Remote Sensing Estimates of Evapotranspiration at a Southwestern Uranium Mill Tailings Site
Pamela Nagler1*, Christopher Jarchow2, William Waugh2, Armando Barreto-Munoz3, Stefanie Herrmann3, Kamel Didan3
1U. S. Geological Survey, Southwest Biological Science Center, Tucson, AZ 85719; pnagler@usgs.gov
2Navarro Research and Engineering, 2597 Legacy Way, Grand Junction, CO 81503; Chris.Jarchow@lm.doe.gov, Jody.Waugh@lm.doe.gov
3University of Arizona, Biosystems Engineering, Tucson, AZ 85719; abarreto@email.arizona.edu; stef@email.arizona.edu; didan@email.arizona.edu
Tamarisk (Tamarix spp.) is a non-native tree that competes with native species for water in riparian corridors of the southwestern U.S. We studied changes in a riparian plant community dominated by tamarisk by measuring long-term trends in greenness and evapotranspiration (ET) at a uranium mill tailings site adjacent to the San Juan River near Shiprock, New Mexico. In August of 2016, we used an unmanned aerial system (UAS) to acquire high-resolution spectral data needed to improve our spatial accuracy in mapping variability in ET measurements. UAS imagery allowed us to monitor changes in phenology, fractional greenness, ET, and on water resources at these sites. We timed ground data and UAS image acquisition with an August 2016 Landsat image to assist with spatiotemporal scaling techniques. The UAS product was correlated not only with Landsat but also Moderate Resolution Imaging Spectrometer (MODIS) satellite imagery. We measured leaf area index (LAI) and sampled biomass on tamarisk, cottonwood (Populus spp.), and willow (Salix spp.) within the UAS acquisition areas to scale leaf area on individual branches to LAI of whole trees of a given genera. Ground validation of vegetation cover types (n=24) was done with a GPS for over a hundred locations. UAS cameras included a Sony Alpha A5100 for species-level mapping (using the 24 vegetation cover types) and a MicaSense Red Edge five-band multispectral camera to map Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI). NDVI and EVI were calibrated across UAS, Landsat and MODIS using regression and ET was calculated using NDVI, EVI, ground meteorological data, and an existing empirical algorithm. Our goal is to scale plant water use acquired from UAS imagery to Landsat and compare with MODIS estimates of ET to provide a time-series documenting long-term trends (through 2018) in plant species cover or vegetation community structure with relationships of ET and groundwater elevation.