Daytime sensible heat flux estimation over heterogeneous surfaces using multitemporal land-surface temperature observations

Equations based on surface renewal (SR) analysis to estimate the sensible heat flux (H) require as input the mean ramp amplitude and period observed in the ramp-like pattern of the air temperature measured at high frequency. A SR-based method to estimate sensible heat flux (HSR-LST) requiring only low-frequency measurements of the air temperature, horizontal mean wind speed, and land-surface temperature as input was derived and tested under unstable conditions over a heterogeneous canopy (olive grove). HSR-LST assumes that the mean ramp amplitude can be inferred from the difference between land-surface temperature and mean air temperature through a linear relationship and that the ramp frequency is related to a wind shear scale characteristic of the canopy flow. The land-surface temperature was retrieved by integrating in situ sensing measures of thermal infrared energy emitted by the surface. The performance of HSR-LST was analyzed against flux tower measurements collected at two heights (close to and well above the canopy top). Crucial parameters involved in HSR-LST, which define the above mentioned linear relationship, were explained using the canopy height and the land surface temperature observed at sunrise and sunset. Although the olive grove can behave as either an isothermal or anisothermal surface, HSR-LST performed close to H measured using the eddy covariance and the Bowen ratio energy balance methods. Root mean square differences between HSR-LST and measured H were of about 55 W m(-2). Thus, by using multitemporal thermal acquisitions, HSR-LST appears to bypass inconsistency between land surface temperature and the mean aerodynamic temperature. The one-source bulk transfer formulation for estimating H performed reliable after calibration against the eddy covariance method. After calibration, the latter performed similar to the proposed SR-LST method.