Assessing multi-risk through impact chains and spatial analysis: experience from the Marrakech-Safi region (Morocco)

Methodologies to assess risk from multiple hazards are increasingly being developed, and a growing body of literature reports implementations of multi-hazard risk assessments. However, fewer studies so far have attempted to complement a multiple hazard perspective with an equally sophisticated approach to assess the other subcomponents of risk, notably multi-vulnerability, in order to represent not just multi-hazard but rather multi-risk dynamics.

Using the impact chains approach, we have developed a participatory multi-risk assessment of the Marrakech-Safi region (Morocco). Floods and droughts, two of the most relevant hazards affecting the region, were considered for the study, with the aim of capturing their effects on diverse elements at risk, such as human security and rural livelihoods.

First, in order to identify the drivers of the most relevant impact-related risks associated with floods and droughts in the region, a set of four impact chains were co-developed with local experts and regional stakeholders during a dedicated workshop. Thereby, each type of risk was narrowly defined (i.e. “ risk of physical harm for the population due to floods”, “risk from loss of infrastructures and properties due to floods”, “risk of economic losses for rainfed agricultural systems due to drought”, “risk of economic losses for irrigated agricultural systems due to drought”), and the principal cause-effect connections between drivers were identified.

As a second step, each impact chain informed the spatial analysis of both single and multi-risk based on secondary data at the municipal level (n = 255 municipalities). The Standardized Precipitation Index (SPI) was used to characterize drought hazard for rainfed (SPI3) and irrigated (SPI12) farmlands, whereas a hydrological model was developed to simulate a 100-year return period flood. Exposure of people to floods was assessed using the WorldPop population distribution dataset, while a regional land use-land cover model was developed to assess exposure of irrigated and rainfed farmlands to drought. For each type of risk, weighted vulnerability indices were computed based on a set of social and environmental indicators, and combined to the hazard exposure assessments via a matrix approach to obtain single-risk classes. Ultimately, the four single-risk exposure and vulnerability scores were combined into a multi-exposure and multi-vulnerability score respectively, which were then used to obtain the final multi-risk classes.

Results show that the vast majority of municipalities in the region are affected by two risks or more, and that multi-vulnerability classes influenced importantly the final multi-risk assessment. The methodology allowed a complex representation of single- and multi-risk, integrating qualitative and quantitative outputs: impact chains proved to be useful at representing the inherent complexities of risk, while the spatial analysis helped to understand regional differences in multi-risk in all components of hazard, exposure and vulnerability. The results of the assessments are expected to support multi-sectoral planning at the regional level. However, further research is needed to understand how to manage the increase in complexity should more hazard and/or more risk typologies be considered, and how to best model the complex interactions emerging from the impact chains in a more dynamic way.