کارایی مدل ترکیبی نسبت فراوانی-ماشین بردار پشتیبان در شناسایی مناطق مستعد سیل آبخیز کلات

نوع مقاله : پژوهشی

نویسندگان

1 دانشجوی دکتری، دانشکدۀ منابع طبیعی، دانشگاه علوم کشاورزی و منابع طبیعی ساری

2 استاد، دانشکدۀ منابع طبیعی، دانشگاه علوم کشاورزی و منابع طبیعی ساری

3 دانشیار، دانشکدۀ منابع طبیعی، دانشگاه علوم کشاورزی و منابع طبیعی ساری

4 استاد‌، مرکز مدل‌سازی پیشرفته و سیستم‌های اطلاعات جغرافیایی، دانشکدۀ مهندسی و فناوری اطلاعات، دانشگاه تکنولوژی سیدنی، NSW، استرالیا

چکیده

جاری شدن سیل آثاری منفی بر محیط زیست، اقتصاد، جوامع انسانی و صنعت دارد. امروزه، کاربرد مدل‌های پیشرفتۀ سیلاب برای شناسایی مناطق حساس و بهبود سیستم مدیریت سیل رشد چشمگیری داشته است. در این میان، تعدادی از محققان با ترکیب برخی مدل‌‌ها به نتایج قابل قبولی برای شناسایی مناطق مستعد سیل دست یافتند. از آنجا که آبخیز کلات از منظر سیلاب به‌خصوص سیلاب‌های اخیر سال 1398 جزء مناطق پرخطر استان خراسان رضوی محسوب می‌شود و تا کنون نیز در آن از تکنیک‌های پیشرفته برای برآورد احتمال وقوع سیل استفاده نشده است، بنابراین مدل ترکیبی نسبت فراوانی- ماشین بردار پشتیبان FR-SVM برای مدل‌سازی سیلاب انتخاب شده و با مدل مستقل SVM مقایسه شد. پس از بررسی‌های صورت‌گرفته 73 نقطۀ سیل‌گیر ثبت شده و 15 عامل مؤثر بر وقوع سیل شامل بارش سالانه، زمین‌شناسی، کاربری اراضی/پوشش زمین، طول شیب، فاصله از رودخانه، تحلیل سایۀ پستی و بلندی‌ها، ارتفاع، شاخص همگرایی، تحدب و تعقر طولی و عرضی، شیب، شاخص قدرت جریان، شاخص زبری توپوگرافی، شاخص رطوبت توپوگرافی و عمق دره، در نظر گرفته شد. ارزیابی مدل‌‌ها توسط معیارهای مختلف سنجش دقت از جمله ضریب کاپا، ریشۀ میانگین مربعات خطا، منحنی مشخصۀ عملکرد سیستم و منحنی میزان پیش‌بینی، صورت گرفت. مدل FR-SVM با منحنی میزان پیش‌بینی 8862/0، دقت زیاد و کارایی بهتری را نسبت به SVM نشان داد. این نتایج می‌تواند برای مدیریت مناطق آسیب‌پذیر سیل و سایر کاربردهای منابع طبیعی استفاده شود.

کلیدواژه‌ها


عنوان مقاله [English]

The Efficiency of an Ensemble Frequency Ratio-Support Vector Machine Model in the Detection of Flood-Prone Areas of the Kalat Basin

نویسندگان [English]

  • Hamzeh Mojaddadi Rizeei 1
  • Mahmoud Habibnezhad Roshan 2
  • Kaka Shahedi 3
  • Biswajeet Pradhan 4
1 PhD Student, Department of Watershed Management, Sari Agricultural Sciences and Natural Resources University, Iran
2 Professor, Department of Watershed Management, Sari Agricultural Sciences and Natural Resources University, Iran
3 Associate Professor, Department of Watershed Management, Sari Agricultural Sciences and Natural Resources University, Iran
4 Professor, Centre for Advanced Modelling and Geospatial Information Systems, Faculty of Engineering and IT, University of Technology Sydney, NSW 2007, Australia
چکیده [English]

Flooding hurts the environment, economy, human communities, and industry. Therefore, comprehensive knowledge on flood probability modeling is essential to identify sensitive areas and to improve flood management systems. Advanced floods models usage has been grown dramatically today. That's why several researchers have integrated some models obtaining acceptable results for identifying flood-prone areas. Since numerous high-risk floods have occurred in the Kalat Basin and no advanced techniques have been used to estimate flood probability, so the Frequency Ratio-Support Vector Machine (FR-SVM) ensemble model was selected for flood modeling. Accuracy and efficiency evaluation, consequently, has been compared with the standalone SVM model. By investigation, 73 floods points were recorded according to recent 2018 end-month floods, and 15 conditioning factors including annual precipitation, geology, land use/land cover, slope length, river distance, analytical hill shading, elevation, convergence index, profile and plan curvatures, slope, stream power index, topographic roughness index, topographic wetness index and valley depth were considered. Models were evaluated by various precision criteria such as kappa coefficient, root means square errors, receiver operating characteristics and precision-recall curve. The FR-SVM model with a precision-recall curve of 0.8862 showed high accuracy and performance than SVM. These results can be used to manage flood-prone areas and other natural resource applications.

کلیدواژه‌ها [English]

  • Flood probability modeling
  • Support vector machine
  • frequency ratio
  • Ensemble model
  • Kalat Basin
[1]. Rizeei HM, Azeez OS, Pradhan B, and Khamees HH. Assessment of groundwater nitrate contamination hazard in a semi-arid region by using integrated parametric IPNOA and data-driven logistic regression models. Environmental monitoring and assessment. 2018; 190(633): 1-17.
[2]. Tehrany MS, Pradhan B, and Jebur MN. Flood susceptibility analysis and its verification using a novel ensemble support vector machine and frequency ratio method. Stochastic Environmental Research and Risk Assessment. 2015; 29: 1149–1165.
[3]. Rizeei HM, Pradhan B, and Saharkhiz MA. An integrated fluvial and flash pluvial model using 2D high-resolution sub-grid and particle swarm optimization-based random forest approaches in GIS. Complex and Intelligent Systems. 2019; 5: 283–302.
[4]. Khosravi Kh, Pham BT, Chapi K, Shirzadi A, Shahabi H, Revhaug I, and et al. A comparative assessment of decision trees algorithms for flash flood susceptibility modeling at Haraz watershed, northern Iran. Science of the Total Environment, 2018. 627: p. 744-755.
[5]. Mojaddadi H, Pradhan B, Nampak H, Ahmad N, and Ghazali AHB. Ensemble machine-learning-based geospatial approach for flood risk assessment using multi-sensor remote-sensing data and GIS. Geomatics, Natural Hazards and Risk. 2017; 8(2): 1080–1102.
[6]. Tehrany MS, Pradhan B, and 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. Journal of Hydrology. 2013; 504: 69-79.
[7]. Chapi K, Singh VP, Shirzadi A, Shahabi H, Bui DT, Pham BT, and et al. A novel hybrid artificial intelligence approach for flood susceptibility assessment. Environmental modelling and software. 2017; 95: 229-245.
[8]. Cao C, Xu P, Wang Y, Chen J, Zheng L, Niu C. Flash flood hazard susceptibility mapping using frequency ratio and statistical index methods in coalmine subsidence areas. Sustainability. 2016; 8: 948-964.
[9]. Ghorbani MA, Azani A, and Naghipour N. A Comparison of Support Vector Machine Performance with Other Intelligent Models in Rainfall-Runoff Simulation. Watershed Management. 2016; 7(13): 99-103. [Persian].
[10].            Tehrany MS, Pradhan B, and Jebur MN. Flood susceptibility mapping using a novel ensemble weights-of-evidence and support-vector machine models in GIS. Journal of Hydrology. 2014; 512: 332–343.
[11].            Rahmati O, Pourghasemi HR, Zeinivand H. Flood susceptibility mapping using frequency ratio and weights-of-evidence models in the Golastan province, Iran. Geocarto Int. 2016; 31: 42–70.
[12].            Hoseinzadeh SR, Khanehbad M, and Khosravi A. Urban Flood Risk Zoning Using Paleo-flood Hydrology Data (Case Study: Kalat Naderi City, Khorasan Razavi). Quantitative Geomorphology Research. 2013; 3(1): 20-36. [Persian].
[13].            Hoseinzadeh SR, Khanehbad M, and Khosravi A. Study of enormous floods in Kalat River using old level evidences. Geographical studies of arid regions. 2014; 5(17): 1-16. [Persian].
[14].            Zare M. March, April and May 2019 Floods and Climate changes in Iran, with special sight to Khuzeshtan Province floods. 2019. The Academy of Sciences Islamic Republic of Iran. http://www.ias.ac.ir/index.php/2015-09-21-08-02-04/1431-mehdi-zare-flood. [Persian].
[15].            Iranian Students’ News Agency (ISNA). Khorasan, 20 April 2019, News Cod: 981-13190-5, Reporter Cod: 15043. https://khorasan.isna.ir/default.aspx?NSID=5andSSLID=46andNID=145439. [Persian].
[16].            Pradhan B. Flood susceptible mapping and risk area delineation using logistic regression, GIS and remote sensing. Journal of Spatial Hydrology. 2010; 9: 1–18.
[17].            Merz B, Thieken AH, and Gocht M. Flood risk mapping at the local scale: concepts and challenges. In: Flood risk management in Europe: innovation in policy and practice. Advances in Natural and Technological Hazards Research. 2007; 25: 231–251.
[18].            Shafizadeh Moghadam H, Valavi R, Shahabi H, Chapi K, and Shirzadi A. Novel forecasting approaches using combination of machine learning and statistical models for flood susceptibility mapping. Journal of environmental management. 2018; 217: 1–11.
[19].            Botzen W, Aerts J, and Van den Bergh J. Individual preferences for reducing flood risk to near zero through elevation. Mitigation and Adaptation Strategies for Global Change. 2013; 18(2): 229-244.
 
[20].            Maier HR, Jain A, Dandy GC, and Sudheer KP. Methods used for the development of neural networks for the prediction of water resource variables in river systems: Current status and future directions. Environmental Modelling Software. 2010; 25: 891-909.
[21].            Khosravi K, Nohani E, Maroufinia E, and Pourghasemi HR. A GIS-based flood susceptibility assessment and its mapping in Iran: a comparison between frequency ratio and weights-of-evidence bivariate statistical models with multicriteria decision-making technique. Natural hazards. 2016; 83(2): 947–987.
[22].            Pike RJ. Diversity in quantative surface analysis progress in Physical Geography. Geomorphology. 2000; 24:1-20.
[23].            Ghanavati A, Saffari A, Beheshti A, and Mansourian A. Flood potential mapping using ensemble Hydrologic model CN-AHP in GIS. Case study: Balkholu River Basin. Natural Geography Journal. 2014; 7(52): 67-80. [Persian].
[24].            Sarhadi A, Soltani S, and Modarres R. Probabilistic flood inundation mapping of ungauged rivers: Linking GIS techniques and frequency analysis. Journal of Hydrology. 2012; 458: 68–86.
[25].            Solaimani K. Urban Flood Hydrology and Quantitative Modeling in GIS and SWMM Environment. 1st ed. Mazandaran. Haraz University. 2015: p 322. [Persian].
[26].            Townsend PA, and Walsh SJ. Modeling floodplain inundation using an integrated GIS with radar and optical remote sensing. Geomorphology. 1998; 21: 295-312.
[27].            Meyer V, Scheuer S, and Haase D. A multicriteria approach for flood risk mapping exemplified at the Mulde River, Germany. Natural Hazards. 2009; 48: 17-39.
[28].            Bui DT, Panahi M, Shahabi H, Singh VP, Shirzadi A, Chapi A, and et al. Novel Hybrid Evolutionary Algorithms for Spatial Prediction of Floods. Scientific Reports. 2018; 8:(15364) 1-14.
[29].            Rokach L. Ensemble-based classifiers. Artificial Intelligence Review. 2010; 33(1-2): 1-39.
[30].            Arabameri A, Pourghasemi HM, and Shirani K. Flood susceptibility Zonation using a novel ensemble Bayesian–AHP model. Case study: Neka Basin, Mazandaran, Iran. Ecohydrology. 2017; 4(2): 447-462. [Persian].
[31].            Chapelle O, Vapnik V, Bousquet O, and Mukherjee S. Choosing multiple parameters for Support Vector Machines. Machine Learning. 2002; 46(1-3): 131–159.
[32].            Samui P. Slope stability analysis: a support vector machine approach. Environmental Geology. 2008; 56(2): 255-267.
[33].            Golshan M, Esmaeely A, and Khosravi Kh. Flood susceptibility evaluation of Talar Basin using FR model. Natural Environment Hazards. 2018; 7(15): 1-16. [Persian].
[34].            Khosravi Kh, Maroufinia E, Nohani E, and Chapi K. Efficiency evaluation of Logistic Regression Model in flood susceptibility mapping. Iranian Natural resources, Watershed Management. 2016; 69(4): 863-876. [Persian].
[35].            Kheyrizadeh M, Maleki J, and Amunia H. Flood hazard zonation in Mardagh Chay Basin using ANP Model. Quantitive Geomorphology Researches. 2013; 1(3): 56-71. [Persian].
[36].            Maroufinia E, Nohani E, Khosravi Kh, and Chapi K. Evaluation of Statistical Index Method in Flood Susceptibility Mapping. Water and Soil Science. 2016; 26(2): 201-214. [Persian].
[37].            Youssef AM, Pradhan B, and Hassan AM. Flash flood risk estimation along the St. Katherine road, southern Sinai, Egypt using GIS based morphometry and satellite imagery. Environmental Earth Scinces. 2011; 62: 611–623.
[38].            Manandhar B. Flood plain analysis and risk assessment of Lothar Khola. MSc Thesis, Tribhuvan University, Phokara, Nepal. 2010. P 65.
[39].            Lee MJ, Kang JE, and Jeon S. Application of frequency ratio model and validation for predictive flooded area susceptibility mapping using GIS. 32nd IEEE International Geoscience and Remote Sensing Symposium(IGARSS), Munich. Germany. 2012; 895–898.
[40].            Nohani E, Darabi F, Maroufinia E, and Khosravi Kh. Evaluation of Entropy Shannon model producing Flood probability and susceptibility mapping in Haraz Basin. Natural Environment Hazards. 2016; 5(10): 99-116. [Persian].
[41].            Darabi H, Shahedi K, and Mardian M. Flood probability and susceptibility mapping using Frequency Ration Model in Pol Doaab Shazand Basin. Journal of Watershed Engineering and Management. 2016; 8(1): 68-79. [Persian].
[42].            Bui DT, Pradhan B, Nampak H, Bui QT, Tran QA, and Nguyen QP. Hybrid artificial intelligence approach based on neural fuzzy inference model and metaheuristic optimization for flood susceptibility modeling in a highfrequency tropical cyclone area using GIS. Journal of Hydrology. 2016; 540: 317-330.
[43].            Hong H, Tsangaratos P, Ilia I, Liu J, Zhua AX, and Chen W. Application of fuzzy weight of evidence and data mining techniques in construction of flood susceptibility map of Poyang County, China. Science of the Total Environment. 2018; 625: 575–588.
[44].            Pallard B, Castellarin A, and Montanari A. A look at the links between drainage density and
flood statistics. Hydrology and Earth System Sciences. 2009; 13(1): 1019–1029.
[45].            Opolot E. Application of remote sensing and geographical information systems in flood management: a review. Research Journal of Applied Science Engineering and Technology. 2013; 6: 1884-1984.