METRIC-2S: A Two-Source Model for Enhanced Partitioning of Evapotranspiration in Agricultural Landscapes

Partitioning evapotranspiration (ET) into soil evaporation (E) and plant transpiration (T) is crucial for accurate water resource management. Traditionally, this has been challenging due to the complexity of the underlying processes. In this study, we develop an approach to enhance the Mapping EvapoTranspiration at high Resolution with Internalized Calibration (METRIC) model, enabling better partitioning of landscape-scale flux components. Named METRIC-2S, this approach introduces a two-source scheme into the original single-source model, using soil and vegetation temperatures for partitioning. These temperatures are used by METRIC to calculate two ET components, one for soil and another for vegetation, subsequently weighted by fractional vegetation cover (fc) to compute E and T. Soil and vegetation temperatures are estimated using the hourglass method, driven by surface temperature and fc. ET estimates from both the original METRIC and the revised METRIC-2S models are compared and validated against eddy covariance measurements over three agricultural sites: an olive orchard, a wheat field, and a mixed wheat/olive plantation. METRIC-2S demonstrates significant improvements in accuracy relative to the original METRIC model across all three sites, with reductions in RMSE from 141 to 63 W/m2 at the olive site, 102 to 83 W/m2 at the wheat field, and 180 to 78 W/m2 at the mixed site. To evaluate the performance of the partitioning scheme, transpiration estimates were compared with available sap flow measurements at the olive orchard site on selected dates coinciding with a Landsat overpass, yielding an RMSE of approximately 22.3 W/m2. While further verification and assessment of component values are necessary, the results suggest that the METRIC-2S approach strikes a good balance between simplicity and improved accuracy.