This research work of this paper completes the authors' activity in the Military Emergencies Unit of Spain in responding to the problem of slope instabilities and how to predict them in the case of heavy rains. This work completes the work of stabilization and stresses made by torrential rains in the south of Spain. The results have been satisfactory.

Flood defences have the potential to be integrated with many other functions. While assessments of multifunctional dikes are usually conservative, using a probabilistic framework allows for synergies to be integrated and the risks to be made explicit. This change leads to a better perspective on the protection level of a flood defence and the effectiveness of reinforcements, allowing for better implementation of multifunctional elements on flood defences.

This study fully considers the characteristics of expressways in mountain areas. The catchment area is considered a prediction unit. The method of slope division is improved, and a method of improving the parameters in the model is proposed. Comparison and analysis with actual observation data show that the method of soil and water loss prediction adopted in this paper has less error and higher prediction accuracy than other models and can satisfy prediction requirements.

We propose in this study a fundamentally new approach for the definition of minimum rainfall required for the initiation of landslides based on satellite-derived antecedent rainfall estimates directly coupled with data on surface susceptibility for landslides. We apply our approach in the western branch of the East African Rift and provide first regional rainfall thresholds for landsliding in tropical Africa.

We introduce a new prediction-oriented method for hazard assessment of volcaniclastic debris flows, based on multiple models. We apply our procedure to a case study of the 1955 Atenquique flow, using three widely used depth-averaged models. Depending on how it is looked at, the exercise provides useful information in either model selection or data inversion. Connecting inverse problems and model uncertainty represents a fundamental challenge in the future development of multi-model solvers.

The study reports on the impact of the assimilation of wind data from airborne Doppler cloud-profiling radar in a kilometre-scale NWP model on predicting heavy precipitation events in the Mediterranean area. The positive impact of the assimilation of such data is particularly evidenced for a heavy precipitation event and results are slightly encouraging over a 45-day period. In addition, the impact of the length of the assimilation window in a 3h-3DVar assimilation system is investigated.

Mitigating landslide risk requires information on landslide hazards on a suitable scale to inform decisions. We develop simple rules to identify landslide hazards and the probability of being hit by a landslide, then test their performance using six existing landslide inventories from recent earthquakes. We find that the best rules are "minimize your maximum look angle to the skyline" and "avoid steep (> 10˚) channels with many steep (> 40˚) areas that are upslope".

Atmospheric ice accretion caused by freezing precipitation leads to severe damage and failure of buildings and infrastructure. This study investigates projected changes to extreme ice loads used to design infrastructure over North America for future periods of specified global mean temperature change using a Canadian regional climate model. Increases in ice accretion for latitudes higher than 40° N are substantial and would have clear implications for future building and infrastructure design.

The main damage characteristics have been reported in Nepal caused by the 2015 Nepal earthquake but not in China. Our investigations suggested that damage caused by the earthquake in Tibet varies with intensity, amount of rock weakened by previous movement, steepness of slope, and lithology. The damage shows directional features mainly developed in the N-trending rifts in southern Tibet. The earthquake-induced landslides and collapses generally occurred where previous ones had taken place.

We study rockfall hazard to a town in an earthquake-prone area, where large trailer-truck-sized boulders are scattered downslope above the town. Mapping boulder locations and sizes (in the field and in past aerial photos) and calculating their predicted trajectories downslope using computer simulation yielded a hazard map for rockfall impact. Hazard is reduced where slope angle is below 10°. Dating rockfalls coincides with past earthquakes and predicts probability for future rockfall.

The potential of W-band radar reflectivity to improve the quality of analyses and forecasts of heavy precipitation events in the Mediterranean area is investigated. The 1D + 3DVar assimilation method has been adapted to assimilate the W-band reflectivity in the Météo-France kilometre-scale NWP model AROME. The results suggest that the joint assimilation of W-band reflectivity and horizontal wind profiles lead to a slight improvement of moisture analyses and rainfall precipitation forecasts.

The study integrates tsunami hazards that create risk into hazard-product risk and hazard-affected risk. Consistent with different factors evaluated for these two risk dimensions, different spatial distributions between them can be formed. The study recommends tsunami shelter accessibility to be enhanced where hazard-product risk is large and where hazard-affected risk is high in a different and targeted way. The tsunami risk situation’s complexity in coastal port cities makes this interesting.

This paper investigates the formation of the Mediterranean hurricane that developed between Balearic Islands and Sardinia in October 1996, with a particular focus on the influence of sea surface temperature. We show that increased sea surface temperatures lead to greater probabilities of appearance and a greater strength of the resulting hurricane, suggesting that the processes for Mediterranean hurricanes at steady state are very similar to tropical cyclones.

A state-of-the-art regional climate model ensemble for Europe is investigated for extreme precipitation intensities. The models poorly reproduce short duration events of less than a few hours. Further, there is poor connection to some known hotspots for extreme cases. The model performance is much improved at 12 h durations. Projected future increases scale with seasonal mean temperature change, within a range from a few percent to over 10 percent per degree Celsius.