ارزیابی و پیش بینی فرسایش شیاری با استفاده از روش داده کاوی و داده های سنجش از دور (منطقه مورد مطالعه: حوضه آبخیز خسویه در استان فارس) (مقاله علمی وزارت علوم)
درجه علمی: نشریه علمی (وزارت علوم)
آرشیو
چکیده
از میان انواع مختلف اشکال فرسایش آبی، فرسایش شیاری یکی از مهمترین عوامل هدر رفت خاک است. کشور ایران نیز با توجه به شرایط اقلیمی و توپوگرافی آن، این نوع از فرسایش خاک به صورت وسیعی خصوصا در مناطق کوهستانی و شیبدار گسترش یافته است. در این پژوهش، به شناسایی عوامل موثر در فرسایش شیاری، پیش بینی و پهنه بندی آن با استفاده از مدل حداکثر آنتروپی (Maxent) در حوضه آبخیز خسویه در استان فارس پرداخته شده است. در ابتدا موقعیت شیارها را با استفاده تصاویر ماهوره ای، عکس های هوایی و بازدیدهای میدانی تهیه گردید و سپس در محیط نرم افزار GIS لایه رقومی شیارها تهیه گردید، در مرحله بعدی شاخص های فیزیوگرافی با استفاده از داده های Tan DEM-X با دقت تفکیک 12.5 متر در نرم افزارSAGA-GIS تهیه گردید. در این تحقیق با عملیات میدانی در منطقه مورد مطالعه و نمونه برداری از خاک نقشه بافت خاک تهیه گردید. لایه نقشه های کاربری اراضی و تراکم پوشش گیاهی نیز با استفاده از تصاویر ماهواره های سنتینل-2 و لندست تهیه گردید. جهت اجرای این مدل، از 70 درصد داده ها برای آموزش و 30 درصد برای تست مدل استفاده شده گردید، که در نهایت تاثیر گذارترین شاخص ها مشخص شدند. جهت اعتبار سنجی مدل جهت پهنه بندی فرسایش شیاری منطقه مورد مطالعه با استفاده از منحنی ROC و مساحت زیر منحنی (AUC) مورد ارزیابی قرار گرفت. نتایج حاصل از این پژوهش نشان داد که شاخص شیب، کاربری اراضی و فاصله از ابراهه تاثیر گذارترین شاخص ها در ایجاد فرسایش شیاری می باشندAssessment and Prediction of Rill erosion by using data Mining and remote sensing data (A case study: Khasoyuh watershed in Fars Provinceو, Iran)
Water erosion is one of the most type of land degradation and desertification in many different climates in the world. Among the different forms of water erosion, rill erosion is one of the most important. This type of water erosion can cause one of the main source of sediments. Our study area is located in Fars province in South of Iran that is effected by difference type of soil loss, same as gully, rill and sheet erosion. For prediction of rill erosion in this research we first digitized the points of rill areas with using GE, filed survey and aerial photos. Different index has been appalled as independent variable, same as, topographic indices, LULC, climate, soil. The topographic Indices have been generated from Tan-DEM-X with 12.5 m in SAGA-GIS. In the next step, after preparing all variables, we have use the Maxent model for prediction of the rill erosion in our study area. In this study we have divided the data of rill erosion points into two categories, its meaning 30% for testing and 70% of the data to training the model. The model used was finally introduced the most effective indicators. In order to validate the model for the zonation of rill erosion, the studied area was evaluated using (AUC) and the area under the curve (ROC) of the curve. The results of this research showed that the slope index and land use and distance from the rivers are the most effective indicators in creating rill erosion. Key words: Rill erosion, Maxent model, zoning, Khasoyuh watershed, GIS Assessment and Prediction of Rill erosion by using data Mining and GIS (A case study: Khasoyuh watershed in Fars Province, Iran)IntroductionSoil erosion is a serious threat to human well-being and life. According to the importance and the way of formation, water erosion is divided into water erosion, Rill erosion includes about 70 to 50% of four groups including erosion, watercourses, Rill and all soil erosion. In many researches, this type of erosion The most important process of sediment production and as a result soil loss has been reported, Rill erosion as the initial stage of hydrological erosion is the importance of water in sloping fields, but it also causes the loss of surface It has many. On the one hand, furrows not only cause significant loss of soil soil nutrients and as a result causes land degradation and pollution of downstream rivers.MethodologyFirst, the location of the Rill through field surveys and satellite images and aerial photos using physiographic maps software (height, slope, slope direction, slope length, relative slope, vertical distance to the network channel, topographic humidity, convergence, length Curvature, flow power, cumulative flow, distance from the waterway, channel depth) were prepared using DEM height digital model data with a resolution of 30 meters in (GIS-SAGA) software Vegetation and land use parameters using Landsat 8 and Sentinel 2 satellite images for 2021 It was prepared in the software (GIS Arc 10.8), the map of soil and geological indicators is also in the software 10.8 ArcGIS software was prepared, after preparing the map of the effective indicators of the model, it was implemented. The basis of the selection of factors It has been investigated in various studies of other researchers about Rill erosion in different parts of the world. In this study, the maximum entropy model is used, which is one of the machine learning technologies (MLTs) that has a high spatial prediction ability in the field. has various environmental sciences. also to check the validity of the final model, the area under the curve index (AUC) for the ROC curve was used. If the level below be. If the curve is more than 0.9, the recognition power of the model is considered excellent.Results In this research, to determine the impact of each variable on the amount of Rill erosion in the region, using the method of percentage participation of each variable It has been used using Maxnet output. The result of this research shows the slope index is one of the most important variable for event of rill erosion in our study area. In the other hand, the area with high slope has more potential for this type of soil loss. The second more important variable is land use land cover in the study area, because area with low vegetation are more prone for eroding of soil by water. In total we can mention the topography condition is the most important criteria for the rill erosion in this watershed.Dissection:The result of the area under the curve (AUC) in this study was equal to 0.947, which indicates that the prediction of the model in zoning the furrow erosion areas is very high, considering the four influencing factors on furrow erosion. including the slope, land use, vegetation and distance from the waterway, the model analysis in this research showed that the slope index has the most impact and importance, and there is an inverse relationship between the slope and areas prone to gully erosion, the lower the slope, the level of gully erosion in the area more than.
تبلیغات
