Statistical-geographical modeling for estimating the yield of rainfed chickpeas in its major cultivation areas in Kermanshah province

Identification of the most effective indicators as well as the effect of spatial and temporal differences on the yield of rainfed chickpeas in major cultivation areas of Kermanshah province were the main objectives of this research. A map of major rainfed chickpea cultivation areas in the province was drawn by crop statistics. Then, the farms of this crop were extracted using the digital map layer of chickpea fields on the received images of Modis sensor from 2000 to 2021. In the next step, 9 plant spectral indices for chickpeas over the four climatic regions of Kermanshah in the flowering growth stage were calculated during a period of 22 years. These 9 spectral indices, together with the variable of total precipitation as independent variables and yield data as a dependent variable entered into the stepwise regression model. The results showed that NDVI and precipitation indices are the most effective indices of yield variability during flowering stage of chickpea in Kermanshah. Furthermore, NDVI, PVI and SAVI indices in the cities of Islamabade Gharb, Dalahou and Songhar are the most effective indices during the studied period, respectively. The results of validation revealed that the statistical model of Kermanshah city was more accurate than other cities. The correlation coefficient of the statistical model for estimating chickpea yield in this city was 0.69 with a standard error of 84 kg/ha-1. In addition, the relative deviation values of the statistical model of this city were less than other models.