Comparative analysis of two evapotranspiration models: Unveiling insights into their dynamics over crop growth cycles and the contribution of land surface temperature to their performance

Evapotranspiration (ET) is the combined result of two highly dynamic processes: transpiration, which is the water loss from plants through their stomata, and evaporation, which is the conversion of water on the surface into water vapor. Thus, ET is a key factor in crop growth and yield. To efficiently estimate ET with high temporal and spatial coverage, satellite data provide essential input, such as input to the well-developed models utilizing the energy balance method. Spaceborne thermal infrared data allow the derivation of land surface temperature (LST), a vital component for many ET models that utilize the energy balance theory. Spaceborne-based modelling of ET using energy balance approaches requires input of atmospheric and surface variables with biophysical parameters of the plants covering the surface. Latent heat flux is a critical component that is modelled, as it is the transfer of energy from the surface to the atmosphere that results from evaporation and the transpiration of water from plants. With the lack of validation data worldwide, estimating ET with more than one model has allowed the identification of suitable input, improvement of assumptions, detection of outliers, and assessment of uncertainty. In this research, two models are used to estimate ET: (1) two-source energy balance (TSEB), (2) Priestley-Taylor Jet Propulsion Laboratory model (PT-JPL). They are two-source models; thus, they consider vegetation and soil to be independent regarding heat flux estimation, yet with distinct characteristics. PT-JPL uses empirical environmental constraints to scale an equilibrium ET to the actual ET, yet it can have bias when there is a saturated evaporating front (i.e., after a heavy rainfall event or irrigation). TSEB attempts to iteratively estimate soil and canopy temperatures. Yet, it tends to overestimate the latent heat flux and underestimate the sensible heat flux in certain cases. This research aims to assess the ET estimation characteristics of the two models throughout several years of full crop growth periods. It aims to understand the impact of specific parametrization on their output and the added value of utilizing a next generation of high resolution LST data, the constellr LST30 data product. constellr LST30 is used as precursor data of the upcoming constellr HiVE thermal satellite constellation. This LST dataset has a spatial resolution of 30m and is utilized as input data for the ET estimation. The modelled latent heat flux is compared to flux tower measurement for this purpose. Flux tower footprints are calculated using the two-dimensional parameterisation Flux Footprint Prediction approach that is based on a scaling of the crosswind distribution of the flux. The outcomes of the research bring information about the suitability of each model to certain environmental and crop conditions and highlights the importance of high quality LST to ET modelling.