Iranian Jornal of Watershed Management Science&Engineering

fa پهنه‌بندی خطر سیلاب با استفاده از مدل MaxEnt (مطالعه موردی: حوضه فاروب رومان، نیشابور) Flood hazard zoning using MaxEnt Model (Case Study: Faroub Roman watershed, Neyshabour) تخصصي Special پژوهشي Research امروزه یکی از ابزارهای پرکاربرد در مطالعات سیلاب، استفاده از سیستم اطلاعات جغرافیایی و داده‌های سنجش از دور است. هدف از تحقیق حاضر تعیین خطر‌پذیری مناطق مستعد سیل در حوزه آبخیز فاروب رومان با استفاده از الگوریتم آنتروپی بیشینه میباشد. حوزه آبخیز فاروب‌رومان با مساحت 40/18637 هکتار در شمال شرق ایران بین طول‌‌های شرقی ´49 ˚58 تا ´03 ˚59 و عرض‌های شمالی ´15 ˚36 تا  ´25 ˚36 واقع شده است. از معیارهای ارتفاع، زمین شناسی، کاربری اراضی، فاصله از رودخانه، شیب، جهت شیب، شاخص رطوبت توپوگرافی، بارندگی، بافت خاک، تراکم زهکشی و انحنای زمین برای پهنه بندی خطر سیلاب با استفاده از مدل MaxEnt استفاده شد. شاخص تفاضل استاندارد شده آب (NDWI) با استفاده از تصویر ماهواره لندست 8 در 23 آگوست 2015 محاسبه شد. مختصات 40 نقطه سیلابی با استفاده از GPS برداشت شد. از 28 نقطه (70 درصد) جهت مدلسازی و 12 نقطه (30 درصد) جهت اعتبارسنجی مدل استفاده شد. سطح زیر منحنی ROC برابر 945/0 بدست آمد که نشان‌دهندۀ پیش‌بینی عالی مدل MaxEnt  در پهنهبندی مناطق سیلابی حوزه فاروب رومان می‌باشد. متغیرهای تراکم زهکشی و طبقات ارتفاعی به ترتیب با 5/39 و 2/39 درصد بیشترین سهم مشارکت را در روند مدلسازی داشتهاند و فاکتور شاخص رطوبت توپوگرافی هیچ تأثیری در روند مدلسازی نداشت.  نتایج آزمون جکنایف نشان داد متغیرهای تراکم زهکشی، ارتفاع و ‌فاصله از رودخانه بیشترین اثر را در میزان AUC داشته است. نقشه پهنهبندی خطر سیلاب نشان داد بیشترین مساحت حوزه آبخیز فاروب‌رومان در طبقۀ بسیار کمخطر (67/88 درصد) و کمترین مساحت حوزه آبخیز در طبقه بسیار پرخطر (36/1 درصد) قرار دارد. مناطق بسیار پرخطر و پرخطر سیلابی در مسیر و اطراف رودخانه فاروب‌رومان بوده که به توجهات مدیریتی جهت پخش سیلاب و همچنین لایروبی و قطع درختان مزاحم و خطرساز در مسیرهای مستعد سیل دارد. Introduction Floods are one of the greatest challenges that affect the economy and people's welfare every year and pose a major threat to human lives and infrastructure worldwide. Climate change projections indicate that the frequency of floods may increase in the future.<ins cite="mailto:Ebrahim" datetime="2026-08-12T12:58"> </ins>Preparing flood zoning maps is a strategic measure and an important step in overall flood risk management to mitigate the impacts of floods.. The Faroub Roman River is one of the main rivers in Neyshabour County. Study of changes in the river's boundary and bed indicates that in past years, part of the Faroub Roman River bed has been allocated for the construction of university units. Current constructions have blocked the two main waterways of this river in Kal Jangal and Kal Nanwa, which exacerbates the risk of flooding in the spring. The aim of this study is to determine the factors affecting floods and prepare a flood hazard zoning map in the Faroub Roman Watershed<ins cite="mailto:Ebrahim" datetime="2026-08-12T12:56"> </ins>using satellite images and the Maximum Entropy. 11 criteria affecting flood occurrence, including elevation, geology, land use, distance from river, slope, aspect, topographic wetness index (TWI), rainfall, soil texture, drainage density, and land curvature (plan curvature and profile curvature) Materials and Methods: The Faroub Roman, with an area of 18,637.40 hectares, is located in the central part of Neyshabour County in northeastern Iran between eastern longitudes 58°49' to 59°03' and northern latitudes 36°15' to 36°25. In this study, 12 criteria affecting flood occurrence, including elevation, geology, land use, distance from river, slope, aspect, topographic wetness index (TWI), rainfall, soil texture, drainage density, and land curvature includes plan curvature and profile curvature, were used for flood hazard zonation using the Maximum Entropy Algorithm. Maps of elevation, slope, aspect, distance from river, drainage density, and land curvature were prepared using the Khorasan Razavi province DEM layer with 30 meters accuracy in Arc GIS 10.8 software. Geology, land use, and soil science maps of the Faroub Roman watershed were prepared from the 1: 100,000 scale maps of the Khorasan natural resources general administration after making corrections based on the Google Earth system and field visits. The precipitation zoning map was drawn using precipitation data from 6 meteorological stations over a 30-year period in ARCGIS 10.8 software using the IDW method. The Topographic Wetness Index (TWI) map was prepared using a simple equation and Digital Elevation Model (DEM) in ARCGIS software. In this study, a satellite image from the OLI sensor on the Landsat 8 satellite dated August 23, 2015, taken two days after the flood occurrence, was used to calculate the Normalized Difference Water Index (NDWI). Then, using the maximum likelihood algorithm and supervised classification in EVVI 5.3 software, a number of points (40 points) were selected as control points over the flood extent.. The ROC curve was used to evaluate the MaxEnt model. The area under the ROC curve is called AUC, which indicates the validity of the model. If its value is higher than 0.7, it indicates acceptable model performance. The importance of variables in flood zoning was assessed using the Jackknife. Results and Discussion After transferring the map of factors affecting flood occurrence and the coordinates of flood distribution points to the MaxEnt software, the maximum entropy algorithm was modeled and the factors affecting flood occurrence were identified. In the model execution process, 40 flood locations were used as the dependent variable, of which 28 flood locations (70%) were used to prepare the flood hazard map and 12 flood locations (30%) were used for model validation. The results indicated that the drainage density variable had the highest contribution to the modeling process at 39.5%, followed by elevation classes at 39.2%, land use at 4.6%, aspect at 4.3%, profile curvature at 4%, and distance from river at 3.2%.  As drainage density increases, water moves more rapidly through the waterways, reducing the opportunity for infiltration. Consequently, the instantaneous maximum discharge increases, leading to inundation. The runoff generated by precipitation in upstream areas collects at the lowest point of the basin; therefore, areas with lower elevation have the highest sensitivity to flood occurrence. Geology, plan curvature, slope, rainfall, and soil texture factors had a weaker influence on the modeling process, and the topographic wetness index (TWI) factor had no effect on the modeling process. Accordingly, the allocated share reflects the role of each factor in the modeling process rather than its relative importance. To investigate the importance of environmental variables in flood hazard zonation, the Jackknife test was performed. Based on the results of the Jackknife test, drainage density, elevation, and distance from river had the greatest effect on the AUC value, respectively. Following the spatial prediction of flood distribution in the region, a flood hazard zonation map was generated by classification within the GIS environment. According to the flood zonation map, the largest area of the Faroub Roman basin is in the very low risk class (88.67%) and the smallest area of the basin is in the very high risk class (1.36%). The maximum entropy evaluation results showed that the area under the ROC curve is 0.945, indicating excellent model prediction. Conclusion  The evaluation of the maximum entropy algorithm in flood hazard zonation in the Faroub Roman basin showed that the area under the ROC curve is 0.945, indicating excellent model prediction and its high capability in predicting and identifying areas prone to flood risk in the Faroub Roman watershed. The flood zonation map indicated that very high risk and high risk areas are located along the river path and the surrounding lands, respectively, where the land use has mostly been converted to orchards and agriculture. These high risk areas require management attention for flood distribution as well as dredging and removing hazardous vegetation in flood prone paths. Given that the universities of Neyshabour have been constructed on lands exposed to flood risk in the Faroub Roman riverbed, it is suggested to use diversion channels upstream to prevent flood hazards and their irreparable damages.   آنتروپی حداکثر, تراکم زهکشی, جک نایف, داده‌کاوی , نیشابور Data Mining, Drainage density, Jackknife, Maximum Entropy, Neyshabour 0 0 http://jwmsei.ir/browse.php?a_code=A-10-489-4&slc_lang=fa&sid=1 Negar Shirzadeh نگار شیرزاده negarshirzad0098@gmail.com No University of Torbat Heydarieh دانشگاه تربت حیدریه maryam azarakhshi مریم آذرخشی m.azarakhshi@torbath.ac.ir Yes University of Torbat Heydarieh دانشگاه تربت حیدریه Ali Golkarian علی گلکاریان golkarian@um.ac.ir No Ferdowsi University of Mashhad دانشگاه فردوسی Mohammad Reza Rahdari محمد رضا راهداری rahdarimr@torbath.ac.ir No University of Torbat Heydarieh دانشگاه تربت حیدریه