A hydrological analysis of the 4 November 2011 event in Genoa

Transcription

1 Nat. Hazards Earth Syst. Sci.,, 7 7, 0 doi:0.9/nhess--7-0 Author(s) 0. CC Attribution.0 License. Natural Hazards and Earth System Sciences A hydrological analysis of November 0 event in Genoa F. Silvestro, S. Gabellani, F. Giannoni, A. Parodi, N. Rebora, R. Rudari, and F. Siccardi, CIMA Research Foundation, Savona, Italy ARPAL Regional Environment Protection Agency, Genova, Italy University of Genova, Genova, Italy Correspondence to: F. Silvestro (francesco.silvestro@cimafoundation.org) Received: 8 April 0 Revised: 9 June 0 Accepted: July 0 Published: September 0 Abstract. On November 0 a flash flood event hit area of Genoa with dramatic consequences. Such an event represents, from meteorological and hydrological perspective, a paradigm of flash floods in Mediterranean environment. The hydro-meteorological probabilistic forecasting system for small and medium size catchments in use at Civil Protection Centre of Liguria region exhibited excellent performances for event, by predicting, 8 h in advance, potential level of risk associated with forecast. It greatly helped decision makers in issuing a timely and correct alert. In this work we present operational outputs of system provided during Liguria events and post event hydrological modelling analysis that has been carried out accounting also for crowd sourcing information and data. We discuss benefit of implemented probabilistic systems for decision-making under uncertainty, highlighting how, in this case, multi-catchment approach used for predicting floods in small basins has been crucial. Introduction In course of a about two weeks, from October to November 0, two devastating flooding events affected Liguria region in norrn Italy, causing death of 9 people and damages of tens of millions of euro to infrastructures, buildings, private and public goods. Some small towns changed ir appearance and it has been estimated that, in some cases, several years are needed to return to conditions prior to flood. The two events had some similar characteristics: y were both characterized by formation of a well organized, very intense and localized finger-shape mesoscale convective system (MCS) like precipitation structure embedded within a general stormy synoptic wear scenario developed by a large Mediterranean perturbation. These convective fingers remained stationary for a significant number of hours on same area (of few square kilometres) pouring very high quantities of rainfall. The meteoradar images, for two events, recorded very similar shapes of such small intense structures. They both developed on sea some tens of kilometres from coast. They were very narrow and elongated towards mainland Appennine range. Moving few kilometres outside shadow of finger-shaped MCS, rainfall depth, at event time scale, was reduced by a factor two or three. The hydrological consequences in terms of basin response were dramatic. Many creeks overflowed ir banks and a large number of mud flows and landslides occurred in quite narrow area hit by each event. The technical authorities in charge of hydrometeorological forecast for Liguria region predicted scenarios with a lead-time of two days. The regional civil protection issued maximum level of alert for most of region including catchments eventually hit by events. Without above mentioned warning number of casualties would have been very high, given intensity and suddenness of hydrological events. The issued alert might have reduced casualties in both events, but in end death toll was unacceptably high. An efficient forecasting system and a skilled civil protection authority issuing general warning, is not enough to avoid casualties. If population does not adequately perceive risk condition, at all levels, from elected authorities to citizens, and do not accept to behave differently than usual in case of alert, it is very difficult to reduce human losses to zero. Published by Copernicus Publications on behalf of European Geosciences Union.

2 7 F. Silvestro et al.: A hydrological analysis of November 0 event in Genoa Fig.. Syntic scheme of alert levels in Liguria region and related expected consequences. For sake of synsis this work is focused on November event. The previous event, on eastern Liguria, had same characteristics. The November event affected central part of Liguria region and city of Genoa. The Bisagno creek and one of its small tributaries, Fereggiano brook, developed most severe ground effects. In following paragraphs we run over chronicle of this case study from a hydrological perspective. The results of forecasting system are shown, toger with interpretation of decision makers. The usefulness of a probabilistic flood forecasting chain that contemplates also multi-catchment approach (Siccardi et al., 00; Silvestro et al., 0) is made evident. The event s ground effects are also estimated by using crowd sourcing information and data, in order to check, also by use of recorded ground effects, granularity of atmospheric event. The technique of crowdsourcing was quite useful in defining extension of ground effects. In this paper crowdsourcing is understood as process that involves outsourcing tasks; in this case task of documenting freely an event to a distributed group of people, in this case population of a city or part of a city affected by a flood. It is not a traditional post flood survey, by interviews or collection of images (Marchi et al., 009). Crowdsourcing mainly relies, in nearly real time, on images or videos distributed on Internet by real time eyewitnesses. The article is organized as follows: in Sect. study area and hydrometeorological forecasting system are described, Sect. provides a meteorological description of event, results of forecast system and chronicle of facts in terms of hydrological modelling and ground effects evaluation. In Sect. discussion and conclusions are presented. Methods Figure Fig.. Study. Study area. In area. different In different grey tones grey five tones alert sub-regions five alertof sub-regions Liguria are shown, toger of with Liguria watershed are shown, of main toger basins. with watershed of main basins. regional offices. In case of Liguria region FC is local Hydro-Meteorological Monitoring Centre (CFMI-PC). CFMI-PC is in charge of hydrometeorological forecasts. It also carries responsibility of nowcasting and monitoring rainfall events. The civil protection of Liguria region has two levels of warning for hydro meteorological risk: Alert and Alert. The Alert is most dangerous level and related scenario carries really severe rainfall events with associated flooding (Fig. ). The issuing of an Alert message contemplates that all actors of civil protection (and mainly municipalities) implement a number of predefined procedures for safeguard of people and goods. The population is requested to behave following a series of simple and practical measures for auto-protection. This level of alert presumes that a very severe event is predicted and it is issued for whole Liguria region no more than times in a year. Liguria is divided into five alert sub-regions (Fig. ), considered homogeneous from a meteorological point of view. They are divided into two groups: one group, south of Apennines divide, has three sub-regions with basins that have ir outlets in Tyrrhenian Sea. The or group has two sub-regions, y include head basins of greater catchments that form Po River, draining to Adriatic Sea. The average size of catchments included within alert sub-regions is in order of ten-one hundred km on south and slightly larger on north... The civil protection system for hydro-meteorological risk forecast In Italian civil protection system, regional offices have responsibility for issuing alert messages related to natural hazards including severe precipitation events. Technical/scientific units, named Functional Centers (FC), assist. The meteorological networks Three different local raingauge networks, with different timing and different working time, were available during this event, toger with precipitation estimation available through Italian National Radar Composite managed by Italian Department of Civil Protection (DPC). Nat. Hazards Earth Syst. Sci.,, 7 7, 0

3 F. Silvestro et al.: A hydrological analysis of November 0 event in Genoa 7 The first raingauge network available, named OMIRL, is official network managed by Civil Protection Agency of Liguria region and part of Italian Raingauge Network of Italian Civil Protection Department (Molini et al., 009). This system provides raingauge measurements with a timestep of min and an average density of raingauge/0 km and a total number of about 0 instruments. The municipality of Genova has set up second network available; it is quite dense but it covers only city and its inland, timestep is min. The third network available is network set up and managed by a non-professional meteorologists association named LIMET (Associazione Ligure di Meteorologia) in Liguria region. LIMET provides raingauge observations freely available over Internet at Web-Site timestep is from to min. The density of this network is highly variable across Regional territory with denser spots in urban areas where Internet high-speed connection and availability of power supply make it easier to instal raingauges. The three networks use different tools, OMIRL uses professional, WMO compliant (accuracy %), systematically maintained raingauges, while LIMET and Municipality networks are composed mainly by low-cost, semiprofessional, albeit reliable raingauges with an accuracy in range %. Despite differences in instrumentation, three networks generally give reliable measurements. Data for rain gauge stations of LIMET and municipality networks located in Fereggiano basin have been checked and corrected for systematic errors, after event, by WMO/CIMO Lead Centre B. Castelli on Precipitation Intensity at University of Genoa. From this point on we will call in following N first ground network, N second and N third, whilst we will call RC radar composite. Observations from N, N and N will be used alone and toger for a better definition of total amount of rainfall that hit Bisagno creek and Fereggiano brook. The Italian National Radar Composite integrates regional system, made up of ten C-band radars (five of m polarimetric) and one transportable X-band polarimetric radar, two systems owned by Italian company for air navigation services (ENAV), and three managed by Meteorological Department of Italian Air Force (AMI). The composite is provided both in terms of radar reflectivity, surface rainfall intensity and surface rain depth with a time step of min and a spatial resolution of km. During Genoa event, unfortunately, closest radar of network was not running due to technical problems. The observations made by or radar covering area gave estimates of spatial structure of event. Due to long distance and to attenuation, it was not possible in real-time to correctly define precipitation amount.. The hydrometeorological forecasting chain As described in Silvestro et al. (0), precipitation forecast for Liguria region is provided by a number of numerical wear prediction models (NWP) and interpreted by expert meteorologists. The experts merge output of different meteorological models ( so called poor man ensemble ) with ir own experience and provide quantitative precipitation forecast on alert sub-regions on predefined time windows. For each alert sub-region a different quantitative precipitation forecast is made. This kind of expert forecast is locally named subjective forecast. The subjective precipitation forecast is used as input into operational probabilistic hydro-meteorological forecasting chain (Siccardi et al., 00; Silvestro et al., 0). The or components of hydro-meteorological chain are downscaling module RainFARM (Rebora et al., 00), and hydrological model DRiFt (Giannoni et al., 00; Gabellani et al., 008). The downscaling module produces high-resolution precipitation fields, say one hundred, by preserving information at large scale derived from a quantitative precipitation prediction and it is in this way able to generate one hundred universes with small scale structures of precipitation that are consistent with radar observations of mid-latitude precipitation events (Rebora et al., 00). Each of one hundred precipitation fields is n used as input into hydrological model of each catchment, down to scales of square kilometer, in order to generate one hundred streamflow scenarios on each catchment. The results are post-processed producing a flood prediction following two different approaches: single-site and multi-catchment (Siccardi et al., 00). The single site approach is applied to all modelled basins, but is mainly useful for those basins that have an area larger than km (Silvestro et al., 0). In this case, probability that a certain flow threshold (or flow with given return period T ) could be exceeded is directly evaluated. For smaller basins a different approach is followed, because ratio between reliable spatial-temporal scale (l met, t met ) of meteorological forecast and scales of hydrological processes (l hydro, t hydro ) is too large (O(l met /l hydro ) > 0 ). As a consequence, forecasting procedure does not allow for any discrimination between different spatial localization. We thus cannot consider every single basin as an independent entity but we consider all basins toger inside domain of l met size. In case of Liguria region forecasting chain, domain of l met size is assumed to be alert sub-region. For each alert sub-region a different forecast is produced. The procedure evaluates probability that, in at least one basin belonging to that specific alert sub-region, flow with given return period T will be exceeded. The procedure does not specify which basin will be stricken, due to uncertainty associated to meteorological forecast. Such a procedure Nat. Hazards Earth Syst. Sci.,, 7 7, 0

4 7 F. Silvestro et al.: A hydrological analysis of November 0 event in Genoa Figure. Finger-like isolated auto-regenerating MCS in front of Genoa coast on Fig.. Finger-like isolated auto-regenerating MCS in front of November th at :0 UTC (Maximum Vertical Reflectivity, RC). Genoa coast on November at :0 UTC (Maximum Vertical Reflectivity, RC). is called a multi-catchment approach (Siccardi et al., 00) Fig.. Results of operational probabilistic flood forecasting and represents an essential paradigm in case of flood forecasts in very small basins both for scientific and political Figure. Results of operational probabilistic flood forecasting chain with subjective chain with subjective precipitation forecast as meteorological input Bisagno creek near mouth (90 km ). The lower panel levels involved in civil protection decision mechanism; precipitation it shows forecast as 80 meteorological % confidence input. intervals Bisagno of creek forecast near discharge mouth and (90 km ). The has been crucial element in prediction of event lower of panel forecast shows peak 80% flows. confidence On intervals x-axis, of forecast UTC discharge time reported. and forecast Note peak November 0 in Genoa. flows. On x that axis lower UTC time confidence is reported. bound Note is that nearly zero. lower The confidence top right bound panel is nearly shows exceedance probability of flow with return period T as zero. The top right panel shows exceedence probability of flow with return period T as function of T for single site approach (Bisagno catchment). For The case study 7 function of T purposes for single of comparison, site approach (Bisagno top left catchment). panel shows For purposes exceedance of comparison 8 top left probability panel shows of exceedence flow withprobability return period of T flow in function with return of Tperiod for T in. Meteorological description 9 function of T for multi-catchment multi catchment approach approach in in alert alert sub-region sub-region containing containing Bisagno Bisagno catchment. The heavy rainfall that triggered deadly flooding 0 of Genova catchment. was most powerful event embedded in a larger sys- tem that persisted on Sourn Europe for almost a week, from November to 8 November. The extra tropical macrostorm originated from extension of 0 Halloween ated with aforementioned finger-like, isolated and self- At γ -mesoscale, torrential event was associ- Nor easter that brought early heavy snowfall on Central and regenerating MCS convective cell triggered in Gulf of Eastern US in last day of October. This system, coming across Atlantic Ocean, regained strength by com- observed by RC (Fig. ). The cell started wandering along Genoa on night of November (0:00 0:00 UTC), as bining with remnants of tropical storm Rina ( eastern coast (0:00 09:00 UTC) of Liguria and finally 8 October, Yucatan and Cuba). Its precipitable water content was strongly enhanced. The system found also a very very high rainfall depth. was stuck over western portion of Genoa hills producing warm Mediterranean, with sea surface temperature (SST) anomaly values, with respect to last 0 yr, around.0. The forecast. C. This was responsible for enhancing latent and sensible heat fluxes from sea to lowest portion of The scenario that hydro-meteorologists of CFMI-PC planetary boundary layer. Both effects continued to increase drew for November 0 was that of a large perturbation with great amounts of precipitation that would af- potential severity of cyclone. An upper-level cold low, centred north-west of Ireland and extended meridionally to Iberian Peninsula, established an intense and moist November. Really intense rainfall on small spatial and temfect central and western part of region on and stream flow from south/souast impinging against Liguria Apennines ridge. The Apennines divide exceeds 00 m The probabilistic hydro-meteorological chain pointed out poral scales was also forecasted. in a very short distance from sea. The west of Liguria was that effects in terms of flooding could have been calamitous. mainly affected by widespread rains, while heavy rainfall was concentrated in centre of region over city Figure shows November hydrological prediction centre of Genoa, extending to west due to a strong convergence of flow of Sirocco wind (warm, humid airflow shows 80 % confidence intervals of predicted discharge for Bisagno creek (area of 90 km ). The lower panel moving to northwest). and predicted peak flows. The lower confidence bound is Nat. Hazards Earth Syst. Sci.,, 7 7, 0

5 F. Silvestro et al.: A hydrological analysis of November 0 event in Genoa 77 Figure. Bisagno creek (blue line and yellow watershed, Area=90 km ) and its tributary Fig.. Bisagno creek (blue line and yellow watershed, Area = 90 km ) and its tributary Fereggiano brook (red area, Area =. km ). Fereggiano brook (red area, Area=. km ). Figure. Location of three analyzed rain gauges: blue and red line shows Fig. Fereggiano. Location riverbed, in of red covered threereach. analysed The yellow rainline gauges: is watershed blue divide. and On red line bottom shows right corner Fereggiano Fereggiano watershed riverbed, is shown in toger red with covered its drainage reach. network. The yellow line is watershed divide. On bottom right corner Fereggiano watershed is shown toger with its drainage network. nearly zero. The spread is large and probability to exceed return period T = 0 yr is about 0 %. The top right panel shows exceedance probability of flow with return period T as function of T for single site approach. For purposes of comparison, top left panel shows exceedance probability of flow with return period T in function of T for multi catchment approach in alert sub-region containing Bisagno catchment. In latter probability to overcome T = yr return period is about 00 % and is 0 % for T = 0 yr. Moreover, a significant probability to exceed flow with return period T = 00 yr (about %) is also showed in prediction. This result warned civil protection officers that in one or more basins, within sub-region, probability of having a flood event was high. The results of forecasting chain and evaluation of CFMI-PC experts lead civil protection of Liguria region to issue highest alert level for alert area B two days in advance. carried under cover is suddenly reduced. The difference overflows and a sudden inundation of urban areas takes place. The Fereggiano brook is last tributary of Bisagno creek on left. The catchment has an area of about. km and it is densely populated (thousand for km ). In last 00 m Fereggiano flows under a cover. The estimated 0 yr return period discharge (about 8 m s ) can not be carried under cover (Provincial Authority of Genoa, 00). The maximum flow carried in condition of free surface flow is about 70 m s. When flow reaches intrados of cover, flow rate is reduced to about 0 m s. The growth of city over final reaches of Bisagno creek and Fereggiano brook during second part of century created hydraulic conditions prone to disastrous floods (Brandolini et al., 0)... Precipitation analysis. The facts.. Study area The most disastrous effects were due to flooding of Bisagno creek and Fereggiano brook (Fig. ). They both cross city centre of Genoa in a really densely urbanized area. The city develops along Bisagno creek for about 0 km inland. The final reach is constricted between two parallel avenues. Along last. km till to mouth river flows under a cover. The maximum flow carried without risk, i.e. in conditions of free surface flow, under cover is between 00 and 700 m s ( yr < T < 0 yr). For discharge values exceeding such limits, flow changes suddenly from condition of free surface flow to condition of pressure flow. If and when transition happens discharge In study area rain gauge density is really high because of presence of stations belonging to all three rain gauge networks described in Sect... We analysed data of three stations located in area where most dramatic effects occurred. As can be seen by Fig. y are close toger, re are about 00 m between N-Fereggiano and N-Fereggiano stations, and less than km between N-Fereggiano, N-Fereggiano and N-Gavette. Despite proximity, re are not negligible differences between N-Fereggiano, N-Fereggiano and N-Gavette simultaneous measurements (Fig. 7). This is due only to rainfall spatial variability. In fact, because such variability was large, a check on instruments was made after event by using methodology described in (Lanza and Vuerich, 009). Revised measurements are available and used for Nat. Hazards Earth Syst. Sci.,, 7 7, 0

6 78 F. Silvestro et al.: A hydrological analysis of November 0 event in Genoa 7 8 Fig. 7. November 0. Comparison of three rain gauges located inside or close to Fereggiano brook at an hourly resolution. On left y-axis hourly rainfall, on right y-axis accumulated rainfall. On x-axis, UTC time is reported. N-Gavette, courtesy of Liguria Civil Protection Agency; N-Fereggiano, courtesy of Lig- Figure Fig Depth-duration-frequency curves curves obtained obtained through through three applied three approaches applied (GEV approaches and Gumbel (GEV site analysis and Gumbel and Regional siteanalysis). analysisn and stands Regional for N- uria Civil Protection Agency and Municipality of Genova; and N- Fereggiano, courtesy of Ligurian Association of Meteorology. Gavette, Analysis). N-N stands Nfor stands N-N-Fereggiano for N-Gavette, N-N stands for N-N- Fereggiano. Figure 7. november 0. Comparison of three rain gauges located inside or close to Fereggiano brook at hourly resolution. On left Y axis hourly rainfall, on right Y axis accumulated rainfall. On x axis UTC time is reported N-Gavette, courtesy of Liguria Civil Protection Agency. N-Fereggiano, courtesy of Liguria Civil Protection Agency and Municipality of Genova. N-Fereggiano, courtesy of Ligurian Association of Meteorology. analysis. After correction of data from N and N stations, re is a very good agreement between two rainfall time-series. We use notation N-N-Fereggiano to indicate data valid for both raingauges available on Fereggiano brook. To provide an estimation of return period T of precipitation, three approaches have been followed. Two are single site analysis based on historical data series of N-Gavette station. The analysis has been carried out using Gumbel and GEV (De Michele and Rosso, 00; Burlando and Rosso, 99) cumulative distribution functions (CDF). The third approach is from a regional analysis made with data series from 90 to 99 (Boni 000). The Gumbel and GEV methods are applied using a more recent data sample of about forty years including precipitation data from 99 to 0 that where evidently not available when regional analysis was performed. The results are reported in Table and Fig. 8 and, as can be noted, estimation of T varies depending on applied method and its uncertainty is sometimes really high. The inspection of Fig. 8 clearly reveals that rain depth recorded for durations of one, three and six hours is a rare event, very rare if compared with regional analysis and slightly less rare in more recent single site analysis... Bisagno creek In order to check granularity of rainfall and relative rarity of event also from point of view of ground effects, hydrograph produced in Bisagno creek and Fereggiano brook has been simulated and compared with observations. The streamflow has been simulated using hydrological model DRiFt (Giannoni et al., 00; Gabellani et al., 008) that is operational in most of Liguria basins. The simulations have been carried out by setting three different configurations depending on rainfall data used as input to model: Configuration A: data from N network only (operational configuration). Configuration B: data from network N integrated with stations belonging to N or N network. The results are reported in Fig. 9 for Bisagno creek. The shape of three hydrographs are really similar, but 7 peak flows largely vary with different sets of input data; using only N data, peak flow is about 0 m s, while using measurements from N or N network, peaks are very similar, about 70 m s. In Table return periods of simulated peak flows derived using Boni et al. (007) are reported. The three peak flow values show T between 0 and 0 yr and y are close to limit value for flooding, this limit has been surely reached since Bisagno flooded. The latter is key information to confirming reliability of simulation... Fereggiano brook In case of Fereggiano brook streamflow has been simulated with a simplified rainfall-runoff model that couple SCS-Curve Number (CN) method (United States Department of Agriculture, 9) for infiltration estimation with Nash Instantaneous Unit Hydrograph (Nash, 97) for discharge calculation. The value of basin scale average CN is 7 and it has been estimated based on a distributed curve number map Nat. Hazards Earth Syst. Sci.,, 7 7, 0

7 F. Silvestro et al.: A hydrological analysis of November 0 event in Genoa 79 Table. Return period T of rain depth observed, during four standard durations by three rain gauges presented in Fig., as estimated following three different methods highlighted in text. Duration Station Rainfall T (yr) T (yr) T (yr) (h) (mm) Regional Analysis GEV (data from Gumbel (data from (data from 90 to 99) 90 to 0) 90 to 0) N-Gavette > > > 00 N-N-Fereggiano > > > Table. Return period of peak flows simulated for Bisagno creek using three different input configurations. Configuration Q (m s ) T (yr) N 0 = 0 N+N or N+N = 70 = 0 Figure 0: Fereggiano brook. Flow simulations. On left y axis streamflow is reported while on right y axis flow contribution per unit area is shown. On x axis UTC time is reported: Fig. 0. Fereggiano brook. Flow simulations. On left y-axis streamflow is reported while on right y-axis flow contribution per unit area is shown. On x-axis UTC time is reported. Fig. 9. Bisagno creek. Simulations with different meteorological networks. Q critic is maximum streamflow that can flow under cover. On left y-axis streamflow is reported while on right y-axis flow contribution per unit area is shown. On x-axis UTC time is reported. available for region. The value is quite high due to very high percentage of urbanized surface in basin. The simulations have been carried out setting only two different configurations (see considerations about rain gauges in Sect...) depending on rainfall data used as an input to model: Configuration A: data from N-Gavette station of official network which is located about. km from stations N-Fereggiano and N-Fereggiano, just on top of catchment used in configuration B. 7 Configuration B: data from N-Fereggiano station located on top of Fereggiano brook catchment. This is also representative of N-Fereggiano station. Figure 0 reports results of simulations. The differences between hydrographs are due to two main reasons: Rainfall depth on a min time step can significantly vary between near gauges due to spatial variability of precipitation. This has a great impact on streamflow simulation on small basins; As already pointed out, N-Fereggiano gauge 9 was just on centre of downpour. It recorded a rain depth significantly larger than N-Gavette station. As a consequence total runoff volume is larger. The estimated peak flow is in range of 0 m s 00 m s, while maximum flow that can be drained under cover is about 70 m s in case of free surface flow and about 0 m s in case of pressure flow. A large part of discharge did overflow above cover (Fig. ). A lot of videos taken by citizens show street above Nat. Hazards Earth Syst. Sci.,, 7 7, 0

8 70 F. Silvestro et al.: A hydrological analysis of November 0 event in Genoa Table. Return period of peak flows simulated for Fereggiano brook using two different rainfall inputs to feed hydrological model. Configuration Q (m s ) T (yr) N-Gavette = 0 00 < T < 00 N-Fereggiano = 00 T = 00 7 Fig.. Fereggiano brook. Flood scenario interpretation using hydrograph estimated by using different Rain Gauge stations. On x-axis UTC time is reported. Figure : Fereggiano brook. Flood scenario interpretation by using hydrograph estimated by using different Rain Gauge stations. On x axis UTC time is reported. cover filled with water flowing at high velocity. Six people died; many stores, restaurants and small firms were damaged as well as a large number of cars. The estimation of return period T associated with peak flow is highly uncertain. In river basin planning (Provincial Authority of Genoa, 00) flows with fixed T are estimated: Q(T = 0 yr) = 8 m s, Q(T = 00 yr) = 0 m s, Q(T = 00 yr) = 90 m s. The simulated peak flows have return periods shown in Table. As can be noticed in comparing Figs. 9 and 0, flow contribution per unit area of Fereggiano brook is considerably larger than that of Bisagno creek. This highlights 0 how most intense core of rainfall event was localized in a reduced area and it affected a very narrow ground strip. This also reflects fact that estimated return periods T of two peak flows are very different, even if both are very rare events. Considering Figs. 9 and, it is clear that percentage of hydrograph that overflowed is largely different in two cases. In case of Bisagno creek, inundation volume (around 0.7 mil. m estimated with configuration C) was % of total flood volume, while in case of Fereggiano brook (around 8. mil. m estimated with configuration B) percentage is 8 %. The experimental evidence and crowd sourcing information (videos, photos, tracks on buildings) compared with hydraulic simulations confirms that level of water that has been reached corresponds to discharge flow with T surely larger than 0 yr. Using both this information and considering two simulations equally reliable, we could state that occurred peak flow had reasonably a T higher than 00 yr. Anor hydraulic verification has been made by checking a video made by a citizen during flooding and available in near real time on The video clearly shows that level of water reaches height of banks, flow hits an obstacle and its kinetic energy transforms into potential energy generating a sort of water jet in vertical direction of about m. The velocity of flow (around that point) can be estimated: V = g h = m s. () Assuming a mean velocity of m s for cross section and estimating area of wet cross section from video and from direct inspection, it is possible to estimate flow as Q = 00 0 m s. () The video refers to times between :00 and :0 on November. By looking at Figs. 9 and it is possible to verify that simulated flow is quite realistic. Conclusions The flooding event occurred on November 0 that hit central part of Liguria Italian region has been analysed from a hydrological point of view. The results of official forecast system have been shown, demonstrating how y helped hydro-meteorologists and decision makers who, valuing also ir experience and knowledge of morpho-climatic characteristics of study area, recognized severity of event and issued maximum level of alert with large anticipation. The rainfall event was very localized, a large amount of precipitation fell in an area of few square kilometres with high intensities in about h. Flooding occurred on two small basins that cross city of Genova (Bisagno creek and Fereggiano brook) while in most of neighbouring catchments re have been only reduced effects. These kinds of events, which are quite common in Mediterranean area, can be hardly forecasted by Numerical Wear Prediction Systems; flood forecast can only be probabilistic, and use of downscaling methodologies remains essential. Moreover, for such small basins multi-catchment approach (Siccardi et al., 00) is fundamental. In such an environment, made by a collection of very small catchments with enhanced flood risk condition along ir streams, decision makers must become familiar with impossibility of identifying a specific catchment as target of warning to be issued: this has an impact both on type of warning to be issued and countermeasures to be prepared; more than Nat. Hazards Earth Syst. Sci.,, 7 7, 0

9 F. Silvestro et al.: A hydrological analysis of November 0 event in Genoa 7 that this concept must be built up in perception of citizens that ultimately interpret flood warnings and apply ir auto-protection measures. The facts on Fereggiano brook evidenced benefit of applying multi-catchment approach during prediction. The multi-catchment approach (see Fig. ) in fact evidenced non-negligible probabilities of exceedance of flow with T = 0 yr (P = %), T = 00 yr (P = %) and T = 00 yr (P = %), this dreadful forecast effectively occurred (see Fig. 0 and Table ). The ground effects have been reconstructed and described by using abundant rainfall observations, hydrological modelling and crowd sourcing information. In particular, simulations on Fereggiano brook have been verified by using videos made by citizens. The differences among simulations performed with official network alone or one where this was complemented with additionally available networks has highlighted anor limitation posed by spatial and temporal scale of se types of events. These differences in fact showed how difficult it is to observe a rainfall distribution consistent with real one and posed question of how to use in operational and official forecasting systems wealth of information that comes from common citizens operating monitoring networks in an organized way. This me has many implications in terms of QA/QC procedures on one side and in terms of liability on or. Despite forecast was correct and alert message had been issued with large anticipation, consequences have been terrible and event claimed six victims. Not all elements of Early Warning System worked effectively. Civil protection plans at municipality level were present, but y proved to be not adequate under real stress conditions. Civil protection plans are however essential in limiting human losses only if population has right perception of prefigured scenarios, so that y can implement properly and timely auto-protection measures once possible risk conditions are communicated. In this event many people ignored that an alert message was issued, and majority of people ignored meaning of that message and behaved as if no risk was pendent. Finally, a relevant number of people (probably not majority) acted inappropriately, as documented in many videos shared during event, enhancing ir exposure to imminent danger. The analysed event and its consequences remind us of what scientists sometimes tend to underestimate: that an Early Warning System does not stop at only timely and accurate forecasts, but all civil protection actors must also be adequately prepared and organized, orwise work of an efficient pool of forecasters and decision-makers can be nullified. The citizens are a fundamental element: y should be better aware of risks related to area where y live and be better prepared to face those risks. The citizens should not perceive mselves as simple customers of civil protection system, but as an active part of system itself. Along se lines, authors think that re is need to involve citizens in phase of designing procedures to be carried out in case of emergency. The increasing availability of handy technology to large public also poses challenge of how to exploit at least a part of citizens as watchers of occurring event as well as data suppliers. Acknowledgements. This work is supported by Italian Civil Protection Department and by Italian Region of Liguria. We acknowledge Liguria region, Municipality of Genoa and LIMET association for providing us with data from ir meteorological observation networks, and Italian Civil Protection Department for providing us with data of Italian National Radar Composite. We acknowledge WMO/CIMO Lead Centre B. Castelli on Precipitation Intensity at University of Genoa for providing us with checked and corrected data for rain gauge stations of LIMET and Municipality Networks located in Fereggiano basin. We are very grateful to all meteorologists and hydrologists of Meteo-Hydrologic Centre of Liguria region, for many useful discussions we had. Edited by: M.-C. Llasat Reviewed by: two anonymous referees References Boni, G.: A physically based regional rainfall frequency analysis: application to a coastal region in Norrn Italy, EGS Plinius Conf. on Mediterranean Storms, Maratea, October 999, 7, 000. Boni, G., Ferraris, L., Giannoni, F., Roth, G., and Rudari, R.: Flood probability analysis for un-gauged watersheds by means of a simple distributed hydrologic model, Adv. Water Res., 0,, doi:0.0/j.advwatres , 007. Brandolini, P., Cevasco, A., Firpo, M., Robbiano, A., and Sacchini, A.: Geo-hydrological risk management for civil protection purposes in urban area of Genoa (Liguria, NW Italy), Nat. Hazards Earth Syst. Sci.,, 9 99, doi:0.9/nhess--9-0, 0. Burlando, P. and Rosso, R.: Scaling and multi-scaling models of depth-duration-frequency curves of storm precipitation, J. Hydrol., 87,, 99. De Michele, C. and Rosso, R.: Guidelines for flood frequency estimation in North-Western Italy, FRAMEWORK Final Project, edited by: Rosso, R., Milano, 00. Gabellani, S., Silvestro, F., Rudari, R., and Boni, G.: General calibration methodology for a combined Horton-SCS infiltration scheme in flash flood modeling, Nat. Hazards Earth Syst. Sci., 8, 7 7, doi:0.9/nhess , 008. Giannoni, F., Roth G., and Rudari, R.: A procedure for drainage network identification from geomorphology and its application to prediction of hydrologic response, Adv. Water Res., 8, 7 8, doi:0.0/j.advwatres.00..0, 00. Lanza, L. G. and Vuerich, E.: The WMO Field Intercomparison of Rainfall Intensity Gauges, Atmos. Res., 9,, doi:0.0/j.atmosres , 009. Marchi, L., Borga, M., Preciso, E., Sangati, M., Gaume, E., Bain, V., Delrieu, G., Bonnifait, L., and Pogancik, N.: Comprehensive Nat. Hazards Earth Syst. Sci.,, 7 7, 0

10 7 F. Silvestro et al.: A hydrological analysis of November 0 event in Genoa post-event survey of a flash flood in Western Slovenia: observation strategy and lessons learned, Hydrol. Process.,, 7 770, doi:0.00/hyp.7, 009. Molini, L., Parodi, A., and Siccardi, F.: Dealing with uncertainty: an analysis of severe wear events over Italy in 00, Nat. Hazards Earth Syst. Sci., 9, 77 78, doi:0.9/nhess , 009. Nash, J. E.: The form of instantaneous unit hydrograph, Int. Assoc. Sci. Hydrol.,,, 97. Provincial Authority of Genoa: River basin planning of Bisagno creek, available at: pdb/bisagno (last access: January 0), 00. Rebora, N., Ferraris, L., Hardenberg, J. H., and Provenzale, A.: The RainFARM: Rainfall Downscaling by a Filtered Auto Regressive Model, J. Hydrometeorol., 7, 7 78, 00. Siccardi, F., Boni, G., Ferraris, L., and Rudari, R.: A hydrometeorological approach for probabilistic flood forecast, J. Geophys. Res, 0, D00, doi:0.09/00jd00, 00. Silvestro, F., Rebora, N., and Ferraris, L.: Quantitative flood forecasting on small and medium size basins: a probabilistic approach for operational purposes, J. Hydrometeorol.,,, 0. United States Department of Agriculture: S. C. S.: National Engeneering Handbook, Section, US Department of Agriculture, Washington, DC, 9. Nat. Hazards Earth Syst. Sci.,, 7 7, 0