Download PDFOpen PDF in browser

Ensemble Numerical Modeling Approach with Social Network Information to Optimize Flood Forecasting

13 pagesPublished: September 20, 2018

Abstract

The rivers in Taiwan are steep, the surface runoff is rushed into ocean quickly with high speeds and large discharges. When the typhoons hit Taiwan with heavy rain, how to predict correct peak time and peak stage of rivers is the most important aim in this research. Taiwan Typhoon and Flood Research Institute will produce a rainfall forecasting every six hours for disaster warning, according to different physical parameters setting. The research site, Xiuguluan River is steepest one of Taiwan central rivers. By cross section data、land use、slope、soils and the rainfall forecasting, we can get results of each member by integrating the physically based on model HEC-HMS and WASH123D.
The research reveals that ensemble numerical modeling can predict precise peak stage of the river by analysis and correction by machine learning system TensorFlow. As for peak time forecasting, it becomes accurate by making use of the open social network information such as facebook、network news、PTT discussion to improve. Moreover, no matter peak time or peak stage, it has highly variation in members. In other words, no member is always the best of typhoons. But we can use the probability flood forecasting to predict and get the best results.

Keyphrases: Flood forecasting, social network information, Taiwan, TensorFlow, Typhoon, WASH123D, Xiuguluan River

In: Goffredo La Loggia, Gabriele Freni, Valeria Puleo and Mauro De Marchis (editors). HIC 2018. 13th International Conference on Hydroinformatics, vol 3, pages 1201--1213

Links:
BibTeX entry
@inproceedings{HIC2018:Ensemble_Numerical_Modeling_Approach,
  author    = {Pin-Hao Liao and Dong-Sin Shih},
  title     = {Ensemble Numerical Modeling Approach with Social Network Information to Optimize Flood Forecasting},
  booktitle = {HIC 2018. 13th International Conference on Hydroinformatics},
  editor    = {Goffredo La Loggia and Gabriele Freni and Valeria Puleo and Mauro De Marchis},
  series    = {EPiC Series in Engineering},
  volume    = {3},
  pages     = {1201--1213},
  year      = {2018},
  publisher = {EasyChair},
  bibsource = {EasyChair, https://easychair.org},
  issn      = {2516-2330},
  url       = {https://easychair.org/publications/paper/zGMZ},
  doi       = {10.29007/7crq}}
Download PDFOpen PDF in browser