Estimation of surface heat fluxes using multi-angular observations of radiative surface temperature

Abstract

The retrieval of canopy and soil component temperatures for estimating evapotranspiration in the two source energy balance (TSEB) model depends on a relatively accurate partitioning of soil/substrate evaporation and canopy transpiration along with the soil and vegetation temperature components. To avoid the need for a Priestley-Taylor based transpiration formulation, this study applies the TSEB model using radiometric land surface temperature observations at multiple view angles from an airborne sensor for estimating soil and canopy temperatures directly. This direct partitioning between soil and canopy temperatures applied with the TSEB formulation improved the agreement between observed and modeled surface heat fluxes, reducing mean absolute percentage error (MAPE) in latent heat fluxes (LE) with flux tower observations from nearly 20% using the original Priestley-Taylor based TSEB model (TSEB-PT) to 15% using TSEB with thermal infrared observations from two substantially different view angles (TSEB-2AG) to nearly 5% using multiple (~6) view angles (TSEB-6AG). Moreover, TSEB-6AG is shown to compute physically realistic spatially-distributed LE for a range of vegetation cover and environmental conditions over the imaged domain. Values of MAPE for sensible heat (H) tended to be larger for all three models due to the fact that tower measurements tended to be located in well irrigated and densely vegetated sites having relatively low H values. This increased accuracy of soil and vegetation component temperature separation using multiangle radiometric temperature observations is useful for evaluating the utility of single and dual view angle thermal radiometer measurements currently available for applying the TSEB model.