This study presents a conceptual framework for addressing temporal variation in natural risk. Numerous former natural risk analyses and investigations have demonstrated that time and related changes have a crucial influence on risk. For natural hazards, time becomes a factor for a number of reasons. Using the example of landslides to illustrate this point, it is shown that: 1. landslide history is important in determining probability of occurrence, 2. the significance of catchment variables in explaining landslide susceptibility is dependent on the time scale chosen, 3. the observer's perception of the geosystem's state changes with different time spans, and 4. the system's sensitivity varies with time. Natural hazards are not isolated events but complex features that are connected with the social system. Similarly, elements at risk and their vulnerability are highly dynamic through time, an aspect that is not sufficiently acknowledged in research. Since natural risk is an amalgam of hazard and vulnerability, its temporal behaviour has to be considered as well. Identifying these changes and their underlying processes contributes to a better understanding of natural risk today and in the future. However, no dynamic models for natural risks are currently available. Dynamic behaviour of factors affecting risk is likely to create increasing connectivity and complexity. This demands a broad approach to natural risk, since the concept of risk encapsulates aspects of many disciplines and has suffered from single-discipline approaches in the past. In New Zealand, dramatic environmental and social change has occurred in a relatively short period of time, graphically demonstrating the temporal variability of the geosystem and the social system. To understand these changes and subsequent interactions between both systems, a holistic perspective is needed. This contribution reviews available frameworks, demonstrates the need for further concepts, and gives research perspectives on a New Zealand example.