기계학습 분류모델을 이용한 하천퇴적물의 중금속 오염원 식별
Identifying sources of heavy metal contamination in stream sediments using machine learning classifiers
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
하천퇴적물은 유역내 다양한 오염원으로부터 발생하는 중금속, 유기물 등 오염물질의 수용체일 뿐만 아니라 수질 오염 및 수생태 악영향을 유발할 수 있는 2차적 오염원이기에 중요한 관리대상이라고 할 수 있다. 오염된 하천퇴적물의효과적인 관리를 위해서는 오염원에 대한 식별과 이와 연계된 관리대책의 수립이 우선되어야 한다. 본 연구는 하천퇴적물내 측정된 다양한 이화학적 오염항목 분포 특성에 기반하여 퇴적물의 주요 오염원을 식별하기 위한 방법으로서기계학습모델의 적용성을 평가하였다. 기계학습 모델의 성능 평가를 위해 전국 4대강 수계내 주요 폐금속광산 및 산업단지 인근에서 수집된 총 356개의 하천퇴적물에 대한 중금속 10개 항목(Cd, Cu, Pb, Ni, As, Zn, Cr, Hg, Li, Al) 과 토양항목 3개(모래, 실트, 점토 비율) 수질항목 5개(함수율, 강열감량, 총유기탄소, 총질소, 총인)를 포함한 총 18개 오염항목에 대한 분석자료를 활용하였다. 기계학습 분류 모델로서 선형판별분석(linear discriminant analysis, LDA)과서포트벡터머신(support vector machine, SVM) 분류기를 사용하여 폐금속광산(‘광산’)과 산업단지(‘산단’) 인근에서의 하천퇴적물 시료의 분류 성능을 평가한 결과, 채취 지점 및 시기별 4가지 경우(비강우시 광산, 강우시 광산, 비강우시 산단, 및 강우시 산단)에 대한 퇴적물 시료의 분류 성능이 우수하였으며, 특히 비선형 모델인 SVM(88.1%)이 선형모델인 LDA(79.5%) 보다 퇴적물을 분류하는데 있어 보다 우수한 성능을 나타냈다. SVM 앙상블 기반 비배타적다중라벨분류기 모델을 이용하여 각 시료채취 지점 상류 유역 1km 반경 내 지배적인 토지이용 및 오염원을 다중 타겟값으로 다중분류 예측을 수행한 결과, 폐금속광산과 산업단지의 분류는 비교적 높은 정확도로 수행하였으나, 도시와 농업지역 등 다른 비점오염원에 대한 분류정확도는 56~60%범위로 비교적 낮게 나타났다. 이는 다중라벨 분류모델의 복잡성에 비해 데이터셋의 크기가 상대적으로 작아서 발생한 과적합에 기인한 것으로 향후 보다 많은 측정자료가 확보될 경우 기계학습 모델을 적용한 오염원 분류의 정확도를 보다 향상시킬 수 있을 것으로 판단된다.
Abstract
Stream sediments are an important component of water quality management because they are receptors of various pollutants such as heavy metals and organic matters emitted from upland sources and can be secondary pollution sources, adversely affecting water environment. To effectively manage the stream sediments, identification of primary sources of sediment contamination and source-associated control strategies will be required. We evaluated the performance of machine learning models in identifying primary sources of sediment contamination based on the physico-chemical properties of stream sediments. A total of 356 stream sediment data sets of 18 quality parameters including 10 heavy metal species(Cd, Cu, Pb, Ni, As, Zn, Cr, Hg, Li, and Al), 3 soil parameters(clay, silt, and sand fractions), and 5 water quality parameters(water content, loss on ignition, total organic carbon, total nitrogen, and total phosphorous) were collected near abandoned metal mines and industrial complexes across the four major river basins in Korea. Two machine learning algorithms, linear discriminant analysis (LDA) and support vector machine (SVM) classifiers were used to classify the sediments into four cases of different combinations of the sampling period and locations (i.e., mine in dry season, mine in wet season, industrial complex in dry season, and industrial complex in wet season). Both models showed good performance in the classification, with SVM outperformed LDA; the accuracy values of LDA and SVM were 79.5% and 88.1%, respectively. An SVM ensemble model was used for multi-label classification of the multiple contamination sources inlcuding landuses in the upland areas within 1 km radius from the sampling sites. The results showed that the multi-label classifier was comparable performance with sinlgle-label SVM in classifying mines and industrial complexes, but was less accurate in classifying dominant land uses (50~60%). The poor performance of the multi-label SVM is likely due to the overfitting caused by small data sets compared to the complexity of the model. A larger data set might increase the performance of the machine learning models in identifying contamination sources.
참고문헌 (26)
- Ali, H., "Environmental chemistry and ecotoxicology of hazardous heavy metals:environmental persistence, toxicity, and bioaccumulation", J. Chem (2019)
- Belousov, A. I., "A flexible classification approach with optimal generalisation performance : support vector machines", Chemometrics and intelligent laboratory systems, 64(1), pp.15~25 (2002)
- 차윤경, "자료기반 물환경 모델의 현황 및 발전 방향", 한국물환경학회지, 36(6), pp.611~620 (2020), 한국물환경학회[DOI]
- Chueh, Y. Y., "Copper concentration simulation in a river by SWAT-WASP integration and its application to assessing the impacts of climate change and various remediation strategies", Journal of environmental management, 279, pp.111613~ (2021)
- 정세웅, "국가 물환경관리정책 지원을 위한 수질모델링 기술의 발전방향", 한국물환경학회지, 36(6), pp.621~635 (2020), 한국물환경학회[DOI]
- dos Santos Simoes, F., "Water quality index as a simple indicator of aquaculture effects on aquatic bodies", Ecological indicators, 8(5), pp.476~484 (2008)
- Guan, Q., "Source apportionment of heavy metals in agricultural soil based on PMF: A case study in Hexi Corridor, northwest China", Chemosphere, 193, pp.189~197 (2018)
- Jiang, Y., "The Characteristics and Source Analysis of Heavy Metals in the Sediment of Water Area of Urban Scenic : A Case Study of the Delta Park in Suzhou City, Anhui Province, China", Polish Journal of Environmental Studies, 30(3) (2021)
- 정명채, "국내 휴폐금속광산의 환경오염 평가 및 향후 관리 방안", 한국자원공학회지, 43(5), pp.383~394 (2006), 한국자원공학회
- Joseph, V. R., "Optimal ratio for data splitting", Statistical Analysis and Data Mining : The ASA Data Science Journal, 15(4), pp.531~538 (2022)
- Kabilan, N., "Surveillance and steering of irrigation system in cloud using Wireless Sensor Network and Wi-Fi module", 2016International Conference on Recent Trends in Information Technology (ICRTIT), pp.1~5 (2016), IEEE
- Kang, J. H., "Factor affecting metal exchange between sediment and water in an estuarine reservoir : a spatial and seasonal observation", J. Environ. Monit, 11, pp.2058~2067 (2009)
- Liber, Y., "Past and recent state of sediment contamination by persistent organic pollutants(POPs)in the Rhône River : Overview of ecotoxicological implications", Science of the Total Environment, 646, pp.1037~1046 (2019)
- Liu, M., "Support vector machine―an alternative to artificial neuron network for water quality forecasting in an agricultural nonpoint source polluted river?", Environmental Science and Pollution Research, 21, pp.11036~11053 (2014)
- Navarro, M. C., "Abandoned mine sites as a source of contamination by heavy metals:a case study in a semi-arid zone", Journal of Geochemical exploration, 96(2-3), pp.183~193 (2008)
- Rodríguez-Seijo, A., "Origin and spatial distribution of metals in urban soils", Journal of soils and sediments, 17, pp.1514~1526 (2017)
- Shen, Z., "Impact of spatial rainfall variability on hydrology and nonpoint source pollution modeling", Journal of Hydrology, 472, pp.205~215 (2012)
- Sojka, M., "Heavy metals in river sediments: contamination, toxicity, and source identification—a case study from Poland", International Journal of Environmental Research and Public Health, 19(17), pp.10502~ (2022)
- Srinivasan, R., "SWAT ungauged : hydrological budget and crop yield predictions in the Upper Mississippi River Basin", Transactions of the ASABE, 53(5), pp.1533~1546 (2010)
- Stoichev, T., "Multiple regression analysis to assess the contamination with metals and metalloids in surface sediments(Aveiro Lagoon", Portugal). Marine Pollution Bulletin, 159, pp.111470~ (2020)
- Syarif, I., "SVM parameter optimization using grid search and genetic algorithm to improve classification performance", TELKOMNIKA(Telecommunication Computing Electronics and Control), 14(4), pp.1502~1509 (2016)
- Wu, H., "Health risk assessment based on source identification of heavy metals : A case study of Beiyun River", China. Ecotoxicology and Environmental Safety, 213, pp.112046~ (2021)
- Yu, H., "A direct LDA algorithm for high-dimensional data—with application to face recognition", Pattern recognition, 34(10), pp.2067~2070 (2001)
- Zhang, X., "Effects of benthic algae on release of soluble reactive phosphorus from sediments : a radioisotope tracing study", Water Sci. Eng, 8(2), pp.127~131 (2015)
- Zhang, Z., "Use of interpretable machine learning to identify the factors influencing the nonlinear linkage between land use and river water quality in the Chesapeake Bay watershed", Ecological Indicators, 140, pp.108977~ (2022)
- Zhu, M., "A review of the application of machine learning in water quality evaluation", Eco-Environment & Health, 1, pp.107~116 (2022)