Temporal Dynamics of Avalanche-Related Territorial Isolation Risk: Application to a Mountain Road Network in the French Alps

Road networks are frequently affected by natural hazards, resulting in disruptions and, in some cases, situations of territorial isolation with a range of direct and indirect consequences. Sparse or tree-like road networks, which are characteristic of mountainous areas, limit alternative routes and thereby increase the risk of isolation. The loss of territorial accessibility represents a major challenge, with indirect impacts on the functioning of societies. However, the specific issue of disrupted territorial accessibility related to avalanches remains poorly studied. Existing studies generally adopt a static approach, overlooking the full range of demographic characteristics of territories. These factors, however, are essential and are rapidly changing under the influence of climate change, which alters both the frequency and intensity of natural hazards and contributes to changes in demographic dynamics. This study proposes a temporal analysis of the risk of territorial isolation. It is based on a diachronic quantification of road network vulnerability, enabling the characterization of risk evolution through spatially explicit indicators. The methodology is based on the calculation of indices that combine an exposure index with demographic data. The objective is to develop a territorial isolation risk index that accounts for the dynamics of both hazards and territories. Application of this method to the Alpes-Maritimes department between 2003 and 2023, for avalanche hazards, reveals a concentration of territorial isolation risk on a limited number of road segments. Several road segments providing access to potentially isolated population centres exhibit high exposure levels, but the risk is primarily concentrated on the access route to the Isola 2000 ski resort, where high avalanche exposure combines with a large served population. Furthermore, although the risk level of this road segment fluctuates over the study period, no significant decrease has been observed since 2003. This approach highlights the operational value of combining exposure indices and demographic data with a temporally dynamic approach to analyse changes in territorial isolation risk.