UP-SCALING PADDY YIELD AT SATELLITE-FOOTPRINT SCALE USING SATELLITE DATA IN CONJUNCTION WITH CCE DATA IN SAHIBGANJ DISTRICT, JHARKHAND
- Department of Land Resource Management, School of Natural Resource Management, Central University of Jharkhand, Brambe-835205, Jharkhand, India
Keywords: Yield prediction, Remote sensing, NDVI, EVI, AquaCrop model, CCE data
Abstract. Agriculture plays a vital role in the economy of India as almost half of the workforce dependent on agriculture and allied activities. Rice is an important staple food and provides nutritious need for the billions of population. Mapping the spatial distribution of paddy and predicting yields at district level aggregation are crucial for food security measures. This study has utilized the time-series MODIS-based Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) data in conjunction with CCE data to derive a statistical model for up-scaling paddy yield at satellite-footprint scale over Sahibganj district in Kharif (monsoon) season 2017. The CCE data were collected from ten random paddy plots. In addition, Area, Production, and Yield (APY) data were collected during harvesting period by interacting with eighty farmers belong to eight villages. The AquaCrop model was also used to simulate the paddy yield for Kharif season. The key results showed that based on the farmers-based yield data, paddy yield was observed as ~3200 kg/hectare, whereas, NDVI and EVI-based yield models based on satellite data showed about 2,960 and 3,530 kg/hectare, respectively. Moreover, multi-regression-based yield model showed the mean yield of 3,070 kg/hectare. With respect to farmers-level yield data, the relative deviation (RD) of yield based on NDVI data was −7.5% (underestimation), while EVI was 10.31% (overestimation). The multi-regression-based yield model and AquaCrop model were underestimated by −4.06 and −10.16%, respectively. Thus, it can be inferred that the multi-regression-based yield was close to farmers-based survey yields. It can be concluded that the satellite databased yield prediction can be reliable with ± 10% of RD. Nevertheless, remote sensing technology can be beneficial over traditional survey method as the satellite-based methods are cost-effective, robust, reliable, and time-saving than the traditional methods.