پهنه‌بندی خطر زمین لغزش با استفاده از مدل رگرسیون لجستیک و آنفیس در حوضه آبریز هشتجین استان اردبیل

شناسایی مناطق حسّاس و تهیه نقشه پهنه بندی خطر زمین لغزش، گام مهمّی در مسیر پیشگیری و کاهش خسارت های ناشی از وقوع این پدیده است. حوضه هشتجین با داشتن چهره کوهستانی و با توجّه به وضعیّت زمین شناسی، لیتولوژی، اقلیمی و انسانی، شرایط لازم را برای شکل گیری حرکات لغزشی دارد؛ بنابراین هدف نوشتار پیش رو پهنه بندی خطر زمین لغزش در این منطقه است. در پژوهش حاضر حسّاسیت زمین لغزش برای حوضه آبریز هشتجین با توجّه به ارزیابی کارایی نتایج حاصل از دو مدل رگرسیون لجستیک و آنفیس برای دستیابی به هدف پژوهش، تجزیه و تحلیل شد. با استفاده از تفسیر عکس های هوایی و بازدید میدانی، مناطق شاهد به عنوان متغیّر وابسته، با جی.پی.اِس ثبت و درادامه عوامل موثّر بر ایجاد زمین لغزش در منطقه شامل شیب، جهت، خطوط ارتفاعی، فاصله از آبراهه، فاصله از گسل، فاصله از جاده، زمین‌شناسی، کاربری اراضی و بارش با توجّه به مرور منابع مختلف، مطالعات میدانی و مشورت با کارشناسان شناسایی شد؛ سپس در محیط آرک جی.آی.اِس لایه ها به مثابه متغیّر مستقل تهیه و با ورود لایه‌های مذکور به محیط نرم افزار ترست، متلب به‌ترتیب مدل رگرسیون لجستیک و آنفیس اجرا شدند. نقشه نهایی خطر زمین لغزش منطقه در پنج کلاس خطرپذیری تهیه شد. در پژوهش حاضر از 25% نمونه های شاهد به عنوان داده های آزمون، به منظور سنجش میزان صحّت مدل های مورد بررسی، استفاده شد. نتایج صحّت سنجی کارایی مدل های پیش گفته با اجرای منحنی راک نشان داد که دقّت مدل آنفیس و رگرسیون لجستیک به ترتیب برابر 88.23% و 86.45% بوده است. نتایج بر اساس مدل انفیس بیانگر آن است که حدود 4854 هکتار،  معادل 20.6% از منطقه هشتجین از نظر مخاطره زمین لغزش در کلاسه زیاد و خیلی زیاد قرار دارند.

Landslide Risk Zoning using Logistic Regression and Anfis Models in Hashtjin Catchment Area in Ardabil Province

Landslides are one of the most important environmental processes, especially in mountainous landscapes. Identifying sensitive areas and preparing a landslide risk zoning map is an important step in preventing and reducing the damage caused by this phenomenon. Hashtjin basin, with its mountainous face and considering the geological, lithological, climatic and human conditions, has the necessary conditions for the formation of landslides; therefore, the current study aims at landslide risk zoning in the given area. Therefore, landslide sensitivity analysis for Hashtjin watershed is evaluated according to the efficiency of the results obtained from two models of logistic regression and Anfis to achieve the research goal. Using the interpretation of aerial photographs and field visits, control areas, as a dependent variable, were recorded by GPS. Then, the factors affecting landslides in the area including slope, direction, elevation lines, distance from waterway, distance from fault, distance from road, geology, land use and rainfall were identified according to various sources, field studies and consultation with experts; Then, layers were prepared as independent variables in GIS Arc environment. Moreover, logistic and ANFIS regression models were implemented by entering the aforesaid layers into TERRSET and MATLAB software environment, respectively. The final landslide hazard map of the area was prepared in 5 hazard classes. In this study, 25% of the control samples were used as test data to measure the accuracy of the studied models. The results of validation of the performance of the mentioned models by performing the ROC curve showed that the accuracy of Anfis model and logistic regression were equal to 88.23 and 86.45%, respectively. The findings from Enfis model reveal that approximately 4854 hectares, equivalent to 20.6% of the Hashtjin area are in high and very high class in terms of landslide risk.Extended Abstract1IntroductionLandslides are defined as the displacement of soil under the influence of gravity. It is one of the most destructive and catastrophic geological hazards that has a lot of harmful impacts on human societies every year, including economic damage, destruction of infrastructure and environmental problems. Identifying landslideprone areas and mass movements is a necessity for natural resource management and development planning. In this regard, mapping landslideprone areas is one of the most important steps required for planning and decisionmaking in the field of land use. Mass movements are in special situations. Iran, with its mainly mountainous topography, high tectonic and seismic activity, climatic and geological diversity, has natural conditions for a wide range of landslides. Hashtjin watershed located in Ardabil province, which is geomorphologically the western and northern part of the mountain basin and consists of high slopes with high slopes, is no exception to this rule. The study of landslide distribution in this basin shows that a lot of landslides have occurred and have high potential from unstable slopes. Based on field observations, several landslides have occurred in this area and more landslides are very likely to occur. In fact, it is necessary to pay attention to the landslides of this time zone.2Materials and MethodsIn order to investigate the status of landslides in the past, as educational sample points, and also to determine the factors affecting its occurrence in Hashtjin region, first the aerial photographs of the region during the period from 2015 to 2020 were reviewed. Then, Aerial reconnaissance and adaptation to terrestrial realities were carried out with the help of the natives of the area to confirm the findings from photo interpretation. By field survey, 41 landslides were identified in the study area and the position and limits of each of them were recorded using a GPS device. At the same time, the factors affecting the occurrence of landslides in the region were determined in the field visit including slope, direction, elevation lines, distance from the waterway, distance from the road, lithology status and land cover / land use. In this study, slope, direction and elevation lines maps were prepared using DEM (digital elevation model) ten meters ASTER. Besides, the waterway map of the area was drawn using ARC / Hydro plugin. The road map of the region was also obtained from the data of the Ministry of Roads and Urban Development. Also, the lithological status map of the region was digitized from the geological map with the scale of 1: 100000. Slope, direction, elevation, lithology and land cover / land use maps that needed to be classified were also classified after scoring.3Results and DiscussionComparison of lithological map (Figure 6) with spatial distribution of landslides (Figure 2) and logistic regression equation (Equation 12) and zoning maps obtained from quantitative research models (Figures 10 and 12) show the decisive effect of lithology on landslide occurrence in the given study area. In terms of the percentage of the area in ​​high and very high risk classes, both methods indicate almost similar results, so that it can be claimed that approximately 23% of the study area is in high and very high risk class, which are mainly in the western part of the area, Hashtjin. A set of conditions has made landslides very likely to occur in parts of the study area. Nevertheless, the dispersion of specific geological formations that provide very favorable conditions for slope instability in the form of mass destruction are of paramount importance. Most of the landslides in the area occurred on the lithological units of the old alluvial sediment units, the basaltic trachea unit and the worn surface andesite trachea unit. The findings from zoning and spatial prediction of both models show the zones corresponding to these units have a high probability of landslides. This poses a serious threat to the settlements and infrastructure located on these zones. The construction of communication roads on these landslidesensitive units has not only increased the risks of landslides, but also triggered slope instability. As a result, slope, slope direction and height play a decisive role in the occurrence of landslides on these lithological units.4ConclusionThe results of Table (No. 3) illustrate that about 6100 hectares, equivalent to 26% of the Hashtjin area is in high and very high class in terms of landslide risk. According to the results presented in Table (No. 4) based on the fuzzyneural model, it is revealed that about 4854 hectares, equivalent to 20.6% of the Hashtjin area is in high and very high classes in terms of slippery land. The results of the unique ROC analysis showed that exactly the fuzzyneural model is better and more acceptable than the logistic regression model.