Monitoring Daily Evapotranspiration at Field-to-Regional Scales: An Application over California Vineyards using Landsat 8

Tuesday, April 21, 2015
Martha C. Anderson1, Kathryn A Semmens1, Feng Gao1, Yun Yang1, Christopher Hain2, William P Kustas1, Joseph G Alfieri1, John H Prueger3, Lynn McKee1 and Wayne Dulaney1, (1)USDA ARS, Beltsville, MD, United States, (2)University of Maryland, CMNS-Earth System Science Interdisciplinary Center, College Park, MD, United States, (3)USDA ARS NLAE, Ames, IA, United States
Abstract:
The California drought of 2014 highlighted the need for robust operational tools for monitoring agricultural water use and crop condition in support of water management decision making, both at the farm scale and at the regional level of irrigation districts. California growers need high resolution information about historical and current water use on their own and surrounding fields to be able to develop reasonable plans for responding to drought and potential water restrictions. Water resource managers require coarser information over larger areas to support credible determinations of water allocations and curtailments. This paper describes a multi-scale remote sensing system for monitoring evapotranspiration (ET) and crop water stress and an application over two vineyard sites near Lodi, California during the 2013-2014 growing seasons.

The system employs a data fusion methodology (STARFM: Spatial and Temporal Adaptive Reflective Fusion Model) combined with multi-scale ET modeling (ALEXI: Atmosphere Land Exchange Inverse Model) to compute daily 30 m resolution ET. ALEXI ET fluxes (4 km resolution, daily) are integrated with ET fluxes from Landsat 8 thermal data (30 m resolution, ~16 day) and Moderate Resolution Imaging Spectroradiometer (MODIS) data (1 km resolution, daily). The high spatial resolution Landsat retrievals are then fused with high temporal frequency MODIS data using STARFM, to produce daily estimates of crop water use that resolve within-field variation in ET for individual vineyards.

Estimates of daily ET generated in two fields of Pinot Noir vines of different maturity agreed well with ground-based flux measurements collected within each field with relative errors of about 15%. Spatial patterns of cumulative ET correspond to yield estimates and indicate areas of variable soil moisture, crop condition, and yield within the vineyard that could require adaptive management strategies. Multi-year ET datacubes (30-spatial resolution, daily timesteps) developed using thermal data from the Landsat archive can provide valuable information for both realtime water management decision-making and for development of longer term drought response plans.

<< Previous Abstract | Next Abstract >>