Soil evaporation monitoring: a possible synergism of microwave and infrared remote sensing

Abstract

Microwave remote sensing allows the measurement of the water content ( θ s) at the soil surface within a layer of a few centimetres. When combined with climatic data, θ s is a relevant quantity to estimate the evaporation of bare soils. The implementation of a simple daily evaporation ( E d) model on bare soils based on a knowledge of θ s is analysed. In order to cover a wide range of soil, soil moisture and climatic conditions, the analysis was carried out on a set of data simulated by a mechanistic model of heat and water flows in the soil. Propagation error analysis on the inputs ( θ s, daily potential evaporation and wind velocity) of the simple model shows that an accuracy of ± 1.5 mm day −1 (95% confidence interval) may be expected when θ s is observed from space-borne microwave sensors. This accuracy is significantly improved (±0.6 mm days) if E d is averaged over a 5 day period. With the simple evaporation model considered, the time interval between microwave observations should not exceed 3 days. However, if the precipitation amounts were known, this time interval could reach 5 days. The soil hydrodynamic properties are included in the simple evaporation model through empirical parameters, which depend strongly on the soil type. We show that the complementarity between thermal infrared and microwave observations could be used to infer the empirical parameters of the simple evaporation model. When microwave and infrared observations are available for the same day, we can inverse the simple evaporation to retrieve its parameters. Even with important errors on the input quantities, the proposed inversion method allows good discrimination between soils. Furthermore, the performances in the estimation of the evaporation are little affected when the simple model is implemented with the retrieved parameters.