Applicability of a nationwide flood forecasting system for Typhoon Hagibis 2019
Floods can be devastating in densely populated regions along rivers, so attaining a longer forecast
lead time with high accuracy is essential for protecting people and property. Although many
techniques are used to forecast foods, sufcient validation of the use of a forecast system for
operational alert purposes is lacking. In this study, we validated the fooding locations and times of
dike breaking that had occurred during Typhoon Hagibis, which caused severe fooding in Japan in
2019. To achieve the goal of the study, we combined a hydrodynamic model with statistical analysis
under forcing by a 39-h prediction of the Japan Meteorological Agency’s Meso-scale model Grid Point
Value (MSM-GPV) and obtained dike-break times for all fooded locations for validation. The results
showed that this method was accurate in predicting foods at 130 locations, approximately 91.6% of
the total of 142 fooded locations, with a lead time of approximately 32.75 h. In terms of precision,
these successfully predicted locations accounted for 24.0% of the total of 542 locations under a food
warning, and on average, the predicted food time was approximately 8.53 h earlier than a given
dike-break time. More warnings were issued for major rivers with severe fooding, indicating that the
system is sensitive to extreme food events and can issue warnings for rivers subject to high risk of
Ma, W.; et al.
You can also access this resource on our Portal
Thank you for recommending this resource.
Share your resources
Are you working to better understand and build community flood resilience? Others can benefit from your knowledge.