پیش‌بینی فرسایش خندقی با استفاده از داده‌های tandem-x و مدل حداکثرآنتروپی (مطالعه موردی: حوضه آبخیز خسویه)

فرسایش خندقی یکی از مخاطره‌آمیزترین انواع فرسایش آبی است که موجب تخریب اراضی و برهم خوردن تعادل منابع زیستی و محیط زیست می‌شود. در این مطالعه، به شناسایی عوامل موثر در فرسایش خندقی، پیش‌بینی و پهنه‌بندی فرسایش خندقی با استفاده از مدل حداکثر آنتروپی (maxent) در حوضه آبخیز خسویه در استان فارس پرداخته شد. ابتدا موقعیت خندق‌ها از طریق بررسی‌های میدانی، عکس‌های هوایی و با استفاده تصاویر نرم‌افزار گوگل ارث تهیه گردید و سپس در محیط نرم‌افزار gis لایه رقومی خندق‌ها نیز به‌صورت نقطه‌ای شد، در مرحله بعدی شاخص‌های فیزیوگرافی حوضه را از داده‌های tandem-x با قدرت تفکیک 12 متر در نرم‌افزار saga-gis تهیه شد. در این تحقیق با عملیات میدانی در منطقه موردمطالعه و نمونه‌برداری از خاک و انجام آزمایش نقشه بافت خاک در نرم‌افزار gis تهیه شد و نقشه کاربری اراضی و تراکم پوشش گیاهی با استفاده از تصاویر ماهواره‌های سنتینل و لندست تهیه گردید و سپس هر یک از شاخص‌ها به خندق‌ها در محیط gis اضافه شد. برای اجرای مدل حداکثر آنتروپی (maxent) از 70 درصد داده‌ها برای آموزش مدل و 30 درصد برای آزمون مدل استفاده شده است. در این مطالعه با استفاده از آزمون جک نایف میزان تاثیر هر یک از شاخص‌ها بر دیگر شاخص‌ها مشخص شده که در نهایت تاثیرگذارترین شاخص‌ها معرفی شد. جهت اعتبار سنجی مدل جهت پهنه‌بندی فرسایش خندقی منطقه موردمطالعه با استفاده از منحنی (roc) و مساحت زیر منحنی (auc) مورد ارزیابی قرار گرفت. نتایج حاصل از این پژوهش نشان داد که شاخص ارتفاع وکاربری اراضی پوشش گیاهی و عمق آبراهه تاثیرگذارترین شاخص‌ها در ایجاد فرسایش خندقی است و میزان auc=0/947 در سطح عالی است.

spatial prediction of gully erosion using tandem-x data and maximum entropy model (a case study: khasoyeh watershed, in southeast of fars province)

1- introduction soil erosion is a global problem that seriously threatens water and soil resources. among water, wind and glacial erosions, water erosion causes extensive damage every year around the world.water erosion is divided into four groups according to the importance and method of formation, including splash erosion, sheet, rill and gully erosion.one of the most important types of water erosion is gully erosion which is a common phenomenon in the climate of arid and semi-arid regions, which affects large areas with different morphological, soil and climatic conditions. in fact, this type of erosion is of special importance due to the production of sediment and soil loss, as well as extensive damage to land, roads and structures. therefore, to prevent the rapid growth of gullies or to minimize the damages caused by them, it is very important to know the morphology of a gully and how it is formed and the reasons for its growth. in fact, by identifying areas prone to gully erosion and identifying the factors affecting it, the risk of this erosion can be reduced. a large part of the southeastern regions of fars province due to special topographic and geological conditions, have been affected by this type of water erosion. the main purpose of this study is to prepare a gully erosion susceptibility map using satellite data and using maxent model to evaluate the ability of this model to predict gully erosion prone areas south of darab city in fars province and also to investigate the impact of each factor in this is the type of soil loss.2- methodologythe study area is located in the khasoyeh watershed, in southeast of fars province in the zagros mountains (zm) (54, 9’ till 54, 42’ e and 28, 18’-28, 39’ n) and covering ca. 139248ha. in this research, we followed the subsequently mentioned working steps to predict the gully erosion susceptible:in the first step the locations of characteristic gullies were digitized using google earth (ge) images, aerial photos and fieldwork. subsequently, we converted the polygons into equally spaced points, which are congruent with the raster cell centroids.ii) in the second step we prepared the predictor variables that are driving the erosion processes. these continuous predictor variables include; physiographic indices including, elevation, topograpghic wettness index (twi), vertical distance to streams, slope, slope direction, hillshade, ls factor, stream power index (spi), flow concentration with spatial resolution of 12 meters were prepared in saga gis. before applying the indices, the dem was preprocessed with low pass filtering to extract artefacts and errors, like local noise and terraces using arcgis 10.8 (esri, 2010). then, the dem was hydrologically corrected eliminating sinks using the algorithm proposed by planchon & darboux (2001). the other applied indeice in this research are including, geological maps, land use, morphology, soil texture, were prepare. in the last stage, after preparing all the effective indices (indepence layers) in the gully erosion of the region, the maximum entropy model was implemented.3- results the results of the applied model showed that the elevation index is the most important factor. in the other words, the result indcated that there is an inverse relationship between the elevation and prone areas to gully erosion because the most gullies area are located at a minimum elevation of 900 meters. the reason for this can be reduced soil depth in high areas. the results of the modeling showed that more than 80% of the area is in the very low risk category and most of these areas are located in the northern parts and highlands. the high-risk areasare mostly located in flat and low elevation areas in southeast of the study area. the roc curves in this study was equal to 0/94 that shows the applied model is considered well distinguish different areas of gully erosion prediction.4- discussion & conclusionsgully erosion is common geomorphological problems in arid and semi-arid regions; therefore, it is essential to develop methods to predict it with simple but highly accurate models. in this research, the spatial prediction of gully erosion in the khasyoue watershed in fars province was carried out by means of a stochastic model and a detailed terrain analysis as well as additional environmental variables. the validation of the maxent model shows that the model prediction for gully is outstanding. percentage of the important variables for prediction of gully erosion indicates that the most influential parameters on gully erosion in the study area are espectively elevation (57.9%), land use (11.6%), vegetation density (9.8%) respectively and the least effective parameters are the direction of flow and catchment area with (0.1%) slope and vertical distance to streams with (0.2%). the simplicity and the relative low data requirements of the applied method allows an effective method also in other regions.