تحلیل توسعه ژئومورفولوژی کارست با استفاده از مدل های منطق فازی و رگرسیون خطی چندمتغیره (توده های کارستی دیمه و پیرغار در استان چهارمحال و بختیاری) (مقاله علمی وزارت علوم)
درجه علمی: نشریه علمی (وزارت علوم)
آرشیو
چکیده
در این پژوهش، تحلیل توسعه ژئومورفولوژی کارست توده های کارستی دیمه و پیرغار در استان چهارمحال و بختیاری با استفاده از مدل سازی پهنه های کارستی با دو رویکرد منطق فازی و رگرسیون خطی چندمتغیره مورد ارزیابی قرار گرفته است.نتایج ارزیابی نقشه های پهنه بندی کارست سطحی تهیه شده از مدل منطق فازی با دولین های کارستی نشان داد که از 185 دولین های شناسایی شده در توده کارستی پیرغار تعداد 102 دولین در پهنه با توسعه زیاد کارست و در توده کارستی دیمه نیز از 140 دولین شناسایی شده 83 دولین در پهنه با توسعه زیاد کارست واقع شده است و همچنین نتایج مدل رگرسیون خطی چند متغیره با دولین های کارستی از 185 دولین شناخته شده در محدوده توده کارستی پیرغار نشان داد تعداد 94 دولین را در پهنه با توسعه زیاد کارست سطحی، همچنین از 140 دولین شناخته شده در محدوده توده کارستی دیمه تعداد 71 دولین در پهنه با توسعه زیاد کارست سطحی، شناسایی شده است. با توجه به این که مدل منطق فازی میزان تراکم بیشتری از دولین ها را در طبقات کارست با توسعه زیاد و متوسط تشخیص داده است و با نتایج صحت سنجی ارقام محاسبه شده مربوط به مساحت زیر منحنی های نام منحنی مشخصه عملکرد (ROC) که نشان داد، مدل منطق فازی برای هر دو توده کارستی در مقایسه با مدل رگرسیون خطی چند متغیره درجه کارآیی بالاتری بر اساس سطح زیر منحنی ROC (AUC) داشته است.Karstification analysis of karst geomorphology using fuzzy logic models and multivariate linear regression (Dimeh and Pirghar karst masses in Chaharmahal and Bakhtiari provinces)
Karstification analysis of karst geomorphology using fuzzy logic models and multivariate linear regression (Dimeh and Pirghar karst masses in Chaharmahal and Bakhtiari provinces) Karstification analysis of karst geomorphology using fuzzy logic models and multivariate linear regression (Dimeh and Pirghar karst masses in Chaharmahal and Bakhtiari provinces) Abstract In this research, to analyze the karst geomorphology of Dimeh and Pirghar karst masses in Chaharmahal and Bakhtiari province located in the Zagros highlands, karst dolin were first modeled with the closed contour line method (CCLS) in karst masses. Then the modeling of karstification zones has been done using two approaches, fuzzy logic and multivariate linear regression. The validation results of the calculated figures related to the area under the receiver operating characteristic (ROC) showed that the fuzzy logic model based on the area under the curve (AUC) of both karst massifs has a higher degree of efficiency compared to the multivariate linear regression model. Introduction Preparation of karst surface karstification map has a significant effect on increasing awareness and understanding of quantitative and qualitative characteristics of underground water resources of karst masses. According to the research records conducted in the region, so far in the studies related to the analysis of karst surface karstification in Dimeh and Pirghar karst masses, the topics of karst geomorphology and surface karstification, need to be studied more. The purpose of this research is to identify dolin, to investigate the role and extent of influence of each of the effective factors in the karstification of karst masses, to prepare zoning maps of areas in terms of karst karstification using fuzzy logic models and multivariate linear regression. In addition, to achieve the effectiveness of the above models in studies related to karst geomorphology. Methodology In this study, to analyze the karst geomorphology of Dimeh and Pirghar karst massifs in Chaharmahal and Bakhtiari province located in the Zagros highlands, karst dolin were first modeled using the closed contour line method (CCLS) in the studied karst mass. And then, by using the modeling of karst dolin with two approaches, fuzzy logic and multivariable linear regression, this evaluation has been done with the help of the parameters of elevation, slope, aspect, lithology, land use, soil, precipitation, temperature and distance from the fault. Finally, in order to evaluate and validate the accuracy of the results presented from fuzzy logic models and multivariate linear regression in predicting the zoning of areas in terms of karst karstification, Receiver Operating Characteristic (ROC) quantitative method has been used. Results and discussion The results of the evaluation of surface karstification zoning maps prepared from the fuzzy logic model with karst dolin showed that out of 185 dolin identified in the Pirghar karst mass, 102 dolin are in the area with high karstification and in the Dime karst mass. In addition, out of 140 identified dolin, 83 dolin are located in areas with high karstification. The results of surface karstification zonation maps using multivariate linear regression model with karst dolin out of 185 known dolin in Pirghar karst massif, 94 dolin in the area with high development of surface karstification and out of 140 known dolin in The range of Dime karst massif has been identified as 71 dolin in the area with high development of surface karstification. By modeling karst dolines in Dimeh and Pirghar karst masses, 185 dolines were identified in the area of Pirghar karst mass and 140 dolines were identified in the study area of Dimeh karst mass and their distribution map was prepared. The calculated figures related to the area under the curves (AUC) showed that these models have high diagnostic power to predict karstification zones. However, the fuzzy logic model has obtained a higher AUC score for both karst masses compared to the multivariate linear regression model and has shown better performance. Conclusion Identification of dolin using digital elevation model modeling, in addition to reducing time and costs, has high accuracy and has many applications for aquifers. The evaluation of surface karstification zoning maps prepared with the validation results of the area under the ROC curves showed that the fuzzy logic model has received a higher AUC score for both karst masses compared to the multivariate linear regression model, which shows that the quality it shows better detection and performance of this model. The results of this research are the same as the results of the studies of Moradi et al. (2016), Mokram et al. (2018), Safari et al. (2018) and Dastranj et al. (2018) are the same. Keywords: karst geomorphology, fuzzy logic model, linear regression model, ROC curve, Chaharmahal and Bakhtiari.
تبلیغات
