Rice Yield Assessment in Nalgonda: A Comparative Study of APSIM and Semi-Physical Models with Remote Sensing Integration

Abstract

Precise estimates of crop yields are crucial for planning and predicting future food supplies and ensuring that the production resources are allocated appropriately. This study demonstrated two approaches: the assimilation of remote sensing with the APSIM model and a semi-physical approach for the spatial rice yield estimation of Nalgonda district, Telangana, during Kharif 2021. APSIM-ORYZA simulated field-level crop production using fundamental input parameters such as soil, weather, and crop management data. The spatial mean yield of the model showed 4925 kg/ha. The semi-physical approach calculated the net primary product using the periodical photosynthetically active radiation (PAR), fraction of absorbed photosynthetically active radiation (FAPAR), crop stress and maximum radiation use efficiency (RUE). The resultant spatial rice yield of the semi-physical approach revealed an average of 4426 kg/ha. In comparison, the government statistics for the Nalgonda district exhibited an average yield of 5026 kg/ha. The crop model showed a deviation of 2% while the semi-physical approach showed 11% from the reported yield. According to the evaluation with observed yields at the field points, the APSIM model correlated linearly with R2 = 0.79, and the root mean square error (RMSE) of 504 kg/ha and SPM showed R2 of 0.76 and an RMSE of 807 kg/ha. These findings suggest that while both approaches hold potential, the APSIM model provided more accurate spatial rice yield estimation.