EVAPOTRANSPIRATION RECOVERY FOLLOWING FIRE IN RIPARIAN FORESTS

Hydrological recovery times from fire disturbance depend on vegetation community characteristics, hydro-geomorphic setting, fire intensity, soil properties, and subsequent climate forcings such as precipitation amount, intensity and frequency. But equally important is the hydrologic metric being measured and used to define recovery. Common metrics include basin runoff, overland flow, evapotranspiration (ET), soil moisture, and sediment production. In this research we compare recovery of riparian forests following ‘bosque fires’ along the Rio Grande in central New Mexico. Riparian vegetation in this region is dominated by native cottonwood and willow, and exotic tamarisk and Russian olive. We compare recovery of ET in stands of cottonwood, tamarisk, and mixed forest using remote sensing and water table monitoring. Landsat 8 images were analyzed for greenness before and following three spring 2016 fires (San Acacia, Escondida, and Bosque del Apache) as well as fires from 2-3 years earlier. We compare results to ET monitoring of a tamarisk fire in 2010 using Mapping EvapoTranspiration with high Resolution and Internalized Calibration (METRIC) analysis of Landsat imagery. We also monitored shallow alluvial aquifer well levels before and after disturbance in two of the 2016 fire areas as well as two unburned control areas. By analyzing daily water table elevation cycles and using specific yield estimates from the well log data, we produced daily ET estimates for the areas. We find that ET recovers most rapidly if tamarisk are present. Precipitation following the fire, while not a controlled variable, appears to have a strong influence on ET recovery. Greenness increases due to herbaceous regrowth are not necessarily tied to increases in total ET as measured in the aquifer. In floodplains that are disconnected from the fluvial system by levees or flow impoundment, we recommend replanting of native vegetation if post-fire recovery of native bosque habitat is valued.