authors: Holger Hofmann, Gang Zhao, Lenny van Bussel, Andreas Enders, Xenia Specka, Carmen Sosa, Jagadeesh Yeluripati, Fulu Tao, Julie Constantin, Helene Raynal, Edmar Teixeira, Balázs Grosz, Luca Doro, Zhigan Zhao, Enli Wang, Claas Nendel, Kurt-Christian Kersebaum, Edwin Haas, Ralf Kiese, Steffen Klatt, Henrik Eckersten, Eline Vanuytrecht, Matthias Kuhnert, Elisabet Lewan, Reimund Rötter, Pier Paolo Roggero, Daniel Wallach, Davide Cammarano, Senthold Asseng, Gunther Krauss, Stefan Siebert, Thomas Gaiser, Frank Ewert.
abstract: Field-scale crop models can be applied at different spatial resolutions but little is known on the response of models to input data aggregation and why these responses can differ across models. We therefore evaluated 13 crop models which were supplied with climate input data of different spatial aggregation. Spatial resolution of climate input data ranged from 1 to 100 km raster and was used with two crops (winter wheat and silage maize) and three production situations (potential, water limited and nitrogen-water-limited growth) to improve the understanding of errors in model simulations related to data aggregation and possible interactions with the model structure. The most important climate variables identified to determining the model-specific input data aggregation on simulated yields (aggregation effects) were mainly related to changes in radiation (winter wheat) and temperature (silage maize). Additionally, aggregation effects were systematic since models differed in the systematic fraction of the aggregation effect, regardless of the extent of the effect (20 to 66 % as compared to 1.7 % for random effects). Climate input data aggregation changed the mean simulated yield over the region up to 0.2 t ha -1 , whereas simulated yields from single years and models differed considerably depending on the data aggregation. This implies that large-scale crop yield simulations
are robust on average but can be systematically biased at higher temporal or spatial resolutions, depending on the model and its parametrization.
journal: Climate Research