Brief Communication : Key papers of 20 years in Natural Hazards and 1 Earth System Sciences

18 To mark the twentieth anniversary of Natural Hazards and Earth System Sciences (NHESS), an interdisciplinary and 19 international journal dedicated to the public discussion and open-access publication of high-quality studies and original 20 research on natural hazards and their consequences, we highlight eleven key publications covering major subject areas of 21 NHESS that stood out within the past 20 years. The papers cover all the subtopics contemplated in the EGU division on Natural 22 Hazards including dissemination, education, outreach and teaching. The selected articles thus represent excellent scientific 23 contributions in the major areas of natural hazards and risks and helped NHESS to become an exceptionally strong journal 24 representing interdisciplinary areas of natural hazards and risks. At its 20th anniversary, we are proud that NHESS is not only 25 used by scientists to disseminate research results and innovative novel ideas but also by practitioners and decision-makers to 26 present effective solutions and strategies for sustainable disaster risk reduction. 27

Hazards including dissemination, education, outreach and teaching. The selected articles thus represent excellent scientific 23 contributions in the major areas of natural hazards and risks and helped NHESS to become an exceptionally strong journal 24 representing interdisciplinary areas of natural hazards and risks. At its 20th anniversary, we are proud that NHESS is not only 25 used by scientists to disseminate research results and innovative novel ideas but also by practitioners and decision-makers to 26 present effective solutions and strategies for sustainable disaster risk reduction. There are currently ten thematic topics covered by NHESS (Table 1). To mark the twentieth anniversary of the journal, we 51 showcase eleven key publications that stood out within the past 20 years. The selection of articles is based on the following 52 criteria: scientific novelty and community impact, and diversity of article types. The process of selecting the key papers was 53 driven by the first author together with the NHESS editorial board and was composed of the following steps: (1) based on high 54 citation and download numbers, as well as the expertise of the thematic editors, a short-list of 30 papers with the highest 55 scientific novelty and community impact was compiled.
(2) Each of these papers was assigned to the NHESS topic it represents, 56 some of them represent multiple topics (e.g., Merz et al., 2010;Peduzzi et al., 2009;Mani et al., 2016). Additionally, also the 57 article type was identified, such as research paper, review, and invited perspective. (3) The NHESS editors discussed the 58 selection with the aim of optimally representing the NHESS topics and reflecting the diversity of article types. The key articles 59 were selected by consensus among the editors. Six of the hazards highlighted in our overview are closely related to weather 60 driven mechanisms that can be amplified by the ongoing climate change to various degrees, as mentioned in the latest IPCC 61 Assessment Report (Ipcc, 2021 278 citations, was one of the initial contributions on was one of the initial contributions on quantitative assessments of risks 70 more correctly be defined as hazard maps. The incorrect use of the terms creates therefore a serious drawback in the overall 103 management of the risk. Typically, assessment of the hazard plays a much prominent part with respect to that regarding the 104 damage, and this results in a mismatch in the quality of the available models and datasets for evaluating the economic damage. 105 Therefore, the thorough review of methods for the assessment of economic flood damage provided by Merz et al. (2010) was 106 and is still of high value for both practitioners and scientists so much so that many new approaches have been. However, we 107 should also mention that many new approaches have been developed in the meantime, since the review was published in 2010. 108

Assessment of storm losses 165
The quantification and forecasting of impacts associated with the occurrence of natural hazards like windstorms or floods is 166 of major importance for society and stakeholders (e.g., Merz et al. (2020)). One of the first efforts to provide a simple but physically based quantification of windstorm associated damage to buildings and infrastructure was the seminal work of Klawa 168 and Ulbrich (2003). The authors considered daily maximum wind gusts from German weather stations, which were scaled by 169 the local 98th percentile to account for local wind conditions and determine the area where damage potentially occurred 170 (windstorm footprint). The scaled wind gusts exceeding the 98 th percentile are cubed (V³) to account for the wind's destructive 171 power, and are weighted with the population density (a proxy for the local insured property). The authors found high 172 correlations between their loss model and the loss data from the German Insurance Association. 173 The loss model by Klawa and Ulbrich (2003) has since proved to be a highly efficient and widely applicable approach, 174 becoming a very popular and easy-to-use socio-economic loss model for insurance applications, and leading to a wide number for setting up an early warning system is the identification of the relationship between the precursors and landslide occurrence 189 (Segoni et al., 2018). A large number of papers have treated this problem by attempting to derive thresholds expressed in the 190 form of a power-law between rainfall event duration and mean intensity or event rainfall (the total rainfall depth accumulated 191 over rainfall event duration), inspired by the pioneering paper by Caine (1980). Not many researchers have questioned this 192 method for decades. 193 With their invited perspective, Bogaard and Greco (2018) discussed some theoretical reasons to move beyond this traditional 194 approach. They stress that thresholds based only on rainfall event characteristics may not sufficiently reflect the hydrological 195 processes occurring along slopes. In particular, intensity-duration thresholds do not allow to explicitly take into account the 196 fact that the triggering rainfall event may be just the final "push" (trigger) after a longer wet period that predisposed the slope 197 to fail (cause). They argued then that the cause-trigger concept may be better represented by hydro-meteorological thresholds. 198 The term hydro-meteorological refers to the fact that these types of thresholds should combine a meteorological variable 199 (rainfall depth) with a hydrological one, reflecting the water storage at the catchment or local scale. 200 Water stored in the unsaturated zone, is however a variable that is more difficult to measure with respect to precipitation. On 201 the other hand, soil moisture information is increasingly becoming available, thanks to remote sensing missions. Reanalysis this improvement remains however quite challenging. Soil moisture presents high spatial and temporal variability, and remote 207 sensing productsas well as reanalysis onesare available only at coarse temporal/spatial resolutions; comparisons with in 208 situ measurements have shown that accuracy issues may be present as well. Notwithstanding such obstacles to deal with, the 209 invited perspective is stimulating scholars to move beyond an approach that remained nearly unquestioned for many years. The findings suggest that serious games have the potential to be effective tools in volcano education for both traditional (school 265 students) and non-traditional (i.e., adults) stakeholder groups. Though serious video games, therefore is a promising 266 communication and educational technique, this approach faces a number of challenges such as expensive and time consuming 267 processes of game development. The study by Mani  Meteorological tsunamis (or simply known as meteotsunamis) are typically recognized as long ocean waves, which have the 286 same frequencies and spatial scales as tsunami waves of seismic origin, but produced by atmospheric processes. They are 287 triggered by extreme weather events atmospheric conditions at the ground or mid-troposphere including severe thunderstorms, 288 squall lines (a sudden violent gust of wind or localized storm, especially one bringing rain, snow, or sleet), storm fronts, 289 hurricanes or instable intense mid-troposphere jets generating atmospheric gravity waves, generated through the rapid changes 290 in barometric pressure, (a few hectopascals over a few minutes) or wind . The similarity between atmospherically generated 291 "meteotsunamis" and seismically generated tsunamis is strong enough that it can be difficult to distinguish one from the other. 292 The article by Monserrat et al. (2006) is one of the very few studies that describes the hazardous phenomena of meteotsunamis 293 in the World Ocean to show the similarities and differences with seismic tsunamis. Analysing several cases, Monserrat and his 294 team found that both tsunamis and meteotsunamis have the same periods, same spatial scales, similar physical properties and affect the coast in a comparably destructive way. In addition, some specific features of meteotsunamis such as the coupling 296 between the moving disturbance and the surface ocean waves make them akin to landslide-generated tsunamis. Monserrat et 297 al. (2006) found that the major difference between the tsunamis and Meteotsunamis is associated with the specific properties 298 (mainly the resonant factors) of corresponding sources. During resonance of the ocean driven by atmospheric forcing, the 299 atmospheric disturbance propagating over the ocean surface is able to generate significant long ocean waves by continuously 300 pumping energy into these waves. This contrasts to seismic tsunamis that can have globally destructive effects without any 301 resonant factor. However, the Meteotsunamis are always local and much less energetic than seismic tsunamis. The destructive 302 meteotsunamis are always the result of a combination of several resonant factors such as Proudman, Greenspan, shelf, harbour. 303 As the probability of occurrence for such a combination is very low, the destructive meteotsunamis are infrequent and observed 304 only at some specific locations in the ocean. 305

Drier conditions in Mediterranean regions 306
The Mediterranean Region is considered a hot-spot of climate change. This qualification is supported by different natural and 307 socioeconomic reasons, being one of them its impact over hydrometeorological hazards, specifically, droughts. Despite the 308 high uncertainty associated to the application of climatic models over the rainfall in this region, there is a high confidence on 309 the drought risk increase (Medecc, 2020), mainly due to precipitation reduction, a negative trend in moisture availability, and 310 warming-enhanced evaporation. In a region where, in average, more than 65% of the freshwater is for agriculture near a 30% 311 is for the direct use of water by the population, and the remaining 5% is for industry, energy and tourism, droughts increase 312 implies that water related intersectoral conflicts are likely to be exacerbated. Even more so if we consider that in 2025 about 313 530 million people will live in the Mediterranean, and that the increase in temperature will lead to an increase in irrigation 314 needs from 4 to 18% (Medecc, 2020). Although today there are already numerous studies at local and regional scale on the 315 observed spatial and temporal evolution of drought conditions, the paper by Sousa  Research Unit. The scPDSI is based on the water budget for a certain period estimated from precipitation, temperature and soil 320 characteristics and self-calibrated from local data. This index was applied to the Mediterranean Region and to four selected 321 sub-regions, homogeneous in terms of drought characteristics and socio-economic relevance, for the period 1900-2000. After 322 a robust analysis the scPDSI showed a clear trend towards drier conditions in most Mediterranean Region. This index 323 reproduced well the strong decadal and inter-annual variability between subregions along all the century and showed how the 324 drought period recorded during the 1940s was extended from Iberia until the Balkans Region. Having in mind that determined 325 synoptic patterns favours the deficit of precipitation and previous literature, and after analysing different major potential for this region. The paper revealed the link between dry periods estimated by scPDSI and the positive phase of the NAO during 328 winter and subsequent climatic seasons over the western Mediterranean, while the Scandinavian index presented a less 329 homogeneous but significant pattern between winter and summer over central Mediterranean. Those teleconnections joined to 330 the influence of the sea surface temperature (SST) anomalies allowed the creation of a stepwise regression model that was able 331 to forecast summer drought conditions six months in advance and was capable of reproducing the observed scPDSI time series 332 fairly well. Although it is a simple algorithm it provides a useful approach to seasonal forecasting of droughts, that can be very 333 useful in a panorama characterized by an increase in dry periods. 334

Conclusion 335
The above articles represent excellent scientific contributions in the major subject areas of natural hazards and risks and helped 336 NHESS to become an exceptionally strong journal representing interdisciplinary areas of natural hazards and risks. consequences, but also the journal is dedicated to the open discussion engaging interdisciplinary scientific communities. At its 340 20th anniversary, we are proud that NHESS is not only used by scientists to disseminate research results and innovative novel 341 ideas but also by practitioners and decision-makers to present effective solutions and strategies for sustainable disaster risk 342 reduction. 343