Nevada Water Science Center


Aquifer Tests

Contact Information

Phil Gardner
Groundwater Specialist
Phone: (775) 887-7664
Email:pgardner@usgs.gov

 

Mailing Address
USGS
Nevada Water Science Center
2730 N. Deer Run Rd.
Carson City, NV 89701

 

Nevada Water Science Center
Information

Home Page Surface Water Groundwater Water Quality Research Contact Us

 

 

(C-20-19)19dcd-1

Primary Investigator: Keith Halford

Well Data

USGS Site ID
Local Name Altitude Uppermost
Opening
Lowermost
Opening
Primary Aquifer Transmissivity
(ft2/d)
390243114012201 (C-20-19)19dcd-1 5079 500 ALLUVIAL FILL 7000

Aquifer Test

All Aquifer Test Files (zip)

Snake Valley North

Aquifer Test (pdf) || Groundwater levels (Groundwater Watch)

Site and Geology

The aquifer test occurred near Baker, NV in Snake Valley where pumpage for irrigation caused measurable drawdowns over more than 100 square miles (Figure 1). The basin fill was pumped largely between 100 and 400 ft below land surface where undifferentiated intervals of silt, sand, and gravel occurred. The pumping interval was overlain by an areally extensive unit that was largely clay and silt that generally extended between 20 and 150 ft below the water table. These predominantly fine-grained, lacustrine deposits were differentiated and will be cited as the fine-grained fill. All other undifferentiated basin fill will be cited as the coarse-grained fill.

The groundwater flow system was interpreted with a thickness of 2,000 ft even though the thickness of basin fill exceeds 10,000 ft near Baker, NV (Welch and others, 2007). The hydraulic conductivity of the heterogeneous basin fill in Snake Valley generally decreases with depth because of increased cementation, induration and occurrence of evaporative deposits. Unconsolidated coarse-grained younger sedimentary rocks occur in the upper 2,000 ft and become indurated with depth. These deposits generally are underlain by Miocene sediments which contain thick anhydrite in southern Snake Valley (Welch and others, 2007).

 

Location of irrigation and observation wells in Snake Valley near Baker, Nevada (Universal Transverse Mercator projection, Zone 11)
Figure 1. Location of irrigation and observation wells in Snake Valley near Baker, Nevada (Universal Transverse Mercator projection, Zone 11).

 

 

Observation wells generally surrounded the pumping wells and irrigated areas (Table 1, Figure 1) and were completed between 50 and 300 ft below land surface (Figure 2). Observation well completions generally were shallower than the pumping well completions.

Perspective from west to east of pumping and observation wells near Baker, NV and vertical discretization of seven-layer model for simulating drawdown during 2000-2003 irrigation seasons
Figure 2. Perspective from west to east of pumping and observation wells near Baker, NV and vertical discretization of seven-layer model for simulating drawdown during 2000-2003 irrigation seasons.

 

 

Table 1. Well location and construction data for observation wells that were used in the Baker irrigation analysis aquifer test (See Figure 1 for well locations).

Well location and construction data for observation wells that were used in the Baker irrigation analysis aquifer test (See Figure 1 for well locations).

Period and Area of Analysis

Water level declines during the 2000-03 irrigation seasons were analyzed because crops had been inventoried and drought conditions existed during this period. Irrigated acreage in Snake Valley was inventoried during 2000, 2002, and 2005 (Wellborn and Moreo, 2007) so irrigated areas were known or could be interpolated within a year. Irrigated crops covered between 3,600 and 4,300 acres during the period of analysis (Figure 1). Most of the irrigation water was assumed to be groundwater because of the drought conditions during the period of analysis.

Annual groundwater pumping for irrigation ranged between 6,200 and 12,900 ac-ft assuming net application rates of 2 or 3 ft/acre (Figure 3). Net application rates are the application rate minus local precipitation and return flow. Annual application rates of 3.5 ft/acre were measured in Smith Valley where the altitude and latitude are similar to Snake Valley (Rush, 1976). Most of the application rate was assumed to be groundwater because annual precipitation was less than 0.5 ft, return flows are about 20 percent of application rates, and surface-water diversions were limited during drought periods. Net application rates for irrigation were estimated to the nearest foot given the uncertain component of surface-water sources and effects of return flow on water-levels.

Annual simulated pumpage from the Baker aquifer test model assuming net application rates of 2 and 3 ft.
Figure 3. Annual simulated pumpage from the Baker aquifer test model assuming net application rates of 2 and 3 ft.

 

Water Levels and Drawdowns

Depth to water in observation wells O1 and O2.
Figure 4. Depth to water in observation wells O1 and O2.

 

Yearly water-level declines were similar during previous drought periods of 1978-80 and 1990-92 and were compared qualitatively to simulated drawdowns. Translation of the seasonal water-level declines measured in well O2 during 1978-80 to the early spring measurements during 2000-02 by adding 22 years to the 1978-80 measurement dates are shown in Figure 4. Seasonal water-level declines measured in well O5 during 1990-92 also were translated from April 1, 1990 to April 1, 2000 for qualitative comparison during the 2000-03 period of analysis. A comparison between translated, continuous records and annual water-levels during late winter shows that annual decline can be estimated accurately from limited records.

Water-level changes were observed in eight wells that were affected by groundwater pumping for irrigation (Table 1, Figure 1). Water levels have been measured as frequently as weekly and as infrequently as annually. Annual and quarterly measurements occurred during the 2000-03 period of analysis (Figure 5). Water levels in the observation wells ranged between 10 and 40 ft below land surface.

 

Water level changes in selected observation wells.  Water-level changes shown for O2-1978x2 were measured 22 years earlier in O2 and shifted down 0.9 ft to show correspondence between measured and translated water level measurements after March 28, 2000.
Figure 5. Water level changes in selected observation wells. Water-level changes shown for O2-1978x2 were measured 22 years earlier in O2 and shifted down 0.9 ft to show correspondence between measured and translated water level measurements after March 28, 2000.

 

 

Drawdowns were estimated by subtracting water levels from water levels that were measured during March 2000. Water-level declines from prior pumping were minor or uncertain enough that further corrections were not justifiable. A drawdown of 0 was assigned to April 1, 2000 when pumping was assumed to begin. Drawdown rates ranged between 0.3 and 1 ft/yr and these extremes were observed in wells O1 and O5, respectively (Figure 5). Well O1 was more than 2 mi from the nearest pumping well and well O5 was within 2,000 ft of the nearest pumping well (Figure 1). Drawdown differences also exist because well O1 is screened across the shallow fine-grained material, whereas well O5 is screened in the deeper coarse sand which is pumped directly by the irrigation wells.

Numerical Analysis

Hydraulic properties of the basin fill were estimated by minimizing differences between simulated and measured drawdowns. Drawdowns were simulated with a three-dimensional, MODFLOW model that simulated seasonal, spatially distributed pumping (Harbaugh and McDonald, 1996). Parameter estimation was performed by minimizing a weighted sum-of-squares objective function with MODOPTIM (Halford, 2006).

The model domain was discretized into 173 rows and 137 columns (Figure 2 and Figure 6). Rows and columns were assigned widths of 1,000 ft where observation wells and the majority of pumping wells existed (Figure 6). The northern edge of the model was 330,000 ft north of the northernmost, 1000-ft wide row. Each successive row north of the northernmost, 1000-ft wide row was 1.1 times wider than the previous row and totaled 37 rows. The model similarly was expanded 330,000 ft from the area of 1000-ft-on-a-side cells to the south, east, and west. All lateral boundaries were specified no-flow. The model grid was oriented north-south in UTM, zone 11 projection for convenience where the lower, left model corner was 644,556 m easting and 4,209,076 m northing.

The absence of alluvium west of the irrigated areas was not simulated (Welch and others, 2007). This simplification was not expected to affect hydraulic property estimates because the mountain front of the Snake Range is between 5 and 10 mis west of the irrigated areas (Figure 6). The western edge of the alluvium also is irregular such that less than a third of the drawdown cone can impinge on this boundary.

 

Numerical model grid, pumping wells, observation wells, and generalized absence of alluvium near Baker, NV.
Figure 7. Numerical model grid, pumping wells, observation wells, and generalized absence of alluvium near Baker, NV.

 

 

The vertical extent of the model was discretized into 7 layers between 0 and 2,000 ft below the water table, which was within 50 ft of land surface around the observation wells. Layers ranged between 5 and 1,600 ft in thickness (Figure 2). Layer 1 was thinnest and simulated the water table. Layer 7 was thickest and occurred beneath the pumping interval in layer 6. The base of the aquifer system was specified no-flow. Changes in the wetted thickness of the aquifer were not simulated because the maximum drawdown near the water table was small relative to the total thickness.

The period between April 1, 2000 and March 31, 2004 was simulated with eight stress periods during the Baker aquifer test. Each stress period was divided into 50 time steps, the initial time step was about 0.0001 d, and successive time steps were 1.3 times greater than the previous time step. Irrigation pumping occurred during stress periods 1, 3, 5, and 7 which were 187, 180, 173, and 178 d, respectively. Irrigated acreage and annual pumpage varied less than 20 percent for a given application rate of 2 or 3 ft (Figure 3). All pumpage was from layer 6 which is between 150 and 400 ft below the water table. Recovery from each irrigation season occurred during the even stress periods when no pumpage was simulated.

Initial conditions of zero drawdown were assigned throughout the model domain. This was consistent with the assumption that water-level declines had largely ceased during the wetter 1997-98 period prior to the 2000-04 simulation period. The coarse-grained and fine-grained fills were distributed with three intervals: 0-20, 20-150, and 150-2,000 ft below the water table (Figure 2). Coarse-grained fill was simulated 0-20 and 150–2,000 ft below the water table and was simulated with layers 1, 2, 6, and 7. Fine-grained fill was simulated 20-150 ft below the water table and was simulated with layers 3, 4, and 5.

The observation wells were simulated as piezometers where simulated observation well depths were assigned discretely to model layers (Table 2). Distinctions between simulating the observation wells as piezometers or well screens of finite length were minor because the minimum thickness of a layer with an observation well was 20 ft and the period investigated was multiple years, rather than less than a day. Wells with unknown depths of completion were assumed to have screens between 80 and 150 ft below the water table. Uncertainties in well construction and simplification of comparison likely caused some misfit between simulated and measured drawdowns.

 

Table 2. Simulated depths of observation wells below the water table and model layer (See Fiure 1 for well locations).

Simulated depths of observation wells below the water table and model layer

 

Observations from well were considered best and weighted more than observations from other wells. Well O1 had more measurements during the 2000-03 simulation period and was more distant from the pumping wells. Water-level changes in O1 were more representative of regional conditions because well O1 was between 2 and 8 mi from the major pumping centers (Figure 1). Drawdowns were affected by a larger volume of aquifer between well O1 and the pumping wells and were much less sensitive to errors in pumping well locations.

Hydraulic Property Estimates

Hydraulic conductivity of coarse-grained fill, hydraulic conductivity of fine-grained fill, and specific yield of the basin fill were estimated. Specific yield was a uniform value throughout the model domain. Vertical–to-horizontal anisotropy was correlated highly with the hydraulic conductivity of fine-grained fill and could not be estimated independently so a uniform value of 0.1 was assigned. Specific storage could not be estimated with limited drawdowns and annual pumpage estimates so a uniform specific storage of 1 x 10-6 ft-1 was assigned throughout the basin fill.

Simulated drawdowns matched measured drawdowns with an unweighted RMS error of 0.8 ft (Figure 7). This is a large error relative to regional drawdowns of 4 ft during the 4-year period that showed some biases at wells. For example, wells O4 and O8 consistently simulated drawdowns being +0.9 and -0.8 ft offset from measured drawdowns (Figure 7). These biases likely resulted from errors in locations of pumping wells, pumping rates, and simulating the basin fill with laterally isotropic and homogeneous units.

Agreement between simulated and measured rates of annual decline was good in all wells except well O8 (Figure 7). This suggests that the hydraulic properties of the basin fill can be interpreted reasonably with homogeneous units where the hydraulic diffusivity or transmissivity divided by specific yield is 50,000 ft²/d. The hydraulic diffusivity estimate is relatively constant for annual application rates between 2 and 3 ft.

Hydraulic property estimates were not affected by the absence of alluvium west of the irrigated areas because simulated drawdowns were less than 0.3 ft in the mountain block (Figure 8). The simulated volume of water that was contributed from the mountain block totaled less than 3 percent of the pumped volume.

Transmissivity of the basin fill ranged between 6,000 and 9,000 ft²/d for annual application rates of 2 and 3 ft, respectively (Table 3). Hydraulic conductivity of the coarse-grained fill was more than 20 times greater than that of the fine-grained fill. Specific yield ranged between 0.12 and 0.18 for annual application rates of 2 and 3 ft, respectively. All hydraulic property estimates nearly increased proportionally with increased application rates.


Table 3. Hydraulic properties estimated with the numerical model.
Hydraulic properties estimated with the numerical model.

 

Hydraulic property estimates are integrated values for an area of about 100,000 acres. Simulated drawdowns exceeded 0.5 ft across 140,000 acres at the end of the 2003 irrigation season (Figure 8). About 80 percent of the water pumped during the 2000–03 simulation period was released from storage where drawdowns exceeded 0.5 ft. Transmissivity and specific yield estimates were affected by differences in annual application rates, but the area affected by pumpage did not change.

The contributing area of the Baker aquifer test was much greater than the contributing area of conventional aquifer tests because the pumped volume ranged between 30,000 and 45,000 ac-ft. This is about 1,000 times greater than the volume pumped during a conventional, high-rate aquifer test. For example, continuously pumping 3,000 gpm during a 3-d period removes less than 40 ac-ft from an aquifer.

Simulated and measured drawdowns in the eight observation wells during 2000–03 irrigation seasons with an annual application rate of 3 ft.  Simulated and measured drawdowns in the eight observation wells during 2000–03 irrigation seasons with an annual application rate of 3 ft.  Simulated and measured drawdowns in the eight observation wells during 2000–03 irrigation seasons with an annual application rate of 3 ft.
Figure 7. Simulated and measured drawdowns in the eight observation wells during 2000–03 irrigation seasons with an annual application rate of 3 ft.

 

 

Simulated drawdowns on March 31, 2004 at the end of stress period 8 after recovering from the 2003 irrigation season.  Annual application rate was 3 ft.
Figure 8. Simulated drawdowns on March 31, 2004 at the end of stress period 8 after recovering from the 2003 irrigation season. Annual application rate was 3 ft.

 


 

 

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo USA.gov logo U.S. Department of the Interior | U.S. Geological Survey
URL: http://nevada.usgs.gov/barcass/index.htm
Page Contact Information: Nevada Water Science Center Web Team
Page Last Modified: September 30, 2010 -->