Evaluation method of chestnut planting suitability based on maximum entropy model

【Objective】Chestnut is a key economic tree species in northern China, possessing both nutritional and medicinal benefits. As one of Beijing’s major specialty agricultural products, chestnut cultivation in Huairou District holds significant economic and ecological value. However, due to its high sensitivity to environmental conditions—such as elevation, temperature, precipitation, soil chemistry, and vegetation cover—its suitability for cultivation varies widely across the region. This research aims to assess the spatial suitability of chestnut cultivation in Huairou District using a comprehensive modeling framework. The main objective is to identify highly suitable and potential expansion areas, thereby supporting local agricultural optimization, ecological protection, and rural revitalization strategies. 【Methods】A total of 439 chestnut distribution points were collected from official databases and field surveys. Sixty-two environmental covariates were assembled, including topographic (elevation, slope, river distance, topographic wetness index), climatic (19 bioclimatic variables from the WorldClim database), soil (physicochemical and trace element properties), and vegetation (NDVI) data. All spatial datasets were resampled to a 30-meter resolution. The Maximum Entropy (MaxEnt) model was employed to assess species-environment relationships. The model was optimized using the ENMeval package in R, evaluating 48 combinations of feature classes (e.g., linear, quadratic, hinge) and regularization multipliers (RM) ranging from 0.5 to 4. The best parameter configuration-linear and quadratic features (LQ) with RM = 1-was selected based on delta. AICc values. Model performance was validated using the ar-ea under the ROC curve (AUC), based on 10 bootstrap replicates. To improve model interpretability and address multicollinearity, Pearson Correlation Analysis was applied to filter environmental variables. Sixteen core variables were retained. Fuzzy membership functions were constructed for each variable, employing sigmoid, parabolic, or discrete functions based on their response curves. The entropy weight method was then used to calculate objective weights for each variable, further revised through expert consultation. For spatial evaluation, land use and soil type data were integrated to create 53, 333 evaluation units. A composite suitability index was computed for each unit using a weighted summation method. The index was classified into three categories: highly suitable (≥ 0.701), moderately suitable (0.509- 0.701), and unsuitable (＜0.509).【Results】The MaxEnt model achieved an AUC value of 0.918, indicating excellent predictive accuracy. Among the 16 key variables, the most influential were distance to river networks (26.3% ), temperature seasonality (17.6% ), elevation (9.4% ), slope (9.3% ), NDVI (7.5%), and parent material (6.8%). These variables reflect the importance of water availability, temperature stability, terrain constraints, and soil- forming processes in chestnut distribution. The suitability assessment indicated that 9.93% of Huairou District was classified as highly suitable for chestnut cultivation and 21.77% as moderately suitable, while the remaining 68.30% was deemed unsuitable. Overlay validation showed that 56.49% of existing chestnut plantations were located in highly suitable zones, 36.61% in moderately suitable zones, and only 6.90% in unsuitable areas, indicating strong consistency between model predictions and actual planting patterns. Further analysis revealed approximately 54 900 hectares of potential chestnut development areas outside current cultivation zones, including 15 300 hectares of highly suitable land and 39 600 hectares of moderately suitable land.【Conclusion】 This study demonstrates the effectiveness of an integrated MaxEnt-fuzzy evaluation-entropy weighting framework for agricultural suitability analysis at a regional scale. By incorporating 62 environmental covariates and optimizing modeling parameters, the study achieved high- precision mapping of chestnut suitability in a complex mountainous environment. The model not only accurately identified current optimal planting zones but also highlighted substantial development potential in underutilized areas. The results suggest that chestnut distribution in Huairou is strongly influenced by terrain, climate variability, and soil nutrient conditions. Particularly, factors such as proximity to water sources, elevation, and temperature seasonality are decisive. These insights are consistent with physiological knowledge of chestnut growth and provide robust guidance for regional agricultural planning. From an ecological perspective, the identified suitable zones overlap significantly with hilly and mountainous areas where chestnut plantations can contribute to water conservation, erosion control, and biodiversity enhancement. From a socio-economic perspective, the spatial zoning results can support targeted policy implementation, rural income generation, and sustainable land-use strategies under the framework of China’s rural revitalization and ecological civilization initiatives. Overall, the study contributes a replicable and data- driven methodology for crop suitability analysis and offers valuable spatial decision support for stakeholders involved in land planning, agricultural modernization, and ecological restoration.