Development of a semi-dynamic flood prediction model based on geographic information systems

Disasters are among the major events that have had the greatest destructive impact on humanity throughout the ages. Increasing population throughout the world pushes people to urbanization, and increasing population directly affects the number of people affected by disasters. The most frequent disasters worldwide are those of meteorological origin. According to reports, while 370 natural disasters occurred between 2002 and 2021, 387 natural disasters were encountered in 2022, resulting in thousands of casualties, millions of people affected and billions of liras of material losses. When the last 20 years are analyzed, floods are in the first place in terms of both the number of occurrences and the number of people affected. Rapidly increasing urbanization causes urban areas to be most affected by floods. An effective urban and disaster management is inevitable to minimize the effects of disasters. Technological developments, information systems and smart city applications, which are the products of these developments, stand out as the most important tool in the effective management of disasters. Within the scope of this thesis, a model that predicts flood risk semi-dynamically using the data available through smart cities and technological developments has been developed. There are many factors that are effective in flood events, these factors were identified through literature research and expert opinions and their weights were calculated with two different Multi-Criteria Decision Making methods, Analytic Hierarchy Process and Hesitant Fuzzy Linguistic Term Sets methods. Depending on continuously updated meteorological forecast data and a set of static criteria, "semi-dynamic" flood-flood risk data were produced with various Geographic Information Systems analyzes. Depending on the data produced based on static criteria and the estimated hourly precipitation and humidity variables, determinations have been made as to which regions may be at risk. According to the flood prediction maps with 100m resolution by evaluating static and dynamic data together; the first and highest risk areas in the study area were determined as the parts of Atakum and Canik districts close to the coast. The results obtained were confirmed by previous flood events. It is aimed to support decision makers by identifying the areas where precautions should be taken with the 3D city model of the region expected to be affected.