Development of Multiple Regression Models and Multiple Linear Percepteron Model For Prediction of Heavy Metal Concentration in Sediment

Öz

In recent years, metal pollution in the aquatic environment has attracted global attention due to its high availability, persistence and environmental toxicity. The behavior of metals in natural water is a function of substrate sediment composition, suspended sediment composition and water chemistry. The sediment is an inseparable and dynamic part of the river basin which includes various habitats and environments. In assessing the quality of sediments, the determination of anthropogenic inputs of heavy metals in aquatic ecosystems due to their long half-life is considered important and may reflect the history of pollution in aquatic ecosystems. Therefore, it is very important to determine and estimate the concentration of heavy metals that affect sediment quality. In this study, it has been realized that heavy metal concentration in mid-Black Sea coastal sea and river sediments can be estimated by using multivariate linear regression (MLR), multivariate polynomial regression (MPR) and multiple layer percepteron (MLP) models. Physico-chemical parameters of sediment samples taken from 5 different points between 2007-2008 for testing and training of models pH, water content (WC), cation exchange capacity (CEC), oxidation reduction potential (ORP), electrical conductivity (Ec), zeta potential (ζP), total carbon (TC), total inorganic carbon (TIC) and total organic carbon (TOC) and heavy metals (Cu, Cr, Cd, Pb, Ni, Fe, Al, Sr, Mn and Cr
In this study, the performance comparisons of MLR, MPR and MLP models were performed to estimate each heavy metal concentration. As a result, physico-chemical parameters were considered as independent variables in the prediction of concentration of heavy metals in sediments, and regression analyzes were performed and it was found that the best results were obtained with MPR model.

Anahtar Kelimeler

• Prediction
• Multivariate Regression
• Artifical Neural Network
• Heavy Metal
• Sediment

Sedimandaki Ağır Metal Konsantrasyonunun Çoklu Değişken Regresyon Modelleri ve Çok Katmanlı Algılayıcı Ağ Modeli ile Tahmini

Öz

Son yıllarda, su ortamındaki metal kirliliğinin fazla bulunması, kalıcılığı ve çevresel toksisitesi nedeniyle küresel olarak dikkat çekmektedir. Metallerin doğal sudaki davranışı, substrat sediment bileşiminin, askıda sediment bileşiminin ve su kimyasının bir işlevidir. Sediman, çeşitli habitatlar ve ortamlar içeren nehir havzasının ayrılmaz ve dinamik bir parçasıdır. Sedimanların kalitesinin değerlendirilmesinde, uzun yarılanma süreleri nedeniyle sucul ekosistemlerdeki ağır metallerin, antropojenik girdilerinin belirlenmesi önemli olarak kabul edilir ve sucul ekosistemlerdeki kirlilik geçmişini yansıtabilir. Bu nedenle, sediman kalitesini etkileyen ağır metallerin konsantrasyonunun belirlenmesi ve tahmin edilmesi oldukça önem taşımaktadır. Bu çalışma kapsamında Orta Karedeniz Kıyı şeridi deniz ve ırmak sedimanlarında ağır metal konsantrasyonunun çoklu değişkenli lineer regresyon (MLR), çoklu değişkenli polinomal regresyon (MPR) ve Çok Katmanlı Algılayıcı Ağ (MLP) modelleri kullanılarak tahminlenebilmesi gerçekleştirilmiştir. Modellerin test ve eğitimleri için 2007-2008 yılları arasında 5 farklı noktadan alınan sediman örneklerine ait fiziko-kimyasal parametreler pH, su içeriği (WC), katyon değişim kapasitesi (CEC), oksidasyon redüksiyon potansiyeli (ORP), elektriksel iletkenlik (EC), zeta potansiyeli (ζP), toplam karbon (TC), toplam inorganik karbon (TIK), toplam organik karbon (TOK) ve ağır metallerin (Cu, Cr, Cd, Pb, Ni, Fe, Al, Sr, Mn ve Cr) konsantrasyonları kullanılmıştır. Çalışmada her bir ağır metal konstantrasyonu tahmin edilmesi için MLR, MPR ve MLP modelllerinin performans karşılaştırlmaları yapılmıştır. Sonuç olarak sedimanlardaki ağır metallerin konsantrasyonun tahminlemesinde fiziko-kimyasal parametreler bağımsız değişkenler olarak kabul edilerek, regresyon analizleri yapılmış ve gerçekleştirilen modeler arasında en iyi sonucun MPR modeli ile elde edildiği ortaya konulmuştur. 

Anahtar Kelimeler

• Tahminleme
• Çoklu Regresyon
• Yapay Sinir Ağları
• Ağır Metal
• Sediman

Kaynakça

1. Arıman S, Bakan G. Assessment Of Heavy Metal Levels In Sediments Of The Mid-Black Sea Coast Of Turkey”. 15th International Conference on Heavy Metals in the Environment, 2010; 19-23 September, pg. 472-474, Gdansk, Poland.
2. Carvalho S, Pereira P, Pereira F, De Pablo H, Vale C, Gaspar MB. Factors structuring temporal and spatial dynamics of macrobenthic communities in a eutrophic coastal lagoon (Óbidos lagoon, Portugal). Mar Environ Res 2011;71:97-110.
3. Clarke KR. Non-parametric multivariate analyses of changes in community structure. Aust J Ecol 1993;18:117–43.
4. Subida, M.D., Berihuete A, Drake P., Blasco, J. Multivariate methods and artificial neural networks in the assessment of the response of infaunal assemblages to sediment metal contamination and organic enrichment. Science of the Total Environment 2013; 450–451; 289–300.
5. Nunes, M., Coelho J, Cardoso P, Pereira M, Duarte A, PardalM. Themacrobenthic community along a mercury contamination in a temperate estuarine system (Ria de Aveiro, Portugal). Sci Total Environ 2008;405:186–94.
6. Pereira P, Carvalho S, Pereira F, de Pablo H, Gaspar MB, Pacheco M, et al. Environmental quality assessment combining sediment metal levels, biomarkers and macrobenthic communities: application to the Óbidos coastal lagoon (Portugal). Environ Monit Assess 2011;184:1-11.
7. Carvalho S, Pereira P, Pereira F, de Pablo H, Vale C, Gaspar MB. Factors structuring temporal and spatial dynamics of macrobenthic communities in a eutrophic coastal lagoon (Óbidos lagoon, Portugal). Mar Environ Res 2011;71:97-110.
8. Nunes M, Coelho J, Cardoso P, Pereira M, Duarte A, PardalM. Themacrobenthic community along a mercury contamination in a temperate estuarine system (Ria de Aveiro, Portugal). Sci Total Environ 2008;405:186–94.

9. Park YS, Chon TS, Kwak IS, Lek S. Hierarchical community classification and assessment of aquatic ecosystems using artificial neural networks. Sci Total Environ 2004;327: 105–22.
10. Alvarez-Guerra M, González-Piñuela C, Andrés A, Galán B, Viguri JR. Assessment of Self-Organizing Map artificial neural networks for the classification of sediment quality. Environ Int 2008;34:782–90
11. Cottenie, A., M. Verloo, L. Kieskens, G. Velgehe and R. Camerlynck., 1982. Chemical analysis of Plants and Soils. IWONL, Brussels, Belgium.
12. Diaz RJ, Rosenberg R. Marine benthic hypoxia: a review of its ecological effects and the behavioural responses of benthic macrofauna. Oceanogr Mar Biol Annu Rev 1995;33: 245–303.
13. Li, X, Wai, O.W.H., Li, Y.S., Coles,B.J., Ramsey, H., Thornton, I., 2000. Heavy Metal Distribution in Sediment Profiles of The Pearl River Estuary, South China. Applied Geochemistry, 15, 567-581.
14. Marengo, E., Gennaro, M.C., Robotti, E., Rossanigo, P., Rinaudo, C., G., M.R., 2006. Investigation of Anthropic Effects Connected with Metal Ions Concentration, Organic Matter and Grain Size in Bormida River Sediments. Analytica Chimica Acta, 560(1-2), 172-183.
15. Kumar N, Singh SK, Srivastava PK, Narsimlu B (2017) SWAT model calibration and uncertainty analysis for streamflow prediction of the tons river basin, India, using sequential uncertainty fitting (SUFI-2) algorithm. Model Earth Syst Environ 3:1–13.
16. Singh H, Singh D, Singh SK, Shukla DN (2017) Assessment of river water quality and ecological diversity through multivariate statistical techniques, and earth observation dataset of rivers Ghaghara and Gandak, India. Int J River Basin Manag, pp 1–14. doi:10.1080/15715124.2017.1300159.
17. Singh SK, Srivastava K, Gupta M, Thakur K, Mukherjee S (2014) Appraisal of land use/land cover of mangrove forest ecosystem using support vector machine. Environ Earth Sci 71:2245–2255.
18. Gautam SK, Sharma D, Tripathi JK, Singh SK, Ahirwar S (2013) A study of the effectiveness of sewage treatment plants in Delhi region. Appl Water Sci 3:57–65.
19. Dixon, B. (2005). Applicability of neuro-fuzzy technique in predicting ground-water vulnerability: A GIS-based sensitivity analysis. Journal of Hydrology, 309, 17–38.
20. Theofanis ZU, Astrid S, Lidia G, Calmano WG (2001) Contaminants in sediments: remobilisation and demobilization. Sci Total Environ 266:195–202.
21. Salas, J. D., Markus, M., & Tokar, A. S. (2000). Streamflow forecasting based on artificial neural networks. In G. Rao & A. R. Rao (Eds.), Artificial neural networks in hydrology. London: Kluwer Academic Publishers.
22. ASCE Task Committee (2000). Artificial neural networks in hydrology, 2. Hydrologic applications. ASCE Journal of Hydrologic Engineering, 5(2), 124–137.
23. Rogers, L. L., Dowla, F. U., & Johnson, V. M. (1995). Optimal field-scale groundwater remediation using neural networks and the genetic algorithm. Environmental Science and Technology, 29(5), 1145–1155.
24. Flood, I., & Kartam, N. (1994). Neural networks in civil engineering I. Principles and understanding. Journal of Computing in Civil Engineering, 8(2), 131–148.
25. Tessier A & Campbell PGC (1988) Comments on the testing of the accuracy of an extraction procedure for determining the partitioning of trace metals in sediments. Anal. Chem. 60: 1475–1476.
26. USEPA (U.S. Environmental Protection Agency), 1999. “SW-846. refernce methodology: Method 3050B. Standard Operating Procedure for the Digestion of Soil/Sediment Samples Using a Hot Plate/Beaker Digestion Technique, Chicago, IL.
27. Golterman, H.L., Sly, P.G., Thomas R.L., 1983. Study of The Realtionship Between Water Quality and Sediment Transport . UNESCO, Technical Papers in Hydrology 26, France.
28. Khwaja, A.R., Singh, R., Tandon, S.N., 2000. Monitoring of Ganga Water and Sediments Vis-A-Vis Tannery Pollution at Kanpur (India): A Case Study. Environ. Monit. Assess., 68, 19–35.
29. Gündoğdu, M. E., 2006. Meteorolojik Parametrelerin Hava Kirliliğine Etkilerinin Yapay Sinir Ağları Modeli İle İncelenmesi, FBE Çevre Mühendisliği, Yüksek Lisans Tezi, İstanbul
30. Kahane, L. (2014). Multiple Regression Analysis. In Regression Basics (pp. 59–78). https://doi.org/10.4135/9781483385662.n4
31. Sinha, P. (2013). Multivariate Polynomial Regression in Data Mining: Methodology, Problems and Solutions. International Journal of Scientific & Engineering Research, 4(12), 962–965