Evaluating the NDVI based Rice and Potato Yield Prediction map Using GIS Geostatistical Environment

Abstract

A yield prediction map is an important element in precision agriculture study for site-specific management. In this situation, NDVI based crop vegetation parameter described a better relationship with crop yield prediction. NDVI values were acquired from optical Sentinel 2B images during a specific phenological time in 2019 and 2020. Fifty agricultural plots are occupied in an area of 300ha for both rice (Kharif) and potato (Rabi) crops, in Tarakeswar Block, Hooghly district, West Bengal, India. The ordinary kriging technique was used to produce NDVI prediction maps using Arc GIS 10.7 software. For validation of NDVI and conforming crop yield, both the crops were verified through geostatistical techniques with the lowest RMSE values. The positive coefficient of correlation between NDVI and crop yield was found as r <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =0.406 for NDVI_rice and r <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =0.692 for NDVI_potato, respectively. Further, at semivariograms analysis the lowest nugget-to-sill ratio found 2.51% for rice yield and 1.52% for potato yield, respectively, described the strong spatial autocorrelation. In the other hand, the highest nugget-to-sill ratio found 11.41% for NDVI_rice and 25.52% for NDVI_potato, respectively, representing moderate to strong spatial dependence. The outcome of this research proposed that NDVI is a good predictor of crop yield within-field management zones for sustainable agricultural planning.