KGS Home Geohydrology Home Reports

Kansas Geological Survey, Open-file Report 2010-18

Ground-Water Model for Southwest Kansas Groundwater Management District No. 3

Gaisheng Liu, Brownie Wilson, Donald Whittemore, Wei Jin, and James Butler, Jr

KGS Open File Report 2010-18
This project was funded by the U.S. Bureau of Reclamation and Kansas Water Office (Kansas Water Plan)

March 2011

Executive Summary

Ground-water levels have been declining during the last few decades in the Ogallala-High Plains aquifer (HPA) in western Kansas, including within Southwest Kansas Groundwater Management District No. 3 (GMD3). The water-level declines have decreased ground-water discharge to the Arkansas and Cimarron rivers, thereby causing decreasing streamflow. One of the Kansas Water Plan (KWP) objectives is to "Reduce water-level declines rates within the Ogallala aquifer and implement enhanced water management in targeted areas." An associated goal of the KWP is to "Conserve and extend the life of the HPA." As a part of planning and management activities, the Kansas Water Office (under a cooperative agreement with the U.S. Bureau of Reclamation) and GMD3 contracted with the Kansas Geological Survey (KGS) to develop a computer model of the HPA in the GMD3 area to further characterize the hydrologic system and water availability. The model will provide more information on water in storage and allow projection of likely aquifer responses to possible future conditions and management scenarios (KWP, Upper Arkansas River Basin High Priority Issue, Management of the HPA).

The KGS constructed a numerical model for a rectangular area of 100 by 150 miles that enclosed GMD3 and extended approximately 6 miles to the north, east, south (into Oklahoma), and west (into Colorado) of the GMD3 boundaries. The active cells included the paleovalley of the Arkansas River in Hamilton and western Kearny counties. The KGS model utilizes MODFLOW, a widely used software program for modeling ground-water flow and stream- aquifer interactions developed by the U.S. Geological Survey. The KWO formed a Technical Advisory Committee to oversee the project, which included staff of the KWO, GMD3, KDA- DWR, and a consulting firm retained by KDA-DWR to provide technical review.

The main focus of the project was the development of a calibrated transient model that simulated ground-water flow and stream-aquifer interactions during the period 1947-2007. Predevelopment conditions were simulated for 1944-1946. The model included 12,083 active model cells (each a mile square), involved one layer, and simulated ground-water flow in the HPA and associated alluvial aquifers. Six recharge zones were used and the types of recharge included that from precipitation, enhancement of precipitation recharge in irrigated land, and return recharge below fields irrigated with ground-water and river water diverted from the Arkansas River. The precipitation applied to each cell varied depending on the distribution for each year across the model area.

Ground-water pumpage from the HPA for Kansas during 1990-2007 was based on reported water-use records, and for earlier years was estimated from regression equations based on a de-trended ratio of water use/authorized quantity versus precipitation and the Palmer drought severity index for 1990-2007. Similar approaches were applied to estimating pumpage in the Colorado and Oklahoma portions of the model, although the procedures varied because the data and data access for pumping records are not as readily available as those for Kansas. The pumpage rate from the HPA increased from 78,000 acre-ft/yr for predevelopment to a maximum of 2,708,000 acre-ft/yr in 1991 and was 1,844,000 acre-ft/yr for 2007 in the modeled area. The percentage of irrigation return recharge was calculated for each year in Kansas counties based on data for changes in irrigation type and applied to adjacent counties in Colorado and Oklahoma. Results from the calibrated model indicated that the long-term recharge from areal precipitation averaged over the model area was 0.41 in/yr during 1946-2007. Stream-aquifer interactions were simulated for the Arkansas and Cimarron rivers and Crooked Creek. Hydraulic conductivity (K) and specific yield (Sy) were estimated using lithologic data from about 15,000 well logs examined by the KGS PST+ (practical saturated thickness) program.

In order to account for the impact of declining water levels on the calculation of K and Sy during the transient period, the calibrated model was broken into six step models: 1) predevelopment, 2) predevelopment to 1966, 3) 1967 to 1976, 4) 1977 to 1986, 5) 1987 to 1996 and 6) 1997 to 2007. In each step model, K and Sy were dynamically updated using the observed water levels for the corresponding time period. During model calibration, the K and Sy values were adjusted by matching streamflows and observed water levels during each step to simulated values. A recharge function with different parameters for each of the six recharge zones was also incorporated into the calibration. The parameter estimation program PEST was employed to optimize parameters during the calibration process.

The model indicates that ground-water pumping has caused substantial decreases in aquifer storage. The storage decline rate started to increase in the 1950s, accelerated in the 1960s to mid-1970s, and then approximately leveled from the late 1970s to 2007, although it varied substantially each year depending on pumping. The accumulated decline in ground-water storage simulated for the entire model area for 1947-2007 is 66,409,000 acre-ft, which comprises 29.3% of the simulated predevelopment storage. The storage decreases have been accompanied by a decrease in streamflow out of the model. Water-level declines in the HPA have resulted in the "capture" of ground water that otherwise would have discharged to streams; without this capture, the aquifer storage loss would have been approximately 12% greater than simulated.

The total storage volumes simulated for the HPA only within the GMD3 area for predevelopment and the end of 2007 are 193,454,000 and 133,622,000, respectively, giving a storage decline of 59,832,000 acre-ft, which is 30.9% of the predevelopment value. The total storage volumes computed for the GMD3 area from measured water levels are 191,216,000 and 133,726,000 acre-ft for predevelopment and 2007, respectively. These values give a storage decrease of 57,490,000 acre-ft, which is 30.1% of the predevelopment volume. The storage volumes from the model and estimated from observations for the GMD3 area differ by only 1.2% and 0.1% for predevelopment and 2007 conditions. The average water-level decline simulated for all the model cells within the GMD3 area is 69.89 ft in comparison with 67.01 ft for the difference between contoured water-level surfaces based on observations in the predevelopment period to 2007.

The calibrated model will be used to simulate ground-water flow and stream-aquifer interactions for future conditions involving continuation and changes in pumping, and different climatic conditions as selected by the KWO and GMD3. A separate report that presents and discusses the results of these scenarios will be prepared.

The complete text of this report is available as an Adobe Acrobat PDF file.

Read the PDF version (13 MB)

Archive of model (458 MB)

Kansas Geological Survey, Geohydrology
Updated March 8, 2011
Comments to
The URL for this page is