Browsing by Author "Kadi, F"

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Determination of alternative forest road routes using produced landslide susceptibility maps: A case study of Tonya (Trabzon), Türkiye
    Kadi, F; Yilmaz, OS
    Firstly, Landslide Susceptibility Maps of the study area were produced using Frequency Ratio and Modified Information Value models. Nine factors were defined and the Landslide Inventory Map was used to produce these maps. In the Landslide Susceptibility Maps obtained from the Frequency Ratio and Modified Information Value models, the total percentages of high and very high-risk areas were calculated as 10% and 15%, respectively. To determine the accuracy of the produced Landslide Susceptibility Maps, the success and the prediction rates were calculated using the receiver operating curve. The success rates of the Frequency Ratio and Modified Information Value models were 82.1% and 83.4%, respectively, and the prediction rates were 79.7% and 80.9%. In the second part of the study, the risk situations of 125 km of forest roads were examined on the map obtained by combining the Landslide Susceptibility Maps. As a result of these investigations, it was found that 4.28% (5.4 km) of the forest roads are in very high areas and 4.27% (5.3 km) in areas with high landslide risk areas. In the last part of the study, as an alternative to forest roads with high and very high landslide risk, 9 new forest road routes with a total length of 5.77 km were produced by performing costpath analysis in with geographic information systems.
  • No Thumbnail Available
    Item
    Detection of temporal changes of treeless forest areas using remote sensing techniques and Google Earth Engine platform: A case study of Trabzon Duzkoy District
    Kadi, F; Yilmaz, OS
    This study aims to detect temporal changes in treeless forest areas using remote sensing techniques on the Google Earth Engine Platform. In this direction, ten treeless forest areas were determined from stand maps. A general study area was determined to include these areas, and the status of treeless forest areas was obtained by classifying the study area with a random forest algorithm on Sentinel-2 images. Then, normalized difference vegetation index (NDVI) time series analyses were performed on Landsat images to reveal vegetation changes in these treeless forest areas. In the classification study conducted with Sentinel-2 images, the land was divided into three classes: forest, treeless forest areas, and bare land. The overall accuracy of the classification study was calculated as 89.46%, and the Kappa value as 0.810. When the obtained treeless forest areas were compared with the areas on the stand map, it was seen that there was an average change of about 52.56% in terms of forest cover for ten regions. As a result of the analyses made with NDVI time series, it was observed that vegetation in treeless forest areas increased and therefore these areas tended to close.