This page will be under continuous updates. Please email me for the related stuff which is not here (I may be able to help you) or come back for the updated version. Last updated on May 13, 1998.

• Objectives
• Methodology
• Data Acquisition
• Basin Boundaries
• DEM Sheets
• Precipitation Data
• RF1 Stream Coverage
• Water Rights
• Data Processing
• Raw Data to DEM conversion
• Size reduction of DEM Grids
• Datum and Projection Conversion
• Grid Cell Size Modification
• Processing RF1 Stream Coverage
• Extracting Water Rights for Basins
• Defining Bay Areas
• Burning-in of stream coverage on DEM
• RF1 Coverage and Water Rights
• Conclusions
• Work in Progress
• Appendix 1

The project was started with three 'initial' objectives. They were

• Investigate the use of 1:250,000 scale (3" arc) DEMs (Digital Elevation Models) for large size river basins of Texas.
• Calculate the water available in the streams and rivers during droughts after taking out water for water rights.
• Develop vectorized watersheds and streams network for use in hydrologic modeling of surface water.

But during the course of the term project presentations it became obvious that the second objective was overlapping with the work of other Luis Aburto, so the objectives were slightly modified to save time & energy. A new dimension to these objectives was added by Kevin Wei's work. The refined objectives, which evolved as a result of the work done by other students, are

• Investigate the use of 3" DEMs for large areas.
• Use of 1:100,000 DLGs (Digital Line Graphs) of stream network instead of RF1s (River-reach Files).
• Delineate watersheds and streams network, which closely match the data for water-rights locations, so HECPREPRO and WRAP can subsequently use it in an ArcView environment.

 Ç

The steps needed to complete the project were

• Data Acquisition
• Process DEMs to produce stream networks
• Develop a shape file of water rights locations in the basins
• Process Digital Line Graphs (Hydrography Layer) of 1:100,000 scale.

 Ç

Three basins of Texas were selected to minimize the amount of work needed. They are

1. San Jacinto Basin
2. Guadalupe Basin
3. Neches Basin

Neches is the largest of them all, whereas San Jacinto is the smallest in terms of area of each basin.

A graphical map showing the basins is shown above.

 The characteristics of the basins and their sizes are explained later.

 Ç

The basin boundaries were selected using the data available on the CD titled 'US Spatial Hydrology Database: prepared by CRWR, UT Austin'. For the availability of this CD contact Dr. David Maidment. Data available on this CD is in Albers Projection for US. They can also be downloaded over the Internet from USGS site.

DEMs for delineating watersheds were downloaded from USGS Eros ftp server. Thirty-four 3" DEM tiles, each 1^o x 1^o in size (1201 columns by 1201 rows of cells) were obtained for three different basins of Texas. Originally I downloaded 34 1^o x 1^o tiles for three basins. Identification of relevant tiles was by overlaying the basin boundaries over county boundaries and than by overlaying those country boundaries by cell boundaries. A graphical image of Neches Basin, the counties, which fall within the boundaries of the basin and the 1^o x 1^o cells that are needed, are shown below. As can be seen, each cell is 2^o wide and 1^o high. Thus Austin has two 1^o x 1^o cells, one as Austin East and the other one as Austin West.

 (Click on the image to see a larger version)

Table showing the DEM sheets needed for each basin.

After completing the process of building DEMs for the three basins, the next step was to get stream coverage and burn it on the DEMs.

RF1 coverage for the three basins was clipped using query builder tool from the RF1 coverage of whole of US available on the CD 'US Spatial Hydrology Database: prepared by CRWR, UT Austin'.

Rainfall data for the US is available from the Oregon State University PRIZM Project site. It was downloaded and saved as an ASCII file with an extension (.asc) and then imported into ArcView using File/Import Data Source (select ASCII Raster) and select the file US_ppt.asc. It will import the data directly into the current view. A view must be open and highlighted (top tab of the view should be deep blue in color) for this to work properly. Again rainfall data can be clipped to match the size of the DEM area using Analysis/Properties and Setting Map Extent equal to the 'Same as the DEM' of the basin. The Cell Size should be set to 'Same as the DEM'. This will help in multiplying the rainfall grid values with the DEM grid values to get the runoff grid.

Water Rights Locations for Texas were obtained from Mr. James Edmonds at TNRCC through Dr. Maidment. It was in Geographic Coordinates and was converted to the Albers Projection in a one step conversion process. Datum was also shifted from NAD 27 to NAD 83 in this one step conversion process. Click here to see the alb.prj file. Copy of the Water Rights for Texas can be had from me.

Ç

 Data Processing

For the purposes of keeping the matters getting too complicated, a separate directory for each basin should be created. Each DEM file is in two forms, compressed and uncompressed. It's better to download uncompressed files, they have a file extension of '.gz', and then uncompress them using the 'gunzip' or 'gnuzip' file compression utilities. Although the files can be downloaded using the ftp command, I used the Windows95 version of this transfer protocol, called wsftp_95. The only difference between the two is that one has a graphical interface while the other is command line based system, and obviously ftp requires UNIX system to run whereas is for the Windows based PC system.

This will expand the file size and the file extension .gz will be removed.

After uncompressing the files, the first step was to limit the size of the data blocks using UNIX command

I had to use this command for all 34 DEM cell-blocks. Until individual DEM sheets were merged together, the commands had to be done on all the files individually. This was an extremely time consuming process.

After delimiting the text the next step was to get DEMs from the files. The command used was (in ArcInfo environment).

  Arc: DEMLATTICE outputfilename dem1 usgs

The input file for this process is the output file obtained after the delimiting process.

At the end of this command a new folder was created, with the name dem1. It had all the files necessary for a grid file. Typcial files in a grid folder are dblbnd.adf, hdr.adf, prj.adf, sta.adf, w001001.adf, vat.adf, w001001x.adf. In addition to this another folder name info was created which contained data files. At the end of this conversion process a brief message tells about the DEM parameters. The text of the message for a conversion process is shown below.

Individual DEMs were obtained for each block of 1^o x 1^o cells and then they were merged together using the GRID function commands (Grid function commands are sub-command functions of ArcInfo. To start Grid, first start the ArcInfo by typing arc and then type grid at the arc prompt).

The newdem file is created after merging individual dem1, dem2 and other files obtained from the Lattice-DEM conversion.

1. Nine individual DEM sheets were merged for San Jacinto basin
2. Twelve individual DEM sheets were merged for the Guadalupe basin
3. Sixteen individual DEM sheets were merged for the Neches basin. Names are given in the table.

Ç

The size of various DEMs obtained after doing all these processes was extremely large and it was impossible to manage them with the limited amount of space available on the system. Also, any operation that had to be done on those DEMs would have taken a lot of time. This was overcome by using changing all the grid values to integers from precision number format. This action alone reduced the size of the files by a factor of 10 approximately.

Here intdem is the filename of the new integer DEM grid created by using the newdem grid from the merging of individual blocks.

The original data and DEMs obtained from USGS are all in geographic coordinates with 'decimal seconds' as units. As I was using the Albers Projection for US. So the data had to be converted to that projection. The coordinates of Albers Projection for US are:

Although conversion of projection is not a complex process, but because this original data was WGS 72 datum based hence the conversion process had to be done in two steps.

The key thing here is that, in the first step the intdem file was converted for the datum only. WGS 72 was shifted to WGS 84. Datum conversion between same types of datum is possible. The output of this datum conversion was wgs84dem. Albers projection is based on NAD 83 datum, which is equivalent to WGS 84 datum.

The output of first conversion (datum conversion only) was used as the input for next step, where the projection itself was converted. For more information one datum, projections and conversion see Exercise 3 on Map Projections.

The projection files, ds_datum.prj and albers.prj are shown in Appendix 1.

Ç

The original cell size of the grid was not exactly 100 meters. It is easier to work with 100-meter cell-size grids. In order to do this again GRID commands were used.

The new grid dem100 obtained from this process will have a cell size of 100 meters. The albdem is the one, which was obtained from the previous step of projection conversion.

The grid obtained after merging the cells was much bigger in area than needed for delineating the watersheds. In order to reduce the area of the grids itself, and only include area needed, the procedure used was simple.

1. Add the DEM theme and the HUC boundary theme of the basin to the view.
2. Make the HUC boundary theme active.
3. Press the (zoom to active theme[s]) button.
4. The view will be zoomed to the active theme(s).
5. Now make the albdem theme active by clicking on the theme name once.
6. Go to Analysis/Properties.
7. Select 'Same as display' for Analysis Extent and 'Same as dem' for the Cell Size. Click OK.
8. Go to Analysis/Map Calculator. Under Layers select albdem and press Evaluate.
9. As a result of calculations a new theme titled 'Map Calculation 1' will be generated.
10. Go to Theme/Convert to Grid to save 'Map Calculation 1' as a permanent theme.

(Click on the image to see a bigger version)

 Ç

To do so, add the RF1 theme to the view from the CD. This theme was for the whole of US. To select the coverage for the selected basin only, use the query builder tool. The strings entered for Neches, San Jacinto and Guadalupe basins respectively were:

All the streams in RF1 coverage for Neches basin were selected. As this selection process is temporary, they were saved by selecting Theme/Convert to Shape File.

This is how the RF1 clipped for the Neches basin looks like.

Ç

Water Rights in Texas were supplied in the form of a shape file. The data was in decimal degrees units and geographic coordinates. This required quite some manipulation to extract the data. The data was in the zip format and after unzipping the files, the following steps were performed to develop point shape files within each basin. They can be obtained from me in Albers Coordinates as it will save a lot of time.

• Using Table/Properties, clicked on all the columns to make them invisible except for WR_ID, Longitude and Latitude.
• The existing format had Latitude column followed by Longitude column. By clicking and dragging the order was changed.
• Exported the table using File/Export/Delimited Text.
• In Excel added zeros to all those Water Rights, which had no values for Latitude and Longitude.
• Saved the file as Tab Delimited File (wr.dat).
• In ArcInfo generated a point coverage using the following commands

1. Arc: generate wrloc
2. Generate: input wr.dat
3. Generate: points
4. Generate: quit
5. Arc: build wrloc points
6. Arc: addxy wrloc
7. Arc: list wrloc.pat ( this will list the generated wrloc file )

• After the wrloc point coverage had been generated. Converted the projection of this coverage to Albers projection using command

• After changing the projection of the water rights coverage, I manually highlighted the ones within the basin boundaries and saved the highlighted ones as a shapefile using Theme/Convert to Shapefile command. The three water rights shape files are available and can be had from me.

This process was needed because, the original water rights coverage was a shape file with quite a few water rights having no values for latitude, longitude or water right ID. To overcome this, the Data was exported to MS Excel and than after processing it by adding values to the cells using a logical function (If B1<>0, B1, 0). This was necessitated because ArcInfo generate function for point coverage generation doesn't accept a space in the data file for a zero.

The water rights shape (point) coverage for individual basins look like this for Neches.

For San Jacinto Basin, their were 252 Water Rights. This is the Water Rights shape file for the San Jacinto Basin.

Guadalupe Basin had 540 Water Rights. This is how the Water Rights shape file for Gaudalupe Basin looked like.

 Ç

Ç

1. Added the RF1 stream coverage [sjrf1.shp] for San Jacinto Basin into the view. Converted the theme into a grid theme by using the command Theme/Convert to Grid. Selected the HUC6 value for cell values. (Any field with non-zero numbers could have been selected for this purpose). Selected Same as SJDEM100 option for the cell size. Named the theme gridstrm. Feature Attributes from the sjrf1.shp were not joined to the new theme.
2. Divided the gridstrm by itself using Analysis/Map Calculator. The logical expression entered was [gridstrm]/[ridstrm]. The result was added to the view as Map Calculation 1. Saved this theme permanently by using the command Analysis/Concert to Grid and assigning name unitstrm. This theme had either one or nodata values. All those places, which had zero value, had nodata assigned after the division of grid by itself.
3. Multiplied the SJDEM100 to unitstrm theme using Analysis/Map Calculator, by entering the logical function [unitstrm] * [SJDEM100]. This resulted in the creation of another Grid, with all the stream cells having a value equivalent to the DEM cell elevation value. Saved this DEM as DEMstrm.
4. Added a value of 2000m to the SJDEM100 Grid, using Analysis/Map Calculator. ([SJDEM100 + 2000]).
5. Saved the above theme as elevDEM. Here is how it looks like.



6. Used the script from Exercise 6 of spring 97 to merge the two themes elevDEM and DEMstrm. This script uses two themes, aGRID and bGRID and they are merged in the order of aGRID onto bGRID. Thus the values of DEMstrm are added to the elevDEM and they are 2000 meter depressed into the surface, thus forcing the water flow direction grid into the RF1 coverage of streams.



7. Next steps were to fill sinks, generate Flow Direction, Flow Accumulation Grids. For the purposes of Stream Definition I chose the value of 10,000 as it gave a drainage area of 100 square kms for the definition purposes of a stream. Links Grid, Outlets Grid were generated.

• Water rights near the bay area are practically inside the bay. This could be due to the fact, that bay area for San Jacinto basin was defined using the INTEGER command of in GRID. This command gave a bay boundary based on all elevation greater than or equal to 1-meter in surface elevation above MSL. So any area with a surface elevation of 0.4 or 0.3 or so meters must have been declared as bay areas. This problem is shown in the figure below.

• Other problems, which have resulted from this type of bay definition, are a thin strip of water coming into the land area from the sea. This is shown below.

• The vectorized stream coverage generated has some streams near the bay area that are small in length. This may have been the result of bay area definition using GRID Integer command.

1. I have already downloaded the DLGs for hydrographic data for 16 of 1:100,000 resolution blocks. The purpose is have a closer match between the Water Rights and the Streams on which they exist. RF1 coverage's spatial resolution of data was not good enough to provide a good stream coverage. The list of files which I had to download, process (unzip, delimit, arc to dlg conversion) is given below. For each of the file, these three processes had to be performed. For the Digital Line Graphs, each 1degree block is divided into eight 15' x 15' blocks which are numbered from 1 to 8. The numbering system is such that block 01 will be on top left. Block 08 is on bottom right. This shown below.

01	02	03	04
05	06	07	08

 

 

List of files

Sixty four files were downloaded and processed for San Jacinto Basin. Each file is a square box of 15 minutes by 15 minutes in size. They come in two different types of formats. One is called the optional format and the other is standard format. The scale of the themes is 1:100,000. A higher resolution map having a scale of 1:24,000 is also available.

Name of the File	Size	Name of the File	Size
agdlg1	176610	ecdlg3	199277
agdlg2	227788	ecdlg4	141986
agdlg3	159888	ecdlg7	137688
agdlg4	197953	ecdlg8	187463
agdlg5	187524	eldlg3	167195
agdlg6	195097	eldlg4	163671
agdlg7	326704	eldlg7	191456
agdlg8	393451	eldlg8	147252
ahdlg1	305993	gvdlg1	281741
ahdlg2	220027	gvdlg2	64927
ahdlg5	88299	gvdlg5	79093
ahdlg6	283173	gvdlg6	2841
bhdlg3	199927	hodlg1	281886
bhdlg4	203387	hodlg2	233548
bhdlg7	188683	hodlg3	146280
bhdlg8	216092	hodlg4	311365
bndlg3	252587	hodlg5	178754
bndlg4	353235	hodlg6	250965
bndlg7	350914	hodlg7	156230
bndlg8	392179	hodlg8	338766
budlg1	322662	hvdlg1	357291
budlg2	130794	hvdlg2	341593
budlg5	273218	hvdlg3	403634
budlg6	182069	hvdlg4	292081
cndlg1	230440	hvdlg5	307061
cndlg2	294833	hvdlg6	298516
cndlg3	205827	hvdlg7	267578
cndlg4	177354	hvdlg8	249496
cndlg5	181562	lsdlg1	226869
cndlg6	240833	lsdlg2	210187
cndlg7	223696	lsdlg5	209457
cndlg8	236749	lsdlg6	160898

Sub-script descriptions are as follows

Bryan E	'bn'	Huntsville E/W	'hv'	Livingston W	'lv'
Brenham E	'bh'	Conroe E/W	'cn'	Beaumont W	'bu'
Eagle Lake E	'el'	Houston E/W	'ho'	Anahuac W	'ah'
El-campo E	'ec'	Angleton E/W	'ag'	Galveston W	'gv'

 

2. For each of these files, a number of steps were performed to get the data into an arc format.

My homepage