A one – source model to estimate sensible heat flux in agricultural landscapes

A one-source model type that combines surface renewal (SR) and similarity theories to estimate the sensible heat flux (H) involving the land surface temperature (LST) and low-frequency measurements, HSR₋LST, was tested over agricultural surfaces with fractions of ground canopy cover that ranged from 39% up to 97% and canopy heights ranging from 0.2 to 1.8 m. The performance was tested under clear sky conditions and for stable and unstable cases during the day. The sensible heat flux was estimated using in situ LST measurements over a homogenous canopy (wheat) under stable conditions. Unstable conditions were met for two growing crops (rain-fed soybean and corn) where HSR₋LST was tested using LST retrieved from Landsat 5 to 7. The sensible heat flux taken as a reference for soybean and corn was determined using the eddy covariance (EC) method and for wheat it was determined using the residual method (the evapotranspiration was measured using a large weighing lysimeter). For wheat HSR₋LST performed slightly better than a calibrated one-source model. For soybean and corn, it performed similarly to one-source and two-source model types. The one-source models included two types, one requiring calibration and the other one self-calibrating, using ‘hot’ and ‘cold’ spots in the surroundings. Further testing for sparse canopies is suggested prior to the development of a two-source model in the framework of SR-LST because HSR₋LST performed similarly to the other models. For corn, HSR₋LST had the closest performance to HEC and for soybean the two-source model (the best) was slightly closer to HEC than HSR₋LST. The root mean square differences (RMSD) comparing HSR₋LST against the reference were in the range 26 (corn) – 46 (wheat) Wm⁻². Therefore, HSR₋LST is an alternative to consider for estimating H because it was competitive and simpler to apply than other thermal remote sensing-based methods.