High resolution flash flood forecasting by combining a hydrometeorological modeling system with a computational fluid dynamics model

Date

2024-4

Author

Turan, Azim

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

82
views

45
downloads

Cite This

Assessing the hydro-meteorological and hydrodynamic aspects together in flood disasters following excessive rainfall can provide valuable insights into the dynamics and consequences of the event. The originality of this study lies in integrating weather forecast models into flow simulations, enabling a continuous analysis of processes. For validation, a real flood event that occurred in the Bozkurt district of Kastamonu province, Turkey, on August 21, 2021, was selected as a case study. Grid rainfall data obtained from meteorological stations in the region were converted into runoff using WRF-WRF HYDRO codes. Subsequently, hydrodynamic analyses at the city scale were conducted using FLOW-3D HYDRO. These analyses examined the hydrodynamic behavior of overflowing streams and observed the effects of channel walls and bridges on flow. In the second step, rainfall forecasts from the WRF model were generated using different WRF physics parameterization options. 24 different WRF physics scenarios were applied to the topography for 24 different fine-scale scenarios using both WRF HYDRO and FLOW-3D HYDRO codes to obtain surface flows, and the results from WRF HYDRO and FLOW-3D HYDRO were compared. Then, surface flows were compared with surface flows obtained from observed rainfall. Thirdly, the results of a rainfall prediction model trained with both observed and hydro-meteorological model scenarios using the Random Forest method were compared. It was found that the rainfall prediction scenarios produced significantly divergent results, while the Random Forest model yielded results close to reality. Finally, flood analyses at the city scale were re-conducted using flood discharges obtained from the Random Forest model and FLOW-3D HYDRO. The analysis revealed that the flood extent areas obtained were very close to those obtained using observed rainfall data.

Subject Keywords

Hydrometeorology, Flood, Simulation, FLOW-3D, WRF, Random Forest

URI

https://hdl.handle.net/11511/109563

Collections

Graduate School of Natural and Applied Sciences, Thesis