l_aburto@mail.utexas.edu

Contents

Introduction

The Texas Water Resources Institute (TWRI), at the Texas A&M University, developed in 1993, and updated in 1996, the Water Rights Analysis Package (WRAP). This is a computer program that simulates the management and use of water resources in a river basin, or multiple-basin region, under a priority-based allocation system. Researchers at TWRI are continuously improving and adding features to WRAP. One of the current efforts is adding to the program the ability to calculate stream flows at ungaged sites based on the flows at gaged sites. To accomplish this WRAP would need information about the watershed area (A), curve numbers (CN), and long-term mean precipitation (M) for all pertinent locations. This information is being calculated manually for the development and testing periods, but it is obvious that a faster and easier method will be necessary when large amounts of data are to be used. The clear choice is to use GIS as a means of getting this information. The scope of this project is investigating and implementing procedures to transfer the information required for this particular use from GIS to WRAP.

Background

Naturalized Flows

The purpose of this project is to help find and manage the information required to calculate naturalized flows at ungaged sites in natural streams. Naturalized or unregulated flows represent natural hydrology without the effects of reservoirs and human water use. In other words, naturalized flows are the flows that would happen in a stream if people didn't use or store part or all of its water. Naturalized flows are very important when water diversion rights need to be managed. They allow estimating the total water that would normally be available along a stream, so it is possible to determine how much water would be available at every point of the stream if water were allocated at certain points. In other words, naturalized flows permit determining the water available in the stream if water were allocated to certain users or not, thus enabling us to evaluate the effects of granting/changing/withdrawing water rights.

Stream flows are composed of two main elements: base flow, that has its origin in ground water, and surface runoff, which is the accumulation of rainfall that drains to the stream. The characteristics of a watershed that affect the base flow are the following:

• The relationship between the groundwater table and the stream channel bottom elevation
• Geologic conditions affecting the flow though the saturated zone
• Soil characteristics and soil moisture conditions affecting flow through the unsaturated zone
• Stream length and channel bed materials
• Vegetation in the watershed and along the streambanks affecting transpiration losses

The characteristics of a watershed that affect the surface runoff are the following:

• Mean precipitation in the watershed
• Watershed drainage area
• Soil type and land cover in the watershed
• Antecedent moisture condition

Although determining in a stream which portion of the flow corresponds to the base flow and which to runoff is complicated, there is seldom the need to do so. It has been found that, to a large extent, base flow can be expected to vary between locations in about the same proportion as surface runoff. Thus, flow distribution methods that deal with total flows are adequate in most cases (Wurbs 1998).

Method for Calculating Flows at Ungaged Sites

There are available several methods for distributing flows from gaged to ungaged sites, ranging from the very simple to the complex and laborious. On the simple side, the most widely used method is the distribution of flow in proportion to drainage area. In this case the streamflow per unit area of watershed is assumed constant, and the naturalized flow at the ungaged site is calculated as the naturalized flow at the gaged site multiplied by the ratio of ungaged to gaged areas. On the other extreme, there are generalized computer models of watershed hydrology that are able to compute sequences of daily or monthly streamflows for a given precipitation unit. The advantage of these systems is the accuracy of its predictions. Their major disadvantage is that they require considerable expertise, time and effort to be used effectively. In between the extremes there are methods like the Natural Resource Conservation Service (NRCS) curve number (CN) method that are relatively easy to use and yield adequate results.

The researchers at the TWRI are investigating which of the methods available is most appropriate to calculate naturalized flows to help in the management of water rights in Texas. They are taking into account the advantages and drawbacks of each method, both regarding accuracy and effort required. A preliminary evaluation suggested that an adaptation to the NRCS curve number method might provide the adequate balance between ease of use and accuracy, but to prove this true it is necessary to assure that the information required is readily available. This is where GIS comes in to play; its role will be to provide the information required to apply this method. To understand the information that GIS will have to deliver, a brief description of the adapted NRCS curve number method is given below, as given by Wurbs (1998).

The standard NRCS curve number method is based on the following relationship between rainfall depth, P in inches, and runoff depth, Q in inches:

Q = 0 if P < 0.2S

To obtain volumes, P and Q (in inches) must be multiplied by the watershed area. The potential maximum retention, S in inches, represents an upper limit of the amount of water that can be abstracted by the watershed through surface storage, infiltration, and other hydrologic abstractions. For convenience, S is expressed in terms of a curve number, CN, which is a dimensionless watershed parameter ranging from 0 to 100. A CN of 100 represents a limiting condition of a perfectly impermeable watershed with zero retention and thus all the rainfall becoming runoff. A CN of zero conceptually represents the other extreme, with the watershed abstracting all rainfall with no runoff regardless of the rainfall amount.

The watershed parameter CN can be determined from empirical information. The NRCS has developed tables of CN values as a function of the watershed soil type, land cover/use/condition, and an antecedent moisture condition. A few examples are presented below.

The hydrologic soil groups refer to the standard NRCS soil classification procedures, where classification A refers to sand and aggregated silts with high infiltration rates, and goes to classification D, that corresponds to soils that swell significantly when wet and have low infiltration rates.

For a watershed with subareas of differnt soil types and land cover, a composite CN is determined by weighting the CN's for the different subareas in proportion to the land area associated with each.

Composite CN = CN₁(A₁/A_total) + CN₂(A₂/A_total) + ... + CN_n(A_n/A_total)

Although the method was developed to determine runoff from single storm events, it might be also appropriate to approximate monthly values. Observations of gaged data indicate that the runoff volume associated with a particular precipitation depth tends to vary greatly between storm events. The CN method estimates the mean runoff associated with a particular precipitation depth and may be significantly in error for a particular rainfall event. However, Goulding (1997) notes that the fit of measured data to the CN relationship improves with aggregation, such that estimating monthly runoff from monthly rainfall has less scatter than for daily values.

The procedure for distributing naturalized flow outlined below is an adaptation of the CN relationship; it distributes all of the flow, including base flow, in the same proportion as runoff. The required data consists of monthly naturalized flows at the gaging station and drainage areas (A) and watershed curve numbers (CN) for both the gage location and the ungaged site. Additionally, the long-term mean precipitation may be input for both the watershed and subwatershed to make a precipitation adjustment, as outlines in step 3 below. The following computations are performed for each month.

Step 1

The flow at the gage, in acre-feet/month, is divided by the drainage area A_gage and multiplied by a unit conversion factor to convert to an equivalent depth Q_gage in inches.

Step 2

Q_gage is input to the curve number equation (above) to obtain P_gage in inches through an iterative method. If necessary, a composite CN must be used. This approximation for precipitation depth is assumed to be applicable to the ungaged subwatershed as well as the gaged watershed.

Step 3

If the long-term mean precipitation varies between the watershed and subwatershed, the precipitation depth may be adjusted by multiplying P_gage by the ratio of the long-term mean precipitation depth of the watershed to that of the watershed to obtain a P_ungaged adjusted in proportion to the mean precipitation:

where M_ungaged and M_gage are the mean precipitation for the ungaged subwatershed and gaged watershed. Otherwise, P_ungaged is assumed equal to P_gage.

Step 4

P_ungaged is input to the CN equation to obtain Q_ungaged in inches. Q_ungaged in inches is multiplied by A_ungaged and a unit conversion factor to convert to flow in acre-feet/month.

The information required to apply this method is thus the gaged flow, the watershed areas, curve numbers, and mean precipitation. The estimation of these parameters can be facilitated by the use of GIS, which is the purpose of this project.

Sources of GIS Information

The first type of information required, gaged flows, is available through the Texas Natural Resource Conservation Commission (TNRCC), which has a database management system called the Interactive Water Rights System. This database contains information about existing water rights, including latitudes and longitudes. The system also has a database of naturalized and unappropriated flows for eight river basins for which models had been developed in the past.

Watershed areas can be derived from topographical information of the basins of interest. Topographical information of the whole country can be found as Digital Elevation Models (DEM) that were developed by the United States Geological Survey (USGS). DEMs are grids representing elevation of the terrain, and can are available at several scales and levels of detail.

Curve numbers can be derived from the Land Use/Land Cover (LULC) data developed by the United States Environment Protection Agency (USEPA) and the soils characteristics included in the STATSGO database developed by the U.S. Department of Agriculture (USDA).

Fortunately, the Center for Research in Water Resources (CRWR) of the University of Texas at Austin has already calculated the curve numbers for all the continental US, and is available in the Hydrologic Spatial Data CD-ROM. This CD also contains the complete DEM for all the continental US, so it will be easier to access this information from this source.

Finally, information regarding precipitation is available from the Oregon State University, from a project called Parameter-elevation Regressions on Independent Slopes Model (PRISM), and can be downloaded online.

Methodology

The steps required for finding the parameters of the Curve Number method for any location of interest, which in this case are the locations of Water Rights in Texas, are described below. In the following directions it is assumed that the user has some experience using ArcView; if in doubt about some of the procedures mentioned please refer to the ArcView User's Manual. Another excellent source to learn how to use ArcView is the set of online exercises prepared for the GIS in Water Resources class by Dr. David Maidment, at the University of Texas at Austin.

1. Preparing the Basic Information

As mentioned, the data required to determine the parameters for an area of interest are (1) a digital elevation model (DEM), (2) a grid of the Curve Number (CN), and (3) a grid of the mean annual precipitation (PPT). Additionally, although not required, coverages of the main river basins (HUCs) and main streams (RF1) are helpful to narrow down the extent of the data to use in the calculations.

As part of this project, all this information was compiled for the whole state of Texas and is included in a compressed format in the Texasdata.zip file. To use the information, the file Texasdata.zip must be decompressed using an utility such as Winzip or Pkunzip, and making sure that the directory structure is reconstructed. For a description of the files included in Texasdata.zip see the Data Dictionary.

Once the information is decompressed, the first step is to open the project that links all the source information. To do this, execute File/Open Project... from the menu, and select the Texasdata.apr project. Since this is the source data and most probably it will be used several times for different basins, it is a good idea to save this project under a different name. To do this, select the Project Window, and execute File/Save Project As... from the menu; then give the project a new name, SNJACIN.APR for example. You should have on your screen

If you wish to view the contents of the Themes included in the project, you can activate them using the button close the the Theme's name. Your screen should look like Figure 1.

Figure 1

2. Selecting Only Relevant Data

The next step is to separate the data that is required for the analysis at hand from the rest of the data. This is necessary to make the computations shorter; if we use all the data for Texas to compute the parameters of one location we are evaluating thousands of cells that are not related to the location of interest.

One way to separate the relevant data from the rest is to first find out where are the point that need to be evaluated. Inserting a theme containing the location of the water rights of interest could do this. In this case, a file has been included with the data set that contains the locations of 10 water rights in the San Jacinto basin. Add this theme to the view using the button; chose the w-rights.shp theme. You can make a zoom to view the Water Rights (WR) using the button. To display the ID number of the WRs make sure that w-rights.shp is the active theme and execute Theme/Auto-Label from the menu; select Albstat_ in the Label Field. If the text is not of an appropriate size you can use the Symbol Window to change the font and size.

Next, make the Texashuc.shp theme active and use Theme/Select by Theme... command. Indicate that the method is to intercept with w-rights.shp. This will select the two HUCs that containt the WRs. Next execute Theme/Convert to Shapefile... to create a new theme that consists only of the two selected HUCs; give a name to the file, for example sanjacinto1.shp. Next convert the new theme to a grid. Before doing that you must indicate that the analysis extent should be the same as the extent of sanjacinto1.shp in the dialog box of Analysis/Properties. Now use the Theme/Convert to Grid command; give the grid a name, sanjacinto2.shp for example, do not join an attribute table, but do include the new theme in the view.

Now select the Texasrf1.shp theme and use Theme/Select by Theme... command. Indicate that the method is to intercept with sanjacinto1.shp. Next execute Theme/Convert to Shapefile... to create a new theme that consists only of the selected river reaches, give a name to the file, for example sanjacinto3.shp, and add the theme to the view.

Next we need to create a grid with a value of 1 in each cell so that multiplying this grid for the grids of the DEM, CN, and PPT will result in grids of these properties the size of the HUCs selected. Execute Analysis/Map Calculator to divide sanjacinto2.shp grid by itself. The Map Calculator window should look like Figure 2. Hit the Evaluate button and close the Map Calculator.

Figure 2

You can now deactivate the Texashuc.shp, Texasrf1.shp, sanjacinto1.shp and sanjacinto2.shp themes, or even delete them from the project with the Edit/Delete Themes command. Rearrange (take to the top) the remaining themes to be able to see the river reaches and the WRs. Your screen should look something like Figure 3.

  Figure 3

Now multiply the newly created theme, Map Calculation 1, by the TexasDEM, TexasCN, and TexasPPT themes. Remember, you must multiply the [Map Calculation 1] field by the [TexasDEM], [TexasCN], and [TexasPPT] fields, one at a time. Use the Theme/Properties command to rename the resulting new grids to DEM, Curve Number, and Precipitation. You must use exactly these names. You can now deactivate or delete the Map Calculation 1, TexasDEM, TexasCN, and TexasPPT themes. If the legends of the themes are too long you can hide them using the Theme/Hide/Show Legend command. If you plan to use this set of data in the future you can save the themes to files of your choice using the Theme/Save Data Set command on each theme. Rearrange the themes to make the WRs and river reaches visible.

We now have all the basic information for the area of interest.

3. Develop flow direction and flow accumulation grids

To be able to delineate the draining areas of the WRs we need to define the Flow Direction and Flow Accumulation grids for our data. To do this, select the DEM theme and use the Hydro/Fill command to fill the sinks that might exist in it. Now you must have a Filled DEM theme. Select it and use the Hydro/Flow Direction command; this will create a Flow Direction theme. Now select the Flow Direction theme and use the Hydro/Flow Accumulation command; this will create a Flow Accumulation theme.

4. Load the script to calculate the parameters

In the Project Window select the scripts icon and press the New button. This will create the Script1 window. From the menu select Script/Load Text File... and load the script included in the data set, cn-ppt.ave. Next compile the script using the button. To avoid having to switch between the view and the script every time the script is going to be used let us create a tool button to access the script. Activate the view and double click in a blank part of the toolbar. This will bring up the Customize dialog. In the Category field select Tools; next press the Tool button. A blank button will appear among the other tools; drag it to a location of your choice. The go to the table in the lower part and double click on Apply and select Script1, then double click on Icon and select an icon of your liking. Close the Customize window using the button. You are now ready to use the script.

5. Use the script to find the parameters

What the script does is to delineate the draining area of a point defined by the user with the mouse. Once the area is defined, a process similar to the one followed to separate the data of interest from all the data of Texas is used to find the CN and PPT grids for that draining area. The script then makes a weighed average of those grids to find the mean CN and PPT. The results are added to a table in Dbase format, which later can be joined to other tables or manipulated in database or spreadsheet programs. The script first looks for the output table in the project, and if it is not present it creates one. However, even if the table exists on disk, but if the table is not present in the project, it will be erased and overwritten; so, it is important to load the table into the project as soon as it is created. Below are the steps that need to be followed.

First, use the button to make a zoom to the first point of interest. In this case lets assume it is the WR 3935. Your screen should look something like Figure 4. Select the Filled DEM theme and press the tool button to which you assigned the script.

Figure 4

With the mouse, select (click on) the cell on which the WR 3935 is located. You have to make sure that you select a cell where the flow accumulation shows that there is a stream; otherwise the draining area might not be delimited correctly. You will be asked for the WR ID number, enter 3935. If you want to see the grids of the draining area, CN and PPT for this WR say Yes when asked if the grids should be included in the view. Figure 5 shows the view including the CN grid for the draining are of WR 3935.

Figure 5

The next step is to bring into ArcView the table where parameters of the WRs are going to be stored. This table by default is called wr-param.dbf. Go to the Project Window, select the Tables icon and press the Add button; then search for the wr-param.dbf file. Once you have the table in the project, select it and execute the Table/Start Editing command. Then select the view, make sure that the tool button assigned to the script and that the Filled DEM are selected. Continue selecting points with the mouse for each WR whose parameters are required, indicating in each case the ID number and if you desire to view the resulting grids.

When you are done, select Table/Stop Editing and say Yes when asked about saving the edits. Depending on the actual cells that you selected, your table should look something like Figure 6.

Figure 6

As mentioned, the resulting table, wr-param.dbf, can be linked or joined to other tables using the WR_id field, or it can be accessed in programs like MS Access or MS Excel.

Conclusions

The objective of the project was accomplished. A procedure to determine in a semi-automatic way the parameters of each water right site for the Curve Number method was developed and is described in this document. The procedure is semi-automatic because it requires input from the user (point location and ID) for each water right site.

It will be difficult to create procedures to determine these parameters in a completely automatic fashion. Observe that the WRs do not fall directly on the streams defined by the flow accumulation grid or the river reaches lines. This is due to the inaccuracy of the data. The resolution of the DEM used in this project is 500 m (each cell has 500 m per side), and the RF1 files were used instead of the more accurate RF3. The choice of data was made based mainly on its availability, but the procedures and scripts developed in this project can be applied directly over more accurate data. However, the difficulty in acquiring the data and the fact that it comes from different sources suggests that it is not likely that in the near future 100% accurate and compatible data will be available.

The principles on which it is based are very simple, mainly map arithmetic and weighed averages, and should be easily understood. However, the procedures, in particular the way to use the script, are rather rigid and should be used carefully and according to the instructions herein presented. Nevertheless, the procedures developed in this project are readily usable and adaptable, and are mainly intended to serve as a starting point towards more sophisticated methods.

Data Dictionary

References

Wurbs, R. A. (1998). Technical Investigation Related to Naturalized Streamflows. Literature Review and Work Plan for Task 1. Prepared for the Texas Natural Resource Conservation Commission by the Texas Water Resource Institute. Texas A&M University.

Back to homepage