Search

International audience; Abstract Climate change is expected to severely affect cropping systems and food production in many parts of the world unless local adaptation can ameliorate these impacts. Ensembles of crop simulation models can be useful tools for assessing if proposed adaptation options are capable of achieving target yields, whilst also quantifying the share of uncertainty in the simulated crop impact resulting from the crop models themselves. Although some studies have analysed the influence of ensemble size on model outcomes, the effect of ensemble composition has not yet been properly appraised. Moreover, results and derived recommendations typically rely on averaged ensemble simulation results without accounting sufficiently for the spread of model outcomes. Therefore, we developed an Ensemble Outcome Agreement (EOA) index, which analyses the effect of changes in composition and size of a multi-model ensemble (MME) to evaluate the level of agreement between MME outcomes with respect to a given hypothesis (e.g. that adaptation measures result in positive crop responses). We analysed the recommendations of a previous study performed with an ensemble of 17 crop models and testing 54 adaptation options for rainfed winter wheat (Triticum aestivum L.) at Lleida (NE Spain) under perturbed conditions of temperature, precipitation and atmospheric CO2 concentration. Our results confirmed that most adaptations recommended in the previous study have a positive effect. However, we also showed that some options did not remain recommendable in specific conditions if different ensembles were considered. Using EOA, we were able to identify the adaptation options for which there is high confidence in their effectiveness at enhancing yields, even under severe climate perturbations. These include substituting spring wheat for winter wheat combined with earlier sowing dates and standard or longer duration cultivars, or introducing supplementary irrigation, the latter increasing EOA values in all cases. There is low confidence in recovering yields to baseline levels, although this target could be attained for some adaptation options under moderate climate perturbations. Recommendations derived from such robust results may provide crucial information for stakeholders seeking to implement adaptation measures.

International audience; This study was aimed to develop, test and provide access to a snow cover model for agrometeorological use (snowMAUS) that would rely strictly on the weather data used by all crop simulation models, i.e., diurnal temperature extremes and total daily precipitation. Such snow model can easily be used to preprocess input data in order to account for the presence or absence of snow cover whenever required by the crop modeler, without necessitating the acquisition of additional data. The snow cover model was tested across 65 sites across Austria with considerable variability in elevation (155–3111 m a.s.l.). In addition 7 sites in the Czech Republic were used to evaluate snowMAUS reliability in assessing frost damage to winter wheat crop. For illustration a case study documenting the benefit of coupling snowMAUS with process based crop model (STICS) was also included in the study. The presented work complements previous snow cover modeling studies with its focus on the development of a simple snow cover model for areas under intensive agricultural use (mostly lowlands) with 65% of stations providing weather data being located at altitudes below 800 m. The presented results proved that SnowMAUS can estimate with reasonable precision snow cover presence/absence, and to a large extent snow cover depth. The model also accurately represented seasonal variations in the number of snow days or in the volume of precipitation in the form of snow. The simplicity of the model and the fact that it relies only on the daily data is a great advantage. We have demonstrated that in areas with a high probability of winter temperatures dropping below frost-tolerance thresholds and that tend to have considerable snow cover (e.g., Central Europe), the information about snow cover presence/absence is essential for estimating potential frost damage to winter crops. The snow cover model tested in this study could be easily implemented in most crop growth models and would enhance their performance in the Central European region. The up-to-date version of the model is freely available to users in the scientific community, which should allow for model testing in different climate conditions and application with various types of agrometeorological models