HydroBASINS

Global watershed boundaries and sub-basin delineations

derived from HydroSHEDS data at 15 second resolution

Technical Documentation Version 1.c

(with and without inserted lakes)
prepared by Bernhard Lehner
([email protected])
July, 2014

1. Overview and background

This documentation accompanies a series of polygon layers that depict watershed boundaries and
sub-basin delineations at a global scale. The goal of this product is to provide a seamless global
coverage of consistently sized and hierarchically nested sub-basins at different scales (from tens
to millions of square kilometers), supported by a coding scheme that allows for analysis of
watershed topology such as up- and downstream connectivity.
Watershed boundaries provide important geospatial units for many applications, but at a global
scale there is a lack of high-quality mapping sources. The HydroSHEDS database (Hydrological
data and maps based on SHuttle Elevation Derivatives; Lehner et al. 2008; for more information
see http://www.hydrosheds.org) provides hydrographic data layers that allow for the derivation
of watershed boundaries for any given location based on the near-global, high-resolution SRTM
digital elevation model. Using this hydrographic information, watersheds were delineated in a
consistent manner at different scales, and a hierarchical sub-basin breakdown was created
following the topological concept of the Pfafstetter coding system (Verdin & Verdin 1999). The
resulting polygon layers are termed HydroBASINS and represent a subset of the HydroSHEDS
database.
The HydroBASINS product has been developed on behalf of World Wildlife Fund US (WWF),
with support and in collaboration with the EU BioFresh project, Berlin, Germany; the
International Union for Conservation of Nature (IUCN), Cambridge, UK; and McGill University,
Montreal, Canada. Major funding for this project was provided to WWF by Sealed Air
Corporation; additional funding was provided by BioFresh and McGill University.
HydroBASINS is covered by the same License Agreement as the HydroSHEDS database, which
is available at http://www.hydrosheds.org. By downloading and using the data the user agrees to
the terms and conditions of this license.
Citations and acknowledgements of the HydroBASINS data should be made as follows:
Lehner, B., Grill G. (2013): Global river hydrography and network routing: baseline data and
new approaches to study the world’s large river systems. Hydrological Processes, 27(15):
2171–2186. Data is available at www.hydrosheds.org.
 HydroBASINS Technical Documentation v1.c

2. Methods and data characteristics

All HydroBASINS layers were derived from World Wildlife Fund’s HydroSHEDS data (Lehner
et al. 2008; Lehner and Grill 2013) based on a grid resolution of 15 arc-seconds (approximately
500 m at the equator). For more information please refer to the Technical Documentation of
HydroSHEDS at http://www.hydrosheds.org. It should be noted that the quality of the
HydroSHEDS data is significantly lower for regions above 60 degrees northern latitude as there
is no underlying SRTM elevation data available and thus a coarser scale DEM has been inserted
(HYDRO1k provided by USGS; see http://gcmd.nasa.gov/records/GCMD_HYDRO1k.html).

2.1 Creation of sub-basin geometry

An important characteristic of any sub-basin delineation is the sub-basin breakdown, i.e. the
decision of when and how to subdivide a larger basin into multiple tributary basins. Standard
GIS tools offer the possibility to break out sub-watersheds at any confluence where the inflowing
branches (i.e., a tributary and its main stem) exceed a certain size threshold, typically measured
as the number of upstream pixels or the upstream catchment area. HydroBASINS follows the
same concept and divides a basin into two sub-basins at every location where two river branches
meet which each have an individual upstream area of at least 100 km2. It should be noted that
this concept still allows for smaller sub-basins to occur, namely the inter-basins between the
tributaries (which can have any smaller
size). Also, sub-basins can grow to sizes
much larger than the 100 km2 threshold if
there is no tributary joining the main stem
for a long distance. This inconsistency due
to “oversized” sub-basins has been
addressed and reduced in HydroBASINS by
forcing additional subdivisions for every
sub-basin larger than 250 km2: these
polygons are split into appropriately sized
sub-basins by introducing break points
along their main stem rivers.

2.2 Nested grouping and coding

A second critical feature of sub-basin

delineations is the way the sub-basins are
grouped or coded to allow for the breakout
of nested sub-basins at different scales, or
to navigate within the sub-basin network
from up- to downstream. One of the easiest
methods for navigation is to provide the ID
of the next downstream object, which
allows for moving from object to object in
order to traverse the network. As for nesting
and topological concepts, the ‘Pfafstetter’
coding system is frequently used due to its

2
 HydroBASINS Technical Documentation v1.c

relative simplicity and ease of application. The basic principle of the Pfafstetter coding is that a
larger basin is sequentially subdivided into 9 smaller units (the 4 largest tributaries, coded with
even numbers, and the 5 inter-basins, coded with odd numbers). Thus, the next finer resolution
of a sub-basin delineation is achieved at the next Pfafstetter level by adding one digit to the code
of the previous level. A more detailed description of the Pfafstetter coding is provided in
literature (e.g., Verdin and Verdin 1999). The concept has successfully been applied both
globally (e.g., HYDRO1k; USGS 2000) and regionally (e.g., Australia; Stein et al. 2014).

The HydroBASINS product follows the Pfafstetter concept and provides levels 1 to 12 globally.
There are several aspects of the Pfafstetter coding, however, which have not been designed for
global applications, thus the following modifications and updates were implemented:

• The first 3 levels of Pfafstetter codes for HydroBASINS were assigned manually. Level 1
distinguishes 9 continents (1 = Africa; 2 = Europe; 3 = Siberia; 4 = Asia; 5 = Australia; 6 =
South America; 7 = North America; 8 = Arctic of North America; 9 = Greenland); Level 2
splits each continent into up to 9 large sub-units; and at Level 3 the largest river basins of
each continent start to break out. From Level 4 onwards, the breakdown follows the
traditional Pfafstetter coding (with further modifications as explained below).

• To provide a unique Pfafstetter code for every basin worldwide, the seeding of first-level
coding numbers, as well as the successive seeding of all basins as they start to break out at
higher Pfafstetter levels, starts at the north-eastern tip of the African continent and continues
counter-clockwise around the continental coasts (Europe, Siberia, Asia, followed by
Australia and the Americas). Islands are inserted into the sequence at appropriate (nearest)
locations.

• The original Pfafstetter scheme was not designed to provide nested coding for islands. To
incorporate islands as sub-units of continents, they have been grouped manually in
HydroBASINS following the hierarchical Pfafstetter concept. For example, at Level 1 New
Zealand is part of the Australian continent, at Level 2 both islands of New Zealand break out
as one unit, and at Level 3 the North and South Islands are separated into their own units.

• Similarly to islands, there is no provision in the Pfafstetter scheme regarding the integration
of endorheic basins (i.e., inland sinks that do not drain to the ocean). In HydroBASINS,
endorheic basins have been grouped and then seeded with initial Pfafstetter codes manually
(similar to islands) to provide a hierarchical nesting. The nesting may start with lumped
groups of endorheic basins at lower levels and ends with the individual endorheic basins at
higher levels. Once an endorheic basin is isolated, the standard Pfafstetter coding is used to
continue the subdivision of the endorheic basin into smaller sub-basins.

• A second issue regarding endorheic basins—in particular for smaller ones—is that at the
lower Pfafstetter levels only large river basins are broken out. Many large river basins,
however, contain small endorheic basins inside them or adjacent to their watershed divide.
For example, the Nile Basin contains many small endorheic basins in the dry middle region
which are not connected to the main river via surface water flows yet are typically considered
to be part of the overall basin (as they may be hydrologically connected via groundwater). In
order to allow for these endorheic basins to be lumped with the larger river basin, and to

3
 HydroBASINS Technical Documentation v1.c

enable topological queries in which the endorheic discontinuities can be eliminated

(traversed) to create contiguous regions, some endorheic sinks have been assigned a ‘virtual’
connection to an appropriate downstream polygon. These virtual connections can be
identified in the attribute table (see details for attribute table below), and the user can decide
whether or not to terminate the routing at an endorheic sink of this type.

• Another structural problem of the Pfafstetter code is apparent for very small coastal basins
that drain into the ocean in between larger river basins. If left untreated, there would be a
very large number of very small coastal basins globally, and the Pfafstetter coding would
start to subdivide them into even smaller units. To avoid this spatial inconsistency, small
coastal basins have been lumped in HydroBASINS to form their own coastal drainage units
between larger river basins. The lumping was performed in a semi-automated procedure with
some manual decisions about their grouping, and small islands close to the coast were
included into the coastal basins where appropriate. Lumped coastal basins were allowed to
reach a maximum individual size of up to 500 km2 (or 700 km2 for the lower quality areas
north of 60ºN).

• Finally, an inherent issue of the Pfafstetter coding is the requirement to break out exactly 9
sub-basins at the next higher level. This can lead to highly inconsistent (and randomly
distributed) sub-basin sizes at the same level. For example, there may be 4 large tributaries
that form 4 large sub-basins, yet the 5 inter-basins may be very small. At the next Pfafstetter
level, the 9 sub-basins each are subdivided again into 9 nested sub-unit, thus the size
discrepancy can be passed along or even amplified (if the small inter-basins in turn contain
more small inter-basins) from level to level. To reduce this inconsistency in sub-basin size,
the breakdown in HydroBASINS was modified (in an automated procedure) to be guided by
level-dependent tolerance ranges of sub-basin areas below which small sub-basins are
allowed to skip a subdivision at the next higher Pfafstetter level. These skips are indicated
with a value of 0 in the Pfafstetter coding digit of the respective level.

2.3 Two formats of HydroBASINS: standard (without lakes), and customized (with lakes)

Version 1.c of HydroBASINS was developed in two formats:

a) In Format 1 (standard; without lakes), all sub-basins follow the standard concept of nested
watersheds and are strictly derived from the underlying flow direction grids of HydroSHEDS
by applying the area thresholds and modification rules as described above. This means that
sub-basins are also created and coded in areas which (in reality) are covered by lakes (e.g.,
inside Lake Victoria). The shape of these particular sub-basins may not reflect true sub-basin
boundaries, but their topology (i.e., interconnectedness) remains valid throughout the lake.
Format 1 has a cleaner coding with a simpler structure than Format 2, and users can visually
overlay additional lake layers to mask the affected sub-basins if needed. Format 1 consists of
12 individual polygon layers representing the 12 Pfafstetter levels. Additionally, an extra
layer (Level 0) is provided which includes all sub-basins at their smallest breakdown with the
full coding sequence of levels 1-12 in the attribute table. Users can derive any Pfafstetter
level directly from the ‘Level 0’ layer by dissolving the sub-basin polygons based on the
desired level code.

4
 HydroBASINS Technical Documentation v1.c

b) Format 2 (customized; with inserted lakes) was designed to specifically accommodate the
special requirements of the BioFresh project and was produced in collaboration with IUCN.
This dedicated version follows the same concept as the standard format, yet a global lake
layer was integrated into the data. Lake polygons were taken from the Global Lakes and
Wetlands Database (GLWD; Lehner and Döll 2004) and were clipped into the sub-basin
polygons of HydroBASINS.
IMPORTANT NOTE: While every effort was made to fully integrate the lakes into the
topological coding structure of HydroBASINS, some specific problems and artifacts were
introduced by the lake features which necessitated a modification from the original coding
scheme. For example, lakes may overlap and thus link multiple sub-basins from different
tributaries; artificial reservoirs may link formerly independent watersheds; or lagoons may
extend into the ocean and/or connect neighboring river basins. Strict up- and downstream
topology is not always possible for these special circumstances, and connectivity may thus be
inconsistent or was actively broken for certain constellations to avoid loops in the search
algorithms. Another problem is that the GLWD lake polygons and the sub-basin polygons of
HydroBASINS are not always well aligned to each other but may expose a spatial shift (in
these cases, typically the local accuracy of the GLWD data is of lower quality). Due to this
shift, some lakes (in particular small ones) may not be correctly registered to their
surrounding sub-basins but may be recognized as ‘downstream’ instead of ‘inside’, or
‘besides’ instead of ‘upstream’, etc. Finally, some connectivity rules are more complex in
Format 2. For example, there may be one or multiple lake polygons entirely within a sub-
basin at a certain Pfafstetter level. In this case, the next downstream polygon of each lake is
defined to be the surrounding sub-basin. This means in turn, however, that the lakes are
assigned to be ‘upstream’ of the surrounding sub-basin, although this is only correct for parts
of the sub-basin, and they have no up- or downstream connectivity among each other, even if
in reality they may be situated along the same river.
Two additional modifications were introduced in the lake version of HydroBASINS upon
request by BioFresh/IUCN:
• Lakes start to appear at certain Pfafstetter levels based on their size. This is to prevent
small lakes from appearing as individual polygons already on low Pfafstetter levels.
Lakes ≥ 1000 km2 appear at level 4 (and higher); lakes ≥ 250 km2 at level 5 (and higher);
lakes ≥ 10 km2 at level 8 (and higher); and lakes ≥ 2 km2 appear at level 11 (and higher).
• If a lake covers the outlet of a sub-basin, the sub-basin is split into a left and a right part
along its main river. This is to accommodate data being assigned to the left or right side
of a lake. This modification is inconsistent with the traditional Pfafstetter scheme in
which both sides have to keep the same Pfafstetter code because they belong to the same
sub-basin. In HydroBASINS, the two sub-basin sides are distinguished by an additional
attribute (‘Side’), which can be L, R, or M (for left, right, or merged if there is no split).
The according side is also reflected in the last digit of the sub-basin ID. Topologically,
the two sides are neither up- nor downstream from each other. However, to keep
connectivity and topological searches consistent, the following rule has been introduced
in HydroBASINS: every right part of a split sub-basin has to first flow into its
corresponding left part, then into the contained lake, and finally to right (or merged) part
of the next downstream sub-basin polygon.

5
 HydroBASINS Technical Documentation v1.c

3. Data format and distribution

3.1 File name syntax

HydroBASINS data are provided as regional tiles in individual polygon shapefiles, one for each
region and each Pfafstetter level. File names follow the syntax:

Hybas_XX_levYY_v1c.shp (standard format; without lakes)

or Hybas_lake_XX_levYY_v1c.shp (customized BioFresh/IUCN format; with inserted lakes)

where XX indicates the region and YY indicates the Pfafstetter level (01-12). The regional extents
are defined by a two-digit identifier:

Identifier Region
af Africa
ar North American Arctic
as Central and South-East Asia
au Australia and Oceania
eu Europe and Middle East
gr Greenland
na North America and Caribbean
sa South America
si Siberia

3.2 Attribute table

Each HydroBASINS shapefile contains an attribute table with the following column structure
and information:

Column Description

Unique basin identifier. The code consists of 10 digits:

• First 1 digit represents the region: 1 = Africa; 2 = Europe; 3 = Siberia; 4 = Asia;
5 = Australia; 6 = South America; 7 = North America; 8 = Arctic (North
America); 9 = Greenland
• Next 2 digits define the Pfafstetter level (01-12). The value ‘00’ is used for the
‘Level 0’ layer that contains all original sub-basins and all Pfafstetter codes (at
Hybas_id all levels); ‘Level 0’ only exists in the standard format of HydroBASINS
(without lakes).
• Next 6 digits represent a unique identifier within the HydroSHEDS network;
values larger than 900,000 represent lakes and only occur in the customized
format (with lakes)
• Last 1 digit indicates the side of a sub-basin in relation to the river network (0 =
noSide; 1 = Left; 2 = Right). Sides are only defined for the customized format
(with lakes).

6
 HydroBASINS Technical Documentation v1.c

Hybas_id of the next downstream polygon. This field can be used for navigation
(up- and downstream) within the river network. The value ‘0’ indicates a polygon
with no downstream connection. Note that small endorheic sinks may have a
‘virtual’ connection to an appropriate downstream polygon to allow for topological
Next_down queries in larger river basins where discontinuities should be eliminated (e.g., the
larger Nile Basin contains smaller endorheic basins that are virtually connected to
the larger basin). Virtual connections can be identified as they carry a value of ‘2’ in
the ‘Endo’ field AND a value larger than ‘0’ in the ‘Next_down’ field. Users can
thus decide whether or not to terminate the routing at endorheic sinks.

Hybas_id of the next downstream sink. This field indicates either the ID of the next
downstream endorheic sink polygon (if there is one) or the most downstream
Next_sink
polygon of the river basin (if there is no endorheic sink in between). This field can
be used to identify the entire, fully connected watershed that a polygon belongs to.

Hybas_id of the most downstream sink, i.e. the outlet of the main river basin. This
field indicates the ID of the most downstream polygon of the river basin and can be
used to identify the entire river basin that a polygon belongs to, including all
Main_bas
associated endorheic basins. Note: small endorheic parts are typically lumped (via
virtual connections) with their corresponding larger basin, while large endorheic
watersheds can form their own basins.

Distance from polygon outlet to the next downstream sink along the river network,
Dist_sink in kilometers. This distance is measured to the next downstream endorheic sink (if
there is one) or (if there is none) to the most downstream sink (i.e. the ocean).

Distance from polygon outlet to the most downstream sink, i.e. the outlet of the
main river basin along the river network, in kilometers. The most downstream sink
or outlet is that of the larger basin (to which smaller endorheic sub-basins may be
Dist_main virtually connected), i.e. either the outlet at the ocean, or the final sink of a large
endorheic watershed which forms its own basin. Note that when small endorheic
basins are lumped with a larger river basin, the virtual linkages are not measured as
true distances but are calculated as direct (zero distance) connections.

Sub_area Area of the individual polygon (i.e. sub-basin), in square kilometers.

Total upstream area, in square kilometers, calculated from the headwaters to the
polygon location (including the polygon). The upstream area only comprises the
Up_area
directly connected watershed area, i.e. it does not include endorheic regions that
may be part of the larger basin through virtual connections.

The Pfafstetter code. For general description see literature (e.g., Verdin and Verdin
1999). The Pfafstetter code uses as many digits as the level it represents. This field
Pfaf_id
can be used to cluster or subdivide sub-basins into nested regions. This field is only
available for levels 1-12 (i.e. not for the ‘Level 0’ layer of the standard format).

Indicates the side of a sub-basin in relation to the river network: L = Left; R =

Right; M = Merged (direction defined looking downstream). This index enables a
Side distinction between the two sides along lake shorelines (see text for more
explanation). Polygons have only been split into left and right parts where lakes
exist. This field is only available in the customized format (with lakes).

7
 HydroBASINS Technical Documentation v1.c

Indicator for lake types: 0 = no Lake; 1 = Lake; 2 = Reservoir; 3 = Lagoon. This

Lake
field is only available in the customized format (with lakes).

Indicator for endorheic (inland) basins without surface flow connection to the
Endo ocean: 0 = not part of an endorheic basin; 1 = part of an endorheic basin; 2 = sink
(i.e. most downstream polygon) of an endorheic basin.

Indicator for lumped coastal basins: 0 = no; 1 = yes. Coastal basins represent
Coast conglomerates of small coastal watersheds that drain into the ocean between larger
river basins.

Indicator of river order (classical ordering system): order 1 represents the main stem
river from sink to source; order 2 represents all tributaries that flow into a 1st order
Order
river; order 3 represents all tributaries that flow into a 2nd order river; etc.; order 0 is
used for conglomerates of small coastal watersheds.

Indicator showing the record number (sequence) in which the original polygons are
stored in the shapefile (i.e. counting upwards from 1 in the original shapefile). The
Sort
original polygons are sorted from downstream to upstream. This field can be used to
sort the polygons back to their original sequence or to perform topological searches.

Pfafstetter codes for all levels (1 to 12). For general description see literature (e.g.,
Verdin and Verdin 1999). The Pfafstetter code uses as many digits as the level it
Pfaf_1 to
represents. These fields can be used to create sub-basins at all Pfafstetter levels by
Pfaf_12
dissolving the polygons accordingly. These fields are only available for the ‘Level
0’ layer of the standard format (without lakes).

3.3 Vector data format and projection

The polygon data sets of HydroBASINS are distributed in ESRI ‘shapefile’ format (ESRI 1998).
Each HydroBASINS shapefile consists of five main files (.dbf, .sbn, .sbx, .shp, .shx).
Additionally, basic metadata information is provided in XML format (.xml). Projection
information is provided in an ASCII text file (.prj). All shapefiles are in geographic
(latitude/longitude) projection, referenced to datum WGS84.

3.4 Data distribution

HydroSHEDS data is available electronically in compressed zip file format from

http://www.hydrosheds.org. [Please note that the former data download site at the EROS Data
Center of USGS at http://hydrosheds.cr.usgs.gov is now discontinued.] To use the data files, the
zip files must first be decompressed. Each zip file includes a copy of the HydroBASINS
Technical Documentation.

8
 HydroBASINS Technical Documentation v1.c

4. Disclaimer and acknowledgement

4.1 License agreement

HydroBASINS is covered by the same License Agreement as the HydroSHEDS database, which
is available at http://www.hydrosheds.org. HydroBASINS data (as defined in the License
Agreement) are free for non-commercial and commercial use. For all regulations regarding
license grants, copyright, redistribution restrictions, required attributions, disclaimer of warranty,
indemnification, liability, waiver of damages, and a precise definition of licensed materials,
please refer to the License Agreement.

By downloading and using the data the user agrees to the terms and conditions of the License
Agreement.

4.2 Acknowledgement and citation

We kindly ask users to cite HydroBASINS in any published material produced using the data. If
possible, online links to the HydroSHEDS website (http://www.hydrosheds.org) should be
provided.

Citations and acknowledgements of the HydroBASINS data should be made as follows:

Lehner, B., Grill G. (2013): Global river hydrography and network routing: baseline data and
new approaches to study the world’s large river systems. Hydrological Processes, 27(15):
2171–2186. Data is available at www.hydrosheds.org.

5. References
ESRI – Environmental Systems Research Institute (1998): ESRI Shapefile Technical Description - An
ESRI white paper. Available at http://www.esri.com/library/whitepapers/pdfs/shapefile.pdf
Lehner, B., Döll, P. (2004): Development and validation of a global database of lakes, reservoirs and
wetlands. Journal of Hydrology 296(1-4): 1-22.
Lehner, B., Grill G. (2013): Global river hydrography and network routing: baseline data and new
approaches to study the world’s large river systems. Hydrological Processes, 27(15): 2171–2186.
Lehner, B., Verdin, K., Jarvis, A. (2008): New global hydrography derived from spaceborne elevation
data. Eos, Transactions, AGU, 89(10): 93-94.
Stein, J.L., Hutchinson, M.F., Stein J.A. (2014): A new stream and nested catchment framework for
Australia. Hydrology and Earth System Science 18: 1917-1933.
USGS – U.S. Geological Survey (2000): HYDRO1k Elevation Derivative Database. USGS EROS Data
Center, Sioux Falls, SD.
Verdin, K.L., Verdin, J.P. (1999): A topological system for delineation and codification of the Earth's
river basins. Journal of Hydrology 218 (1-2): 1-12.