بهینه‏ سازی مدل دراستیک با استفاده از ماشین بردار پشتیبان و شبکۀ عصبی مصنوعی به‌منظور ارزیابی آسیب ‏پذیری ذاتی آبخوان دشت اردبیل

نوع مقاله : پژوهشی

نویسندگان

1 دانشجوی کارشناسی ارشد هیدروژئولوژی، دانشکدۀ علوم طبیعی، دانشگاه تبریز

2 استادیار گروه علوم زمین، دانشکدۀ علوم طبیعی، دانشگاه تبریز

3 استاد گروه علوم زمین، دانشکدۀ علوم طبیعی، دانشگاه تبریز

4 دانشجوی دکتری هیدروژئولوژی، دانشکدۀ علوم طبیعی، دانشگاه تبریز

چکیده

با توجه به افزایش جمعیت و توسعۀ کشاورزی در دشت اردبیل، ارزیابی آسیب‏پذیری آبخوان این دشت برای مدیریت منابع آب زیرزمینی و جلوگیری از آلودگی آب‏های زیرزمینی ضروری است. در این پژوهش آسیب‏پذیری آبخوان دشت اردبیل در برابر آلودگی با استفاده از روش دراستیک بررسی شد. در مدل دراستیک هفت پارامتر مؤثر در آسیب‏پذیری شامل عمق آب زیرزمینی، تغذیۀ خالص، محیط آبخوان، محیط خاک، توپوگرافی، محیط غیر‌اشباع و هدایت هیدرولیکی، به‌صورت هفت لایۀ رستری با مقیاس 30000: 1 تهیه شد و بعد از رتبه‏دهی و وزن‏دهی شاخص دراستیک محاسبه شد که برای دشت اردبیل بین 82 تا 151 به‌دست آمد. سپس به‌منظور بهینه‏سازی مدل دراستیک از مدل ماشین بردار پشتیبان، شبکۀ عصبی پیشرو و شبکۀ عصبی برگشتی استفاده شد تا بدین‌طریق بتوان به نتایج دقیق‏تری از ارزیابی آسیب‏پذیری دست یافت. به این منظور پارامترهای دراستیک به‌عنوان ورودی مدل و شاخص دراستیک به‌عنوان خروجی مدل تعریف شدند و مقادیر نیترات مربوطه به 2 دستۀ آموزش و آزمون تقسیم شد. شاخص دراستیک مربوط به مرحلۀ آموزش با مقادیر نیترات مربوط تصحیح شد و بعد از آموزش مدل، در مرحلۀ آزمون نتایج مدل‏ها با استفاده از مقادیر نیترات ارزیابی شد. نتایج نشان داد که هر سه مدل هوش مصنوعی توانایی زیادی در ارزیابی آسیب‏پذیری آبخوان دارند، اما در این بین، مدل ماشین بردار پشتیبان در مرحلۀ آزمون برای هر سه قسمت شرقی، غربی و جنوبی دشت با کمترین مقدار RMSE به‌ترتیب 74/6، 93/3 و 78/3 و بیشترین مقدار R2 به‌ترتیب 73/0، 79/0 و 72/0نتایج بهتری را در‌بر‌داشت. بر‌اساس این مدل، قسمت‏های شمالی و غربی دشت پتانسیل آلودگی بالایی دارد و باید محافظت بیشتری از این مناطق صورت گیرد.
 
 
 
 
 

کلیدواژه‌ها

موضوعات


عنوان مقاله [English]

Optimization of DRASTIC Model by Support Vector Machine and Artificial Neural Network for Evaluating of Intrinsic Vulnerability of Ardabil Plain Aquifer

نویسندگان [English]

  • Maryam Gharekhani 1
  • Ata Allah Nadiri 2
  • Asghar Asghari Moghaddam 3
  • Fariba Sadeghi Aghdam 4
1 MSc. Student, Faculty of Science, University of Tabriz
2 Faculty of Science, University of Tabriz, Tabriz
3 Faculty of Science, University of Tabriz, Tabriz.
4 PhD student, Faculty of Science, University of Tabriz, Tabriz.
چکیده [English]

With respect to population growth and agricultural development in Ardabil Plain, vulnerability assessment of the plain aquifer is necessary for management of groundwater resources and the prevention of groundwater contamination. In this study, vulnerability of Ardabil plain aquifer to pollution was evaluated by DRASTIC method. DRASTIC model was prepared by seven effective parameters on vulnerability, including groundwater depth, net recharge, aquifer media, soil media, topography, impact of vadose zone, and hydraulic conductivity as seven raster layers at 1:30000 scales. Then DRASTIC index was calculated after ranking and weighting that it was obtained 82 to 151 for Ardabil plain. The support vector machine (SVM), feedforward network (FFN) and recurrent neural network (RNN) models were adapted for optimizing the DRASTIC model to obtain the most accurate results of vulnerability evaluation. For this purpose, the DRASTIC parameters and the vulnerability index were defined as inputs data and output data respectively for models, and nitrate concentration data were divided in two categories for training and testing.DRASTIC index in training step was corrected by the related nitrate concentration, and after model training, the output of model in test step was verified by the nitrate concentration. The results show that 3 models of artificial intelligence are able to assessment of aquifer vulnerability, but the Support vector machine (SVM) with the least value of RMSE for all Eastern, Western and Southern parts of the plain is 6.74, 3.93 and 3.78, respectively and the highest value of R2 is 0.73, 0.79 and 0.72, respectively had the best results in the test step.According to this model, the northern and western parts of the plain are classified as high pollution potential areas and should be more protection of these areas.
 
 
 

کلیدواژه‌ها [English]

  • Aquifer Vulnerability
  • DRASTIC model
  • Artificial Intelligence
  • Support vector machine
  • Ardabil Plain
1[. فرج‏زادۀ اصل، منوچهر؛ محمدی، عثمان؛ 1390، «پهنه‏بندی آسیب‏پذیری آب‏های زیرزمینی با کمک الگوریتم‏های فازی عصبی (مطالعۀ موردی: استان زنجان)»، سنجش از دور و GIS ایران، سال سوم، شمارۀ اول، ص 1-18.
]2[. کرد، مهدی؛ 1393، «مدل‌سازی عددی آبخوان دشت اردبیل و مدیریت آن با استفاده از بهینه‏سازی برداشت آن»، رساله دکتری، دانشگاه تبریز.
]3[. ندیری، عطاالله؛ 1392، «مقایسه کارایی مدل‌های عددی و هوش مصنوعی در مدیریت آبخوان‌ها (مطالعۀ موردی: دشت تسوج)»، رسالۀ دکتری، دانشگاه تبریز.
[4]. Al-Abadi, Alaa M; Al-Shamma’a, Ayser M; Aljabbari, Mukdad H; 2014, A GIS-based DRASTIC model for assessing intrinsic groundwater vulnerability in northeastern Missan governorate, southern Iraq, Appl Water Sci.
[5]. Aller, Linda; Bennett, Truman; Lehr, Jay.H; Petty, Rebecca.; and Hackett, Glen;1987, DRASTIC: A Standardized System for Evaluating Ground Water Pollution Potential Using Hydrogeologic Settings, EPA 600/2-87-035. U.S. Environmental Protection Agency, Ada, Oklahoma.
[6]. Antonakos, Andreas.K; and Lambrakis, Nikolaos. I; 2007, Development and testing of three hybrid methods for the assessment of aquifer vulnerability to nitrates based on the drastic model, an example from NE Korinthia, Greece, Journal of Hydrology, pp. 288– 304.
[7]. ASCE Task Committee on Application of Artificial Neural Networks in Hydrology, 2000, Artificial neural network in hydrology, part I and II. J. Hydrol. Eng., 5(2), pp. 115-137.
[8]. Asefa, Tirusew; Kemblowski, Mariush; McKee,Mac; Khalil, Abedalrazq; 2006, Multi-time scale stream flow predictions: The support vector machines approach, Journal of Hydrology, 318, pp. 7–16.
[9]. Babiker,Insaf, S; Mohamed, Mohamed, A.A; Tetsuya, Hiyama;and Kikuo, Kato; 2005, A GIS-basedDRASTIC model for assessing aquifer vulnerability in KakamigaharaHeights, Gifu Prefecture, central Japan, Sci Total Environ, vol 354, pp.127–140.
[10]. Barzegar, Rahim; Asghari Moghaddam, Asghar; Baghban, Hamed; 2015, A supervised committee machine artificial intelligent for improving DRASTIC method to assess groundwater contamination risk: a case study from Tabriz plain aquifer, Iran, Stoch Environ Res Risk Assess.
[11]. Behzad, Mohsen; Asghari, Keyvan; and Coppola, Jr,Emery.A; 2010, Comparative Study of SVMs and ANNs in Aquifer Water Level Prediction. J. Comput. Civ. Eng, vol 24, pp. 408-413.
[12]. Chen, Bo.Juen; Chang, Ming.Wei; and Lin, Chih.Jen;2004, Load forecasting using Support Vector Machines: A study on EUNITE Competition 2001, IEEE Transactions on Power Systems, 19(4) ,pp. 1821–1830.
[13]. Chitsazan, Nima ; Nadiri, Ata Allah; Tsai, Frank, T-C, 2015, Prediction and structural uncertainty analyses of artificial neural networks using hierarchical Bayesian model averaging, Journal of Hydrology, 528, pp. 52-62.
[14]. Cortes, Corinna; and Vapnik, Vladimir; 1995, Support-vector networks. Machine learning, 20: 3, pp. 273-297.
[15]. Cristianini, Nello; and Shawe-Taylor, John; 2000, An Introduction to Support VectorMachines. Cambridge University Press, New York, USA.
[16]. Dibike, Yonas, B; Velickov, Slavco; Solomatine, Dimitri; and Abbot, Michael, B; 2001, ModelInduction with Support Vector Machines-Introduction and Applications. J. Comp. Civil Engin. ASCE, vol 15, pp. 208-216.
[17]. Dixon, Barnali; 2005a, Applicability of neuro-fuzzy techniques in predicting ground-water vulnerability: a GIS-based sensitivity analysis, Journal of Hydrology, vol 309, pp. 17-38.
[18]. Dixon, Barnali; 2009, A case study using support vector machines, neural networks and logistic regression in a GIS to identify wells contaminated with nitrate-N, Hydrogeology Journal, vol 17, pp. 1507–1520.
[19]. Fausett, Laurene; 1994, Fundamentals of neural network. Prentice Hall, Englewood Cliffs, N. J. 461 Pages.
[20]. Fijani, Elham; Nadiri, Ata. Allah; Asghari, Moghaddam, Asghar; Tsai, Frank, T-C; and Dixon, Barnali; 2013, Optimization of DRASTIC Method by Supervised Committee Machine Artificial Intelligence to Assess Groundwater Vulnerability for Maragheh-Bonab Plain Aquifer, Iran, Journal of hydrology, vol 530, pp. 89-100.
[21]. Haykin, Simon; 1994, Neural networks: a comprehensive foundation,Macmillan College Publishing, New York.
[22]. Hong, Wei, Chiang; 2011, Traffic flow forecasting by seasonal SVR with chaotic simulated annealing algorithm, Neurocomputing, Vol 74, pp. 2096-2107.
[23]. Jovanovic, N. Z; Adams S; Thomas A; Fey M; Beekman H. E; Campbell R; Saayman I; Conrad J; 2006, Improved DRASTIC method for assessment of groundwater vulnerability to generic aqueousphase contaminants, WIT Transactions on Ecology and the Environment, Vol 92, waste Management and the Environment III, pp. 393–402.
[24]. McCulloch, Warren, S; Pitts, Walter; 1943, A logic calculus of the ideas immanent in nervousactivity. Bulletin of Mathematical Biophysics, vol 5, pp. 115-133.
[25]. McLay, C.D.A., Dragten, R., Sparling, G., and Selvarajah, N., 2001, Predicting groundwater nitrate concentrations in a region of mixed agricultural land use: a comparison of three approaches, Environmental Pollutants, vol 115, pp. 191-204.
[26]. Mohammadi, Kourosh; Niknam, Ramin;and Majd, Vahid, Johari; 2009, Aquifer vulnerability assessment using GIS and fuzzy system: a case study in Tehran-Karaj aquifer, Iran, Environ Geol, vol 58, pp.437–446.
[27]. Mousavi, S.F. Amiri, M.J. Gohari, A.R. and Afyuni, M. 2011, Estimation of Nitrate Concentration Using Fuzzy Regression Method and Support Vector Machines, World Applied Sciences Journal, 12 (6), pp. 774-782.
[28]. Nadiri, Ata Allah; Fijani, Elham; Asghari Moghaddam, Asghar; 2013, Supervised committee machine with artificial intelligence for prediction of fluoride concentration, IWA Publishing.
[29]. Panagopoulos, George; Antonakos, Andreas; and Lambrakis, Nicolaos; 2005, Optimization of DRASTIC model for groundwater vulnerability assessment, by the use of simple statistical methods and GIS, Hydrogeology Journal.
[30]. Raghavendra, Sujay; and Chandra, Deka, Paresh; 2014, Support vector machine applications in the field of hydrology: A review, Elsevier, Applied Soft Computing,vol 19, pp. 372-386.
[31]. Secunda, S., Collin, M.L., and Melloul, A.J., 1998, Groundwater vulnerability assessment using a composite model combining DRASTIC with extensive agricultural land use in Israel’s Sharon region, Journal of Environmental Management, vol 54, pp. 39-57.
[32]. Seifi, Akram; 2010, Developing of expert system to prediction of dailyevapotranspiration by support vector machine and compare results to ANN,ANFIS and experimental method. M.Sc. Thesis, Department of WaterEngineering, Tarbiat Modarres University, Tehran, Iran (In Persian).
[33]. Shabri, Ani; and Suhartono; 2012, Streamflow forecasting using least-squares support vector machines, Hydrological Sciences journal, 57(7), pp. 1275-1293.
[34]. Suykens, Johan, A, K; Van, Gestel, Tony; Brabanter, Jos, De; De, Moor, Bart; Vandewalle, Joos; 2002, in:Least Squares Support Vector Machines, World Scientific Publishing, Singapore.
[35]. Tayfur, Gokmen; Nadiri, Ata Allah; Asghari Moghaddam, Asghar; 2014, Supervised intelligent committee machine method for hydraulic conductivity estimation, Water resources management, 28 (4), pp. 1173-1184.
[36]. Vrba, Jiri; and Zoporozec, Alexander; 1994, Guidebook on mapping groundwater vulnerability, International Contributions to Hydrogeology, Verlag Heinz Heise GmbH and Co, KG.
[37]. Yin, Jiajian;2011, LogP prediction for blocked tripeptides with amino acids descriptors (HMLP) by multiple linear regression and support vector regression, Procedia Environmental Sciences, vol 8, pp. 173–178.