Heyelan Risk Haritalaması Üzerine Yarı Sayısal Bir Değerlendirme
Tetik Biçer, Çiğdem
xmlui.mirage2.itemSummaryView.MetaDataShow full item record
Following the earthquakes, landslides, which correspond serious obstacle to sustainable development with the negative consequences, are the second destructive type of disaster causing the highest loss of lives and properties in Turkey. The most effective work struggling with the negative effects of landslides can be achieved by the dissemination of landslide inventory, susceptibility, hazard and risk studies created by reliable and healthy database on landslides. In this study, it was aimed at assessing the landslide risk at the cenral Kahramanmaraş, and the parameters and methodologies were selected on the scientific landslide literature basis. According to the landslide literature between 1990 and 2016, the most commonly used parameters considered in these studies were appeared as slope, lithology, aspect, topographical elevation, drainage characteristics, land use, curvature, distance to tectonic structures and NDVI (Normalized Difference Vegetation Index), while the most frequently used analyses methods were determined as LR (Logistic Regression), FR (Frequency Ratio) and AHP (Analythical Hierarchy Process) based on the most actual literature including the period between 2015 and 2016. Firstly, landslide susceptibility maps were produced by 9 prepatory parameters and 3 different analysis methods for the Central Kahramanmaraş. Performances of the so produced maps were evaluated by ROC (Relative Operating Characteristics) method, and the AUC (Area Under Curve) values were calculated 0.828 for LR, 0.862 for FR and 0.793 for AHP, respectively. In the process of producing the landslide hazard map for the study area, the susceptibility map produced by the FR method was used. Rainfall data were used as the triggering parameter during the hazard assessment, and were analyzed by Gumbel distribution. Of the 214 mapped landslides in the study area, 22 landslides were exactly dated and their cumulative rainfall values were calculated. In this study, some new approaches related to the rainfall analysis and vulnerability assessments have been put forward, and can give rise to increase its originality. Of these, as a new approach, based on these calculated values, considering the rainfall intervals, the probabilitiy of the rainfall occurrences were determined for 2, 5, 10, 25, 50 and 100 years. For the vulnerability analysis, a relative evaluation was performed for risk assessments. All the findings were combined at the final stage to calculate the landslide risk of the study area. The calculated risk value corresponds to the specific risk, and it was determined that most of the centrum had landslide risk. For this reason, it is suggested that the produced maps should be taken into consideration by decision makers and local administrations, and should be used in possible future works.
Showing items related by title, author, creator and subject.
Baştuğ, Gizem (Fen Bilimleri Enstitüsü, 2018)Landslides occurring as a consequence of conurbation related to increasing population are - in addition to geological, topographical and climatic conditions - causing serious amounts of life and property loss worldwide. ...
Heyelan Duyarlılık Haritalarının Üretilmesinde Örneklem ve Doğrulama Stratejilerinin Değerlendirilmesi (Gelibolu Yarımadası'nın Doğu Kesimi) Dağdelenler, Gülseren (Fen Bilimleri Enstitüsü, 2013)The purposes of this study are to evaluate the sampling strategies used in landslide susceptibility mapping studies in the literature and to assess validation indices. In addition, validation indices of the models created ...
Yapay Sinir Ağları ile Heyelan Duyarlılık Haritalarının Üretilmesinde Farklı Algoritmaların Kullanımının Araştırılması Can, Aslı (Fen Bilimleri Enstitüsü, 2014)This study aims to investigate the effects and performances of the uses of different learning algorithms in the training of artificial neural networks on the resulting maps by applying the artificial neural networks (ANN) ...