Estimating evapotranspiration and yield of wheat and maize croplands through a remote sensing-based model

Estimation of evapotranspiration (ET) and yield over cropland is essential for water resources management and food security. Particularly the estimation at long-term and regional scales still needs to be investigated. In this study, a remotely sensed and water-carbon coupled model was combined with a process-based crop growth simulation method to simultaneously estimate ET and yield. The developed model was tested for two of the most important crops (wheat and maize) with years of observations. Leaf area index determined by various remote sensing indices and individually calibrated parameters for wheat and maize were considered to improve crop ET and gross primary production (GPP) estimates. The parameters well reflected different physiological properties that C4 maize has a smaller stomatal conductance coefficient and a larger sensitivity in response to initial light than C3 wheat. Comparing with 20 years of measurements, the root mean square error (RMSE) of ET and GPP for winter wheat was 0.57 mm d-1 and 1.65 g C m- 2 d-1, and for maize were 0.80 mm d-1 and 2.92 g C m- 2 d-1, respectively. Besides, crop growth simulation including biomass accumulation and allocation agreed well with the measurements. Coefficient of determination was mostly larger than 0.66, and RMSE of yield was 554.7 and 1346.6 kg hm-2 for wheat and maize, respectively. Finally, the spatiotemporal crop water productivity (WP) of winter wheat and summer maize in the North China Plain during 2001-2018 were quantified. The increased WP of winter wheat while the larger WP of maize than that of wheat were discovered. The developed model helps to examine regional crop ET and yield and provides critical information on agricultural water management.

Automated summary

In this study, a combination of remote sensing and water-carbon coupled model was used to estimate evapotranspiration (ET) and yield of wheat and maize. The model produced accurate estimates by considering various remote sensing indices and individually calibrated parameters. The simulation results of crop growth and yield were in good agreement with measurements, with coefficient of determination mostly larger than 0.66. The developed model also quantified the spatiotemporal crop water productivity (WP), revealing increasing WP of winter wheat and larger WP of maize compared to wheat in the North China Plain during 2001-2018.