The Agricultural Resources Utilization and Zoning Team at the Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences (CAAS), recently advanced crop planting optimization. Their research, "Evaluating spatial heterogeneity and suitability patterns for soybean cultivation in Heilongjiang: A productivity-oriented approach," was published in the Journal of Cleaner Production.

China’s reliance on soybean imports, exceeding 80% for ten years, threatens food security. Given the limited arable land, expanding soybean cultivation in suitable regions is considered a key strategy to enhance domestic self-sufficiency. Focusing on Heilongjiang Province, China’s major soybean-producing region, the study integrated time-series remote sensing vegetation indices to construct a grid-scale soybean yield prediction model, and systematically compared the predictive performance of LightGBM and Random Forest (RF) models.

The RF model showed higher accuracy. By incorporating the slope of vegetation index trends during critical growth stages—particularly pod-setting and maturity—the model reduced the root mean square error (RMSE) by 9.8% relative to static indices alone, thereby significantly improving yield prediction accuracy.

Building on these predictions, the research further integrated spatial autocorrelation analysis with an improved MaxEnt model to identify high-yield areas as suitable for soybean cultivation and to quantify the associated potential for carbon emission reduction. The results reveal a mismatch between resource endowment and actual planting: in 2021, only 65.3% of the provincial soybean planting area was in suitable zones. This indicates that concentrating planting in high-suitability areas could increase yield per unit area, reduce agricultural input waste, and achieve certain carbon emission reductions.

This research advances traditional MaxEnt models, which rely solely on occurrence points, by incorporating productivity as a key indicator of cultivation suitability. By integrating natural environmental conditions and anthropogenic activity characteristics, a more robust and accurate framework for evaluating soybean planting suitability was established. The findings provide important scientific guidance and technical support for expanding soybean cultivation in suitable regions, optimizing crop layouts, and enhancing food self-sufficiency in China.

The study's first author is Associate Researcher Hu Mengmeng, with Researcher Yin Changbin as corresponding author. The research was supported by the National Key Laboratory of Efficient Utilization of Dryland and Semi-Arid Farmland in Northern China, the National Natural Science Foundation of China, the National Key R&D Program, and the National Social Science Fund.

Original link: https://doi.org/10.1016/j.jclepro.2025.147264