Comparison of Modis and Proxy-viirs Derived Evapotranspiration Estimates for Improved Agricultural Best Practices Assessment

The AnnAGNPS pollution model uses evapotranspiration as one of the inputs to predict runoff, sediment, nutrients, and pesticide contributions from land to streams. This modeling tool has been modified to accept evapotranspiration estimates derived from remotely sensed imagery, produced by the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor. This research evaluated the next generation of moderate resolution sensor, the Visible Infrared Imager Radiometer Suite (VIIRS), as an alternative source of remotely sensed evapotranspiration estimates for use as input into the AnnAGNPS pollution model. Two distinct sets of daily evapotranspiration maps were produced based on remotely sensed parameters, one set from MODIS and another set from proxy-VIIRS images. NDVI and LST values from both datasets, MODIS and proxy-VIIRS, were integrated with discrete AnnAGNPS evapotranspiration values based on field stations through the use of the VI-LST Triangular Space Method and linear multivariate regression. Daily regression models were developed using the information from the existing field stations as independent variables and the NDVI and LST pixel values corresponding to the field stations, as dependent variables. The developed models were then used to create two sets of daily evapotranspiration maps, providing data where field stations are not found. Comparisons between the two evapotranspiration maps were performed. Preliminary results indicate that the next generation of moderate resolution sensors will continue to provide AnnAGNPS an alternative source of gridded evapotranspiration estimates to overcome the limitations imposed by localized field stations.