16:15 〜 16:30
[AHW29-04] Improvement of Rainfall-Runoff Modeling Using Detailed Slope Surface Flow Observation
キーワード：Rainfall-Runoff Modeling, Infiltration Process, Surface Flow Observation, Distributed Hydrological Model
Accurate flood prediction in small mountainous catchments requires accurately representing surface-to-subsurface flow interactions. Rainfall-runoff processes during flash floods in steep mountainous catchment like in Japan have been well reproduced by the lateral flow-dominated hillslope modeling based on kinematic wave-based distributed hydrological model. On the other hand, recent studies argue that vertical infiltration processes also play a significant role.
On these backgrounds, the study applied a kinematic wave-based distributed hydrological model 1K-DHM to the Shigaraki Experimental Catchment using high-resolution (5-meter) Digital Elevation Model (DEM) data and analyzed the model performance based on both river discharge as well as slope surface flow data taken by time-lapse cameras as a new type of spatial observation data. Model verification was performed using three flood events (2021, 2022, 2023) for seven different parameter sets. All parameter sets reproduced the peak flow and overall hydrograph and slope surface flow duration for the largest flood event in 2022, whereas peak flow and timing and flow attenuation were overestimated by most parameters. Two parameter sets showed better results. However, they overestimated flood volumes and surface flow duration, and parameter values are unreasonable, revealing the limitation of the present model structure.
To overcome the above-revealed model limitation, the study introduced vertical infiltration components by the Green Ampt model to resolve inconsistencies in surface flow representations. The improved model refined the peak discharge estimation while better represented surface flow duration and recession curves. Also, infiltration adjustments resulted in a balanced distribution between surface and subsurface runoff, reducing unrealistic surface flow persistence.
On these backgrounds, the study applied a kinematic wave-based distributed hydrological model 1K-DHM to the Shigaraki Experimental Catchment using high-resolution (5-meter) Digital Elevation Model (DEM) data and analyzed the model performance based on both river discharge as well as slope surface flow data taken by time-lapse cameras as a new type of spatial observation data. Model verification was performed using three flood events (2021, 2022, 2023) for seven different parameter sets. All parameter sets reproduced the peak flow and overall hydrograph and slope surface flow duration for the largest flood event in 2022, whereas peak flow and timing and flow attenuation were overestimated by most parameters. Two parameter sets showed better results. However, they overestimated flood volumes and surface flow duration, and parameter values are unreasonable, revealing the limitation of the present model structure.
To overcome the above-revealed model limitation, the study introduced vertical infiltration components by the Green Ampt model to resolve inconsistencies in surface flow representations. The improved model refined the peak discharge estimation while better represented surface flow duration and recession curves. Also, infiltration adjustments resulted in a balanced distribution between surface and subsurface runoff, reducing unrealistic surface flow persistence.