ارزیابی چند مدل هوش مصنوعی برای پیش‌بینی کیفیت آب زیرزمینی (مطالعۀ موردی: دشت گرو)

نویسندگان

1 بیابانزدایی، دانشکده کشاورزی و منابع طبیعی، دانشگاه هرمزگان، بندرعباس

2 گروه برنامه ریزی مدیریت و آموزش محیط زیست، دانشکده محیط زیست دانشگاه تهران

3 مهندسی ابخیزداری، دانشکده منابع طبیعی و علوم دریایی دانشگاه تربیت مدرس تهراان

10.22052/6.17.27

چکیده

مدل­سازی مناسب کیفیت آب­زیرزمینی از ابزارهای مهم برنامه­ریزی و تصمیم­گیری در مدیریت منابع آب است. در این مطالعه به منظور مدلسازی تغییرات متغیرهای کیفی آب­زیرزمینی دشت گرو از داده­های 14 چاه در  دوره آماری (1388 تا 1395) استفاده شد. متغیرهای Na، Mg، Ca،SO4، Cl و  HCO3به عنوان متغیر مستقل و EC، SAR، TDS و TH به عنوان متغیر وابسته در نظر گرفته شد. از روش­های ماشین بردار پشتیبان، شبکه عصبی مصنوعی و شبکه عصبی-فازی تطبیقی برای مدل­سازی متغیرهای کیفی آب­زیرزمینی استفاده شد. به منظور تخمین کیفیت آب­زیرزمینی کل داده­ها به صورت تصادفی به دو دسته آموزشی (80 درصد کل داده­ها) و آزمایشی (20 درصد کل داده­ها) تقسیم شد. نتایج حاصل از مدل­سازی متغیرهای کیفی آب­زیرزمینی در دشت گرو نشان داد که شبکه عصبی-فازی تطبیقی در متغیرهای EC (99/0=R2، 13/109= RMSE و 99/0 =CE)، SAR (98/0=R2، 28/0= RMSE و 98/0 =CE) و TH (99/0=R2، 49/0= RMSE و 99/0 =CE) نسبت به دو روش شبکه عصبی مصنوعی و ماشین بردار پشتیبان عملکرد بهتری دارد و در متغیر TDS مدل شبکه عصبی مصنوعی (99/0=R2، 13/109= RMSE و 99/0 =CE) نسبت به دو مدل دیگر کارایی بهتری داشته است. به منظور پهنه­بندی تغییرات کیفیت آب­زیرزمینی از مدل­های انتخاب شده بر اساس دو طبقه­بندی کیفیت آب شرب شولر و کشاورزی ویلکوکس استفاده گردید. نتایج حاصل از پهنه­بندی براساس طبقه­بندی آب شولر نشان داد که متغیر TDS داری سه طبقه نامناسب (1/21%)، بد (59/74%) و غیرقابل شرب (31/4%) و متغیرTH  دارای 4 طبقه خوب (85/0%)، قابل قبول (48/23%)، نامناسب (55/67%) و بد (12/8%) می­باشد. نتایج پهنه­بندی بر اساس طبقه­بندی ویلکوکس نیز نشان داد که متغیر EC داری سه طبقه عالی (41/9%)، خوب (79/89%) و متوسط (8/0%) و متغیر SAR دارای دو طبقه عالی (19%) و خوب (81%) می­باشد.

کلیدواژه‌ها


عنوان مقاله [English]

Assessment of some Artificial Intelligence (AI)-based models for groundwater quality prediction (case study: Garoo plain)

نویسندگان [English]

  • Mojtaba Dolati Kordestani 1
  • Ahmad Nohegar 2
  • saeid Janizadeh 3
1
2
3
چکیده [English]

Introduction
Today, a significant portion of the water consumption in Iran, especially in the drinking sector, is provided by water resources. Exploitation of groundwater resources requires knowledge of the quantitative and qualitative status of aquifers. By determining the chemical quality of groundwater, an estimate of the health status of these water resources can be obtained and, depending on its state, the type of use is determined. In this regard, direct and indirect methods can be used to understand the qualitative characteristics of water. Direct methods, despite their high precision, require a high size of observational data, involves substantial time and cost. Hence, numerous indirect methods have been developed for simulating natural systems and estimating their parameters using a computer based on complex calculations. The main advantage of these methods is the ability to learn time series and prediction. One of these methods is modeling or hydrological simulation. The modeling of groundwater quality is an important tool for planning and decision-making in the management of water resources. The goal of this research is to identify the ability of intelligent model of Support Vector Machines (SVM), Artificial Neural Network (ANN), and Adaptive Neuro-Fuzzy Inference System (ANFIS) for modeling groundwater quality variables (EC, SAR, TDS, and TH) in Gero plain and zoning these variables. Therefore, it can provide an appropriate management tool for controlling quality parameters for drinking and farming.
Material and methods
In this study, data from 14 wells over the 2008-2016 period was used in order to model the variations in quality variables of Gero plain groundwater. The observed values for Na, Mg, Ca, SO4, Cl, and HCO3 are considered as independent variables and values of EC, SAR, TDS, and TH are considered as dependent variables. An SVM, an ANN, and an ANFIS design were used to model groundwater quality. Input data are randomly divided into two sets such that 80% of data are assigned to the training set and the remaining data (20%) form the test set.
Results
Results showed that the ANFIS system had the best performance in the estimation of EC (R2 = 0.99, RMSE=109.13, CE=0.99), SAR (R2 = 0.98, RMSE=0.28, CE=0.98), and TH (R2 = 0.99, RMSE=0.49, CE=0.99) among considered methods for the modeling of groundwater quality. Results also indicated that the ANN had the best performance in estimating TDS (R2 = 0.99, RMSE=109.13, CE=0.99). Furthermore, Schoeller and Wilcox water quality classifications, for drinking and agricultural water, were respectively employed to perform groundwater quality zoning based on outcomes of the considered methods. According to Schoeller classification, TDS has three classes: inappropriate (21.1%), bad (74.59%), and non-dirking (4.31%) and TH variable has four class: good (0.84%), acceptable (23.48%), inappropriate (67.55%), and bad (8.16%). According to Wilcox classification, EC has three classes: excellent (9.41%), good (89.79%), middle (0.8%) and SAR has two classes: excellent (19%) and good (81%).
Discussion and Conclusion
ANFIS for a better estimation of EC, SAR, and TH variables outperforms two models of ANN and SVM. The ANFIS system, using the if-then rules, describes that these rules are implemented in a network structure that can be used for learning algorithms used in ANN. Due to this structure, the fuzzy-comparative neural network model has more transparency for analysis and interpretation. The zoning of qualitative variables (TDS and TH) based on the classification of Schoeller drinking water showed that in the TDS variable, the groundwater quality has three classes: bad, inappropriate, and non-drinkable, with the most inadequate plain, southeastern plain bad status and the west of the plain has a terrible situation. The TH zoning map presents that the plain is in good, acceptable, inappropriate, and bad classes. The most part of the plain is in the inappropriate class, the west of the plain in the bad class, and the southeast plain is in an acceptable class. The results of zoning the variables EC and SAR based on the Wilcox agricultural classification showed that groundwater quality is acceptable four agricultural purposes. Therefore, it is essential to take measures to improve the quality of drinking water in the region.

کلیدواژه‌ها [English]

  • Garoo plain
  • Support Vector Machine
  • Artificial Neural Network
  • adaptive neuro-fuzzy
  • zoning and groundwater quality
Abghari, H., 2008. Intelligent prediction methods based on wavelet and neural network models between monthly river discharges. Ph.D. Dissertation, Faculty of Natural Resources, University of Tehran.173pp. (In Persian). 2. Adib, A., Zamani, R., 2015. Spatial variability analysis of groundwater quality indicators Dezful plain using geostatistical method. Journal of Water Resources Engineering 8:1-12. (In Persian). 3. Afifi, E., Yamani, M., Hasanzadeh, Y., 2012. Hydrogeomorphology basin Garu plain (Hormozgan Province). Journal of Geographical land 9(35): 61-76. (In Persian). 4. Askari Marnani, S., Chitsazan, M., Mirzayi, Y., 2001. Investigation of Water Quality in Firoozabad Sub-Chachment in View of Domestic and Agricultural Usage using GIS. P 1-8, the 8 th International Congress on River Engineering, Shahid Chamran University, Iran. (In Persian). 5. Chowdhury, M., Alouani, A., Hossain, F., 2010. Comparison of ordinary kriging and artificial neural network for spatial mapping of arsenic contamination of groundwater. Stochastic Environ. Res. and Risk Assess 24(1): 1-7. 6. El-Shafie, A., Jaafer, O., Seyed, A., 2011. Adaptive neuro-fuzzy inference system based model for rainfall forecasting in Klang River, Malaysia, Int. J. Phys. Sci. 6(12): 2875-2888. 7. Hosseini, S.M., Borhani, R., 2009. The application of artificial neural network in estimating the river yield by minimum temperature and discharge (case study: Hamoon basin). The First International Conference of Water Crisis, 10-12 March. Zabol University. (In Persian). 8. Huiqun, M., Ling, L., 2008. Water quality assessment using artificial neural network .In International Conference on Computer Science and Software Engineering 5-13 December, USA. 9. Imrie, C.E., Durucan, S., Korre, A., 2000. River flow prediction using artificial neural networks :generalisation beyond the calibration range. Journal of Hydrology 233: 138–153. 10. Isazadeh, M., Arabzadeh, R., Darbandi, S., 2016. Performance Evaluation of Geostatistical Methods and Artificial Neural Network in Estimation of Aquifer Quality Parameters(Case Study: Qorveh Dehghan Plain). J. Water and Soil Sci. 20(77): 197-210. (In Persian). 11. Jalali, M., Karami, S., Marj, A.F., 2016. Geostatistical Evaluation of Spatial Variation Related to Groundwater Quality Database: Case Study for Arak Plain Aquifer, Iran. Environmental Modeling & Assessment 21(6):707-719. 12. Khodai, K., Shahsavari, A.A., Etebari, B., 2006. Vulnerability assessment Jovin aquifer with DRASTIC and GODS methods. Iranian Journal of Geology 2(4): 73-87. (In Persian). 13. Kisi, O., Shiri, J., Tombul, M., 2012. Modeling rainfall-runoff process using soft computing techniques. Computers & Geosciences 51: 108-117. 14. Mirzavand, M., Ghasemiyeh, H., Sadatinejad, S.J., Akbari, M., 2015. Simulation of changes in groundwater quality using artificial neural network (case study: Kashan aquifer). Journal of Iranian Natural Resource 68 (1): 159-171. (In Persian). 15. Moosavi, V., Vafakhah, M., Shirmohammadi, B., Behnia, N., 2013. A wavelet-ANFIS hybrid model for groundwater level forecasting for different prediction periods. Water resources management 27(5):1301-1321. 16. Nourani, V., Alami, M.T., Vousoughi, F.D., 2016. Self-organizing map clustering technique for ANN-based spatiotemporal modeling of groundwater quality parameters. Journal of Hydro informatics 18(2):288-309. 17. Riad, S., Mania, J., Bouchaou, L., Najjar, Y., 2004. Predicting catchment flow in a semi-arid region via an artificial neural network technique. Hydrological Processes Journal, 18(13): 2387–2393. 18. Shirmohammadi, B., Vafakhah, M., Moosavi, V., Moghaddamnia, A., 2013. Application of several data-driven techniques for predicting groundwater level. Water resources management 27(2):419-432. 19. Taheri Tizro, A., Voudouris, K., Vahedi, S., Spatial variation of groundwater quality parameters: a case study from a semiarid region of Iran. International Bulletin of Water Resources & Development 1(3): 1-14. 20. Vafakhah, M., 2012. Application of artificial neural networks and adaptive neuro-fuzzy inference system models to short-term streamflow forecasting. Canadian Journal of Civil Engineering 39(4), 402-414. 21. Vafakhah, M., Mohseni saravi, M., Mahadavi, M., Alavipanah, S.K., 2011. Snowmelt Runoff Prediction by Using Artificial Neural Network in taleghan watershed. Journal of Iran-Watershed Management Science & Engineering 5(14): 23-36. (In Persian). 22. Wang, W.C., Chau, K.W., Cheng, C.h.T., Qiu, L., 2009. A comparison of performance of several artificial intelligence methods for forecasting monthly discharge time series. Journal of Hydrology 374(34), 323-331. 23. Yu, X., Liongs, S.Y., 2006. Forecasting of hydrologic time series with ridge regression in feature space. Journal of Hydrology 332 (3-4), 290-302. 24. Zare Abyaneh, H., Bayat, M., Akhavan, S., Mohamadi, M., 2011. Estimation of nitrate in groundwater Hamedan-Bahar plain using artificial neural network and data separation effect on prediction accuracy. Journal of Ecology 37(58): 129-140. (In Persian). 25. Zehtabian, Gh., Janfaza, E., Mohammad asgari, H., 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 17(1): 61-73. (In Persian).