Yield Estimation of Winter Wheat in Pre-harvest Season by Satellite Imagery Based Regression Models

Authors : Ediz ÜNAL, Hakan YILDIZ, Ali MERMER, Metin AYDOĞDU

Pages : 390-403

View : 38 | Download : 13

Publication Date : 2020-12-31

Article Type : Research Paper

Abstract :Early crop yield estimates could provide up-to-date information on supply, demand, stocks, and export availability through which governing bodies can make better agricultural management plans. This study aims to develop a yield model estimating pre-harvest winter wheat yield at both tillering and flowering stages using a multiple linear regression approach based on the relationship between actual yield and satellite derived crops’ phenological parameters. Four crop parameters insert ignore into journalissuearticles values(NDVI, Cumulative NDVI, LAI and FPAR); were regressed in combination to find the best applicable model. Regression results showed that correlations for all models among the variables of the flowering period are higher than that of tillering insert ignore into journalissuearticles values(0.63>0.53);. The mean RMSE’s of the observed vs predicted yields for tillering period was 645.9 kg ha-1 and 574.5 kg ha-1 for flowering period. The optimal developed model which consists of NDVI and CNDVI variables provided 76% and 79% of predicting accuracy 3 and 1.5 months before harvest respectively.
Keywords : MODIS, Remote sensing, Vegetation index, Wheat, Yield prediction

ORIGINAL ARTICLE URL