Predicting the site specific soil N supply under winter wheat in Germany

The expected amount of plant nitrogen (N) at harvest which originates from soil N supply is of high relevance for N fertilization planning. Due to mineralization–immobilisation turnover processes, soil N supply is influenced by N fertilization which complicates its assessment. The soil N supply consists of two components: the soil mineral N measured at early spring and the ‘effective’ N mineralization (Min^eff) under winter wheat (Triticum aestivum L.). Min^eff was defined as the difference between crop N uptake (N^crop) at harvest and N supply. Our aim was the identification and quantification of climate and site-related factors in order to achieve an improved assessment of the site-specific (long term average) Min^eff. We used N rate experiments from 411 collective seasons, carried out at 98 sites across Germany in order to analyze the impact of climate and site-related factors on Min^eff. Quadratic curves were fitted in order to describe the grain N uptake as a function of N supply. A fixed marginal N efficiency was defined in order to analyze Min^eff at a reasonable N supply. Starting with estimates for Min^eff as function of preceding crop, we found that climate (average temperature during May, annual rainfall) and site-related factors have a significant influence on Min^eff. In order to ensure that the regression model is transferable to unknown sites, a “leave one site out” cross validation was carried out. Compared to considering preceding crop only (reference), the regression model reduced the RMSE by 9.5 (calibration) or 8.3 (cross validation) kg N/ha.