Use of multiple regression analysis to estimate average corn yields using selected soils and climatic data

Abstract A systematic method for predicting corn yields was developed using research plots, selected soil properties and climatic data. :Multiple regression analyses were employed to generate coefficients for the relative contributions of selected soil properties and climatic data from five experimental research plots located in New York State. Yield data were collected from 2 to 19 years, along with climatic data. The soils at each of the plots were mapped, described, sampled, characterized, classified and interpreted. Linear equations, with coefficients generated from the multiple regression analyses, were developed from 43 years of corn data at five research plots. The predictive linear equation was tested at an independent experimental plot with 19 years ofactual corn yield data. The resulting linear relationship between actual and estimated corn yields had an r2 of 0·52. The model did, however, predict an average corn yield (for 19 years) within 194 kg ha−1 (3 %) of the actual average corn yield. The equations have the potential to predict the average 20-year crop yields under high management for all soil map units in New York State.