Reconstruction of daily evapotranspiration under cloudy sky constrained by soil water budget balance

Abstract

Evapotranspiration (ET) plays a significant role in fields of agriculture, hydrology, and meteorology, including the irrigation water allocation, water resources management, drought monitoring and climate change studies. However, because of contamination of cloud cover, the retrieval of daily ET from thermal infrared remote sensing data under cloudy sky bristles with difficulties. In this paper, we proposed a physical and practical method for daily ET reconstruction under cloudy sky that was constrained by soil water budget balance with investigation of the soil water stress relationship between available water fraction and potential ET ratio. This method was tested on both totally field data from 2001 to 2015 at 10 AmeriFlux sites and MODIS remote sensing data from 2009 to 2010 at the Yucheng site in China. Results showed that the reconstructed daily ET under cloudy sky agreed well with the in-situ eddy covariance measurements, with the mean average error (MAE) ranging from −12.20 W/m2 to 16.77 W/m2, the root mean square error (RMSE) varying from 9.47 W/m2 to 28.97 W/m2 at the 10 AmeriFlux sites, the MAE of 2.62 W/m2, and the RMSE of 35.21 W/m2 at the Yucheng site. Overall, the proposed method is promising in reconstructing daily ET under cloudy sky and can be a good alternative to the conventional interpolating/extrapolating methods. Future work may focus on the further investigation of the proposed method over a wider range of atmospheric, soil and vegetation conditions.