Geospatial Technology for Crop Yield Forecasting-A Case Study of SW Uttarakhand

Abstract

AbstractCrop yield forecasting is important for policy decisions at the national and regional levels. In the present study, the yield of rice, Wheat, and Sugarcane crops has been predicted for South-west Uttarakhand. Three major crop-grown districts of SW Uttarakhand, namely Haridwar, Dehradun, and Pauri Garhwal, were taken under the study. MODIS-derived Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) parameters were used for crop yield forecasting. Stepwise Regression and Least Absolute Shrinkage and Selection Operator (LASSO) methods were used for forecasting using YIELDCAST-DSS software developed by India Meteorological Department (IMD) with Indian Agricultural Statistics Research Institute (IASRI). The meteorological parameters viz. rainfall, maximum and minimum temperature, the maximum and minimum relative humidity for the last 20 years (2001–2020) were used. The crop yield forecasting has been done for F2 stage (vegetative growth stage) and F3 stages (pre-harvest stage). NDVI helps to provide information about the overall state of vegetation cover and differentiate vegetation from other types of Landcover. EVI is used to quantify the greenness of vegetation, and it is more sensitive to dense vegetation cover. The results revealed a reduction in uncertainties in crop yield forecasting by assimilation of remote sensing data in the crop model. The study shows a good result of yield forecasted by using NDVI and EVI as an independent parameter compared to the other climatic variables without NDVI and EVI. In most models, the EVI outperformed NDVI and climatic variables alone.KeywordsCrop yield forecastingRemote sensingStepwise regressionLASSONDVIEVI