[A-HW21] Surface and subsurface hydrologic models: Technical advances and applications for water management

[AHW21-08] Introduce social media data to hydraulic model for flood hazard simulation in Mobara city, Japan

*Mao Ouyang¹, Shunji Kotsuki¹, Yuka Ito², Tomochika Tokunaga² (1.Chiba University, 2.The University of Tokyo)

Keywords:Social media data, Hydraulic model, Flood

Intensive rainfall-induced fluvial floods have caused catastrophic repercussions in many urban areas. One of the major challenges in urban flood simulations is lack of field observations, such as spatial and temporal changes of inundation areas and their water depths, to evaluate the performance of numerical models. This study implemented the hydraulic model, consisting of both surface flow and soil infiltration, to investigate the flood events in Mobara city, Japan. The hydraulic model was calibrated by the flood on July 1, 1970, and then validated against the flood on October 25, 2019. Simulated flood extent and temporal changes of water depths agreed well with the observations from social media data, suggesting that the hydraulic model was appropriate for the flood simulation in the study area. Comparison of the two flood events indicated that the areal extent of the flood was greatly expanded in around 50 years, which was responsible for the changes of land use/land covers, topographies, and input river water levels. Sensitivity experiments were conducted by keeping all but one parameter in the hydraulic model to disentangle their influences on flood extents and volumes. Land use/land cover change caused the change in the flood extent to around 23 %. Change in topographies showed around 12 % change in flood extent and 39 % change in flood volume. The change of flood volume was quite sensitive to the input river water level, e.g., 58 % of flood volume decreased due to around 1.2 m decrease of input river water level. Our study highlights the possibility of employing hydraulic model and social media data to reduce the uncertainty in flood simulation, and the necessity of considering the temporal changes of land use/land covers, topographies, and input river water levels for flood mapping in urban cities. This presentation will include our recent research on the parameter estimation of geophysical models.
Reference: Mao Ouyang, Shunji Kotsuki, Yuka Ito, Tomochika Tokunaga, Employment of hydraulic model and social media data for flood hazard assessment in an urban city, Journal of Hydrology: Regional Studies, 44 (101261), 2022, doi: 10.1016/j.ejrh.2022.101261.