Combining an ecological model with remote sensing and GIS techniques to monitor soil water content of croplands with a monsoon climate

Soil water is an important factor affecting photosynthesis, transpiration, growth, and yield of crops. Accurate information on soil water content (SWC) is crucial for practical agricultural water management at various scales. In this study, remotely sensed parameters (leaf area index, land cover type, and albedo) and spatial data manipulated using the geographic information system (GIS) technique were assimilated into the boreal ecosystem productivity simulator (BEPS) model to monitor SWC dynamics of croplands in Jiangsu Province, China. The monsoon climate here is characterized by large interannual and seasonal variability of rainfall causing periods of high and low SWC. Model validation was conducted by comparing simulated SWC with measurements by a gravimetric method in the years 2005 and 2006 at nine agro-meteorological stations. The model-to-measurement R2 values ranged from 0.40 to 0.82. Nash-Sutcliffe efficiency values were in the range from 0.10 to 0.80. Root mean square error (RMSE) values ranged from 0.028 to 0.056m3m-3. Simulated evapotranspiration (ET) was consistent with ET estimated from pan evaporation measurements. The BEPS model successfully tracked the dynamics and extent of the serious soil water deficit that occurred during September-November 2006. These results demonstrate the applicability of combining process-based models with remote sensing and GIS techniques in monitoring SWC of croplands and improving agricultural water management at regional scales in a monsoon climate.