Regularised canonical correlated prediction of fertility responses to weather for optimised artificial insemination strategies in Murciano-Granadina does

This study examined how climatic variability affects reproductive performance in Murciano-Granadina does, analysing 21,757 animals and 32,693 artificial insemination (AI) records collected from 2010 to 2019. Comprehensive weather data (temperature, rainfall, wind speed and gusts, wind direction, barometric pressure, photoperiod and altitude) were integrated with fertility outcomes to model environment–reproduction interactions. Fertility was assessed using three indexes: daily fertility rate, fertility per buck batch per day, and fertility per semen type (fresh vs. frozen). Fresh semen achieved higher mean fertility (56.17%) than frozen semen (37.75%) with lower variability (SD 16.52 vs. 15.96), confirming its superior integrity. Regularised canonical correlation analysis (rCCA) revealed strong multivariate links between climatic factors and fertility. For frozen semen, two canonical functions explained 66.22% of shared variance (F1 = 0.230; F2 = 0.164), indicating sensitivity to wind instability and low temperatures. Fresh semen showed a more stable pattern, with the first function alone explaining 96.79% of variance (F1 = 0.277); fertility was optimal at minimum temperatures >11.35 °C, maximum <25.7 °C, altitudes <489 m and rainfall <2.5 mm. Wind gusts above 4.5 m/s negatively affected conception, especially for cryopreserved semen. Cross-validation confirmed model robustness, with higher predictive accuracy for frozen semen (CV = 0.642) than for fresh semen (0.247), reflecting greater environmental sensitivity of cryopreserved gametes. These findings show that climatic data can inform AI scheduling, enabling semen-type-specific management to improve fertility, accelerate genetic gain and enhance the sustainability of breeding programs under variable Mediterranean conditions.