The Phaseolus vulgaris L. landraces distribution patterns in Madeira: a modelling approach with maxent

Abstract Common bean (Phaseolus vulgaris L. ) is a widely cultivated legume and a key source of dietary protein globally. This study employed MaxEnt modelling to predict the potential distribution of P. vulgaris landraces on Madeira Island, Portugal, where local landraces are predominant. A dataset of 247 occurrence records (2000—2010) was integrated with environmental variables, including altitude, temperature (average, minimum, and maximum), precipitation, evapotranspiration parameters, and soil types. The model demonstrated strong predictive performance (AUC = 0.843) with no evidence of overfitting. Jackknife analysis identified altitude as the most influential variable, contributing 76.7% to the model's gain. Although temperature—particularly maximum temperature—was also important, its effects were closely correlated with altitude, which drives temperature gradients across the island. While precipitation alone had a limited influence, the most suitable areas aligned with zones of higher accumulated rainfall, suggesting a complementary role when thermal and altitudinal conditions are favourable. Response curves indicated that P. vulgaris favours lower altitudes, warmer climates (16—23 °C), and regions with elevated real evapotranspiration. Suitability maps confirmed that the northern coastal areas, characterized by low elevations, moderate temperatures, and higher precipitation, offer the most favourable conditions for cultivation. These findings provide valuable insights for the conservation and sustainable management of P. vulgaris agrobiodiversity in Madeira.