تعیین پتانسیل آب زیرزمینی با استفاده از مدل‌های شبکۀ عصبی مصنوعی، جنگل تصادفی، ماشین بردار پشتیبان و رگرسیون خطی (مطالعۀ موردی: حوضۀ آبخیز دریاچۀ ارومیه)

نوع مقاله : پژوهشی

نویسندگان

1 بخش تحقیقات اقتصادی-اجتماعی، مرکز تحقیقات و آموزش کشاورزی و منابع‏ طبیعی زنجان، سازمان تحقیقات، آموزش و ترویج کشاورزی، زنجان

2 بخش تحقیقات حفاظت‏ خاک و آبخیزداری، مرکز تحقیقات و آموزش کشاورزی و منابع‏ طبیعی خراسان‏ رضوی، سازمان تحقیقات، آموزش و ترویج کشاورزی، مشهد

3 مدیرعامل شرکت سایه‌گستر دشت البرز، کرج

4 دکترای آبخیزداری، دانشگاه علوم کشاورزی و منابع طبیعی گرگان

چکیده

هدف از تحقیق حاضر، تعیین مناطق دارای پتانسیل آب زیرزمینی با استفاده از مدل‏های شبکۀ عصبی مصنوعی (ANN)، جنگل تصادفی (RF)، ماشین بردار پشتیبان (SVM) و رگرسیون خطی (GLM) است. در تحقیق حاضر از 14 پارامتر شامل ارتفاع، شیب، جهت شیب، انحنا، فاصله از آبراهه و گسل، تراکم آبراهه و گسل، لیتولوژی، متوسط بارندگی، کاربری اراضی، شاخص موقعیت توپوگرافیک (TPI)، موقعیت شیب نسبی (RSP) و شاخص رطوبت توپوگرافیک (TWI) برای بررسی پتانسیل آب زیرزمینی استفاده شده است. از مجموع 10624 چشمه، به صورت تصادفی 70 درصد به عنوان داده‏های آزمون و 30 درصد به عنوان داده‏های اعتبارسنجی طبقه‏بندی شدند. همچنین، برای تعیین مهم‏ترین پارامترها از مدل RF استفاده شد. تست هم‏خطی بین پارامترها با استفاده از نرم‏افزار SPSS انجام شد. از منحنی تشخیص عملکرد نسبی برای قدرت پیش‏بینی مدل‏ها و شاخص سطح سلول هسته (SCAI) به منظور دقت تفکیک بین طبقات استفاده شد. نتایج نشان داد بین پارامترها هم‏خطی وجود ندارد. نتایج مدل RF نشان داد به‌ترتیب پارامترهای ارتفاع، کاربری اراضی، شیب، فاصله از گسل، TWI و لیتولوژی مهم‏ترین عوامل تأثیرگذار بر پتانسیل آب زیرزمینی هستند. همچنین، بر اساس منحنی ROC در هر دو بخش آموزش (915/0=AUC) و اعتبارسنجی (909/0=AUC)، مدل ANN دارای بیشترین دقت بودند و مدل‏های RF، SVM و GLM در رده‏های بعدی قرار گرفتند. همچنین، نتایج شاخص سطح سلول هسته نشان داد هر چهار مدل با دقت مناسبی به تفکیک طبقات پرداخته‏اند. بر اساس مدل ANN، 4/31 درصد حوضه ‌پتانسیل آب زیرزمینی زیاد و خیلی زیاد دارد.

کلیدواژه‌ها


عنوان مقاله [English]

Determination of Groundwater Potential Using Artificial Neural Network, Random Forest, Support Vector Machine and Linear Regression Models (Case Study: Lake Urmia Watershed)

نویسندگان [English]

  • Alireza Rabet 1
  • Ali Dastranj 2
  • Sorena Asadi 3
  • Omid Asadi Nalivan 4
1 Economic, Social and Extension Research Department, Zanjan Agricultural and Natural Resources Research and Education Center, AREEO, Zanjan, Iran
2 Soil Conservation and Watershed Management Department, Agricultural and Natural Resources Research Center of khorasan Razavi,AREEO, Mashhhad, Iran
3 Managing Director of Sayehgostar Dasht Alborz company, Karaj, Iran
4 Ph.D. Graduate in Watershed Management Sciences and Engineering, Gorgan University of Agricultural Sciences and Natural Resources, Iran
چکیده [English]

The purpose of this study is to determine the areas with groundwater potential using artificial neural network (ANN), random forest (RF), support vector machine (SVM) and linear regression (GLM) models. In the present study, 14 parameters groundwater potential including altitude, slope, slope direction, curvature, distance to stream and fault, stream and fault density, lithology, average rainfall, land use, topographic position index (TPI), relative slope position (RSP) and topographic wetness index (TWI) were used. From a total of 10,624 springs, randomly 70% as test data and 30% as validation data were classified. The RF model was also used to determine the most important parameters. Alignment test between parameters was performed using SPSS software. The Receiver operating characteristic was used to Predictive power of models and the Seed Cell Area Indexes (SCAI) was used to accurately distinguish between classes. The results showed that there is no alignment between the parameters. The results of RF model showed that the parameters of height, land use, slope, and distance from fault, TWI and lithology are the most important factors affecting groundwater potential, respectively. Also, based on the ROC curve in both training (0.915) and validation (0.909), the ANN model had the highest accuracy and the RF, SVM and GLM models were in the next categories. Also, the results of the seed cell area index showed that all four models have separated the classes with appropriate accuracy. According to the ANN model, 31.4% of the basin has high and very high groundwater potential.

کلیدواژه‌ها [English]

  • Artificial Neural Network
  • Groundwater Potential
  • Lake Urmia
  • Machine learning
  • Random forest
[1]. Jothibasu A, Anbazhagan S. Modeling groundwater probability index in Ponnaiyar River basin of South India using analytic hierarchy process. Model. Earth Syst. Environ. 2016; (2): 109.
[2]. Rahmati O, Pourghasemi HR, Melesse AM. Application of GIS-based data driven random forest and maximum entropy models for groundwater potential mapping: a case study at Mehran Region, Iran. Catena. 2016; (137): 360–372.
[3]. Arabameri AR, Sohrabi M, Rezaei KH, Yamani M, Shirani K. Simulation of Najaf-Abad watershed groundwater using data driven ensemble model EBF-Index of entropy. Journal of Water and Soil Conservation. 2018; 25(2), 25-48. ]Persian[
[4]. Davoodi Moghaddam D, Rezaei M, Pourghasemi HR, Pourtaghie ZS, and Pradhan B. Groundwater spring potential mapping using bivariate statistical model and GIS in theTaleghan watershed, Iran. Arabian Journal Geoscience. 2015; 8(2), 913-929.
[5]. Chen W, Li H, Hou E, Wang S, Wang G, Panahi M, et al. GIS-based groundwater potential analysis using novel ensemble weights-of-evidence with logistic regression and functional tree models. Science of the Total Environment. 2018; (634): 853–867.
[6]. Agarwal R, Garg PK. Remote Sensing and GIS based groundwater potential & recharge zones mapping using multi criteria decision analysis making technique. Water Resources Management. 2016 ;( 30): 243-260.
[7]. Rahmati O, Naghibi SA, Shahabi H, Bui DT, Pradhan B, Azareh A, et al. Groundwater spring potential modeling: Comprising the capability and robustness of three different modeling approaches. Journal of Hydrology. 2018; (565): 248–261.
 
[8]. Chen W, Pradhan B, Li S, Shahabi H, Rizaei H, Hou E, Wang S. Novel Hybrid Integration Approach of Bagging-Based Fishers Linear Discriminate Function for Groundwater Potential Analysis. Natural Resources Research.https://doi.org/10.1007/s11053-019-09465-w.
[9]. Mogaji KA, Lim HS. Application of Dempster-Shafer theory of evidence model to geoelectric and hydraulic parameters for groundwater potential zonation.Journal of Astronomy and Geophysics. 2018; 7(1): 134-148.
[10].            Naghibi SA, Moghaddam D, Kalantari B, Pradhan B, Kisi O. A comparative assessment of GIS-based data mining models and a novel ensemble model in groundwater well potential mapping. Journal of Hydrology. 2017; (548): 471–483.
[11].            Lee S, Song KY, Kim Y, Park I. Regional groundwater productivity potential mapping using a geographic information system (GIS) based artificial neural network model. Hydrogeology Journal. 2012; (20): 1511–1527.
[12].            Naghibi SA, Pourghasemi HR. A comparative assessment between three machine learning models and their performance comparison by bivariate and multivariate statistical methods for groundwater potential mapping in Iran, Water Resources Management. 2015; 29(14): 5217–5236.
[13].            Mousavi SM, Golkarian A, Naghibi SA, Kalantari B, and Pradhan B. GIS-based Groundwater Spring Potential Mapping Using Data Mining Boosted Regression Tree and Probabilistic Frequency Ratio Models in Iran. Geosciences. 2017; 3(1): 91-115.
[14].            Lee S, Hyun Y, Lee M. Groundwater Potential Mapping Using Data Mining Models of Big Data Analysis in Goyang-si, South Korea. Sustainability. 2019; (11): 1678.
[15].            Haghizade A, Moghaddam D, Pourghasemi H. GIS-based bivariate statistical techniques for groundwater potential analysis.Journal of Earth System Science. 2017; (126): 109.
[16].            Thapa R, Gupta S, Guin S, Kaur H. Assessment of groundwater potential zones using multi-influencing factor (MIF) and GIS: a case study from Birbhum district, West Bengal. Appl. Water Sci.2017; (7): 4117-4131.
[17].            Miraki SH, Zanganeh S, Chapi K, Singh VP, Shirzadi A, Shahabi H, Pham BT. Mapping Groundwater Potential Using a Novel Hybrid Intelligence Approach. Water Resources Management.2019; (33): 281–302.
[18].            Kordestani MD, Naghibi SA, Hashemi H, Ahmadi K, Kalantar B, Pradhan B. Groundwater potential mapping using a novel data-mining ensemble model. Hydrogeology. 2018; (27): 211–224.
[19].            Golkarian A, Naghibi SA, Kalantar B, Pradhan B. Groundwater potential mapping using C5.0, random forest, and multivariate adaptive regression spline models in GIS. Environmental Monitoring and Assessment. 2018; 190(3): 149.
[20].            Golkarian A, Rahmati O. Use of a maximum entropy model to identify the key factors that influence groundwater availability on the Gonabad Plain, Iran. Environmental Earth Sciences. 2018; (77): 369.
[21].            Shafiei N, AmirAhmadi A, Rahmani A. Finding potential sources of underground water catchment model Shannon entropy Nour plain mamasani. Journal of Geographic space. 2019; 19(66): 291-304. ]Persian[
[22].            Zabihi M, Pourghasemi HR, Behzadfar M. Groundwater Potential Mapping using Shannon's Entropy and Random Forest Models in the Bojnourd Township. Journal of EcoHydrology. 2015; 2(2): 221–232. ]Persian[
[23].            Chowdhury A, Madan C, Bimal C. Integrated remote sensing and GIS-based approach for assessing groundwater potential in West Medinipur district, West Bengal, India. International Journal of Remote Sensing. 2009; (30): 231–250.
[24].            Yeh HF, Cheng YS, Lin H, Lee C. Mapping groundwater recharge potential zone using a GIS approach in Hualian River, Taiwan. Sustainable Environment Research. 2016; (26): 33–43.
[25].            Gaur S, Chahar R, GraillotD. Combined use of groundwater modeling and potential zone analysis for management of groundwater. International Journal of Applied Earth Observation and Geoinformation. 2011; (13): 127–139.
[26].            Manap MA, Nampak H, Pradhan B, Lee S, Soleiman WNA, Ramli MF. Application of probabilistic-based frequency ratio model in groundwater potential mapping using remote sensing data and GIS. Arab. J. Geosci. 2012; 7(2): 711-724.
[27].            Ayazi MH, Pirasteh S, Arvin AKP, PradhanB, Nikouravan B, Mansor S. Disasters and risk reduction in groundwater: Zagros mountain southwest Iran using geo-informatics techniques. Dis. Adv.2010; (3): 51-57.
[28].            Samy I, Shattri M, Bujang BK, Ahmad R.M. Structural geologic control with the limestone bedrock associated with piling problems using remote sensing and GIS: a modified geomorphological method. Environ. Earth Sci. 2011; 66: 8. 2185-2195.
[29].            Tehrany MS, Pradhan B, Jebur MN. Spatial prediction of flood susceptible areas using rule based decision tree (DT) and a novel ensemble bivariate and multivariate statistical models in GIS. J. Hydrol. 2013; (504): 69-79.
[30].            Lee S, Ryu JH, Lee MJ, Won JS. The application of artificial neural network to landslide susceptibility mapping at Janghung, Korea. Mathematical Geology. 2006; 38(2): 199-220.
[31].            Caniani D, Pascale S, Sdao F, and Sole A. Neural networks and landslide susceptibility: a case study of the urban area of Potenza. Natural hazards. 2008; (45): 55-72.
[32].            Peng L, Ruiqing NBH, Xueling W, Yannan Z, Runqing Y. Landslide Susceptibility Mapping Based on Rough Set Theory and Support Vector Machines: A Case of the Three Gorges Area, China. Geomorphology. 2014; (204): 287–301.
[33].            Tien Bui D, Pradhan B, Owe L, Inge R. LandslideSusceptibility Assessment in Vietnam Using Support Vector Machines, Decision Tree, ANaive Bayes Models. Mathematical Problems in Engineering. 2012; 1-26.
[34].            CataniF, Lagomarsino D, Segoni S, Tofani V. Landslide susceptibility estimation by random forests technique: sensitivity and scaling issues. Natural Hazards Earth System Science. 2013; 13: 2815-2831.
[35].            Nicodemus KK. Letter to the Editor: On the stability and ranking of predictors from random forest variable importance measures. Briefings in Bioinformatics. 2011; (12): 369-373.
[36].            Youssef AM, Pourghasemi H.R, Pourtaghi Z.S,Al-Katheeri M.M. Landslide susceptibility mapping using random forest, boosted regression tree, classification and regression tree, and general linear models and comparison of their performance at Wadi Tayyah Basin, Asir Region. Saudi Arabia Landslides. 2015; DOI 10.1007/s10346-015-0614-1.
[37].            Arulbalaji P, Padmalal D, Sreelash K. GIS and AHP techniques Based Delineation of Groundwater potential Zones: a case study from southern Western Ghats, India. Scientific Reports, 2019; 9: 2082.
[38].            Al-Abadi A, Al-Temmeme A, Al-Ghanimy A. A GIS-based combining of frequency ratio and index of entropy approaches for mapping groundwater availability zones at Badra-Al-Gharbi–Teeb areas, Iraq, Sustain. Water Resources Management. 2016; 2(3): 265-283.
[39].            Deng F, Deng Z,Lv D,Wang D, Duan H,Xing Z. Application of remote sensing and GIS analysis in groundwater potential estimation in west Liaoning Province, China. Journal of Engineering Research. 2016; 4(3): 1-17.
[40].            Muniraj K, Jesudhas CJ, Chinnasamy A. Delineating the Groundwater Potential Zone in Tirunelveli Taluk, South Tamil Nadu, India, Using Remote Sensing, Geographical Information System (GIS) and Analytic Hierarchy Process (AHP) Techniques. Proceedings of the National Academy of Sciences, India Section A: Physical Sciences. 2019; Doi.org/10.1007/s40010-019-00608-5.