Simulation of Rice Yield and its Components Using SWAP Model and Remote Sensing Technology for Optimal Use of Water and Soil

Given the importance of soil and water resources in the development of sustainable agriculture, increasing world population and the growing need for crop production, predicting crop yields using plant simulation models and remote sensing technology is very crucial. The aim of this study was to estimate the yield of rice components including straw, paddy and biomass of Hashemi cultivar during different growth stages with SWAP model and to provide regression equations by extracting NDVI and SAVI plant indices from Sentinel-2 and Landsat-7 and 8 satellite images. It was done in the National Rice Research Institute. Comparison of statistical variables indicated that the mean values of coefficient of determination (R2) and model efficiency factor (EF) in estimating the yield of rice components in different stages of growth with SWAP model were more than 0.70 and 0.90, respectively, and with an error of 1.93 to 6.54% was equivalent to 134.21 to 470.43 kg/ha. The slight difference between the measured and simulated values showed that the SWAP model estimates the rice yield in the study area with appropriate accuracy. The results also showed that the extracted NDVI and SAVI indices with very good accuracy estimate the yield of rice components at different stages of growth. However, the highest amount of correlation was related to the reproductive development stage. Finally, R2 for NDVI at different growth stages as well as  the entire growth period for straw, paddy, and biomass were higher than the SAVI index, revealing more accuracy of NDVI than SAVI.