Characterization of spatial and temporal combinations of climatic factors affecting yields: An empirical model applied to the French barley belt

The adaptation of genotypes to environmental conditions is one of the main levers for maintaining an acceptable level of production, both in terms of quality and quantity. To breed suitable genotypes and for the farmers to choose the most adapted one to his farm conditions, the factors affecting production must be precisely characterized. Here, we analyzed the impacts of the climatic factors on winter barley yield in 35 departements (French geographic units) over 25 years, by partial least squares (PLS) regression analysis. Using ascendant hierarchical clustering based on PLS results, we defined the main combinations of climatic factors affecting yield in the French barley belt, hereafter referred as climatic-stress patterns (CPs). Four CPs captured 27% of total yield variability and widely differ in term of yield. Crops experienced low winter rainfall and few days with heat stress (31.8% of environments- mean yield of 7.2 t ha(-1)), high winter frost levels and high amounts of precipitation during stem elongation (34.7% of environments- mean yield of 6.5 t ha(-1)), high temperature during grain filling either with low vernalization temperatures (28.9% of environments-mean yield of 6.2 t ha(-1)), or high winter rainfall (4.6% of environments -mean yield of 5.5 t ha(-1)). Two-thirds of the French departements experienced all the four CPS over the years studied. Three clusters of regions with homogeneous frequencies of the CPs were identified. The regions with similar climatic-stress patterns could help breeders to design genotypes better adapted to the different local French growing conditions. It could also help farmers to choose the most appropriate cultivars to grow.