Abstract

Evapotranspiration (ET) in riparian areas is a poorly understood component of the regional water balance in the Platte River Basin, where competing demands have resulted in water shortages in the ground-water/surface-water system. From April 2002 through March 2006, the U.S. Geological Survey, Nebraska Platte River Cooperative Hydrology Study Group, and Central Platte Natural Resources District conducted a micrometeorological study of water and energy balances at two sites in central Nebraska near Odessa and Gothenburg to improve understanding of ET rates and factors affecting them in Platte River riparian forests. A secondary objective of the study was to constrain estimates of ground-water use by riparian vegetation to satisfy ET consumptive demands, a useful input to regional ground-water flow models.

Both study sites are located on large islands within the Platte River characterized by a cottonwood-dominated forest canopy on primarily sandy alluvium. Although both sites are typical of riparian forests along the Platte River in Nebraska, the Odessa understory is dominated by deciduous shrubs, whereas the Gothenburg understory is dominated by eastern redcedars. Additionally, seasonal ground-water levels fluctuated more at Odessa than at Gothenburg. The study period of April 2002 through March 2006 encompassed precipitation conditions ranging from dry to wet.

This study characterized the components of the water balance in the riparian zone of each site. ET was evaluated from eddy-covariance sensors installed on towers above the forest canopy at a height of 26.1 meters. Precipitation was measured both above and below the forest canopy. A series of sensors measured soil-moisture availability within the unsaturated zone in two different vertical profiles at each site. Changes in ground-water altitude were evaluated from piezometers. The areal footprint represented in the water balance extended up to 800 meters from each tower.

During the study, ET was less variable than precipitation. Annual ET fluctuated about 7 percent from the 4-year mean, ranging from about 514 to 586 millimeters per year (551 on average) at the Odessa site and 535 to 616 millimeters per year (575 on average) at the Gothenburg site. Conversely, annual precipitation fluctuated by about 35 percent from the 4-year mean, ranging from 429 to 844 millimeters per year at Odessa and 359 to 791 millimeters per year at Gothenburg. Of this precipitation, 14 to 15 percent was intercepted by the forest canopy before it could infiltrate into the soil.

For the 4-year period, annual ground-water recharge from the riparian measurement zone averaged 76 and 13 millimeters at Odessa and Gothenburg, respectively, to satisfy the water balance at each site. This indicates that, from an annual perspective, ground-water reductions caused by ET may be minimal. This effect varied somewhat and primarily was affected by fluctuations in precipitation. Ground-water discharge occurred during the driest study year (2002), whereas ground-water recharge occurred from 2003 to 2005. These results do not exclude ground water as an important source of water to riparian vegetation—especially to phreatophytes that have the capability of directly using water from the saturated zone—during periods of high ET in the summer, particularly during periods of lower than normal precipitation. However, the calculations indicate that, on an annual (or longer) net-flux basis, ground-water use by riparian forests is likely to be balanced by periods of recharge from excess precipitation at other times of the year. In contrast to more arid settings, where scientific literature indicates that ground water may supply a large fraction of the water used for ET by riparian vegetation, precipitation along the Platte River of Nebraska was great enough—and generally greater than ET—that most or all of the annual ET demand was satisfied by available precipitation.

Crop coefficients developed for 15-day and monthly periods from the measured data predicted ET within 3.5 percent of actual annual ET; however, daily ET was underpredicted on days of increased ET and overpredicted on days of low ET. These crop coefficients can be used to extrapolate riparian-forest ET along the Platte River in conjunction with atmospheric data from other climate stations in central Nebraska.

Regression models of simple and multiple-linear relations between explanatory variables and ET indicated that the relation of ET to environmental factors was different on days with precipitation than on dry days. At Odessa, ET was affected by vapor-pressure deficit, solar radiation, leaf-area index, and depth to water regardless of precipitation conditions, but was also affected by air temperature on days without precipitation, suggesting energy limitations on ET on days without precipitation. At Gothenburg, ET was always a function of vapor-pressure deficit, solar radiation, and leaf-area index, but, as with Odessa, air temperature also became important on days without precipitation.

Despite depths to ground water of less than 2 meters and phreatophytic vegetation, measured ET was substantially less than potential ET (based on the modified Penman method), consistent with plant-stomatal regulation of ET in response to environmental and meteorological factors. Although annual ET rates generally were similar, the two sites exhibited different intraannual soil-moisture regimes that had a corresponding effect on ET and vegetation vigor. Smaller seasonal declines in ground-water levels and a lack of understory shrubs at the Gothenburg site as compared to the Odessa site may explain why Gothenburg ET was comparatively greater later in the summer and was not dependent on depth to water (as identified by the multiple-linear regression model). These differences also may explain why, during years of increased precipitation, ET rates increased at Odessa but not at Gothenburg.

Posted June 8, 2009