Flood Detection and Susceptibility Mapping Using Sentinel-1 Time Series, Alternating Decision Trees, and Bag-ADTree Models

摘要: Flooding is one of the most damaging natural hazards globally. During past three years, floods have claimed hundreds lives and millions dollars damage in Iran. In this study, we detected flood locations mapped areas susceptible to using time series satellite data analysis as well a new model bagging ensemble-based alternating decision trees, namely, bag-ADTree. We used Sentinel-1 for detection analysis. employed twelve conditioning parameters elevation, normalized difference’s vegetation index, slope, topographic wetness aspect, curvature, stream power lithology, drainage density, proximities river, soil type, rainfall mapping floods. ADTree bag-ADTree models were susceptibility mapping. software Sentinel application platform, Waikato Environment Knowledge Analysis, ArcGIS, Statistical Package Social Sciences preprocessing, processing, postprocessing data. extracted 199 flooded areas, which tested global positioning system ensure that correctly. Root mean square error, accuracy, area under ROC curve validate models. Findings showed root error was 0.31 0.3 techniques, respectively. More findings illustrated accuracy obtained 86.61 model, while it 85.44 method. Based on AUC, success prediction rates 0.736 0.786 algorithm, order, these proportions 0.714 0.784 ADTree. This study can be good source information crisis management area.