Real-time slope water table forecasting by multi-tank model combined with dual ensemble Kalman filter

Jun Xiong, Tomofumi Koyama, Satoshi Nisiyama, Yuzo Ohnishi, Kenji Takahashi

Research output: Contribution to journalArticlepeer-review

2 Citations (Scopus)

Abstract

The multi-tank model is a helpful tool for rainfall-groundwater-runoff analysis since it can represent a nonlinear transport behavior and give predictions very quickly. On one hand, because of its simple and quick calculation, the tank model is useful in water table or stream flow prediction; on the other hand, by reason of its simplicity, its model error is relatively big, the model adjustment through the time variation is necessary. A new data assimilation methodology for online tuning of model parameters and states in the multi-tank model by taking into account new observations was developed, and important model parameters and states are retrieved simultaneously based on dual ensemble Kalman filter. A description of applying multi-tank model to simulate rainfall infiltration process in the slope is presented, and also the implementation of the dual ensemble Kalman filter approach for slope water table forecasting is demonstrated. Based on the observation data, the performance of the new method is studied. Finally, the factor of slope stability could be predicted based on filtered water tables. In this way, an early warning system of slope stability during rainfall could be built, which is helpful for disaster prevention.

Original languageEnglish
Pages (from-to)81-89
Number of pages9
JournalLandslides
Volume8
Issue number1
DOIs
Publication statusPublished - Mar 2011
Externally publishedYes

Keywords

  • Data assimilation
  • Dual ensemble Kalman filter
  • Groundwater table
  • Multi-tank model
  • Optimization
  • Rainfall
  • Slope stability

ASJC Scopus subject areas

  • Geotechnical Engineering and Engineering Geology

Fingerprint

Dive into the research topics of 'Real-time slope water table forecasting by multi-tank model combined with dual ensemble Kalman filter'. Together they form a unique fingerprint.

Cite this