Regional Integrated Meteorological Forecasting and Warning Model for Geological Hazards Based on Logistic Regression
Article
Figures
Metrics
Preview PDF
Reference
Related
Cited by
Materials
Abstract:
Regional Integrated Meteorological Forecasting and Warning Model for G eological Hazards is constructed based on Logistic regression of the probabiliti es of geological hazard occurrences. Dynamic forecasting of the probability of r ainfallinduced geological hazard occurrence is thus realized. Rainfall observa t ions and numerical rainfall forecasts are used as dynamical inputs of the model. Total information produced by information model that integrates factors of geos ciences is used as the static predictor of the model. The model forecasts probab ilities of geological hazard occurrences on a grid, and releases warning message s in the form of a 5level hierarchy. Validation of the model with observationa l data for the year 2004 shows that 80% of the geological hazards of the year hav e been identified as Warning Level 3 or above. The operation of the model during the rainy season of the year 2006 has made forecasts and warnings with high acc uracy for geological hazards triggered by Typhoon Bilis. All these show that the model can satisfy the requirements of warning operations, and is an effective w ay to improve the meteorological warnings for geological hazards.