Investigation of the weather conditions during the collapse of the Morandi Bridge in Genoa on 14 August 2018

On 14 August 2018, Morandi Bridge in Genoa, Italy, collapsed sending vehicles and tons of rubble to the ground about 40 m below and killing 43 people. Preliminary investigations indicated poor design, questionable building practices and insufficient maintenance or a combination of these factors as a possible cause of collapse. However, at the time of collapse, a thunderstorm associated with strong winds, lightning and 15 rain was developed over the city. While it is still not clear whether or not it played a role in this disaster, the present paper documents the weather conditions during the collapse and analyzes in detail a downburst that occurred at the time of the collapse a few kilometers from the bridge. The thunderstorm is analyzed using direct and remote measurements in an attempt to describe the evolution of the cumulonimbus cloud as it approached the coast from the sea. The detected downburst is investigated using a lidar scanner and the anemometric 20 network in the Port of Genoa. The paper shows that the unique lidar measurements enabled a partial reconstruction of the gust front shape and displacement velocity. The Weather Research and Forecasting (WRF) simulations, carried out with three different forcing conditions, forecasted the cumuliform convection at larger scales but did not accurately replicate the downburst signature at the surface that was measured by radar, lidar, and anemometers. This result demonstrates that the localized wind conditions during the collapse time could 25 not be operationally forecasted.


Introduction
Morandi Bridge in Genoa, Italy, named after its designer Riccardo Morandi, was built in the period [1963][1964][1965][1966][1967] and it collapsed on 14 August 2018 at 11:36 a.m.local Italian time (0936 UTC).The collapse caused 43 fatalities.Morandi was known for his unconventional cable-stayed bridges that featured an unusually low span-30 to-stay ratio.As for the Morandi Bridge, he often used pre-stressed concrete instead of steel for the stays.
Preliminary investigations after the collapse indicated poor design, questionable building practices and insufficient maintenance, or a combination of them, as possible causes of collapse.At the time of collapse, a violent thunderstorm was striking the city, but it is still not clear whether or not it played a role in this disaster due to the lack of meteorological measurements at the time of collapse in the vicinity of the bridge.Even if a 35 quantitative evaluation of weather conditions is not available, it is evident based only on the video available very close to the accident (released by the Italian Finance Guard on July 1, 2019 and taken from a security collapse are still under investigation and largely unknown, this paper intends to provide the needed information on the weather conditions at different spatiotemporal scales prior to and during the collapse. The rest of this paper is organized in the following manner.Section 2 describes the data and numerical simulations that were used in the analysis of weather conditions on 14 August 2018 over Genoa and the broader region.Section 3 analyzes the weather scenario prior to and during the bridge collapse with the emphasis on 125 thunderstorm that occurred on that day over Genoa.This section analyzes the meteorological precursors for the observed thunderstorm as well as the local scale observations of thunderstorm characteristics making use of weather station data (anemometer, thermometer, etc.), Doppler radar and lidar measurements.Section 4 describes the spatiotemporal evolution of the thunderstorm using the WRF simulations and presents several derived quantitative indices used to characterize the severity of thunderstorm winds.Section 5 provides a 130 concluding discussion on the most relevant findings presented in this study.

Meteorological Data
Data sources at different spatial scales have been used to describe this event from the synoptic to the local scale.
The synoptic meteorological conditions have been analyzed by means of the Global Forecast System (GFS) 135 analyses of the National Centers for Environmental Prediction (NCEP), available on a 0.5° × 0.5° grid every 6 h.The development of the cumulonimbus cloud that approached Genoa from southwest has been monitored through the meteorological radar of Liguria Region, which belongs to the Italian Meteo-radar Network and has a range of approximately 100 km.The associated lightning strikes have been recorded through the Blitzortung network (http://www.blitzortung.org/)as well as the LAMPINET system (Biron et al., 2006;Biron, 2009; De 140 Leonibus et al., 2010), managed by the Italian Air Force.The local weather conditions have been monitored using the meteorological stations of the Meteo-Hydrological Observatory of Liguria Region, operated by the Regional Agency for Environmental Protection (ARPAL) and the weather METAR station of the Genoa Airport.Lastly, the gust front evolution in time and space has been reconstructed by means of the anemometric stations available in the Port of Genoa managed by the Port Authority, while a lidar scanner property of the 145 University of Genoa has documented a downburst that occurred in the rear flank of the cumulonimbus cloud.
Table 1 lists the meteorological stations available in the area that was affected by the investigated thunderstorm.
With the exception of GEPOA and GEPVA stations, all other stations measure precipitation (P), but only three of these stations measure wind speed and direction (W).All available measurements are either averaged or cumulated over 1 h period.

150
Table 2 lists the anemometric stations available within the port area.Unfortunately, wind speed measurements were available with a resolution of 1 m s -1 only, whereas wind direction was available with 1° resolution.The sampling rate was about 4 s, which allows the evaluation of gust front movement in space and its evolution in time.However, higher sampling frequency is needed in order to precisely measure the actual maximum wind Anemometer 07, which is the closest one to the bridge, is 2.7 km southward from the center of the bridge, while the meteorological station GEBOL is 3.2 km northward.

WRF Model Setup
The Weather Research and Forecasting (WRF) model (Skamarock et al., 2008) is a compressible nonhydrostatic model with mass-based terrain-following coordinates that was developed at the National Center for 195 employed with a cycling update technique following the aforementioned procedure in order to assimilate radar reflectivity into the WRF model.The data assimilation was performed following the same configurations setup of Lagasio et al. (2019) and by using the modified direct operator presented in Eq. 11 in Lagasio et al. (2019).
All the simulations were performed with the same set of physical parameterizations that have already been successfully tested in the study of similar events (Fiori et al., 2017;Lagasio et al., 2017;Lagasio et al., 2019).

200
The surface layer was modelled using the MM5 scheme that uses the stability functions from Paulson (1970), the next 20 min it moved in the northeast direction, and between 0930 and 0940 UTC it was situated above the city.The south cell, on the other hand, stayed stationary during the same time period.One of the interesting features of the north precipitation cell is its deformation that is manifested as deeper propagation of the central precipitation zone in the north direction.Between 0920 and 0940 UTC, the isoecho of 40 dBZ propagated 245 approximately 6800 m inland (i.e., at 0920 UTC, this isoecho was at the coastline).Therefore, the northward propagation velocity of precipitation cell that traversed Genoa around the bridge collapse time was about 5.7 m s -1 .The northward movement of the precipitation cell is due to the orographic channeling that is inferred from Fig. 6.The mountains on both sides of two valleys that traverse Genoa rise to almost 1000 m above sea level, while the bottom of the valley is below 100 m.The location of the Morandi Bridge in the west valley of Genoa 250 is shown in Fig. 6 in respect to the precipitation cell.The influence of orography and river valley on thunderstorm propagation was investigated in Ćurić et al. (2003).The authors concluded that thunderstorms influenced by orographic effects are more compact in comparison to thunderstorms over a flat terrain.This difference is caused by the smaller and unidirectional (only along the direction of valley) supply of low-level moisture in the orographic case.However, the numerical study of Ćurić et al. (2003) was carried out for a 255 different geographic region (the Balkan Peninsula) and more research is needed on the influence of orography on gust fronts and thunderstorm propagation, in particular in coastal regions.
The 5-min Surface Rainfall Intensity (SRI) estimates from the national mosaic of the Italian Meteo-Radar Network is here used to evaluate the storm cell motion towards the coast in the period between 0900 and 1000 UTC.The application of a region growth algorithm to the binary image rain/no rain (Gonzalez and Woods, 260 2002) showed that the cell was isolated from the rest of the precipitation field observed at larger scales.The centre of mass of the precipitating cell was identified and the cell centroid displacement between subsequent images was used to evaluate wind speed and direction of the cell.
The results in Fig. 7 demonstrate that the precipitation cell was moving north-westward until 0925 UTC, when the cell reached the coast.Once landed, the cell intensity decayed, its shape from approximately elliptical 265 changed to an arrowhead-like shape pointing northward as it wedged along the Polcevera Valley and the translation speed lowered down to almost 2 m s -1 and backed almost 90° from 0925 to 0940 UTC.The translational speed of the cell approaching the coast were on average around 4 m s -1 with a peak speed above 6 m s -1 between 0915-0925 UTC.This pronounced convective activity above and south from Genoa was also detected by the Blitzortung network 270 for lightning strikes detection (Fig. 8a).We observe that the lightning strikes were organized along a southnorth line that stretched out from the north part of Corsica to north of Genoa.These results show that the lighting was predominantly concentrated over the convective line situated above the sea and the number of lightning strikes above Genoa was less significant.
The position of lightning strikes measured through LAMPINET has an accuracy of 500 m (Fig. 8b,c).In the 275 bridge collapse hour, lightning strikes were mostly recorded over the sea (Fig. 8b) and later over the coast and east part of Genoa (Fig. 8c).The largest number of strikes occurred between 09:35 and 09:40 UTC at a distance of 4-5 km from the bridge, in the area around the Old Port of Genoa.The spatial extent of the cumuliform clouds (Fig. 5) was predominantly in the west-east, which is not necessarily the dominant direction of the area characterized with the highest frequency of lightning strikes around Genoa (Fig. 8b,c).We find that the lightning 280 strikes were not observed in the vicinity of the Morandi Bridge (thick black line).Therefore, the possibility that a lightning strike triggered the bridge collapse is minimal.

Local Observations
The standard METAR meteorological measurements from the Air Force weather station of the Genoa Airport are shown in Fig. 9.The temporal resolution of data is not consistent throughout the records, but on average the 285 wind velocity and air temperature data are available at every 20-30 min, while the availability of pressure data is at every approximately 2.5 hours.Figure 6a shows that the wind direction at the time of bridge collapse was continuously and abruptly changing for almost 360° (i.e., from 350° to approximately 10°) in the counterclockwise direction.At the exact time of bridge collapse, the wind was blowing from 140° (southeast).
Around the same time interval, the mean wind speed increased from approximately 2 m s -1 to 7.7 m s -1 at the 290 time of collapse (Fig. 9a).Afterwards, the wind speed dropped down to about 4.5 m s -1 at 1000 UTC.Bearing in mind that the values are 10-min averages, a mean velocity of ~8 m s -1 demonstrates that the event was characterized with relatively low wind speeds, as this value corresponds to return periods lower than 2 years in Liguria (Zhang et al., 2018).This is further confirmed by observing the recorded wind gust of 16 m s -1 at 0950 UTC (14 min after the collapse), which is the overall maximum value that was measured from 0940 to 0950 295 UTC.Note that, in the time interval between 0940 and 0950 UTC, the mean (maximum) wind speed values measured at ARPAL stations GEPOA and GEPVA (see Fig. 2) were 7.9 (16.1) m s -1 and 5.9 (15.5) m s -1 , and their prevalent wind direction from south and west-southwest, respectively.
An important validation that the event was a thunderstorm downburst comes when the wind velocity records are analyzed in conjunction with air temperature and pressure data (Fig. 9b,c).In this particular case, the air 300 temperature is a more useful quantity due to the higher temporal resolution of these data.Figure 9b shows that a rapid decrease of temperature occurred simultaneously with the increase of surface wind speed and abrupt shifts in wind direction (Fig. 9a).A sudden decrease of temperature at 0850 UTC coincides precisely with both the increase of wind speed and abrupt change of wind direction.In the next hour (until 0950 UTC), the temperature dropped for 4°C.This relationship between air temperature and wind velocity demonstrates that 305 the increase of wind speed was caused by a spread of cold air in the form of a gust front that originated from the thunderstorm cloud (Charba, 1974;Wakimoto, 1982;Mueller and Carbone, 1987;Droegemeier and Wilhelmson, 1987;Lompar et al., 2018).The subsequent increase of surface air pressure (Fig. 6c) is also in accordance with the kinematics and dynamics of gust fronts, but the low temporal resolution of data prevents a more comprehensive analysis.

310
The accumulated surface precipitation in the collapse hour increased in the eastward direction (Fig. 10).While precipitation was not observed in the west parts of the port, the east side received 33.8 mm h -1 of rain between 0900 and 1000 UTC.The weather station situated in the same valley as the Morandi Bridge (GEBOL; Fig. 10) and approximately 3.2 km north of the bridge received 12 mm h -1 of rain.Therefore, the zone of the strongest precipitation coincided with the area characterized with the highest activity of lightning strikes (Fig. 8).We 315 also observe that the heavy rain (i.e., > 10 mm h -1 ) was localized at the GEBOL, RIGHI and CFUNZ stations (Fig. 10) thereby indicating that the Morandi Bridge probably received around 10 mm h -1 of rain in the collapse hour.Hail and graupel were not observed in this period at any of the weather station in the Genoa region.
Here, we discuss the high-frequency velocity records from the anemometers located along the Port of Genoa.
In the hour centered around the collapse time (0936 UTC), for the majority of stations the wind direction was 320 continuously changing in the counterclockwise direction starting around the north direction at the beginning of the hour and returning back to the north direction at the end of the hour (about 1000 UTC) (Fig. 11).Although the counterclockwise temporal change of wind direction is observed at all eight anemometers, the shift is less pronounced at 08, 09 and 11.These three anemometers are concentrated in the west part of the port area and  Similarly, the wind speed records also exhibit noticeable differences between the west and east parts of the port (Fig. 11).The differences between the peak wind speed close the bridge collapse time and the wind speed at the beginning of that hour are higher at the anemometer stations located in the east part of the port (Fig. 11eh).The velocity time series in the west and central parts of the port (Fig. 11a-d), on the other hand, show pronounced nonstationary signature prior to the velocity peak that occurred around the collapse time.Figure 11 335 shows clearly that the maximum wind speed, which occurs before 0936 UTC for stations 07, 08, 09, 11, and after the collapse time for stations 01, 02, 03, 16, slightly shifts in time from the westernmost to the easternmost anemometer.This can be interpreted like the signature of the gust front passage that follows the cloud translation from south-west to north-east over approximately 15 km of coast.
Overall, these results show that the thunderstorm produced a macroburst that originated over the sea and 340 approached the anemometers as well as the bridge from prevailingly south direction.Recently, Burlando et al. (2018) also demonstrated that the majority of thunderstorm downbursts in this region are generated over the sea and then advance towards the shore.A similar trend in terms of downburst formation and movement was also reported by Burlando et al. (2017) for a downburst event in the Port of Livorno, Italy, on 1 October 2012.As already discussed for the METAR measurements presented in Fig. 9, the macroburst was relatively low-345 intensity as the maximum velocity recorded by all stations was in the order of 15 m s -1 .

Lidar Gust Front Detection and Analysis
Approximately 15 min prior to the bridge collapse, i.e. between 0920 and 0935 UTC, the thunderstorm downburst was also captured by the Doppler lidar installed in the Port of Genoa (Fig. 12).Note that the timing is coherent with the maximum wind speeds recorded by anemometer 11, as shown in Fig. 11.The lidar scanned 350 four different elevation angles (#) that are between 2.5° and 10° above the horizontal plane with a 2.5° increment.These results are particularly interesting as Fig. 12 is among the first few sets of published lidar velocity measurements of a thunderstorm downburst.Unfortunately, the heavy rain that occurred right behind the gust front of the macroburst affected the lidar measurements reducing its range of acquisition from about 5 km at 0915 UTC to 1-2 kilometers at 0930 UTC.

355
This thunderstorm downburst was characterized with velocities exceeding 20 m s -1 (Fig. 12).The highest positive velocities (i.e., towards lidar) are observed at the lowest # and between 0915 and 0923 UTC.In addition, Fig. 12 shows a propagation of the region characterized with the maximum velocity towards lidar (observed for all # angles).The direction of the maximum velocity is from 150° until approximately 0923 UTC (Fig. 12i) and then it starts shifting in the clockwise direction eventually reaching ~170° (Fig. 12k).The 360 width of this region is about 3.5 km.The azimuthal orientation ($) of the zone with the maximum velocity seems to be independent of height which indicates that the advancing gust front is not (significantly) inclined in the # − $ plane.However, it is particularly important to note that with increasing #, this region of the maximum velocity is inclined towards lidar.For example, when benchmarked against Fig. 12a (# =  and # = 7.5°, respectively.Moreover, the same trend is found in the next set of four velocity slices shown in Fig. 12e-h.That is, at # = 2.5° (Fig. 12e) the front is approximately 2000 m away from lidar, while at # = 10° (Fig. 12h), the front is already above lidar (or within 300 m away from the instrument).The intermediate 370 distances between these two are found at the other two elevations.
This displacement of the region of the maximum velocity with height is a characteristics of the radially advancing gust front of cold air in front of the thunderstorm cloud (Charba, 1974;Wakimoto, 1982;Mueller and Carbone, 1987;Droegemeier and Wilhelmson, 1987;Lompar et al., 2018).Further, Fig. 12 enables an estimate of the displacement velocity of the gust front towards lidar.This analysis is shown in Fig. 13 that 375 shows the height of the gust front (black dashed line in a-g) above lidar at different radial distances from the instrument along $ = 151°.For the height corresponding to 120 m ASL, (i.e., 115 m above lidar level + 5 m ASL, which is the height of the lidar ASL), we estimate the displacement velocity (/ 0 ) of the gust front to be / 01 = 6.3 m s 61 towards the instrument.A similar displacement velocity, / 07 = 6.9 m s 61 , is obtained at 70 m ASL (i.e., 65 m above lidar).Thus, an average displacement velocity is evaluated to be / 0 = 6.6 m s 61 .

380
Note that this velocity is about 1-2 m s -1 larger than the estimated velocity of the north-eastward propagation of the precipitation cell in Doppler radar images (Fig. 6).The displacement velocities calculated above are very similar to the one estimated by Mueller and Carbone (1987) by tracking radar reflectivity and Doppler velocity of a gust front associated to a thunderstorm outflow measured in Denver, Colorado, in 1984.They obtained a value of 6.9 m s -1 .

385
When combined with the meteorological measurements from the Genoa Airport weather station (Fig. 9) and precipitation measurements in the area (Fig. 10), this displacement velocity can be used to estimate the mean height of the cold inflow (Charba, 1974).First, we estimate the mean air densities of the ambient and gust front air masses.The air density (9) is calculated using the equation of state for wet air: where : 0 and > are the pressure of dry air and water vapor (in Pa), < is the air temperature (in K), ; 0 = 390 287.058J kg -1 K -1 is the gas constant of dry air, and ; ?= 461.495J kg 61 K 61 is the gas constant of water vapor.Since the relative humidity during the investigated time period was 100%, > is equal to the saturation water vapor pressure (> F ) because dew point is equal to air temperature.The relationship between > F and air temperature can be expressed through the August-Roche-Magnus formula as (Lawrence, 2005): where > FG = 610.94Pa , L 1 = 7.625, L 7 = 243.04℃(Alduchov and Eskridge, 1996) and M is the air 395 temperature (in °C).Further, the pressure of dry air is calculated using the Dalton's law of partial pressures: 400 Then, we use the equation for displacement velocity (/ 0 ) of gravity currents (e.g., von Kármán, 1940;Middleton, 1966;Daly andPracht, 1968 Charba, 1974) to estimate the mean depth of the cold inflow: / 0 = RSTU 9 7 − 9 1 9 1 (4) where 9 1 and 9 7 are the densities of warm (ambient) and cold (gust front) air masses (in kg m -3 ), T = 9.8053 m s 61 is the gravitational acceleration at 44°N and sea level, R = 0.77 (Seitter, 1983;Droegemeier and Wilhelmson, 1987) is the constant that represents the ratio of internal to gravitational forces (Charba, 1974), and U is the unknown mean depth of the cold inflow (in m).The value of R is uncertain and other figures were proposed too (Middleton, 1966;Daly and Pracht, 1968).Solving Eq. ( 4) for U yields the value of U = 464 m.A value of 478 m is obtained if the air is assumed to be dry.For instance, Charba (1974) used Eq. ( 4) to estimate the displacement velocity of 22.7 m s -1 , but the measured value of U in his study was 1350 m.
Smaller values of / 0 and U were reported in Goff (1976).Besides the value of R, the largest uncertainty 410 associated with the above analysis is related to the usage of surface air densities in Eq. ( 4) instead of the mean air density along the height of gust front inflow.Unfortunately, temperature and pressure profiles at this location are not available for this event.Since the air density difference 9 7 − 9 1 is fairly constant with height (Charba, 1974) and 9 1 (W > 0) < 9 1 (W = 0), the obtained value represents the upper limit of U. The higher wind speeds associated with the gust front passage (Fig. 11 and Fig. 12) in comparison to displacement velocity of the gust 415 front (Fig. 13) are expected and reported elsewhere too (Charba, 1974).According to Goff (1976), the slowly moving gust fronts such as the one reported in this present paper, are usually associated either with intensifying storms and accelerating outflow or with dissipating storms and decelerating outflows with respect to the storm.
On the other hand, gust front velocities are usually the highest during the mature stage of a thunderstorm cloud.
The gust front leading edge was inclined in the direction of propagation (i.e., towards lidar) due to the increase 420 of wind speed with height in the outflow (dashed line in Fig. 13).The angle of inclination depends on density differences between ambient and gust front air masses and wind speed, among other factors.In the analyzed case, the inclination angle is -10.1° from the horizontal plane, with the minus sign indicating the negative inclination in respect to the measurement location.That is, the gust front surge line increases for 177 m km -1 , which is very similar to the result of Charba (1974), who obtained the value of 150 m km -1 for a gust front that 425 occurred in central Oklahoma, United States.This comparative analysis indicates that the surface roughness might not be the dominant factor that governs the slope of the gust front leading edge (i.e.., gust front surge line) because similar values are obtained for two profoundly different surface roughnesses.However, this subject deserves more research and larger sample of data in order to confirm this speculation.
Further reconstruction of this event is possible by applying various relationships between the height of cold 430 inflow (U) and other features of the gust front.According to a number of studies that investigated gust fronts at full scale or using physical experiments and numerical simulations (e.g., Goff, 1976;Charba, 1974;(Wakimoto, 1982;Droegemeier and Wilhelmson, 1987), the height of the leading edge ([; called gust front head) is usually . This further enables us to sketch the likely shape of the gust front that occurred over the sea approximately 15 min prior to the bridge collapse (Fig. 14).The characteristic heights are indicated in figure 435 with the height of 159 m being the first lidar slice at # = 10° (see Fig. 13).The height of the gust front nose (265 m) is estimated from Charba (1974), where the author found that this height is approximately U/1.8.
However, other relationships are reported in Goff (1976); hence there is uncertainty associated with this value.
Note also that the height of the gust front nose is not the height of maximum horizontal wind speed which is however much lower and closer to the ground, as also commented before according to the fact that the highest especially for small-scale phenomena like single cell thunderstorms.Some deep convective spots, with VMI values above 40 dBZ, occur in all the three WRF simulations at 0830 and 0955 (Fig. 17d,g,j and e,h,k), but these structures are rather small, organized along narrow stripes from southwest to northeast, and they do not 485 resemble the structures in Fig. 17a,b whereas they are more similar to the convection that develops later after noon (Fig. 17c at 1325 UTC,).WRF-IFS at 1240 UTC seems to be able to produce a convective storm with better spatial agreement with the one observed four hours before at 0830 UTC (Fig. 17a), producing wind speeds as high as 16 m s -1 , which is the same maximum value observed in anemometric measurements at the collapse time (Fig. 11).

490
According to the latter consideration, we assume that WRF-IFS at 1240 UTC is the most reliable representation of the thermodynamic conditions to produce strong single cell thunderstorms similar to the one observed during bridge collapse.Hence, the corresponding skewT-logp diagram has been computed at position (Lon 8.8° E, Lat 44.4° N) in front of the Genoa harbor (Fig. 18).The diagram presents a significant MUCAPE (Most Unstable Convective Available Potential Energy) equal to 2934 J kg -1 , which is even higher at 0830 UTC (3427 J kg -1 ) 495 and 0955 UTC (3678 J kg -1 ) despite no strong convective cells are detected earlier in the morning.Precipitable water is 30 mm for the WRF-IFS experiment, thus corresponding to a scenario prone to the occurrence of high precipitation.This figure is similar to the amount of rain measured at the CFUNZ station in the east part of Genoa (Fig. 10).Considering the SWEAT and Showalter Index (SWI) values, it seems that in the WRF-IFS simulation both indicate a low probability for severe thunderstorm to occur.Similar conclusions can be drawn 500 when considering the Total Totals as well as the K and Lifted Index values.Wind shear and storm relative helicity (SREH) depict atmospheric conditions that are not prone to develop supercell-like structures.Finally, it is worth highlighting that the value of the Microburst Windspeed Potential Index (MWPI), which by definition ranges between 0 and 5, here is equal to a value of 3.2 that corresponds to thermodynamic conditions prone to the occurrence of downbursts (Pryor, 2015).

505
All considered, the current WRF simulations with three different forcing conditions were not capable of fully replicating the transient wind characteristics in the bridge collapse hour.WRF-GFS simulations were not able to produce a maximum wind speed of approximately 16 m s -1 , which was the wind gust measured at the nearby station of Genova Airport as well as the other stations in the port area around the collapse time (Fig. 9), while WRF-IFS produced a thunderstorm cell similar, in terms of intensity of wind speed and reflectivity, to the one 510 that occurred around 0930 UTC, roughly at the same location but a few hours later.This demonstrate that the WRF forecast of radar reflectivity, thermodynamic properties of the atmosphere at larger scales, and wind field at the local scale can be used potentially as a useful tool for determining the meteorological precursors and contributing factors to the development of thunderstorms in this region.Also, in perspective these simulations might be used to simulate gust fronts and downbursts at much finer scale adopting proper downscaling 515 techniques or coupling WRF with other higher resolution models (Chen et al., 2011;Solari et al., 2012;Burlando et al., 2007).

Conclusions
This article documents and analyzes the weather conditions prior to and during the collapse of the Morandi Bridge on 14 August 2018.This disastrous event that occurred at 0936 UTC caused 43 casualties.Since the 520 forensic investigation of collapsed bridge has not pinpoint the exact factors that caused the collapse, the goal of this paper is to provide the contributions in terms of observed and modelled weather conditions around the

535
In addition, this study presents the unique lidar measurements of downburst outflow in the form of gust front that advanced in ahead of the parent cumulonimbus cloud.Furthermore, we developed and applied a technique for gust front surge reconstruction using lidar measurements and theoretical relationships that analytically link the main parameters of this phenomena.The displacement velocity of gust front that was measured approximately 20 min prior to the bridge collapse was estimated to be 6.6 m s -1 and the height of the cold pool 540 behind the gust front head was calculated to be at 464 m.The front was advancing from the sea towards the land and the observed features were measured approximately 10-15 km south from the bridge.In addition, the lidar velocities showed multiple high velocity spots behind the leading edge of the gust front.

550
In conclusions, this paper demonstrated the existence of a severe thunderstorm in the Genoa region during the collapse of the Morandi Bridge on 14 August 2018.However, while the strong and non-synoptic wind conditions in the form of thunderstorm downburst and gust front, as well as precipitation and lightning, were all present in the minutes prior to and during the bridge collapse, this study does not firmly conclude that the severe weather was the major triggering factor for the collapse.However, we do acknowledge that the 555 documented weather conditions might have played some role in the failure of the Morandi Bridge.

Author contributions
MB supervised and designed the study, and was responsible for data collection.MB and DR performed data analysis and manuscript preparation and editing.DR performed literature review.ML and AP performed numerical simulations, analyzed numerical results and edited the corresponding sections in this paper.GB 560 analyzed radar reflectivity and storm motion.

Figure 2
Figure 2 shows the position of all meteorological and anemometric stations included in Tabs.s 1 and 2, as well as the METAR station and the lidar scanner location.Note that the lidar scanner is almost side by side with anemometer 11 and thus the two symbols overlap each other.The position of Morandi Bridge is also shown in blue.The bridge was located in a suburban and industrial area along the Valley of Polcevera stream, which has a south to north extension of approximately 10 km and riversides as high as 600 m ASL in the northernmost 170 https://doi.org/10.5194/nhess-2019-371Preprint.Discussion started: 25 November 2019 c Author(s) 2019.CC BY 4.0 License.https://doi.org/10.5194/nhess-2019-371Preprint.Discussion started: 25 November 2019 c Author(s) 2019.CC BY 4.0 License.
https://doi.org/10.5194/nhess-2019-371Preprint.Discussion started: 25 November 2019 c Author(s) 2019.CC BY 4.0 License.thebridge collapse, i.e. 0936 UTC, was approximately 180° at the 07 station.At the same time, in the east stations, i.e. 03, 01, 02, 16, the wind was blowing from the third quadrant, which is between 180° and 270°, whereas the west stations, i.e. 08, 09, and 11 recorded the wind direction from the second quadrant, between 90° and 180°.This spatial distribution of wind directions resembles the cloud-scale (~10 km) radial outflow produced by the gust front which was spreading from the cloud base. 330 2.5°), this region is closer to lidar in Fig. 12d (# = 10°) than it is in Fig. 12e (# = 2.5°), despite the fact that the velocity https://doi.org/10.5194/nhess-2019-371Preprint.Discussion started: 25 November 2019 c Author(s) 2019.CC BY 4.0 License.slice in Fig. 12e is later in time than the one in Fig. 12d.The continuous advancements of this leading front between the elevations of # = 2.5° and # = 10° are nicely observed in Fig. 12b and Fig. 12c for # = 5° 3) since > = > F (i.e., the relative humidity of 100%).Here, : is the measured (total) atmospheric pressure of wet air (in Pa).Plugging in the temperature drop of 4 K [from 296.15K (23.2°C) to 292.15 K (19.0°C)] and the pressure rise of 130 Pa (1.3 mb) into the above equations results in the air densities of cold and warm air being ~1.2024 kg m -3 and ~1.1833 kg m -3 , respectively.
440 velocities in Fig. 12 are observed for lidar scans with # = 2.5°.The slope of ~70° of the gust front edge above the height of the nose is adapted from Goff (1976), whereas the gust front head height (928 m) is estimated https://doi.org/10.5194/nhess-2019-371Preprint.Discussion started: 25 November 2019 c Author(s) 2019.CC BY 4.0 License.through the relationship [ ≅ 2U.The region behind the gust front head is characterized by pronounced turbulence due to the Kelvin-Helmholtz instability (Britter and Simpson, 1978) and therefore there is no a clear shape of the gust front edge in that region (notice that the blue dotted lines are not connected behind the gust https://doi.org/10.5194/nhess-2019-371Preprint.Discussion started: 25 November 2019 c Author(s) 2019.CC BY 4.0 License.
https://doi.org/10.5194/nhess-2019-371Preprint.Discussion started: 25 November 2019 c Author(s) 2019.CC BY 4.0 License.collapse time.Therefore, the analyses in this study are based on direct meteorological measurements from the weather stations in the area, wind velocity records from eight anemometers installed along the Port of Genoa, remote satellite and radar observations and Global Forecast System (GFS) analysis.The local weather station 525 of Genoa Airport recorded the wind gust of 16 m s -1 few minutes after the collapse time.The time records of air temperature, pressure and wind velocity from this weather station, as well as the higher frequency velocity measurements from the eight anemometers located along the coastline showed the well-known signature of transient thunderstorm conditions during the collapse hour.This observation was confirmed using the satellite (i.e.cloud top heights) and radar (i.e., composite radar reflectivity) data.The radar data showed the existence 530 of two cells close to Genoa with the strong radar reflectivity exceeding 55 dBZ.The satellite images, on the other hand, depicted a convective line that stretched in the south-north direction and extended from Corsica tonorther Italy (thus crossing Genoa).The event was also characterized with the high frequency of lightning strikes, but they were not recorded exactly in the close vicinity of the Morandi Bridge but a few kilometers east of it.
The numerical simulations were carried out using the Weather Research and Forecasting (WRF) model with three different initial and boundary conditions: (1) IFS (from the European Center for Medium-Range Weather 545 Forecast); (2) GFS (from the National Centers for Environmental Prediction) and (3) GFS-DA.The GFS-DA simulation uses the assimilated radar reflectivity data into the GFS analysis.While all three numerical simulations under-estimated the maximum velocity during the bridge collapse hour, the simulated radar reflectivity and thermodynamic indices demonstrated that these products can be used as precursor alerts for severe thunderstorm weather in this region.

Figure 1 .
Figure 1.Two frames of the video recorded by a security camera located nearby and to the west of the bridge taken during the collapse of Morandi Bridge (a) and immediately after (b).

Figure 2 .Figure 3 .
Figure 2. Topographic map of the Genoa region showing the position of Morandi Bridge (blue line), ARPAL meteorological stations (green circles), Port Authority anemometric stations (red squares), airport METAR station 795

Figure 4 .
Figure 4. Mean sea level pressure (contours) and tropopause height (shaded contours) over Europe from GFS 800

Figure 5 .
Figure 5. Cloud top heights above Italy and surrounding regions from the MSG (Meteosat Second Generation) 805

Figure 6 .
Figure 6.Radar reflectivity (dBZ) measured by the Ligurian Doppler radar.The panels (a) to (c) show three 810

Figure 7 .
Figure 7. Rain cell observed offshore the Port of Genova on 4 August 2018 between 0900 and 0955 UTC.The arrow 815

Figure 8 .
Figure 8. (a)Lightning strikes recorded between 0930 and 10:00 UTC on 14 August 2018 over Italy and the 820

Figure 9 .
Figure 9. Meteorological measurements from the Genoa Airport weather station on 14 August 2018: (a) wind speed (black line; primary _-axis) and wind direction (grey line; secondary _-axis); (b) air temperature at 2 m above ground; and (c) sea level pressure at 2 m above ground.In (a), the asterisk symbol shows the velocity peak.The vertical red (dotted) lines indicate the bridge collapse time.

Figure 11 .
Figure 11.Wind speed (black dots; primary _-axis) and wind direction (grey dots; secondary _-axis) from 8 anemometers in the Port of Genoa.The vertical red (dotted) lines indicate the bridge collapse time.Panels (a)-(d) represent the stations west and south from the Morandi Bridge, while the panels (e)-(g) are the stations southeast from the bridge (see Fig. 2 for details).

Figure 13 .
Figure 13.Post-processed lidar data shown the height of gust front at different elevation angles (symbols) and 845

Figure 14 .
Figure 14.Conceptual model of the gust front that occurred prior to the collapse of the Morandi Bridge in Genoa on 850

Figure 15 .
Figure 15.Doppler velocity measured during the time of bridge collapse by the WSL400s lidar located in the Port of Genoa.The beam elevation (%) and the scanning time interval in MM:SS (blue text; the hour in all plots is 09 UTC) provided in each figure.

Figure 16 .
Figure 16.Maximum 5-min wind speeds at 10 m above ground at the eight anemometers location in the Genoa port (see Fig. 2 and Fig. 11 for anemometer locations and measurements, respectively).The vertical dashed lines show the time instances used for the analysis of radar reflectivity and thermodynamics diagrams.

Figure 17 .
Figure 17.Radar reflectivity (dBZ) measured by the Ligurian Doppler radar.VMI maps from observations (a-c) and WRF simulations (d-l).The position of Morandi bridge is indicated by the black cross.

Figure 18 .
Figure 18.WRF-IFS thermodynamic diagram at 1240 UTC in the center of the convective structure reported in Fig.17f.