基于随机森林的区域土壤重金属污染影响因素识别研究

崔释文; 陈坚; 廖磊; 周睿; 钱江波; 黄国鑫

doi:10.13205/j.hjgc.202508018

基于随机森林的区域土壤重金属污染影响因素识别研究

doi: 10.13205/j.hjgc.202508018

崔释文^1,2,
陈坚²,
廖磊³,
周睿¹,
钱江波⁴,
黄国鑫^2, ,

1. 吉林大学新能源与环境学院, 长春 130021;
2. 生态环境部环境规划院, 北京 100041;
3. 核工业二九〇研究所, 广东韶关 512029;
4. 浙江科环环境工程技术有限公司, 杭州 311200

基金项目:

国家重点研发计划（2023YFC3708901）

详细信息

作者简介:
崔释文（2000—），女，硕士研究生，主要研究方向为土壤和地下水污染治理。1176668387@qq.com

通讯作者:
黄国鑫（1980—），男，研究员，主要研究方向为土壤和地下水污染防治。huanggx@caep.org.cn

计量
- 文章访问数: 55
- HTML全文浏览量: 15
- PDF下载量: 1
- 被引次数: 0
出版历程
- 收稿日期: 2024-12-03
- 录用日期: 2025-01-20
- 修回日期: 2025-01-14

Identification of influencing factors of regional soil heavy metal pollution based on random forest

1. College of New Energy and Environment, Jilin University, Changchun 130021, China;
2. Chinese Academy of Environmental Planning, Beijing 100041, China;
3. Research Institute No. 290, CNNC, Shaoguan 512029, China;
4. Zhejiang Kehuan Environmental Engineering Technology Co., Ltd., Hangzhou 311200, China

摘要

摘要: 准确识别重金属污染来源是土壤重金属污染防治的重要前提。由于缺乏土壤重金属污染影响因素有关信息，污染溯源分析的效果常受到限制。以广东省某典型工业地区为研究对象，基于577个采样点的土壤重金属浓度实测数据和18项环境协变量数据，采用随机森林（RF）和双变量局部空间自相关方法，识别了土壤Cd、Pb和Cr浓度的影响因素，确定了18项影响因素对Cd、Pb和Cr浓度的定量贡献，并提出了有针对性的土壤环境管理对策。结果表明：3种重金属浓度的预测模型决定系数分别为0.93（Cd）、0.97（Pb）、0.93（Cr）时，RF预测模型性能最佳；3种重金属相对浓度空间分布存在差异；Cd的主要影响因素为铁路（贡献率0.119），Pb的主要影响因素为土壤pH值（贡献率0.099），Cr的主要影响因素为危险废物集中处置场（贡献率0.100）；相较于研究区的其他地区，中部的3种重金属浓度较高，人类活动复杂，且他们与其最主要影响因素的高-高区较多。在实施土壤重金属污染源头防控时应优先关注中部。
- 区域尺度 /
- 土壤 /
- 重金属 /
- 随机森林 /
- 影响因素 /
- 双变量局部莫兰指数
Abstract: Accurate identification of heavy metal pollution sources is an important prerequisite for soil pollution prevention and control. However， due to the lack of information on the influencing factors of heavy metal pollution， the effectiveness of pollution source tracing analysis is often constrained. In this paper， a typical industrial area was taken as the study area. On the basis of the actual measurements of soil heavy metal concentrations from 577 sampling points and data on the 18 environmental covariates， random forest （RF） and bivariate local spatial autocorrelation methods were applied to identify the influencing factors of the Cd， Pb， and Cr concentrations， determine the quantitative contributions of the 18 influencing factors to these heavy metal concentrations， and further propose targeted strategies for soil environmental management in an industralized study area. The results showed that the optimal performance of the RF prediction models for the three heavy metals was achieved when the coefficient of determination （R²） reached 0.93 with a root mean square error （RMSE） of 0.43 mg/kg for Cd， 0.97 with an RMSE of 48.57 mg/kg for Pb， and 0.93 with an RMSE of 18.57 mg/kg for Cr，namely. There were differences in the spatial distributions of the relative concentrations of the three heavy metals. The regions with relatively high Cd concentration were concentrated in the central and southern parts of the study area， whereas relatively high Pb relative concentration were predominantly found in the central and eastern parts. The regions with relatively high Cr relative concentration were concentrated in the southwestern and northeastern parts. Railway was identified as the most significant factor influencing Cd concentration with a contribution rate of 0.119. Soil pH was identified to be the most significant factor influencing Pb concentrations with a contribution rate of 0.099. The hazardous waste disposal site was identified to be the most significant factor influencing Cr concentrations with a contribution rate of 0.100. Compared to the rest of the study area， the central region exhibited higher concentrations of Cd， Pb， and Cr， more complex human activity， and a greater number of high-high cluster zones， where Cd， Pb， and Cr concentrations and their most significant influencing factors were concentrated. When source prevention and control measures for heavy metal pollution in soil were implemented， particular attention should be prioritized towards the central region of the study area.
- regional scale /
- soil /
- heavy metals /
- random forest /
- influencing factors /
- bivariate local Moran index

HTML全文

参考文献(34)

[1]	王勇淏.土壤重金属污染特征与状况调查分析[J].西部探矿工程，2024，36（10）：12-15. WANG Y H. Investigation and analysis of soil heavy metal pollution characteristics and conditions [J]. West-China Exploration Engineering，2024，36（10）：12-15.
[2]	董志芬，梁策，谢容生，等.有色冶炼企业地块非饱和带重金属分布迁移特征及影响因素[J].有色金属（冶炼部分），2023（11）：135-143. DONG Z F， LIANG C， XIE R S， et al. Distribution and migration characteristics and influencing factors of heavy metals in unsaturated zone of metallurgical enterprise[J]. Site Nonferrous Metals（Extractive Metallurgy），2023（11）：135-143.
[3]	ADHIKARI K， MANCINI M， LIBOHOVA Z， et al. Heavy metals concentration in soils across the conterminous USA： Spatial prediction，model uncertainty，and influencing factors [J]. Science of the Total Environment，2024，919：170972.
[4]	VASUDHEVAN P，MANIKANDAN E，JONATHAN M P，et al. Pollution assessment and source apportionment of metals in paddy field of Salem，South India[J]. Environmental Earth Sciences， 2022，81（6）：184.
[5]	张雨涵，李瑜，官开江，等.重庆某农业集中区土壤重金属污染评价与来源解析[J].中国土壤与肥料，2024（7）：10-18. ZHANG Y H，LI Y，GUAN K J，et al. Pollution assessment and source analysis of soil heavy metals in a concentrated distribution area of agricultural in Chongqing[J]. Soil and Fertilizer Sciences in China，2024（7）：10-18.
[6]	杨昱莹，刘亮，陈明，等.长三角地区南京市表土重金属污染特征及源解析[J].中国环境科学，2024，44（7）：3910-3918. YANG Y Y，LIU L，CHEN M，et al. Characterization and source analysis of topsoil heavy metal pollution in Nanjing，Yangtze River Delta Region[J]. China Environmental Science，2024，44（7）：3910-3918.
[7]	岳聪，黄顺寅，屠春宝，等.植物修复重金属镉污染土壤研究进展[J].现代农业科技，2024（17）：129-135. YUE C，HUANG S Y，TU C B，et al. Research progress on phytoremediation of soil contaminated by heavy metal cadmium [J]. Modern Agricultural Science and Technology，2024（17）： 129-135.
[8]	李军，高占栋，马利邦，等.黄河兰州段城市饮用水源地土壤重金属污染及其溯源的多指标综合分析[J].环境科学， 2024，45（11）：6723-6733. LI J，GAO Z D，MA L B，et al. Multiproxy comprehensive analysis for source apportionment and pollution of heavy metals in urban drinking-water source soils from the Lanzhou Reach of the Yellow River[J]. Environmental Science，2024，45（11）：6723-6733.
[9]	赵恒谦，常仁强，金倩，等.河北西石门铁矿区土壤重金属污染空间分析及风险评价[J].岩矿测试，2023，42（2）： 371-382. ZHAO H Q，CHANG R Q，JIN Q，et al. Spatial analysis and risk assessment of soil heavy metal pollution in the Xishimen iron mining area of Hebei Province[J]. Rock and Mineral Analysis， 2023，42（2）：371-382.
[10]	宋炉生，孙振洲，胡晶，等.废弃铁矿及下游农田土壤重金属污染特征及来源解析[J].环境工程，2024，42（10）： 155-164. SONG L S，SUN Z Z，HU J，et al. Pollution characteristics and source apportionment of heavy metals in an abandoned iron ore and downstream farmland soil[J]. Environmental Engineering， 2024，42（10）：155-164.
[11]	董曼慧，夏卫生，周浩，等.湖南省地块尺度土壤重金属污染及成因研究[J].环境科学与技术，2023，46（增刊1）：60-66. DONG M H，XIA W S，ZHOU H，et al. Study on the pollution status and causes of heavy metals in soil at plot scale in Hunan Province[J]. Environmental Science & Technology，2023，46 （S1）：60-66.
[12]	孙思静，董春雨，张好，等.基于PMF模型和地理探测器的土壤重金属源解析及影响因素分析[J].环境科学，2024，45（9）：5474-5484. SUN S J，DONG C Y，ZHANG H. Source and influencing factor analysis of soil heavy metals based on PMF model and GeoDetector[J]. Environmental Science，2024，45（9）：5474-5484.
[13]	耿治鹏，宋颉，王春林，等.污染场地土壤重金属污染空间特征分析—以某搬迁电镀厂为例[J].环境工程技术学报， 2023，13（1）：295-302. GENG Z P，SONG J，WANG C L，et al. Spatial characteristics of soil heavy metal pollution in polluted sites：taking a relocated electroplating factory as an example [J]. Journal of Environmental Engineering Technology，2023，13（1）：295-302.
[14]	ZHENG J， WANG P， SHI H， et al. Quantitative source apportionment and driver identification of soil heavy metals using advanced machine learning techniques[J]. Science of the Total Environment，2023，873：162371.
[15]	SHI T，MA J，WU X，et al. Inventories of heavy metal inputs and outputs to and from agricultural soils：a review[J]. Ecotoxicology and Environmental Safety，2018，164：118-124.
[16]	许洋，陈健松，王志栋，等.基于多源异构数据的典型场地土壤重金属污染模拟预测研究[J].环境科学学报，2023，43（9）：357-68. XU Y，CHEN J S，WANG Z D，et al. Simulation and prediction research of heavy metal pollution in soil of typical sites based on multi-source heterogeneous data [J]. Acta Scientiae Circumstantiae，2023，43（9）：357-68.
[17]	李珊，杨济妮，苏贵金，等.基于Catboost算法的中国典型农业区重金属污染特征及影响因素分析[J].环境化学，2024， 43（10）：3377-3387. LI S，YANG J N，SU G J，et al. Analysis of heavy metal pollution characteristics and influencing factors in China's typical agricultural areas based on Catboost algorithm [J]. Environmental Chemistry，2024，43（10）：3377-3387.
[18]	GUI H， YANG Q， LU X， et al. Spatial distribution， contamination characteristics and ecological-health risk assessment of toxic heavy metals in soils near a smelting area[J]. Environmental Research，2023，222：115328.
[19]	冯锋，王育红，左雨芳.基于RF-XGBoost的土壤镉污染影响因子及空间分布研究[J].农业环境科学学报，2023，42（4）： 811-819. FENG F，WANG Y H，ZUO Y F. A study on factors that influence the spatial distribution of soil cadmium pollution based on RF-XGBoost [J]. Journal of Agro-Environment Science， 2023，42（4）：811-819.
[20]	王爱琼，萨日娜，张巧丽，等.基于随机森林的川崎病诊断关键实验室指标预测研究[J].检验医学与临床，2024，21（19）：2887-2891. WANG A Q，SA R N，ZHANG Q L，et al. Prediction of key laboratory indicators for diagnosis of Kawasaki disease based on random forest [J]. Laboratory Medicine and Clinic，2024，21（19）：2887-2891.
[21]	BI Z，SUN J，XIE Y，et al. Machine learning-driven source identification and ecological risk prediction of heavy metal pollution in cultivated soils[J]. Journal of Hazardous Materials， 2024，476：135109.
[22]	ZHOU M，LIAO B，SHU W，et al. Pollution Assessment and Potential Sources of Heavy Metals in Agricultural Soils around Four Pb/Zn Mines of Shaoguan City，China[J]. Soil & Sediment Contamination，2015，24（1）：76-89.
[23]	Zhen J，Pei T，Xie S. Kriging methods with auxiliary nighttime lights data to detect potentially toxic metals concentrations in soil [J]. Science of the Total Environment，2019，659：363-371.
[24]	李辉，翟星，李琛曦，等.河北省泥石流灾害易发性云模型评价方法：以邢台赵沟村泥石流为例[J].科学技术与工程， 2024，24（25）：10884-10891. LI H，ZHAI X，LI C X，et al. Development law of debris flow disaster and the evaluation method of its susceptibility： take mudslide in Zhaogou Village，Xingtai as an example[J]. Science Technology and Engineering，2024，24（25）：10884-10891.
[25]	GUO Q， LI B， CHEN Y， et al. Intelligent Model for the Compressibility Prediction of Cement-Stabilized Dredged Soil Based on Random Forest Regression Algorithm [J]. KSCE Journal of Civil Engineering，2021，25（10）：3727-3736.
[26]	张慧，栾思雨，丛蓉.东北黑土区典型县域耕地质量对耕地水田化的空间响应[J].水土保持研究，2024，31（01）： 327-334. ZHANG H，LUAN S Y，CONG R. Spatial response of cultivated land quality to paddy farming in typical counties of northeast black soil region[J]. Research of Soil and Water Conservation，2024， 31（1）：327-334.
[27]	刘宝涛，杨兴龙，郭佳.东北地区城镇化发展与土地健康利用空间关联特征[J].水土保持研究，2018，25（4）：386-392. LIU B T，YANG X L，GUO J. Spatial correlation characteristics between urbanization development and land health utilization[J]. Research of Soil and Water Conservation， 2018， 25（4）： 386-392.
[28]	CHEN D，WANG X，LUO X，et al. Delineating and identifying risk zones of soil heavy metal pollution in an industrialized region using machine learning [J]. Environmental Pollution， 2023， 318：120932.
[29]	LI J，ZHONG M，JIANG L，et al. Assessing and mitigating health risks of workers exposed to volatile organic compounds in contaminated soils during active pit excavation：Accounting for exposure variability and uncertainty [J]. Journal of Cleaner Production，2024，469：143227.
[30]	钱江波，陈涤，王夏晖，等.基于机器学习的区域土壤重金属污染风险诊断[J].环境工程，2023，41（12）：296-303. QIAN J B，CHEN D，WANG X H，et al. Risk diagnosis of heavy metal pollution in regional soil based on machine learning[J]. Environmental Engineering，2023，41（12）：296-303.
[31]	HUANG G，WANG X，CHEN D，et al. A hybrid data-driven framework for diagnosing contributing factors for soil heavy metal contaminations using machine learning and spatial clustering analysis [J]. Journal of Hazardous Materials， 2022， 437： 129324.
[32]	DING Y， YAO X， SU Z， et al. Research and engineering application of efficient pretreatment technology of initial rainwater in artificial wetland of the coastal industrial park[J]. Fresenius Environmental Bulletin，2022，31（6A）：6290-6304.
[33]	HOU S，ZHENG N，TANG L，et al. Effect of soil pH and organic matter content on heavy metals availability in maize（Zea mays L.） rhizospheric soil of non-ferrous metals smelting area [J]. Environmental Monitoring and Assessment， 2019， 191（10）：634.
[34]	龙於洋，邱钧健，宋禹翮，等.危险废物焚烧灰-渣共热处理对Cr和Pb的原位固定[J].环境科学学报，2023，43（6）： 390-397. LONG Y Y，QIU J J，SONG Y Y，et al. In-situ immobilization of Cr and Pb by co-thermal treatment of hazardous waste incineration ash-slag[J]. Acta Scientiae Circumstantiae，2023， 43（6）：390-397.