The evolution of coastal hazards in the context of climate change has been addressed at the regional scale by studying the height and frequency of extreme sea levels (ESL). However, sea level is not the only factor determining the hazard changes that can be used at this scale. Therefore, this article proposes an assessment method of coastal hazard changes integrating other determinants: geographical configurations (continental or island), tidal regimes and meteo-oceanic event types. This method, applied to the coasts of France (mainland and overseas), reveals significant differences in the evolution of coastal hazards: coasts subjected to high tidal ranges and storms (e.g., Atlantic, English Channel and North Sea) will experience a relatively moderate evolution of the hazard, thanks to «training» for the future conditions that present-day high variations constitute. Conversely, the microtidal shorelines of temperate latitudes (e.g., those of the Mediterranean) benefit from only a small variability generated mainly by storm surges, and are therefore poorly prepared for sea level rise. The situation of the small tropical islands is of particular concern: with the passage of cyclones these territories are subjected to very energetic sea states, but by their form, the surges remain moderate, which constitutes, as well as the low tidal ranges, a limiting factor for preparing for sea level rise. In addition to this approach at the regional level, geological, sedimentary and biological evolutions, as well as local hydraulic phenomena, should be considered to assess at a finer spatio-temporal scale the hazard changes.

Soil erosion is impacted by climate and land use changes which need to be quantified to assess future risks and to design efficient soil conservation measures. The Coupled Model Intercomparison Project Phase 5 (CMIP5) simulations have provided the basis for most such assessments and yet are being gradually superseded by more recent simulations from Phase 6 (CMIP6). The High-Resolution Model Intercomparison Project (HighResMIP) experiment in CMIP6 adds value over the downscaled CMIP5 simulations by improving process representation in the global climate system. Our study investigates and compares high-resolution model simulations from CMIP6 against CMIP5. Model evaluation for the reference period (1986–2005) indicates that the CMIP6 model outperforms the regional climate models (RCM) from CMIP5 for better circulation simulations, but both overestimate soil erosion in China. The average projected soil erosion increases by 27.85 from CMIP5 and 20.03 t·hm-2·a-1 from the CMIP6 model with remarkable geographical heterogeneity. Soil erosion is projected to decrease in black soil regions, purple soil regions, and karst regions from CMIP6 results, which is opposite to the increasing trend found in those regions from CMIP5. Land use and climatic changes contributed 51.68% and -5.92% respectively from CMIP5 simulations while 35.74% and -13.77% from CMIP6 to the increased soil erosion rate. The negative contribution of land use change is gradually intensified with the CMIP6 model representing finer-scale processes of converting land-use type into cropland, pasture, and urban land. Overall, the CMIP6 projections provide a less severe soil erosion situation while addressing the need to pursue soil conservation more.