USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007–5205

Back to Table of Contents

Summary and Conclusions

To address concerns over increased development in Carson Valley, the U.S. Geological Survey (USGS) recently made estimates of ground-water inflow to basin-fill aquifers of Carson Valley from the adjacent Carson Range and Pine Nut Mountains. Ground-water inflow was estimated to range from 22,000 acre-ft/yr using a water-yield method, to 40,000 acre-ft/yr using a chloride-balance method. Because of the relatively large range in these estimates, and uncertainties in each method, watershed models were developed for perennial watersheds of the Carson Range and Pine Nut Mountains to provide an independent estimate of ground-water inflow to Carson Valley

The model used for this study was the Precipitation-Runoff Modeling System (PRMS), a physically based, distributed-parameter model. A Geographic Information System (GIS) program, the Weasel Toolbox was used to manage spatial data and characterize model drainages and to develop HRUs.

Models were developed first for the four perennial watersheds having gaged daily mean runoff; Daggett Creek, Fredericksburg Canyon, Pine Nut Creek, and Buckeye Creek watersheds. Models were then developed for 10 ungaged perennial watersheds in the Carson Range using the models developed for either the Daggett Creek or the Fredericksburg Canyon watersheds as an index model. Selection of the gaged watershed to be used as an index model was based on similarity between the major bedrock type and the percentage of precipitation that becomes runoff in each watershed. Finally, models were developed for 10 ephemeral watersheds in the Carson Range and a large area of ephemeral runoff near the Pine Nut Mountains to estimate the quantity of ephemeral runoff tributary to Carson Valley and the potential for ground-water inflow from ephemeral watersheds. Because the ephemeral runoff is uncertain, selection of the index model was based only on the bedrock type underlying the ephemeral watershed. The Buckeye Creek watershed model was used as an index model for the area of ephemeral runoff near the Pine Nut Mountains because of the similarity of geologic units exposed in the two areas. Model calibration was constrained by daily mean flows for four gaged watersheds in Carson Valley and ten ungaged watersheds in the Carson Range estimated in a previous study. The models were further constrained by annual precipitation volumes estimated in a previous study to provide estimates of ground-water inflow using similar water input.

The calibration periods were water years 1990–2002 for watersheds in the Carson Range, the period for which reconstructed runoff was available for ungaged perennial watersheds, and water years 1981–97 for Pine Nut Creek and Buckeye Creek watersheds. Simulations for water years 1990–2002 were then made using the calibrated models for Pine Nut Creek and Buckeye Creek watersheds to obtain daily mean values for the 1990–2002 period.

The watershed models were affected by the assumption that the ungaged perennial and ephemeral watersheds are hydrologically similar to the index watersheds, the scale of available soil data, the adequacy of available climate data and accuracy of precipitation estimates using the linear-relations, and the accuracy of the reconstructed daily mean runoff used for calibration of the ungaged perennial watersheds.

The error of the resulting models was determined for watersheds having gaged runoff data. Mean annual bias for the period of record for Daggett Creek and Fredericksburg Canyon watersheds was negligible, relative error was about 6 and 12 percent, respectively, and RMSE was about 1 and 6 in., respectively. A satisfactory overall fit was obtained between simulated and measured runoff for Daggett Creek for the 1993–98 period of high runoff (and periods of higher ground-water inflow); however, runoff is underestimated for the drier years (1990–92, 1994).

Overall, mean annual water budget volumes indicate that the Daggett Creek and Fredericksburg Canyon watershed model simulations are in agreement with previous estimates for the 1990–2002 period. The mean annual runoff efficiency using the simulated runoff and simulated precipitation was 20 percent for Daggett Creek watershed and 45 percent for Fredericksburg Canyon watershed, both comparable to the efficiency calculated with the measured flow. With the exception of the April–June season, calibration statistics for the seasonal aggregates are considered satisfactory for both the Daggett Creek and Fredericksburg Canyon watershed simulations.

Error for the period of record for the Pine Nut and Buckeye Creek watersheds was greater, in part due to the sensitivity of the calibration statistic for days with zero and low runoff whereby differences in runoff were assigned a large relative error. In addition, the spatial and temporal distribution of precipitation on the east side of Carson Valley may not be adequately represented due to the more localized nature of storm events, and the lack of high-altitude precipitation data. The mean annual runoff efficiency for the 1990–2002 period using the simulated runoff and simulated precipitation was 2 percent for Pine Nut Creek and 2 percent for Buckeye Creek, both comparable to the efficiency calculated with measured flow.

The Pine Nut Creek watershed model underestimates runoff for all but the wet years in the latter part of the record, while adequately simulating the bulk of spring runoff for most years. The Pine Nut Creek watershed model has a large negative bias, and under-estimates mean annual runoff on average by 11 percent. Overall. the simulated hydrograph for Buckeye Creek indicates a much quicker response to precipitation than measured at the gage and the Buckeye Creek model tends to over-estimate runoff. For water years 1981–97, the mean annual error is about –5 percent when water year 1994 is removed. For the February–April period, the Buckeye Creek model has a large negative bias and 5 percent relative error. The bias and error of the calibrated models are within generally accepted limits for watershed models, indicating the simulated rates and volumes of runoff and ground-water inflow are reasonable.

The daily mean values of precipitation, runoff, ET, and ground-water inflow simulated from the watershed models were summed to provide annual mean rates and volumes for each year of the simulations, and mean annual rates and volumes for water years 1990–2002. Daily mean and annual mean simulated runoff matched measured or reconstructed runoff reasonably well for most watersheds, with the exceptions of Monument Canyon, Sheridan Creek, and Jobs Canyon. The difference between mean annual simulated and measured or reconstructed runoff was 11 percent or less, also with the exceptions of Monument Creek, Sheridan Creek, and Jobs Canyon.

Model results were consistent with a previous study, in that no subsurface flow was simulated from Mott Canyon, Monument Creek, and Sheridan Creek watersheds, and subsurface flow may be taking place from the Stutler Canyon to the Sheridan Creek watershed. However, the watershed models simulated ground-water inflow from the Water Canyon and Fredericksburg Canyon watersheds, and ground-water inflow from ephemeral watersheds in the Carson Range and Pine Nut Mountains. The simulated mean annual volume of ground-water inflow for the 14 perennial watersheds totaled 14,400 acre-ft. The annual ground-water inflow simulated from the 10 ephemeral watersheds of the Carson Range totaled 3,500 acre-ft, whereas a previous study assumed ground-water inflow from ephemeral watersheds of the Carson Range was negligible. Simulated ground-water inflow from the ephemeral watershed near the Pine Nut Mountains was 5,700 acre-ft.

Simulated mean annual runoff rates ranged from about 5 to 7 inches for ephemeral watersheds using Daggett Creek as an index station and from 9 to 13 inches for ephemeral watersheds using Fredericksburg Canyon as an index station. Mean annual ephemeral runoff from the Carson Range simulated from the models totaled 9,900 acre-feet. For the area of ephemeral runoff near the Pine Nut Mountains, the mean annual ephemeral runoff rate simulated from the model was 0.1 inch for a total runoff of 800 acre-feet.

The mean annual ground-water inflow to basin-fill aquifers of Carson Valley from the Carson Range and the Pine Nut Mountains simulated by the watersheds models for water years 1990–2002 was 35,000 acre-feet, with 19,000 acre-feet from the Carson Range and 16,000 acre-feet from the Pine Nut Mountains. The simulated mean annual inflow varied over an order of magnitude, from 7,800 acre-feet during dry conditions (water years 1990–92), to 76,000 acre-feet during wet conditions (1995–97).

The uncertainty in simulated annual runoff and ground-water inflow from ephemeral watersheds of the Carson Range caused by selection of the index model ranged from 13,000 to 26,000 acre-feet, considerably less than the range from dry to wet conditions. Because bedrock type appears to affect the volumes of runoff and ground-water inflow from gaged and ungaged perennial watersheds, volumes simulated for the ephemeral watersheds using the appropriate index model were considered to be the best and most reasonable estimate.

Total mean annual ground-water inflow simulated by the models is 38,000 acre-feet with the addition of ground-water inflow from Eagle Valley and recharge from precipitation on eolian sand and gravel deposits, 1,700 acre-feet, and ground-water recharge from precipitation on the western alluvial fans, 500 acre-feet. The estimate of 38,000 acre-feet for mean annual ground-water inflow simulated from the watershed models, was in close agreement with that obtained from the chloride-balance method, 40,000 acre-feet, but was considerably greater than the estimate obtained from the water-yield method, 22,000 acre-feet from a previous study. The similar estimates obtained from the watershed models and chloride-balance method, two relatively independent methods, provide more confidence that they represent a reasonably accurate volume of mean annual ground-water inflow to Carson Valley. However, the two estimates are not completely independent because they use the same distribution of annual precipitation.

Annual ground-water recharge of the basin-fill aquifers in Carson Valley ranged from 51,000 to 54,000 acre-feet using estimates of ground-water inflow to Carson Valley simulated from the watershed models combined with estimates of other ground-water budget components from a previous study. Estimates of mean annual ground-water discharge ranged from 44,000 to 47,000 acre-feet, including an increase of 3,000 acre-ft/yr in the amount of net ground-water pumping from the estimate in a previous study.

The low range estimate for ground-water recharge, 51,000 acre-feet per year, is most similar to the high range estimate for ground-water discharge, 47,000 acre-feet per year. Thus, an average annual volume of about 50,000 acre-feet is a reasonable estimate for mean annual ground-water recharge to and discharge from the basin-fill aquifers in Carson Valley. However, the results of watershed models indicate that significant interannual variability in the volumes of ground-water inflow is caused by climate variations. During multi-year drought conditions, the watershed simulations indicate that ground-water recharge could be as much as 80 percent less than the mean annual volume of 50,000 acre-feet per year.

Back to Table of Contents