Estimating ET Partitioning Using a Thermal-Based Two-Source Energy Balance Model in Vineyards

As water supplies for agricultural production become more restricted due to overuse and drought, particularly in arid regions, there is a concerted effort to improve irrigation methods in order to reduce the amount of water lost through soil evaporation (E) versus transpiration (T) since for many crops T is correlated to biomass production and ultimately yield. The recent drought in California significantly impacting crop production is a prime example of where reliably estimating T versus E and developing techniques to reduce E for irrigated croplands are needed for improving water conservation and agricultural sustainability. While a number of measurement techniques have been developed to estimate T and E, they are very difficult to extrapolate from the local patch scale to field and certainly to the landscape and regional scales. This study presents the application of the Two-Source Energy Balance (TSEB) model for estimating evapotranspiration (ET) that explicitly partitions ET to T and E using land surface temperature (LST), which is available from satellites, giving it regional scale application. An overview of the modeling approach is presented along with a few examples demonstrating the utility of TSEB for partitioning ET between E and T. However, the focus will be on recent efforts in applying the TSEB model in vineyards, which have wide row spacing, often strongly clumped leaf area with most of the biomass concentrated in the upper half of the canopy. Revisions to the model algorithms for radiation and wind divergence through the canopy layer and inter-row air-space are compared to ground-truth observations for vineyard sites in California. The TSEB model is run with tower-based LST as well as airborne and satellite LST imagery. The results on flux output from TSEB using the various LST observations are analyzed and compared to ET observations from eddy covariance measurements as well as T/ET ratios derived from the high frequency turbulence data.