Using Landslide Movement to Differentiate Landslide Hazard and Risk

Landslide-time-series records (landslide catalogue) are necessary to determine the frequency component of landslide magnitude/frequency relationships required to calculate the probability of a landslide occurring (the landslide hazard). Fast landslides, defined as landslides that have a total cumulative movement of > 1.0 m in 24 hours, often have a single movement episode and get a single entry in a landslide catalogue. Slow landslides, defined as cumulative displacement < 0.1 m in 24 hours, often have multiple movement episodes through time and this can result in multiple entries in a landslide catalogue. If a landslide catalogue combining both fast and slow landslide movements is analysed to determine the frequency of landsliding, the resulting landslide frequency will be dominated by the frequency of the small episodic movements of slow landslides. As landslide frequency is required to derive landslide hazard, this can lead to an inappropriate determination of landslide hazard. Slow landslides are usually pre-existing landslides and it is often possible to identify them using geomorphic criteria, prior to movement being observed. A map or inventory of pre-existing landslides can be used to identify potential slow landslides as slow landslides will be a subset of the landslide inventory population (some inventory landslides are the ‘preserved’ remains of fast landslides). Slow landslide movement is site specific and information on individual landslide behaviour over time is required to determine the probability of movement occurring. For slow landslides, the landslide hazard is the probability of movement occurring at a specific site. Fast landslides are usually first-time landslides and the specific sites at which they may occur are seldom identifiable prior to failure. The probability of fast landslide movement is potentially treatable using probabilistic techniques to spatially differentiate landslide hazard. For fast landslides, calculating the landslide hazard requires determining the probability of landslide movement occurring anywhere. Differentiation of landslides based on movement parameters recognises that different datasets are required to determine the hazard (or the probability of movement occurring) for fast and slow landslides. A landslide catalogue is a required input for determining -fast-landslide hazard, while a landslide inventory can be used to identify potential slow landslides which then must be examined to determine their hazard or probability of movement. Differences in the frequency of landslide movement for fast and slow landslides differentiate a landslide population into two subsets with different risk profiles. Landslides causing fatalities in New Zealand (over 350 deaths from 50 individual landslides) have all been fast landslides. Fast landslides are a risk to both life (life safety) and infrastructure (dollar cost). Small episodic movements of slow landslides are not a direct life-safety risky (no such deaths or injuries are recorded in New Zealand), although they often damage infrastructure (dollar cost).