The importance of prehistoric data and variability of hazard regimes in natural hazard risk assessment - examples from Australia

摘要

Natural hazards are normally viewed as events that occur randomly over time. This precept usually forms the basis for the development of the hazard magnitude-recurrence interval relationship used in risk assessments. However, hazard variability doesnot always conform to this relationship especially over longer time intervals. Non-stationarity can be common with some hazards and those periods where the variability and/or mean (magnitude/frequency) remain constant are referred to here as hazard regimes. Shifts from one regime to another occur at a variety of time scales from centuries to millennia. Regime shifts are often only discernible by examining longer-term records which usually include prehistoric data. Risk assessments frequently ignore these regime shifts and estimates of the risks associated with tropical cyclones, tsunami, terrestrial floods and landslides in Australia have been both under-estimated and exaggerated when such assessments have been based solely upon short historical records. Examples of these regime shifts and their significance for natural hazard risk assessment are presented here.