A comprehensive nitrogen fertilizer management model for winter wheat (Triticum aestivum L.)

Winter wheat at 47 locations in Montana was topdressed with nitrogen (N) in the spring of 1970, 1971, 1972 and 1973.A stepwise multiple regression technique was utilized to generate a comprehensive N fertilizer management model to predict potential grain yield, N fertilizer requirements, grain protein with additions of N fertilizer, grain yield and grain protein without spring N additions and residual soil NO3-N after harvest.The data were organized into two groups based on soil NO3-N to 4 ft.Locations having soils which contained less than 120 lbs NOg-N/4 ' were designated as group I and remaining locations as group II.With group I data highly significant equations were generated for the entire N fertilizer management model.Independent variables for potential yield prediction were growing season rain-fall, evaporation rates during the first half of the growing season and soil organic matter.Soil NO3-N, potential yield, evaporation rate during the first half of the growing season and available soil water were the important factors for predicting N fertilizer requirement.The variables useful in predicting grain protein were potential yield, soil NO3-N, N fertilizer rate, soil organic matter and growing season rainfall.For comparison with potential yield, grain yield equations were generated from check plot data; the important independent variables were soil NO3-N, evaporation rate during the first half of the growing season, growing season rainfall and soil organic matter.Similarly grain protein was predicted; important factors were soil NO3-N, growing season rainfall, grain yield and soil temperature at 50 cm.Equations for the group II data were eractic because of insufficient data for analysis and response to added N was uncertain.Data from groups I and II were combined and equations developed similar to group I; only the protein functions were nonsignificant.A modeling system to predict residual soil NO3-N after harvest was generated.The equations developed from the check plots were highly significant; the important variables include soil NO3-N, soil water, soil temperature at 50 cm, evaporation rate and grain protein.Equations generated from 80 to 180 -N treatments were nonsignificant.The modeling system applies to winter wheat producing areas of Montana where excellent stands of recommended varieties are present, an alternate' crop-fallow management system is practiced and P fertilizer is drilled with the seed.