ABSTRACT Two severe flood events hit Liguria Region in less than 15 days: on the 25th of October ... more ABSTRACT Two severe flood events hit Liguria Region in less than 15 days: on the 25th of October the Cinque Terre area was interested by a severe Mediterranean storm that caused 9 deaths and only ten days later, on November the 4th, the city of Genoa was rocked by severe flash floods that killed six people. During these events a hydrometeorological probabilistic forecasting system for small and medium size basins was operational at the Civil Protection Center of Liguria Region. This system shown excellent performances for these events by indicating, 24 hours in advance, the potential level of hazard associated to the forecasted events and helped the Liguria Region decision makers in issuing a timely and correct alert. In this work we present in detail the operational outputs of the system provided during the Liguria events and the overall performance of the system in the period 2010-2011. We discuss strong points and drawbacks of probabilistic systems for decision-making under uncertainty conditions, such as their impact in terms of missing/false alarms and their actual effectiveness in driving the decision process in comparison with deterministic forecast in same conditions.
ABSTRACT The city of Genoa, which nestles between the Tyrrhenian sea and the Apennine mountains (... more ABSTRACT The city of Genoa, which nestles between the Tyrrhenian sea and the Apennine mountains (Liguria, Italy) was rocked by severe flash floods on 4th November 2011. About 500 millimeters of rain - a third of the average annual rainfall - fell in six hours. Six people were killed. The raging waters uprooted trees, swept cars, shattered shops and flooded the town center. The synoptic-scale meteorological system, which was responsible for the Genoa flash flood, raged from West Virginia to Maine from 29th until 30th October and was blamed for at least 13 deaths. It moved across the Atlantic Ocean, generated floods that killed 5 people in Southern France, and finally it arrived over the Liguria sea and produced the severe rainfall, which we present here. Cloud-permitting simulations (1 km) of the finger-like convective system responsible for this torrential event over Genoa have been performed using Advanced Research Weather and Forecasting Model (ARW-WRF, version 3.2.1). Two different microphysics (WSM6 and Thompson) as well as three different convection closures (explicit, Kain-Fritsch, and Betts-Miller-Janjic) have been combined to gain a deeper understanding of the physical processes underlying the observed torrential event. Comparison between modeling findings and observed fields provided by raingauge data, satellite data, and radar data have undertaken. The paper presents the results and gathers conclusions.
ABSTRACT The estimation of rainfall fields, especially its spatial distribution and position is a... more ABSTRACT The estimation of rainfall fields, especially its spatial distribution and position is a crucial task both for rainfall nowcasting and for modeling catchment response to rainfall. Some studies of literature about multisensor have suggested that discharge estimations are improved when radar and rain gauge data are combined to estimate input rainfall fields.Sinclair and Peagram (2004) have proposed the Conditional Merging (CM) technique, a merging algorithm which extract the information content from the observed data and use it within an interpolation method to obtain the rainfall maps. The idea is to combine the "real", but punctual amount of rainfall measured by raingauges with the structure of covariance and correlation of rainfall maps estimated from remote sensors (radar network or satellite constellation). In this work is studied an enhanced algorithm based on CM, called Modified Conditional Merging.. The main innovation respect to classical CM is the estimation of the structure of covariance and the length of spatial correlation λ, for every raingauge, directly from the cumulated radar rainfall fields. The domain of application is the Italy, where are both available a dense network of raingauge measurements (about 2500 stations) and a QPE estimated by the Italian Radar composite. The MCM algorithm can be used in real-time over the whole domain to produce hourly the optimal rainfall maps. An application to several test cases together with the evaluation of algorithm performances are presented and discussed.
ABSTRACT Quantitative precipitation estimation is a crucial issue especially for predicting flash... more ABSTRACT Quantitative precipitation estimation is a crucial issue especially for predicting flash floods in the small catchments typical of the Mediterranean area. In Italy, Civil Protection has the duty of managing and generating reliable QPE products useful for warning and monitoring purposes in case of severe events over the Italian territory. To this end a dense network of raingauges (about 2000 stations with hourly timestep) is available together with a precipitation field obtained from the National Radar Composite. This work aims at considering and comparing different techniques available for operationally merge rainfall estimation from different sensors . In particular will be presented a comparison of the Modified Conditional Merging algorithm (presented at this conference by Pignone et al.) with a Bayesian procedure that exploits the radar quality maps available for two test sites of the Italian Radar Composite (Mt. Settepani and Il Monte). The aim of this work is twofold, from one side is to test the Bayesian algorithm, its robustness and applicability in an operational context, and, on the other side, to define the best procedure for comparing and evaluating different data merging algorithms potentially useful for Civil Protection applications.
ABSTRACT The forecast of flash floods is sometimes impossible. In the last two decades, Numerical... more ABSTRACT The forecast of flash floods is sometimes impossible. In the last two decades, Numerical Weather Prediction Systems have become increasingly reliable with very relevant improvements in terms of quantitative precipitation forecasts. However, some types of events, those that are intense and localized in small areas, are still very difficult to predict. In many cases meteorological models fail to predict the volume of precipitable water at the large scale. Despite the application of modern probabilistic chains that uses precipitation downscaling algorithms in order to forecast the streamflow, some significant flood events remain unpredicted. This was also the case with an event which occurred on 8 and 9 June 2011 in the eastern part of the Liguria Region, Italy. This event affected in particular the Entella basin, which is quite a small watershed that flows into the Mediterranean Sea. The application of a hydrological nowcasting chain as a tool for predicting flash-floods in small and medium size basins with an anticipation time of a few hours (2-5) is here presented. This work investigated the "behaviour" of the chain in the cited event and how it could be exploited for operational purposes. The results in this particular case were encouraging.
ABSTRACT The estimation of rainfall fields, especially its spatial distribution and position is a... more ABSTRACT The estimation of rainfall fields, especially its spatial distribution and position is a crucial task both for rainfall nowcasting and for modeling catchment response to rainfall. Some studies of literature about multisensor datafusion prove that combining data from raingauges and radar represents the best way to obtain an enhanced ad more reliable estimation of QPE and of the associated river discharge. Sinclair and Peagram (2004) have proposed the Conditional Merging (CM) technique, a merging algorithm which extract the information content from the observed data and use it within an interpolation method to obtain the rainfall maps. The raingauges provide a punctual measure of the ground-observed rainfall while the remote sensors (radar network or satellite constellation) supply rainfall estimation maps which give an idea of the correlation and structure of covariance of the observed field. In this work is presented an algorithm called Modified Conditional Merging that is based on CM and which is used for real-time estimation of the optimal rainfall maps. The area of interest is Italy, where are both available a dense network of raingauge measurements (about 2000 stations) and a QPE estimated by the Italian Radar composite. The main innovation respect to classical CM is to estimate the structure of covariance and the length of spatial correlation λ, for every raingauge, directly from the cumulated radar rainfall fields. An application to several test cases together with the evaluation of algorithm performances are presented and discussed.
ABSTRACT The forecast of floods related to intense rainfall events is one of the main themes of a... more ABSTRACT The forecast of floods related to intense rainfall events is one of the main themes of analysis in hydrometeorology and a key issue for Civil Protection systems. In this work we present a hydrometeorological probabilistic forecast system for small and medium size basins, designed for operational applications. The probabilistic approach presented here allows to face the problems related to the reduced dimension of the basins and to properly account for uncertainty sources in the prediction chain. Starting from Quantitative Precipitation Forecasts (QPF) provided by the regional center which is in charge of hydrometeorological predictions in Liguria Region, the system is able to issue probabilistic warnings both following a catchment-based criterion (single site) or following an area-based approach (multi-catchment).
An intriguing prospective in distributed modeling is the possibility to incorporate remotely sens... more An intriguing prospective in distributed modeling is the possibility to incorporate remotely sensed data, such as evaporative fraction and soil moisture conditions, with the purpose of initializing, driving, updating or calibrating models. In this work a parsimonious distributed model devoted to small catchments continuous discharge simulation is presented. Soil type and land use information are used to calibrate a modified
ABSTRACT On the 4th November 2011 a flash flood event hit the area of Genoa with dramatic consequ... more ABSTRACT On the 4th November 2011 a flash flood event hit the area of Genoa with dramatic consequences. Such an event represents, from the meteorological and hydrological prospective, a paradigm of flash floods in the Mediterranean environment. The performances of an early warning system in use at the Ligurian Hydrometeorological Centre are presented and discussed together with its ex-post evaluation. The evaluation was conducted by using official and unofficial rain gauge networks in order to understand the added value of a denser observation network in case of localized convective rainfall, which is a recurrent feature in these types of events. The huge number of crowd sourcing information (videos and photos) available in near real time during the event were used to monitor, from a different angle, the flood ground effects. In the post event they were used to quantitatively estimate flood magnitude in ungauged sites for hydrological modeling validation.
ABSTRACT On the 4 November 2011 a flash flood event hit the area of Genoa with dramatic consequen... more ABSTRACT On the 4 November 2011 a flash flood event hit the area of Genoa with dramatic consequences. Such an event represents, from the meteorological and hydrological perspective, a paradigm of flash floods in the Mediterranean environment. The hydro-meteorological probabilistic forecasting system for small and medium size catchments in use at the Civil Protection Centre of Liguria region exhibited excellent performances for the event, by predicting, 24-48 h in advance, the potential level of risk associated with the forecast. It greatly helped the decision makers in issuing a timely and correct alert. In this work we present the operational outputs of the system provided during the Liguria events and the post event hydrological modelling analysis that has been carried out accounting also for the crowd sourcing information and data. We discuss the benefit of the implemented probabilistic systems for decision-making under uncertainty, highlighting how, in this case, the multi-catchment approach used for predicting floods in small basins has been crucial.
ABSTRACT Two severe flood events hit Liguria Region in less than 15 days: on the 25th of October ... more ABSTRACT Two severe flood events hit Liguria Region in less than 15 days: on the 25th of October the Cinque Terre area was interested by a severe Mediterranean storm that caused 9 deaths and only ten days later, on November the 4th, the city of Genoa was rocked by severe flash floods that killed six people. During these events a hydrometeorological probabilistic forecasting system for small and medium size basins was operational at the Civil Protection Center of Liguria Region. This system shown excellent performances for these events by indicating, 24 hours in advance, the potential level of hazard associated to the forecasted events and helped the Liguria Region decision makers in issuing a timely and correct alert. In this work we present in detail the operational outputs of the system provided during the Liguria events and the overall performance of the system in the period 2010-2011. We discuss strong points and drawbacks of probabilistic systems for decision-making under uncertainty conditions, such as their impact in terms of missing/false alarms and their actual effectiveness in driving the decision process in comparison with deterministic forecast in same conditions.
ABSTRACT The city of Genoa, which nestles between the Tyrrhenian sea and the Apennine mountains (... more ABSTRACT The city of Genoa, which nestles between the Tyrrhenian sea and the Apennine mountains (Liguria, Italy) was rocked by severe flash floods on 4th November 2011. About 500 millimeters of rain - a third of the average annual rainfall - fell in six hours. Six people were killed. The raging waters uprooted trees, swept cars, shattered shops and flooded the town center. The synoptic-scale meteorological system, which was responsible for the Genoa flash flood, raged from West Virginia to Maine from 29th until 30th October and was blamed for at least 13 deaths. It moved across the Atlantic Ocean, generated floods that killed 5 people in Southern France, and finally it arrived over the Liguria sea and produced the severe rainfall, which we present here. Cloud-permitting simulations (1 km) of the finger-like convective system responsible for this torrential event over Genoa have been performed using Advanced Research Weather and Forecasting Model (ARW-WRF, version 3.2.1). Two different microphysics (WSM6 and Thompson) as well as three different convection closures (explicit, Kain-Fritsch, and Betts-Miller-Janjic) have been combined to gain a deeper understanding of the physical processes underlying the observed torrential event. Comparison between modeling findings and observed fields provided by raingauge data, satellite data, and radar data have undertaken. The paper presents the results and gathers conclusions.
ABSTRACT The estimation of rainfall fields, especially its spatial distribution and position is a... more ABSTRACT The estimation of rainfall fields, especially its spatial distribution and position is a crucial task both for rainfall nowcasting and for modeling catchment response to rainfall. Some studies of literature about multisensor have suggested that discharge estimations are improved when radar and rain gauge data are combined to estimate input rainfall fields.Sinclair and Peagram (2004) have proposed the Conditional Merging (CM) technique, a merging algorithm which extract the information content from the observed data and use it within an interpolation method to obtain the rainfall maps. The idea is to combine the "real", but punctual amount of rainfall measured by raingauges with the structure of covariance and correlation of rainfall maps estimated from remote sensors (radar network or satellite constellation). In this work is studied an enhanced algorithm based on CM, called Modified Conditional Merging.. The main innovation respect to classical CM is the estimation of the structure of covariance and the length of spatial correlation λ, for every raingauge, directly from the cumulated radar rainfall fields. The domain of application is the Italy, where are both available a dense network of raingauge measurements (about 2500 stations) and a QPE estimated by the Italian Radar composite. The MCM algorithm can be used in real-time over the whole domain to produce hourly the optimal rainfall maps. An application to several test cases together with the evaluation of algorithm performances are presented and discussed.
ABSTRACT Quantitative precipitation estimation is a crucial issue especially for predicting flash... more ABSTRACT Quantitative precipitation estimation is a crucial issue especially for predicting flash floods in the small catchments typical of the Mediterranean area. In Italy, Civil Protection has the duty of managing and generating reliable QPE products useful for warning and monitoring purposes in case of severe events over the Italian territory. To this end a dense network of raingauges (about 2000 stations with hourly timestep) is available together with a precipitation field obtained from the National Radar Composite. This work aims at considering and comparing different techniques available for operationally merge rainfall estimation from different sensors . In particular will be presented a comparison of the Modified Conditional Merging algorithm (presented at this conference by Pignone et al.) with a Bayesian procedure that exploits the radar quality maps available for two test sites of the Italian Radar Composite (Mt. Settepani and Il Monte). The aim of this work is twofold, from one side is to test the Bayesian algorithm, its robustness and applicability in an operational context, and, on the other side, to define the best procedure for comparing and evaluating different data merging algorithms potentially useful for Civil Protection applications.
ABSTRACT The forecast of flash floods is sometimes impossible. In the last two decades, Numerical... more ABSTRACT The forecast of flash floods is sometimes impossible. In the last two decades, Numerical Weather Prediction Systems have become increasingly reliable with very relevant improvements in terms of quantitative precipitation forecasts. However, some types of events, those that are intense and localized in small areas, are still very difficult to predict. In many cases meteorological models fail to predict the volume of precipitable water at the large scale. Despite the application of modern probabilistic chains that uses precipitation downscaling algorithms in order to forecast the streamflow, some significant flood events remain unpredicted. This was also the case with an event which occurred on 8 and 9 June 2011 in the eastern part of the Liguria Region, Italy. This event affected in particular the Entella basin, which is quite a small watershed that flows into the Mediterranean Sea. The application of a hydrological nowcasting chain as a tool for predicting flash-floods in small and medium size basins with an anticipation time of a few hours (2-5) is here presented. This work investigated the "behaviour" of the chain in the cited event and how it could be exploited for operational purposes. The results in this particular case were encouraging.
ABSTRACT The estimation of rainfall fields, especially its spatial distribution and position is a... more ABSTRACT The estimation of rainfall fields, especially its spatial distribution and position is a crucial task both for rainfall nowcasting and for modeling catchment response to rainfall. Some studies of literature about multisensor datafusion prove that combining data from raingauges and radar represents the best way to obtain an enhanced ad more reliable estimation of QPE and of the associated river discharge. Sinclair and Peagram (2004) have proposed the Conditional Merging (CM) technique, a merging algorithm which extract the information content from the observed data and use it within an interpolation method to obtain the rainfall maps. The raingauges provide a punctual measure of the ground-observed rainfall while the remote sensors (radar network or satellite constellation) supply rainfall estimation maps which give an idea of the correlation and structure of covariance of the observed field. In this work is presented an algorithm called Modified Conditional Merging that is based on CM and which is used for real-time estimation of the optimal rainfall maps. The area of interest is Italy, where are both available a dense network of raingauge measurements (about 2000 stations) and a QPE estimated by the Italian Radar composite. The main innovation respect to classical CM is to estimate the structure of covariance and the length of spatial correlation λ, for every raingauge, directly from the cumulated radar rainfall fields. An application to several test cases together with the evaluation of algorithm performances are presented and discussed.
ABSTRACT The forecast of floods related to intense rainfall events is one of the main themes of a... more ABSTRACT The forecast of floods related to intense rainfall events is one of the main themes of analysis in hydrometeorology and a key issue for Civil Protection systems. In this work we present a hydrometeorological probabilistic forecast system for small and medium size basins, designed for operational applications. The probabilistic approach presented here allows to face the problems related to the reduced dimension of the basins and to properly account for uncertainty sources in the prediction chain. Starting from Quantitative Precipitation Forecasts (QPF) provided by the regional center which is in charge of hydrometeorological predictions in Liguria Region, the system is able to issue probabilistic warnings both following a catchment-based criterion (single site) or following an area-based approach (multi-catchment).
An intriguing prospective in distributed modeling is the possibility to incorporate remotely sens... more An intriguing prospective in distributed modeling is the possibility to incorporate remotely sensed data, such as evaporative fraction and soil moisture conditions, with the purpose of initializing, driving, updating or calibrating models. In this work a parsimonious distributed model devoted to small catchments continuous discharge simulation is presented. Soil type and land use information are used to calibrate a modified
ABSTRACT On the 4th November 2011 a flash flood event hit the area of Genoa with dramatic consequ... more ABSTRACT On the 4th November 2011 a flash flood event hit the area of Genoa with dramatic consequences. Such an event represents, from the meteorological and hydrological prospective, a paradigm of flash floods in the Mediterranean environment. The performances of an early warning system in use at the Ligurian Hydrometeorological Centre are presented and discussed together with its ex-post evaluation. The evaluation was conducted by using official and unofficial rain gauge networks in order to understand the added value of a denser observation network in case of localized convective rainfall, which is a recurrent feature in these types of events. The huge number of crowd sourcing information (videos and photos) available in near real time during the event were used to monitor, from a different angle, the flood ground effects. In the post event they were used to quantitatively estimate flood magnitude in ungauged sites for hydrological modeling validation.
ABSTRACT On the 4 November 2011 a flash flood event hit the area of Genoa with dramatic consequen... more ABSTRACT On the 4 November 2011 a flash flood event hit the area of Genoa with dramatic consequences. Such an event represents, from the meteorological and hydrological perspective, a paradigm of flash floods in the Mediterranean environment. The hydro-meteorological probabilistic forecasting system for small and medium size catchments in use at the Civil Protection Centre of Liguria region exhibited excellent performances for the event, by predicting, 24-48 h in advance, the potential level of risk associated with the forecast. It greatly helped the decision makers in issuing a timely and correct alert. In this work we present the operational outputs of the system provided during the Liguria events and the post event hydrological modelling analysis that has been carried out accounting also for the crowd sourcing information and data. We discuss the benefit of the implemented probabilistic systems for decision-making under uncertainty, highlighting how, in this case, the multi-catchment approach used for predicting floods in small basins has been crucial.
Uploads
Papers by Francesco Silvestro