Papers by Francesco Lo Conti
Le precipitazioni intense costituiscono uno dei principali pericoli naturali perch\ue9 sono all\u... more Le precipitazioni intense costituiscono uno dei principali pericoli naturali perch\ue9 sono all\u2019origine di processi, come innesco di frane o piene improvvise, che possono rappresentare una grave minaccia per la vita umana. Il problema di determinare la variazione spaziale delle precipitazioni intense e in particolare, di indagare sulle relazioni che intercorrono tra queste e la morfologia del territorio, \ue8 molto importante soprattutto per gli studi connessi alla realizzazione di efficienti sistemi di allerta e di allarme. Tuttavia la variabilit\ue0 delle piogge intense con la morfologia \ue8 scarsamente studiata in idrologia. In questo lavoro si intende affrontare l\u2019argomento a scala regionale, assumendo che le precipitazioni intense siano rappresentate dalle curve di probabilit\ue0 pluviometrica che forniscono il quantile Tennale di assegnata durata come prodotto di un coefficiente di crescita in frequenza per una relazione di potenza che serve a riscalare le medie ora...
The EGU General Assembly, 2016
The extreme events have large impacts on society and are likely to increase under climate change.... more The extreme events have large impacts on society and are likely to increase under climate change. For design and management decisions, particularly around hydraulic infrastructures, accurate estimates of precipitation magnitudes are needed at different durations. In this paper, the regional frequency analysis has been implemented and applied to precipitation data recorded in Sicily, Italy. Annual maximum series for rainfall durations of 1, 3, 6, 12 and 24 h provided by about 130 rain gauges were used. The Regional Frequency Analysis (RFA) has been used to identify the homogeneous regions using Principal Component Analysis (PCA) followed by a clustering analysis, through k-means, aimed to identify regional groups. Two regional probability distributions have been used in order to derive the Depth-Duration Frequency (DDF) curves: lognormal distribution with three parameters (GNO) and generalized extreme value distribution (GEV). The regional parameters of these distributions were estimated using the L-moment ratios approach while the relative bias and relative RMSE have been calculated using a simulation study of regional L-moment algorithm for the assessment of the accuracy.
Extreme rainfall events have large impacts on society and are likely to continue to do so under p... more Extreme rainfall events have large impacts on society and are likely to continue to do so under predicted future climate change. Indeed, extreme precipitations show intensification in many regions of the world and this is of key importance to society as a result of the large impact of flooding. Thus, for planning and management decisions of the hydraulic infrastructures, accurate estimates of precipitation magnitudes at different durations are needed. Moreover, extreme precipitation events represent an increasing threat to society under global warming and are among the most serious challenges. In fact, future climate change is likely to lead to the change of extremes events that will become more frequent and could have significant impacts according to the outcomes of the IPCC 5th report. Therefore, the research of more accurate tool to build the Depth-Duration- Frequency (DDF) curves under future climate change is justified by the engineering applications that need more reliable and correct estimates of extreme precipitation in order to reduce the damage and loss of human lives. Starting from this premise, the primary objective of the present study is to explore response of the rainfall ex- tremes under possible future climate change in Sicily using an ensemble of outputs of General Circulation Models (GCMs). The future climate scenarios are generated using a stochastic downscaling technique based on the hourly weather generator, AWE-GEN. This methodology allows for the downscaling of an ensemble of climate model outputs deriving the frequency distribution functions of factors of change for several statistics of precipitation. The stochastic downscaling is carried out using simulations of GCMs adopted in the IPCC 5th report, for the future periods of 2046-2065 and 2081-2100. The evaluation of extreme precipitations has been carried out for two differ- ent climate scenarios Representative Concentration Pathways (RCP) 4.5 and 8.5, related to the future increase of the CO2, assessing the impact of the climate change embracing scenarios from the moderately optimistic one to the most pessimistic one, respectively. The generalized extreme value (GEV) frequency distribution was selected for the DDF curves derivation for different return periods. The analysis of climate change of extreme precipitation has shown an increase of the quantiles with a simultaneous decrease of mean annual precipitation. Moreover, the precipitation for the highest duration has shown an increase lower than that relative to the precipitation for the shortest duration
Recently, the Department of Civil, Environmental, Aerospace Engineering, and Materials (DICAM) of... more Recently, the Department of Civil, Environmental, Aerospace Engineering, and Materials (DICAM) of the University of Palermo (Italy) has installed several devices for the monitoring of precipitation for the urban area of Palermo. These devices include a single polarization X-band weather radar, a rain gauge network spread over the urban area, and a laboratory site where advanced precipitation devices (an optical disdrometer, a weight rain gauge, and a weather station) are available. Given the ensemble of measurements retrieved by sensors, a set of models for the combination of data has been developed in order to exploit their joint usage. The disdrometer information have been exploited for the calibration of both the radar equation and the Z-R relation. The availability of rain gauges data is considered for the implementation of a correction procedure that is aimed to the improvement of congruence between ground rain references measurements, and the radar estimates. The application f...
Recent literature shows several examples of simplified approaches that perform flood hazard (FH) ... more Recent literature shows several examples of simplified approaches that perform flood hazard (FH) assessment and mapping across large geographical areas on the basis of fast-computing geomorphic descriptors. These approaches may consider a single index (univariate) or use a set of indices simultaneously (multivariate). What is the potential and accuracy of multivariate approaches relative to univariate ones? Can we effectively use these methods for extrapolation purposes, i.e., FH assessment outside the region used for setting up the model? Our study addresses these open problems by considering two separate issues: (1) mapping flood-prone areas and (2) predicting the expected water depth for a given inundation scenario. We blend seven geomorphic descriptors through decision tree models trained on target FH maps, referring to a large study area (∼ 10 5 km 2). We discuss the potential of multivariate approaches relative to the performance of a selected univariate model and on the basis of multiple extrapolation experiments, where models are tested outside their training region. Our results show that multivariate approaches may (a) significantly enhance floodprone area delineation (accuracy: 92 %) relative to univariate ones (accuracy: 84 %), (b) provide accurate predictions of expected inundation depths (determination coefficient ∼ 0.7), and (c) produce encouraging results in extrapolation.
Proceedings of the 1st International Conference on Internet of Things and Machine Learning - IML '17, 2017
In this paper, a novel convolutional neural network model for blind deconvolution of images is pr... more In this paper, a novel convolutional neural network model for blind deconvolution of images is proposed. The structure of the model is based on two sub models devoted, respectively, to deblurring and denoising of an input image. The model has been designed to restore a picture affected by different kinds of noise. The main innovation is the introduction of a regularization term in the training cost function, based on a blurred/non-blurred classification tool. Results show interesting features of the model, particularly regarding the robustness of results. The comparison with other state-of-the-art models confirms the value of the model proposed in this study.
Environmental Modelling & Software, 2017
Integrative information models for filling/reconstructing hydro-climatic time-series are required... more Integrative information models for filling/reconstructing hydro-climatic time-series are required for a variety of practical applications. A GIS-based model for a rapid and reliable assessment of monthly timeseries of several key hydro-climatic variables at the basin scale, is here developed as plug-in and applied to the entire region of Sicily (Italy). The plug-in, once the desired basin outlet section and time-window are selected, uses appropriate spatial techniques and algorithms to identify its drainage area and estimate the corresponding mean areal rainfall and temperatures time-series. A recent regional regressive rainfall-runoff model is successively applied for the assessment of the runoff time-series. Finally, a consolidated temperature-based method is applied to estimate monthly potential evapotranspiration time-series, while, actual evapotranspiration and soil moisture storage time-series are derived through a classical water balance model. The tool, supported by a preliminarily developed database, includes automatic procedures for data retrieving and processing and a user friendly interface.
Environmental Earth Sciences, 2017
Critical rainfall thresholds for landslides are powerful tools for preventing landslide hazard. T... more Critical rainfall thresholds for landslides are powerful tools for preventing landslide hazard. The thresholds are commonly estimated empirically starting from rainfall events that triggered landslides in the past. The creation of the appropriate rainfall-landslide database is one of the main efforts in this approach. In fact, an accurate agreement between the landslide and rainfall information, in terms of location and timing, is essential in order to correctly estimate the rainfall-landslide relationships. A further issue is taking into account the average moisture conditions prior the triggering event, which reasonably may be crucial in determining the sufficient amount of precipitation. In this context, the aim of this paper is exploiting historical landslide and rainfall data in a spatial database for the derivation of critical rainfall thresholds for landslide occurrence in Sicily, southern Italy. The hourly rainfall events that caused landslides occurred in the twentieth century were specifically identified and reconstructed. A procedure was proposed to automatically convert rain guages charts recorded on paper tape into digital format and then to provide the cumulative rainfall hyetograph in digital format. This procedure is based on a segmentation followed by signal recognition techniques which allow to digitalize and to recognize the hyetograph automatically. The role of rainfall prior to the landslide events was taken into account by including in the analysis the rainfall occurred 5, 15 and 30 days before each landslide. Finally, cumulated rainfall duration thresholds for different exceedance probability levels were determined. The obtained thresholds resulted in agreement with the regional curves proposed by other authors for the same area; antecedent rainfall turned out to be particularly important in triggering landslides.
Journal of Hydrology, 2015
Summary The aim of this study is to evaluate the potential use of a low-cost single polarization ... more Summary The aim of this study is to evaluate the potential use of a low-cost single polarization X-band weather radar, verified by a disdrometer and a dense rain gauge network, installed as a supporting tool for hydrological applications and for monitoring the urban area of Palermo (Italy). Moreover, this study focuses on studying the temporal variability of the Z–R relation for Mediterranean areas. The radar device is provided with an automatic operational ground-clutter filter developed by the producer. Attention has been paid to the development of blending procedures between radar measurements and other auxiliary instruments and to their suitability for both meteorological and hydrological applications. A general scheme enveloping these procedures and achieving the combination of data retrieved from the weather radar, the optical disdrometer, and the rain gauge network distributed within the monitored area has been designed. The first step of the procedure consists in the calibration of the radar equation by comparing the match between the radar raw data and the disdrometer reflectivity. The second step is the calibration of the Z–R relationship based on the retrieval of parameters that optimize the transformation of disdrometer reflectivity into rainfall intensity, starting from the disdrometer rainfall intensity measurements. The Z–R calibration has been applied to the disdrometer measurements retrieved during a 1 year observation period, after a preliminary segmentation into separated rainfall events. This analysis allows for the characterization of the variability of the Z–R relationship from event to event, deriving some considerations about its predictability as well. Results obtained from this analysis provide a geographical specific record, for the Mediterranean area, for the study of the spatial variability of the Z–R relationship. Finally, the set of operational procedures also includes a correction procedure of radar estimates based on rain gauge data. Each application has been evaluated with reference to its suitability in supporting proper monitoring usage, with emphasis on the readiness of observations, and hydrological modelling, where the robustness of the quantitative estimates is focused.
The possibility to study the precipitation dynamics with advanced and specific tools is an import... more The possibility to study the precipitation dynamics with advanced and specific tools is an important task of the research activity addressing the understanding, the modeling, and the managing of rainfall events. Over the last years, the hydrology laboratory of the Department of Civil, Environmental, Aerospace Engineering, and Materials (DICAM) at the University of Palermo, has installed several instruments for the monitoring and the study of precipitation within the urban area of Palermo (Italy). The main instrument of this system is the X-band weather radar, which allows monitoring the precipitation fields with high resolution in space and time. This instrument is supported by a rain gauges network of 18 tipping bucket gauges spread over the observed area, a weight rain gauge, an optical disdrometer, and a weather station. The information provided by different devices can be combined in order to integrate different data and correct errors. In particular, the disdrometer is able to ...
This paper presents a comparative analysis between rain-gauge storm tracking techniques in order ... more This paper presents a comparative analysis between rain-gauge storm tracking techniques in order to achieve a better knowledge of the rainfall dynamics over an urbanized area. The temporal and spatial distribution and kinematics of short term rainfall are recognized as one of the most important reasons in error production in rainfall-runoff on urban catchments. The uncertainty due to rainfall variability can greatly affect urban drainage modeling performance and reliability thus reducing the confidence of operators in their results. Modeling representations of urban catchments and drainage systems are commonly adopted for surface flooding forecasting and management and an adequate knowledge of rainfall spatial and temporal variability should be considered as a fundamental step for a robust interpretation of the physical processes that take part in urban areas during intense rainfall events. The starting basis of such studies is usually given by a network of high resolution raingauges disseminated inside and around the examined urban area. One of the raingauge techniques used is based on simulating the storm motion by visualizing the sequence of the rainfall patterns obtained using rain-gauge data and on spatial correlation. The storm speed and direction are obtained using the rain-gauge method by tracking the advance of the maximum rainfall intensity in time and space. A second method is based on the identification, for each gauge, of the time of occurrence of some significant features such the time of onset of a storm or the time of peak. A third method is based on the classical idea of spacetime autocorrelation function; This function describes the way in which the correspondence between the rainfall patterns at two points in space-time reduces as the distance between two points is increased. The analysis has been carried out on the basis given by high resolution rainfall data collected over Palermo urban area (Italy). The urban area has a surface of around 30 km 2 and it is mainly distributed on North West-South East direction. The monitoring network is made of 10 tipping bucket raingauges. Bucket volume is equivalent to 0.1 mm rainfall. Raingauges have been uniformly distributed over the urban areas allocating them mainly over public buildings and school in order to allow for easy access. The network has been put in place in January 2006 and it is still working. Data is monthly collected by the operator that also provide for clock synchronization and ordinary maintenance and cleaning. An accurate analysis of the results of this comparison between the techniques has been carried out and, since the city of Palermo is not covered by any meteorological radar, the analysis of storm dynamics will allow to create a system monitoring hydrometeorological conditions which operates on time basis using the information coming from the raingauge network as forecast triggers.
ABSTRACT La conoscenza della distribuzione spaziale e temporale delle piogge di breve durata, non... more ABSTRACT La conoscenza della distribuzione spaziale e temporale delle piogge di breve durata, nonché la loro cinematica, sono alcuni tra i fattori più importanti che stanno alla base dell’approssimazione dei modelli di trasformazione afflussi-deflussi nei bacini urbani. L'incertezza dovuta alla variabilità spaziale della pioggia può influenzare, ad esempio, le performance dei modelli di drenaggio urbano. Un'adeguata conoscenza della variabilità spaziale e temporale delle precipitazioni può essere considerata un passo fondamentale per un'esatta interpretazione dei processi idrologici di base che avvengono nei sistemi di drenaggio urbano durante gli eventi di pioggia intensi. L'analisi di tali informazioni può essere condotta disponendo di una rete di pluviografi ad alta risoluzione distribuiti all'interno ed intorno l'area di studio con la quale osservare la dinamica degli eventi di pioggia. I metodi tipici con cui si effettua l'analisi di storm tracking, consistono nell'identificazione, per ogni strumento, di alcuni elementi significativi all'interno dello ietogramma come l'istante iniziale dell’evento meteorico o l'istante di picco e in analisi di cross-correlazione tra i dati dei diversi strumenti sulla base della loro distribuzione nello spazio. Nel presente studio viene descritta l'analisi condotta sulla base dei dati di precipitazione ad alta risoluzione raccolti nell'area urbana di Palermo da una rete di monitoraggio, costituita da 10 pluviografi posta in opera nel gennaio del 2006 e tuttora in funzione. L’analisi di storm tracking viene effettuata con il metodo della correlazione per una serie di eventi significativi, approfondendo al contempo l’analisi dell’influenza del passo temporale di aggregazione dei dati di pioggia e del numero di strumenti disponibili.
La valutazione della suscettibilit\ue0 al dissesto rimane uno degli approcci pi\uf9 utilizzanti e... more La valutazione della suscettibilit\ue0 al dissesto rimane uno degli approcci pi\uf9 utilizzanti e pi\uf9 efficaci per l'analisi della pericolosit\ue0 da frana. Come noto, la correlazione tra il fenomeno fisico ed i fattori predisponenti sulla base degli eventi accaduti in passato \ue8 il punto chiave di tale analisi. I metodi statistici, uniti con le tecnologie GIS, si sono rilevati in questi anni tra gli strumenti pi\uf9 idonei e pi\uf9 efficaci per la valutazione e la modellazione di tale correlazione. Tuttavia, questi metodi richiedono spesso ipotesi restrittive circa la distribuzione statistica dei dati che spesso non vengono rispettate. Per tale motivo si sono anche sviluppate metodologie alternative basate sull\u2019utilizzo delle Reti Neurali che hanno mostrato risultati migliori in termini di bont\ue0 del modello. Le problematiche connesse all\u2019utilizzo di tali modelli sono legate alla definizione dell\u2019architettura della rete, delle funzioni di interconnessione, nonch\ue9 alla scelta del dataset per la fase di apprendimento. Nel presente lavoro \ue8 stata affrontata in particolar modo l\u2019ultima problematica, applicando la metodologia su un piccolo bacino situato nella Sicilia orientale, dove una serie di eventi storici sono stati documentati nel corso degli anni e per il quale esistono gi\ue0 delle mappe di suscettibilit\ue0 derivate con altri metodi
ABSTRACT Negli ultimi anni i sistemi di misurazione e di stima delle precipitazioni, sono stati a... more ABSTRACT Negli ultimi anni i sistemi di misurazione e di stima delle precipitazioni, sono stati arricchiti dalla possibilità di ottenere stime attraverso dati acquisiti da sensori satellitari. I prodotti di stima così ottenuti risultano comunque spesso fortemente distorti. Per tale ragione le stime sulla base di soli dati satellitari, vengono ulteriormente elaborate attraverso l'utilizzo di dataset che raccolgono dati di precipitazione provenienti da reti pluviografiche al fine di ridurre la distorsione che caratterizza i dati satellitari. È auspicabile che questi dataset siano quanto più precisi al fine di fornire dati adatti alla correzione dei dati satellitari e in particolare alla riduzione della distorsione. Nello studio di seguito riportato viene descritta un'analisi di confronto tra il principale di tali dataset globali, cioè le stime di precipitazione del Global Precipitation Climatology Centre (GPCC) e quelle fornite da una rete pluviografica caratterizzata da una maggiore densità relativamente al territorio siciliano. I risultati evidenziano la presenza di forti discrepanze tra le due stime, localizzate in particolare nelle aree a quota maggiore. L'ipotesi di un non adeguato campionamento delle precipitazioni in tali aree viene confermata dall'osservazione delle localizzazioni delle stazioni utilizzate dal GPCC e da un'analisi dei risultati del campionamento effettuato con tale rete.
ABSTRACT The interest for spatial interpolating climatic datasets, as precipitation and temperatu... more ABSTRACT The interest for spatial interpolating climatic datasets, as precipitation and temperature, arises from different needs, ranging from their usage for hydrological models to the building of meteoclimatic atlas of spatially distributed data. In the area of Sicily (Italy) the spatial distribution of these variables is related to the extremely variable morphology of the area. Measurements of precipitation and temperature are collected by means of related networks in a set of points. While simple deterministic interpolation methods usually produce just the spatial distribution of the variable of interest, implicitly relying on the data spatial autocorrelation and manually selecting a few parameters, more complex statistical models, can produce the uncertainty associated with the estimate and model its different components, like those related to the measurement error and the parameters selection. With reference to the area of Sicily, mean annual precipitation and temperature data have been modeled using a hierarchical bayesian spatial model. Highlights about the insights provided by the hierarchical model are commented, in particular with reference to the uncertainty associated with spatial random effects.
ABSTRACT I modelli empirici per la derivazione di soglie pluviometriche critiche di innesco frana... more ABSTRACT I modelli empirici per la derivazione di soglie pluviometriche critiche di innesco frana risultano tra le metodologie più utilizzati utilizzate in letteratura volte a fornire strumenti di pre-allarme in vista di un evento meteorico critico. Esse identificano il valore minimo o massimo di precipitazione necessaria per innescare il processo franoso ad una fissata durata e sono generalmente descritte da una relazione funzionale di potenza sul piano Intensità-Durata della pioggia, i cui parametri vengono stimati empiricamente dall'analisi degli eventi meteorici storici che hanno innescato eventi franosi. Per poter condurre una simile analisi è dunque fondamentale la creazione di un opportuno database di eventi storici franosi e dei corrispondenti eventi di pioggia scatenanti. Il presente studio descrive la creazione di un accurato dataset per il territorio Siciliano a partire da un database di frane storiche e da uno dei dati pluviometrici. Il dataset delle frane è stato derivato dall'archivio on line del Sistema Informativo sulle Catastrofi Idrogeologiche (SICI) ideato nell'ambito del progetto AVI (Aree Vulnerabili Italiane). I dati pluviometrici sono stati ricavati dalla rete di monitoraggio della regione Sicilia, gestita dall'Osservatorio delle Acque–Agenzia Regionale per i Rifiuti e le Acque (OA-ARRA). La creazione di un simile dataset richiede un'accurata analisi delle informazioni temporali e spaziali dei due database, non sempre facilmente reperibili.
Uploads
Papers by Francesco Lo Conti