Satellite-based crop coefficient and evapotranspiration using surface soil moisture and vegetation indices in Northeast Asia

摘要

Accurate estimation of the crop coefficient (K-c) is crucial for estimating actual crop evapotranspiration (ETa) and planning appropriate irrigation management for efficient crop yield. In this study, satellite-based K-c values were estimated at cropland and mixed forest sites based on the dual crop coefficient approach using merged soil moisture from the European Space Agency as an indicator of evaporation from soil, as well as the Normalized Vegetation Index (NDVI) and the Leaf Area Index (LAI) to explain the effect of transpiration from plants. Comparison of the seasonal patterns and Pearson's correlation coefficient (r) of NDVI, LAI, and surface soil moisture with K-c indicated that it was reasonable to use the three variables as independent variables to estimate K-c. Based on these results, the satellite -based K-c estimated using NDVI, LAI, and soil moisture (Case 1) was compared with the K-c calculated from NDVI and LAI (Case 2) and the flux towers at the significance level of 0.05. The statistical results confirmed that the K-c estimated from Case 1 (Bias: -0.012 to 0.053, RMSE: 0.144 to 0.172, and r: 0.463 to 0.800) showed better agreement with the observed K-c than that estimated from Case 2 (Bias: -0.058 to 0.088, RMSE: 0.146 to 0.221, and r: 0.434 to 0.788). Among the three variables, soil moisture had the greatest impact on the rice paddy, while the NDVI showed the highest influence on the mixed forest. Based on these results, K-c estimated from Case 1 was multiplied by MODerate resolution Imaging Spectroradiometer (MODIS)-based potential crop evapotranspiration and compared with the latent heat flux from flux towers. ETa showed reasonable bias (cropland: -0.224 to 1.364, mixed forest: 0.711 to 1.055), RMSE (cropland: 1.952 to 2.126, mixed forest: 1.085 to 1.878) and r (cropland: 0.529 to 0.832, mixed forest: 0.850 to 0.909) at all of the study sites. After validation of the satellite -based K-c approach under various vegetation types and climate conditions, this approach can be employed not only for developing adequate water and agricultural management plans, but also for analyzing and predicting crop yield productivity and agricultural drought.