ارزیابی عملکرد روش‌های شبکۀ عصبی مصنوعی و زمین‌آمار در شبیه‌سازی پارامترهای کیفی آب‌های زیرزمینی (مطالعۀ موردی: شهر کوهپایه، استان اصفهان)

نوع مقاله: مقاله پژوهشی

نویسندگان

1 استادیار گروه محیط زیست، دانشکدۀ منابع طبیعی، دانشگاه تهران

2 کارشناس ارشد محیط زیست، دانشکدۀ منابع طبیعی، دانشگاه تهران

3 کارشناس ارشد بیابان‌زدایی، دانشکدۀ منابع طبیعی، دانشگاه تهران

4 دانشجوی دکتری آبخیزداری، دانشکدۀ منابع طبیعی، دانشگاه تهران، ایران

چکیده

آب‌های زیرزمینی، مهم‌ترین منبع آب مصرفی در مناطق خشک و نیمه‌خشک در بخش‌های مختلف از قبیل کشاورزی، صنعت و شرب است. مدیریت این منابع آبی نسبت به آب‌های سطحی مشکل‌تر و پرهزینه‌تر است. به همین دلیل باید به دنبال روش‌هایی معقول و مقرون به صرفه برای مشخص‌کردن وضعیت این آب‌ها بود. در این مطالعه از روش‌های زمین‌آماری کریجینگ و کوکریجینگ و همچنین شبکۀ عصبی پروسپترون چند‌لایه به‌منظور برآورد پارامترهای کیفی -SO42، TDS، Ca و TH استفاده شد تا ضمن مقایسۀ این روش‌ها با هم بهترین روش نیز در این زمینه انتخاب شود. بدین منظور از داده‌های 50 حلقه چاه دشت کوهپایۀ استان اصفهان استفاده شد. به‌منظور ارزیابی عملکرد روش‌های مذکور در شبیه‌سازی پارامترهای مطالعه‌شده از خطای جذر میانگین مربعات (RMSE) و ضریب همبستگی استفاده شد. نتایج حاصل از مقایسۀ سه روش نشان داد که در مورد همۀ پارامترها، شبکۀ عصبی پروسپترون چند‌لایه با RMSE کمتر و ضریب همبستگی بالاتر دقت بهتری نسبت به روش‌های کریجینگ و کوکریجینگ دارد و بین دو روش زمین‌آماری کریجینگ و کوکریجینگ نیز، روش کوکریجینگ با RMSE کمتر و ضریب همبستگی بالاتر عملکرد بهتری نسبت به روش کریجینگ در برآورد همۀ پارامترهای مطالعه‌شده از خود نشان داد.

کلیدواژه‌ها


عنوان مقاله [English]

The evaluation of the functionality of artificial neural network and geostatistical methods in simulation of quality parameters of groundwater (Case study: Koohpayeh, Isfahan)

نویسندگان [English]

  • Amir Hossein Hamidian 1
  • Majid Atashgahi 2
  • Mohammadreza Hajihashemijazi 3
  • Abolhasan Fathabadi 4
1 Assistant Professor, Department of Environment, Faculty of Natural Resources, University of Tehran, I.R. Iran
2 MSc. Graduated, Department of Environment, Faculty of Natural Resources, University of Tehran, I.R. Iran
3 MSc. Graduated, Department of Rehabilitation of Arid and Mountainous Regions, Faculty of Natural Resources, University of Tehran, I.R. Iran
4 PhD. Candidate, Department of Rehabilitation of Arid and Mountainous Regions, Faculty of Natural Resources, University of Tehran, I.R. Iran
چکیده [English]

Groundwater resources are the most important sources of water consumption (agriculture, industry and drinking water) in arid and semi-arid regions. The management of these water resources is more expensive and difficult than surface waters; therefore more complicated and economic methods are needed for determination of their quality and quantity. In this study geostatistical methods of kringing and cokriging and multilayer perceptron (MLP) artificial neural network model were used to estimate the quality parameters of SO42-, TDS, Ca and TH. The methods were compared in order to understand which one is the best method for this estimation. Data obtained from 50 wells located in Koohpayeh plain in Isfahan province was used for this study. For estimation of the functionality of these methods in simulation of the parameters, RMSE and correlation coefficient were used. The results showed that for all of the studied parameters, MLP with a lower RMSE and higher correlation coefficient showed the highest precision followed by cokriging. Kriging showed to have the lowest precision in stimulation of the quality parameters.

کلیدواژه‌ها [English]

  • MLP Neural Network
  • Kriging
  • Cokriging
  • RMSE
  • Koohpayeh
- Abedini,M., and Poladi, J.,2003. comparison of artificial neural network by others methods in special estimation of daily rainfall. 6th International Conference on Civil Engineering. May, 5 to 7. Isfahan University of Technology.

- Dagostino, V.; E.A. Greene; B. Passarella and G. Vurro. 1998. Spatial and temporal study of nitrate concentration in groundwater by means of coregionalization. Environmental Geology. 36: 285-295.

- Dehghani, A.A. M., Asgari., Mosaedi, A., 2009. Comparison of Geostatistics, Artifitial Neural Networks and Adaptive Neuro-Fuzzy Inference System Approaches in Groundwater Level Interpolation (Case study: Ghazvin aquifer). J. Agric. Sci. Natur. Resour., Vol. 16(Special issue 1-b), 2009.

- Hajihashemijazi, M.R., Atashgahi,M. and Hamidian, A. H.,2011. Spatial distribution of ground water pollution maps as a tool for management of this water resource. 7th National Seminar on Watershed Management Sciences and Engineering. April 27–29, 2011. Noor, Iran

- Hajihashemijazi, M.R., Atashgahi,M. and Hamidian, A. H.,2011. Spatial estimation of groundwater quality factors using geostatistical methods (case study: Golpayegan plain). Iranian Journal of Natural Resources , Natural Environmental Journal. Vol.63,No.4,2011,PP.347-357.

- Hasani-Pak, A.A., 1998. Geostatistics. Tehran University Press. 360p.

- Haykin S., Neural Network, Pearson Education, 2004.

- Kholghi, M., Hoseini, S.M. 2009. Comparison of groundwater level estimation using ordinary kriging and Neural-Phazy methods. Environment Modeling and Assessment Journal. 6: 729_753.

- Marofi, S.; A. Toranjeyan and H. Zare Abyaneh. 2009. Evaluation of geostatistical methods for estimating electrical conductivity and pH of stream waters in Hamedan-Bahar plain. Journal of Water and Soil Conservation, 16: 169-187.

- Misaghi, F.., Mohammadi, K.. 2002. Estimation of groundwater levels using conventional interpolation techniques and comparison with geostatistics technique, twenty-first meeting on Earth Sciences, Geological Survey and Mineral Exploration of Country, p. 588 to 590.

- Noori, R., Ashrafi, Kh.,and Ajarpour,A., 2008.Comparison of ANN and PCA based multivariate linear regression applied to predict the daily average concentration of CO: a case study of Tehran. Journal of the Earth & Space Physics, Vol. 34‏, No. 1, 2008.

- Ramirez, M. C. V., H. F. C. Velho and N. J. Ferreira. 2005. Artificial neural network technique for rainfall forecasting applied to the São Paulo region. J. Hydrol. 301: 146-162.

- Rizzo, D.M. and J.M. Mouser. 2000. Evaluation of Geostatistics for Combined Hydrochemistry and Microbial Community Fingerprinting at a Waste Disposal Site: 1-11

- Rizzo, D.M., and Dogherty, D.E. 1994. Characterization of aquifer properties using Artificial Neural Networks: Neural Kriging. Water Resour. Res. 30:2. 483-497

- Salajegheh, A., Fathabadi, A., Mahdavi, M., 2009. Investigation on the efficiency of neuro-fuzzy method and statistical models in simulation of rainfall-runoff process. Journal of Range and Water shed Management, Iranian Journal of Natural Resources, Vol 62, No.1, 2009.pp65-79.

- Taghizade Mehrjerdi, R.; M. Zareian Jahromi; Sh. Mahmodi; A. Heidari and F. Sarmadian. 2009. Reviewing methods of spatial interpolation to investigate the underground water quality parameters, Rafsanjan plain. Journal of Iranian Watersheds Science and Engineering. 2: 63-70.

 - Yesilnacar,M.I., et al.2008. Neural network prediction of nitrate in groundwater of Harran Plain, Turkey. Environ Geol: 56:19–25.DOI 10.1007/s0025400711365.

- Zehtabian, Gh., Janfaza, E., Mohammad asgari, H., and Nematollahi, M.J., 2010 Modeling of ground water spatial distribution for some chemical properties (Case study in Garmsar watershed). Iranian journal of Range and Desert Reseach, Vol. 17 No. (1).