Quantifying the Effect of Thinning Vegetation on Evapotranspiration in a Mountainous Watershed through Remote Sensing: Improving Water Balance Estimates for Managed Aquifer Recharge

Abstract:

A long-term water balance study in an experimental watershed of the Sacramento Mountains in New Mexico monitors the impact of thinning vegetation on groundwater recharge. The study objective is to evaluate if thinning forest vegetation will increase groundwater recharge in the mountains to provide larger regional flows to aquifers in surrounding basins. In the semi-arid Southwest, evapotranspiration (ET) makes up 75 to 95% or more of the total water budget. The variability of daily vegetation transpiration and solar radiation with time of year and the effects of complex terrain create a seasonal and spatial variability of ET that is not well quantified in mountainous regions. Through applying the remote sensing model METRIC (Mapping Evapotranspiration with High Resolution and Internalized Calibration) to satellite imagery from the LANDSAT satellite, we calculate high-resolution maps of ET for the Sacramento Mountains watershed area to quantify spatially-distributed estimates of ET before and after thinning to provide improved estimates for determining the water balance and the effect on recharge. METRIC calculates ET through applying an energy balance spatially across an image to estimate ET for each pixel (30m x 30m). Differences in ET are calculated between thinned and control plots in the watershed before and after thinning with the net impact of thinning on ET for an image determined with standard statistical tests following a Before-After Control-Impact (BACI) approach commonly used in environmental impact assessment studies. Estimates of ET from METRIC indicate a net decrease in ET in the first year after thinning for all of the thinned plots but show significant variability (~2 – 12 %) between areas with different terrain characteristics. The impact of surface parameters such as slope, aspect, or albedo among others are currently being examined using multivariate statistical analysis methods to improve the understanding of the spatial and temporal variability of ET for regional forest vegetation groups. Future insights based on such estimates could help in identifying areas that have the largest response to land management practices such as tree thinning.