.,

S. African Jl. of Surveying Sur. Rev.

Mapping,

Featherstone, W.E. & Olliver.J.G. 1994. A new gravimetric determination o f the geoid o f the British Isles. 32, No. 254 (O ctober) :464-478. Fell, P.J. 1980. A comparative analysis o f GPS range, doppler and interferometric observations for geodetic positioning. No. 54:564-574.

Bull. Geod.

Ferguson, J. 1990. H ow the flat earth got round: an ancient history. Summer, pp. 8-9.

Surveyor,

Ontario Land

Furguson, J. 1991. What is height anyway?

Ontario Land Surveyor, Northpoint,

Spring, pp. 17-19.

Fischer, I. 1974. A continental datum fo r mapping and engineering in South America. FIG, Washington, September. Fowler, R. 1992. All about coordinate systems. pp. 24-28. Summer, pp. 20-24; Fall,

Frei, E. & Schubernigg, M. 1992. GPS surveying techniques using the fast ambiguity resolution approach (F A R A ). 34th Australian Surveyors Congress, Cairns. Frodge, S.L. et al., 1994. Results o f real-time testing o f GPS carrier phase ambiguity resolution on-the-fly. FIG Congress, Melbourne. Paper TS 606.3. Garland, G.D. 1977.

The earths shape and gravity.

Pergamon Press, Toronto.

Gilliland, J.R. 1984. Comparison o f height derived from a combined solution o f doppler and free air geoid data with A H D values. 32, No. 3:186-192. Goad, C.C. 1989. On the move with GPS.

Aust. Sur. P. O. B., Apr./May,

pp. 36-44.

Greggor, K.N. 1993. GPS in Lesotho-the heighting problem and rapid height monitor ing by GPS. /. (S. Africa) 27, No. 3:43-58.

Inst. Mine Surveyors

Hajela, D. 1990. Obtaining centimeter-precision heights by GPS observations over small areas. Jan./Feb., pp. 55-59.

GPS World,

Hannah, ). 1985. The global positioning system the positioning tool o f the future. Z. Aug., pp. 268-281.

N.

Sur.,

I laug, M.D. et al. 1980. A simplified explanation o f doppler positioning. S. No. 1:29-45.

& M.

40,

BiDiiograi > tly

Helm er, G. 1991. Which way is up in earth orbit? IVo/essional Surveyor, Inly/Aug.. pp. 39-44. Hoar, G.J. 1982. Satellite surveying. Magnavox, Torrance, GA. Hogarth, B. 1990. The global positioning system. C. E. Sur. Suppl., May, pp. 2-4. Hollmann, R. & Welsch, W.M. 1994. GPS for precise engineering surveying applications, FIG Congress, Melbourne. Paper TS 606.1. Homes, G.M. 1992. Changing technology-monitoring sea level change. Aust. Sur. 37, No. 1 (M arch): 13-22. Hothem , L.D. et al. 1984. GPS satellite surveying-practical aspects. Can. Sur. 38, No. 3:177-192. Hui, P.J. 1982. On satellite signal processing techniques applicable to GPS geodetic equipment. Can. Sur. 36, No. 1:43-54. Jackson, J.E. 1980. Sphere, spheroid and projections fo r surveyors. Granada, London. King-Hele, D.G. 1960. Satellites and space research. Routledge & Kegan Paul, London. King-Hele, D.G. 1963. The shape o f the earth. Institute o f Navigation, London. King-Hele, D.G. 1969. The shape of the earth. Royal Aircraft Establishment, Farnborough. Kinlyside, D.A. 1994. The geocentric datum o f Australia transform or readjust? FIG Congress, Melbourne. Paper TS 506.1. Kleusberg, A. & Langley, E.B. 1990. The limitations o f GPS. GPS World, Mar./Apr., pp. 50-52. Lachapelle, G. 1986. GPS-current capabilities and prospects fo r multi-purpose offshore surveying. FIG, Toronto. Paper 502.2. Langley, R.B. 1990. Why is the GPS signal so complex? GPS World, May /June, pp. 56-59. Langley, R.B. 1992. Basic geodesy for GPS. GPS World, Feb., pp. 44-49. Leach, M.P. & Cardoza, M.A. 1994. GPS monitor stations, reference stations and networks. FIG Congress, Melbourne. Paper TS 504.2. Leick, A. 1993. Geodesy in review. ACSM Bulletin, Jan/Feb., pp. 45-48. Leick, A. 1994. GPS Satellite surveying, 2nd ed. Wiley Interscience, New York. Levallois, J.J. 1988. Mesurer la Terre. 300 ans de geodesie franaise. Presses de Ponts et Chausses, Paris. Lippold, H.R. 1980. Readjustment o f the National geodetic vertical datum. S. & M. 40, No. 2:155-164. Macdonald, A.S. & Christie, R.R. 1991. From miles to millimetres: the story o f geodesy at Ordnance Survey, 1791-1991. Sur. Rev. 31, No. 241:126-147. Manning, J. & Harvey, W. 1994. Status o f the Australian geocentric datum. Aust. Sur., March, pp. 28-33. McClusky, S. et al. 1994. The Papua New Guinea satellite crustal motion surveys. Aust. Sur., Sept., pp. 194-214. McLintock, D.N. 1987. GPS-Land Survey applications. RICS/Royal Institution o f Navi gation meeting. London. Meade, B.K. 1983. Latitude, longitude and ellipsoid height changes N A D 27 to pre dicted NAD 83. 5. & M. 43, No. 1:65-71. Milbert, D.G. 1992. GPS and geoid 90 the new level rod. GPS World, Feb., pp. 38-43. Moritz, H. 1990. The figure o f the earth. 'Theoretical geodesy and the earth's interior. Wichmann, Karlsruhe.

Mueller, 1.1. 1981. Inertial survey systems in the geodetic arsenal. Bull. Geod. No. 55:272-285. Mueller, 1.1., Hannah, J. & Pavlis, D. 1981. Inertial technology for surveying. FIG, Montreux. Paper 501.4. Nash, D., W hiffen, P. & Dixon, K. 1995. Surveying Kazakhstans rich resources with GPS. GPS World, Feb., pp. 22-30. Olliver, J.G. 1981. Satellite-derived geoids fo r Great Britain and Ireland. Sur. Rev. 26, N o. 202:161-179. Olsen, N.T. 1993. Understanding the differences between an ellipsoid, a geoid and a spheroid. G.I.M., May, pp. 64-65. Ordnance Survey. 1995a. Th e ellipsoid and the transverse mercator projection. Geodetic Information. Paper 1. Ordnance Survey. 1995b. National grid/ETRF 89 transformation parameters. Geodetic Information. Paper 2. Parkinson, B.W. 1979. The global positioning system (N A V STA R ) Bull. Geod. No. 53:89-108. Pelletier, M. 1990. La Carte de Cassini. Presses de Ponts et Chausses, Paris. Poretti, G. & Beinat, C.M.A. 1994. GPS surveys on Mount Everest. GPS World, Oct., pp. 33-44. Rapp, R.H. 1994. Separation between reference surfaces o f selected datums. Bull. Geod. 69:26-31. Rizos, C., Stolz, A. & Masters, E.G. 1984. Surveying and geodesy in Australia with GPS. Aust. Sur. 32, No. 3:202-224. Schwarz, K.P. 1981. Error characteristics of inertial survey system s. FIG, Montreux. Paper 512.1. Seeber, G. 1993. Satellite geodesy: foundation, methods and applications, de Gruyter, Berlin and New York. Seeger, H. 1994. EUREF: T h e new European Reference Datum and its relationship to WGS84. FIG Congress, Melbourne. Paper TS 506.4. Sherwin, T. 1995. A look at some o f the main tidal constituents. Hydrographic J. No. 75 (Jan.):15-19. Smith, J.R. 1969. The developm ent o f two standards. Sur. Rev. No. 153:133-146. Smith, J.R. 1983. Geodim eter 1947-1983. Geotronics, Huntingdon, UK. Smith, J.R. 1987. From plane to spheroid. Landmark Enterprises, Rancho Cordova, CA. Smith, J.R. 1996. Capt. R. S. W ebb MBE, RFA (1892-1976): From Shropshire to Paarl via Geodesy and Lesotho. S. African Council fo r Professional and Technical Surveyors, Durban Smith, J.R. 1997. Everest: The man and the mountain. Landmark Enterprises, Rancho Cordova, CA. In press. Snyder, J.P. 1993. Flattening the earth: Two thousand years of map projections. Univ. o f Chicago Press. Soler, T. & Hothem , L.D. 1988. Coordinate systems used in geodesy: basic definitions and concepts./. Surveying Engineering 114, No. 2:84-97. Sprent, A. 1992. VLBI, GPS and Greenhouse. Aus. Sur. 37, No. 1:23-32. Stanbridge, M.J. 1979. Doppler survey a review o fth e techniques, systems, equipment, computation and typical results. Conference o f Commonwealth Surveyors. Paper B2. Cambridge, UK.

Stansell, T.A. 1983. The Transit navigation satellite system. Magnavox, Torrance, < .A Stead, J. & Holtznagel, S. 1994. A I I I ) heights from (IPS using AUSGEOID93. Au\t Sur., March, pp. 21-27. Stolz, A. et al. 1983a. Australian baselines measured by radio interferometry. Aust. Sin , 31, N o 8:563-566. Stolz, A. et al. 1983b. Geodetic surveying with quasar radio interferometry. Aust. Sin 31, No. 5:305-314. Stolz, A. & Masters, E.G. 1982. Studying the tectonics o f Australia by satellite lasei ranging. Aust. Sur. 31, No. 1:34-44. Stolz, A. & Masters, E.G. 1983. Satellite laser range measurements o f the 3200 km Orroral-Yarragadee baseline. Aust. Sur. 31, No. 8:557-562. Strasser, G. 1974. The toise, the yard and the metre; the struggle fo r a universal unil o f length. 17th Australian Survey Conference, Melbourne. Thomas, T.L. 1982. T h e six methods o f finding north using a suspended gyroscope. Sur. Rev. 26, No. 203:225-235; No. 204:257-272. Thompson, S.D. Everymans guide to satellite navigation. A R IN C Research Corporation. Treftz, W.H. 1981. An introduction to inertial positioning as applied to control and land surveying. S. & M. 41, No. 1:59-67. Vincenty, T. 1987. Geoid heights for GPS densification. ACSM Bull., Dec., pp. 25-26. Wells, D.E. et al. 1982. Marine navigation with NAVSTAR/Global positioning system (GPS) today and in the future. Can. Sur. 36, N o 1:9-26. Wells, D.E. et al. 1986. Guide to GPS positioning. Canadian GPS Associates, Univ. ol N ew Brunswick. Wells, D. & Kleusberg, A. 1990. GPS: a multipurpose system. GPS World, Jan./Feb., pp. 60-63. Wilson, J.l. & Christie, R.R. 1992. A new geodetic datum fo r Great Britain. T h e Ordnance Survey Scientific GPS Network SCINET 92. Ordnance Survey.

A A U U U l II I d

n_/i

James (Jim) Smith was educated at Devizes Grammar School and S. W. Essex Technical College (now the University o f East London). There he qualified lor membership o f the Royal Institution o f Chartered Surveyors in the Land Survey Division, where he has been active on various committees and the I )ivisional Council for nearly 30 years. H e was closely involved in development o f the monthly journal Land and Minerals Surveying and its successor Surveying World. H e at present assists with the contents o f The Civil Engineering Surveyor. Internationally he was secretary o f Commission 6 (Engineering Surveying) o f the FIG (International Federation o f Surveyors) for nine years and is now secretary to its ad hoc Commission on the History o f Surveying. He has served on British Standards Institution and International Standards Organization working groups on surveying. H e was a principal lecturer in land surveying in the civil engineering depart ment at the University o f Portsmouth, Portsmouth, England, for nearly 25 years until taking early retirement in 1990. H e is the author o f several books and numerous booklets and technical papers, and is form er editor o f a series o f books on Aspects o f Modern Surveying. He has recently had published Capt. R. S. Webb MBli, RFA (1892-1976): from Shropshire to Paarl via Geodesy and Lesotho, and has Everest: the Man and the Mountain awaiting publication.

'

Index

Aberration, 61, 63 Acadmie Royale des Sciences, 40, 51 founding of, 19 Accelerometer, 176, 177 Airy, Sir George Biddell (1801-1892), 31, 113, 121 spheroid of, 31 Al-Mamun, Caliph (786-833), 15 method of, 14 Altimetry, satellite, 107 Altitude: definition of, 51 Ambiguity: carrier beat phase measurement of, 165 initial phase, 165 pseudo-range measurement of, 165 Angle: measure of, 44, 65 Anomaly: Bouguer, 119, 202 definition of, 115 free-air, 119, 202 isostatic, 121 Antispoofing (AS), 166 Arago, Dominique Franois Jean (1786-1853), 180 Arc: Al-Mamun, 15 Archimedes, 7 Aristotle, 7 18th century, 20 Kratosthenes, 7 F r .r tl, 17

Frisius, 17 I-Hsing, 13 Lapland, 21, 23, 29 Mason and Dixon, 24 measurement of, 129 Peru, 21, 23, 29, 79, 99 Picard, 17 Poseidonius, 11 Snellius, 17 Struve, 95 Archimedes, (287-221 B.C.), 7 Arc to chord correction, 73 Aristotle, (c384-322 B.C.), 7 Ascension Island, 161 Aslakson, Carl, 126, 180 Azimuth: astronomical, 85, 87 definition of, 51 determination of, 55, 89 Barometer: aneroid, 79 mercury, 79 Baseline: measurement of, by, bimetallic rods, 65 glass rods, 65 metal bars, 65 steel chain, 65 Bearing: definition of, 61 Benchmark, 73 Benoit, Justin-Mirande-Rene (1844-1922), 44 Bergstrand, Erik (1904-1987), 67, 126, 180

Bessel, Briedrich Wilhelm (1784-1846), 31 earth quadrant, 43 spheroid parameters, 31 Bird, John (1709-1776): scale of, 25 Borda, Jean-Charles de (1733-1799), 40, 41, 43 circle, 41 Metre des Archives, 42 Bouguer, Pierre (1698-1758), 21, 41, 79, 99, 111, 113, 202 Boundary, International, 95 Bowie, William (1872-1940), 121 Bradley, Rev James (1693-1762), 23, 61, 180 ' Brahe, Tycho (1546-1601), 17 Brisson, Mathurin-Jacques (1723-1806), 41 Bureau Internationale des Poids et Mesures, 44 Camera, Baker-Nunn, 139 Cassinis, The, 19, 23, 39 Cayenne, 172 Celsius, Anders (1701-1744), 21 Chandler, S C, 63 period, 63 Channel tunnel, 199 Charlemagne. (742-814), 38 Charles II, King (1630-1685), 51 Chimborazo, Mount, 111 Chronometer, 55 Clairaut, Alexis-Claude (1713-1765), 21 Clarke, Alexander Ross (1828-1914), 29, 31 earth quadrant, 43 spheroid parameters, 31 Clock: atomic, 59 cesium, 161 correction for, 91 hydrogen, 161 maser, 161 pendulum, 56

quartz crystal, 59, 161 rubidium, 161 Collimation, plane of, 73 Colorado Springs, 161 Columbus, Christopher (1446-1501), 15, 21 Condorcet, Marie-Jean-AntoineNicolas Caritat, Marquis de (1743-1794), 40 Continental drift, 185 Convergence, definition of, 71, 197 Coordinate: astronomical, 85 transformation, 93 state plane, 197 Corps Lgislatif, 42 Coriolis effect, 107, 176 Correction: for height, 119 for topography, 119 Cte-dOr, Claude Antoine, Prieur de la, 39 Cubit, 36, 37 Assyrian, 38 Hashimi, 38 Royal, 38 Curvature, 79 Datum: African, 99 Astrogeodetic, 89 before W orld War II, 93 Cape, 200 connection, 93 discrepancies between, 91 ED 50, 96, 206 European, 95, 96, 97, 133, 135 Geodetic, 83, 85, 93, 205 in Great Britain, 205 horizontal, 83, 101 Indian, 95, 96, 97, 108 Mean sea level, 99 NAD 27, 95, 108 NAD 83, 96, 108 North American, 95, 133, 135 orientation of, 85, 87

II l U t J A

Datum: ( contd ) Origin, 85 OSGB 36, 206 OSGB 70(SN), 206 OS(SN)80, 108 preferred, 95, 97, 131, 135 Pulkovo, 96, 129, 200 Satellite, 91 shift, 108, 200 single astronomical, 85 Tokyo, 95, 96, 97, 133, 135 transformation, 93, 206 parameters, 93 U.S. Standard, 95 vertical, 83, 99, 101 WGS 72, 97, 108 WGS 84, 108 Day: apparent solar, 56 definition of, 56 mean solar, 56 Declination: definition of, 49 Deflection: astrogeodetic, 133 gravimetric, 111, 133 Degree: centesimal, 44 sexagesimal, 44 Delambre, Jean-BaptisteJoseph (1749-1822), 41 Diego Garcia, 161 Direction: definition of, 51 Dixon, Jeremiah (1733-1779), 24, 25 method of, 24, 25 Doppler, Johann Christian (1803-1853), 141 count, 145 method, 99, 107, 139, 141, 202 shift, 151 Transit, 149, 151 Earth: axis, 85

center o f mass, 145 crustal movement, 156, 185 curvature, radii of, 19 density, 201 determination as a sphere, 6 flattening, 111 gravity field, 113 mass, 91 orbit of, 61 parameters, 31 rotation of, 91 rotation parameters, 156 shape, controversy of, 19, 21 velocity of, 61 Echo: sounding, 79 Sounder, 80 Ecliptic, 56 Edgar, King (944-975), 38 Edward I, King (1239-1307), 38 Eiffel tower, 203 Electromagnetic distance measurement (ED M ), 69, 97, 181 Electrooptical methods, 180, 183, 185 Elevation: effect of, 28 Elizabeth I, Queen (1533-1603), 38 Ellipse: elements of, 30 Ellipsoid: axis o f rotation, 85 Bessel, 96, 97, 192 best fitting, 89, 97, 135 choice of, 97 Clarke 1866, 97 Clarke 1880, 192, 200 Clarke 1886, 192 defined, 32, 33 derivation of, 129 Everest, 96, 97, 102 International, 96, 97, 192 normal, 87, 89 o f revolution, 33

Ellipsoid: ( contd ) orientation of, 85, 89 parameters, 31, 97 reference, 87, 91 relation to geoid, 34, 75, 85 suitable, 124 WGS, 129 Etvs, Baron von Roland (1848-1919), 172 effect, 172 Ephemeris: astronomical, 57 broadcast, 149 precise, 149 Equator, circumference of, 40 Equinoctial points, 56 Equinox: precession of, 61 Equipotential surface, 35, 75, 77, 103, 108 Error: sources of, 169, 170 European Reference Frame (EUREF), 136 European Space Agency (ESA), 161 European Terrestrial Reference Frame 1989 (ETRF 89), 136, 206 Essen, L, 180 Eratosthenes, (276-195 B.C.), 7, 11, 37 method o f determination, 8, 9 Everest: Mount, 79, 202 Sir George (1790-1866), 29 spheroid parameters, 31 Fernel, Jean Franois (1497?1558), 17 Figure, strength of, 165 Fischer, Irene, spheroid parameters, 31 Fizeau, Armand-Hippolyte-Louis (1819-1896), 180 Flamsteed, John (1646-1719), 51

Foot, 37 Amsterdam, 38 English, 25, 38 French, 25, 41 Geodetic, 39 National, 40 Roman, 38 Royal, 38 Foucault, Jean-Bernard-Leon (1819-1868), 180 Frequency standard, 161 Frisius, Reiner Gemma (1508-1555), 17 triangulation of, 16 Gal, 113, 171 Galilei, Galileo (1564-1642), 171 Geodesic, 73 Geodesy: definition of, 2 function of, 47 International Association of, 3 Geodetic Reference System: 1967 (GRS 67), 115 1980 (GRS 80), 136 Geodimeter, 67, 139, 180 characteristics of, 69 Geoid, 34, 35, 55, 91, 115, 129, 202 gravimetric, 135 height above, 75, 131 model, 206 separation, 75, 77, 85, 89, 198, 199, 203, 205 surface, 107 undulation, 75, 111, 121, 133, 135, 199, 201 absolute, 131 Geoid90, 35 Geoid93, 35 Gill, Sir David (1843-1914), 44 Global Positioning System (GPS), 59, 163, 198, 199, 202, 205, 206 codes, 166 heighting by, 200 positioning methods, 163-165 differential, 167

Global Positioning System (GPS) ( conl'd) phase measurement, 169 relative, 167 real time, 167 Glonass, 161 Gon, 44 Gotthard tunnel, 199 Graaff-Hunter, James de (1882-1967), 133, 203 Grad, 44 Grade, 44 Graham, George (1675-1751), 25 Gravimeter, 172 Gravimetric: method, 121, 129 orientation, 124, 131 Gravitation, Law of, 101 Gravitational constant, 91 Gravity: absolute measure, 171 acceleration of, 113 anomalies, 115, 121 definition of, 115 data, 124 collection of, 127 direction of, 77 field, 111, 176 formula, 115 Meter, 172 observed, 123 relative, 172 theoretical value of, 115 values, 124, 127 Greggor, K N, 200 Greeks, ancient, 129 Grid: arbitrary local, 71 change of, 205 line, 171 national, 71 Gyroscope, 174, 178 Gyrotheodolite, 174, 175, 199 Harrison, John (1693-1776), 55 Hawaii, 161

Hayford, John (1868-1925), 31, 119,121 spheroid parameters, 31 Height: barometric, 79, 163 datum, 101 dynamic, 75 ellipsoidal, 75, 77, 198 errors in, 203 geoidal, 75, 135, 201 GPS, 77 orthometric, 75, 77, 198, 201 satellite, 107, 201 trigonometric, 79 Heiskanen, Weiko Aleksanteri (1895-1971), 121 Helmert, Friedrich Robert (1843-1917), 31, 121, 129 spheroid parameters, 31 Henry VII, King (1457-1509), 38 Hipparchus, (cl80-125 B.C.), 61, 63 Hiran, 97, 126, 129, 135 Hour angle: definition of, 51 Hutton, Charles (1737-1823): Abridged Philosophical Transactions, 25 Huygens, Christiaan (16291695), 39 Hypsometer, 79 I-Hsing, (682-727), 13 method of, 12 Inch, 36 British, 44 Chinese, 13, 37 present day, 38 U.S., 44 India, Survey of, 113 Inertial surveying, 176 Initial point, 131 Interferometry, Very long baseline, 103, 137, 141, 155 International Earth Rotation Service (IERS, TR F ), 103, 137

International spheroid parameters, 31 International Terrestrial Reference System (IT R S ), 136, 137 International Union for Geodesy and Geophysics (IU G G ), 137 Invar: invention of, 65 Isostasy, 113 Italian-Chinese expedition, 205 IUG G spheroid parameters, 31 Kater, Henry (1777-1835), 172 Krassovsky, Theodosy Nicolaevich (1878-1948), 31, 129 spheroid parameters, 31 Kwajalein, 161 La Caille, lAbbe Nicholas-Louis (1713-1762), 41 La Condamine, Charles-Marie de (1701-1774), 21, 41 LAGEOS, 185 Lagrange, Toseph-Louis de (1736-1813), 40 Laplace, Pierre-Simon, Marquis de (1749-1827), 40, 87 equation, 87, 89 station, 87 Lapland: arc of, 21, 29 Laser ranging, 107, 137, 156, 185 Latitude: astronomical, 89 determination of, 49, 53 geodetic, 85, 89 Length, units of, 36 Lesotho Highlands Water Scheme, 200 Leveling: geodetic, 73, 105 precise, 73 spirit, 73 Li, Chinese, 37

Light: velocity of, 44, 59, 61, 67, 143, 167, 179, 180 wavelength of, 36, 43 Ligne, 36 Paris, 41 Line: weighted, 80 Links, Intercontinental, 126 Longitude: astronomical, 87 determination of, 49, 53 geodetic, 85, 87 Louis XVI, (1754-1793), 39 Lunar Laser Ranging (L L R ), 137 Map projections, 29 Maskelyne, Rev Nevil (1732-1811), 111 Mason, Charles (1728-1786), 24 method of, 24, 25 Mass: deficiency of, 123 distribution of, 11 Maupertuis, Pierre-Louis Moreau de (1698-1759), 21, 23, Mechain, Pierre-Francois-Andr (1744-1804), 4 Mekometer, 69 Meridian: Canary Islands, 53 central, 73 Greenwich, 53, 55 Paris, 53 Meter, 37, 39-44 calculation of, 41 definition, 44 definitive, 41 Legal, 42 prototype, 44 provisional, 41 redefined 1960, 4 redefined 1983, 44, Metre: des Archives, 41, 42

M ich elson , A lb e rt A (1 8 5 2 -1 9 3 1 ),

44, 180 Microwave instruments, 133, 185 Mile: Arab, 15, 37 Italian, 15, 37 Roman, see Mile, Italian Swedish, 37 Milligal, definition of, 113 Module, 4 Monge, Gaspard (1745-1818), 40 Mouton, Gabriel (1618-1694), 38 NAD 27 parameters 31 NAD 83 parameters, 31 Napoleon, 43 Navigation Satellite (N AVSAT), 161 Navigation System using Tim e and Ranging (NAVSTAR), 159, 166 Network: first order, 71 national geodetic, 89 second order, 71 third order, 71 Newton, Sir Isaac (1642-1727), 19, 21, 23, 63, 101 Law, 101 North: grid, 53, 71 magnetic, 53, 73 true, 53, 71 Number: geopotential, 75 Nutation, 61, 63 Observations: circummeridian, 53 Observatory: Greenwich, 51, 57 Paris, 19 Ordnance Survey Geodetic Reference System 1980 (OSGRS 80), 136 Orientation: astrogeodetic, 85, 87, 91, 93 ellipsoid to geoid, 85

Origin, 83 Australian, 89 Helm ert Tower, 96 Kalianpur, 96 Manchurian, 96 Meades Ranch, 97 Russian Pulkovo, 96, 129, 200 Tokyo, 96 Oscillator rate, 91 Pendulum, 171 half second, 172 reversible, 172 seconds, 39, 40, 172 Peru: arc of, 21, 29 Petrie, Sir William Matthew Flinders (1853-1942), 38 Picard, PAbbe Jean (1620-1682), 17, 39 Pied du Roi, 38, 39 Point positioning, 145 Polar motion, 63 Pole Star, 49, 55 altitude of, 53 Poseidonius, (cl3 5 50 B.C.), 10, 11 Positioning: horizontal, astronomical, 49 methods, 47 vertical, 73 methods, 47 Pratt, Archdeacon John Henry (1809-1871), 113, 121 Precession, 61, 63 Precise Positioning Service (PPS), 166 Pretoria, University of, 165 Prieur, Claude Antoine, du Vernois (later de la Cote-dO r) 39 Projection, 73, 187 Cassini, 206 conformal, 187 equidistant, 187 equal area, 18

091

Projection (contd) Gauss-Kriiger, 18 Lambert conical orthomorphic, 195, 197 Mercator, 19 Transverse Mercator, 191, 192, 193, 197, 206 Universal Transverse Mercator (U T M ), 191 World, 191 Ptolemy, Claudius (100-178), 15, 187 Pythagoras, (c580-500 B.C.), 7 Quadrant, 36 meridian, 40 Radio wave, velocity of, 143, 179 Range values, 166 Rayon astronomique, 39 Refraction, 79, 151, 152, 200, 205 atmospheric, 203 lateral, 199 Riccioli, Giovanni Battista (1598-1671), 38 Richer, Jean (1630-1696), 172 Right ascension: definition of, 49 Rocket: V2, 176 Roemer, O laf (1644-1710), 180 Royal Greenwich Observatory, 51, 57 Royal Society, 40, 51 founding of, 19 Satellite: 139 altimetry, 10, Challenger, 159 Explorer I, 139 Jupiter, 180 Laser Ranging (SLR), 107, 137, 156, 185 methods, 47 point positioning by, 145 Sputnik, 138, 139

tracking, optical, 139 radar, 139 radio, 139 translocation by, 145, 147 Vanguard, 129, 139 Schehallion, Mount, 111 Sea level, 27 Australia, 99 China, 99 connection to, 99 Europe, 99 local mean, 103 mean, 103, 107 monitoring of, 103 observations for, 103 recording, 105 United Kingdom, 99 United States, 99 Sequential Collation o f Range (SECOR), 151, 152 Shoran, 97, 126 Short arc translocation, 147 Signal: broadcast, 59 time, 55 Snellius, Willibrord (1580-1626), 17, 63 triangulation of, 16 Sound: velocity of, 80 Southern Cross, 49 Spherical: excess, 73 triangle, 73 Spheroid: definition of, 33 oblate, 18, 19, 22, 27 definition, 32 parameters, 31 prolate, 18, 19 definition, 32 Stade, 36 equivalents to, 37 Standard: natural, 44

Standard: ( contd ) Positioning Service (SPS), 166 Royal Society brass, 25 Star: apparent position of, 61 catalogue, 51 Ephemerides, 51 latitude of, 61 longitude of, 61 positions of, 49 Starfix, 161 STARLETTE, 185 Station: astrogeodetic, 131 Stokes, Sir George Gabriel (1819-1903), 121 integral, 123, 133 Stonehenge, 38 Sun: imaginary means, 56 mean, 56 semidiameter, 53 Superconducting Super Collider (SSC), 198 (t-T), 73 Tallyrand, Perigord Charles Maurice de (1754-1838), 40 Telegraph: electric, 55 Telegraphy: wireless, 55 Tellurometer, 67 characteristics of, 69 Theodolite, 65, 79 electronic, 205 Kine-, 139 Thermometer: boiling point, 79 Tide: cause of, 101 diurnal, 105 gauge, 99 air-acoustic, 105 automatic, 105 lowest astronomical, 105 A

neap, 105 spring, 105 Tim e, 56 atomic, 59 international, 59 ephemeris, 57 equation of, 56 GPS, 59 Greenwich apparent sidereal, 57 Greenwich mean, 57 local, 55, 57 apparent, 57 Moscow, 161 other forms of, 59 sidereal, 57 apparent, 57 solar, 56 standard, 57 universal, 57, 161 coordinated, 59 Toise, 37, 38 du Chtelet, 38, 39 French, 25 du Nord, 39 du Prou, 39, 42 Transformation: 163 two point, 108 seven point, 108 Translocation, 145 Traversing, 65, 69, 71 Triangulation, 16, 63 adjustment, 89 networks, 65, 91 Triangulateration, 67 Trilatration, 65, 67 Tsicada, 161 United States Coast and Geodetic Survey (USCGS), 126, 197 Vandenberg, 161 Vening Meinesz, Felix Andries (1887-1966), 121, 123, 133 Vernois, Prieur Du, 39

Vertical: deflection (deviation) of, 85, 87, 89, 91, 129, 131, 177, 199, 201, 203 local, 87 Very Long Baseline Interferometry (V L B I), 137, 155 Virga, 39 Voltaire, Franois Marie Arouet de (1694-1778), 23 Wadley, Trevor (1920-1981), 67 W orld Geodetic System (W GS), 124, 129 derivation of, 124 WGS 58, 133, 135 WGS 59, 133, 135 WGS 60, 124, 126, 135 WGS 66 parameters, 31

WGS 72 param eters, 31, 136, 151 169 WGS 84 parameters, 31, 33, 136, 200, 201, 206 Wren, Sir Christopher (1632-1723) 39 Yard: redefined 1946, 44 standard, 36, 38, 43 Year: sidereal, 56 tropical, 56 Zenith: astronomical, 91 geodetic, 91 Zero velocity update (Z U P T ), 177, 178

Geodesy made easyan accessible introduction to a rapidly changing field

Geodesy is the science that deals w ith the earth's figure and the interrelationship o f selected points on its surface. This user-friendly introduction presents w hat m ight be called "geodesy w ith o u t mathematics." A revised edition of the author's Basic Geodesy\ Introduction to Geodesy avoids com plex formulas and equations and provides a broad overview for students and professionals who require a basic understanding o f the subject.

Traditional survey positioning techniques Geodetic systems, including horizontal and vertical geodetic datums Physical geodesy The w orld geodetic system Satellite geodesy The Global Positioning System Gravity measures Electromagnetic distance measurement Projections Examples o f modern projects

former principal lecturer in the Civil Engineering Department at the University of Portsmouth, is a member of the Land Survey Division o f the Royal Institution o f Chartered Surveyors, where he has been active for nearly 30 years.

Cover Design: Lynn Cole

WILEY-INTERSCIENCE

John W iley & Sons, Inc. Professional, Reference and Trade Group 605 Third Avenue, New York, N.Y. 10158-0012 New York Chichester W einheim Brisbane Singapore Toronto