Quantifying germination cardinal temperatures of ten forage legumes using non-linear regression models

Both germination percentage and germination speed were low in hot and cold temperatures outside the range of 15-25°C for these species. Notably, significant inter-species variation in thermal requirements was identified. Alfalfa (<i>Medicago sativa</i>), yellow medick (<i>Medicago falcata</i>), and erect milkvetch (<i>Astragalus adsurgens</i>) demonstrated high thermal plasticity, characterized by low minimum (base) temperature (<i>T_b</i> ), high maximum temperature (<i>T_m</i> ), and wide thermal ranges. Conversely, sweet clover (<i>Melilotus officinalis</i>) exhibited a preference for cooler thermal regimes, with the lowest optimum temperature (<i>T_o</i> ) and <i>T_m</i> thresholds. Niuzhizi (<i>Lespedeza potaninii</i>) and white clover (<i>Trifolium repens</i>) were identified as thermophilic species, requiring higher temperatures for optimal germination, as evidenced by their higher <i>T_o</i> and <i>T_m</i> thresholds, whereas species of the genus <i>Vicia</i> and red clover (<i>Trifolium pratense</i>) were best adapted to moderate thermal environments. Furthermore, germination speed-related parameters were more sensitive to temperature fluctuations than germination percentage. Among the tested models, the intersected-lines model showed superior performance for predicting the cardinal temperatures of the ten forage legumes, as indicated by a lower root mean square of error (RMSE) and a higher coefficient of determination (<i>R²</i>). No statistically significant relationships between cardinal temperatures and germination parameters were observed. These findings provide a scientific basis for improving sowing practices and regional species selection, which is vital for forage legume cultivation and production. The identification of cardinal temperatures is also crucial for the development of plant growth and biomass prediction models that simulate growth and estimate yield under current and future climate scenarios.