Exploring key local influencing factors of arsenic content in suburban soils of Changchun city using the GeoShapley approach

ZOU Xin-ying; YAN Qing-wu; WU Zi-hao; ZHU Yuan-li; CHEN Yi-yun

PDF(1912 KB)

China Environmental Science ›› 2026, Vol. 46 ›› Issue (1) : 289-300.

Exploring key local influencing factors of arsenic content in suburban soils of Changchun city using the GeoShapley approach

ZOU Xin-ying¹, YAN Qing-wu¹, WU Zi-hao¹, ZHU Yuan-li¹, CHEN Yi-yun²

Author information +

History +

Abstract

sing 200soil As samples collected from suburban Changchun, this study combined the extreme gradient boosting tree (XGBoost) and the GeoShapley algorithm to explore the heterogeneity and nonlinear relationship between As and environmental factors, and to identify the local influencing factors and sources of As. The main conclusions are as follows: ① The XGBoost model had a fitted R² of 0.958, which was better than the geographically weighted regression and multiscale geographically weighted regression models, with fitted and predicted R² values of 0.958 and 0.729, respectively. ② The results of GeoShapley showed that the average contribution rate of manganese content ranked first and was the most significant characteristic variable affecting soil As, while the spatial location characteristic ranked second in importance, confirming the necessity of considering spatial coordinates. ③ The spatial distribution of GeoShapley values of influencing factors revealed that manganese content, alkali-hydrolyzable nitrogen, and sulfur content were the primary influencing factors for local As content at most sampling points, accounting for 99% of the samples. This indicated that parent material and agricultural production are the main sources of As. ④ The nonlinear relationship plots showed that there was a threshold effect between As content and various environmental variables, especially when the manganese content was higher than 2g/kg, the alkali-hydrolyzable nitrogen was higher than 175mg/kg, and the sulfur content was lower than 20mg/kg, As reached the peak, which should be paid special attention to in pollution prevention and control. This paper confirms the superiority of GeoShapley algorithm and the necessity of considering the spatial location, and provides data support for the governmental departments to formulate the policy of localized prevention and control of soil As pollution.

Key words

soil arsenic content / suburban area / GeoShapley / extreme gradient boosting tree / local influencing factor

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

ZOU Xin-ying, YAN Qing-wu, WU Zi-hao, ZHU Yuan-li, CHEN Yi-yun. Exploring key local influencing factors of arsenic content in suburban soils of Changchun city using the GeoShapley approach[J]. China Environmental Science. 2026, 46(1): 289-300

References

[1] 纪冬丽,孟凡生,薛浩,等.国内外土壤砷污染及其修复技术现状与展望[J]. 环境工程技术学报, 2016,6(1):90-99. Ji D L, Meng F S, Xue H, et al. Situation and prospect of soil arsenic pollution and its remediation techniques at home and abroad[J]. Journal of Environmental Engineering Technology, 2016,6(1):90-99.
[2] 梁富威,肖丽莎,李镇城,等.中国土壤砷污染与植物修复的研究进展[J]. 生态科学, 2024,43(6):250-258. Liang F W, Xiao L S, Li Z C, et al. Research progress of arsenic contamination in soil and phytoremediation in China[J]. Ecological Science, 2024,43(6):250-258.
[3] Chen Z, Zhang S, Geng W, et al. Use of geographically weighted regression (GWR) to reveal spatially varying relationships between Cd accumulation and soil properties at field scale[J]. Land, 2022,11(5):635.
[4] Wang Z, Xiao J, Wang L Q, et al. Elucidating the differentiation of soil heavy metals under different land uses with geographically weighted regression and self-organizing map[J]. Environmental Pollution, 2020, 260:114065.
[5] Yang L X, Meng F H, Chen M, et al. Elucidating the spatial determinants of heavy metals pollution in different agricultural soils using geographically weighted regression[J]. Science of the Total Environment, 2022,853:158628.
[6] Song X D, Brus D J, Liu F, et al. Mapping soil organic carbon content by geographically weighted regression: A case study in the Heihe River Basin, China[J]. Geoderma, 2016,261:11-22.
[7] Jas K, Dodagoudar G R. Explainable machine le-arning model for liquefaction potential assessment of soils using XGBoost-SHAP[J]. Soil Dynamics and Earthquake Engineering, 2023,165:107662.
[8] Wang J, Deng Y, Huang Z, et al. Identification of driving factors for heavy metals and polycyclic aromatic hydrocarbons pollution in agricultural soils using interpretable machine learning[J]. Science of the Total Environment, 2025,960:178384.
[9] Song X, Sun Y, Wang H, et al. Uncovering soil heavy metal pollution hotspots and influencing mechanisms through machine learning and spatial analysis[J]. Environmental Pollution, 2025,370:125901.
[10] Huang H, Huang S, Du Q. Evaluation of soil heavy metal pollution based on K-means and SVM[J]. International Journal of Environmental Science and Technology, 2023,20(11):12015-12024.
[11] Xu X, Liu W, Jia T, et al. Investigation of relationship between polycyclic aromatic hydrocarbons and human activities in urban soils of China using machine learning methods[J]. Journal of Cleaner Production, 2023,404:136839.
[12] Yan Y, Yang Y. Revealing the synergistic spatial effects in soil heavy metal pollution with explainable machine learning models[J]. Journal of Hazardous Materials, 2025,482:136578.
[13] Alam S M K, Li P, Rahman M, et al. Key factors affecting groundwater nitrate levels in the Yinchuan Region, Northwest China: Research using the eXtreme Gradient Boosting (XGBoost) model with the SHapley Additive exPlanations (SHAP) method[J]. Environmental Pollution, 2025,364:125336.
[14] 郑家桐,王鹏,石航源,等.基于CatBoost模型和SHAP解释方法的土壤重金属影响因素与程度定量分析[J]. 环境科学学报, 2023, 43(4):448-456. Zheng J T, Wang P, Shi H Y, et al, Quantitative analysis of the influence factors and extent of soil heavy metals based on CatBoost model and SHAP interpretation method[J]. Acta Scientiae Circumstantiae, 2023,43(4):448-456.
[15] Zhou B, Chen G, Yu H, et al. Revealing the nonlinear impact of human activities and climate change on ecosystem services in the karst region of southeastern Yunnan using the XGBoost–SHAP model[J]. Forests, 2024,15(8):1420.
[16] Li Z Q. GeoShapley: A game theory approach to measuring spatial effects in machine learning models[J]. Annals of the American Association of Geographers, 2024,114(7):1365-1385.
[17] 刘旭峰,陈文明,杨益民,等.基于GWR模型的湖南省耕地非粮化驱动因子及空间分异特征研究[J]. 湖南师范大学自然科学学报, 2025,48(2):93-101. Liu X F, Chen W M, Yang Y M, et al. Research on driving factors and differentiation characteristics of the spatial pattern of non-grain cultivated land in hunan province based on GWR model[J]. Journal of Natural Science of Hunan Normal University, 2025,48(2):93-101.
[18] 祝新明,宋小宁,冷佩,等.多尺度地理加权回归的地表温度降尺度研究[J]. 遥感学报, 2021,25(8):1749-1766. Zhu X M, Song X N, Leng P, et al, Spatial downscaling of land sur-face temperature with the multi-scale geographically weighted reg-ression[J]. National Remote Sensing Bulletin, 2021,25(8):1749-1766.
[19] 李春泽,张超,张皓源,等.基于XGBoost与地理加权回归的吉林省西部土壤盐渍化反演[J]. 中国农业大学学报, 2024,29(2):1-10. Li C Z, Zhang C, Zhang H Y, et al, Inversion of soil salinization in western Jilin Province based on XGBoost and Geographically Wei-ghted Regression[J]. Journal of China Ag-ricultural University, 2024,29(2):1-10.
[20] 罗妍,王枞,叶文玲.基于XGBoost和SHAP的急性肾损伤可解释预测模型[J]. 电子与信息学报, 2022,44(1):27-38. Luo Y, Wang C, Ye W L. An interpretable prediction model for acute kidney injury based on XGBoost and SHAP[J]. Journal of Electronics & Information Technology, 2022,44(1):27-38.
[21] 王桂兰,赵洪山,米增强.XGBoost算法在风机主轴承故障预测中的应用[J]. 电力自动化设备, 2019,39(1):73-7+83. Wang G L, Zhao H S, Mi Z Q. Application of XGBoost algorithm in prediction of wind motor main bearing fault[J]. Electric Power Automation Equipment, 2019,39(1):73-77,83.
[22] 刘瑞雪,李佳轩,李云.水库消落带植物多样性空间格局预测模型及环境解释——基于XGBoost-SHAP模型框架[J]. 生态学报, 2024,44(21):9652-9669. Liu R X, Li J X, Li Y. Prediction model and environmental interp-retation for the spatial pattern of plant diversity in the water-level fluctuation zone of reservoir based on XGBoost and SHAP[J]. Acta Ecologica Sinica, 2024,44(21):9652-9669.
[23] 崔帅,刘烁然,王寅,等.吉林省旱地土壤有效硫含量及其与土壤有机质和全氮的关系[J]. 中国农业科学, 2022,55(12):2372-2383. Cui S, Liu S R, Wang Y, et al. Soil available sulfur content in Jilin province and Its correlation with soil organic matter and soil total nitrogen[J]. Scientia Agricultura Sinica, 2022,55(12):2371-2383.
[24] 李桂华,王俊松,齐洁.上海住宅价格时空分异格局及其影响因素——基于地理加权回归模型的分析[J]. 地域研究与开发, 2024,43(6):74-81. Li G H, Wang J S, Qi J. Spatio-temporal differentiation pattern of housing prices in Shanghai and its influencing factors: Analysis based on GWR model[J]. Areal Research and Development, 2024,43(6):74-81.
[25] 赵慧,李新国,靳万贵,等.基于地理加权回归模型的博斯腾湖湖滨绿洲土壤盐分离子含量高光谱估算[J]. 土壤, 2021,53(3):646-653. Zhao H, Li X G, Qi W G, et al. Hyperspectral estimation of soil salt ion contents in lakeside oasis of bosten lake based on Geographical weighted regression model[J]. Soils, 2021,53(3):646-653
[26] 栗珂珂,周詹杭,王真.中国土壤侵蚀的社会经济空间驱动因素研究:基于多尺度地理加权回归模型的分析[J]. 华中农业大学学报, 2024,43(6):29-38. Li K K, Zhou Z H, Wang Z. Socioeconomic spatial driving factors of soil erosion in China: An analysis based on a multi-scale geogr- aphically weighted regression model[J]. Journal of Huazhong Agricultural University, 2024,43(6):29-38.
[27] 吴子豪,刘耀林,冯向阳,等.基于多尺度地理加权回归的土壤镉污染局部影响因子分析[J]. 地球信息科学学报, 2023,25(3):573-587. Wu Z H, Liu Y L, Feng X Y, et al. Analysis of local influencing factors of cadmium pollution in Soil by using multi-scale geographically weighted regression[J]. Journal of Geoinformation Science, 2023,25(3):573-587.
[28] 姚雯雯,赵新正,马海涛,等.基于多尺度地理加权回归模型的中国城市网络外部性的空间异质性比较研究[J]. 地理研究, 2024,43(5): 1073-1087. Yao W W, Zhao X Z, Ma H T, et al. Ancomparative study on spat-ial heterogeneity of China's urban network externality based on Multiscale Geographically Weighted Regression Model[J]. Geo- graphical Research, 2024,43(5):1073-1087.
[29] 胡贵贵,杨粉莉,杨联安,等.基于主成分和机器学习的土壤有机质含量空间预测建模[J]. 干旱区地理, 2021,44(4):1114-1124. Hu G G, Yang F L, Yang L A, et al. Spatial prediction modeling of soil organic matter content based on principal c-omponents and m-achine learning[J]. Arid Land Geography, 2021,44(4):1114-1124.
[30] 马国林,丁建丽,韩礼敬,等.基于变量优选与机器学习的干旱区湿地土壤盐渍化数字制图[J]. 农业工程学报, 2020,36(19):124-131. Ma G L, Ding J L, Han L J, et al. Digital mapping of soil saliniz-ation in arid area wetland based on variable optimized selection and machine learning[J]. Transactions of the Chinese Society of Agricultural Engineering, 2020,36(19):124-131.
[31] 沈佳丽,陈颂超,胡碧峰,等.基于机器学习的江汉平原土壤有机碳预测及制图[J]. 农业资源与环境学报, 2023,40(3):644-650. Shen J L, Chen S C, Hu B F, et al. Prediction and mapping soil organic carbon in Jianghan Plain by machine learning[J]. Journal of Agricultural Resources and Environment, 2023,40(3):644-650.
[32] Ke E, Zhao J, Zhao Y. Investigating the influence of nonlinear spatial heterogeneity in urban flooding factors using geographic explainable artificial intelligence[J]. Journal of Hydrology, 2025,648:132398.
[33] Zhu Y L, Liu B, Jin G, et al. Identifying the local influencing factors of arsenic concentration in suburban soil: A multiscale geographically weighted regression approach[J]. Toxics, 2024,12(3):229.
[34] Zhang R J, Jiang L H, Dong T H, et al. Effects of geographical and soil factors on soils arsenic levels: a case study in typical arsenic- contaminated paddy fields based on machine learning[J]. Environmental Management, 2025,75(9):2272-2286.
[35] 安礼航,刘敏超,张建强,等.土壤中砷的来源及迁移释放影响因素研究进展[J]. 土壤, 2020,52(2):234-246. An L H, Liu M C,Zhang J Q, et al. Sources of Arsenic in Soil and Affecting Factors of Migration and Release: A Review[J]. Soils, 2020,52(2):234-246.
[36] 黄文妍,杨文俊,汤叶涛,等.土壤环境中铁锰氧化物对重金属的吸附络合机制及其影响因素研究进展[J]. 地球环境学报, 2024,15(11):1-13. Huang W Y, Yang W J, Tang Y T, et al. Research progress on the mechanism for the adsorption and complexation of heavy metals onto iron and manganese oxides and its influencing factors in soil environment[J]. Journal of Earth Environment, 2024,15(11):1-13.
[37] 王鑫浩,范芳,张洛红,等.锰氧化物对环境中重金属吸附及影响因素研究进展[J]. 应用化工, 2018,47(1):155-9+64. Wang X H, Fan F, Zhang L H, et al. Research progress on adsor-ption of heavy metals by manganese oxides and its influencing factors[J]. Applied Chemical Industry, 2018,47(1):155-9+64.
[38] 李月芬,王冬艳,汤洁,等.吉林西部土壤砷的形态分布及其与土壤性质的关系研究[J]. 农业环境科学学报, 2012,31(3):516-522. Li Y F, Wang D Y, Tang J, et al. Speciation of soil arsenic and its correlation with soil properties in western Jilin province,China[J]. Journal of Agro-Environment Science, 2012,31(3):516-522.
[39] Zhai Y, Cao X, Xia X, et al. Elevated Fe and Mn concentrations in groundwater in the Songnen Plain, Northeast China, and the Factors and Mechanisms Involved[J]. Agronomy, 2021,11(12):2392.
[40] Yang Z P, Lu W X, Long Y Q, et al. Assessment of heavy metals contamination in urban topsoil from Changchun City, China[J]. Journal of Geochemical Exploration, 2011,108(1):27-38.
[41] 张彦儒,孙思佳,王雅君,等.MnSO₄与钝化材料联合施用对镉污染稻田的修复效果[J]. 农业环境科学学报, 2025,https://link.cnki. net/urlid/12.1347.S.20250529.1642.004. Zhang Y R, Sun S J, Wang Y J, et al. Effect of combined application of MnSO₄ and passivation materials on the remediation performance of cadmium-contaminated rice fields[J]. Journal of Agro- Environment Science, 2025, https://link.cnki.net/urlid/12.1347.S.20250529.1642.004.
[42] 郭炳跃,杨锟鹏,张璟,等.二氧化锰/氨基改性生物炭对铅、镉复合污染土壤的钝化修复研究[J]. 生态与农村环境学报, 2023,39(3): 422-428. Guo B Y, Yang K P, Zhang J, et al. Research on the remediation effect of manganese dioxide/aminomodified biochar on Pb and Cd composite contaminated soil[J]. Journal of Ecology and Rural Environment, 2023,39(3):422-428.
[43] 贾正雷.土壤砷和氮含量的空间变异及其相互关系研究[D]. 广州:华南农业大学, 2016. Jia Z L. Study on spatial variability and relationship of soil arsenic and soil nitrogen[D]. GuangZhou: South China Agricultural University, 2016.
[44] 成杭新,李括,李敏,等.中国城市土壤化学元素的背景值与基准值[J]. 地学前缘, 2014,21(3):265-306. Cheng H X, Li K, Li M, et al. Geochemical background and base-line value of chemical elements in urban soil in China[J]. Earth Science Frontiers, 2024,21(3):265-305.
[45] Barkhordari M S, Qi C. Prediction of soil arsenic concentration in European soils: A dimensionality reduction and ensemble learning approach[J]. Journal of Hazardous Materials Advances, 2025,17:100604.