Water-Resources Investigations Report 01-4239
Ground-Water Discharge Determined from Measurements of Evapotranspiration, Other Available Hydrologic Components, and Shallow Water-Level Changes, Oasis Valley, Nye County, Nevada
Previous section: Introduction | Next
section: Ground-Water Levels
Table of Contents | Conversion Factors and Acronyms | Report Home Page
Ground water discharges in or leaves Oasis Valley by means of five major processes: (1) springflow, (2) transpiration by local vegetation, (3) evaporation from soil and open water, (4) subsurface outflow, and (5) withdrawal for local water uses. Of the four natural processes, springflow is the most visible form of discharge. As ground water emerges from the many springs and seeps scattered about Oasis Valley, it either is captured in local marshes and small pools or is channeled into free-flowing drainages. Once at the surface, water evaporates to the atmosphere or infiltrates valley-fill deposits. Little surface water flows out of Oasis Valley except during short periods (lasting less than a month) that follow occasional, intense rainstorms (U.S. Geological Survey, 1993-95).
Most of the spring and surface flow that is not evapotranspired infiltrates the valley-fill deposits and recharges the alluvial aquifer (fig. 5). In addition to this recharge, the alluvial aquifer is recharged from below by diffuse or fault-associated upward flow from the welded-tuff aquifer. Other than the occasional influx of water from rainfall or surface inflow into the valley-fill deposits, these two sources provide most of the recharge maintaining the alluvial aquifer. Although data are limited, rainfall or surface inflow from areas outside the borders of the valley-fill deposits most likely is evaporated before it can recharge the shallow ground-water system, thus it is considered of lesser importance.
Water stored locally within the alluvial aquifer in areas in which the water table is at or near land surface becomes available for use by plants. Evapotranspiration (ET) is a composite term for two processes: (1) the evaporation of water from bare soil or from bodies of surface water, and (2) transpiration, a biological function of plants in which water is released to the atmosphere through the stomata of plant tissue. ET is the primary process by which ground water is removed from the alluvial aquifer. Seasonal changes in ET may be responsible for seasonal fluctuations in the local water table -- generally observed as a declining water table in the summer and fall, and a rising water table in the winter and spring (Laczniak and others, 1999).
Some portion of natural ground-water discharge leaves Oasis Valley as subsurface outflow. This ground water flows through a narrow veneer of the valley-fill alluvium at the southernmost extent of Oasis Valley into adjacent valley-fill deposits in the Amargosa Desert. An additional quantity of ground water is withdrawn from wells and springs in Oasis Valley to satisfy local water supply requirements.
7One method of estimating the natural loss of ground water from Oasis Valley is to estimate the ET from areas of ground-water discharge. An estimate of ET includes water losses from the regional welded-tuff aquifer both by diffuse or preferential, fault-associated upward flow into the alluvial aquifer and by spring and seep flow. The ET estimate includes most spring and seep flow because this water either evaporates or infiltrates the subsurface, recharging the alluvial aquifer where it eventually evapotranspires or leaves Oasis Valley by other discharge processes.
As part of their reconnaissance study of Oasis Valley, Malmberg and Eakin (1962) estimated annual ET to be 2,000 acre-ft. They calculated annual ET as the product of the acreage and the average ET rate of local phreatophytes. Based on studies by W.A. Beetem and R.A. Young, Blankennagel and Weir (1973, p. 21) reported that annual ground-water discharge from Oasis Valley might exceed the Malmberg and Eakin (1962) estimate by a factor of two or more. This discrepancy, combined with results from recent studies (Johnson, 1993; Nichols and others, 1997; Laczniak and others, 1999) suggesting that ET rates for local phreatophytes may be higher than those used by Malmberg and Eakin (1962, p. 25), provided the basis for initiating a study to re-evaluate and more rigorously quantify ET and other ground-water discharge in Oasis Valley. An improved quantification of ground-water discharge would significantly help in formulating an understanding of ground-water flow and aid ongoing development of a flow model for the Death Valley regional ground-water flow system.
The method used to quantify ET from Oasis Valley follows an approach similar to that used by Laczniak and others (1999) to estimate ET from the nearby Ash Meadows discharge area (fig. 1). The approach assumes that total ET can be quantified by summing estimates of annual ET computed for areas of similar plant cover (in terms of type and density) and soil cover (in terms of type and moisture content). These areas of similar vegetation and soil cover here-after are referred to as ET units. Annual ET from each ET unit is computed as the product of the unit's acreage and ET rate.
The major difference between the Malmberg and Eakin (1962) method and the approach used in this study is the set of specific techniques used to identify the major ET units, determine their spatial distribution, and estimate their associated ET rates. Malmberg and Eakin identified and delineated one generalized ET unit from vegetation and soil maps that were constructed using standard field techniques, and estimated an ET rate for this unit from rates determined for similar phreatophytes growing elsewhere in the southwestern United States (Lee, 1912; Robinson, 1958; White, 1932; Young and Blaney, 1942). The technique used in this study refines their approach by incorporating satellite imagery and remote-sensing techniques to better delineate and discriminate ET units, and by determining ET rates for each ET unit using long-term micrometeorological data collected at numerous sites within the Oasis Valley and nearby Ash Meadows discharge areas. In addition, because local vegetation and soil-moisture conditions are largely a consequence of the availability of ground water, water levels were measured to define the depth and seasonal fluctuation of the water table.
ET units were identified and mapped in Oasis Valley through a procedure by which spatial changes in vegetation and soil covers were determined from remotely sensed spectral reflectance data. The procedure discriminates ET units on the basis of spectral similarities identified from Landsat Thematic Mapper (TM) imagery for major vegetation and soil covers within the Oasis Valley discharge area.
TM imagery is acquired by satellites equipped with sensors that measure reflected solar and emitted radiation from the Earth's surface and that scattered from the atmosphere. Measurements are made within seven wavelength bands spanning discrete parts of the visible and infrared regions of the electromagnetic spectrum. Each band is referred to as a TM channel. Six TM channels (1, 2, 3, 4, 5, and 7) measure reflected solar radiation in the visible through short-wave infrared regions (fig. 6). A seventh band, channel 6, which measures thermal energy radiated by the Earth, was not used in this study.
Spectral data received by satellite sensors are transmitted to earth as digital numbers, each denoting the reflectance of the wavelengths across each of the TM channels from a small area of the earth's surface (fig. 6). The surface area scanned by the sensor for TM channels 1, 2, 3, 4, 5, and 7 measures about 100 ft by 100 ft. Each square-shaped area is referred to as a picture element, or pixel; the dimensions of each pixel define the spatial resolution of the imagery. One basic advantage of these digital data is that they can be mathematically manipulated, processed, and analyzed.
Satellite data have long been used to identify and delineate different land covers (Anderson and others, 1976, p. 2; American Society of Photogrammetry, 1983, p. 23-25). Vegetation, water, and soil covers have distinct spectral properties and can be identified by characteristic patterns or signatures defined by their spectral-response curves (fig. 6). A detailed analysis of the shape, slope, and absorption features within a land cover's spectral-response curve often can be used to identify differences in vegetation type, density, and health, as well as differences in soil type and moisture content (Goetz and others, 1983, p. 576-581). Past studies have shown that ET rates throughout the Great Basin region vary with vegetation and soil covers -- in general, the denser and healthier the vegetation or the wetter the soil, the greater the rate of evapotranspiration (Ustin, 1992; Laczniak and others, 1999; Nichols, 2001). The procedure used to identify and map ET units in Oasis Valley takes advantage of this relation and the characteristic patterns in the spectral response of differing vegetation and soil covers, particularly those associated with the evapotranspiration of ground water.
The process of identifying pixels on the basis of patterns in their reflectance spectra is referred to as a classification. If pixels are grouped to represent specific land covers, the classification is called a land-cover classification, and, if grouped to discriminate vegetation, is referred to as a vegetation classification. Whatever the classification type, each different group defines a specific class. The procedure presented here ultimately groups pixels into unique ET units, and is referred to as an ET-unit classification.
The TM data used to classify ET units within the Oasis Valley area was imaged June 13, 1992 (scene identification number LT5040035009216510, fig. 7). The decision to use June 1992 imagery was based on (1) June being a period of high vegetation vigor, (2) 1992 having slightly above-normal precipitation, and (3) the desire for consistency with other recent studies of ET from discharge areas in the Death Valley regional flow system (Laczniak and others, 1999, 2001). Although the procedure used here is similar to that used by Laczniak and others (1999) in the Ash Meadows discharge area, it differs in the number of images and image dates used by the classification process. Laczniak and others (1999) used two TM images, one acquired in June 1992 and the other in September 1992; this study used only the June image. The June imagery was used to represent conditions of near-maximum plant vigor and high moisture and the September imagery to represent conditions of high plant stress (dormancy) and low moisture. Results and insights gained from the Ash Meadows study indicated that a single-date classification would be adequate for discriminating ET units. The imagery was corrected for atmospheric effects using the method described by Chavez (1989).
The first step in the overall procedure reduced the number of pixels used to develop spectral statistics by constraining the area of interest to that of the discharge area (pl. 1). The outer extent of the discharge area was defined using a modified soil-adjusted vegetation index (MSAVI; Qi and others, 1994) developed from the June imagery. The MSAVI uses TM channels 3 and 4 to compute a vegetation index that increases the dynamic range of the vegetation signal by minimizing background influences from the soil. This outer boundary was refined based on information gathered during numerous field visits early in the study. Only pixels within this boundary were classified during the processing.
The classification procedure first identified the different spectral signatures present within the discharge area. The different spectral signatures present within the TM imagery were identified using an un-supervised approach (Lillesand and Kiefer, 1987). This approach identified 188 spectral signatures on the basis of statistical similarities between reflectance values in TM channels 1, 2, 3, 4, 5, and 7. Each signature defined a unique spectral-response curve characterized by statistical variables representing a different set of reflectance values. An example illustrating differences in the spectral signatures of different vegetation and soil covers and their associated response curves as convolved over TM channels 1, 2, 3, 4, 5, and 7 is shown in figure 6.
Next, the procedure associated each pixel within the imagery to one of the identified spectral-response curves. This association was made using the maximum likelihood classification technique (Lillesand and Kiefer, 1987, p. 685-689). This technique compares reflectance values of each pixel against those defining each of the unique spectral-response curves to calculate the statistical probability of a pixel being represented adequately by a spectral-response curve. The procedure assigns each pixel to the spectral-response curve having the greatest statistical probability.
The next step in the procedure was to group spectral-response curves into clusters that best represent the ET units within the discharge area. Each group is referred to as a spectral cluster and can be discriminated by the differences in the characteristic shape defined by the slope and amplitude of the cluster's spectral-reflectance curves. These different shapes result primarily from differences in the amount of spectral absorption and scatter over a particular TM channel.
Spectral-response curves initially were grouped into the seven ET units similar to those delineating areas of ET in the Ash Meadows area (Laczniak and others, 1999, table 3). The placement and groupings of spectral curves were done on the basis of similarities in the statistics defining their reflectance values and on similarities in vegetation and soil conditions noted in the field. Field observations made during many visits to identify actual vegetation and soil conditions within pixels resulted in significant modifications of the groupings initially made based solely on similarities in reflectance values. Ultimately, this dual approach grouped about 130 of the unique spectral signatures into 8 clusters (fig. 8) representing the different vegetation and soil conditions consistent with areas of ground-water ET in Oasis Valley. The 8 clusters given in figure 8 include one additional cluster to those given by Laczniak and others (1999, table 3) for Ash Meadows. This added cluster was included to account for sparse to moderately dense shrubland vegetation not present in the Ash Meadows area. These shrubland communities are dominated by greasewood, rabbitbrush, wolfberry, or some combination thereof (table 1). The eight clusters identify the different ET units within those areas of Oasis Valley dominated by open water, phreatophytes, and moist bare soil (table 1). The remaining 60 or so spectral signatures were associated with pixels falling in areas dominated by sparse upland desert vegetation or in more xeric habitats and were used to discriminate those areas of no substantial ground-water ET. The spectral cluster for this ET unit is inclusive of a variety of spectral-response curves having no identifiable characteristic pattern.
Most spectral-response curves included within a cluster exhibit a similar characteristic shape (fig. 8). The two primary exceptions are the clusters representing ET units 4 (dense meadow and woodland vegetation) and 7 (moist bare soil). Each of these clusters contains spectral-response curves exhibiting two or more distinct characteristic shapes. These multiple patterns result from the inclusion of more than one vegetation or soil type within the cluster. For example, ET unit 4 includes both dense grasses and trees, each of which exhibits a different spectral response (fig. 8). Both spectral responses are included in the cluster because their ET rates are assumed to be similar based on vegetation density. In the case of ET unit 7, its cluster includes multiple soil types usually distinguished by color or wetness. Although spectral responses associated with these soil types vary, they were grouped into one cluster based on the assumption that ET rates are similar.
The final step in the classification procedure was to digitally associate each pixel with an ET unit by assigning a number to each pixel. All pixels outside the boundary of the discharge area and those pixels within the discharge area associated with an area of no substantial ground-water ET were assigned a value of zero. The remaining pixels were assigned a value of 1 through 8 in accordance with their associated ET unit. This process created a raster image of classified ET units. The image was resampled to a finer resolution (60 ft by 60 ft) for consistency with results presented for other major discharge areas in the Death Valley regional flow system (Laczniak and others, 1999, 2001). Lastly, the image was filtered to remove spuriously and anonymously classified pixels. Filtering was performed only on those classes representing ET units 6, 7, and 8 (sparse to moderately dense grassland, moist bare soil, and sparse to moderately dense shrubland, respectively). The filtering process replaced spuriously classified pixels (areas of a few pixels or less) and filled single-pixel gaps by assigning them to the ET unit of their nearest neighbors. This process resulted in a change of less than 3 percent within any of the classified ET units.
The acreage of each ET unit, as computed from the filtered raster image, is listed in table 1. Total ET unit acreage for the Oasis Valley is 3,426 acres (table 1). About 35 percent of this acreage is sparse to moderately dense grassland (SGV) and 26 percent is sparse to moderately dense shrubland (SSV). Denser vegetation types, including dense meadow and woodland vegetation (DMV), moderately dense to dense grassland vegetation (DGV), and dense wetland vegetation (DWV), make up about 35 percent of the total area. Wetter ET units, open water (OWB), submerged and sparse emergent aquatic vegetation (SAV), dense wetland vegetation (DWV), and moist bare soil (MBS), make up less than 5 percent of the total area.
Some difficulty was encountered trying to discriminate between the two grassland ET units, sparse to moderately dense grassland (SGV) and moderately dense to dense grassland (DGV). Laczniak and others (2001, fig. 6 and table 1) also classified two grassland units in the Oasis Valley discharge area but described them as sparse grassland and dense to moderately dense grassland. The major difference between these grassland classifications is in the placement of two spectral-response curves representing a moderately dense grassland cover (fig. 8). Laczniak and others (2001, fig. 6) placed these curves in the cluster representing the denser grassland unit (DGV, ET unit 5, fig. 8). Their placement relied primarily on a single field visit and groupings developed during the Ash Meadows study (Laczniak and others, 1999). After numerous field visits to Oasis Valley were made throughout a 2-year period, the placement of these two curves was deemed most appropriate within the sparser grassland cluster (SGV, ET unit 6, fig. 8). Including these curves in the denser grassland cluster greatly underestimated the sparse grassland acreage observed in Oasis Valley while overestimating that of dense grassland. Placing these curves in the sparser grassland ET unit accounts for a major part of the difference between sparser grassland acreage computed for Oasis Valley by the two studies (1,215 acres by this study, table 1; and 962 acres by Laczniak and others, 2001, table 2).
Another matter of difficulty in the classification was differentiating between the moderately dense to dense grassland (DGV, ET unit 5) and moist bare soil (MBS, ET unit 7) classes. This difficulty is illustrated by similarities in the spectral-response curves within their two respective clusters (clusters 5 and 7, fig. 8). Both ET units have similar soil-moisture characteristics with the primary difference being their vegetation density. The lower density vegetation curves included in the cluster representing ET unit 5 (DGV) are similar to curves included in ET unit 7 (MBS). The somewhat large variation in shape of these curves is attributed to spectral differences resulting from mineralogical differences in the different soil covers. The placement of the more similar curves was determined primarily by conditions observed in the field. Considering the relatively low acreages covered by these two ET units (340 acres by ET unit 5 and 102 acres by ET unit 7) any error in classification would result in a minimal error in the calculation of annual ET.
Total ET unit acreage estimated by Laczniak and others (2001, table 2) was 3,473 acres compared with 3,426 acres estimated in this study (table 1). The small difference in total acreage and other minor differences in ET unit acreage between these two studies are attributed to the more rigorous association of spectral curves and observed field conditions and the filtering method applied in this study. The 3,426 acres classified in this study compares reasonably well with the 3,800 acres of phreatophytes estimated by Malmberg and Eakin (1962, p. 25). The differences in acreage between these two studies could be the result of vegetation changes stemming from increased development, increased local pumpage, or a changing climate but more likely result from differences in delineation methods.
The ET units, as defined and delineated, are not intended to be exact but rather to serve as generalizations of the long-term average vegetation and soil conditions within the discharge area. The accuracy of the final ET-unit classification is difficult to assess because vegetation and soil conditions throughout the Oasis Valley area are not homogeneous, and transitions from one condition to another are not abrupt but rather gradual, often occurring over broad zones. Another factor contributing to the difficulty in assessing the accuracy of mapped ET units is that vegetation and soil conditions can change during the year and from one year to the next. Despite these complications, the accuracy of the classification was assessed.
The overall performance or accuracy of a classification procedure can be described in terms of the percentage of sites correctly classified (Lillesand and Kiefer, 1987, p. 692-694). A correctly classified site is one in which the same ET unit is assigned both through field observation and by the classification procedure. The accuracy of the classification was assessed by evaluating 58 sites. Selected sites typically were within an area of 6 or more pixels of the same ET unit to provide an aerially more consistent depiction of vegetation and soil conditions. Areas used to develop the relations between spectral signatures and ET units were avoided. Access also played a major role in the selection of sites, in that much of the discharge area is in private ownership. Sites were selected to place more emphasis on the ET units having the greatest acreage. Each ET unit was represented by at least one site. Sites instrumented to collect micrometeorological data used to compute ET rates (pl. 1) in Oasis Valley were in-cluded in the assessment. Field observations included a minimum of one visit to examine and document site conditions. Each site was described, photographed, and later evaluated and assigned independently by two individuals to one of the eight ET units. The few discrepancies in assignments were resolved through discussion and site re-visitation.
Results of the accuracy assessment are presented as an error matrix in table 2 (Story and Congalton, 1986). The overall accuracy of the classification is 88 percent (ratio of the number of sites classified correctly to the total number of sites evaluated) and the average accuracy of individual classes is 91 percent. Both of these values are above the acceptability criterion of 85 percent established by Anderson and others (1976, p. 5). Most classification errors are associated with misclassifications between UCL and the sparser vegetation units, SSV and SGV (table 2). The low performance of these two units is attributed primarily to the difficulty in spectrally discriminating upland desert from sparse vegetation. These sparser vegetation classes are similar in that they often are dominated by open desert and therefore have only limited leaf area. Any misclassifications between these three units over the entire discharge area are expected to average out. Most other classification errors can be attributed to assessment sites being located near a transitional zone that only subtly defines the boundary between the three units in question.
Evapotranspiration is a process by which water from the Earth's surface is transferred to the atmosphere. This transfer requires that water change state from a liquid to a vapor, a process that consumes energy. As a result, any change in the rate of water loss by ET is reflected by a change in energy. This relation between water loss and energy consumption is the basis for energy budget methods used to estimate ET rates.
Energy at the surface of the earth can be expressed in terms of an energy budget that balances incoming and outgoing energy fluxes. Assuming negligible energy use by biological processes and limited storage of heat by the plant canopy, the energy budget for conditions typical of Oasis Valley can be expressed mathematically in terms of principal component energy fluxes as:
Rn = H + G + 
E
where
Rn is net radiation (energy per area per time);
H is sensible heat flux (energy per area per time);
G is subsurface heat flux (energy per area per time); and
E is rate of water evaporation (mass per area per time).
Net radiation (Rn) is the principal source of the energy available at the surface of the earth and is the algebraic sum of the incoming and outgoing long- and short-wave radiation. Subsurface heat flux (G) is the rate of change at which heat is stored in the soil or water profile directly beneath the earth's surface. Net radiation and subsurface heat flux can be measured or computed in the field using readily available instrumenta-tion. The difference between Rn and G is the energy available at the earth's surface.
Sensible heat flux (H) is the energy that goes into heating the air and is proportional to the product of the
temperature gradient and the turbulent transfer coefficient for heat. Latent heat flux (E) is the energy consumed for
evapotranspiration and is proportional to the product of the vapor pressure gradient and the turbulent transfer
coefficient for vapor. Neither H nor E
can be determined directly unless the turbulent transfer coefficients are known. Because turbulent transfer coefficients
are difficult to determine, indirect methods have been developed to solve the energy budget. One indirect method
developed by Bowen (1926) uses the ratio between sensible and latent heat flux (H/E). This ratio and the method that uses this ratio to solve the energy
budget are referred to as the Bowen ratio. A detailed derivation of the method and its supporting equations and
parameters are given in Laczniak and others (1999). Using this method, ET can be calculated directly from measurable
micrometeorological data. This method along with the required micrometeorological data provided the primary means by
which ET rates were estimated for the different vegetation and soil environments found in the Oasis Valley discharge
area.
Five sites were selected and instrumented to measure ET. Each site represented an area dominated by a different vegetation type and soil condition. In addition to local vegetation and soil conditions, other factors influencing the selection and location of a site were year-round accessibility, landowner cooperation, and adequate fetch. Generally, fetch (defined as the distance between the sensor and the upwind edge of the environment of interest) implies a homogeneous mix of vegetation types, soils, surface water, or some combination thereof. Sites were located such that the fetch was at least 100 times the height of the highest temperature-humidity sensor (Campbell, 1977). The location and general description of the five sites selected for instrumentation are given in table 3 and plate 1. An additional site at Fairbanks Meadows, originally established as part of a study conducted in Ash Meadows by Laczniak and others (1999), also was maintained as part of this study (table 3).
Each site was equipped with the instrumentation required to measure or compute the micrometeorological data needed to calculate the energy-budget fluxes. Figure 9 presents a schematic showing the typical instrumentation used to determine ET, while figure 10 presents photographs of four actual installations. A typical installation consisted of a net radiometer to measure net radiation, two solid-state air temperature/ humidity probes to measure air temperature and relative humidity, two anemometers to measure windspeed, two infrared temperature transducers to measure soil and plant canopy temperatures, and a set of thermocouples and heat flux plates to compute soil heat flux. Instrument pairs were used to measure vertical difference of a particular variable between two reference heights.
Micrometeorological data required to solve the energy budget by the Bowen ratio method were collected at each ET site for a period of 1 year or more. A minimum period of 1 year was required to evaluate and document seasonal fluctuations in ET rates and compute an annual ET value. Additional years of data were acquired at most sites to better assess annual changes in ET that may result from climatic variations, such as differences between dry and wet years. The period of data acquisition for each instrumented site is given in table 3.
The micrometeorological data collected throughout the study were stored as 20-minute averages computed from measurements made during 10- or 30-second sampling intervals. This collection procedure produced large amounts of data most of which are not presented in this report but are available on request from the USGS's Las Vegas Subdistrict Office. Some gaps occur in the record as a result of instrument failures or the instability of the Bowen ratio (Laczniak and others, 1999). ET values were calculated for each 20-minute period from measured and computed energy fluxes and summed to compute daily ET. Daily values were computed only for days having 68 or more 20-minute computations. Shown in figure 11 are micrometeorological data acquired to solve the energy budget using the Bowen ratio method for the 5-day period, June 7-11, 1997, at the Springdale site (SDALE). Energy-budget fluxes and daily ET calculated from these micrometeorological data are shown in figure 12.
Daily ET calculated by the Bowen ratio method at SDALE for 1997 is shown in figure 13. The minimum calculated daily ET was near zero on Julian day 13 (January 13) and the maximum was nearly 0.29 in. on Julian day 187 (July 6). The mean of the daily ET values is 0.106 in. Annual ET for 1997 was 38.7 in. and was computed by adding the daily ET values. Although the plot of daily ET values shows a general pattern defined by higher rates throughout the late spring and summer months, significant daily variability is apparent. Daily variability is due mainly to short-term changes in weather patterns. Smoothing the annual ET curve using an eighth-order polynomial fitted to daily ET values reduced daily variability, while reasonably maintaining the annual value of ET as calculated directly from the daily values. The smoothed ET curve allows for clear graphical comparisons of ET rates computed at different sites and in different years.
Smoothed ET curves developed from data collected at each of the instrumented ET sites are shown on plate 1. An estimate of the average annual ET at each site was computed by integrating daily ET measured over a 1- or 2-year period and is given in table 3. Estimated average annual ET rates differed among ET units and ranged from 3.14 ft over dense meadow vegetation (SDALE) to 0.62 ft over sparse shrubland vegetation (UOVUP). A graph combining all smoothed ET curves for the period of data collection is shown in figure 14A. Annual precipitation measurements from 1996 to 2000 are compared to the long-term average in figure 14B (National Weather Service, station name: Beatty 8N, station number: 260718-4). Annual precipitation for the 5-year period ranged from 4.7 in. in 1999 to 12.6 in. in 1998 with a long-term average of about 6.3 in. (Desert Research Institute, Western Regional Climate Center, electronic data accessed at <http://www.wrcc.dri.edu/summary/climsmnv.html> on June 17, 2001).
The graph of aggregate ET curves (fig. 14A) shows the spatial and temporal differences in ET computed for five sites in Oasis Valley and one site in Ash Meadows. Individual curves show differences in computed daily and annual ET rates between ET units and between sites located within the same ET unit. The intra-unit differences in ET rates at sites located within SGV and SSV were expected considering that the sites were located to evaluate differences in ET rate in areas of different vegetation density. As would be expected, more densely vegetated areas had the larger ET rate. Although temporally limited, ET rates exhibited daily and annual variations. The largest annual variation occurred in 1998. The high ET rates during 1998 are consistent with the much higher-than-normal precipitation measured that year (fig. 14B) and likely is a response to increased water availability during that year. The ET curve for MOVAL peaks slightly earlier than curves at other sites. The early peak is explained by the site's location along the Amargosa River. This site is inundated by streamflow in the late winter and early spring, whereas other sites have no similar source of water during this period.
An estimate of the mean annual ET from Oasis Valley was computed by summing estimates of the mean annual ET from each of the ET units. ET-unit estimates of the mean annual ET were computed as the product of a unit's acreage and its average ET rate. The average ET rate of an ET unit was determined by averaging all ET rates calculated for sites located within the unit. Site-specific ET rates were calculated from micrometeorological data collected at 5 ET sites in Oasis Valley (table 4) and 9 ET sites in nearby Ash Meadows (Laczniak and others, 1999, table 7). Average ET rates computed from sites in Ash Meadows are considered appropriate for calculating ET rates for ET units in Oasis Valley because vegetation, soil, and meteorological conditions are similar at both locations. A unit having only one ET site within its boundary was assigned an ET rate equal to that of the rate calculated for the lone site. With one exception, the average ET rate of units having two or more sites located within their boundary was computed as the arithmetic mean. The exception was for SSV (sparse shrubland vegetation), where the ET rate was computed as an area-weighted average to reflect the dominance within the unit of the vegetation found at the UOVLO site.Average ET rates for individual ET units range from 1.2 ft/yr for SSV to 8.6 ft/yr for OWB and SAV (table 4). Estimates of mean annual ET range from 8.6 acre-ft at OWB to 2,700 acre-ft at DMV (table 5). The estimate of the mean annual ET from Oasis Valley is 7,800 acre-ft (table 5).
Estimates of mean annual ET include precipitation falling on the area that evaporates or recharges the shallow ground-water flow system and later is evaporated or transpired from within the area. Because the precipitation component of ET is not derived from ground water, it must be removed prior to estimating ground-water discharge. The precipitation component was removed by decreasing the ET rate by an amount equivalent to the average annual precipitation. The remaining ET is assumed to be that derived from ground water. Removing all the average annual precipitation reasonably assumes that no precipitation leaves the Oasis Valley discharge area as runoff during average conditions.
Mean annual precipitation was estimated from bulk precipitation measurements collected during the study and long-term measurements taken at National Weather Service station Beatty 8N. The average annual precipitation determined from long-term measurements (1972-99) was 6.3 in. (electronic data accessed at <http://www.wrcc.dri.edu/summary/climsmnv.html> on June 17, 2001). Based on this average and bulk precipitation measurements, a reasonable estimate of mean annual precipitation for the Oasis Valley area is 6 in. (fig. 14B). Mean annual ground-water ET rates were estimated by subtracting the mean annual precipitation (0.5 ft) from the mean annual ET rate (table 5). As applied, this adjustment assumes that the only source of water other than ground water is the rain falling directly on an ET unit's surface. This assumption discounts as potential sources any water originating from the infiltration of local surface runoff or precipitation falling on the surface of areas of no substantial ground-water ET. Although a limitation, these assumptions are considered reasonable because local surface runoff is minimized by (1) the fractured nature of the volcanic ridges within the area, and (2) low and infrequent rainfall. In addition, limited available data indicate that much of the local surface runoff occurring throughout the region evaporates before entering the discharge area.
Mean annual ground-water ET from Oasis Valley was estimated by summing the mean annual ground-water ET from each ET unit. Mean annual ground-water ET from each ET unit was computed as the product of the unit's acreage and mean average ground-water ET rate. Estimates of mean annual ground-water ET from individual ET units range from 8.1 acre-ft at OWB and SAV to 2,300 acre-ft at DMV (table 5). The estimate of the mean annual ground-water ET from Oasis Valley is 6,000 acre-ft (table 5).
The estimate of mean annual ground-water ET differs by a factor of 3 from that of Malmberg and Eakin (1962, p. 25). Their estimate of 2,000 acre-ft assumes that there are 3,800 acres of phreatophytes in Oasis Valley and an average ET rate of 0.5 ft/yr, whereas the 6,000 acre-ft estimated in this study assumes that there are 3,426 acres of phreatophytes and moist bare soil and an average ET rate of 1.7 ft/yr (table 5). There is a difference of about 10 percent in the estimated acreages. The primary discrepancy between the two estimates, however, is the result of the difference in the estimated average ET rate. Although the accuracy of one rate estimate versus the other is difficult to evaluate, the more localized nature of the data and more rigorous method used in this study are likely to result in a more accurate estimate of the ET rate.
The accuracy of the estimate of ground-water discharge via ET is limited by the assumptions inherent in the classification procedure and the energy-budget method (Bowen ratio) used to compute daily ET. The classification procedure identified 3,426 acres of Oasis Valley as an area from which ground water is being lost by evapotranspiration. The remaining portion of Oasis Valley is assumed to be an area of no substantial ground-water loss. This assumption, although strongly supported by this area's lack of vegetation, dryness of soil, and greater depths to the water table, could result in some error in the estimate of ground-water discharge by ET. Although the remaining portion of the valley is large (about 30,000 acres), the rate of ground-water discharge by ET is likely less than 0.01 ft/yr (Andraski, 1997, p. 1913), thus the volumetric loss would be minimal.
ET-unit acreage was delineated on the basis of TM imagery acquired in 1992. Precipitation data reported for nearby weather stations indicate that rainfall for 1992 was slightly above normal, a level that may have produced healthier vegetation and moister soils. Classifying ET units on the basis of multiple years of imagery would likely result in acreage estimates more representative of the long-term average.
Other limitations include (1) the assumption that all springflow is ultimately evaporated or transpired from within the bounds of one of the delineated ET units; (2) the short-term nature of the data used to compute mean values; (3) the limited number of sites used to estimate ET from each ET unit, (4) the uncertainty in the adjustment applied to remove precipitation from ET estimates, and (5) local ground-water recharge from areas outside ET unit boundaries. The mean annual ET estimates of each ET unit (table 5) were computed from Oasis Valley and Ash Meadows data typically acquired over a period of 2 or more years. Although the period of data collection included years of varying climatological conditions, variations are fairly small in the annual ET rates computed from one year to the next and between sites within the same ET unit. ET estimates determined from longer-term data and additional ET-site installations would help refine, improve, and provide more confidence in any estimate of mean annual ground-water discharge.
The adjustment applied to remove precipitation from estimated ground-water discharge discounts as potential recharge sources (1) water originating from infiltration of local surface runoff, or (2) precipitation falling on the surface of areas of no substantial ground-water ET. This infiltration probably is minimal and most likely evaporates before entering the shallow ground-water system. Any amount of infiltration that does flow into areas of ground-water discharge may be balanced by any ground water that flows out of areas of substantial ground-water discharge into unclassified areas. The lack of available data about infiltration limits our ability to adjust ET rates for it and may result in larger estimated ground-water discharge values. However, these estimated values may still be compared with previous ground-water discharge estimates in Oasis Valley that also were not adjusted for these infiltration processes (Malmberg and Eakin, 1962).
At the southern and western boundaries of the Oasis Valley discharge area, the alluvial and welded-tuff aquifers thin and pinch out against less-permeable Paleozoic and Precambrian siliciclastic rocks (fig. 2). The boundary between these aquifers and less-permeable rocks forces ground water to flow either toward land surface, where it is evapotranspired, or out of Oasis Valley, where it is termed "subsurface outflow." The most likely pathway for subsurface outflow is through the alluvial aquifer via the alluvium-filled channel of the Amargosa River at the Amargosa Narrows (figs. 3 and 15). This subsurface outflow would move southward into the Amargosa Desert. Other potential but less likely pathways for subsurface outflow are to the southeast across a ridge separating Oasis Valley from Crater Flat (Fridrich and others, 1999) and to the south beneath the Bullfrog Hills. Flow through the basement confining unit beneath the alluvial aquifer is considered negligible considering its very low permeability (Fridrich and others, 1999). The range of hydraulic conductivity values of this basement confining unit probably is three to seven magnitudes of order less than that of the alluvial aquifer (Winograd and Thordarson, 1975; Bedinger and others, 1989).
Assuming that all subsurface flow occurs through the alluvium within the Amargosa River channel and knowing the hydraulic gradient, cross-sectional geometry, and hydraulic conductivity of the channel fill material, an estimate of subsurface outflow can be calculated using Darcy's law.
Darcy's law as modified by Heath (1989, p. 12) can be expressed as:
Q = 0.0084 K A (dh/dl),
where
Q is quantity of ground water flow, in acre-feet per year;
K is hydraulic conductivity, in feet per day;
A is cross-sectional area through which flow occurs, perpendicular to the direction of flow, in square feet;
(dh/dl) is the hydraulic gradient, in feet per foot; and
0.0084 is the factor to convert cubic feet per day into acre-feet per year.
The hydraulic gradient through the Amargosa Narrows area was calculated from depth-to-water measurements and spring altitudes in the vicinity of Amargosa Narrows (see "Spring Discharge" section; plate 2). Based upon the calculated gradient, ground-water flow generally is to the south. The gradient changes as ground water travels from Beatty south-southeast to the Amargosa Narrows. Gradients are about 0.010 ft/ft at section B-B', 0.017 ft/ft near the mouth of the Amargosa Narrows, and 0.0044 ft/ft south of the Amargosa Narrows. Although some seasonal fluctuations in the water table occur, the effect on the hydraulic gradient is negligible.
Geophysical information, including data from seismic refraction and downhole geophysical logging (D.L. Berger and A.R. Robledo, U.S. Geological Survey, written commun., 1999), geologic mapping, and lithologic logs, was used to estimate the thickness and cross-sectional area of the alluvium-filled channel at the Amargosa Narrows area. One in-line and three transverse cross-sections at the Amargosa Narrows illustrate the variability in the thickness and cross-sectional area of the alluvium (fig. 15). The cross-sectional area of saturated alluvium is approximately 230,000 ft2 at cross-section B-B' north of the Amargosa Narrows, 88,000 ft2 at cross-section C-C' at the northern entrance of the Amargosa Narrows, and 248,000 ft2 at cross-section D-D' south of the Amargosa Narrows. The alluvial aquifer at the Amargosa Narrows is somewhat funnel-shaped in that its width decreases and thickness increases approaching the Amargosa Narrows area. At Amargosa Narrows, the aquifer's width is at its narrowest; upon exiting both its width and thickness increase.
Results of aquifer tests conducted in the Death Valley regional ground-water flow system in basin-fill deposits similar to those found at the Amargosa Narrows were used to estimate the hydraulic conductivity of the alluvium (W.A. Belcher, U.S. Geological Survey, oral commun., 2001). Based on these tests, a range defined by two standard deviations from the geometric mean of hydraulic conductivity is 2 to 10 ft/day. The range may be biased toward higher values because most wells tested were constructed for ground-water production and may not best represent common basin-fill deposits.
Substituting a hydraulic gradient of 0.017 ft/ft, a cross-sectional area of 88,000 ft2, and the range of 2 to 10 ft/day for hydraulic conductivity into Darcy's law results in a computed outflow that ranges between 30 and 130 acre-ft/yr. This subsurface outflow estimate is limited by the accuracy of the calculated or estimated hydraulic gradient, cross-sectional area, and hydraulic conductivity. The estimate is less than the previous estimate of 400 acre-ft/yr (Malmberg and Eakin, 1962), but still supports the concept of limited subsurface outflow when compared to total ground-water discharge in Oasis Valley. The difference between these estimates probably is related in part to the more rigorous quantification of parameters affecting subsurface outflow applied in this study. Additional test boreholes in the alluvium, seismic-refraction surveys, and geophysical borehole logging are needed to better quantify the parameters needed to estimate subsurface outflow from Oasis Valley.
Ground water is withdrawn from wells scattered throughout Oasis Valley. The largest single user of ground water is the Beatty Water and Sanitation District (BWSD), which supplies water to most homes and businesses within the town of Beatty. Outside of Beatty, springs and non-municipal wells supply most of the homes and ranches in Oasis Valley with water for irrigation, livestock, and domestic uses.
The BWSD pumps ground water from seven wells. Location, construction, and open-interval data for these wells are given in table 61. Six of these wells are in Oasis Valley -- Beatty Middle, Summit, and Upper Indian wells are in the Bullfrog Hills, and Beatty wells 1, 2, and 3 are in the town of Beatty. A seventh well, EW-4, is southwest of Beatty in the Amargosa Desert. Table 7 gives monthly and annual ground-water withdrawals by well for 1996 to 1999 (J.C. Weeks, BWSD, written commun., 2000).
Total annual ground-water withdrawal from the six BWSD wells in Oasis Valley declined from 410 acre-ft in 1996 to 179 acre-ft in 1999 (fig. 16). Water pumped from outside Oasis Valley at well EW-4 compensated for much of this decrease, but overall withdrawal still declined. Production from this well started in the fall of 1997 and increased from 115 acre-ft in 1998 to 155 acre-ft in 1999 (table 7).
Ground-water withdrawal by the BWSD varies in response to seasonal water demands (fig. 17). The largest demands occur in summer (July through September) and the smallest in winter (January through March). The seasonal variation in ground-water withdrawal from Oasis Valley has lessened (fig. 17) since 1997. Although demands continued to vary seasonally in 1998 and 1999, the larger summer needs were met primarily by water pumped from well EW-4 outside of Oasis Valley.
Water-level measurements indicate that the reduction in municipal ground-water withdrawal from Oasis Valley may have affected local water levels. Water levels measured in the Lower Indian Springs Well in Bullfrog Hills have risen since February 1999 (pl. 2). This rise is consistent with decreased ground-water withdrawal from the Beatty Upper Indian and Beatty Middle wells (table 7).
Field reconnaissance and Nevada Division of Water Resources drilling records identified approximately 15 springs and 20 non-municipal wells that supply water to individual homes and ranches in Oasis Valley. A reasonable estimate of annual ground-water withdrawal consumed from each of these sources is 1 acre-ft (Coache, 1999). Based on this consumption rate and the number of supply sources, a reasonable estimate of the annual non-municipal use of ground water from Oasis Valley is 35 acre-ft. Estimates of the total annual ground-water withdrawal from Oasis Valley, computed by combining municipal and non-municipal estimates, are 440 acre-ft in 1996 and 210 acre-ft in 1999. Monitoring non-municipal water consumption would improve the accuracy of the estimate.
Spring discharge was measured periodically at sites throughout Oasis Valley from 1996 to 1999. Measurement sites were distributed geographically throughout the valley (pl. 2; table 8). About half the sites were located where springflow could be measured near or at a single spring orifice. Measuring locations for these sites, referred to as spring sites, were natural outlet channels or man-made outflows such as plastic or metal pipe draining the springhead or pool. Other springs and seeps were more difficult to measure because the discharge is diffuse. Discharge from these springs and seeps was measured at a more downgradient location, referred to as a channel site, where flow converged and channelized. Typically these channel flow measurements included contributions from a combination of nearby springs, seeps, and local shallow ground-water inflow. Evapotranspiration and other ground-water recharge may occur upstream from these sites.
Periodic spring-discharge measurements were made monthly from November 1996 to September 1997, quarterly from October 1997 to September 1998, and semi-annually from October 1998 to September 1999. On occasion, a site could not be measured because of difficulties in accessing the site. More frequent measurements were made at selected channel sites to better evaluate seasonal and annual changes in discharge. Table 9 gives the maximum and minimum measured discharge and the magnitude of discharge fluctuation at each site for each year of data collection. Measurements may not be indicative of the actual minimum or maximum discharge due to their periodic nature. Local precipitation affects channel site measurements but has little or no effect on spring site measurements (table 9).
Differences between discharge measurements at spring and channel sites are evident by comparing values in table 9 and figure 18. The annual maximum discharge at channel sites typically occurs in winter or early spring (January to April), coincident with minimum ET and maximum seasonal precipitation, whereas the annual minimum occurs in late spring through early fall (April to September), coincident with increasing or maximum ET. At spring sites, the timing of annual maximums and minimums is not as consistent. The annual fluctuation in discharge at channel sites is larger and more variable than at spring sites. The annual fluctuation typically was greater than 40 gal/min at channel sites, and less than 10 gal/min at spring sites.
Larger annual fluctuations observed at channel sites are attributed primarily to seasonal changes in ET by riparian vegetation. The small annual fluctuation in discharge at spring sites measured at or near spring orifices in bedrock indicates that regional springflow is nearly constant. Discharge measured at channel sites, which typically receive contributions from multiple springs and seeps issuing from valley-fill and regional springflow, decreases as ET increases and ambient soil moisture decreases. Although channel flow decreases, it is uncertain whether springflow from valley-fill deposits decreases in response to decreasing regional inflow, or because riparian vegetation is transpiring water from areas adjacent to the channel. Assuming constant regional inflow, decreased channel flow is attributed to increased ET.
Estimated annual discharge from springs in Oasis Valley, calculated from published estimates and measurements of springflow (Thordarson and Robinson, 1971; White, 1979), is approximately 3,000 acre-ft. This estimate excludes flow from numerous seeps or springs where discharge measurements were impractical or unavailable. This estimated ground-water discharge from springs is 3,000 acre-ft less than estimated ground-water discharge from ET (see "Estimates of Annual Evapotranspiration" section). Differences are attributed to the exclusion of non-measurable springs and seeps and to diffuse and fault-associated upflow into the alluvial aquifer from the underlying welded-tuff aquifer.
An estimate of annual ground-water discharge from Oasis Valley was computed by summing estimates of the mean annual ground-water ET, subsurface outflow, and ground-water withdrawal. Although seep and spring discharge are not considered directly in the estimate, they are indirectly accounted for in the estimate of ET. Most spring and seep flow evaporates or recycles back into the shallow ground-water flow system where it later is evaporated or is transpired by the local vegetation. The approach used to estimate ground-water discharge assumes that all spring and seep flow is evaporated or transpired locally, or is withdrawn by municipal or non-municipal wells within Oasis Valley. The ET estimate also includes any upward leakage (diffuse or fault-associated upflow) of water from the underlying welded-tuff aquifer into the shallow ground-water flow system.
Estimated mean annual natural ground-water discharge from Oasis Valley is 6,100 acre-ft of which 6,000 acre-ft is ground-water ET and about 80 acre-ft, the mean of the estimated range, is subsurface out- flow (table 10). Total estimated ground-water discharge, which includes both natural ground-water discharge and ground-water withdrawal, ranged from 6,500 acre-ft in 1996 to 6,300 acre-ft in 1999. When combined, subsurface outflow and ground-water withdrawal account for less than 10 percent of the total ground-water discharge -- the remainder being attributable to ET.
1 Well EW-4 is not shown in the figures or on the plates that accompany this report.
Previous section: Introduction | Next
section: Ground-Water Levels
Table of Contents | Conversion
Factors and Acronyms | Report Home Page