Spatial Prediction of wheat crop yield Using Digital Soil Mapping in Gotvand, Khuzestan Province

In this research 110 observed crop yields were correlated with auxiliary variables (DEM and Landsat images) using genetic programming (GP) in Gotvand area (Khuzestan Province). Then, the spatial prediction map of wheat crop yield was calculated using the obtained equation. Wrapper algorithm identified some more important auxiliary variables including NDVI, SAVI, wetness index and channel network based level. RMSE, coefficient of determination and lin's concordance coefficient of GP (1) with all auxiliary data were obtained 525.11, 0.87 and 0.82, respectively. Moreover, results indicated GP (2) with auxiliary data selected by wrapper algorithm could also reasonably predict wheat crop yield (RMSE, coefficient of determination and lin's concordance coefficient, 530.82, 0.86 and 0.79, respectively). It is, therefore, recommended using the same approach to predict spatial distribution of crop yields in the future studies.