Under extreme rainfall conditions, multiple hazardous processes such as landslides, debris flows and flooding may occur simultaneously or sequentially, posing high risks to human lives and properties. So far, the study on multi-hazard analysis and hazard interactions under extreme storms has been rather limited, particularly in the urban environment.

The principal objectives of this thesis research are to identify spatial characteristics of severe rainstorms in Hong Kong; to simulate shallow slope failures, flash floods, and debris flows in urban settings under extreme storms using a cell-based approach; and to investigate hazard interactions and their cascading effects.

A Hong Kong Storms and Landslides Database (HKSlid) has been constructed based on historical records of rainfall,...[ Read more ]

Under extreme rainfall conditions, multiple hazardous processes such as landslides, debris flows and flooding may occur simultaneously or sequentially, posing high risks to human lives and properties. So far, the study on multi-hazard analysis and hazard interactions under extreme storms has been rather limited, particularly in the urban environment.

The principal objectives of this thesis research are to identify spatial characteristics of severe rainstorms in Hong Kong; to simulate shallow slope failures, flash floods, and debris flows in urban settings under extreme storms using a cell-based approach; and to investigate hazard interactions and their cascading effects.

A Hong Kong Storms and Landslides Database (HKSlid) has been constructed based on historical records of rainfall, landslides, affected buildings and casualties. A strong positive correlation exists between the rainfall intensity and the occurrence frequency or magnitude of the landslides. Nearly one thirds of the landslides can transform into debris flows. The slope safety system has proven to be effective under normal weather conditions, which is indicated by a sharp decrease in affected buildings and casualties since 2000.

The spatial correlation characteristics of three large storms in Hong Kong are quantified in terms of the maximum rolling 4-h, 24-h and 36-h rainfall amounts using geostatistical methods. The rainfall amounts in the three large storms are observed to be strongly correlated spatially within 5 km and likely to be correlated within 25 km.

The potential hazards from slope failures under extreme storms are predicted using a physically-based method. The method is validated using landslides in two past heavy storms. The landslide scenarios on western Hong Kong Island under three extreme storm events are predicted using the model. As storms become more severe, acute slope-failure consequences may occur. Under extreme storms of 44%, 65% and 85% of the 24-h probable maximum precipitation (PMP), 0.5-1.7% of the natural terrain on western Hong Kong Island may lose stability.

The rainfall-runoff processes associated with drainage networks at different times are simulated to detect the flooding-hazard-prone area on Hong Kong Island. The model is verified using the historical scenario in 2008. The verified model is applied to evaluate the flooding consequences including flood intensity, affected area and so on under different rainfall scenarios. The urban area will face serious flooding if severe rainstorms occur. The maximum flow velocity can reach a hazardous level as a result of steep slopes in the study area.

A computational scheme is developed for simulating possible scenarios of urban debris flows considering building blockage effects and bed erosion. The scheme is evaluated using two historical debris flow cases on Hong Kong Island and used to predict probable future scenarios. The multiple debris flows are simulated and the source materials are determined based on the simulated volumes of loose materials from slope failures. The presence of densely populated buildings and the bed erosion enlarge the debris-flow intensity. Large impact pressures develop on the building facades that face the main path of the debris flow. These stress-concentration locations correspond to the largest perpendicular velocities. The less space left for the debris flow, the larger the impact pressures on the buildings are. Deposition in front of buildings increases the impact pressure.

The simulation results of slope failures, flooding and debris flows under different storm scenarios are integrated to detect potential hazard interactions. The cascading effects of debris flows on flooding are simulated. The erosion volumes of multiple debris flows increase the flow volumes significantly. The erosion and deposition processes can change the topography significantly, which in turn affects the flooding or other hazard process.

The inundated areas in the low-lying areas increase if the channel morphology is changed by debris flows, though the flow depths in the mid-mountain areas are not significantly affected. The patterns, values and directions of the flow velocity are also changed. More buildings and streets are affected by larger and faster debris flows. The interactions among the debris flows and flooding should not be neglected, especially when severe rainstorms occur. The maximum flow depths in the downstream low-lying areas are enlarged if drainage inlets are blocked by debris flows. The areas next to the target catchment are also affected because more water is diverted into the neighbouring catchment.

It is evident that multi-hazards and their interactions will significantly increase the slope risks. Their cascading effects are affected by the magnitudes of rainfall and other factors. Integrating the single hazards without considering their interactions tends to underestimate the impacts of these hazards.