基于机器学习的重金属毒性及生态风险预测

Abstract
Figure/Table
References
Related Citation (15)

Download: PDF (1508 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract This study focused on the toxicity of typical heavy metals in soil, including cadmium (Cd), copper (Cu), lead (Pb), and zinc (Zn), and summarized their effects on the model organisms, earthworms. A total of 113 datasets encompassing the median effective concentration (EC₅₀) of heavy metals on earthworm reproduction, along with corresponding soil physicochemical properties, were compiled from the published literature. The correlation between various datasets was analyzed to reveal the influence of soil physicochemical factors on the biotoxicity of heavy metals. Five machine learning algorithms, including Random Forest (RF), Gradient Boosting Decision Tree (GBDT), Extreme Gradient Boosting (XGBoost), K-Nearest Neighbor (KNN), and Support Vector Regression (SVR), were employed to develop predictive models for biotoxicity of heavy metals based on soil characteristics, ultimately selecting the best-performing model for predicting potential ecological risk thresholds of heavy metals in Chinese soils. The results indicate significant variation in heavy metal toxicity across different soils, with the toxicity trend for earthworm reproduction ranking as follows: indicate significant variation in heavy metal toxicity across different soil types, with the toxicity ranking for earthworm reproduction as Cd>Cu>Pb≈Zn. The effects of soil physicochemical properties on heavy metal toxicity varies depending on the specific heavy metal. Specifically, soil pH emerged as a key factor influencing the toxicity of Pb and Cd, contributing 57.2% and 69.0% respectively, while cation exchange capacity and organic matter content were found to be the primary influencing factors for the bio-toxicity of Cu and Zn. The performance of the machine prediction models for biological toxicity of heavy metals based on soil physicochemical factors was compared and analyzed in terms of model fit and prediction accuracy. Among the predictive models, the XGBoost model performed well for predicting the bio-toxicity of Cd, Cu, and Zn, while the RF model demonstrated higher accuracy in predicting Pb bio-toxicity, achieving R² values of 0.939 and 0.886 for training and testing sets, respectively. Furthermore, the potential ecological risk thresholds of heavy metals in soils across 34provinces in China were evaluated with the selected models, revealing significant regional differences in potential ecological risks. The findings provided a new strategy for accurate prediction and rational assessment of heavy metal ecological toxicity and potential ecological risk based on soil physicochemical properties.

Key words： machine learning soil heavy metal physicochemical properties biotoxicity ecological risk assessment

Received: 16 May 2024

PACS:

X53

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	LI Guo-feng
	YU Jin-qiu
	WANG Hong
	CHI Hai-feng
	LIN Shan-na
	CAI Chao

Cite this article:

LI Guo-feng,YU Jin-qiu,WANG Hong等. Prediction of heavy metal toxicity and ecological risk based on machine learning methods[J]. CHINA ENVIRONMENTAL SCIENCECE, 2024, 44(12): 7001-7010.

URL:

http://www.zghjkx.com.cn/EN/ OR http://www.zghjkx.com.cn/EN/Y2024/V44/I12/7001

[1] Jaishankar M, Tseten T, Anbalagan N, et al. Toxicity, mechanism and health effects of some heavy metals [J]. Interdisciplinary toxicology, 2014,7(2):60-72.
[2] Liu H, Wang H, Zhao H, et al. Speciation, bioaccumulation, and toxicity of the newly deposited atmospheric heavy metals in soil- earthworm (Eisenia fetida) system near a large copper smelter [J]. Science of The Total Environment, 2024,924:171700.
[3] 宋玉芳,周启星,许华夏,等.重金属对土壤中小麦种子发芽与根伸长抑制的生态毒性[J]. 应用生态学报, 2002,13(4):459-462. Song Y F, Zhou Q X, Xu H X, et al. Eco-toxicity of heavy metals on the inhibition of seed germination and root elongation of wheat in soils [J]. Chinese Journal of Applied Ecology, 2002,13(4):459-462.
[4] 王巍然,林祥龙,赵龙,等.我国20种典型土壤中锌对白符跳虫的毒性阈值及其预测模型[J]. 农业环境科学学报, 2021,40(4):766-773. Wang W R, Lin X L, Zhao L, et al. Toxicity threshold and prediction model for zinc in soil-dwelling springtails in Chinese soils [J]. Journal of Agro-Environment Science, 2021,40(4):766-773.
[5] 李宁,郭雪雁,陈世宝,等.基于大麦根伸长测定土壤Pb毒性阈值,淋洗因子及其预测模型[J]. 应用生态学报, 2015,26(7):2177-2182. Li N, Guo X Y, Chen S B, et al. Toxicity thresholds and predicted model of Pb added to soils with various properties and its leaching factors as determined by barley root-elongation test [J]. Chinese Journal of Applied Ecology, 2015,26(7):2177-2182.
[6] Li B, Zhang H T, Ma Y B, et al. Relationships between soil properties and toxicity of copper and nickel to bok choy and tomato in Chinese soils [J]. Environmental Toxicology & Chemistry, 2013,32(10):2372-2378.
[7] 高超,李霁,刘征涛,等.土壤铅镉铬暴露下赤子爱胜蚓的回避行为和急性毒性[J]. 环境科学研究, 2015,28(10):1596-1601. Gao C, Li Q, Liu Z T, et al. Avoidance Behavior and Acute Toxicity of Eisenia fetida under Exposure to Lead, Cadmium and Chromium in Soil [J]. Research of Environmental Sciences, 2015,28:1596-1601.
[8] Chen K Y, Chen H X, Zhou C L, et al. Comparative analysis of surface water quality prediction performance and identification of key water parameters using different machine learning models based on big data [J]. Water Research, 2020,171,115454.
[9] Lanno R P, Oorts K, Smolders E, et al. Effects of soil properties on the toxicity and bioaccumulation of lead in soil invertebrates [J]. Environmental Toxicology and Chemistry, 2019,38(7):1486-1494.
[10] Fan J T, Huang G X, Chi M H, et al. Prediction of chemical reproductive toxicity to aquatic species using a machine learning model: An application in an ecological risk assessment of the Yangtze River, China [J]. Science of the Total Environment, 2021,796,148901.
[11] 莫俊超,姚洪伟,曹峰.机器学习预测有机化学品对稀有鮈鲫急性毒性[J]. 中国环境科学, 2024,44(8):4661-4673. Mo J C, Yao H W, Cao F. Prediction of acute toxicity of organic chemicals on rare minnow using machine learning [J]. China Environmental Science, 2024,44(8):4661-4673.
[12] Ban M J, Lee D H, Shin S W, et al. Identifying the acute toxicity of contaminated sediments using machine learning models [J]. Environmental Pollution, 2022,312,120086.
[13] 聂红梅.基于多源数据和机器学习算法的土壤水分预测[D]. 西安:西北大学, 2020. Nie H M. Prediction of Soil moisture Based on Multi-source data and Machine learning algorithm [D]. Xian: Northwest University, 2020.
[14] 丁玮祺,吾木提·艾山江,阿不都艾尼·阿不里,等.基于机器学习的干旱区土壤盐渍化定量估算[J]. 土壤通报, 2022,53(5):1038-1048. Ding W Q, Umut Hasan, Abdugheni Abliz, et al. Quantitative Estimation Soil Salinization in Arid Areas Based on Machine Learning [J]. Chinese Journal of Soil Science, 2022,53(5):1038-1048.
[15] 周志华.机器学习[M]. 北京:清华大学出版社, 2016. Zhou Z H. Machine learning [M]. Beijing: Tsinghua University Press, 2016.
[16] 刘靖宇,李若怡,梁永春,等.基于特征优选和机器学习的塔里木盆地东缘绿洲土壤镉元素含量预测及健康风险评价[J]. 环境科学, 2024,45(8):4802-4811. Liu J Y, Li R Y, Liang Y C, et al. Soil Cadmium Prediction and Health Risk Assessment of an Oasis on the Eastern Edge of the Tarim Basin Based on Feature Optimization and Machine Learning [J]. Environmental Science, 2024,45(8):4802-4811.
[17] Zhang J W, Wang X N, Li J, et al. Bioavailability (BA)-based risk assessment of soil heavy metals in provinces of China through the predictive BA-models [J]. Journal of Hazardous Materials, 2024,465: 133327.
[18] Lu H, Li H M, Liu T, et al. Simulating heavy metal concentrations in an aquatic environment using artificial intelligence models and physicochemical indexes [J]. Science of The Total Environment, 2019,694:133591.
[19] Finizio A, Vighi M. Predicted No Effect Concentration (PNEC) [J]. Encyclopedia of Toxicology, 2014:1061-1065.
[20] Zhang L L, Shen L N, Qin S, et al. Quinolones antibiotics in the Baiyangdian Lake, China: Occurrence, distribution, predicted no- effect concentrations (PNECs) and ecological risks by three methods [J]. Environmental Pollution, 2020,256:113458.
[21] 孙振军,刘玉庆,李文立.温度,湿度和酸碱度对蚯蚓生长与繁殖的影响[J]. 莱阳农学院学报, 1993,10(4):297-300. Sun Z J, Liu Y Q, Li W L. Effects of temperature moisture and acidity on the growth and reproduction in Eisenia Fetida [J]. Journal of Laiyang Agricultural College, 1993,10(4):297-300.
[22] 刘海龙,王玉军,宣亮,等.土壤基本理化性质对外源镉蚯蚓慢性毒性的影响[J]. 农业环境科学学报, 2016,35(2):225-233. Liu H L, Wang Y J, Xuan L, et al. Effects of soil physico-chemical properties on chronic toxicity of cadmium to earthworm Eisenia fetida [J]. Journal of Agro-Environment Science, 2016,35(2):225-233.
[23] 贺萌萌.北京褐潮土中的镉对玉米幼苗和蚯蚓的毒性效应[D]. 杨凌:西北农林科技大学, 2014. He M M. Toxic effects of cadmium on maize seedlings and earthworm in fluvo-aquic soils of Beijing [D]. Yangling: Northwest Agriculture & Forestry University, 2014.
[24] Spurgeon D J, Stürzenbaum S R, Svendsen C, et al. Toxicological, cellular and gene expression responses in earthworms exposed to copper and cadmium [J]. Comparative Biochemistry and Physiology Part C: Toxicology & Pharmacology, 2004,138(1):11-21.
[25] Duan X W, Xu M, Zhou Y Y, et al. Effects of soil properties on copper toxicity to earthworm Eisenia fetida in 15Chinese soils [J]. Chemosphere, 2016,145:185-192.
[26] Amorim M J B, Scott-Fordsmand J J. Toxicity of copper nanoparticles and CuCl2salt to Enchytraeus albidus worms: Survival, reproduction and avoidance responses [J]. Environmental Pollution, 2012,164:164-168.
[27] Criel P, Lock K, Van Eeckhout H, et al. Influence of soil properties on copper toxicity for two soil invertebrates [J]. Environmental Toxicology and Chemistry, 2008,27(8):1748-1755.
[28] Ruyters S, Salaets P, Oorts K, et al. Copper toxicity in soils under established vineyards in Europe: A survey [J]. Science of The Total Environment, 2013,443:470-477.
[29] De Barros Amorim M J, Römbke J, Schallnaß H-J, et al. Effect of soil properties and aging on the toxicity of copper for enchytraeus albidus, enchytraeus luxuriosus, and folsomia candida [J]. Environmental Toxicology and Chemistry, 2005,24(8):1875-1885.
[30] 周娟,颜增光,蒋金炜,等.几种典型土壤中铜对赤子爱胜蚓的毒性差异比较研究[J]. 生态毒理学报, 2008,3(4):394-402. Zhou J, Yan Z G, Jiang J W, et al. Comparative study on toxicity differences of copper to earthworm Eisenia fetida in four typical soils [J]. Asian Journal of Ecotoxicology, 2008,3(4):394-402.
[31] Simões B, Mazur N, Correia M, et al. Ecotoxicity test as an aid in the determination of copper guideline values in soils [J]. Ciência Rural, 2020,50:6.
[32] Zhang L, Verweij R A, Van Gestel C A M. Effect of soil properties on Pb bioavailability and toxicity to the soil invertebrate Enchytraeus crypticus [J]. Chemosphere, 2019,217:9-17.
[33] Alexandrino R, Lima F, Martins G, et al. Lead acetate ecotoxicity in tropical soils [J]. Ecotoxicology, 2021,30:1-14.
[34] Savard K, Berthelot Y, Auroy A, et al. Effects of HMX-Lead Mixtures on Reproduction of the Earthworm Eisenia Andrei [J]. Archives of environmental contamination and toxicology, 2007,53:351-358.
[35] Zhang L, Gestel C. The toxicity of different lead salts to Enchytraeus crypticus in relation to bioavailability in soil: Bioavailability, toxicity of lead salts to enchytraeids [J]. Environmental Toxicology and Chemistry, 2017,36(8):2083-2091.
[36] Wijayawardena A, Mallavarapu M, Naidu R. Bioaccumulation and toxicity of lead, influenced by edaphic factors: using earthworms to study the effect of Pb on ecological health [J]. Journal of Soils and Sediments, 2017,17:1064-1072.
[37] Qin L, Wang L, Sun X, et al. Ecological toxicity (ECx) of Pb and its prediction models in Chinese soils with different physiochemical properties [J]. Science of The Total Environment, 2022,853:158769.
[38] Davies N A, Hodson M E, Black S. Is the OECD acute worm toxicity test environmentally relevant? The effect of mineral form on calculated lead toxicity [J]. Environmental Pollution, 2003,121(1):49-54.
[39] Scheffczyk A, Frankenbach S, Jänsch S, et al. Comparison of the effects of zinc nitrate-tetrahydrate and tributyltin-oxide on the reproduction and avoidance behavior of the earthworm Eisenia andrei in laboratory tests using nine soils [J]. Applied Soil Ecology, 2014,83:253-257.
[40] Gomes S I L, De Boer T E, Van Gestel C A M, et al. Molecular mechanisms of zinc toxicity in the potworm Enchytraeus crypticus, analysed by high-throughput gene expression profiling [J]. Science of The Total Environment, 2022,825:153975.
[41] Lock K, Janssen C. Modeling zinc toxicity for terrestrial invertebrates [J]. Environmental toxicology and chemistry/SETAC, 2001,20: 1901-1908.
[42] Camargo Carniel L S, Niemeyer J C, Iuñes D E Oliveira Filho L C, et al. The fungicide mancozeb affects soil invertebrates in two subtropical Brazilian soils [J]. Chemosphere, 2019,232:180-185.
[43] Alves M, Oliveira Filho L C, Nogueira P, et al. Influence of ZnO Nanoparticles and a Non-Nano ZnO on Survival and Reproduction of Earthworm and Springtail in Tropical Natural Soil [J]. Revista Brasileira de Ciência do Solo, 2019,43:e0180133.
[44] Novo M, Lahive E, Díez-Ortiz M, et al. Toxicogenomics in a soil sentinel exposure to Zn nanoparticles and ions reveals the comparative role of toxicokinetic and toxicodynamic mechanisms [J]. Environmental Science: Nano, 2020,7:1464-1480.
[45] Hobbelen P H F, Koolhaas J E, Van Gestel C A M. Risk assessment of heavy metal pollution for detritivores in floodplain soils in the Biesbosch, The Netherlands, taking bioavailability into account [J]. Environmental Pollution, 2004,129(3):409-419.
[46] Garg P, Satya S, Sharma S. Effect of heavy metal supplementation on local (Allolobophora parva) and exotic (Eisenia fetida) earthworm species: A comparative study [J]. Journal of environmental science and health Part A, Toxic/hazardous substances & environmental engineering, 2009,44:1025-1032.
[47] 周娟.我国典型土壤中铜和镍对赤子爱胜蚓(Eisenia fetida)的毒性效应研究[D]. 郑州:河南农业大学, 2010. Zhou J. Study on toxicity of copper and nickel to Eisenia fetida in typical Chinese soils [D]. Zhengzhou: Henan Agricultural University, 2010.
[48] Tang Y Q, Wang C, Holm P E, et al. Impacts of biochar materials on copper speciation, bioavailability, and toxicity in chromated copper arsenate polluted soil [J]. Journal of Hazardous Materials, 2023,459: 132067.
[49] Owojori O J, Reinecke A J, Rozanov A B. Influence of clay content on bioavailability of copper in the earthworm Eisenia fetida [J]. Ecotoxicology and Environmental Safety, 2010,73(3):407-414.
[50] Spurgeon D, Lofts S, Hankard P, et al. Effect of pH on metal speciation and resulting metal uptake and toxicity for earthworms [J]. Environmental toxicology and chemistry / SETAC, 2006,25:788-796.
[51] 杜彩艳,祖艳群,李元.pH和有机质对土壤中镉和锌生物有效性影响研究[J]. 云南农业大学学报, 2005,20(4):539-543. Du C Y, Zu Y Q, Li Y. Effect of pH and organic matter on the bioavailability Cd and Zn in soil [J]. Journal of Yunnan Agricultural University, 2005,20(4):539-543.
[52] Ghosh S, Ojha P K, Carnesecchi E, et al. Exploring QSAR modeling of toxicity of chemicals on earthworm [J]. Ecotoxicology and Environmental Safety, 2020,190:110067.
[53] 潘泳兴,陈盟,王櫹橦,等.基于RF和MLR的土壤重金属影响因素分析及生物有效性预测[J]. 农业环境科学学报, 2024,43(4):845-857. Pan Y X, Chen M, Wang X T, et al. Analysis of influencing factors and bioavailability prediction of soil heavy metals based on RF and MLR [J]. Journal of Agro-Environment Science, 2024,43(4):845-857.
[54] 毕卫冬,丁昌峰,周志高,等.基于机器学习方法的花生镉富集系数预测[J]. 农业环境科学学报, 2024,43(6):1230-1238. Bi W D, Ding C F, Zhou Z G, et al. Prediction of cadmium bioconcentration factor for peanuts based on machine-learning methods [J]. Journal of Agro-Environment Science, 2024,43(6):1230-1238.
[55] 颜增光,谷庆宝,周娟,等.构建土壤生态筛选基准的技术关键及方法学概述[J]. 生态毒理学报, 2008,3(5):417-427. Yan Z G, Gu B Q, Zhou J, et al. A synoptic review of the technical tips and methodologies for the development of ecological soil screening benchmarks [J]. Asian Journal of Ecotoxicology, 2008,3(5):417-427.
[56] Zhang J W, Liu Z T, Tian B, et al. Assessment of soil heavy metal pollution in provinces of China based on different soil types: From normalization to soil quality criteria and ecological risk assessment [J]. Journal of Hazardous Materials, 2023,441:129891.
[57] 张媛,王美娥,谢天,等.澳门特别行政区土壤自然消减能力与污染背景[J]. 环境科学学报, 2021,41(2):583-596. Zhang Y, Wang M E, Xie T, et al. Natural attenuation capacity and urban background contamination in soil in Macao [J]. Acta Scientiae Circumstantiae, 2021,41(2):583-596.
[58] Zhang H B, Luo Y M, Zhao Q G, et al. Residues of organochlorine pesticides in Hong Kong soils [J]. Chemosphere, 2006,63(4):633-641.
[59] 简士濠,蔡呈奇,许正一,等.台湾都市土壤之特性与土壤分类[C]//中国土壤学会第十一届全国会员代表大会暨第七届海峡两岸土壤肥料学术交流研讨会.北京:中国土壤学会, 2008:463-469. Jian S H, Cai C Q, Xu Z Y, et al. Characteristics and soil classification of urban soils in Taiwan [C]//The 11th National Assembly for Soil Science Society of China and also the 7th Cross-Straits Academic Exchange Seminar on Soil and Fertilizer. Beijing: Soil Science Society of China, 2008:463-469.
[60] 宋文恩,陈世宝.基于水稻根伸长的不同土壤中镉(Cd)毒性阈值(ECx)及预测模型[J]. 中国农业科学, 2014,47(17):3434-3443. Song W E, Chen S B. The toxicity thresholds (ECx) of cadmium (Cd) to rice cultivars as determined by root-elongation tests in soils and its predicted models [J]. Scientia Agricultura Sinica, 2014,47(17):3434-3443.
[61] Qin L Y, Sun X Y, Yu L, et al. Ecological risk threshold for Pb in Chinese soils [J]. Journal of Hazardous Materials, 2023,444(A): 130418.