Examining the impact of shallow groundwater on evapotranspiration from uncultivated land in Colorado's Lower Arkansas River Valley

Abstract

The Lower Arkansas River Valley (LARV) in Southeastern Colorado is an important agricultural region of the state. However, more than a century of intensive irrigation has raised the water table under the area causing agricultural and environmental problems including water logging, soil salinization, and the leaching of salts and selenium into waterways. These issues could be mitigated by improving irrigation practices, lining irrigation canals, installing subsurface drainage systems, and implementing other strategies that lower the water table. A lower water table is expected to increase crop productivity in some areas, and it is also likely to reduce non-beneficial evapotranspiration (ET) from uncultivated lands that are interspersed with the irrigated fields. Various studies suggest that a lower water table tends to result in lower ET rates, but a quantitative analysis is lacking for conditions that are representative of the uncultivated lands in the LARV. The overarching goal of this research is to determine the contribution of groundwater upflux to the total ET for uncultivated areas in the LARV and to quantify the dependence of ET on water table depth as well as other site properties. Three field sites in the LARV were selected for detailed study. One site is a retired field near the Arkansas River that has a shallow water table due to nearby irrigation. The second site is a naturally vegetated area at the edge of the alluvial valley that has a shallow water table due to its proximity to an irrigation canal. The third site is another naturally vegetated field that lies between an irrigation canal and a creek. These sites are mainly vegetated by grasses and forbs. ET was estimated at a 30 m x 30 m resolution from Landsat satellite imagery using the Remote Sensing of Evapotranspiration (ReSET) method. This method also produces estimates of the Normalized Difference Vegetation Index (NDVI), which is a measure of vegetation greenness, at the same resolution. Water table depths were measured at a total of 84 monitoring wells divided between the three sites. Other variables were measured in the field including precipitation, gravimetric soil moisture, and soil salinity. Clear relationships between ET and water table depth at one of the sites suggest that ET rates decrease with deeper water tables, but ET rates can remain high even when the water table is relatively deep (greater than 2 m). Water balance analyses indicate that groundwater upflux contributes between 80% and 86% of the total ET at these sites. Further study is needed to quantify the water savings that might be achievable for the region as a whole if the water table was lowered through engineering intervention.