Maize yield prediction using NDVI derived from Sentinal 2 data in Siddipet district of Telangana state

Abstract

Maize is a short duration crop. Accurate and up-to-date assessment of the spatial distribution of cultivated area and its production is a key information requirement for all stake holders including policy makers, farmers and consumers. The study utilized Sentinel-2A/2B satellite based NDVI (Normalized Difference Vegetation Index) data to study yield variability of maize in Siddipet district of Telangana state, in 2019. NDVI is a quantitative measurement of crop vigour, which denotes the crop biomass and health status. A regression analysis was developed between observed yield from crop cut experimental plots and seasonal maximum NDVI of maize area. Results showed that there was statistically significant relation (R2=0.87) between NDVI and observed yield at respective field plots. Validation of yield model with Root Mean Squared Error (RSME) computed for all observed and predicted yields was 0.50 explaining higher accuracy of the yield model. The average seasonal rainfall received in different mandals of Siddipet district was between 727 to 799 mm and during the crop growth period the yields were recorded in the range of 4.2 to 5.0 t ha-1. This showed that the distribution of rainfall has played a major role in yield prediction. The 31st Standard Week Rainfall (SWR) coinciding with the peak vegetative growth showed a significant positive correlation (r=0.68***) with maize yield data at p<0.001.