Energy partitioning between plant canopy and soil, performance of the two-source energy balance model in a vineyard

Abstract

Partitioning of evapotranspiration (ET) into soil water evaporation and transpiration allows separate assessment of soil and plant water, energy, and carbon exchange. Remote sensing-based models are ideally suited to monitor ET over large areas, but ET partitioning estimates vary widely. The objective of this study was to evaluate the two-source energy balance (TSEB) model for seasonal ET partitioning using total, soil, and vine canopy energy balance fluxes measured over a vineyard in the Negev desert in Israel. Energy fluxes were evaluated with the original TSEB and three adapted versions using (1) measured soil heat flux, (2) optimized plant transpiration parameterization, and (3) measured soil and vine temperatures instead of composite surface temperature as model inputs. Optimization of plant transpiration parameters revealed a model tendency to underestimate transpiration due to underestimation of available energy and potential transpiration. Adaptations included, among others, accounting for higher leaf radiation absorption expected in dense clumped canopies, which increases available energy, and increasing the Priestley-Taylor coefficient from 1.26 to 2, which increases potential transpiration. While the original TSEB gave reasonable total energy fluxes, the vine energy fluxes were greatly underestimated. Both soil heat flux and plant transpiration adaptations improved modeled vine energy fluxes throughout the season under both well-watered and water-stressed conditions. However, the performance of the TSEB version using measured soil and vine temperatures was inferior to applying the standard TSEB with composite temperature. While daily energy fluxes could be estimated with reasonable accuracy, sub-daily fluxes proved to be more challenging and merit further research. Finally, changes in ET partitioning with canopy development and in response to water stress could be detected quite well albeit it with an underestimation of the transpiration fraction of ET, which, on average, amounted to 39% using standard TSEB and 6% with optimized plant transpiration parameters.