Mapping cropland suitability in China using optimized MaxEnt model

Assessing farmland suitability is a crucial step towards the rational utilization of land resources and ensuring sustainable agricultural development. Despite China's position as a major producer and consumer of agricultural products, a comprehensive map of farmland suitability based on crop suitability has yet to be produced. Additionally, significant changes have occurred in China's agricultural production space over the past few decades, and research on the correlation between these changes and farmland suitability is lacking, posing a serious threat to agricultural security and sustainable land use. This study aims to investigate the spatial suitability of six crops, namely rice, wheat, maize, soybean, rape, and cotton, and to determine the corresponding range of farmland suitability. Additionally, this research aims to analyze the degree of matching between land cover and farmland suitability in both current and long-term time series to identify inappropriate land use and the potential for flexible farmland allocation. In this study, we applied the optimized maximum entropy (MaxEnt) model to simulate the spatial distribution of crop suitability by integrating six crop occurrence records and environmental variables. The land suitability space was determined by computing the suitability results for all six crops using the maximum-value compositing algorithm. The optimized MaxEnt model significantly enhanced the model's performance and transferability. Environmental variables influenced the proportion of suitable area for the six crops, with soybean > rape > maize > wheat > rice > cotton in descending order. In China, 64.96% of the land was suitable for cultivation, with 22.64% being highly suitable. In 2020, the farmland areas that were highly suitable, moderately suitable, poorly suitable, and unsuitable for crop cultivation accounted for 72.02%, 21.51%, 4.25%, and 2.22% of the total farmland area, respectively. The current farmland area that was highly suitable reached 135.93 million ha, which was sufficient to meet China's cultivated land preservation goal of 1.8 billion mu (120 million hectares). Of the woodland, grassland, and unused land, 28.64 million ha, 24.53 million ha, and 2.38 million ha, respectively, were found to be highly suitable for conversion into cropland. Over the past 21 years, built-up land has occupied 8.39 million ha of cropland, 86.65% of which was high-suitability cropland. In addition, 1.35 million ha of unused land were converted to farmland, with 69.80%, 17.32%, and 9.11% being highly suitable, moderately suitable, and poorly suitable for crop cultivation, respectively, while 3.77% of the converted unused land was unsuitable for planting. The optimized MaxEnt model shows significant potential for predicting the distribution of crop suitability. Crop suitability is determined by a combination of climatic, topographic, geological, pedological, and hydrological factors. Currently, the distribution of cropland in China matches well with its suitability, and there is still sufficient suitable land available for agricultural production. However, the implementation of the "Farmland Balance Policy" has not effectively curbed the loss of high-quality farmland, and the available reserve resources for cultivation are becoming scarce. The crop suitability spatial analysis in our study may assist farmers in adjusting their crop cultivation strategies, and the farmland suitability spatial analysis may aid managers in reconciling conflicts between farmland protection and economic development to ensure sufficient land is available for agricultural production.