Executive Summary

Ecological Recovery Foundation (ERF) mapped 1,840 possible spring locations in a 71-county project area. Of the total, the Texas Water Development Board (TWDB) provided its database of 796 springs (TWDB, 2005) which included springs submitted by the U.S. Geological Survey (USGS) from their 2003 study (USGS, 2003a). ERF provided 859 springs from its own research (ERF, 2005) and another 185 Geographic Names Information System (GNIS) springs from a query of their place names database (USGS, 2003b).

Fourteen of the 71 counties in the study area in the TWDB database have no records of springs. From the three data sources used in this study (TWDB, ERF, and GNIS), only five counties have no records. Predictably, the counties with the most springs occur in the Texas Hill Country and are underlain by the Edwards and Edwards Trinity aquifers.

ERF completed an optional task at the request of TWDB. The purpose was to refine the data so that TWDB could select target springs more effectively for further investigation. ERF compiled historical spring records and flow discharge records for this task. TWDB has records of 383 springs with flow measurements. By looking at additional flow data, ERF found an additional 185 springs in USGS flow data records (USGS, 2003a). A total of 488 springs from the two sources have at least one flow record; 105 of these have multiple flow measurement values. The 1,044 springs in the ERF and GNIS databases have no known records of discharge.

ERF found that many of the springs’ location data from TWDB are in question for two main reasons. First, some of the locations of historical springs (from the 1930s and 1940s) were best-guessed digitally from very poor information that used approximate mileage in an approximate cardinal direction from a nearby town (Heitmuller, pers. comm.). Second, the locations of approximately 100 springs were described only in degrees and minutes (Brune, 1975). After conversion to decimal degrees, the locations could be offset by as much as a mile. In general, the actual locations of many of the 1,840 possible springs are questionable.

ERF also questions the total number of springs in the project area. In looking at data sources and considering the questionable mapping due to insufficient or incorrect data, ERF staff speculates that there may be fewer springs in the project area than are compiled in their database. ERF eliminated duplicate springs from among the three spring sources but only did so for springs within 300 feet of each other. Springs mapped farther apart may also be duplicates. Conversely, ERF, in past field studies, found springs that have not been recorded by any agency or organization. These and other questionable data questions can only be answered in the field.

Introduction

Springs are the least studied and least known of Texas’ water resources. In the 1970s, Gunnar Brune compiled data on springs in 183 of Texas’ 254 counties (Brune, 1981). Although the TWDB and the USGS have flow data and other information on numerous springs in the state, researchers recognize that Mr. Brune’s study is one of the most comprehensive compilations of spring data statewide.

The TWDB provided a grant to gain more information on springs in the remaining 71 counties (Table 1) as part of its goals to improve conceptual models and boundary conditions in groundwater and surface water availability models, to provide a more accurate database of springs throughout the state to interested users, and to gain a better understanding of the interrelationship of surface geology and springs.

The study includes five tasks: springs compilation, selection of 400 target springs for further study, pre-field activities, field activities, and post-field consolidation. This report completes the requirements for Task 1 under TWDB Contract No. 2006-001-069.

Table 1

Individual activities performed under Task 1 were to:

1.         Overlay information layers containing springs recorded by the TWDB on top of county and quadrangle boundaries utilizing ArcView 9.1. 

2.         Visually scan all topographic maps for springs. Mark additional springs digitally and enter them into a database. Give each spring a unique identification number.

3.         Transfer spring locations and identification numbers to paper copies of the same USGS topographic maps for use in the field. (NOTE: Although not included in the contract, ERF also queried GNIS springs and placed them into the project database.)

4.         Provide the spring database in dbf format to TWDB.

5.         Transfer spring locations onto Bureau of Economic Geology Geologic Atlas of Texas (GAT) maps in pdf format and provide to TWDB. 

ERF has completed all tasks except for the transfer of spring locations to paper copies on approximately 65 unavailable quad sheets (with the concurrence of TWDB).

Optional Scope

In its proposal to TWDB, ERF suggested an optional task to delineate further those springs that might be of interest to the agency. TWDB elected to include this option although they were not able to increase the budget. ERF agreed to provide information as readily available. The scope of the optional task and the results are in the body of this report.

Task 1 Methodology

ERF, through earlier independent research in 2002, compiled approximately 1,500 springs in the 71 counties addressed by this project. ERF compiled its spring data from three sources:

After ERF’s independent research was completed, the TWDB acquired more springs data from the USGS. Before the USGS springs were added to the existing TWDB database, USGS and TWDB staff double-checked the data for possible duplications (Hopkins, pers. comm.). Staff determined that the locations of all the springs were different but did not verify through actual field visits.

For this present study, ERF acquired spring data from the TWDB in September of 2005 (that included the added USGS data) and produced new digitized maps of those springs (utilizing ArcView 9.1), again overlain on seamless USGS 7 ½-minute topographic maps. ERF retained the GNIS and ERF springs from their earlier study and added them to the same maps. In an attempt to discern if locations from the new data (i.e., TWDB springs) were duplicates of the old data (i.e., GNIS or ERF springs), ERF examined all TWDB springs that were in close proximity to springs from the other sources.

Staff checked all ERF springs within 300 ft of TWDB springs. If a TWDB and an ERF symbol were found for the same spring, ERF considered one spring location a duplicate and deleted the ERF spring. Of the 55 duplicate springs, staff deleted 50 of the duplicate springs.

Of the five remaining ERF/TWDB springs that were in close proximity to each other, each TWDB spring appeared between two spring symbols on the USGS topographic maps, whereas each of the spring symbols contained[j2]  a digitized ERF location.

As ERF did not know which of the two springs (or both) was the TWDB spring, staff left the TWDB symbol in place and one of the ERF symbols was arbitrarily deleted.

Both Kerr and Kimble counties had one double spring; Bandera County had three double springs. Double springs will require verification in the field if selected as target springs for field visits.

ERF applied the same technique to determine if the compiled database had duplicate TWDB/GNIS springs; none were found.

Once staff resolved the duplicate springs issue, all springs were given a unique 3-digit identification number (ID), county by county (leading zeros are not shown). Spring locations and IDs were transferred from the digital topographic maps to paper maps for use in the field.

As ERF will be conducting field studies on only 400 of the possible 1,840 springs, placing all the spring locations on paper quads may not seem a particularly efficient use of staff time and budget. In practice, the effort was very beneficial as each spring was examined and numerous data issues came to light (discussed in Research Issues). 

Once all of the issues were resolved to the satisfaction of ERF and TWDB, staff acquired the Bureau of Economic Geology’s GAT maps in digital form and overlaid the springs on the GAT maps. ERF provided springs on GAT maps for each county in pdf format and the springs’ locations in ArcMap shape files to TWDB.

Research Issues

1.         The most perplexing data question involved the GNIS data. These data were in ERF’s original database in 2002, along with TWDB and ERF data. When the USGS submitted its spring survey data to TWDB, ERF staff assumed that most, if not all, of the GNIS spring symbols would be replaced by TWDB spring symbols because the GNIS springs are in the national USGS database. This was not the case. When ERF replaced its original TWDB data with TWDB’s newer data, much of the GNIS data remained the same. As described in Task 1 Methodology, staff examined TWDB and ERF spring symbols that were within 300 feet of each other. None of the 185 GNIS springs appear to be duplicates of the TWDB data.

2.          In several instances, both the TWDB and GNIS databases appeared to have their own duplicate records for springs.

• Kendall County (USGS Kendalia quadrangle). Three TWDB springs were identified in the same location—Edge Falls—with identical coordinates, identical state well identification numbers, and other identical data fields. ERF deleted the two higher numbered springs (49 and 50). Note: The USGS database has separate state well IDs for these springs. ERF found this information after the springs were deleted from the TWDB database, with concurrence of TWDB staff. In the Optional Scope subtask, ERF used flow records from the more extensive USGS database. This resulted in a discrepancy of two springs between the two databases (see Table 8 and the ArcMap dbf attributes table).
• Brazos County (USGS Wellborn quadrangle). TWDB springs 4 and 5 had identical identification information. ERF deleted Spring 5.
• Grimes County (USGS Carlos quadrangle). TWDB springs 6 and 7 had identical identification information. ERF deleted spring 7.
• Angelina County (USGS Zavala quadrangle). Two identification numbers seemed to apply to only one GNIS spring. ERF deleted spring 3 and retained spring 2.
• Kerr County (Bee Caves Creek quadrangle). GNIS spring 96 (Muskhog Spring) was in the vicinity of two spring symbols on the map, about 500 ft apart. ERF arbitrarily labeled one of the springs ERF 164 and assumes that the GNIS spring is the other location.
• Coke County (Tennyson quadrangle). The GNIS database indicated a place name spring in the northwest corner of the map. ERF could not find a spring name or symbol on the map but left the data as is.

3.         When looking at springs by county using the 3-digit ID numbers, the highest number does not necessarily indicate the number of springs found in that county. After all springs were given unique IDs, ERF staff verified the existence of each spring on USGS digital and paper topographical maps. Some were deleted (see Research Issue No. 2), so gaps appear in the numbering sequence.

4.         Many of the TWDB springs (        ) do not appear in or even near a spring symbol (      ) on the topographic maps. In a number of instances, a spring appears on a hillside far away from a water course with no apparent path to a creek. Others appear on relatively level ground (between two widely separated contour lines) near a water course but with no indication of a path to the waterway.

ERF can only speculate that because most of the spring coordinates were added to the TWDB database before handheld GPS units were available, field measurements were less accurate than they are now. In cases where the spring symbols are not near a watercourse, those sites might actually be wells that were accidentally labeled springs under Well Type. These “springs” could be checked against the TWDB Water Information Integration and Dissemination mapping application for information that might help determine their veracity, but this is not within the scope of this project.

5.         Many of the springs’ locations are in question. Locations of springs in old Texas State Board of Water Engineer (TSBWE) reports from the 1930s and 1940s were best-guessed digitally from very poor information that used approximate mileage in an approximate cardinal direction from a nearby town. For example, a spring location might be described as “3.5 miles SE of San Saba.” This spring and others from the TSBWE reports were mapped digitally by the USGS. If the spring location could not be readily matched to an existing spring on the topographic map, the USGS placed the spring according to description and where they normally would occur (e.g. “streambeds, small ponds, forested areas, or in streams of the same name”) in the general vicinity (Heitmuller, pers. comm. and USGS, 2003a).

TWDB staff believes that historical spring data from an old TWDB report (Brune, 1975) is included in their groundwater database. If so, they converted Brune’s coordinate locations (degrees and minutes) to decimal degrees. The locations of those springs could be offset as much as a mile. The actual locations of these springs (and the resultant positions on BEG GAT maps) are questionable. 

Results

ERF staff found a possible 1,840 springs in its 71-county project area using data compiled from the TWDB well identification database (TWDB, 2005), from a query of the USGS GNIS, and from ERF’s independent research. The breakdown by source is:

TWDB             =          796

ERF                 =          859

GNIS               =          185

Table 2 shows the counties for which TWDB had no springs; Table 3 shows counties with no known springs from any of the three sources; and Table 4 shows the 71 counties and the number of springs found in each one. Tables 5 and 6 are lists of the TSBWE springs and the Brune (1975) springs, respectively. The state well identification number assigned to each is also included. The springs in the latter two tables are the springs with the least credible locations. Tables 5 and 6 are attached to this report in Excel spreadsheet format (Table 5 and Table 6)

Table 2

TWDB Counties with No Recorded Springs

Source: TWDB, 2005

Table 3

Counties with No Recorded Springs from All Data Sources

Sources: TWDB, 2005; ERF, 2005; and USGS, 2003b

Table 4

Number of Springs by County

Optional Scope

In its application to the TWDB, ERF suggested that TWDB staff select target springs based upon their locations on the GAT maps. As an option, ERF suggested additional criteria to select the target springs. TWDB declined to fund the optional scope but requested that ERF provide the data gratis. As a result, ERF offers this compilation of data based upon the optional scope.

1.         Determine which springs in the 71-county project area were included in the TWDB Report 189, “Major and Historical Springs of Texas” (Brune, 1975).

NOTE: Data from Brune is presented separately in this report as ERF cannot assure TWDB that the springs in Brune’s report have been matched correctly to later TWDB state well IDs.

ERF first created a table of all springs in the 71-county project area from Brune (1975) by county, spring name, number of discharge measurements, and the highest and lowest of those measurements (Table 7, attached to this report in Excel spreadsheet format). Flow measurements were converted from cfs to gpm. The data in Brune’s 1975 report may provide interesting historical data to the TWDB as some of the measurements go back to the 1890s.

Springs in Brune’s report did not have state well numbers, but TWDB staff think that the springs are in their groundwater database (Hopkins, pers.comm.). To correlate the springs, ERF looked at the more descriptive table of springs from the USGS (2003a) rather than using the TWDB database. Springs were matched through spring name (which may have changed in time) in each county and then assigned the state well ID from the same table. ERF then converted Brune’s coordinate measurements (degrees and minutes) into decimal degrees and overlaid the resultant approximate locations onto the springs digital base map as an additional layer. ERF staff looked at some, not all, of the Brune springs on the digital maps to assure that those springs were at least in the approximate location as the state well ID assigned to them. The numbered springs are in a separate layer in ArcMap and are represented by s. These springs are also noted on Table 7 along with the approximate distance and direction from the TWDB spring. ERF could not satisfactorily correlate the TWDB springs with the ones mapped through ERF’s conversions, but most were near other springs (approximately 2,000 to 7,000 ft).

NOTE: ERF used 0 seconds in their coordinate conversions. If TWDB used 30 seconds, then their mapped springs and ERF’s mapped springs would not overlap each other.

2.         Determine which of the total number of springs have any other historical flow spring data (from the 1930s and 1940s Texas State Board of Water Engineers reports and later TWDB reports, if readily available). Note: the locations of these springs may be difficult to correlate with the mapped springs in Task 1.

Since the USGS (2003a) discharge tables have more exhaustive flow data than the TWDB database, ERF used that as the basis for determining flow histories of springs. Flow records from the TWDB and the TSBWE are included in the USGS database. Table 8 (attached in Excel spreadsheet format) shows the total number of spring flow records by county in the 71-county project area. The table breaks down the data further into TWDB and TSBWE flow records. ERF converted flows from cfs to gpm as needed (rounded to the nearest whole number). High and low flows from each data source are also shown. Several springs in the “State Well ID” column are marked with an asterisk. These indicate springs from the TWDB database that were not found in the more comprehensive USGS database.

NOTE: The USGS database contains two springs that were deleted from the TWDB (2005) database used in this study. The USGS database (Table 8) has 490 springs with one or more flow records. The mapped springs (ArcMap) show only 488 (see Research Issue No. 2). For this portion of the study, ERF assumes that 488 is the correct number.

Staff compared all flow discharge values subjectively to ensure that flow records were not represented more than once across the three data sources. If two flow values were the same, ERF checked the dates of each record to ensure the records were not carried forth from one agency database to another. ERF staff deleted obvious duplicate records from the table. Several others were questionable. For some records with identical flow values, TWDB was one reporting agency with 1940s collection dates but TWDB was not the agency name at that time. The other records came from TSBWE files with a slightly different data collection dates (e.g., 9/43 vs. 12/43). ERF did not delete these seemingly duplicate records.

ERF also determined which of the springs in the TWDB database had either one or multiple flow records. On the GAT maps generated by ArcMap, springs with one or more flow records are represented by      and      symbols, respectively. All other TWDB springs (        ), ERF springs (     ), and GNIS springs (    ) are assumed to have no flow measurements on record at the TWDB or USGS.

NOTE: To be consistent with the GAT maps, ERF suggests that the same symbology be used on other maps that TWDB generates with this project’s ArcGis database.

Spring flows have been measured at 488 TWDB springs. Dove Creek Spring in Irion County has the highest recorded flow at 24,684 gpm (with a low of 350 gpm) from 352 measurements. Table 9 shows 15 springs with the highest recorded flows, their lowest recorded flows, and the number of recorded measurements.

Table 9

Springs with the Highest Recorded Flows in the 71-County Project Area

Sources: USGS (2003a), TWDB (2005), and TSBWE (within the USGS data)

3.         Determine which of the springs currently in the TWDB database are being monitored.

The TWDB is not currently monitoring any springs in the 71-county project area although they are working with the Texas Parks and Wildlife Department to collect springs data (Hopkins, pers. comm.). The USGS is taking discreet (non-continuous) flow measurements at Dove Creek Springs in Irion County, San Saba Springs in San Saba County; and Fort McKavett Springs in Menard County (Heitmuller, pers. comm.).

4.         Apply data from steps 1, 2, and 3 to further prioritize the target springs delineated on geologic maps.

Springs overlaid on the Bureau of Economic Geology’s GAT maps and in the ArcMap shape files (submitted to TWDB on cds) are shown in five categories:

TWDB springs with no flow records =

TWDB springs with one flow record =

TWDB springs with multiple flow records =

ERF springs =

GNIS springs =

ERF and GNIS springs are thought to have no flow measurements from the sources described earlier unless they are duplicates of TWDB springs (see Research Issue No. 2) or have flow records from entities other than TWDB or USGS. Maximum and minimum flow values are given in Table 8, along with the number of records for those identified springs.

NOTE 1: Symbol colors may be different on some county GAT maps as the base colors were too similar. Colors for each county are noted on the maps’ legends.

NOTE 2: To be consistent, ERF suggests that the same symbology be used on other maps that TWDB generates with this project’s ArcGis database.

5.         Develop a list of springs by county that would be most likely to yield useful information from targeted geologic formations, from springs with known historically high flows and/or historical value; from springs with historical flows (but not necessarily high flows), and from springs already being monitored by the TWDB or USGS.

ERF declines to provide a comprehensive list with all the parameters outlined in option item No. 5 as all that data is provided separately in the Optional Scope tasks 1 through 4.

Attached Tables

• Table 5 - TWBWE Mapped Springs with Questionable Locations

• Table 6 - Brune (1975) Springs in the 71-County Project Area with Questionable Locations

• Table 7 - Brune Spring Flow Records in the 71-County Project Area

• Table 8 - Spring Flow Records from USGS, TWDB, and TSBWE

References

Spring Sources

Ecological Recovery Foundation. 2002. Unpublished research.

Ecological Recovery Foundation. 2005. Unpublished research.

Texas Water Development Board. September 2005. Water well data.

U.S. Geological Survey. 2003a.Database of Historically Documented Springs and Spring Flow Measurements in Texas. Open-File Report 03-315. Authored by Franklin T. Heitmuller and Brian D. Reece.

U.S. Geological Survey. 2003b. Geographic Names Information System. Reston, VA. (database query).

Water Flow Sources

Brune, Gunnar. 1975. Major and Historical Springs of Texas. Report 189. Texas Water Development Board.

Capitol Environmental Services. 2002. Springs and flows from selected Texas State Board of Engineers reports (unpublished data).

Texas Water Development Board. 2005. Springs queried from the TWDB water well inventory (in spreadsheet and dbf formats)

U.S. Geological Survey. 2003a. Database of Historically Documented Springs and Spring Flow Measurements in Texas. Open-File Report 03-315. Authored by Franklin T. Heitmuller and Brian D. Reece

Other References

Brune, Gunnar. 1981. Springs of Texas, Volume I. Branch-Smith, Inc. Ft. Worth. (Reprinted in 2002 by Texas A&M Press).

Heitmuller, F. 2005. Personal Communication. Geographer. U.S. Geological Survey, Austin, Texas.

Hopkins, J. 2005. Personal Communication. Manager, Groundwater Monitoring Section, Groundwater Resources Division, Texas Water Development Board.

[j1] Place caption below table.
[j2] represented?