This theme lies at the heart of the project. As part of the production process, it deals with the challenges of translating probabilistic forecasts of physical hazards into impact-based risk assessments, taking account of hazard generation in human systems, the exposure of individuals and communities to the hazard, and the vulnerability of those affected. Conversely, as part of the specification of the need for warning services, it identifies the role that weather forecasts might play in enabling individuals and communities to minimise hazard impacts. Quantitative models that relate the occurrence, severity, duration or other important aspects of the physical hazard to consequences of significance for human systems need to be developed, evaluated, peer reviewed and shared. 

Hazards result from the interaction of weather with other natural phenomena and human systems. Often the damage is caused by interactions between weather and the built environment created by humans – e.g. in windstorms it is often loose building materials, while in floods it may be engineering failures. 

Physical hazards may affect a variety of social, health or economic systems through death and injury; physical and mental illness; damage to buildings or infrastructure; contamination of land; or disruption to business, government and welfare services. Many of these effects are highly non-linear with poorly understood thresholds beyond which impacts may be irreversible. In particular, the transition from the normal response of distress to the abnormal one of physical and mental illness and its consequent loss of economic productivity and cost of treatment and care, is poorly understood and needs to be better studied. 

The impact of a hazard depends on the extent to which individuals, businesses, communities & infrastructure are exposed to it, i.e. how likely is the hazard to impact on the receptor? This depends on the relation of receptors to the hazard, e.g. what traffic is using a road during a severe wind storm, how many aircraft are flying a route that is affected by thunderstorms, or which utilities are affected by flooding. Access to current socio-economic data is a key requirement to enable exposure models to be developed, evaluated and shared between academic institutes.

The impact of a hazard also depends on the vulnerability of exposed individuals, businesses, communities and infrastructure, i.e. how sensitive is the receptor to the impact? This depends on how the receptor responds, what reserves they have to call on, and how well prepared they are. Vulnerability data and models depend on a detailed understanding of the cultures of groups of people within the population. Generic relations between the characteristics of such groups and their behaviours can provide first-order information, but they are usually not directly applicable in specific high impact weather event situations. However, detailed studies in representative locations, synthesised across contexts, can enable the development and documentation of best practice that can be applied by National Meteorological and Hydrological Services (NMHSs) and disaster reduction agencies. Differing responses to different types of hazard need to be identified, in particular between slow onset and sudden onset impacts.

Expertise in the science required for this theme is widely disbursed in academia and in the user domain, and bringing these diverse communities together will be a challenge. This challenge is common to the climate and seasonal forecasting endeavours, where the emphasis is on policy rather than management responses. Links with the Climate Impacts community in WCRP and beyond will be established to ensure that expertise gained here is made available to be used in these broader contexts.