Field-specific simulations of net N mineralization based on digitally available soil and weather data. I. Temperature and soil water dependency of the rate coefficients

Including field-or even site-specific estimates of current net N mineralization into N fertilizer strategy is essential in order to further reduce N surpluses while maintaining crop yields, but adequate estimates are not available. Simulation models could account for many influencing factors, yet are not easily adjustable to different soil and site characteristics. Nowadays important input data for N mineralization models are digitally available. Thus, our objectives were (1) to experimentally determine specific temperature and soil water dependency functions for the rate coefficients of net N mineralization that could be allocated via digitally mapped data and (2) to find out the least necessary discrimination between soils. Specific and general functions for the rate coefficients of two organic N pools with first-order kinetics were derived using laboratory long-and short-term incubations from a broad variety of soils. Functions were evaluated using comparisons to field incubations of undisturbed soil columns from 27 sites. Interestingly, a differentiation between specific functions of not more than three soil groups was necessary for quite accurate simulations (r(2) = 0.87, P<0.001; RMSE = 23 kg N ha(-1), n-RMSE = 29%). The two criteria for grouping, soil texture (loess vs. sandy/loamy classes) and humus content class (applies only to temperature functions for sandy textures), can be taken from digital soil maps. Field studies, especially under suboptimal water contents, with plant cover and N-fertilization, will have to further prove the applicability of the derived functions. Pedotransfer functions for the pool sizes also based on digitally available data are needed for automatically calculating specific estimates of net N mineralization.