Building a ensemble flood forecasting system using a numerical weather prediction data

DOI

Bibliographic Information

Other Title
  • 数値気象予報データを用いたアンサンブル洪水予測システムの構築

Abstract

Numerical Flood Prediction is a method to predict future state of rivers by using numerical models and forecast values as input. This enables predictions to have a much longer lead time. However, due to uncertainty caused by input values and model parameterization, ensemble simulation is suggested.<br>  In this study, a regional ensemble flood prediction system considering precipitation input uncertainty was built and hindcasted a real flood event. The spatial and temporal reslutions are 0.05 degrees and 1 hour, respectively.   Verification experiment forced by Radar precipitation reproduced annual discharge and each peaks well. In the hindcast for Kinu-River flood due to Kanto-Tohoku heavy rain, the system simulated the occurrence of a flood 35 hours before. The number of members which predicted the flood increases with initial forecast time, where 43 out of 51 members predicted floods 11 hours before the actual flood while only 17 did so 35 hours before.

Journal

Keywords

Details 詳細情報について

  • CRID
    1390001205712460928
  • NII Article ID
    130005176103
  • DOI
    10.11520/jshwr.29.0_36
  • Text Lang
    ja
  • Data Source
    • JaLC
    • CiNii Articles
  • Abstract License Flag
    Disallowed

Report a problem

Back to top