Damaging Hydrogeological Events (DHE) occur due to a combination of predisposing factors, triggering causes and the presence of vulnerable elements (Petrucci and Polemio, 2009). A DHE is made of two components: a Rainfall Event (RE) and a subsequent Damage Event (DE), resulting from floods and landslides triggered by rain. Depending on both rivers and slopes conditions, the same rain amount either can or cannot cause damage. Moreover, damage occurrence depends on the geographical distribution of vulnerable elements such as population, lifelines and urbanized sectors. DHE can affect entire regions and may last for more than one day, and, accordingly, they can show high variability, both in space and in time. For this reason, it can be difficult to compare causes and effects of these events in a systematic way.

Over the last few years, in some countries, such as Italy, there has been an increase in both the recurrence and the severity of DHE's damage. Indeed, while during the past century the DHE were identified following the year of occurrence (e.g., the event of 1966), nowadays, the events are often identified by means of both the year and the month of occurrence, given that a region is frequently affected by more than one event per year. In several areas of the world, an increasing number of damaging events, related to increasing frequency and intensity of extreme precipitation have been highlighted (e.g. Alexander et al., 2006; Min et al., 2011; Westra et al., 2013). Climate models also indicate that further increases in extreme precipitation should be expected over the next decades (Pall et al., 2006; O'Gorman and Schneiderb, 2009; Toreti et al., 2013), with moist regions which are expected to be wetter, and dry regions drier (Chou and Neelin, 2004; Wentz et al., 2007; Zhang et al., 2007; Allan and Soden, 2008). In this framework, a great deal of research has been carried out on daily precipitation, by means of various methods of analysis, such as indices, percentiles, thresholds, and extreme value theory (Groisman et al., 2005; Reiser and Kutiel, 2010; Toreti et al., 2010; Durão et al., 2010; Caloiero et al., 2016). Several studies focused their attention on rain variability, evidencing positive trends in daily rain intensity and a tendency toward higher frequencies of heavy and extreme rainfall (Kunkel et al., 1999; Plummer et al., 1999; Osborn et al., 2002; Villarini et al., 2011; Caloiero, 2014, 2015).

In addition to the changes in the climatic features, the literature also focuses on the analysis of historical DHE's scenarios (Brázdil et al., 2011; Bullón, 2011). Actually, this analysis can supply useful information for the management of future events, providing an overview of the critical sites during past events that, without preventive measures, can be damaged again in future events. This information can be useful in decision support, emergency management and implementation of prevention and response measures. Moreover, they can play an important theoretical guiding role in meteorological disaster prediction, and future disaster prevention and reduction (Guan et al., 2015). In fact, an insight into past disasters allows assessing the performances of multiple mitigation strategies (Day and Fearnley, 2015) and, consequently, to help in the management of future events in which the simultaneous occurrence of landslides and floods must be taken into account. Especially for floods frequency estimation, the value of historical data is generally recognised in different countries, but practical methods for systematic and routinely inclusion of these data into risk analysis are rarely available, even though harvesting data on past extreme events could improve the reliability of flood risk assessments (Kjeldsen et al., 2014). Additionally, the analysis of historical DHE that affected a region offers didactic information to share with the population in order to improve the consciousness of risks affecting the areas in which they live, or can be used to calibrate loss estimation models (Papathoma-Köhle et al., 2015).

This study compares two events that affected the same area at a temporal distance of 15 years, highlighting similarities and differences in triggering rain, resulting damage and emergency management practices. In the Sect. 2, we summarise the methodology pointed out in previous papers to classify DHE according to severity of both triggering rain and resulting effects. Section 3 introduces the study area, while Sects. 4 and 5 describe the two compared DHE. Then, Sect. 6 presents the analytical comparison of the studied events, and Sect. 7 draws some conclusive remarks.

To perform the comparison of two DHE, we have to take in mind that each event presents two components: the rain and the damage indirectly caused by rain, by means of phenomena as landslides and floods. Unfortunately, in the study area there are not hydrometric data that can be included in the analysis. For this reason, in order to perform this comparison we applied the methodology proposed in Aceto et al. (2016), based on the assessment of a damage score and a rainfall score combined to represent the exceptionality of the events. To assess the damage score, we evaluate three damage indicators (IDA, DI, and NoV). The Index of Damaged Area (IDA) is the sum of the areas of the damaged municipalities divided by the area of the region. It is the percentage of regional area that was damaged and represents a proxy of the actual damaged area. Obviously, the antecedent condition, i.e. in terms of terrain saturation, and the vulnerability and the resilience of the territory of the region are crucial elements affecting the level of damage of a DHE. In fact, the characteristics of Calabrian towns and villages in the past were different. For example, one of the most damaging DHE occurred in Calabria on 1953, and killed around 80 people. Several victims died in shacks realized to shelter people affected by previous DHE, which occurred just two years before, on 1951. Because of damage caused by the 1953 DHE, two villages (Brancaleone Superiore and Caulonia hamlet) were definitively abandoned (Petrucci and Pasqua, 2014). The Damage Index (DI) is a relative evaluation of direct damage. According to the elements damaged in the Calabria region throughout the historical series of regional DHEs, the damageable elements are sorted into seven types (roads and railways, houses, public buildings, services, productive activities, hydraulic works, and people) and their values range from 0 to 1, on an arbitrary scale, while the levels of loss are set as: 1 = high; 0.75 = medium high; 0.5 = medium; and 0.25 = low. The sum of the products of the damaged elements by the respective levels of loss is the DI value, which is finally multiplied by a scale factor of 10 to avoid decimal values. The Number of Victims (NoV) is used as a further indicator of damage severity. Finally, the damage score (Dscore) is evaluated as in the Eq. (1), where IDAj, DIj and NoVj are the damage indicators of the event j, and IDAmax⁡, DImax⁡ and NoVmax⁡ are the maximum value of each damage indicators evaluated basing on the historical series of severest Calabrian DHEs:

Dscore=IDAjIDAmax⁡+DIjDImax⁡+NoVjNoVmax⁡/3 For any further and detailed information about the evaluation of the damage indicators, the interested reader can easily refer to Aceto et al. (2016).

The Rainfall Score (Rscore) (Caloiero et al., 2014; revised by Aceto et al., 2016) is based on the return period (T) of the maximum daily rain recorded in each gauge during the days in which the damage event occurred. For each event, given a series of return period classes (selected according to the ones used in the Calabrian Hydro-geological Plan), the Rscore is calculated by evaluating the percentage of gauges falling within these classes, as in the Eq. (2): Rscore=c0⋅∑i=1ni⋅PTi∑i=1nPTi where n is the number classes, i is a return period class, P(T)i is the percentage of gauges falling within each i class, and c0 is a graphic scale factor.

Basing on the combinations of Dscore and Rscore, DHE can be classified as: A.

Ordinary events, characterized by low damage and rain values;

Located at the toe of the Italian peninsula, Calabria has a surface of 15 080 km2; on average, its altitude is 597 m a.s.l., and its tallest relief is 2266 m a.s.l. The reliefs over 500 m a.s.l. high occupy 42 % of the regional area, while hills between 50 and 500 m a.s.l. cover 49 % of the territory, and only 9 % of the region is under 50 m a.s.l. (Fig. 1a). The rugged morphology is due to the still active tectonic uplift that has been started in Quaternary. The geological structure is made of allochthonous crystalline rocks (Palaeozoic to Jurassic), stacked during the middle Miocene (Tortorici, 1982) over carbonate units (Ogniben, 1973), with Neogene flysch filling the tectonic depressions. Tectonic stresses and climatic conditions worsened the rocks characteristics, and disposed slopes to mass movements. Still due to the rapid uplift, the fluvial system is mainly made of ephemeral streams that cannot be gauged. Actually they are dry during summer and undergo severe flash floods in autumn-winter: during floods, active channels move from side to side of the riverbed, making difficult to find the right place to install the gauges. The climate is Mediterranean, presenting sharp contrasts due to the position of the region within the Mediterranean Sea and its orography (Coscarelli and Caloiero, 2012). Specifically, warm air currents coming from Africa and high temperatures affect the east side, leading to short and heavy rainfall, while on the west side, western air currents cause milder temperatures and higher precipitation. Cold and snowy winters and fresh summers with some precipitation are typical of the inner sectors (Caloiero et al., 2015). Numerical simulations of severe weather showed the crucial role played by orography, which enhances rainfall in localized spots (Federico et al., 2003a, b) or forces secondary cyclogenesis which persists over the east side (Federico et al., 2007). The frequent DHE have been causing damage and victims throughout the years (Petrucci and Pasqua, 2012), as highlighted by both historical research (Petrucci and Versace, 2005, 2007; Palmieri et al., 2011; Gullà et al., 2012; Pasqua and Petrucci, 2016), and reports about more recent events (Petrucci et al., 2010; Caloiero and Petrucci, 2014).

The event occurred on September 2000 is known as Soverato event, after the name of the municipality in which 13 people were killed by the flood of Beltrame River. The event lasted three days, during which 27.0 % of regional territory (109 municipalities, 3922.7 km2) was affected (Fig. 1c). The synoptic pattern of the event concerns: (a) an area of high pressure over central Europe, and (b) a low pressure over Sicily Island, originated from the merge of cold/dry air coming from North Atlantic and warm/moist air laying over the Mediterranean, particularly over the Strait of Messina and Calabrian east coasts. The mixing of these two air masses, having different temperature and humidity, originated severe precipitation that were amplified by the orographic effect. The result was a series of thunderstorms over the east coast, from north to south (Fig. 2). Fifty-six rain gauges (35.6 % of the gauges working during the event), managed by the regional Multi-Risk Functional Centre, recorded daily rain higher than 100 mm and, among these, 12 gauges recorded values higher than 200 mm. The maximum value of daily rain recorded was 301.6 mm, and it was recorded on the southeast sector of the region (Fig. 3). It was equal to about five times the September average monthly value of the region, and represents 27.1 % of the mean total annual rainfall. As to what concerns the return period of maximum daily rainfall, 13 rain gauges showed return periods of 10/20 years, 4 of 20/50 years, 2 of 50/100 years, 4 of 100/200 years and one gauge higher than 200 years (Fig. 3).

The central government, with different ordinances, allocated 491 million of euros to repair damage to public and private sectors (communication of Calabrian Civil Protection Unit).

The second analysed event occurred between 30 October and 1 November 2015, and affected 79 municipalities of Calabria (19.3 %), with an area of 2578.9 km2 (17.0 % of regional area) (Fig. 1c). While an area of high pressure with values up to 1040 hPa was located over central Europe, a low pressure moved from central Tyrrhenian Sea towards the Strait of Sicily, causing a marked instability over southern Italy. Particularly on the east coast, low-level eastern flow enhanced intensity and frequency of storms (Fig. 2). 39 rain gauges recorded daily rain higher than 100 mm, but 21 reached rain values higher than 200 mm, and 4 higher than 300 mm, with a maximum value of 370.4 mm (Fig. 3). These values, registered on the southeast sector of the region, were exceptional for the period, with maximum daily values more than three times than the average monthly value. Moreover, in 3 rain gauges, in just one day, more than 33 % of the mean annual rainfall was recorded. In 8 rain gauges, daily rain showed return periods of 10/20 years, in 7 of 20/50 years, in 7 of 50/100 years, in 4 of 100/200 years, and in one rain gauge showed a value higher than 200 years (Fig. 3).

The central government, with different ordinances, allocated about 130 million of euros to repair damage to private and public sectors (communication of Calabrian Civil Protection Unit).

Both the DHE were characterized by a 3-day duration, but they occurred in two months characterised by different climatic features, in fact, in Calabria the September average daily rain in is 64.1 mm, while in October it rises to 123.3 mm. The total rain, expressed as mm no. of gauges-1 km-2, was higher for the 2015 than for the 2000 event (Table 1). To facilitate the comparison, in Fig. 6 we presented the descriptive features of the two events as ratio between the 2015 and 2000 values. The red line marking the value “1” represents the case in which the value of 2015 was equal to the one of 2000. A value above or under this line, indicates that the analysed feature was higher in 2015 or in 2000, respectively. The diagrams of Fig. 6a, b, c, represent the features describing the rain event. Results show that the area affected by daily rain between 50 and 100 mm was higher in 2015, and the area with more than 200 mm was considerably larger for 2015 than for 2000, showing the greater severity of the 2015 event in terms of stronger daily intensity on a larger area (Fig. 6a). From the analysis of the maxima (Fig. 6b), both 1-day and 2-day values were greater in 2015, while 1, 3, and 6 h were considerably higher in 2000 than in 2015. Thus, the 2000 event was characterised by stronger rain intensity, which caused flash floods in some small basins of the region such as the Beltrame River, while the 2015 event by a greater rain duration, as confirmed by the 2-day rain, equals to 505.2 and 687.2 mm respectively. The latter value is particularly relevant, being more than five times the average October monthly rain, and about 60 % of the mean annual rainfall. As far as the exceptionality of return period of daily rain, the compared events were equals in terms of number of gauges recording return periods of 100/200 years and greater than 200 years. Nevertheless, 2015 event showed the number of gauges with return periods of 20/50 years and 50/100 years considerably greater than 2000 event (Figs. 3; 6c).

From a geographical point of view, both the event focused on the east regional sector, as the majority of damaging event affecting the region, generally caused by a barometric minimum moving from the Tyrrhenian Sea, to the South of Italy (Petrucci and Polemio, 2009). Nevertheless, the 2000 event affected almost the entire region in a spotty way, while the 2015 event mainly hit the southeast sector (Fig. 7). As regards the damage, the 2000 event affected an area larger than 2015 event, and accordingly, a greater number of phenomena caused damage (Fig. 6d) to the different types of vulnerable elements (Fig. 6e), even if services and roads and railways were the most affected types. This is confirmed by the amount of funds allocated to repair damage, larger in 2000 than in 2015 event.

The damage in 2000 event (Table 2) was strongly increased by the high number of victims (13) with respect to 2015 (1), even if all the victims of 2000 occurred in a single accident that could be avoided if the campsite was located far from the river and/or the event was forecasted in advance allowing people evacuation.

According to the data gathered and indexes assessed to characterize both RE and DE, using expressions (1) and (2), the 2000 event was classified as catastrophic while 2015 as major catastrophic event (Fig. 8). This result is mainly due to the value of the Rscore, which is higher in 2015 than in 2000, while an opposite behaviour has been obtained for the Dscore. In practice, at daily scale, the rainfall event was more severe in 2015 than in the 2000, even if damage caused by the 2000 event was higher than in 2015 case. One factor that can explain this difference is the management of the events, especially in terms of alert and evacuation measures. Actually, in 2000, Civil Protection in Italy had an organisation different from nowadays, both at national and regional level: there were any alert system for meteorological events, and thus no alert for possible damaging rain was issued to the central-southern sector of the region before the event. The 2000 event represented a break: after the Soverato event, the Italian government recognised the urgent need of a more strength and effective organisation of Civil Protection regional units, based on the support of real-time meteorological data and effective rainfall forecasting systems. In fact, after this event, the Civil Protection Ordinance no. 3081 was issued and it established the Multi-Risk Functional Centre (Regional Agency for Environment Protection), a regional office in charge to systematically collect and elaborate data measured by hundreds of automatic weather meteorological stations of the regional gauges network. In 2015, the Multi-Risk Functional Centre, after the information transmitted by the National Office of Civil Protection and the outcomes of their meteorological data elaborations, transmitted official bulletins to Civil Protection, forecasting the intense rain approaching. Basing on this information, the Civil Protection Unit alerted majors of the municipalities where rain was forecasted, allowing them to be prepared to the emergency management and thus avoiding greater damage.

The Civil Protection rainfall forecasting systems is then paramount in a region such as Calabria, which is characterized by small basins and influenced by orographic precipitation that can cause flash floods. In fact, currently, predictions of flash floods rely predominantly on the weather forecast, although the time and the place where heavy rain may occur are still difficult to predict. Indeed, meteorological forecast gives the overall information that a given region can be affected by heavy rainfall causing flash floods in small basins. In order to overcome such a problem, after 2001, in Calabria, when a meteorological alert is issued, each of the majors of the municipalities involved by the alert starts a monitoring of the flash flood prone basins and the critical points belonging to their municipality, such as bridges and underpasses, in order to avoid causalities.

This work presents a comparative analysis of two events, both affecting Calabria region (southern Italy) at a distance of 15 years, on 2000 and 2015. In particular, a methodological approach is proposed, by organising a systematic comparative analysis of all the qualitative and quantitative features that are used to describe the analysed events, even using severity indexes introduced in previous works. The same approach can be used to describe and compare similar events occurred in different countries and epochs.

The analysis highlight the following points:

The 2000 event was classified as catastrophic, according to the classification proposed by Aceto et al. (2016), mainly due to the exceptionality of the damage.

It presented the highest rain intensities in 1, 3 and 6 h, concurring to trigger flash floods, storm surges and landslides.

The 2015 event was classified as major catastrophic, according to the classification proposed by Aceto et al. (2016), mainly due to the exceptionality of the daily rain.

It presented the highest total amount of rain and the highest values of daily intensity on a larger area: the two days rain reached five times the regional October monthly average, mainly triggering floods and storm surges