Tailored seasonal climate forecasts for crop breeding in the Nordic and Baltic regions

As climate change drives a northward shift in agro-climatic zones across Europe, it presents both risks and opportunities for agricultural production in the Nordic regions. Plant breeding plays a key role in adaptation strategies by enabling the development of climate-resilient crop varieties and exploiting novel growing conditions to secure yields. The NorBalFoodSec project aims at increasing food security in the Nordic and Baltic regions by advancing knowledge on how to better adapt crop breeding and agricultural production to future climates. As part of this effort, tailored seasonal climate forecasts for agri-food production are developed and their applicability and value in supporting crop breeders’ planning and decision-making in crop management are evaluated.  In this study, the predictability of key variables, i.e. temperature and precipitation for growing season, and the reliability assessment of the developed seasonal forecasts tailored for agri-food productions are presented.

To investigate the predictability limits of seasonal forecasts in the Nordic and Baltic region, we post-processed and evaluated the skill of temperature and precipitation from ECMWF’s SEAS5 seasonal forecast system using reforecasts for 1981-2016 and the ERA5 reanalysis dataset as reference. The analysis employed the open source CSTools package for R, which implements widely used methods from literature, ranging from the simple bias removal to the ensemble calibration methods that correct the bias, the overall forecast variance and ensemble spread. For precipitation, downscaling approaches such as the RainFarm stochastic method were tested to generate and assess higher-resolution fields. Furthermore, we explored EMOS (ensemble model output statistics), a nonhomogeneous regression technique widely used in short-range weather forecasting but less common in the post-processing of longer-range forecasts. Based on verification results, the most effective bias adjustment methods were applied to reduce the systematic errors in temperature and precipitation.

The post-processed variables were then used to develop growing season indicators, selected in close collaboration with crop breeders to meet their specific needs, such as the start of growing season, growing degree days, mean temperature, total precipitation and dry spell. The value of these seasonal forecasts is assessed using historical forecasts for 2017-2026 with a focus on years featuring hazardous conditions for key crops: cereal (barley), forage (red clover) and tubers (potatoes). Ultimately, these forecasts aim to support crop breeders in planning and decision-making for improved crop management.