Water-Resources Investigations Report 01-4234
Estimates of Evapotranspiration from the Ruby Lake National Wildlife Refuge Area, Ruby Valley, Northeastern Nevada, May 1999-October 2000
Micrometeorological data used in estimating the energy budget were collected at nine sites that represented five of the most aerially extensive habitats in the refuge (table 1). The source and amount of water consumed by ET, in part, is a function of the conditions at each site. Daily ET rates, computed by summing ET calculations made for each 20-minute period, are given in appendices 1-5. The period of data collection at Bowen-ratio sites began in early summer 1999 and ended in November 2000. Data at eddy-correlation sites were acquired at different times during the summer of 2000.
The open-water and bulrush-marsh sites were selected primarily to measure the amount of water consumed by ET in the wetland area (fig. 2). The Bowen-ratio method was used to estimate daily ET for more than 540 consecutive days at both sites (apps. 1 and 2). Ruby Lake is the primary source of water for ET in the wetland. Water within the lake is derived principally from springs discharging along the west and southwest side of the refuge and beneath the lake, and from precipitation that falls directly on the lake.
The open-water site was located in the extreme southern part of the South Marsh (figs. 2 and 5A; table 1). Historically, the South Marsh has remained at least partially flooded during prolonged dry periods while other water bodies in the refuge desiccate. In June 1985, open water covered about 1,030 acres in the South Marsh (Nichols, 2000, C17). During this study, the water level at the open-water site initially was 4.3 ft in May 1999, but fell 1.5 ft by September 1999. During the winter (October 1999-April 2000) the water level rose about 1.0 ft, but dropped 2.3 ft by September 2000. The bulrush-marsh site was located in a moderately dense stand of bulrush (Scirpus robustus) with scattered cattails (Typha spp.) in the southern part of the North Marsh (figs. 2 and 5B; table 1). The initial water level at the bulrush-marsh site was about 3.0 ft in May 1999 and declined by 1.6 ft in September 1999. In March 2000 the water level was about 2.8 ft and dropped 1.6 ft by September 2000. The extent of the wetland area decreased by about 4,500 acres between March and September 2000.
In general, surface-water levels throughout the wetland were 1.0 ft lower in the summer of 2000 (May-September) than in the summer of 1999 (USFWS, written commun., 2000). This decline in water level is largely due to the smaller amount of precipitation during the 1999-2000 winter (October 1999-April 2000) than during the preceding winter (October 1998-April 1999; fig. 3).
Meadows along the western and southern parts of the refuge are found in association with springs and areas of frequent flooding from rising ground water. Eddy-correlation instrumentation was set up in a meadow in the southern part of the refuge, less than 1/4 mi from the South Marsh (figs. 2 and 6A; table 1). Daily ET at the meadow site was computed from data collected continuously from May 26 through August 29 except for 19 days at the end of July when the data logger malfunctioned (app. 5). Plants at the meadow site consist primarily of sedges (Carex spp.), rushes (Juncus spp.), and some grasses and herbaceous species. Depth to ground water beneath the site was estimated to be less than 2 ft during the period of data collection and the soil generally was moist. Although the meadow site has been subject to periodic flooding in years of above-average precipitation, the site was not flooded during this study.
Mixed phreatophytic shrubs and associated areas of bare soil are found in a broad expanse along the east, northeast, and southeast sides of the refuge that is not subject to flooding (fig. 7). ET typically exceeds seasonal precipitation in these areas because the plants have access to ground water for transpiration. Five sites were selected to estimate ET from various mixtures of phreatophytic shrubs using both the Bowen-ratio and eddy-correlation methods (fig. 2; table 1). The plant species of interest at the phreatophyte sites include saltgrass (Distichlis stricta), rubber rabbitbrush (Chrysothamnus nauseosus), basin wildrye (Elymus cinereus), greasewood (Sarcobatus vermiculatus) and big sagebrush (Artemesia tridentada spp. tridentada). At each phreatophyte site about 30 to 35 percent of the area was vegetated and the remaining area was bare soil. During data collection the depth to ground water at the five phreatophyte sites ranged from less than 5 ft to nearly 20 ft (table 1). Ground-water levels measured in wells near the phreatophyte sites dropped on average about 2.4 ft between March and September 2000.
Bowen-ratio instrumentation was set up at the phreatophyte-1 site (fig. 7A) and daily ET was computed continuously for 502 days (app. 3). Data were collected at the remaining four phreatophyte sites using two similar sets of eddy-correlation instrumentation (app. 5). In late August, the eddy-correlation instrumentation initially set up at the meadow site was moved to the northeastern part of the refuge (phreatophyte-2) until mid-September; data was collected there for 20 days. The second set of eddy-correlation instrumentation began collecting data in late May at the phreatophyte-3 site, then was moved in late July to the phreatophyte-4 site in the same area as the phreatophyte-1 site to compare daily ET rates with those computed using the Bowen-ratio method. The second set of eddy-correlation instrument again was moved in late August to the phreatophyte-5 site (fig. 7B), where it remained until mid-September, providing daily ET rates for 19 days.
The desert-shrub upland habitat occupies the higher areas, mostly on the west and northeast sides of the refuge where the depth to ground water is too great to support phreatophytes (fig. 2). In the desert-shrub upland, the source of water for ET is soil moisture derived from precipitation. The depletion of soil moisture typically is equal to ET when precipitation does not occur. The desert-shrub upland site was located in the southwestern part of the refuge on a piedmont slope at an altitude of about 6,080 ft (fig. 6B). The Bowen-ratio method was used and daily ET was computed for 512 consecutive days. Dominant plants include black sagebrush (Artemisia nova) and green rabbitbrush (Chrysothamnus viscidiflorus). The soils at this site were dry and depth to ground water probably was greater than 80 ft. Although ET in the desert-shrub upland area is minor compared to the wetter habitats, it significantly reduces the amount of annual precipitation available for deep percolation and ground-water recharge.
Daily and monthly ET rates computed at the four Bowen-ratio sites are presented graphically in figures 8-11. Fluctuations in daily and monthly estimated ET are much more pronounced at the open-water and bulrush-marsh sites (figs. 8 and 9) than at the phreatophyte-1 and desert-shrub upland sites (figs. 10 and 11). Daily fluctuations in ET at a given site are caused by changes in cloud cover and other short-term changes in weather patterns. Differences in ET estimates among sites are, in part, a function of the spatial and temporal differences in the availability of water for ET. The annual variability in daily estimates of ET at the Bowen-ratio sites during the 2000 water year (October 1999-September 2000) are given in table 2.
The lowest average rates of daily ET among the Bowen-ratio sites during the 2000 water year were estimated at the phreatophyte-1 and desert-shrub up-land sites. In comparison, the average daily ET rates estimated at the open-water and bulrush-marsh sites, where standing water was continuously available for evaporation, were about four to five times greater (table 2). Daily ET at the phreatophyte-1 and desert-shrub upland sites ranged from less than 0.010 in/d during the winter to a maximum of about 0.146 in/d and 0.160 in/d, respectively, in May (apps. 3 and 4). At the open-water and bulrush-marsh sites minimum daily ET rates also were less than 0.010 in/d in the winter, but maximum rates of 0.464 in/d and 0.396 in/d, respectively, occurred in July. The timing of the maximum daily ET rates at the phreatophyte-1 and desert-shrub upland sites reflects the above-average precipitation that occurred during May, which was preceded by two months of below-average precipitation (fig. 3).
A comparison of daily average ET rates (fig. 12) and monthly totals (figs. 8B and 9B) for the open-water and bulrush-marsh sites shows that ET increased more rapidly from February through April at the open-water site than at the bulrush-marsh site. This less-rapid increase in ET at the bulrush-marsh site is attributed to shading by dead plant material from previous years. Shading can reduce evaporative losses by partitioning energy to sensible heat at the expense of latent heat (Bidlake, 2000, p. 1315). Shading effects also are apparent in the comparison of winter to summer ET (table 2). The winter ET estimate at the open-water site (23.85 in.) is almost twice that at the bulrush-marsh site (13.18 in.); however, ET is similar at the two sites during summer. Laczniak and others (1999, p. 33) suggest that, in vegetated areas, shading by dead vegetation reduces winter evaporation by maintaining relative humidity near saturation and decreasing air exchange. Shading also is somewhat decreased in the summer by the higher angle of the sun, allowing direct evaporation from the water to make up more of the summer ET.
Temporal differences in water source and availability also appear to cause variations in daily ET between those sites where plants rely solely on soil moisture and those that use a combination of soil moisture and ground water. Beginning in November 1999 and up through May 2000 estimates of average daily ET were similar at the phreatophyte-1 and desert-shrub upland sites (fig. 12). Average daily ET nearly doubled at both sites in May as a result of increased soil moisture from precipitation (fig. 13). As the moisture content in the shallow soils decreased following the May precipitation, average daily ET at the upland site also decreased while average daily ET at the phreatophyte-1 site reached a summer maximum in June (fig. 12). At phreatophyte-1, where the water table is shallow, plants are able to use ground water directly to supplement the soil moisture available for transpiration. Daily ET at the phreatophyte-1 site remained relatively constant through July and gradually de-creased as available energy decreased (fig. 14).
Although daily ET data estimated at the eddy- correlation sites are limited in duration, some general statements and comparisons can be made. Total ET estimated at the meadow site for a span of 84 days, about 55 percent of the summer period, was 10.73 in., more than two-thirds of the annual ET of 15.89 in. estimated at the phreatophyte-1 site (table 2). The daily energy budget closure at the meadow site, calculated as the difference between available energy and turbulent fluxes, averaged 23.9 W/m2 (table 3). This residual suggests that about 17.2 percent of the available energy at the meadow site was not accounted for by measurements of the turbulent fluxes and the imbalance is due either to measurement or computational error.
The remaining eddy-correlation sites provided daily estimates of ET for periods of 19 to 51 days from areas with various mixtures of phreatophytic shrubs (table 1). In a comparison of average daily ET computed at the phreatophyte sites, the eddy-correlation method generally provided a lower value for ET than was provided by the Bowen-ratio method (fig. 15). On corresponding days, average daily ET at the phreatophyte-2 site was 0.012 in/d less than at the phreatophyte-1 site. At the phreatophyte-3 site, ET was 0.021 in/d less than at phreatophyte-1. At the phreatophyte-5 site, ET was 0.009 in/d less than at phreatophyte-1.
The best correlation between methods was at the phreatophyte-1 and phreatophyte-4 sites; on corresponding days, ET computed at the phreatophyte-4 site was only 0.001 more than at phreatophyte-1. Although the two sites were about 100 ft apart, the average difference in available energy was 3 percent, suggesting that similar available-energy conditions existed at both sites. Energy-budget closure at phreatophyte-4 was 20.1 W/m2 with a 16.3-percent relative closure (table 3). The average energy-budget closure for the remaining three eddy-correlation sites ranged from 33.1 W/m2 computed at phreatophyte-5 to 61.4 W/m2 computed at phreatophyte-2. The energy-budget closure generally was positive, indicating that either available energy was overestimated or turbulent fluxes were underestimated (Sumner, 1996, p. 18).
ET rates presented represent total ET at a given site and as such include the volume of precipitation that fell during the data-collection period at each site that was consumed by plants. ET rates estimated for one particular habitat are assumed, in this study, to be representative of ET rates in similar habitats throughout the refuge. The uncertainty in ET computed by the Bowen-ratio method is a composite of errors introduced in measuring net radiation and subsurface-heat flux, and in measuring air temperature and relative humidity at two heights. One potential source of error is the instrumentation used to measure the variables needed to compute the flux components. Tomlinson (1995, p. 15) suggests, based on instrument error analysis, that about a 12-percent change in the final ET estimate would be expected if all instruments varied by a maximum amount. Analysis of energy-budget closure data computed at the eddy-correlation sites (table 3) indicates that the measured turbulent fluxes were not sufficient to account for the measured available energy. Although there are techniques to account for these discrepancies in the energy budget (Bidlake and others, 1993; Sumner, 1996, p. 13), they were not applied to the eddy-correlation-estimated ET in this study.
Previous Section: Methods of Estimating Evapotranspiration
Next Section: Annual Evapotranspiration in Ruby Lake National Wildlife Refuge
Table of Contents | Home Page