**1. Introduction**

A new era has unequivocally emerged that has brought climate change and its impacts to the foreground of scientific research. There is growing evidence that weather and climate extremes (i.e., hazards) are increasing in frequency, intensity, spatial coverage and duration, indicating the need for a more meticulous investigation and a better physical understanding of the processes governing the state of the climate and its future evolution [1–3]. The adverse impacts of extreme events, evidenced in the reported data of disaster implications, e.g., [4–6], are also an active subject of climate research of paramount importance [7–11]. Noteworthily, anthropogenic effects are emerging as the underlying cause of the weather and climate extremes [1,12–18].

Over recent years, research works have consistently reported that climate change aggravates climate hazards, amplifying the risks of various impacts (river and coastal floods, wildfires, droughts, landslides, etc.) [19–23]. Several studies on temperature and precipitation extremes have provided important findings on the regional variability of the impacts of climate change across Europe, e.g., [24–29]. Forzieri et al. [30] reported that the risks of wildfires, windstorms and inland flooding would increase in Europe, with varying degrees of change across regions, while the most dramatic rise is predicted to be in damages in southern Europe caused by heatwaves, droughts and coastal floods. The report of PESETA IV [31] consolidated those findings and indicated "a clear north–south divide, with the southern regions in Europe being much more impacted by the effects of extreme heat, water scarcity, drought, forest fires and agriculture losses". The estimated

**Citation:** Vlachogiannis, D.; Sfetsos, A.; Markantonis, I.; Politi, N.; Karozis, S.; Gounaris, N. Quantifying the Occurrence of Multi-Hazards Due to Climate Change. *Appl. Sci.* **2022**, *12*, 1218. https://doi.org/10.3390/ app12031218

Academic Editor: Jason K. Levy

Received: 15 December 2021 Accepted: 21 January 2022 Published: 24 January 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

patterns of climate-change developments urge for more efficient risk management of climate-related extremes and disasters in order to significantly advance climate-change adaptation, particularly in the most vulnerable regions. This consequently implies the need for a reliable quantification of the probability of extremes in the current and future climate. Identifying the climate vulnerabilities of key societal systems should be based on a detailed knowledge of projected climate-change hazards and the factors affecting the likelihood of each one for the selected assessment of the region of interest.

According to the report of the United Nations Office for Disaster Risk Reduction [32], the term "multi-hazard" is used to promote risk reduction and disaster management, and denotes hazardous events that may occur simultaneously or cumulatively over time. One of the most challenging research questions is the harmonization of risk metrics to allow the comparison of risks across hazards, regions, time, assets, or sectors [33]. Establishing a harmonized risk understanding would pave the way to a multi-hazard risk assessment, introducing interactions and cascading effects as well as providing some analytical interpretations of the compound and systemic risks. This would lead to more credible scenarios for describing future disaster events in terms of their magnitude and probability based on the validated scientific knowledge that can benefit from high-resolution climate projections.

The single-hazard risk assessment is a proven methodology, but shifting to multihazards is not a linear or easily understood process, as a multi-hazard risk analysis is not just the sum of single hazard risk examinations and thus, comparability of the singlehazard results is strongly needed [34]. Due to the diversity of the hazard characteristics' complex relationships, triggering effects, climate-changing mechanisms, compounds and interactions could be potentially established [35–37].

In this work, the occurrence of hazards due to climate change was determined for Greece using data that were dynamically downscaled to a very high resolution by the Advanced Weather Research and Forecasting (WRF-ARW) model [38], initially produced by the EC–Earth Global Climate Model (GCM) [39]. The hindcast period covered the years from 1980 to 2004, while for the future projections, two different periods, i.e., 2025–2049 (near future) and 2075–2099 (far future), were studied using the Intergovernmental Panel on Climate Change (IPCC) Representative Concentration Pathways (RCPs), RCP4.5 and RCP8.5, following the recommendations of the EU National Risk Assessment [33] and similar studies in the US [40]. The RCP scenarios demonstrated a significant convergence in their emission pathways in the near future and considerable deviations towards the end of the century [41].

It has been established that high-resolution, dynamic-downscaling models applied to regional climate assessments can be implemented to assess the climate-change impacts on extremes, especially in areas with complex topography and local scale effects [42–45]. Here, we sought to provide the first step towards a comprehensive multi-hazard risk assessment for the country based on high-resolution model data to support training and preparatory activities for disaster risk reduction (DRR). The analysis focused on four critical climate hazards for Europe: heat and cold extremes, flash floods, and windstorms, each one described by a climate indicator (Section 2.2).

The scope of this work was to carry out a very detailed assessment of the most significant hazards that have occurred in Greece in the past and to predict their evolution in the future considering the impact of climate change. This is a highly valuable process as disaster management should also take into account the (non-)stationary characteristics of climate change. The current study examined these parameters for Greece using veryhigh-resolution climate simulations at 5 km. Furthermore, one of the goals was to identify a common categorization framework across different climate hazards, which allowed a direct and coherent prioritization of the hazards and their evolution due to climate change considering complex patterns due to local geographic conditions. The produced hazard data could readily be applied to the generation of multiple scenarios with various likelihoods of occurrence in order to obtain a more complete picture of risk [32], accounting for climate-change projections (IPCC). In addition, this work was based on the recommenda-

260

tions set by the UNDRR/ISC Sendai Hazard Definition and Classification Review Technical Report [46] and intended to introduce the climate dimensions and dynamic evolution of risk harmonization that was missing from such assessments [33].

Section 2 focuses on the details of the data used and the methodology that was developed to estimate the probability of the occurrence of extreme values of the variables. Section 3 presents the results and discussion of the analysis applied to the quantification of risks and the likelihood of hazard evolution due to climate change. Finally, the final section concludes the paper.

### **2. Materials and Methods**

### *2.1. Area of the Study and Model Datasets*

The study area included the country of Greece. The country presents several climatic variations, always in the Mediterranean climate frame, due to the influence of its vivid geomorphologic complexity (interplay of mountainous regions and plains, extended coastline, and numerous islands) on the different atmospheric-pressure dependencies from the Atlantic, central Mediterranean area, Eurasia and North Africa. This enhances the need for higher-resolution climate modeling to resolve the topography features more effectively.

In the present work, climate data of EC–Earth (1.125◦ horizontal resolution originally) downscaled by the WRF-ARW version 3.6.1 model to 5 <sup>×</sup> 5 km<sup>2</sup> were employed at a temporal resolution of 6-h. The hindcast climate simulations have been extensively evaluated in our previous works, whereby exhaustive quantitative validation of the highly resolved fields of temperature, precipitation, wind speed and solar radiation were performed for our observations [45,47–50]. The WRF-ARW modeling domain covering Greece comprises a grid of 185 × 185 cells in the horizontal and 40 levels in the vertical that are arranged according to terrain, following the hydro-static-pressure vertical coordinates (up to ~50 mbars). A more detailed description of the WRF model setup and physical parameterization schemes can be found in [47].

For the simulations of future years under the influence of climate change, the two IPCC greenhouse-gas-emission scenarios, RCP4.5 and RCP8.5, were selected as they constitute the most commonly used scenarios by impact-assessment modelers. In particular, RCP4.5 represents an increase in the radiative forcing of the atmosphere of 4.5 W/m<sup>2</sup> relative to the pre-industrial era with a profile of greenhouse-gas emissions increasing until the midcentury (~2050) and stabilizing thereafter until the end of the century (2100). On the other hand, RCP8.5 is considered to be the most extreme scenario with greenhouse-gas emissions increasing sharply until the end of the century, implying at its end a radiative forcing of 8.5 W/m<sup>2</sup> relative to the pre-industrial era. The future time periods in the RCP4.5 and RCP8.5 simulations were selected in order to study the projected climate-change effects on hazard dynamics, both in the middle and near the end of the 21st century. For the present analysis, the model-downscaled data that were used and the corresponding 25-year-period slots are presented in Table 1.

corresponding periods. **Dataset Period**

**Table 1.** The global (EC–Earth) model datasets downscaled by WRF model used in the study and

