Search

Although hail is a well-known meteorological hazard, it is hard to find long records of hail observed at the ground with high spatial resolution. Most hail climatologies are based on remote sensing observations or inhomogeneous networks of human observers. In the plain of Friuli Venezia Giulia (northeast Italy), a hailpad network of 367 stations has operated since 1988. During the 1988–2016 warm seasons, 7782 hailpads were impacted by hailstones and more than one million dents were observed and automatically analyzed, even though only 63% of them were associated with valid hailstone dents. In this work, this large quantity of direct hail observations is used to build a hail climatology in terms of hailstone size, areal density, and flux of kinetic energy. From the empirical distributions of data collected, it is possible to fit statistical distributions to the different hailstone/hailpad behaviors. In particular, it is also possible to find an approximate estimation of the flux of kinetic energy based only on the largest hail diameter observed on the hailpad. Last, temporal and spatial distributions of these characteristics are investigated. Hailstones are larger along a southwestern-to-northeastern alley, which is parallel to the main pre-Alpine crest, with the very largest sizes being more frequent on the southwestern corner. The only hail climate change signal that one can infer from the analysis of these multidecadal trends is that, in more recent years, hailstorms seem to produce fewer and larger hailstones, on average.

Measuring hailstorms is a difficult task due to the rarity and mainly small spatial extent of the events. Especially, hail observations from ground-based time-recording instruments are scarce. We present the first study of extended field observations made by a network of 80 automatic hail sensors from Switzerland. The main benefits of the sensors are the live recording of the hailstone kinetic energy and the precise timing of the impacts. Its potential limitations include a diameter dependent dead time which results in less than 5 % of missed impacts, and the possible recording of impacts not due to hail which can be filtered using a radar reflectivity filter. We assess the robustness of the sensors measurements by doing a statistical comparison of the sensor observations with hailpads observations and we show that despite their different measurement approaches, both devices measure the same hail size distributions. We then use the timing information to measure the local duration of hail events, the cumulative time distribution of impacts and the time of the largest hailstone during a hail event. We find that 75 % of local hailfalls last just a few minutes (from less than 4.4 min to less than 7.7 min, depending on a parameter to delineate the events) and that 75 % of impacts occurs in less than 3.3 min to less than 4.7 min. This time distribution suggests that most hailstones, including the largest, fall during a first phase of high hailstone density, while a few remaining and smaller hailstones fall in a second low density phase.

A new lightning-flash and convective initiation climatology is developed over the Alpine area, one of the hotspots for lightning activity in Europe. The climatology uses cloud-to-ground (CG) data from the European Cooperation for LIghtning Detection (EUCLID) network, occurring from 2005 to 2019. The CG lightning data are gridded at a resolution of approximately 2km and 10min. A new and simple method of identifying convective initiation (CI) events applies a spatiotemporal mask to the CG data to determine CI timing and location.Although the method depends on a few empirical thresholds, sensitivity tests show the results to be robust. The maximum activity for both CG flashes and CI events is observed from mid-May to mid-September, with a peak at the end of July; the peak in the diurnal cycle occurs in the afternoon. CI is mainly concentrated over and around the Alps, particularly in northern and northeastern Italy. Since many thunderstorms follow the prevailing mid-latitude westerly flow, a peak of CG flashes extends from the mountains into the plains and coastal areas of northeastern Italy and Slovenia. CG flashes and CI events over the sea/coast occur less frequently than in plains and mountains, have a weaker diurnal cycle, and have a seasonal maximum in autumn instead of summer.

Hailstorms are relatively frequent in the Friuli Venezia Giulia region, northeastern Italy and, for that reason, a network of manual hailpads has been set up there since the late 1980s. On July 4, 2007, a record number of hailpads in the network were impacted by hail (115 out of about 360 total stations). To the best of the authors' knowledge, no other single hailstorm case-study has information coming from such a large number of hailpads impacted by hail. These hailpads are analysed using automatic software, able to interpret each hail dent as an ellipse. From these data, the distribution of hailstone diameters, areal hailstone density on the pads and total kinetic energy flux are computed and analysed in the present study. A parallel analysis of radar maximum reflectivity and cloud-to-ground lightning shows that the evolution of the hailstorm can be divided into three different phases, the first being the most severe in terms of lightning and hail. Finally, numerical simulations performed with a non-hydrostatic mesoscale model, implementing a prognostic module for the explicit simulation of hail (WRF-HAILCAST), show that the simulation system provides useful information on the severity of the event in terms of maximum hailstone diameter and reflectivity shape.

International audience; Moisture is a key variable for determining the precipitation, thus the notion of the "moisture control" of the precipitation follows. The present chapter examines this notion from a probabilistic perspective. Here, it is important to distinguish between the probability of precipitation and the expected intensity of precipitation, under a given moisture state. It is suggested that the former increases in a more pronounced manner than the latter with the increasing moisture in the atmosphere. For this demonstration, local rain-gauge station data and a shared sounding over Friuli Venezia Giulia, NorthEast Italy are analyzed. Soundings provide a column precipitable water, which is used as a measure of the moisture in the atmosphere. Implication of the study is that moisture is more crucial in increasing a chance of precipitation, while it "controls" the precipitation intensity only in a lesser extent.

A bstract Friuli Venezia Giulia (FVG, NE Italy) is an area of maximum rainfall in the whole Alpine chain territory, reaching more than 3200 mm of mean annual rain in the Julian Prealps. According to recent climatological studies, the same area is also one of the European spot in recent lightning climatologies, meaning that convective rain plays an important role in the total rainfall. A network of 104 raingauges placed around the FVG territory is used to extract the absolute maximum rain accumulated every 6 hours in four subareas of FVG. In an attempt to improve the original ECMWF maximum rain, these data have been targeted to develop 32 statistical downscaling models, according to the period of the day, of the year and specific sub-area. ECMWF 6-hour rain forecasts available for all the gridpoints encompassed in the FVG territory and some derived variables (absolute values, anomalies, standardized values, plus mean, max and SD in time and/or space) have been used as predictors. With respect to a previous version of this work, here also the instability pseudo-indices (derived from the vertical profile with the maximum vertical resolution available in the ECMWF hybrid levels) are used as candidate predictors. Moreover, also non-linear methods, namely neural networks, are implemented, together with exhaustive multiregression models. Results show that the 32 models improve -on average- R2 of 12% on the validation sample and of 5% on the 2017 test sample, with respect to the ECMWF rain forecast, but the improvement is particularly notable during the convective season (18%).

This work evaluates numerous thunderstorm predictors and investigates the use of artificial neural networks (ANNs) for identifying occurrences of thunderstorms in reanalysis data. Environmental conditions favorable for deep, moist convection are derived from 6-hourly ERA-Interim reanalyses, while thunderstorm occurrence in the following 6 h over Finland is derived from lightning location data. By taking advantage of the consistency and large sample size (14 summers) provided by the reanalysis, complex multivariate models can be trained for a robust estimation of convective weather events from model data. This and other methods are used to yield information on the most effective convective predictors in a multivariate setting, which can also benefit the forecasting community. The best ANN found uses 15 inputs and received a Heidke skill score (HSS) of 0.51 on an independent test sample. This is a substantial improvement over the best predictor when used alone, the most unstable lifted index (MULI) with HSS = 0.40, the multivariate model having fewer false alarms in particular. After MULI, the most important ANN input was relative humidity near 700 hPa. Dry air aloft was associated with significantly lower thunderstorm probability and flash density regardless of convective available potential energy (CAPE). Other important parameters for thunderstorm development were vertical velocity and low-level θe advection. Finally, the Peirce skill score indicates a clear meridional gradient in skill for categorical forecasts, with higher skill in northern Finland. This analysis suggests that the difference in skill is real and associated with a steeper thunderstorm probability curve in the north, but further studies are needed for a physical explanation.

During the first HyMeX Special Observation Period (SOP1) field campaign, the target site of north‐eastern Italy (NEI) experienced a large amount of precipitation, locally exceeding the climatological values and distributed among several heavy‐rainfall episodes. In particular, two events that occurred during the last period of the campaign drew our attention. These events had common large‐scale patterns and a similar mesoscale setting, characterised by southerly low‐level flow interacting with the Alpine orography, but the precipitation distribution was very different. During Intensive Observing Period IOP18 (31 October–1 November 2012), convective systems were responsible for intense rainfall mainly located over a flat area of the eastern Po Valley, well upstream of the orography. Conversely, during IOP19 (4/5 November 2012), heavy precipitation affected only the Alpine area. In addition to IOP18 and IOP19, the present study analyses other heavy‐precipitation episodes that display similar characteristics and which occurred over NEI during the autumn of recent years. A high‐resolution (2 km grid spacing) non‐hydrostatic NWP model and available observations are used for this purpose. The two different observed precipitation patterns are explained in terms of interaction between the impinging flow and the Alps. Depending on the thermodynamic profile, convection can be triggered when the impinging flow is forced to rise over a pre‐existing cold‐air layer at the base of the orography. In this situation a persistent blocked‐flow condition and upstream convergence are responsible for heavy rain localized over the plain. Conversely, if convection does not develop, flow‐over conditions are established and heavy rain affects the Alps. Numerical parameters proposed in the literature are used to support the analysis. Finally, the role of evaporative cooling beneath the convective systems is evaluated. It turns out that the stationarity of the systems upstream of the Alps is mainly attributable to persistent blocked‐flow conditions, while convective outflow slightly modifies the location of precipitation.

This study analyzes the spatial distribution and temporal characteristics of cloud-to-ground lightnings (C2G) in the North East of Italy and the neighboring areas of Austria, Slovenia and Croatia. The dataset consists of about 6.5 millions C2G flash records, both positive and negative, observed between January 1995 and December 2011 by the "Centro Elettrotecnico Sperimentale Italiano-Sistema Italiano Rilevamento Fulmini'' (CESI/SIRF), part of the European Cooperation for Lightning Detection (EUCLID) Network. The results show that C2G lightnings concentrate in the foothill regions on the southern flank of the Eastern Alps with a maximum of discharge frequency of 10 lightnings per km2 per year. The number of C2G strokes varies with the period of the year: the most active period for lightning starts in April and lasts through November with the highest number of C2G strokes happening during the summer months of July and August, with maximum spatial density slightly moving from the mountain to the coastal area. The least frequency of C2G strokes is observed during wintertime. The mean diurnal C2G lightning activity for the whole domain shows a peak around 16:00–17:00 UTC and reaches a minimum around 07:00–09:00 UTC; the mean spatial distribution follows different patterns depending on the period of the day.

On 12 September 2012 a sequence of convective events hit the northeastern part of Italy and in particular the eastern part of Veneto and the plain of Friuli Venezia Giulia regions. During the day at least two events could be classified as supercells, the first one being also associated with a heavy hailfall. After a few hours, a third storm system, resembling a squall-line, although of limited dimensions, swept over the area.This event occurred during the first Special Observing Period (SOP1) of the HyMeX project as IOP2 (Intense Observing Period) and - also for this reason - many observations managed by different institutions were collected, including Doppler radar, extra-soundings, sodar and surface stations. Moreover, EUMETSAT was conducting its first experimental 2.5-minute rapid scan with the MSG-3 satellite, with data available from early morning until 0900 UTC of the IOP2 day.Several mesoscale models were run during the HyMeX SOP to support the field operations. A comparison between simulations of two high-resolution models (MOLOCH and WRF) is presented here and shows the capability in forecasting the intense convective activity in the area, although the exact temporal evolution of the systems was missed. Model simulations also provide useful insights concerning the mesoscale conditions conducive to the development of the convective systems.Finally, a diagnostic tool (Corfidi and Bunkers vectors) is applied using the model wind field, in order to infer further information on the temporal evolution of the convective cells. © 2014 Elsevier B.V.

The main object of this work is to study the lightning climatology in the Po Valley in Italy and how it varies in time (interannual, annual, weekly, and daily time scales) and space (sea coast, plains, and mountain areas) and how that is related to topographic characteristics and anthropogenic emissions. Cloud-to-ground (CG) lightning in the target area is analyzed for 18 yr of data (about 7 million records). It is found that the Julian Prealps of the Friuli Venezia Giulia region are one of the areas of maximum CG lightning activity across all of Europe. During spring lightning activity is more confined toward the mountainous regions, whereas during summer and even more during autumn the lightning activity involves also the coastal region and the Adriatic Sea. This is due to different triggering mechanisms acting in different topographic zones and during different periods of the year and times of the day. In analogy to previous studies of lightning done in the United States, a weekly cycle is also identified in the area of interest, showing that on Friday the probability of thunderstorms reaches its maximum. After conducting a parallel analysis with monitoring stations of atmospheric particulates (diameter ≤ 10 μm: PM10) and sounding-derived potential instability, the results presented herein seem to support the hypothesis that the weekly cycle in the thunderstorm activity may be due to anthropogenic emissions.

The Friuli Venezia Giulia region (northeastern Italy, hereafter FVG) is an area subject to frequent storm occurrences. The Regional Meteorological Observatory of Friuli Venezia Giulia (hereafter OSMER) manages a spatially dense network of meteorological stations, that provides measures with a maximum sampling frequency of 5 min. Moreover there is a C-band Doppler Polarimetric Radar that provides complete scans every 5 min during the convective events, or every 30 min otherwise. This work aims to characterize the skill of one or more predictors, that are the measurements of such meteorological stations (e.g. wind magnitude, equivalent potential temperature, moisture transport, etc.), when forecasting the highest radar Vertical Maximum Intensity reflectivity (hereafter VMI) measured on the plain and coast of FVG, used as signature of the storm occurrence and intensity. The dataset here used consists of 5-minutes maximum time-resolution VMI (used to define each event) on plain and coast areas of FVG, and 5-minutes time-resolution observations derived from 31 stations, considering a time-lag varying from 30 min to 3 h before the VMI observation. The VMI dataset spans over a period 11 years long (2000–2010) and consists of about 100,000 cases. Both a multiregression study to forecast the maximum VMI value and a classification study of the occurrence of VMI events exceeding certain thresholds have been performed.

In a previous work, the hailpad data collected over the plain of the Friuli Venezia Giulia region in northeast Italy during the April–September 1992–2009 period were studied through a bivariate analysis with 52 sounding-derived indices from the Udine–Campoformido station (WMO code 16044). The results showed statistically significant relations but, nevertheless, were not completely satisfactory from a practical point of view. In the current work, a prognostic multivariate analysis is performed, using linear and nonlinear approaches, finding the best results with an ensemble of neural networks. For the hail occurrence–classification problem, a novel method for combining binary classifiers (a variant of the Mojirsheibani major voting algorithm) is introduced. For the hail extension–regression problem the ensemble is built by choosing the members with a bagging algorithm, but combining them with a linear multiregression, in order to increase the forecast variability.

In this work the hailpad data collected during the April–September period of the years 1992–2009 by a network of approximately 360 stations on the plain of the Friuli Venezia Giulia region (northeast Italy) were analyzed. The frequency distribution of the hail cases is studied for the different years, for the yearday, and for the time of day, considering four 6-h-long periods of the day. Hail is found to be more frequent during the 1100–2300 UTC time interval and in the months of June and July even though during the 18 years analyzed there does not seem to be a defined trend in hail probability. The spatial distribution shows a maximum of probability along the foothills in particular in the northwest. In the middle of this area there is also the Udine–Campoformido radiosounding station, which performed four soundings per day. Each case (6-h-long period) is then associated with the 52 indices derived from the sounding launched at the beginning of that period. These sounding-derived indices are not all statistically independent, as found by looking at their correlation matrix and aggregating the most-correlated indices in three groups. A diagnostic bivariate analysis between each index and the hit hailpads in 6 h was done as a preliminary attempt to evaluate the utility of these indices for forecasting hail. It is found that some measures of instability (like updraft, hail diameter, and lifted index) seem to have more skill than the other indices when classifying the hail-event occurrence. When estimating the number of hit hailpads, the best correlation is obtained by the indices belonging to the “lifted index family.”

In this work, 40 different indices derived from real soundings and the corresponding ECMWF model forecasts for the same location (near Udine, northeast Italy) are compared. This comparison is repeated for more than 500 days, from June 2004 to December 2005. The comparison is performed in terms of linear correlation, bias, root-mean-square error, and relative absolute error. The results show that the quality of agreement varies considerably among the different indices and lead times. It is found that only 15 indices have strong correlations (R2 always larger than 0.5) that hold up until the fourth forecast day (+96 h). The indices having the best agreement are usually representative of the thermal fields while those having the worst agreement are usually representative of the wind fields, particularly in the low levels. It is also found that the correlations for all 40 of the indices change for the 4 times of the day (one sounding every 6 h), with the 1200 UTC comparison being usually the best one. Also, at 1800 and 0000 UTC the potential instability is strongly underestimated by the model. To test if these errors depend on the vertical resolutions, a similar comparison has been done between the observed sounding reduced in the World Meteorological Organization TEMP format (as assimilated by the NWP model) and the originals at high vertical resolution. It is found that the indices derived from the TEMP format vary from those with the highest resolution, in particular the indices that depend on the details of the low-level profile, such as the level of free convection and the initial height of the lifted parcel. Nevertheless, only the shear has a very low correlation (R2 = 0.15) with the corresponding index derived from the high-resolution soundings, because it is very sensitive to the number of vertical levels used.

Binary classifiers are obtained from a continuous predictor using a threshold to dichotomize the predictor value into event occurrence and nonoccurrence classes. A contingency table is associated with each threshold, and from this table many statistical indices (like skill scores) can be computed. This work shows that the threshold that maximizes one of these indices [the Peirce skill score (PSS)] has some important properties. In particular, at that threshold the ratio of the two likelihood distributions is always 1 and the event posterior probability is equal to the event prior probability. These properties, together with the consideration that the maximum PSS is the point with the “most skill” on the relative operating characteristic curve and the point that maximizes the forecast value, suggest the use of the maximum PSS as a good scalar measure of the classifier skill. To show that this most skilled point is not always the best one for all the users, a simple economic cost model is presented.

A neural network-based scheme to do a multivariate analysis for forecasting the occurrence and intensity of a meteo event is presented. Many sounding-derived indices are combined together to build a short-term forecast of thunderstorm and rainfall events, in the plain of the Friuli Venezia Giulia region (hereafter FVG, NE Italy). For thunderstorm forecasting, sounding, lightning strikes and mesonet station data (rain and wind) from April to November of the years 1995–2002 have been used to train and validate the artificial neural network (hereafter ANN), while the 2003 and 2004 data have been used as an independent test sample. Two kind of ANNs have been developed: the first is a “classification model” ANN and is built for forecasting the thunderstorm occurrence. If this first ANN predicts convective activity, then a second ANN, built as a “regression model”, is used for forecasting the thunderstorm intensity, as defined in a previous article. The classification performances are evaluated with the ROC diagram and some indices derived from the Table of Contingency (like KSS, FAR, Odds Ratio). The regression performances are evaluated using the Mean Square Error and the linear cross correlation coefficient R . A similar approach is applied to the problem of 6 h rainfall forecast in the Friuli Venezia Giulia plain, but in this second case the data cover the period from 1992 to 2004. Also the forecasts of binary events (defined as the occurrence of 5, 20 or 40 mm of maximum rain), made by classification and regression ANN, were compared. Particular emphasis is given to the sounding-derived indices which are chosen in the first places by the predictor forward selection algorithm.

In this paper several indices derived from atmospheric sounding data are used to develop a short-term thunderstorm forecast tool. The Udine (Italy) sounding (WMO code 16044, taken every 6 h, hereafter 6 h) is used to describe the initial conditions in which a thunderstorm may develop in the Friuli Venezia Giulia region. The tool forecasts the convective activity in the 6 h after the sounding is launched. Sounding, lightning, and mesonet station data, from April to September of the years 1995–2002, were used to train and validate artificial neural networks (ANNs) and the data of 2003 were used as an independent test sample. Special emphasis was given to avoid one of the major ANN problems, that of data overfitting, which requires limiting the possible complexity of the ANN, that is, the total number of inputs (predictors) and of hidden neurons. To select the best input set, the training–validation mechanism was repeated in eight different ways. Two types of ANNs were developed: the first is a classification model and is built for forecasting the thunderstorm occurrence, defined as the report of at least one cloud-to-ground lightning strike. If this first ANN forecasts convective activity, then a second ANN, built as a regression model, is used for forecasting the thunderstorm intensity. The relative operating characteristic (ROC) diagram and some parameters derived from a contingency table were used to compare the different classification performances. The linear cross-correlation coefficient, R2, was used to assess the performance of the regression ANNs. The ANN inputs were not the only “raw” sounding-derived index values; in addition, two derived sets of variables, corresponding to the z scores and the “posterior probability” transformed data, were used as preprocessing methods. The last method yields the best results. The tool reported on here has been used operationally since 2001 in the regional meteorological office of the Friuli Venezia Giulia region [Osservatorio Meteorologico Regionale/Agenzia Regionale per la Protezione dell’Ambiente (OSMER/ARPA)].

The relative operating characteristic (ROC) diagram is often used to assess the performance of a classification system, like the categorical forecast of an event occurrence. Categorical forecasting can be obtained by imposing a threshold on a continuous variable in order to make it dichotomous. In practice this threshold could be varied to create different contingency tables. From each table, it is then possible to derive many statistical indices and skill scores, which are functions of the chosen threshold. The ROC curve is obtained by plotting two of these indices: probability of detection (POD) versus probability of false detection (POFD). In this work a simple approximation for another of these indices, the odds ratio (O), is proposed. Thus, O is parameterized as a function of POFD and that leads to a parameterization of all the theoretical ROC curves. Using this approximation, it is also possible to derive the theoretical maximum Hanssen and Kuipers skill score (KSS) and the theoretical maximum Heidke skill score (HSS), for each ROC. It is found that the maximum HSS depends explicitly on the database event frequency (α), while the KSS seems independent of it. Out of the approximation framework, some general properties of ROC points corresponding to the maximum KSS, to the maximum HSS, and to the BIAS = 1 condition have also been found. It is also suggested that many of these performance measures are influenced by the event frequency, which must be taken into account when comparing classifiers made for different databases. Another interesting outcome of this study is that it is shown how the KSS is also equitable (in the sense introduced by Gandin and Murphy) for a generic “cost ratio” (λ) between miss and false alarm cases, not only for the original case λ = 1.

On March 26, 2001 a tornado was observed on the Friuli plain (Northeastern Italy), close to the foothills. The event took place in the afternoon, 15:45 UTC and it lasted for about 15 min. Fortunately no significant damages and injuries have been recorded since the tornado hit an uninhabited area with little vegetation and cultivation owing to the season. The whole event was observed and photographed by witnesses at a distance of about 1 km from the tornado outset to its end. Radar and mesonetwork data revealed the presence of a supercell mesocyclone having a radius of 4.5 ± 0.5 km that lasted for about half an hour and it was moving 30 ± 8° right to the average tropospheric flow; The event occurred in a synoptic situation dominated by a strong jet in the upper troposphere associated with a well-defined baroclinic zone (front). Mesoscale data revealed modest instability, but unusually high for the early spring period in the area. Vertical wind shear was very close to the 10 -2 s 1 that is significantly high for stormy days in Friuli. The surface conditions reveal that a thermal boundary was present that could have played an important role in the formation of the mesocyclone. The thermal boundary was the result of earlier storms' outflows and their interaction with the complex orography of the region. Comparison between horizontal relative vorticity parallel to the low level thermal boundary and the vertical relative vorticity of the mesocyclone suggests that such a boundary could have been a factor for the onset of the supercell. The lightning activity recorded during the event was characterized by a decrease of negative cloud-to-ground lightning (CG) frequency and an increase of positive (CG) lightning frequency during the observation of the event.

In the previous work, “Evaluating the sounding instability with the Lifted Parcel Theory” (hereafter ESI also published in this issue), we described a program which implements the Lifted Parcel Theory in three different ways. In this work, we build a climatology of the sounding-derived instability indices based on 5050 different soundings, made in Udine (WMO code 16044) from April to September 1995–2001. Every 6 h, we associate a sounding with the presence of cloud-to-ground lightning strikes and also with a continuous variable, CALCA6h, built using lightning strikes number, rain and wind gust, measured in the plain of our region. For each of the about 30 indices, we studied the skill score for forecasting the presence of thunderstorms and the linear cross-correlation for forecasting the thunderstorm “intensity”. We also introduce the use of some relatively new indices: the maximum buoyancy, the cloud precipitable water, the mean horizontal vapour flux and the standard deviation of the sounding vertical velocity. Finally, we discuss which indices and which of the three methods are the best in our region for forecasting convective events.

The Lifted Parcel Theory is a simple method for evaluating the instability of a sounding and computing a number of instability indices, for forecasting convective events. Very different results are obtained, depending on the choice of the initial parcel to lift and on the method used for evaluating the buoyancy (e.g., using the virtual or cloud-virtual temperature). A new method, which retains the condensed water inside the rising parcel and freezes it progressively and continuously, is investigated. For each of the three methods implemented, the coherent adiabatic processes are applied, and finally, the different results are compared.

On the morning of June 4th 1999, a severe weather event took place in San Quirino, a small village of Friuli-Venezia Giulia in the northeast of Italy. This village is located near the piedmont of the Alps, 40 km west from Udine and 60 km north from Venice. Around 0900 UTC (1100 local time), a thunderstorm with an intense hail fall affected the area of San Quirino. A few minutes later (around 0920 UTC, source: a farmer), a funnel cloud from a cumulonimbus touched the ground, producing damages to houses, trees and sheds. The damaged area was quite narrow (about 300 m) and short (less than 10 km). No injuries to people were reported. In spite of the smallness of the area interested by the phenomenon, this storm is studied here starting from the synoptic scale, moving to the mesoscale and finishing with the storm scale, trying to underline its characteristics. These analyses, especially those coming from the Doppler radar images, bring us to the conclusion that the San Quirino episode was produced by a supercell storm.

The Special Observation Period (SOP1), part of the HyMeX campaign (Hydrological cycle in the Mediterranean Experiments, 5 September–6 November 2012), was dedicated to heavy precipitation events and flash floods in the western Mediterranean, and three Italian hydro-meteorological monitoring sites were identified: Liguria–Tuscany, northeastern Italy and central Italy. The extraordinary deployment of advanced instrumentation, including instrumented aircrafts, and the use of several different operational weather forecast models, including hydrological models and marine models, allowed an unprecedented monitoring and analysis of high-impact weather events around the Italian hydro-meteorological sites. This activity has seen strong collaboration between the Italian scientific and operational communities. In this paper an overview of the Italian organization during SOP1 is provided, and selected Intensive Observation Periods (IOPs) are described. A significant event for each Italian target area is chosen for this analysis: IOP2 (12–13 September 2012) in northeastern Italy, IOP13 (15–16 October 2012) in central Italy and IOP19 (3–5 November 2012) in Liguria and Tuscany. For each IOP the meteorological characteristics, together with special observations and weather forecasts, are analyzed with the aim of highlighting strengths and weaknesses of the forecast modeling systems, including the hydrological impacts. The usefulness of having different weather forecast operational chains characterized by different numerical weather prediction models and/or different model set up or initial conditions is finally shown for one of the events (IOP19).