Abstract: Due to the more frequent use of crop models at regional and national scale, the effects of spatial data input resolution have gained increased attention. However, little is known about the influence of variability in crop management on model outputs. A constant and uniform crop management is often considered over the simulated area and period. This study determines the influence of crop management adapted to climatic conditions and input data resolution on regional-scale outputs of crop models. For this purpose, winter wheat and maize were simulated over 30 years with spatially and temporally uniform management or adaptive management for North Rhine-Westphalia ((similar to)34 083 km(2)), Germany. Adaptive management to local climatic conditions was used for 1) sowing date, 2) N fertilization dates, 3) N amounts, and 4) crop cycle length. Therefore, the models were applied with four different management sets for each crop. Input data for climate, soil and management were selected at five resolutions, from 1 x 1 km to 100 x 100 km grid size. Overall, 11 crop models were used to predict regional mean crop yield, actual evapotranspiration, and drainage. Adaptive management had little effect (< 10% difference) on the 30-year mean of the three output variables for most models and did not depend on soil, climate, and management resolution. Nevertheless, the effect was substantial for certain models, up to 31% on yield, 27% on evapotranspiration, and 12% on drainage compared to the uniform management reference. In general, effects were stronger on yield than on evapotranspiration and drainage, which had little sensitivity to changes in management. Scaling effects were generally lower than management effects on yield and evapotranspiration as opposed to drainage. Despite this trend, sensitivity to management and scaling varied greatly among the models. At the annual scale, effects were stronger in certain years, particularly the management effect on yield. These results imply that depending on the model, the representation of management should be carefully chosen, particularly when simulating yields and for predictions on annual scale.

Abstract: Highlights • The parameters to be fixed were consistent across sites. • Model calibration must be performed separately for each specific case. • Possible to reduce model parameters from 66 to 45. • Strong model reductions must be avoided. • The error term for the training data were characterised by timing (phase shift). Abstract Proper parameterisation and quantification of model uncertainty are two essential tasks in improvement and assessment of model performance. Bayesian calibration is a method that combines both tasks by quantifying probability distributions for model parameters and outputs. However, the method is rarely applied to complex models because of its high computational demand when used with high-dimensional parameter spaces. We therefore combined Bayesian calibration with sensitivity analysis, using the screening method by Morris (1991), in order to reduce model complexity by fixing parameters to which model output was only weakly sensitive to a nominal value. Further, the robustness of the model with respect to reduction in the number of free parameters were examined according to model discrepancy and output uncertainty. The process-based grassland model BASGRA was examined in the present study on two sites in Norway and in Germany, for two grass species (Phleum pratense and Arrhenatherum elatius). According to this study, a reduction of free model parameters from 66 to 45 was possible. The sensitivity analysis showed that the parameters to be fixed were consistent across sites (which differed in climate and soil conditions), while model calibration had to be performed separately for each combination of site and species. The output uncertainty decreased slightly, but still covered the field observations of aboveground biomass. Considering the training data, the mean square error for both the 66 and the 45 parameter model was dominated by errors in timing (phase shift), whereas no general pattern was found in errors when using the validation data. Stronger model reduction should be avoided, as the error term increased and output uncertainty was underestimated.