Machine Learning-Based Insights into Environmental Determinants of Morchella importuna Growth in Muğla, Türkiye

Morchella species are edible mushrooms with high nutritional, ecological, and economic significance. Due to overharvesting and habitat loss, their natural populations face serious threats, which has led to growing interest in controlled cultivation. Although several studies have investigated Morchella growth, particularly M. importuna, key uncertainties remain regarding the environmental and nutritional requirements for optimal cultivation. Previous studies have largely focused on experimental approaches, but systematic machine learning analyses remain scarce. This study aims to fill this gap by applying machine learning algorithms to identify the most critical soil and climatic variables affecting M. importuna growth. To achieve this objective, using soil and climatic data from three localities in Türkiye, 20 features and M. importuna growth were analyzed by following a three-step procedure: (i) applying regression analysis based on eleven methods, (ii) selecting best fit model by using three goodness of fit measures, and (iii) calculating feature importances based on permutation importance and Shapley additive explanations methods. Gradient Boosting emerged as the best-performing algorithm, highlighting lime (%), phosphorus, precipitation, magnesium, minimum temperature, and maximum temperature as the most important predictors of growth. These results provide quantitative evidence that can guide cultivation strategies for morels, contributing to both conservation and commercial production.