of the Ground Water Levels
in the Albuquerque Area
Eric Robinson
May 9, 2003
CE 547 - GIS in Water Resource Engineering
Professor Julie Coonrod
University of New Mexico
LINKS TO GRAPHICS – Appendix A, Appendix B, Appendix C, Appendix D, Appendix E, APPENDIX F, Appendix G, and Appendix H.
Text of report is below
Introduction: This assignment was the final project for the GIS for Water Resource Engineering class (CE 547) offered the spring semester 2003 at the University of New Mexico. This topic was chosen out of curiosity and a desire to utilize two sets of easily available and widely used data to answer the specific questions detailed below. As an introductory GIS project, I felt the need to use mostly pre-existing sets of data to answer a fairly simple pair of questions.
Objective: To determine whether the ground water levels in the Albuquerque area fluctuate on a seasonal basis and to look at long-term trends within a given season. Although the data sets used are not necessarily considered to be the best to answer the subject questions, they were utilized as a test of sorts of frequently and easily available data.
Hypothesis: 1). The lack of surface irrigation in the valley in the winter will result in a decline in the winter water table, while in the heights the reverse may occur as the large City of Albuquerque wells are pumping at a much reduced rate relative to the summer pumping rate. I do not believe the connection will be as strong in the heights however due to fact that most of the well data are either based on one-time well drilling events which would most likely be farther away from the City wells or because in the case when actual well measurements are used, most hydrologists prefer to measure wells after pumping has been suspended for a given amount of time to allow the water table to stabilize. 2). I expect the long-term trends within each season to reflect the observed draw down in the local aquifer over time.
Data Sources: Two sets of well data were downloaded from the Office of the State Engineer’s Website – www.seo.state.nm.us/water-info/gis-data. The location accuracy of both sets is limited to the historical requirement of describing a well location within a ten-acre description which yields a radius of 372 feet. The first set of data is well log data which essentially consists of well logs submitted by the well drillers shortly after drilling the well. The depth-to-water information from the well logs has an accuracy of only +/- 5 feet (Carr, Jerry E., hydrologist, State Engineer’s Office, pers. com.) and also lacks elevation data. The second set of data was assembled by the Espanola Basin Technical Advisory Group (hereafter EBTAG) and includes well data as far south as Albuquerque (see Appendices A and B for 1950s and 1990s well location map, respectively). The EBTAG data set has elevations, both ground level and water level. It should be noted that for the most part neither of these data sets are long-term records of water depths in wells. Rather, they are mostly a one-time record of the water depth as recorded by the well driller. It is possible and likely that there is some overlap in the two sets of data as they both rely, at least in part, on the State Engineer’s database, W.A.T.E.R.S.
The other data was from rgis.unm.edu and includes land use coverages from 1955 and 1992 as well as one digital elevation model (#178). This set of data was used for display only.
Projections Used: NAD 1983 UTM Zone 13N, Transverse Mercator, False Easting: 500,000.0, False Northing: 0.0, Central Meridian: -105.0, Scale Factor: 0.9996, Latitude Of Origin: 0.00.
Software Environment: The ArcView environment within ArcGIS was used exclusively. Also for data analysis purposes, only the raster environment was utilized.
Procedure: 1). Starting with the database files associated with the two data sets, columns were added and the data was classified in groups by decade and season with “winter” being December, January, and February and “summer” being May, June, and July. Well data from all other time periods was excluded as was well data lacking a well completion date and a depth-to-water measurement. The decade and season period were then selected using the “Select by Attributes” option to create and export 1) four new database files for the EBTAG data – 1950s winter, 1950s summer, 1990s winter, and 1990s summer and 2) two new database files using the well log data – 1993-2002 winter and 1993-2002 summer. With regard to the EBTAG data only the 1950s and 1990s data were used for analysis for contrast purposes. There was very little data before the 1950s. The various EBTAG-based database files were then displayed as shapefile points, then splined to obtain a grid. Then using the “Raster Calculator” the summer grid of each decade group was subtracted from the winter grid. This yielded two output grids showing the seasonal variation of each decade group. Also using the Raster Calculator, the 1950s season subtracted the same 1990s season yielding two grids, one showing the long-term winter variations and the other the long-term summer variations. Regarding the well log data, the two database files, 1993-2002 winter and 1993-2002 summer, were displayed as shapefile points, then only the wells within the valley floor were manually selected due to the lack of elevations. Then a flat valley was assumed and a similar procedure was performed as described for the EBTAG data using the Raster Calculator.
Results
Log Well Data – As seen in Appendix C, the stated hypothesis is not necessarily validated by this calculation.
EBTAG Well Data – The 1950s seasonal variation in ground water levels is shown in Appendix D. This result is contrary to the stated hypothesis; however, the 1950s data set is rather thin. Therefore, the value of the results is limited. The 1990s seasonal variation in ground water levels is shown in Appendix E. Despite significantly more wells, there is no clear pattern and comparing this result with that of the well log data, the two mostly overlapping time periods seem to partially contradict each other (see APPENDIX F). The result of the long-term winter water levels is shown in Appendix G. The result is that, in general, the 1950s winter water levels were higher than the 1990s winter water levels which is in accordance with the hypothesis. The result of the long-term summer water levels as shown in Appendix H is less conclusive, again due in large part to a shortage of good data.
Conclusion
The seasonal variation shown in the results did, for the most part, not coincide with the stated hypothesis. The reason for this appears to be that the quantity of the data, particularly the 1950s data set, is insufficient to answer the question asked in this report. Additionally, the quality of data, both the location of the wells and the accuracy of the depth to water information, most likely proved to be a limitation to the analysis. A comparison of the two 1990s seasonal variation data sets shows just this. Another factor is that the aquifer and the ground water- surface water interactions are much more complex than anticipated.
Additional Work
The conclusions reached in this report are based primarily on a visual inspection of the output grids. It would be appropriate to statistically analyze the sets of results to more specifically quantify the validity or lack thereof. Additionally, to confirm or deny the conclusions reached above the same questions should be answered using the best set of data available- long-term, bi-seasonal well measurements.