An improved satellite‐based approach for estimating vapor pressure deficit from MODIS data

Abstract Vapor pressure deficit (VPD) is an important variable widely used in ecosystem and climate models. In this paper, an improved satellite‐based approach to estimating VPD was presented that uses several remote sensing products coupled with field measured data. The proposed method exploits an optimized algorithm to derive near‐surface actual vapor pressure ( e a ) from Moderate Resolution Imaging Spectroradiometer (MODIS) data and upgrades Smith's (1966) methodology for estimating e a . The proposed new algorithm for calculating e a was evaluated against in situ measurements at 119 validation sites in China for 2 months in 2013. The mean absolute error (MAE) and root‐mean‐square error (RMSE) were less than 0.25 kPa and 0.33 kPa, respectively. The near‐surface air temperature ( T a ), which is an important input data for calculating VPD, was estimated from satellite‐retrieved land surface temperature, and had an RMSE of less than 2.5 K. The estimated VPD values were validated with ground observation data from the Heihe River Basin for 5 months in 2012 and for all of China for August 2013. A coefficient of determination ( R 2 ) of 0.912, MAE of 0.27 kPa, and RMSE value of 0.32 kPa were achieved for the 2012 test data, and corresponding values of 0.88, 0.278 kPa, and 0.367 kPa for the 2013 test data. These results are promising, especially considering the comparatively high spatial resolution (1 km) of the VPD map estimated from the satellite data. Potential applications include global evapotranspiration estimation, fire warning, and vegetation analysis.