From a Comprehensively Calibrated Mechanistic Crop Model to Field Applications.

Crop growth models have become a commonly used tool to formulate and evaluate scientific hypotheses but, on the other hand, they may also be used to improve agricultural productivity. In the context of crop growth under drought stress and/or irrigation under limited water resources, the accurate simulation of crop development, yield and soil water dynamics is of vital importance. Hence, for reliable model predictions, a precise model calibration and validation using field observations is essential. Subsequently, in contrast to empirical model approaches, robustly calibrated mechanistic crop models can be used for extrapolation for other conditions (climate, soils) or management strategies. However, farmers tend to apply empiric approaches for irrigation management purposes. The objective of this work was to bridge the gap between scientific simulation studies using comprehensively calibrated mechanistic crop models and practical applications, e.g. to solve common management issues in irrigated crop production. To demonstrate the benefits of the application of a more comprehensive modelling approach, drip irrigated white cabbage was chosen as the crop of interest because it turned out to be barely investigated with respect to crop growth modelling. In the study, a mechanistic crop growth model was calibrated based on comprehensive experimental field data using a multi-criteria approach and Pareto front analysis for the selection of the best set of crop and soil parameters. The field data was derived from an experiment conducted in Germany wherein five irrigation treatments were tested. Data collection included continuous measurements of soil tension at three soil depths, weekly measurements of LAI, plant heights, leaf-N-content and stomatal conductivity as well as biomass partitioning, rooting depth, harvested yields and duration of the growing season. For model calibration, one out of five irrigation treatments – a tension controlled drip irrigated one – was applied. The model was then validated against the other four irrigation treatments. The conducted model calibration significantly enhanced the accuracy of the model predictions and thus was afterwards applied to a scenario analysis to find out best management practices for white cabbage in terms of economic returns and optimal irrigation scheduling.