Crop yield outlooks in the Iberian Peninsula: Connecting seasonal climate forecasts with crop simulation models

Seasonal climate prediction can potentially contribute to achieving a more resilient cropping system management. This can help alleviate food insecurity and the economical sustainability of farming at large. For this purpose, seasonal climate forecasts are used to generate crop forecasts. This study assesses two methods for linking seasonal climate forecasts with crop models to improve crop yield predictability in the Iberian Peninsula (IP). Crop models usually require daily weather data and therefore, we tested two methods to disaggregate seasonal climate forecasts into daily weather realizations: (1) a conditional stochastic weather generator (predictWTD) and (2) a simple forecast probability resampler (FResampler1). These methods were evaluated under three seasonal rainfall forecasts by analyzing the impacts on rainfed wheat yield and on irrigation requirements and yields of maize crop. In addition, we estimated the gross margins ((sic) ha(-1)) and the production risks associated with contrasting scenarios of seasonal rainfall forecasts (dry and wet). Both methods provided comparable predictability and therefore, both seem feasible options for using seasonal forecasts to establish yield forecasts and irrigation requirements. The large impact of crop prices and of irrigation cost on gross margins for both crops suggests that using a combination of information on expected market prices and crop forecast based on seasonal climate forecasts can be an effective tool for farmer's decision-making, especially under dry forecast situation and/or in locations with low annual precipitation. These methods can help to quantify the benefits and risks from the seasonal weather forecasts to farmers in the IP. The anticipation of risks and the opportunity that skillful climate and crop forecast provide allows for windows of opportunity to prepare and preempt mitigating actions. (C) 2016 Elsevier Ltd. All rights reserved.