Dataset Selection: C01

Authors:

Dr. David Maidment
Center for Research in Water Resources, The University of Texas at Austin

Summary:

Primary Reference:

Five-Minute, 1/2º, and 1º Data Sets of Continental Watersheds and River Networks for Use in Regional and Global Hydrologic and Climate System Modeling Studies

• Integrated data set
• Design
• Source
• Additioinal references
• File lists

Data Set Description

INTEGRATED DATA SET

Data­Set Citation:

Projection:

Spatial Representation:

(2) Tabular values of runoff data.

Temporal Representation:

Data Representation:

Mixed numeric and ASCII tabular runoff data.

Layers and Attributes:

Runoff data with associated geographic and political reference information.

Data Set Description

DESIGN

Variables:

Origin:

Tabular runoff data based on Perry et al. [1996] and UNESCO [1974].

Geographic Reference:

Geographic Coverage:

Geographic Sampling:

Time Period:

Temporal Sampling:

Data Set Description

SOURCE

Source Data Citation:

Gorny, A. J., and R. Carter, World Data Bank II General User's Guide, Central Intelligence Agency, Washington, D.C., 1987.

Row, L. W., D. A. Hastings, and P. K. Dunbar, TerrainBase Worldwide Digital Terrain Data Documentation Manual, National Geophysical Data Center, Boulder, Colo., 1995.

Contributor:

Distributor:

Date of Production:

Lineage & Contacts:

1. Data Production:

Dr. Jay Famiglietti
Department of Geological Sciences
The University of Texas at Austin
Mail Code C1100
Austin, TX 78712, USA
2. Data Integration:

John J. Kineman and Joshua Klaus
NOAA National Geophysical Data Center
325 S. Broadway, E/GC1
Boulder, CO 80303 USA
fax: (303) 497-6513
Email: jkineman@ngdc.noaa.gov
Web: http://www.ngdc.noaa.gov/seg/eco.

Data Set Description

ADDITIONAL REFERENCES

Bryan, F. O., B. G. Kauffman, W. G. Large, and P. R. Gent, 1996: The NCAR CSM Flux Coupler. NCAR Technical Note NCAR/TN-424+STR, National Center for Atmospheric Research, Boulder, Colorado.

Coe, M. T., Simulating continental surface waters: an application to Holocene Northern Africa, J. Clim., 10(7), 1680-1689, 1997.

Environmental Systems Research Institute (ESRI), Inc., ARC/INFO Version 7.1.2, 1997.

Gorny, A. J., and R. Carter, World Data Bank II General User's Guide, Central Intelligence Agency, Washington, D.C., 1987.

Hornberger, G. M., Data and analysis note: A new type of article for Water Resources Research, Water Resour. Res., 30(12), 3241-3242, 1994.

Kite, G. W., A. Dalton, and K. Dion, Simulation of streamflow in a macroscale watershed using general circulation model data, Water Resour. Res., 30(5), 1547-1559, 1994.

Liston, G. E., Y. C. Sud, and E. F. Wood, Evaluating GCM land surface hydrology parameterizations by computing river discharges using a runoff model: application to the Mississippi basin, J. Appl. Meteorol., 33, 394-405, 1994.

Miller, J. R. , G. L. Russell, and G. Caliri, Continental-scale river flow in climate models, J. Clim., 7, 914-928, 1994.

O'Donnell, G., B. Nijssen, and D. P. Lettenmaier, A simple algorithm for generating streamflow networks for grid-based, macroscale hydrological models, accepted Hydrol. Processes.

Oki, T., and Y. C. Sud, Design of Total Runoff Integrating Pathways (TRIP); A global river channel network, Earth Interactions, 2, EI013, 1998.

Perry, G. D., P. B. Duffy, and N. L. Miller, An extended data set of river discharges for validation of general circulation models, J. Geophys. Res., 101(d16), 21,339-21,349, 1996.

Row, L. W., D. A. Hastings, and P. K. Dunbar, TerrainBase Worldwide Digital Terrain Data Documentation Manual, National Geophysical Data Center, Boulder, Colo., 1995.

Russell, G. L., and J. R. Miller, Global river runoff calculated from a global atmospheric general circulation model, J. Hydrol., 117, 241-254, 1990.

Sausen, R., S. Schubert, and L. Dumenil, A model of river runoff for use in coupled atmosphere-ocean models, J. Hydrol., 155, 337-352, 1994.

United Nations Educational Scientific and Cultural Organization (UNESCO), Discharge of selected rivers of the world, vol. I, II, III (parts I, II, III, IV), UNESCO, Paris, France, 1985.

U.S. Geological Survey, GTOPO30, http://edcwww.cr.usgs.gov/landdaac/gtopo30/gtopo30.html, Earth Resource Observation System Data Center, Sioux Falls, S.D., 1996.

U.S. Geological Survey, HYDRO1k, http://edcwww.cr.usgs.gov/landdaac/gtopo30/hydro/, Earth Resource Observation System Data Center, Sioux Falls, S.D., 1998.

Vörösmarty, C. J., B. Moore III, A. L. Grace, M. P. Gildea, J. M. Melillo, B. J. Peterson, E. B. Rastetter, and P. A. Steudler, Continental scale models of water balance and fluvial transport: an application to South America, Global Biogeochem. Cycles, 3(3), 241-265, 1989.

Wessel, P., and W. H. F. Smith, The Generic Mapping Tools (GMT) Version 3.0 Technical reference and cookbook, SOEST/NOAA, 1995.
 

Data Set Description

FILE LISTS

• Data Set Files
• Reprint Files
• Source Files (examples)

Five-Minute, 1/2º, and 1º Data Sets of Continental Watersheds and River Networks for Use in Regional and Global Hydrologic and Climate System Modeling Studies

Land/Sea Mask

Description:

Structure:

Series:

System Files:

Notes:

1. Complete legend included in datadesc.txt
2. See technical report for information on file manipulation
3. Mask manually corrected for some interior points with elevations below 0 meters
4. Due to the large size of the *.e00 files they will not be available for download
5. The ASCIIGRID command will convert the text file to a grid for ARC/INFO accessibility

Flow Direction Data

Description:

Structure:

Series:

System Files:

Notes:

1. Complete legend included in datadesc.txt
2. See technical report for information on file manipulation
3. Flow direction calculated based on direction of steepest descent
4. Flow direction data altered in Antarctica to ensure valid flow direction information. Flow direction set to north if not originally to north , northwest, or northeast in Antarctica only
5. Due to the large size of the *.e00 files they will not be available for download
6. The ASCIIGRID command will convert the text file to a grid for ARC/INFO accessibility
7. The 5-minute vector files will not be available for download

Flow Accumulation Data

Description:

Structure:

Series:

System Files:

Notes:

1. Complete legend included in datadesc.txt
2. See technical report for information on file manipulation
3. Flow accumulation calculated as the number of grid cells that are drained through a cell
4. Due to the large size of the *.e00 files they will not be available for download
5. The ASCIIGRID command will convert the text file to a grid for ARC/INFO accessibility
6. The 5 minute vector files will not be available for download

Rivers Delineation

Description:

Structure:

Series:

System Files:

Notes:

1. Complete legend included in datadesc.txt
2. See technical report for information on file manipulation
3. Rivers digitized from CIA World Data Bank II
4. Rivers can be redelineated at any threshold desired using flow accumulation information
5. Due to the large size of the *.e00 files they will not be available for download
6. The ASCIIGRID command will convert the text file to a grid for ARC/INFO accessibility

55 Large Watersheds Delineation

Description:

Structure:

Series:

System Files:

Notes:

1. Complete legend included in datadesc.txt
2. See technical report for information on file manipulation
3. Watersheds selected based on number of grid cells with additional watersheds included in Russell and Miller [1990]
4. Due to the large size of the *.e00 files they will not be available for download
5. The ASCIIGRID command will convert the text file to a grid for ARC/INFO accessibility

Internally Draining Regions

Description:

Structure:

Series:

System Files:

Notes:

1. Complete legend included in datadesc.txt
2. See technical report for information on file manipulation
3. Only the large internal drainage regions of Northern Africa, Australia, and Central Eurasia were delineated
4. To be used as a mask for other studies
5. Due to the large size of the *.e00 files they will not be available for download
6. The ASCIIGRID command will convert the text file to a grid for ARC/INFO accessibility

19 Large-Scale Drainage Regions

Description:

Structure:

Series:

System Files:

Notes:

1. Complete legend included in datadesc.txt
2. See technical report for information on file manipulation
3. Based on land area - water body pairings for NCAR CSM project
4. Due to the large size of the *.e00 files they will not be available for download
5. The ASCIIGRID command will convert the text file to a grid for ARC/INFO accessibility

19 Large-Scale Drainage Regions Including Water Bodies

Description:

Structure:

Series:

System Files:

Notes:

1. Complete legend included in datadesc.txt
2. See technical report for information on file manipulation
3. Extension of the 19 drainage regions accomplished by nearest existing value assignment with artificial divisions imposed as can be seen at straight line contacts
4. Due to the large size of the *.e00 files they will not be available for download
5. The ASCIIGRID command will convert the text file to a grid for ARC/INFO accessibility

Lakes Delineation

Description:

Structure:

Series:

System Files:

Notes:

1. Complete legend included in datadesc.txt
2. See technical report for information on file manipulation
3. Lakes gridded at 5-minute resolution from CIA World Data Bank II as 0% or 100% coverage
4. 1/2-degree and 1-degree resolution files taken as the percentage of the coarser grid cell covered by lakes
5. Due to the large size of the *.e00 files they will not be available for download
6. The ASCIIGRID command will convert the text file to a grid for ARC/INFO accessibility

Runoff Data

Description:

Structure:

Series:

System Files:

Notes:

1. Data taken from Perry et al. [1996], and UNESCO [1985]
2. Headers included in file

Five-Minute, 1/2º, and 1º Data Sets of Continental Watersheds and River Networks for Use in Regional and Global Hydrologic and Climate System Modeling Studies

• Investigator's Report
• Data Integration Report

Investigator's Report

Stephen Graham

1. Introduction

In this Technical Report we describe the methods for the production and derivation of the data sets associated with Graham, S.T., J. S. Famiglietti, and D.R. Maidment, Five-Minute, 1/2º, and 1º Data Sets of Continental Watersheds and River Networks for Use in Regional and Global Hydrologic and Climate System Modeling Studies, 1999.

All of the analysis completed in this research was conducted using the ARC/INFO Version 7.1.2 software and the associated GRID package. The source data for the geographical analysis was taken from the National Geophysical Data Center TerrainBase Global DTM Version 1.0 [Rowet al., 1995] and the CIA World Data Bank II [Gorny and Carter, 1987]. Supplementary annual average streamflow data was taken from Perryet al. [1996] and UNESCO [1974], and is included for validation purposes.

The TerrainBase DTM first was converted into an ARC/INFO grid format from its original image format, so that the individual pixels could be accessed for analysis. This was accomplished by using the command IMAGEGRID. Because IMAGEGRID does not support conversions of images formatted as signed integers, an additional step had to be taken to complete the conversion to elevation values. This conversion was accomplished by issuing the following command:

out_grid = con ( in_grid >= 32768 , in_grid - 65536 , in_grid )

where in_grid is the original converted DEM and out_grid is the new corrected elevation DEM with the negative elevation values properly represented.

The rivers and water bodies from the CIA data base also had to be converted into an ARC/INFO grid format. The delineation of rivers and water bodies was derived from the CIA data distributed with The Generic Mapping Tools (GMT) version 3.0 [Wessel and Smith, 1995]. An associated GMT program, named SHOREDUMP, was used to convert the rivers and water bodies to an ASCII file of line segments. This ASCII file was then used as input to the ARC/INFO LINES command to generate a coverage of the hydrological information. The coverage was then converted to gridded data at a 5-minute resolution using the LINEGRID command so that these data could be used in the grid based geographical analysis. Some manual correction was necessary at this point to ensure rivers were connected with coastal outlets, and to separate rivers that could not be represented at 5-minute resolution as separate entities, due to their close proximity.

Lakes were then extracted and gridded in the same way rivers were from the CIA data. These lakes were gridded at 5-minute resolution as a 0 or 1, effectively producing a 0% or 100% areal coverage of inland water for each 5-minute grid cell.

2.2.1) Determination of land/sea mask.

The first step in the data analysis was the generation of a land/sea mask. This was accomplished by selecting the DEM grid cells having elevation values greater than 0 meters and assigning them as 'land' grid cells designated with a value of 1. Those grid cells that were designated as 'sea' grid cells were assigned a 'nodata' value so that they would not be included in the analysis of land hydrological parameters. After this initial land/sea mask selection, some manual correction was carried out to include areas which had elevation values less than or equal to 0, but which were still desirable for inclusion as land grid cells for analysis. Examples of such corrected points are inland areas near coasts which may be below sea level, but which should not be considered as points for ultimate water accumulation. The main purpose for the creation of the land/sea mask was to define the analysis area, and to designate 'sea' points as points for water to accumulate, and to be passed to coupled ocean models.

2.2.2) Geolocation of rivers.

Once the river delineation had been gridded in ARC/INFO the rivers were then 'burned in' to the DEM. This process entails reducing the elevation values in grid cells that correspond to an existing river network grid. The elevations of the river grid cells are reduced by an appropriately large number in order to improve the resulting automated river delineation. This elevation difference increases the probability that water will flow toward and along the predetermined river channels by increasing the slope toward the river channels in their immediate vicinity. This process also reopens narrow river paths that may be averaged out in the DEM.

2.2.3) Filling of artificial depressions.

The next step in this analysis, is the filling of inland depressions. This process is required for water to flow continuously across the land surface and to water bodies where it can accumulate and be passed to other models. At this point internal sinks can also be identified as locations for water accumulation, and which should not therefore be filled. The depressions are identified and iteratively filled up to their lowest outlet points, until all such depressions are eliminated. The result of this process is a DEM that has no internal sinks, except those desired by the user.

2.2.4) Calculation of flow directions.

Once the DEM has been filled so that closed depressions are removed, the direction of flow at every grid cell can be determined. Flow directions are derived by calculating the direction of steepest downward slope between each cell and its 8 neighboring cells. ARC/INFO assigns a number to each direction as a power of 2 corresponding to the following scheme: 1 for E, 2 for SE, 4 for S, 8 for SW, 16 for W, 32 for NW, 64 for N, and 128 for NE.

2.2.5) Calculation of flow accumulations.

Once the flow direction information has been derived from the filled DEM, flow accumulation data can then be derived. Flow accumulation data describe the number of grid cells whose surface water would flow through the current cell, and is also known as the upslope drainage area. Flow accumulation is determined by using the flow direction information to determine all cells that are upstream of the current cell.

2.2.6) Selection and delineation of rivers.

A delineation of DEM defined rivers was then created from the flow accumulation data. A grid cell was identified as a river grid cell if it had a flow accumulation value greater than or equal to a specified threshold value. Any threshold value may be used, and should be altered as is appropriate for each specific case or use. Areas can also be excluded by deselecting grid cells that do meet the criteria, but are undesirable in selected regions.

2.2.7) Selection and delineation of watersheds.

Watersheds were delineated from the river grid cells identified above. The WATERSHED function was used to identify all cells that drained to each set of river, or source cells, based on the flow direction data. Watersheds were also identified for internally draining regions by selecting closed depressions as the source cells for the WATERSHED function.

The 19 large-scale drainage regions were also defined using the WATERSHED function, however in this case the source cells were strips of coastal cells. All coastal cells were identified and divided into groups as to the boundary between pairings of land masses with water bodies.

These 19 drainage regions were extended to those cells not selected for analysis as land cells by using the EUCALLOCATION function. Some artificial divisions were imposed between the 19 basins in oceanic areas where common modeling or other simple divisions existed for ease of use.

Having completed the analyses at 5-minute resolution, the DEM was then averaged to 1/2 degree and 1-degree resolution by taking the average of the elevation values of all grid cells to be included in each of the coarser grid cells. The same analyses described above were undertaken at these coarser resolutions.

In the case of the lakes data however, a different method was used. The high resolution lakes data were summed for each coarser grid cell and included as a percent of total grid cell areal coverage.

Gorny, A. J., and R. Carter, World Data Bank II General User's Guide, Central Intelligence Agency, Washington, D.C., 1987.

Graham, S.T., J. S. Famiglietti, and D.R. Maidment, Five-Minute, 1/2º, and 1º Data Sets of Continental Watersheds and River Networks for Use in Regional and Global Hydrologic and Climate System Modeling Studies, Water Resources Research, 35(2), 583-587, 1999

Perry, G. D., P. B. Duffy, and N. L. Miller, An extended data set of river discharges for validation of general circulation models, J. Geophys. Res., 101(d16), 21,339-21,349, 1996.

Row, L. W., D. A. Hastings, and P. K. Dunbar, TerrainBase Worldwide Digital Terrain Data Documentation Manual, National Geophysical Data Center, Boulder, Colo., 1995.

United Nations Educational Scientific and Cultural Organization (UNESCO), Discharge of selected rivers of the world, vol. I, II, III (parts I, II, III, IV), UNESCO, Paris, France, 1985.

Wessel, P., and W. H. F. Smith, The Generic Mapping Tools (GMT) Version 3.0 Technical reference and cookbook, SOEST/NOAA, 1995.

Joshua Klaus

Data Integration Report:

File Manipulation between ArcInfo, ArcView and Idrisi

Since there are various GIS platforms for analyzing the data, we have provided three distinct file structures upon which the user can work with: Arc/Info, ArcView, and Idrisi 2.0 for windows.  In order for the user to have the flexibility to use one of these platforms we have allowed the accessibility to all these file structures: Arc/Info export files, ArcView shapefiles and Idrisi image files.  We produced the various formats by the following procedure:

1. We used the "Import 71" option within ArcView to import the ArcInfo export file into ArcView. This created two directories within the destination directory: one being an info directory and the other named the same as the filename designation.
2. We created a New project with a New View in ArcView 3.1.
3. We added a Theme to the Project by switching the Data Source Type to Grid Data Source and selecting the correct file. This produced a View with a theme corresponding to the imported file, once the theme was selected.
4. Next we converted the theme to a Shapefile.  This created an additional theme similar to the previous with a *.shp extension.  There was also a *.dbf file created which was used later when converting from vector to raster in Idrisi.
5. Within Idrisi for windows version 2.0 we set the Environment to the ArcView working directory containing the shapefile just made. Then we selected Import/Export from the file directory and selected Software-specific formats within that directory.  Then we selected SHAPEIDR to convert the shape file to an Idrisi file. This created a converted vector file (*.vec) as well as a converted attribute database for Idrisi (*.mdb).
6. We then linked the database to the vector file by using Database Workshop found in Database Query under the Analysis menu.
7. We created a blank image through the use of the INITIAL command found under the Data Entry drop down menu.
8. We then selected "Define spatial parameters individually."  We typed in an Output image; selected integer for data type; binary for file type; and 0 for initial value.
9. We selected the appropriate number for columns and for rows as well as bounding coordinates.  Then we selected the appropriate reference system and units.  This produced a black empty image that was correctly structured (you can view the *.doc file for confirmation).
10. We selected Raster/Vector Conversion and chose POLYRAS.
11. We assigned the values from the database to the image with the ASSIGN command under the Data Entry menu.  This created our Idrisi image files.
(Please note that the file naming conventions have changed in the recently released Idrisi 32. i.e. *.doc files are now *.rdc  However there are file conversion procedures in place for converting Idrisi 2.0 files to Idrisi 32)

To correct for any possible errors when converting from a vector file to a raster image we re-imported the original *.txt files into IDRISI with the import module SSTIDRIS.  This module required us to first strip off the headers to the *.txt files.  Since the 5 minute data required programming to strip the header we utilized the resources of Idrisi Technical Support.  They were working on a module that does a direct import into Idrisi from an Arc/Info text file.  Then, the 1/2 degree, 1 degree, and 5 minute data were imported correctly into Idrisi without errors.