Estimating risks induced by interacting natural hazards remains a challenge for practitioners. An approach to tackle this challenge is to use multivariate statistical models. Here we evaluate the efficacy of six models. The models are compared against synthetic data which are comparable to time series of environmental variables. We find which models are more appropriate to estimate relations between hazards in a range of cases. We highlight the benefits of this approach with two examples.

Awu is a little-known volcano in Indonesia, and paradoxically it is one of the deadliest volcanoes on Earth. Some of its recurrent intense eruptions have induced world-scale impacts. The pulverization of a cooled lava dome and its conduit plug have allowed lake water injection into the conduit, leading to explosive water–magma interaction. The past vigorous eruptions were likely induced by these phenomena and it is a possible scenario for future events.

Hydrological impacts of climate change are generally performed by following steps from global to regional climate modeling through data tailoring and hydrological modeling. Usually, the climate–hydrology chain primary focuses on medium to large catchments. To study impacts of climate change on small catchments, a high-resolution regional climate model and hydrological model are required. The results from high-resolution models help in proposing specific adaptation strategies for impacts.

Rock avalanches are extremely dangerous, causing much damage worldwide. The Masiere di Vedana is a rock avalanche deposit (9 km², 170 Mm³) in NE Italy. We dated it back to late Roman to early Middle Ages. Identified drivers are the overall structural setting, exceptional rainfall events and seismic shakings. No exceptional event is required as a trigger. When dealing with heavily deformed bedrocks, especially in inhabited areas, the occurrence of a huge event like this must be considered.

There is increasing interest and need to analyze the contribution of anthropogenic climate change to negative impacts of climate change. We study the case of glacial lake Palcacocha in Peru, which poses a significant flood risk to the city of Huaraz. We found that greenhouse gas emissions; strong urbanization processes without appropriate land use planning; and social, cultural, political, and institutional factors all contribute to the existing flood risk.

A catastrophic flash-flood event hit the northeastern part of Mallorca in 2018, causing 13 casualties and impacting on the international opinion in one of the most important tourist resorts. The analysis of the rainfall–runoff processes illustrated an unprecedented flashy behaviour in Europe triggering the natural disaster. UAVs and hydrogeomorphological precision techniques were used as a rapid post-catastrophe decision-making tool, playing a key role during the rescue searching tasks.

A localized empirical model consisting of multilevel parameters has been built to evaluate the flood vulnerability of residential buildings in a heritage community of Kuala Lumpur, Malaysia. A new economic loss model is developed to quantify the flood risk in terms of replacement cost, taking into account both specific vulnerability and a normalized depth–damage ratio function. The findings provide multiscale flood-resistant strategies for the protection of individual residential buildings.

In this study, we corrected Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation of Water Resources (APHRODITE) in the Yarlung Tsangpo–Brahmaputra River Basin using both linear and nonlinear methods, and their influences on resulting extreme precipitation indices were assessed. Results showed that all methods were able to correct mean precipitation, but their ability to correct wet-day frequency and coefficient of variation were markedly different.

This study highlights the use of the software Flow-3D in reproducing landslide-generated impulse waves. Due to the available data and the possibility of comparing the results with other previous works, a numerical modelling investigation on the 1958 Lituya Bay tsunami event is proposed. It is noted that the rockslide impact into the waterbody has a key role in the wave initiation and thus its propagation. The concept used in this work can be applied to prevent such phenomena in future.

Outlined here is a multi-disciplinary framework for analysing and evaluating the nature of vulnerability to, and capacity for, flood hazard within a complex urban society. It provides scope beyond the current, reified, descriptors of flood risk and models the role of affected individuals within flooded areas. Using agent-based modelling coupled with the LISFLOOD-FP hydrodynamic model, potentially influential behaviours that give rise to the flood hazard system are identified and discussed.

In the Canary Islands, a link between rainfall and rockfall occurrence is found for most of the year, except for the warm season. Empirical rainfall thresholds for rockfalls are first proposed for Gran Canaria and Tenerife, and the dependence of the thresholds on the mean annual rainfall is discussed. The use of thresholds in early-warning systems might contribute to the mitigation of the rockfall hazard in the archipelago and reduce the associated risk.

This work presents a study of the biggest mudflow event in 20 years in Chilean Patagonia, which resulted from an avalanche in the Cordon Yelcho. We integrate in situ geotechnical tests and numerical modeling to model the Villa Santa Lucía mudflow event. Our results suggest that the initial soil water content is sufficient to transform the landslide and scoured soil into a mudflow. Therefore, knowing the soil characteristics is crucial to evaluating the impact of landslides in the study area.

In 2018 windsurfer’s mast broke about 1 km offshore during a scirocco storm in the northern Adriatic. He was drifting in severe conditions until he eventually beached alive and well in Sistiana (Italy) 24 h later. We conducted an interview with the survivor to reconstruct his trajectory. We simulate his trajectory in several ways and estimate the optimal search-and-rescue area for a civil rescue response. Properly calibrated virtual drifter properties are key to reliable rescue area forecasting.

The present study estimated the Kerman–Baghin aquifer vulnerability using DRASTIC and ‎composite DRASTIC (CDRASTIC) indices with the aid of geographic information system (GIS) techniques. The aquifer vulnerability maps indicated very similar results, ‎identifying the north-west parts of the aquifer as areas with high to very high vulnerability. According to the results, parts of the studied aquifer have a high vulnerability and ‎require protective measures.

Forecasting of daily fire weather indices driven by the ECMWF ensemble prediction system is shown to have a good skill up to 10 d ahead in predicting flammable conditions in most regions of the world. The availability of these forecasts through the Copernicus Emergency Management Service can extend early warnings by up to 1–2 weeks, allowing for greater proactive coordination of resource-sharing and mobilization within and across countries.