Mapping evapotranspiration over viticultural landscapes using multi-sensor remote sensing Dr. Martha Anderson USDA-ARS, USA

Thermal infrared (TIR) and visible/near-infrared (VNIR) surface reflectance imagery from remote sensing can be effectively combined in surface energy balance models to map evapotranspiration (ET) and vegetation health, with broad applications in agriculture and water resource management. In viticulture, geospatial information about actual crop water use and stress at daily timesteps and sub-field scales could significantly inform deficit irrigation strategies that aim to maintain target levels of vine stress in different stages of grape development. Historically, the Landsat satellites have been the workhorse for field-scale ET mapping, providing routine high-quality imagery in all the required bands at scales of human land and water management (30-100m). To augment the temporal sampling of Landsat (8-16 days with 2 or 1 platforms), additional sources of medium-resolution thermal imaging (e.g., ECOSTRESS, VIIRS, MODIS) can be integrated into the ET timeseries via image sharpening and data fusion techniques. In this paper we present case study applications of multi-sensor 30-m ET image times-series over viticultural areas in California, USA, spanning a gradient in climate conditions. Findings regarding instrument and model performance and the value of increased TIR temporal sampling afforded by multi-sensor fusion will be discussed.