Long-term weather records are often stored in archives still in paper format. These data come from different sources, and they represent a key starting point to understand the climate of the past, a reference to validate climate models and an input data for reanalysis. This problem concerns both rich regions and data-sparse areas. In recent decades, numerous climate data rescue programs have begun in many countries worldwide. These projects aim to preserve data by digitizing, transcribing and analyzing them, in order to make them accessible to the scientific community. One of these projects is presented in this study, ReData (Recovery of Data), launched by the Meteonetwork Association in collaboration with the University of Milan in 2017. The aim of this project is to scan and digitize the daily weather bulletins available from 1879 to 1940 relative to the Italian country plus some colonies and surrounding territories edited by the Italian Royal Central Meteorological Office. The scanning process is finished, resulting in a collection of 99,518 pages (about 200GB) that will be soon available. The digitization is in progress through the Zooniverse platform (https://www.zooniverse.org/projects/meteonetwork/redata). Here, volunteers from all over the world can contribute to the digitization of a station with data for 12 variables in about 1 minute per day only. Since December 2024, when the project has been launched on Zooniverse, with a mean of about 5000 classification per week and a number of active participants per week stabilized between 50 and 100, we have digitized more than 10 years for 37 stations. Moreover, considering the good quality of the digitization (for about 99% of the data two of the three digitization are equal) it has been decided to move from three digitizations for each day to two digitizations to reduce the whole transcription.

Precipitation extremes are a significant natural hazard that has caused considerable destruction in Italy over the past decade. However, our understanding of the effects of climate change on these extremes remains incomplete, with unclear trends in the intensity and frequency of precipitation extremes. Part of this uncertainty results from internal variability in atmospheric circulation, which is key in triggering high-­ impact precipitation events. To address this issue, here we develop a comprehensive classification of the Weather Types (WTs) associated with November 1984 to October 2024 Extreme Precipitation Events (EPEs) from CERRA in the 156 operational warning areas used by the Italian Department of Civil Protection, by applying Self-Organising Maps to sea level pressure and 500 hPa geopotential height. We identify six different WTs associated with Italian EPEs, corresponding to different large-­ scale dynamical drivers: Atlantic cyclone over France/northern Tyrrhenian Sea (WT1), Mediterranean cyclone over Central Italy (WT2), Western Mediterranean cyclone associated with upper level trough over Iberia (WT3), Westerly zonal flow (WT4), upper-­ level cut-­ off low (WT5), and Mediterranean cyclone centered over the Tyrrhenian Sea (WT6). The relevance of these WTs for different warning areas is evaluated through composites of moisture transport, the probability of EPEs given a specific WT and a seasonality analysis. The annual frequency of extreme precipitation events exhibits a statistically significant increasing trend (Mann-­ Kendall, p < 0.05), corresponding to an average rise of more than two additional events per year per decade. The positive trend in precipitation extremes occurs for five out of the six WTs, with the largest increase occurring for WT4. These results add to the existing knowledge of drivers of extreme precipitation events in Italy, providing an understanding of underlying large-­ scale atmospheric circulation and a database of weather types to investigate the role of anthropogenic climate change in climate model simulations.

Understanding the interplay between meteorological and agricultural droughts is crucial for assessing the impacts of compound climate extremes. While the link between temperature, precipitation, and soil moisture has been largely investigated for drought assessment, their physical properties and role in drought type are region-dependent and influenced by different drivers that are still not fully understood. In this study, we combine satellite and reanalysis datasets to understand the atmospheric drivers for the transition from meteorological to agricultural drought, and address the physical characteristics of major drought events in Europe. Particular attention is given to the role of land-atmosphere interaction for compound drought events. Preliminary results highlight where and under what conditions the persistence indices can serve as robust proxies for linking meteorological and agricultural droughts, analyze favorable weather conditions that contribute to drought amplification, and finally address knowledge gaps in this aspect.

Anthropogenic climate change is profoundly altering the global water cycle, increasing both the frequency and severity of extreme precipitation events. This issue is particularly acute in the Mediterranean basin, a recognized climate-change hotspot where heavy rainfall and flooding already impose severe societal risks, as demonstrated by the May 2023 Emilia-Romagna and October 2024 Bologna floods.

This study validates and applies the newly produced Destination Earth IFS–FESOM dataset to investigate the thermodynamic contribution of anthropogenic warming to mean and extreme precipitation over Italy. First, it examines how climate change influences precipitation intensity, assessing whether heavy rainfall events are intensifying at rates consistent with thermodynamic expectations. Second, it explores how this intensification is modulated by large-scale circulation patterns, evaluating differences across Weather Types (WTs) associated with Italian extremes. Finally, it analyzes how anthropogenic warming affects the magnitude and spatial structure of recent flood events, using spectrally nudged storyline experiments to isolate the thermodynamic contribution of climate change in the Emilia-Romagna (May 2023) and Bologna (October 2024) floods.

Two background-climate simulations of the Destination Earth IFS–FESOM dataset were analyzed: Control, representing a 1950s climate, and Historical, representing present-day conditions. Both were produced using spectral nudging toward ERA5, ensuring consistent large-scale circulation while allowing small-scale processes to evolve freely. The simulations were validated against multiple benchmarks (ERA5, CERRA, and ArCIS), confirming their ability to realistically reproduce Mediterranean precipitation variability.

Analyses of precipitation changes employed complementary diagnostics, including quantile behavior, weather-type composites, and storyline case studies. Results reveal a clear thermodynamic amplification of rainfall under warming, particularly in northern Italy and during autumn. The precipitation response depends on the quantile, with the strongest increases occurring in the most extreme events, and is modulated by atmospheric circulation patterns associated with different Weather Types.

Overall, this work confirms that extreme precipitation is intensifying under anthropogenic warming and demonstrates the explanatory power of the storyline framework in attributing observed flood magnitudes to climate change. These findings highlight the urgent societal need to anticipate and adapt to escalating hydrological risks in the Mediterranean region.

Historical meteorological data are crucial for understanding a region’s climate. This research analyzes spatial patterns and trends of daily precipitation in a transect between the Po Plain near Verona, Italy, and Tyrol.

A dataset of daily precipitation was created spanning 1923-2024. Many station’s data were manually digitized from historical annual reports. The entire dataset was quality-checked with suitable procedures, and suspicious data were verified. The time series were homogenized using the R library Climatol.

Climatological precipitation indices were calculated on the homogenized dataset. Additionally, some indices were spatially interpolated for three normal reference periods (1931-1960, 1961-1990, and 1991-2020). Data normality was first verified. Then, an iterative procedure calculated experimental semivariograms to determine the best interpolation parameters. Finally, the kriging with external drift algorithm was executed.

Results confirm established climatic features of pre-alpine precipitation distribution. Notably, however, novel spatial patterns have emerged. The Precipitation Concentration Index suggests that, over the past century, locally the domain has seen an increase in highly irregular precipitation distribution. Furthermore, R95pTOT, R99pTOT, R95p, and R99p indices indicate spatial shifts and significant variations within the domain. The synoptic-scale weather regime variability is currently hypothesized to be the driving force behind this observed pattern.

In conclusion, the present study underlines the importance of reliable historical data from meteorological measurements and encourages the systematic digitization and publication of paper-based annual reports.

Reanalyses are widely used by the scientific community to assess precipitation trends across a broad range of disciplines. In particular, high-resolution regional reanalyses offer unique opportunities to monitor small-scale processes and evaluate the impacts of extreme precipitation events. Nevertheless, previous studies based on ERA5 and other global reanalyses have highlighted limitations in the temporal consistency of long-term annual precipitation estimates, calling for caution regarding their suitability for detecting robust climatological signals.

This study extends the analysis of temporal consistency to four state-of-the-art convection-permitting reanalyses specifically developed for the Italian domain (MORE, CHAPTER, MERIDA HRES, and VHR-REA_IT) by dynamically downscaling ERA5 using different numerical models (MOLOCH, WRF, and COSMO). Long-term seasonal precipitation trends are evaluated against a homogenized observational dataset (UniMi/ISAC-CNR), specifically designed for climate analysis and ensuring temporal consistency. This comparison allows for a detailed quantification of uncertainties in precipitation trends across regions and seasons, helping to disentangle the climate signal from potential artefacts. Such inhomogeneities may be inherited from ERA5 in some cases or further introduced during the assimilation of local observations through techniques such as observational nudging. Finally, similarities and differences among the various reanalyses are discussed in the context of the specific architecture and configuration of each product.

By providing an uncertainty assessment of seasonal precipitation trends, this work will deliver crucial information to reanalysis users on the extent to which trend results can be reliably employed in climate studies and impact assessments. At the same time, it will offer valuable feedback to reanalysis developers, supporting the design of reanalysis products that meet the highest standards of temporal consistency for Italy.

El-Niño Southern Oscillations (ENSO) is the most important variability mode of the climate system on the interannual time scale. In addition to its significant impact on the climate system, it has been suggested that ENSO will play a crucial role in shaping the effects of climate change on future weather and climate extremes. The ability of climate models to correctly reproduce ENSO variability is a strict test of the reliability of their future projections. Particularly, ENSO provides the context for the study of the climate system response to unforced radiative feedbacks, which in turn provides information on the long-term feedbacks related to climate change. This study evaluates the variability of outgoing longwave radiation (OLR) driven by El Niño-Southern Oscillation (ENSO), as simulated by a subset of climate models participating in the Coupled Model Intercomparison Project Phase 6 (CMIP6). This is achieved by using spectral OLR fluxes derived from satellites observations. The first part of the work uses two datasets of spectral fluxes derived from the Infrared Atmospheric Sounding Interferometer and the Atmospheric Infrared Sounder, to investigate the spectral ENSO feedback. Then, ENSO spectral signature is calculated for CMIP6 models using spectral radiative kernels and climate models outputs of water vapor, surface and air temperature fields. The results obtained are compared with those from observations to identify potential biases in the climate models reproduction of the time lag and peak of the radiative response to ENSO driven sea surface temperature anomalies that were not apparent from the broadband observations.

In the climate system many components interact with each other in a nonlinear manner, producing variability over a wide range of spatial and temporal scales. These variabilities, called internal, cannot be traced back to direct responses to external forcings, and are generally chaotic in nature. However, despite the holistic nature of the climate, modeling some of its components separately from others can be fully justified by the immense complexity of the system. Of course, modeling studies of this kind must focus on phenomena and processes that are known not to interact through strong feedbacks with other climate components. Modeling the ocean system through a hierarchy of models (from low-dimensional ones up to high-resolution three-dimensional ones) is, in fact, of great interest in order to disentangle internal variability (here more commonly referred to as “intrinsic”) from that directly connected to atmospheric forcing.

In this context, this communication presents modeling results regarding the intrinsic variability of the Gulf Stream (GS) in the North Atlantic and the Mediterranean Sea (MS). In the first case, the focus is on the separation of the GS at Cape Hatteras due to inertial overshooting, on its dependence on the jet intensity, on the presence of a critical transition and on its possible effect on the AMOC in a scenario of global climate change. As regards the MS, preliminary results are presented based on ensemble simulations using the FESOM-C finite element model.

This research is carried out within the INVMED (“Investigation of the intrinsic variability of the Mediterranean Sea”) project funded by PRIN-2022.

The C3S2_520_CNR project improves the evaluation and quality control (EQC) function of the Copernicus Climate Change Service (C3S) by providing efficient and transparent quality assurance of climate data sets in the Climate Data Store (CDS). The main objective is to answer the question “How, and how well, can I use these data for my purposes?” To this end, the EQC material is organised in a hierarchical structure so that users can find high-level information on fitness for purpose with application examples, detailed requirements that data, metadata and documentation must meet, and scientific assessments with explicit examples of data use. The approach is strongly user-oriented and combines interactive self-assessment tools, stakeholder consultation and continuous feedback to ensure the reliability, usability and long-term sustainability of quality information for climate datasets.

Di fronte a eventi meteorologici intensi sempre più frequenti, determinati dal riscaldamento globale in atto, diventa indispensabile aumentare la consapevolezza collettiva sul sistema climatico. In questo contesto, la scuola e l’università assumono un ruolo chiave nel processo di educazione climatica, offrendo a studenti e insegnanti l’opportunità di comprendere il sistema climatico terrestre, le sue cause e le conseguenze dei cambiamenti che lo caratterizzano. Le attività di recupero e digitalizzazione dei dati meteorologici storici possono rappresentare un importante strumento educativo. A tale scopo, tra i Percorsi per le Competenze Trasversali e l'Orientamento (PCTO) di Sapienza Università di Roma, è stato avviato nel 2023 il progetto Non perdiamo il “Tempo”: adotta una serie storica di dati meteorologici di Roma.

L’attività proposta riguarda la trascrizione di un anno di dati del Registro Meteorologico dell’osservatorio centenario del Collegio Romano (fondato in Roma nel 1787, come Torre Calandrelli), e il successivo confronto dei dati di temperatura con quelli misurati a distanza di 100 anni. Tale percorso si propone di: (i) avvicinare gli studenti e studentesse alla riscoperta dei dati meteorologici di elevato interesse storico-scientifico conservati su supporti cartacei ormai dimenticati; (ii) contribuire al recupero del dato storico mediante la digitalizzazione e, quindi, rendere le persone coinvolte coscienti dell’importanza della raccolta delle osservazioni a lungo termine per studiare la tendenza evolutiva del clima. Ad oggi sono stati coinvolti circa 65 liceali provenienti da due scuole superiori di secondo grado di Roma e provincia.

Il presente contributo descriverà la struttura delle attività del PCTO, i risultati emersi nel biennio di realizzazione e una valutazione del loro impatto formativo.

Italy has long played a pivotal role in the development of meteorology, from inventing key instruments to establishing one of the earliest international observation networks. Over the past three centuries, this legacy has generated a vast and valuable archive of meteorological data preserved in Italian repositories. While numerous initiatives have helped to safeguard parts of this heritage, a large portion of the records still exists only in paper form. These collections are vulnerable to deterioration, placing at risk data of inestimable scientific value for meteorology, climate science, and climate change assessments.

This study highlights recent projects in which the Italian Association of Atmospheric Sciences and Meteorology (AISAM) has been central to national data rescue efforts.

The first is Cli-DaRe@School, a Citizen Science initiative launched in 2022 to digitize previously untapped Italian meteorological observations not yet available in digital format. The project focused on four monographs published by the Italian Hydrographic Service containing monthly temperature (1926–1955) and precipitation (pre-1950) data. Over two academic years, more than 500 students from 10+ high schools contributed to the digitization, producing around 7,931 station records.

The second, Dieci e Lode (2023–2025), targeted the recovery and digitization of meteorological records from former Italian colonies and territories. These included data collected in regions administered by Italy at various times between the late 19^th and early 20^th centuries, such as Eritrea, Somalia, Ethiopia, Libya, the Dodecanese Islands, Albania, Dalmatia, and Istria. The project carried out extensive archival searches to retrieve meteorological observations from these areas and periods.

The third initiative, Cli-DaRe@Images, launched in 2024 and still ongoing, combines education and awareness-raising on climate change with historical data recovery. It engages high school students in the Trentino region to digitize meteorological records preserved in the San Bernardino Library in Trento.

In current practice, short-range atmospheric dispersion modelling is typically conducted using meteorological data obtained from local measurements, typically gathered using "conventional" electro-mechanical sensors, such as cup-vane anemometers and other similar devices.

Recently, wind sensors, such as three-dimensional ultrasonic anemometers, have become more widely available, allowing for higher-resolution measurements.

In a recent paper, the authors investigated the effect of using meteorological data from conventional or ultrasonic anemometers on modelled results in the case of a short Summer-time measurement campaign at Como and found significant ground concentration differences, especially evident in low-speed regimes.

This study extends this result by evaluating the modelled concentration differences over a longer time span at multiple sites in Lombardy equipped with both cup-vane and 3d sonic anemometers, the same used in the previous work.

The new test sites are characterized by different wind statistics and this allowed exploring the effect of different wind speed regimes. The new results confirm the former paper conclusions and reinforce them showing to what extent model outputs may change due to the use of wind sensors based on different technologies.

Volcanic ash released during Etna’s paroxysms poses a significant hazard to aviation at nearby airports, with Catania Airport being the most affected. This is because volcanic ash has a relatively low melting temperature compared to the operational temperature of jet engines, and its ingestion can cause engine failure. Consequently, accurate forecasting the volcanic cloud height, ash concentration, and atmospheric dispersion is crucial. To achieve this, plume-rise models are employed, with one-dimensional (1D) models commonly used for operational forecasting. Among these, FPLUME (Folch et al., 2016) is considered the most comprehensive, accounting for the effects of wind, moisture, latent heat, and variable entrainment coefficients. FPLUME requires several input parameters for the initial conditions, including the mass eruption rate (MER), exit ash velocity, exit ash temperature, exit water fraction, and meteorological variables.

Here we present the results from the analysis of 34 eruption events that occurred between 2011 and 2024. MER was retrieved using the Volcanic Ash Radar Retrieval (VARR) technique (Marzano et al., 2012; Mereu et al., 2015), based on data from an X-band weather radar located in Catania. Meteorological soundings were obtained from ERA5 reanalysis dataset. To account for uncertainties in the initial conditions, a Monte Carlo simulation was performed, running FPLUME 10,000 times for each event while perturbed input parameters. The results show that for events with MER > 8 × 10⁵ kg/s, FPLUME accurately predicts volcanic cloud heights. However, for weaker events, FPLUME consistently underestimates the cloud height. This discrepancy is more pronounced during eruptions occurring under unstable atmospheric conditions, characterized by high Convective Available Potential Energy (CAPE) and very low wind speeds.

These findings suggest that for weak paroxysms, atmospheric processes—particularly convection, which FPLUME does not currently represent—play a dominant role in plume evolution. To address this limitation, a modified version of FPLUME was developed, incorporating a parameterization of convective vertical velocity through CAPE, as well as a revised formulation of vortex entrainment that depends on wind speed rather than the Richardson number. This new model reduced the error between observed and simulated cloud heights, especially for weaker eruptions.

Climate services connect atmospheric science to societal decision-making in the face of rapid climate change. Current provision reveals fundamental tension between commercial interests and equitable access, prompting our advocacy for transformation from market-only models toward a knowledge commons approach that balances innovation with universal accessibility.

Commercial trends in climate services create concerning information disparities. Our analysis across multiple countries demonstrates that profit-driven models frequently exclude vulnerable populations from essential climate data due to cost barriers, creating a paradox where those most exposed to climate risks have least access to adaptation knowledge. This commercialization of publicly-funded research fundamentally weakens the social contract between science and society. In several Western nations, data access restrictions deliberately create lucrative markets, enabling corporations to develop superior risk assessments while constraining public access, thereby reinforcing existing socioeconomic disparities.

However, our research recognizes the legitimate role of private sector innovation in climate services. Drawing on team members' business experience, we acknowledge that commercial applications can flourish alongside equitable data access principles, particularly when specialized tools and interfaces build upon freely accessible core datasets.

Empirical evidence illustrates these inequities' real-world consequences. Asymmetric forecast access undermines adaptation capacity (Lemos et al., 2007), while agricultural insurance applications can paradoxically disadvantage resource-poor farmers despite actuarial benefits (Carriquiry and Osgood, 2012). Developing nations face additional institutional barriers that further widen access gaps (Buontempo et al., 2018).

Promising solutions emerge from European models, particularly the Copernicus Climate Change Service (C3S), which successfully balances open data provision, rigorous quality control (Dee et al., 2024), and stakeholder engagement (Buontempo et al., 2022) while enabling commercial innovation.

Through examination of restriction ethics, successful governance frameworks, and differential adaptation outcomes, we propose a tiered system ensuring universal core service access complemented by specialized commercial offerings. This approach recognizes that publicly-funded scientists must serve broader public interests while accommodating private engagement through mechanisms like open data mandates.

Our institutional analysis, economic modeling, and comparative case studies provide guidance for funding agencies and policy frameworks, advancing both knowledge commons theory and climate justice implementation to ensure climate services enhance collective resilience, particularly for the most vulnerable populations.

Cities are typically affected by the urban heat island (UHI) effect, that is the positive temperature difference between the urban area and the surrounding rural areas. Furthermore, in the context of a changing climate the increase in frequency and intensity of heat waves add up, exacerbating the overheating to which city dwellers are subjected, with severe consequences on health and society.

A full understanding of the distribution of meteorological variables within the perimeter of a metropolitan area is therefore essential to investigate the several issues affecting the urban population (e.g., outdoor thermal stress), so as to identify the neighbourhoods that suffer most from each of these issues, especially during heat waves. Numerical models such as the Weather Research and Forecasting (WRF) model are fundamental tools for this purpose as they guarantee thermodynamic fields with high spatial-temporal resolution and with continuous spatial coverage.

This work presents an innovative WRF configuration recently developed and tuned for the city of Rome (Italy), including a detailed characterization of the urban land use based on the Local Climate Zones at high resolution (grid horizontal size of 500 m in the innermost domain). Such configuration has been applied here to the simulation of a heat wave event that occurred in Rome during the exceptionally hot year 2022 testing different planetary boundary layer schemes and urban canopy models available in WRF.

Simulated values of temperature at 2 meters height and wind speed at 10 meters height were compared with the acquisitions of the 13 weather stations of the ASTI-Network, showing a good agreement for the different Local Climate Zones. Average bias is mostly less in module than 0.5°C, RMSE is about 1°C and the Pearson correlation coefficient is higher than 0.95 for temperature; average bias is mostly less in module than 1 m/s, RMSE is about 1m/s and the Pearson correlation coefficient is higher than 0.8 for wind speed.

These statistical parameters reveal that this tool is suitable for the exploration of UHI-related vulnerabilities in the city of Rome, even during heat waves.

The high thermal inertia of massive buildings (thick masonry, concrete, or stone walls with minimal openings) naturally attenuates and delays outdoor climate variability, offering a potential sustainable solution for the passive preservation of climate-vulnerable heritage collection. This research proposes an approach combining multiple statistical techniques for indoor climate (named also as microclimate) characterization, and for assessing existing buildings as passive or low-energy conservation spaces. The approach is applied to multi-years series of indoor temperature (T) and relative humidity (RH), including the mixing ratio (MR) of moist air, available at the Library Section for Special Collections (LSSC), located within the WWII massive bunker “Dora I” (Trondheim, Norway). Five years of microclimate data (2020-2024), are analyzed to investigate the intrinsic properties of the time series relevant for the microclimate characterization in conservation spaces housing climate vulnerable collections (paper-based objects housed within Dora I). It is found that indoor climate conditions remain stationary, with no discontinuities or trends over five years. Indoor T and MR are delayed of 70 and 50 days and attenuated by a factor of 0.2 and 0.4 compared to outdoor values, demonstrating high thermal inertia and moisture buffer. Findings confirm the suitability of heavyweight, passively conditioned massive buildings in ensuring stable indoor conditions for heritage conservation, supporting sustainable reuse strategies as an energy-saving solution. Given the widespread availability of similar massive structures (often former industrial, military, or infrastructural buildings), their potential is particularly promising.

In questo studio si indaga la variazione a lungo termine dell’esposizione della popolazione al biossido di azoto (NO2) nell’area urbana di Napoli, Italia, durante il periodo 2013–2022. La ricerca integra dati ad alta risoluzione temporale dalla rete regionale di monitoraggio di riferimento con misure ad alta risoluzione spaziale raccolte durante la campagna di citizen science “NO2, NO grazie!”, condotta nel febbraio 2020 con campionatori passivi a basso costo. L’obiettivo principale è stato quello di generare stime ad alta risoluzione delle concentrazioni di NO2 per valutarne l’andamento temporale, l’esposizione della popolazione e l’impatto sanitario in un ambiente urbano densamente popolato.

E' stato utilizzato un modello Random Forest basato sull’uso del suolo (Land Use Random Forest, LURF) per estrapolare i dati sperimentali della campagna di citizen science e per calibrare e armonizzare questi dati con la rete di monitoraggio ufficiale. Questo approccio ha permesso di stimare la concentrazione media annuale di NO2 su una griglia spaziale fine per ciascun anno del periodo di studio. La metodologia ha tenuto conto delle incertezze delle stime di esposizione e ha permesso di valutare la conformità agli standard della Comunità Europea (EC) e alle linee guida sulla qualità dell’aria dell’Organizzazione Mondiale della Sanità (OMS).

I risultati evidenziano che, nonostante una lieve tendenza al ribasso delle concentrazioni di NO2 nel decennio, una larga parte della popolazione napoletana è esposta a livelli superiori sia al nuovo valore guida dell’OMS (10 ug/m3) sia agli obiettivi intermedi, sempre indicati dall'OMS. Nella maggior parte degli anni, circa due terzi della popolazione ha sperimentato valori medi annuali di NO2 superiori a 40 ug/m3, con una riduzione temporanea osservata solo nel 2020 durante il lockdown per l'epidemia da COVID, quando la percentuale di esposti sopra questa soglia è scesa al 6%. La valutazione dell’impatto sanitario, basata su funzioni di risposta concentrazione-effetto consolidate, ha stimato che circa 1.300 decessi all’anno a Napoli sarebbero attribuibili all’esposizione a NO2 sopra la soglia di 10 ug/m3, con una riduzione significativa osservata solo nell’anno del lockdown.

Lo studio dimostra che le campagne di citizen science co-progettate, se integrate con il monitoraggio ufficiale e la modellazione statistica avanzata, possono migliorare notevolmente la risoluzione spaziale e l’affidabilità delle valutazioni sull’esposizione all’inquinamento urbano da NO2. I risultati sottolineano la persistenza di un’elevata esposizione a NO2 a Napoli e l’importanza di politiche di riduzione delle emissioni più rigorose e continuative per raggiungere la conformità agli standard internazionali sulla qualità dell’aria e tutelare la salute pubblica.

Urban Heat Island (UHI) and Urban Pollution Island (UPI) are two interlinked phenomena that exacerbate the vulnerability of cities to climate change, extreme weather events, and deteriorating air quality. While UHI refers to the excess temperature in built-up areas compared to their rural surroundings, UPI describes the accumulation of atmospheric pollutants within the urban canopy. Understanding their interaction is crucial for improving thermal comfort, air quality, and urban sustainability.

This study investigates the UHI–UPI relationship in Rome (Italy) by combining in-situ measurements of meteorological variables and pollutant concentrations using an integrated approach that combined multiple statistical techniques. Data collected over the period 2018–2023 include air temperature, humidity, and wind speed from dense meteorological stations, together with PM₁₀, PM_2.5, NO, NO₂, and O₃ from the ARPA Lazio air quality network.

Results reveal marked temporal variability of UPI intensity, with aerosol and nitrogen oxides peaking in winter under stagnant conditions, and ozone reaching maxima in summer. Seasonal cycles are mirrored by UHII, which is strongest in summer and weakest in winter. Lag-correlation analysis reveals the most significant coupling when nocturnal UHII is shifted backward by one day, indicating that daytime emissions accumulate overnight. Regression analyses identify daily mean air temperature and wind speed as key drivers of UHI–UPI interaction. Spearman correlations indicate negative associations of UHII with NO, PM₁₀, NO₂, and PM_2.5, and a positive correlation with O₃. Case studies (atmospheric stagnation, calm wind days, cold-air pooling nights, and heatwaves) confirm that atmospheric dynamics strongly modulate the intensity and persistence of both phenomena.

These findings highlight the complexity of UHI–UPI feedback and stress the importance of integrated monitoring and modelling frameworks. Science-based insights from this analysis provide valuable guidance for urban planning and environmental management aimed at reducing combined thermal and pollution stress in Mediterranean cities.

Solar ultraviolet (UV) radiation is a major risk factor for a variety of short- and long-term health conditions affecting the skin, eyes and the immune system. At the same time, controlled UV exposure is essential for human health, as it is the primary source of vitamin D. Individual exposure levels are influenced not only by ambient UV irradiance, typically monitored by fixed broad band radiometers, but also by the orientation of exposed body parts, outdoor exposure duration, activity patterns and interactions with the surrounding environment. Reliable quantification of personal UV exposure therefore requires the use of dosimeters, portable UV-sensitive devices - based on radiation-sensitive materials or electronic sensors such as photodiodes - that can be worn by individuals. At high altitudes, where solar UV irradiance is enhanced by elevation, snow cover, and clean atmospheric conditions, both residents and visitors are at increased risk of overexposure. To investigate this issue, a field campaign was carried out from July to August 2025 in the Mont Avic Natural Park, an Alpine protected area in the Aosta Valley (northern Italy). Ten volunteers, including park rangers and seasonal outdoor workers, wore calibrated electronic dosimeters at the wrist to measure personal UV exposure during their daily activities in high-altitude (1500–2700 m a.s.l.) environments. The collected exposure data are analyzed in relation to ambient UV levels measured by a fixed broadband radiometer and classified by type of outdoor activities. The individual exposure was compared to the occupational threshold limit value of UV radiation of 1 SED (1 standard erythemal dose is equivalent to an erythemal effective radiant exposure of 100 Jm^-2) as recommended by the International Commission on Non Ionizing Radiation Protection (ICNIRP). To our knowledge, this study represents one of the first assessments of occupational UV exposure among park rangers, a group that can also serve as a representative proxy for other cohorts of population such as mountain hikers. The results are beneficial for developing targeted sun protection guidelines for both occupational groups and mountain visitors.

TRIGGER (Solutions for mitigating climate-health induced risks) https://project-trigger.eu/ is an Horizon Europe project started in September 2022 dealing with the complex interlinkages between extreme weather and health effects. It involves 22 partners from several European countries working together to understand how to mitigate the incidence of several diseases such as cardio-vascular disorders in a warming climate. Specifically, the project focuses on achieving a better integration between personal health protection and the environment in which choices at personal level can be made to mitigate climate-related health risks. TRIGGER's engines are the Climate-Health Connections Labs (CHC Labs): five selected Labs built in European cities, strategically distributed from south to north Europe to capture the diversity in climate. The role of CHCL is to act as hub for the various TRIGGER activities. Each represents a specific environment and climate-related risks ranging from heat waves to air pollution. Each Lab co-design and implement clinical studies, namely the CrossCLAVIS (cross-sectional study), the LongCLAVIS (longitudinal study) and a retrospective study (RetroCLAVIS) to gather new information about climate-related health conditions and use refined climate and health indicators to understand criticalities and work on mitigation of those. TRIGGER is one of the 6 Horizon Europe projects, BlueAdapt, CATALYSE, CLIMOS, HIGH Horizons, IDAlert, and TRIGGER, form the climate change and health cluster. All cluster projects address climate change-induced health risks and help increase preparedness and adaptation by creating synergies, sharing experiences and knowledge, and developing aligned communication actions to maximise impact of climate change. In this presentation we focus on the findings of the first 3 years of projects with emphasis on the identifications of meteo-climate indicators used to infer health effects as well as the sensitivity to spatial and temporal scales. Starting from heat-waves cases, a methodology is presented to downscale relevant variables to formulate fine-grain novel indicators of heat exposure. The presentation will provide some insight on the relevance of data requirements to advance climate science in the health domain.

The monitoring of atmospheric water vapor is crucial in understanding and predicting severe weather conditions such as heavy rainfalls. This work presents a cost-effective and near real-time approach to estimate Precipitable Water Vapor (PWV) using low-cost dual frequency Global Navigation Satellite System (GNSS) receivers and open-source software to create a near real-time monitoring and alerting service for extreme weather conditions. The positioning method is based on real-time Precise Point Positioning (PPP) with State Spatial Representation (SSR), a messaging standard provided by the International GNSS service (IGS) to disseminate corrections for orbits, clocks, phase-biases and ionospheric delays in real-time. In this process, the Zenith Total Delay (ZTD), i.e., the delay experienced by GNSS signals while crossing the atmosphere, is estimated and PWV can be computed, thanks to the availability of ZTD estimates, and surface pressure and temperature data that are estimated through Global Pressure and Temperature (GPT-3) models. The service allows users to receive warnings about potential extreme weather events based on the sharp increase of PWV in a short period of time, which often indicates an imminent severe weather event. The system structure integrates GNSS data processing, atmospheric data estimation and dissemination of timely and reliable outputs. It also uses cloud computing for data logging to enable further analysis, and supports the potential integration of additional data sources for scalability. This framework offers an affordable and adaptable solution for enhancing weather monitoring, ensuring global applicability, thanks to the use of PPP technique.

The study was carried out within “PAIN AND GAIN - Positioning and INtelligent Alarms supported by a New Dense GNSS Affordable Infrastructure” Project n. 2022P8C7ZA, funded by European Union-Next Generation EU within the PRIN 2022 program (D.D. 104-02/02/2022 Ministero dell’Università e della Ricerca).

Nontuberculous mycobacteria (NTM) are opportunistic agents, with main sources of infection being soil and water. Higher population densities and tropical/subtropical climates were associated with increased pulmonary NTM infection (pNTM). Aim of this study was to assess whether pNTM was associated with climate parameters (temperature, humidity, wind speed, and precipitation) and urbanization.

We conducted a case-control study evaluating meteorological conditions (in the 12 months before the first NTM isolate for every patient) and urbanization in the municipalities of residence of 1,061 adults with pNTM enrolled in an Italian multicenter observational study (2013-2022), compared with a random sample of 10,000 adults from the Italian population.

Climate conditions of the municipalities involved in the study for the period of interest (January 2013- December 2022) were reconstructed using an interpolation technique called MLRLI (Multi-linear regression with local improvement).

The input data comes from various networks of meteorological weather stations, spread all over the Italian territory, measuring temperature at 2-meter height, relative humidity, wind speed and direction, barometric pressure and precipitation. Among these networks it is notable to mention the MNW network (Meteonetwork). Other networks (regional, private and aerodrome networks) are also included in the present analysis. For the rain precipitation field, satellite data are also included, from the GPM (Global Precipitation Measurement) mission.

Results show that densely populated areas were at higher risk for pNTM, even after adjusting for age, sex and meteorological parameters in a multivariable mixed-effects logistic regression model. Hotter, poorly ventilated climates with less precipitation were at higher risk for NTM infection, with OR (Odds Ratio) 95%; CI (Confidence Interval) 0.90 for each km/h increase in mean annual windiness; CI 0.99 for each 10 mm increase in mean annual precipitation; CI 1.21 for each °C increase in mean annual temperature.

This contribution presents the Three-Dimensional Urban Canopy Model (3DUCM), a novel three-dimensional urban parameterization, and its implementation in the Weather Research and Forecasting (WRF) model (WRF/3DUCM). Traditional urban sub-grid scale parameterizations implemented into mesoscale meteorological models typically consider a simple two-dimensional canyon geometry representative of the average urban morphology of the mesoscale model grid cell. On the other hand, 3DUCM is a single-layer urban canopy model that takes into account every single building of the urban area, also considering canyon orientation and explicit crossroads modeling, i.e., three-dimensional effects. Model output incorporates not only average grid cell variables, but also predicted fields down to the building scale.

In this work, WRF/3DUCM is tested in the city of Rome during a summer heat wave. Simulations benefit from detailed information on the geometric characteristics of the city, whereas physical properties of urban materials are assigned based on the local climate zone (LCZ) framework. Simulation results obtained with WRF/3DUCM are compared with those from simulations performed with WRF coupled with the default single-layer and multi-layer urban canopy parameterizations and evaluated against an observational dataset. The results highlight the potential of incorporating single-building details to improve the representation of urban microclimates with mesoscale meteorological models.

A rigorous definition for risk as the product of three factors (hazard, vulnerability and damage) is proposed. The application of this formalism requires: (1) the determination of the hazard as the probability density function (pdf) of a set of variables that correspond to the quantities that are potentially harmful for the conservation of cultural heritage; (2) the determination of the vulnerability as the conditional pdf of the damages produced by the potentially harmful conditions; (3) the damage, i.e., the value of the elements at risk or the costs to their restoration. Examples are given to test the application of this approach, through the processing of indoor and/or outdoor data and the use of dose-response (or loss, or damage) functions, to different study cases.

Recognizing the growing urgency of climate change, many regional administrations are developing tailored adaptation strategies based on the specific characteristics of their territories, including climate patterns, local atmospheric dynamics, topography, and, importantly, environmental and socio-economic factors. However, effectively leveraging this wealth of information to design robust climate strategies remains a challenge. This study presents the comprehensive approach used to establish the knowledge foundation for the mitigation and adaptation strategy of the Trentino region. We highlight three main steps in this process, emphasizing the collaborative effort of multiple stakeholders: 1) Preparing a scientific report reviewing current knowledge and identifying research gaps regarding climate change impacts in the region; 2) Disseminating a synthesis of this information to the general public; 3) Creating reference climate scenarios.
The literature review revealed substantial variation in both the scope and depth of analyses across different scientific sectors concerning climate change impacts in Trentino. This heterogeneity was reflected in the informative report, which employed a structured methodology to simplify and synthesize technical information without compromising scientific accuracy. For the climate scenarios, we considered multiple factors, including topography, data availability, and stakeholder needs. A layered approach was applied, incorporating spatially aggregated results, statistical downscaling, and high-resolution numerical simulations using a convection-permitting model. The study projects a general increase in temperatures, with more pronounced rises in winter minima and summer maxima. Extreme precipitation events are expected to become more frequent and intense, while seasonal patterns vary depending on the emissions scenario considered.

Preventive conservation involves strategies and practices that aim to minimise the risk of damage and losses to cultural heritage before deterioration occurs. Within this framework, microclimate — the climate conditions surrounding an object — plays a key role as it directly influences preservation.

Therefore, investigating microclimates is not merely about collecting data; it requires a meticulous analysis of datasets rather than simply calculating statistical parameters or verifying compliance with thresholds. It demands a thorough understanding of the complex interactions between the environment and the materials, along with an awareness of the climate-induced mechanisms responsible for deterioration.

This integrated approach is essential for assessing past and present conditions that have led to the object's current conservation status, and for guiding the identification of the most appropriate preservation actions in view of future scenarios. In this context, it is crucial to consider the growing impact of climate change and the pressure of mass tourism, both of which pose new challenges to the cultural heritage protection.

The Climate4Heritage laboratory of the Physics Department at Sapienza University of Rome has investigated the microclimate in more than 20 case studies over the past 15 years. This contribution analyses the challenges addressed during this period, from instrumental issues to data analysis, and discusses how combining theory with experiences in microclimate studies can lead to effectively and scientifically support preventive conservation of cultural heritage.

Insurance claim verification is the process through which insurance companies assess and validate the legitimacy and accuracy of claims before approval. This often involves complex procedures aimed at identifying inconsistencies, errors, or potential fraud.

Catwink is a new tool developed for the Italian context, designed to improve the efficiency of claim verification related to natural hazards such as floods, landslides, and earthquakes.

In this work, we present the Catwink components focused on evaluating the plausibility of claims caused by floods and landslides. The methods leverage multiple rainfall datasets and have been calibrated using recent extreme weather events. System performance has been assessed both through selected case studies and statistical analyses.

I medicane sono cicloni mediterranei che acquisiscono caratteristiche tipiche dei cicloni tropicali. Sebbene si verifichino mediamente pochi eventi ogni anno, i medicane hanno riscosso un crescente interesse poiché sono spesso associati a forti impatti (piogge intense, tempeste di vento, mareggiate) che colpiscono soprattutto le coste densamente popolate del bacino del Mediterraneo.

I prodotti satellitari non solo permettono di monitorare l’evoluzione e identificare le caratteristiche chiave dei medicane, ma forniscono anche preziose stime di precipitazione sopra il mare, dove questi cicloni si sviluppano e dove le osservazioni sono solitamente scarse. Nell’ambito del progetto “MEDICANES – Earth Observations as a cornerstone to the understanding and prediction of tropical-like cyclone risk in the Mediterranean” finanziato dall’Agenzia Spaziale Europea (https://medicanes.isac.cnr.it), vengono raccolti diversi prodotti di stime di pioggia da satellite. Lo scopo di questo studio è indagare le potenzialità dell’assimilazione di questi dati in un modello meteorologico ad alta risoluzione “convection-permitting”, al fine di migliorare la simulazione numerica dei medicane. L’assimilazione sfrutta una tecnica di nudging che modifica progressivamente i profili di umidità specifica del modello a seguito del confronto tra pioggia osservata e simulata.

Utilizzando due intensi eventi recenti come casi di studio, i medicane Ianos e Apollo, l’assimilazione si è mostrata capace di migliorare le simulazioni della traiettoria e dell’intensità dei cicloni, inducendo una modifica nei processi diabatici, che può propagarsi e influenzare anche la dinamica a larga scala.

Il radar meteorologico è uno degli strumenti che si presta maggiormente per supportare le tecniche per la previsione di eventi precipitativi con un’elevata risoluzione spaziale e temporale (ossia il nowcasting). Attualmente, esistono numerosi metodi di nowcasting che utilizzano in input serie temporali di echi radar, sfruttando anche tecniche di intelligenza artificiale. Le performance di tali metodi, tuttavia, sono influenzate da vari fattori e le loro capacità previsionali devono essere investigate localmente per applicazioni specifiche a livello nazionale. Questo lavoro si propone quindi di confrontare varie tecniche di nowcasting (tradizionali, basate sulla stima del campo di moto, e più recenti, considerando algoritmi di intelligenza artificiale) sul territorio italiano, al fine di stabilire una metodologia per fissare una soglia minima accettabile di performance (MAP) da utilizzare nell’analisi delle prestazioni di futuri metodi di nowcasting per quantificare il miglioramento rispetto allo “stato dell'arte”.
I risultati presentati sono ovviamente validi per l'Italia, anche se l'approccio è generale e potrà essere riprodotto in altre aree geografiche. Più nel dettaglio, come dataset di test sono stati adoperati circa due anni di stime radar di precipitazione, da gennaio 2022 a febbraio 2024, rivelate dal mosaico radar nazionale ogni 5 minuti, con una risoluzione spaziale di 1 km. Il dataset è stato filtrato mediante l’introduzione di un nuovo processo di screening che consente di selezionare in modo automatico le immagini contenenti un numero minimo di punti griglia distribuiti in modo omogeneo e consistente con i pattern spaziali tipici della precipitazione. Il risultato più evidente è la grande variabilità nelle prestazioni di ogni metodo testato. Inoltre, è stata rilevata una dipendenza stagionale e geografica delle performance. Durante la stagione estiva, infatti, quando i temporali convettivi sono più frequenti, le capacità predittive di tutti i metodi risultano ridotte e ciò suggerisce la necessità per il futuro di implementare moduli di nowcasting specifici per la convezione, possibilmente ottimizzati localmente, per migliorare le capacità di previsione in caso di instabilità atmosferica.

Il progetto DiToNA (Digital Twin of North Adriatic) mira a implementare una piattaforma di gemello digitale per la regione dell’Alto Adriatico, integrando modelli numerici avanzati di previsione meteorologica, oceanografica e ondosa, in grado di assimilare dati osservativi, inclusi quelli satellitari ad alta risoluzione, e produrre simulazioni ad altissima fedeltà spazio-temporale. DiToNA è cruciale per un’area vulnerabile a eventi meteo-marini estremi, quali cicloni mediterranei, la Bora, mareggiate e fenomeni di acqua alta che colpiscono città costiere e infrastrutture dell’Adriatico settentrionale.

La piattaforma combina i modelli WRF (atmosfera), ROMS (oceano) e SWAN (onde) tramite il framework COAWST, consentendo simulazioni tridirezionali e processi multi-scala tra atmosfera, oceano e onde. Le griglie di simulazione prevedono risoluzioni fino a 1 km nelle aree costiere più sensibili e una dettagliata stratificazione verticale nei primi metri sopra e sotto la superficie, per una rappresentazione accurata di scambi termici e turbolenze. L’innovazione chiave di DiToNA risiede nell’uso di tecniche di data assimilation variazionale (3DVAR) e nell’integrazione di dati da radar satellitari SAR, che migliorano significativamente l’inizializzazione dei modelli in aree marine.

La validazione, effettuata tramite indicatori quantitativi (bias, MAE, RMSE, CSI, POD), evidenzia benefici dell’assimilazione soprattutto nelle fasi iniziali della simulazione e nei modelli accoppiati, pur rilevando persistenti criticità nelle previsioni di precipitazioni estreme e una tendenza alla sovrastima degli eventi più intensi. Le analisi condotte su casi di studio emblematici, come il ciclone Detlef (novembre 2019), confermano la capacità del sistema di riprodurre eventi eccezionali, sottolineando la necessità di ulteriori miglioramenti, in particolare tramite reti osservative più dense e strategie di calibrazione dinamica. DiToNA si propone dunque come riferimento nazionale ed europeo per la previsione e la mitigazione degli impatti meteomarini nelle coste mediterranee.

As a result of greenhouse gases mitigation strategies to face climate change, renewable energy plants have greatly expanded during the last decades. Concurrently, extreme weather events are increasing in frequency and intensity, posing significant challenges for renewable energy generation and grid stability. Deep convection, capable of generating strong wind gusts, heavy rainfall and hailstorms, is often responsible for substantial damage to electrical infrastructure. As a result of the small temporal and spatial scales in which convection typically develops, its associated phenomena are difficult to predict accurately.

It follows that improving the ability of weather forecasting models to predict them is essential for making disaster prevention measures more effective.

This study used the regional model WRF-ARW to investigate the capabilities of forecasting a convective event that occurred in Northern Italy on August 28th, 2025. During the event, strong wind gusts uprooted several panels from a large PV plant in Verretto (PV). This event was analyzed using several high-resolution simulations conducted in both reanalysis and forecast mode, with three nested domains at resolutions of 12 km, 4 km and 1.3 km, respectively. As far as reanalysis mode is concerned, ERA5 was used, while two simulations using GFS and IFS drivers were carried out in forecast mode to assess the relative performance of the different boundary conditions in reconstructing this event.

Various meteorological variables were examined to evaluate the performance of each simulation. Convective parameters related to the genesis and development of thunderstorms, as well as surface wind speed and vertical velocity, were analyzed to assess the simulation’s relative performance in accurately reconstructing this event in space and time.

This study has also deepened the ability of regional downscaling in the detection of this localized event at different spatial resolutions.

These preliminary results are useful for guiding the selection of the most appropriate model configuration, such as grid settings and parameterizations, and boundary conditions for describing destructive and localized convective phenomena, which pose an even greater threat to energy infrastructure.

Urbanisation is responsible for deep modifications to the environment, altering the energy balance and local microclimate. In particular, urban areas are known for the so-called heat island effect, associated with a reduction in thermal comfort and, more generally, a series of negative effects on human physical and mental health.

Vegetation (shading, evapotranspiration), soil (support for vegetation, interface between vegetation and water, water reserve), water (integrated management returned to the soil directly or indirectly) and surface cover (thermoradiative properties, evapotranspiration) are considered the main factors capable of counterbalancing the heat island effect, contributing to the creation of urban cool islands. However, the impact of these factors on the urban microclimate is variable and is affected by the mutual influence of the various factors.

In this study, we will examine the impact of the main factors contributing to the development of urban cool islands, using the TEB-Surfatm model on several case studies in France. Parameters such as the colour and thermal properties of coatings, the height and the rate of coverage of the vegetation, water requirements associated with the presence of vegetation will be taken into account in order quantify their impact on the microclimate. The results of this research, developed in collaboration with Vinci Constructions, will be used to create resilient urban environments, reducing the impact of the urban heat island effect.

We present a computationally efficient regional weather prediction system based on the Weather Research and Forecasting (WRF) model and its data assimilation component, WRFDA. The system generates twice-daily 24-hour forecasts (00 and 12 UTC) over the European Alps and surrounding regions at 3.5 km resolution. Surface observations are assimilated using the 3D-Var algorithm, and forecast performance is evaluated against independent surface and radiosonde measurements. The assimilated observations are obtained from both conventional sources (SYNOP observations) and the dataset hosted by Hypermeteo S.r.l., which comprises observations from publicly and privately owned surface stations that are usually not included in assimilation routines. Assimilating these observations results in substantial improvements in near-surface temperature and humidity forecasts compared to control simulations without assimilation. In the first six forecast hours, mean temperature and humidity errors are reduced by up to 0.26 K and 0.18 g kg⁻¹, while the spread of errors decreases by 7–10%. The comparison of forecasts against radiosondes indicates, however, that these surface-based adjustments can increase forecast biases within the planetary boundary layer (PBL). At the same time, the assimilation of surface observations leads to an overestimation of the accumulated precipitation. We show that these side effects can be effectively mitigated by considering a different assimilation algorithm. The assimilation also exhibits different effects across terrain types, with greater benefits in lowlands than in mountainous regions, likely due to limitations in how static covariances spread observational information along terrain-following coordinates. Moreover, the forecast skill displays a clear diurnal pattern, with larger temperature errors occurring over mountains during the day and over plains at night.

Mediterranean Hurricanes (hereafter Medicanes) are high-impact, short-fused events whose rapid intensification and small spatial extent challenge conventional nowcasting systems. Reliable, automatic detection and tracking from geostationary satellite imagery could improve situational awareness and contribute to earlier warnings. Within the MEDICANES project funded by the European Space Agency (ESA), we apply a selfsupervised vision transformer called VideoMAE for learning high-level spatiotemporal features from geostationary satellite IR image sequences. Our aim is to both automatically detect Medicanes and track them from short video clips. The dataset consists of airmass RGB composites over the Mediterranean basin, built by means of the IR channels images collected by the Rapid Scan Service (RSS) of the Spinning Enhanced Visible InfraRed Imager SEVIRI on board the Meteosat Second Generation (MSG) geostationary satellite. We perform self-supervised pretraining of VideoMAE for representation learning, and fine-tune the pretrained model for downstream tasks: binary classification for cyclone presence, and regression for the cyclone-eye center localization. The pretrained-then-fine-tuned VideoMAE model yields robust detection task performance on the imbalanced test set, with overall accuracy exceeding 90% and POD above 95%. The model captures key cyclone dynamics, such as rotational cloud patterns and evolving airmass contrasts without manual feature engineering. The study shows that VideoMAE can provide a scalable and efficient approach to satellite video analysis, demonstrating robust transfer to cyclone detection despite severe class imbalance and delivering consistent center localization. These results suggest that transformer-based video models are well suited for operational use in rare-event detection, offering a practical path toward automated monitoring and early warning of Medicanes.

La previsione meteorologica probabilistica è una tecnica che utilizza ensemble forecasting, ovvero insiemi di simulazioni con condizioni iniziali e parametrizzazioni leggermente diverse per quantificare l'incertezza previsionale.

L'Ensemble Prediction System (EPS) dell'European Centre for Medium-Range Weather Forecasts (ECMWF) genera 50 membri ogni 6 ore. L'utilizzo operativo di questo ensemble per fornire condizioni iniziali e al contorno a modelli regionali ad alta risoluzione richiede una riduzione del numero di membri compatibile con le risorse computazionali disponibili. Si pone quindi il problema della selezione ottimale dei membri da utilizzare.

Questo studio presenta e valuta un metodo di riduzione dell'ensemble che preserva la consistenza statistica del campione originale. La metodologia si basa sulla riduzione dimensionale tramite Principal Component Analysis (PCA) delle variabili prognostiche del modello globale a livelli isobarici selezionati, seguita da clustering nello spazio latente. Da ogni raggruppamento viene estratto il membro più prossimo alla centroide. La numerosità dell’ensemble ridotto è stabilita priori con il numero di raggruppamenti prescritto. I membri selezionati vengono utilizzati come condizioni iniziali e al contorno per altrettante simulazioni con il modello ad area limitata WRF-ARW, mantenendo invariati i parametri del modello per ogni membro.

Il metodo è stato testato su cinque eventi meteorologici severi che hanno coinvolto il territorio italiano. Lo stesso dominio è stato usato per ogni evento. I risultati dimostrano che l'approccio proposto mantiene meglio la dispersione statistica dell'ensemble originale rispetto a una selezione casuale, offrendo una soluzione computazionalmente efficiente e facilmente implementabile per sistemi di previsione probabilistica operativi con numero arbitrario di membri.

Ensemble-based data assimilation algorithms can be used for the objective estimation of the optimal values of uncertain empirical constants in parameterization schemes. This is accomplished by state augmentation: empirical parameters are appended to the model state vector and updated on the basis of flow-dependent ensemble covariances with observable quantities. The method has been used so far as a way of increasing ensemble spread and accounting for model errors in the assimilation process. In this study, we show that parameter estimation results can be useful also for parameterization design, but only under rather restrictive conditions. The error variance of the assimilated observations needs to be as low as that of the state perturbations induced by the estimated parameter, and parametric uncertainty must be the dominant contributor to the total forecast uncertainty. We illustrate the methodology with examples dealing with parameterizations of boundary-layer turbulence and gravity-wave drag.

Accurate seasonal forecasting of water table elevation is essential for the effective management of water resources in unconfined aquifers, especially in the context of climate variability and human-induced pressures. This study presents a novel methodology for predicting water table elevation on seasonal timescales. This approach couples reanalysis and seasonal forecast data of soil moisture with a calibrated non-linear transfer model. This approach uses ERA5 reanalysis and SEAS5 seasonal forecasts to estimate the flux towards the aquifer and predict the water table elevation. A case study in the Umbria region of central Italy demonstrates the model's ability to simulate and predict monthly water table fluctuations. Two modelling strategies were compared: a static calibration approach (OPT 1) and a dynamic calibration approach (OPT 2), in which model parameters were updated by considering different time periods. Both options yielded skilful forecasts across lead times of 1 to 6 months, with OPT_2 showing slightly improved stability in forecast performance metrics. The results confirm the feasibility of incorporating seasonal climate forecasts into operational groundwater prediction frameworks. As expected, the accuracy of the forecasts is limited by the accuracy of the precipitation predictions, particularly during autumn and winter. The proposed framework establishes the basis for anticipatory aquifer management and early warning systems in the context of changing hydroclimatic conditions.

The interpolation of sparse ground data for the initialization of numerical weather forecast, for model validation, and for data assimilation is sometimes performed with geostatistical interpolation. Standard theoretical variograms are often applied. The characteristics of some intense convective events have triggered the idea of a new formula for a theoretical variogram. Strengths and possible difficulties occurring in the application of this variogram model are analyzed.

The Monin–Obukhov Similarity Theory (MOST) provides the framework for describing turbulent exchanges of momentum and scalars in the atmospheric surface layer. Despite its wide application, systematic deviations between MOST predictions and observations persist under non-ideal conditions, such as strong instability, weak turbulence, surface heterogeneity, and the existence of sources and sinks for momentum or scalars. In this study, we quantify the magnitude and structure of these uncertainties using a Bayesian framework applied to virtual potential temperature gradients measured over Bosco Fontana, a deciduous broadleaf forest in Marmirolo (MN), Italy. We reformulate the classical MOST relationship for the mean virtual potential temperature difference as a hierarchical Bayesian model that explicitly separates random measurement noise from structured model error. The hierarchical term allows deviations from MOST to vary systematically with atmospheric stability, while still sharing a common underlying distribution. Posterior estimates are obtained using Markov Chain Monte Carlo sampling, and model diagnostics are used to identify regions of parameter space where MOST assumptions fail. Building on this, we introduce a latent process inference approach to simulate unobserved mechanisms that contribute to MOST shortcomings. These latent processes act as data-driven hypotheses, enabling the formulation of new parametric corrections and extensions of similarity relationships under weakened assumptions. Results and proposed model extensions will be presented in the poster.

Ozone deposition to vegetated surfaces is commonly estimated by 1-D dry deposition models which implement ozone uptake by leaf stomata and ozone disruption on leaf cuticles, branches and soil.

These models use resistance networks where ozone flux is considered as an electric current and ozone concentrations as electric potentials.

Simple models treat the vegetation as a single big-leaf with certain features, while more recent models divide the canopy vertically into multiple layers with a certain number of leaves or branches.

In this work, the total ozone deposition predicted by the two types of models was compared with the vertical ozone flux measured in a broadleaf forest from 2013 to 2022.

Moreover, the ozone concentrations calculated at the top of the canopy using the two model schemes were used to calculate the stomatal uptake and the phytotoxic ozone dose with the indicator (POD1) adopted by UN/ECE to assess the ozone risk for vegetation in Europe.

Aerodynamic resistances were calculated hour by hour according to the MOST and the stomatal resistances were modeled according to Jarvis (1976), while cuticular and soil resistances were kept as constant. Simulations were performed under well-mixed conditions and under different stability conditions, as well as under real soil water content and full water supply for roots, and under isothermal vertical profile of temperature or with measured temperature profile within the canopy.

Results revealed discrepancies between the modeled and the measured total ozone deposition which were attributed to the chemical sink of O3 due to the NO in the trunk space, as found by Finco et al (2018), but that were not modeled in both deposition schemes. Once this missing chemical sink was added, the multilayer model performed better than the single-layer one and predictions were close to the measured fluxes.

Instead, the ozone doses predicted as POD1 by the two models were similar and quite close to the measured ones. However, when the ozone dose absorbed by the whole plant is required - instead of the dose absorbed by a single leaf at the top of the canopy, as defined by UN/ECE- the multilayer model predicted doses that were 40% greater than those predicted by the single-layer model.

Data series from October 2005 to September 2025 from the ISAC-Lecce Micrometeorological Station database (www.basesperimentale.le .isac.cnr.it ) have been analyzed with main attention to the surface water budget and possible 20-years trends. The results are presented as 6-months averages dividing the hydrological year in wet season (October-March) and dry season (April-September), after applying post processing corrections on the half-hour surface fluxes in the database and using the energy budget closure as validation. 20-year trends in surface and soil temperature and surface fluxes show a pronounced increase of all temperatures and sensible heat fluxes, together with a decrease of the latent heat fluxes and a marked increase of the surface-air temperature difference. Together with the observed precipitation tendency to migrate towards the dry season, the observed trends imply a decrease in the calculated annual net infiltration, in agreement with the decreasing trend of groundwater levels measured in the last years in two wells of the underlying aquifer.

Direct measurements of greenhouse gas fluxes (CO₂ and N₂O) were conducted to evaluate the mitigation potential of agricultural management over a pear crop under real cultivation conditions in Emilia-Romagna. Net CO₂ fluxes were measured using the eddy covariance technique, while N₂O emissions were monitored via dynamic chambers coupled with a high-precision gas analyzer.

The pear orchard in Conselice (RA) was impacted in the first year by the May 2023 flood, which disrupted normal crop function. Post-flood, the system shifted from a CO₂ sink to a net source, and N₂O emissions increased due to anaerobic conditions promoting denitrification. The anaerobic soil conditions, elevated temperatures, and nutrient input from floodwaters (e.g., sewage, manure) likely enhanced denitrification, driving this response. Although nitrogen input during the flood could not be quantified, the event acted as a large-scale fertilization, disrupting the planned GHG balance assessment

Measurements continued through 2024, capturing critical data on flux variability and environmental response over a more representative year.

This contribution presents an overview of the activities and results of the INTERFACE project. The project aims to quantify the non-closure of the surface energy balance at different sites in the Alpine environment, where processes related to the lack of closure, i.e., advection due to the development of thermally-driven circulations, are expected to be particularly significant. This objective is addressed by combining flux station and unmanned aerial system (UAS) measurements. The use of the UAS allows spatially distributed measurements around the eddy-covariance sites, which are crucial for the estimation of advection.

The analysis of eddy-covariance data from various sites representative of different Alpine contexts (e.g., valley floor, valley slope, mountain top) and climatic settings (North and South of the main Alpine crest) allows a systematic quantification and comparison of the characteristics of the surface energy balance, including the lack of closure. Particular attention is given to the evaluation of the role of thermally-driven circulations in the non-closure of the surface energy balance, selecting, by means of objective criteria, days with well-developed slope and valley circulations.

The INTERFACE project contributes to the TEAMx international research programme, which aims to improve our understanding of exchange processes in the atmosphere over mountains.

GREEN-POLIS is a two-year research project involving the University of Rome ‘La Sapienza’, Trento and Bologna, founded in 2023 under the PRIN 2022 scheme by the Italian Ministry of University and Research. The project studies the efficacy of selected Nature-Based Solutions (NBSs) in mitigating the negative effects caused by urban climate change, specifically the Urban Heat Island (UHI) and the Urban Pollution Island (UPI). The analyses are conducted in a multiscale perspective, ranging from the street and building scale to the neighbourhood scale, up to the city scale. The final objective is to provide evidence-based analysis, grounded in a rigorous scientific approach, to build a robust and systemic knowledge of the most effective urban NBSs, their potential benefits, as well as the possible side effects.

Such a challenge is addressed by implementing an investigation that starts with the detailed analysis of microscale effects of NBSs on the in-city ventilation, temperature, and pollutant dispersion in prototypical urban geometries, through the use of laboratory experiments and high-resolution numerical simulations. Subsequently, specific building-scale models are applied to improve urban numerical parameterisations for mesoscale meteorological models, allowing for the analysis of an extensive implementation of NBSs within an urban area.

Selected neighbourhoods and the entire city of Bologna, a recognised hotspot of climate change and air pollution, are taken as a practical case study that will be investigated experimentally through an ad hoc experimental campaign, and numerically through building-resolved simulations as well as mesoscale simulations at the city scale. The overall objective of the project will be discussed, giving an overview of the different results obtained. Additionally, key findings obtained through high-resolution simulations at building scales will be discussed.

The Eddy Covariance method relies on the assumption that the wind, temperature, and scalar data series fed as input are stationary up to order 2.

Order 1 stationarity is usually enforced by subtracting the “trends” from the original signals and calculating the second-order moments (variances and covariances) on the residuals so obtained.

To date, many definitions are available for these trends, and various identification methods have been described for each.

The question then arises as to whether these different definitions and computing approaches impact the final eddy covariance results and to what extent.

Studies on this subject have been conducted in the past, for example, within the AmeriFlux community and with respect to the FLUXNET network.

There is still interest in extending these results to micro-meteorological networks for environmental protection, such as SHAKEUP by ARPA Lombardia.

This study was then conducted using data from SHAKEUP sites, characterized by heterogeneous contexts, including suburban and rural areas in Lombardy.

The best results were achieved when site and context dependencies were considered when choosing the method.

This study investigates the phenomenon of the surface urban heat island (SUHI), which refers to the elevated land surface temperatures (LST) observed in urban areas compared to surrounding rural landscapes. SUHI is one of the most obvious indicators of anthropogenic alteration of the environment and directly influences local microclimates, human health, and urban sustainability. Understanding its spatial and temporal patterns is essential for developing effective adaptation and mitigation strategies in the context of global climate change.
The research focuses on a comparative analysis of several large Italian metropolitan areas, using high-resolution thermal data acquired by ECOSTRESS (Ecosystem Spaceborne Thermal Radiometer Experiment on Space Station). This sensor offers a spatial resolution of 70x70 meters and provides frequent observations throughout the diurnal cycle, allowing the detection of fine-scale thermal variations in different urban fabrics and times of day. These capabilities allow for a more detailed characterization of the intensity, persistence, and variability of the urban heat island effect than is possible with sensors with lower temporal and spatial resolution.

To better interpret the observed thermal patterns, LST data derived from ECOSTRESS are integrated with land use and land cover information from the Copernicus Urban Atlas. This approach allows for the assessment of SUHI gradients at different levels of urbanization and building density, distinguishing the specific thermal footprints associated with industrial zones, compact urban cores, and suburban or peri-urban areas.

In addition, a long-term temporal analysis is performed to assess the potential of ECOSTRESS data for monitoring SUHI dynamics over multiple seasons and under extreme weather conditions such as heat waves. By identifying areas most prone to excessive overheating, the study provides crucial insights into urban vulnerability and thermal inequality.

Ultimately, this work contributes to improving the understanding of urban thermal environments and their spatio-temporal complexity, providing a solid scientific basis for evidence-based urban planning. The results support the design of climate-sensitive strategies, including the enhancement of green infrastructure, cool materials, and adaptive urban morphologies, aimed at strengthening urban resilience in a time of global warming.

This work is supported by: Progetto “RETE - Resilience of the Electric Transmission grid to Extreme events” (PNRR innovation grants) (CUP F83C22000740001);

Acknowledgements: This work is supported by ICSC – Centro Nazionale di Ricerca in High Performance Computing, Big Data and Quantum Computing, funded by European Union – NextGenerationEU (CUP F83C22000740001).

This study investigates experimentally how surface roughness affects turbulence and vertical mixing at a sheared density interface forming at the top of a gravity current flowing on a flat surface. The experiments were conducted in a water channel using the lock-exchange technique. Two cases were considered: one in which the channel bottom was smooth and the other when it was made rough by means of a series of parallelepiped elements about one-sixth the height of the gravity current. Feature-Tracking and Planar Laser-Induced Fluorescence techniques were used to measure fluid velocity and
density. The results show a general sensitivity of the mean and turbulent flow variables on the surface characteristics. For the rough case, the density interface is thinner and some of the main parameters involved in the turbulence kinetic energy equation, e.g., buoyancy P, production B and its dissipation rate ε are higher. Similarly, turbulent diffusivities of mass and momentum also depend on roughness, although their ratio remains virtually unchanged going from the smooth to the rough surface. Furthermore, a balance between P, B, and ε was observed in both cases across the whole density interface, thus verifying the applicability of established buoyancy laws proposed in the past (Osborne
1980). Finally, some of the results are consistent with existing field data and numerical simulations, e.g., mixing coefficient agrees reasonably well with literature data irrespective of surface characteristics.

Snow cover plays an essential role in regulating the Earth’s climate but it has significant impacts on human well-being in several parts of the world (e.g. source of freshwater for agriculture and human consumption, source of energy for hydroelectric power). In this study the distribution of snow cover variables over the whole Italian territory which includes the southern part of the Alps and the Apennines chain between 2000 and 2022 using MODIS data acquired from Terra and Aqua platform are analyzed. After preprocessing the data to obtain a binary snow/no-snow field, the start (SOS), length (LOS), and end (EOS) of the snow season were calculated. The LOS mean values which range from 0 to 365 days show the highest values over the Alpine chain with a mean value of about 90 days for elevations above 500 m a.s.l. Conversely, the lowest values are seen over the Po Plain area with about 5 days for elevations lower than 500 m a.s.l. Moving to the south, the Apennine region show higher values again for higher elevations with a mean value of 6 days in the West region and to 10 days in the East region. For all regions LOS clearly depends on elevation, but the large variability in values at the same altitude highlights the influence of other factors (e.g., slope, aspect, latitude, and longitude). Regarding the temporal evolution, the east region of the Apennines is the only region where the series shows a significant trend of -3.2 days per decade. When different elevation bands are considered the LOS series shows a significant negative trend only at elevations higher than 3500 m a.s.l. especially due to the signal observed over the Alps of about -5.1 and -0.6 days per decade.

To further explore snow cover changes, ERA5-Land reanalysis snow cover was analyzed. A good correlation between MODIS-derived snow metrics and reanalysis over the 21-year period was found. Given this, ERA5-Land snow cover trends across its entire time (1951-2022) was further evaluated, offering a longer-term perspective on snow cover variability in the region.

I radar meteorologici sono strumenti fondamentali per la misura e la stima della grandine e sono in grado di fornire informazioni chiave per dedurre la severità delle precipitazioni convettive. Ad oggi, gli approcci più innovativi si avvantaggiano di radar a doppia polarizzazione, che permettono di misurare la precipitazione grandinigena attraverso la riflettività differenziale e altri parametri polarimetrici (e.g. Bechini and Chandrasekar, 2015). Allo stesso tempo, sono state proposte molteplici tecniche che si basano sulla singola polarizzazione. Questi metodi utilizzano misure di riflettività orizzontale per individuare alcune variabili “proxy” dei processi fisici connessi allo sviluppo della grandine. Tuttavia, le misure da radar sono influenzate da alcuni errori sistematici che, localmente, possono rendere incerta la stima del rischio grandine. Per superare queste limitazioni, sono stati sviluppati alcuni algoritmi basati su un’opportuna integrazione di dati radar con altre variabili meteorologiche, come i dati provenienti da strumentazione in-situ o da satellite (e.g. Kunz and Kugel, 2015; Capozzi et al., 2018).

In questo studio, si propone un metodo che possa sfruttare la combinazione dei dati radar con i dati di fulminazione, con il fine ultimo di individuare i temporali grandinigeni in evoluzione. Recentemente, è stata condotta un’analisi sulla relazione che si instaura tra i chicchi di grandine e l’attività di fulminazione nelle nubi convettive sul territorio italiano, dalla quale sono emersi risultati molto promettenti (Vermi et al., 2025). In particolare, l’analisi effettuata si è concentrata su una caratteristica specifica dell’attività di fulminazione, ovvero il cosiddetto “lightning jump” (LJ), che si può definire come un brusco aumento del numero totale di fulmini, che si verifica tipicamente nelle prime fasi di sviluppo di un temporale. Per il 77% dei casi di studio indagati, il segnale di LJ è in grado di classificare correttamente un temporale grandinigeno. Inoltre, la rilevazione del lightning jump permette di definire una serie di indicatori come il numero di LJ misurati, la loro intensità e l’anticipo con cui si prevede la grandinata (in minuti), utili ad ipotizzare quale sarà il diametro massimo dei chicchi di grandine. Infatti, tutti i trend di questi indicatori crescono all’aumentare delle dimensioni dei chicchi di grandine e divengono ancor più robusti in caso di grandine grossa (diametro dei chicchi ≥ 5 cm): in tali circostanze, si rilevano mediamente 7 LJ per evento, un aumento nell’attività di fulminazione di 35 volte in poche decine di minuti e circa 60 minuti di anticipo tra l’osservazione del primo LJ e la caduta dei chicchi di grandine al suolo. A partire da questi risultati, opportune combinazioni di radar a singola o doppia polarizzazione con le variabili connesse alle fulminazioni potrebbero condurre ad un miglioramento delle performance di algoritmi che si occupano del riconoscimento e del monitoraggio dei temporali grandinigeni in modalità operativa (e.g. Capozzi et al., 2022).

Riferimenti bibliografici

Bechini, R. and Chandrasekar, V., 2015. A Semisupervised Robust Hydrometeor Classification Method for Dual-Polarization Radar Applications. Journal Of Atmospheric and Oceanic Technology, 32, 22-47. https://doi.org/10.1175/JTECH-D-14-00097.1

Capozzi, V., Picciotti, E., Mazzarella, V. and Marzano, F. S., 2018. Fuzzy-logic detection and probability of hail exploiting short-range X-band. Atmospheric Research, 201, 17-33. https://doi.org/10.1016/j.atmosres.2017.10.006.

Capozzi, V., Mazzarella, V., De Vivo, C., Annella, C., Greco, A., Fusco, G. and Budillon, G., 2022. A Network of X-Band Meteorological Radars to Support the Motorway System (Campania Region Meteorological Radar Network Project). Remote Sensing, 14(9), 2221. https://doi.org/10.3390/rs14092221

Kunz, M. and Kugel, P.I.S., 2015. Detection of hail signatures from single-polarization C-band radar reflectivity. Atmos. Res., 2015, 153, 565-577. https://doi.org/10.1016/j.atmosres.2014.09.010.

Vermi, F., Capozzi, V., Monte, G., Budillon, G. and Laviola, S., 2025. Lightning jump as precursor of very large hail occurrence: first evidence in the Italian territory. Bull. of Atmos. Sci.& Technol, 6, 21. https://doi.org/10.1007/s42865-025-00104-2.

In the framework of the Hail Hazard in the Mediterranean (H2Med) project, the Multi-sensor Approach for Satellite Hail Advection (MASHA) technique is applied to study severe hail events occurred in Italy during the last years. MASHA reconstruction of hail patterns allows to identify the severity of events, the trajectory of storm and the lifetime of hail clusters into the clouds. Each event is then investigated through the Principal Component Analysis and cluster analysis in aim to identify the large-scale atmospheric conditions that trigger and reinforce hail and super hail events classified by MASHA. The multivariate statistical approach based on Principal Component Analysis and cluster analysis is applied to some atmospheric fields and thermodynamic indices derived from ERA5 reanalysis. Then, a set atmospheric types favouring hail formation is provided. Finally, future changes of the occurrence of large and extreme hail events over the whole Mediterranean basin is derived from CMIP6 climate model projections for the 21^st century under different SSPs scenarios.

The Alpine orography plays a crucial role in modulating convective storm dynamics, yet the initiation and development of severe convection, including hail-producing storms, remain only partially understood. This study presents a case study analysis of convective events observed in the Alpine region, using national composite radar data provided by the Department of Civil Protection, the MASHA satellite product (Laviola et al., 2025, in submission) for the estimation of hail probability, and ground reports from the European Severe Weather Database (ESWD). MASHA is a hybrid advanced satellite technique capable of detecting the hail-bearing convection developing in the Mediterranean basin every 5 min at very high spatial resolution (3-5 km). The selected cases, characterised by significant hail episodes, are analysed to study the spatio-temporal evolution of convective cells and to assess the capabilities of the MASHA method in detecting hail probability. By integrating these three methods, the aim is to assess the consistency and limitations of satellite algorithms in complex terrain environments. This research is part of the broader TIM (Thunderstorm Intensification from Mountains to Plains) project promoted by ESSL, with the aim of improving the understanding of convective phenomena in mountain environments and their implications for impact prediction and mitigation.

Building on this qualitative validation, the MASHA dataset is now being used to perform a preliminary climatological analysis of hail events over the Alpine arc for the recent 5–6 summer seasons. This extended investigation aims to identify spatial patterns and temporal variability of hail occurrence, providing new insights into the influence of complex orography on convective activity.

Dal mese di maggio 2025, gravissimi incendi boschivi hanno colpito vaste aree del territorio canadese, in particolare le province di Saskatchewan, Manitoba e Ontario. Tali eventi sono stati ricondotti alle severe condizioni di temperatura ed aridità del suolo connesse al cambiamento climatico. Come conseguenza, ingenti quantitativi di particolato derivante dagli incendi sono stati immessi in atmosfera, interessando la troposfera e la bassa stratosfera. Gli intensi venti occidentali che caratterizzano le alte quote delle medie latitudini hanno trasportato i fumi attraverso l’Atlantico settentrionale, fino ad interessare, a partite dai primi di giugno, l’Europa centro-occidentale, sia in quota che negli strati più bassi. L’analisi correlata delle back-trajectories ottenute dal modello Hysplit, delle elaborazioni di immagini satellitari e dei dati prodotti dalla modellistica CAMS, ha permesso di individuare alcune delle fasi salienti del trasporto di aerosol derivante dagli incendi canadesi, evidenziando poi il sopraggiungere di una non trascurabile componente di particolato di origine sahariana. Sfruttando i dati registrati dalla rete di stazioni AirQino (https://www.airqino.it), è stata svolta un’analisi delle serie temporali dei valori di concentrazione di PM10 e PM2.5 misurati durante le prime due decadi del mese di giugno. Le stazioni AirQino, che integrano sensori per la rilevazione dei principali inquinanti atmosferici, di gas serra e di parametri meteorologici, attualmente garantiscono una notevole copertura del territorio italiano e sono operativi con alcuni punti di misura in Francia (Cannes e Marsiglia), Spagna (Barcellona), Ungheria (Budapest e Debrecen) e Romania (Bucarest). In particolare, le misure su base oraria di concentrazione di PM10 e PM2.5 hanno permesso di valutare l’evoluzione temporale e spaziale degli eventi connessi al trasporto dei fumi provenienti dagli incendi canadesi e del contributo desertico sahariano.

The Boundary-layer Air Quality-analysis Using Network of INstruments (BAQUNIN) supersite has been operational since 2017, providing high-quality atmospheric composition data for both scientific research and satellite validation activities. BAQUNIN involves the coordination and the synergistic exploitation of a number of ground-based instruments operated at the Sapienza University campus (downtown Rome), CNR-ISAC (Tor Vergata), and CNR-IIA (Montelibretti), i.e., across urban, semi-rural, and rural sites around Rome. The supersite, promoted by the European Space Agency (ESA) and the EUropean METeorological SATellite system (EUMETSAT), is designed to validate current and future satellite products (trace gases, greenhouse gases, aerosols, clouds) and to investigate the physical processes driving boundary layer evolution in complex Mediterranean urban environments.

Within the BAQUNIN framework, the Raymetrics Aerosol Profiler (RAP) represents a key instrument continuously operating at the Sapienza site since 2021. RAP is a single-wavelength elastic lidar (1064 nm) acquiring vertical profiles of aerosol backscatter and extinction up to 15 km with high spatial (3.75 m) and temporal (10 s) resolution.

Following a 2024 refurbishment, RAP has been upgraded with full scanning capability, enabling novel observation strategies. In particular, the Horizontal Pointing Mode (HPM) allows the system to probe the lowest portion of the urban boundary layer over distances up to 5 km, a region typically inaccessible to conventional lidars and ceilometers, at very high spatial and temporal resolution (3.75m, 10 sec).

Operating RAP in HPM provides unique observational capabilities for the validation of aerosol loads and atmospheric corrections in very high-resolution satellite optical missions, as it captures spatial and temporal scales directly comparable to those of satellite products. Thus, RAP strengthens BAQUNIN’s role as a unique infrastructure for investigating aerosol dynamics in the urban boundary layer and for supporting the calibration and validation of new-generation satellite observations.

Convection plays a crucial role in redistributing energy within Earth’s atmosphere and is often associated with cloud formation and severe weather events, including hailstorms that can cause significant damage to infrastructure and property. In recent decades, Italy and the broader Mediterranean Basin have experienced a rising trend in such extreme events, highlighting the need for improved observational capabilities and retrieval methodologies to analyze convective storms, particularly those producing hail.

This is the primary objective of the PRIN 2022 project “Convection Characterization via Synergistic GEO and LEO Satellite Observations”, which aims to investigate convection using data from the EarthCARE (EC) mission—featuring the highly sensitive 94-GHz Cloud Profiling Radar (CPR) with Doppler capability—alongside observations from the METEOSAT Rapid Scan Service (RSS).

This study presents analyses of convective case studies that occurred over the Mediterranean region in 2024 and 2025. Convective clouds are examined using convection products from the EUMETSAT Satellite Application Facility in support of nowcasting and very short-range forecasting, derived from Meteosat RSS data. Additional insights into updrafts and overshooting tops are obtained through EarthCARE CPR measurements, including radar reflectivity and vertical velocity profiles. The Multi-sensor Approach for Satellite Hail Advection (MASHA), a novel multi-instrument technique designed for real-time tracking of hail-producing clouds, further complements the analysis.

Key case studies are discussed to assess the complementarity and effectiveness of combining active and passive, GEO and LEO satellite observations, with particular focus on the challenges of interpreting Doppler velocity data for identifying updrafts within convective systems.

Snow precipitation plays a crucial role in the global water cycle and energy balance of Earth's climate system, particularly in Polar regions, where it significantly impacts the ice mass balance of polar caps and ice sheets.

The EarthCARE mission, a collaborative effort between ESA and JAXA, is designed to capture the microphysical variability of clouds and precipitation. Its W-band (94 GHz) Cloud Profiling Radar (CPR) enables tracking the formation and evolution of precipitation. But quantitative snowfall remote sensing presents challenges due to the highly variable microphysical and electromagnetic properties of ice crystals and aggregates on small spatial and temporal scales, which is not fully captured but the retrieval algorithms. Moreover, spaceborne radars cannot sample the lowest atmospheric layers because of ground clutter. This makes the estimation of the snowfall at the surface very challenging when significant variations in precipitation occur within the few hundred meters above the ground (Bracci et al., 2022).

To address this, a methodology for estimating reflectivity and Doppler velocity at 94 GHz from K-band (24 GHz) Doppler spectra collected by a Micro Rain Radar (MRR, Metek) and coincident disdrometer observations has been developed and tested in Antarctica (Bracci et al., 2023). This method, known as K2W, exploits the synergy between two commonly available Antarctic instruments to validate satellite-based W-band radar products.

With the release of EarthCARE’s first L2 CPR products for December 2024 and January 2025, this approach has been replicated using EarthCARE overpasses near the Italian Antarctic station “Mario Zucchelli” (MZS), where an MRR and a disdrometer have been operational since 2016 in the framework of the project “Antarctic Precipitation Properties” (APP) of the Italian National Antarctic Research Program (PNRA), also integrating data from a weather station and a ceilometer from the ENEA Italian Antarctic Meteo-Climatological Observatory (IAMCO).

Snowfall events at MZS coinciding with an EarthCARE overpass (point-to-line distance < 20 km) happened on seven occasions since the start of EarthCARE data delivery, with five virga occurrences and at least one good match (April, 30th 2025). A comparison of retrieved physical quantities from ground-based and satellite observations is presented and critically analysed. A statistical analysis on the long time series quantifies the expected frequency of virga occurrence at MZS.

Negli ultimi anni, l’utilizzo dei segnali provenienti dalle costellazioni di Global Navigation Satellite Systems (GNSS) si è affermato come una tecnica affidabile per la stima continua del contenuto di vapore acqueo atmosferico, attraverso la determinazione del ritardo zenitale troposferico (Zenith Tropospheric Delay, ZTD). Nell’ambito di alcuni progetti INTERREG (PROTERINA-3 Évolution, PROTERINA4Future) è stata realizzata una nuova infrastruttura GNSS-meteo per il monitoraggio integrato delle condizioni atmosferiche su mare tramite sistemi installati a bordo di una flotta di navi di linea operativa sull’alto Tirreno. Tale infrastruttura combina reti di ricevitori GNSS permanenti con stazioni meteorologiche di superficie, consentendo la generazione in tempo quasi reale di prodotti come ZTD, Integrated Water Vapor (IWV) e parametri termoigrometrici superficiali, con elevata risoluzione temporale e spaziale.

Al fine di validare le prestazioni del sistema e quantificare l’accuratezza dei prodotti derivati, è stata organizzata una campagna di radiosondaggi sperimentali a basso costo presso le stazioni della rete, con il rilascio di palloni meteorologici dotati di sensori per la misura diretta dei profili verticali di temperatura, pressione e umidità relativa. Le osservazioni dei radiosondaggi, considerate lo standard di riferimento per la calibrazione e la validazione dei modelli atmosferici, sono state confrontate con le stime simultanee fornite dalla rete GNSS e dai sensori a terra.

Il presente lavoro descrive da un lato le prestazioni del sistema GNSS-meteo di osservazione, il primo operativo per un lungo periodo (oltre 4 anni), dall’altro il sistema di radiosondaggio a basso costo sviluppato appositamente per la validazione. Sono stati analizzati i dati di confronto per i lanci effettuati durante differenti stagioni e con diverse condizioni atmosferiche. I risultati mostrano una buona correlazione tra i valori di IWV derivati da GNSS e quelli stimati dai radiosondaggi. Il lavoro presenta infine anche un confronto tra queste osservazioni ed i dati ottenuti da modelli di rianalisi (MERRA-2, ERA5). Questa validazione dimostra l’affidabilità di entrambi i sistemi di misura - GNSS-meteo e radiosondaggi sperimentali - ponendoli come utili strumenti da utilizzare in maniera complementare o in alternativa a sistemi di misura convenzionali.

I pluviometri e i disdrometri sono strumenti fondamentali per ottenere informazioni dirette sulle caratteristiche delle precipitazioni in un determinato sito. Tuttavia, le loro misure possono essere influenzate dalla presenza del vento. In questo studio sono stati identificati e quantificati gli effetti indotti da questo fattore ambientale. In particolare, per valutare le prestazioni del disdrometro in condizioni di vento, sono stati analizzati i dati raccolti da due disdrometri Thies Clima LPM e da sensori di vento installati nelle città di Pescara e Roma.

Il set di dati copre per Pescara il periodo da luglio 2021 ad agosto 2024, sebbene includa interruzioni significative, mentre per Roma gli interi anni 2023 e 2024. In primo luogo, lo studio presenta le principali caratteristiche dei siti in termini di distribuzione del vento e della pioggia, nonché le loro distribuzioni congiunte. Quindi, vengono quantificati gli effetti del vento sulle misure dei disdrometri in termini di errore sistematico associato alla stima della DSD (Drop Size Distribution).

I risultati indicano che a Pescara le DSD non corrette dal vento differiscono, in media, del 10.6 % in termini di errore medio assoluto percentuale rispetto alle DSD corrette e a Roma del 6.3 %. Le correzioni sulle DSD si riflettono sui valori di intensità di precipitazione. In questo caso, le differenze tra le intensità di precipitazione ottenute da DSD non corrette per effetto del vento e da DSD corrette sono di 0.26 mm/h per Pescara e 0.23 m/h per Roma in termini della radice dell’errore quadratico medio. Tali differenze risultano statisticamente significative.

Successivamente, poiché il vento ha effetti anche sulla misura pluviometrica, e tale effetto varia al variare delle dimensioni delle gocce (e quindi della DSD) i dati disdrometrici corretti sono stati utilizzati per correggere le misure di precipitazione di un pluviometro posto a qualche metro dal disdrometro di Roma e di due pluviometri installati nei pressi del disdrometro di Pescara, nonché di una rete di 25 pluviometri nella regione Calabria. Gli effetti della correzione sono valutati confrontando i valori corretti con quelli non corretti. Queste differenze sono risultate statisticamente significative.

Understanding temperature variations within the complex Urban Canopy Layer (UCL) is challenging due to limitations and discrepancies between temperature measurements taken in urban canyons or in other more practical nearby positions, such as rooftops. The key question is then how much these measurements differ and what factors contribute to these differences. According to the guidance of the World Meteorological Organization (WMO), measurements within urban canyons are recommended, while rooftop observations are not encouraged for urban monitoring due to “potentially anomalous microclimatic conditions”. Questions about the representativeness of rooftop data are particularly relevant given the increasing number of rooftop sensors deployed through citizen science. This study aimed to address this knowledge gap by comparing temperatures within the UCL using two sensors: one located on a rooftop, and the other positioned within the canyon. The temperature difference between these two nearby locations followed a clear diurnal cycle, peaking at over 1 °C between 12:00 and 16:00 local time, with the canyon warmer than the rooftop. This daytime warming was primarily driven by solar radiation and, to a lesser extent, by wind speed, but only under clear-sky conditions. During the rest of the day, the temperature difference remained negligible.

This methodology was extended to a broader network of low-cost temperature sensors located near rooftop stations deployed in the city of Rome, and similar results were found.

In many practical applications (e.g., wind energy assessment, atmospheric pollutant dispersion modelling, and others), knowledge of wind speed is of paramount importance, and errors in its measurement may hamper the application usefulness of technical results.

To date, most wind speed measurements have been conducted using cup anemometers. These sensors are known to show measurement distortions under slow wind conditions, the frequency of which is significant at locations worldwide.

This study aimed to investigate the impact of these measurement distortions at sites characterized by different wind speed statistical distributions, and was conducted by comparing the 10-min average speed from cup and 3D sonic anemometers at various SHAKEUP sites (ARPA Lombardia).

The results show a cup anemometer-induced change in the measured wind statistics, particularly at low wind speed regimes. The medium and high wind speed regimes showed substantial agreement.

Possible strategies for mitigating the statistical perturbation are also considered and discussed in view of applications (e.g., dispersion modeling).

In recent years, data-driven approaches have emerged as alternatives to traditional physics-based retrievals, taking advantage of machine learning techniques such as learnable pseudoinverse, random forests, or deep learning architectures. These methodologies generally rely on training datasets derived from simulations or observational databases and exploit non-linear relationships between measured radiances and atmospheric parameters without explicit forward modeling. In this work, a deep learning architecture, based on the latent twins approach (originally introduced in theoretical works such as Chung et al., 2025), is applied to IASI spectra, with the goal to assess the robustness of this method for retrieving atmospheric profiles, including temperature, water vapor, ozone, surface emissivity, and surface temperature, in real-world clear-sky conditions. The algorithm is first applied on synthetic radiances derived from the NWP SAF database using the fast radiative transfer code sigma-IASI/F2N (Masiello et al, 2024). After validating the architecture on synthetic data, the algorithm is applied to IASI Level 1C observations, along with their corresponding Level 2 products which serve as reference to evaluate the reconstruction accuracy of the autoencoder-based retrieval. The retrieval performances are discussed along with possible strategies to provide an error analysis for the reconstructed thermodynamical profiles.

Continuous monitoring of pollutants and greenhouse gas vertical concentrations in the atmosphere allows for air quality studies of a region and for long-term studies of trends and seasonal behaviors. Ground-based remote sensing instruments are usually exploited for this scope.

Since 2021, CNR-ISAC has managed two MAX-DOAS SkySpec-2D, one in Rome Tor Vergata and the other in San Pietro Capofiume, located in the middle of the Po Valley. These instruments measure scattered atmospheric spectra in the UV-VIS range. From them, we retrieve the total vertical column densities of NO2 and O3. We also apply our algorithm DEAP to retrieve the tropospheric profiles of NO2 and aerosol extinction, and their integrated value of tropospheric concentration and aerosol optical depth. These instruments are compliant with the FRM4DOAS network, an ESA activity aimed at harmonizing operations from DOAS instruments all around the world. Since mid-2024, in the framework of the PNRR-EMM project, we expanded the net of SkySpec in Italy by installing two new instruments, one in Monte Cimone at 2165m a.s.l., and the other in Bologna at the CNR-ISAC facility. Here we also installed a Fourier Transform Spectrometer (FTS), the EM27/SUN. It measures direct-sun spectra in the NIR range, from which we retrieve total columns and dry-air mole fractions of CO2, CO, CH4, and H2O. The instrument is part of the COCCON network, which supports the TCCON network, made of IFS125 HR FTS. By the end of 2025, we plan to install one IFS125 HR in the Bologna CNR facility in the frame of the project PNRR-ITINERIS.
In addition to air quality studies, these measurements are useful for satellite validation purposes. This activity is important as satellite data quality must continuously be upheld by ground-based measurements.

Here we will report an overview of the instruments' operations, the retrieved products, and comparisons against satellites.

Meteonetwork (MNW) is a Citizen Weather Stations (CWSs) initiative established in 2002 and continuously operating since then. It provides high spatial-density meteorological coverage across Italy. This study evaluates the consistency between MNW and National Oceanic and Atmospheric Administration (NOAA) observations for temperature (T) and pressure (P). NOAA stations located in Italy were selected as reference sites, while MNW stations within a 10 km radius were paired accordingly. A total of 49 NOAA and 93 MNW stations were analyzed using data collected on five fixed days per month from January 2023 to April 2025. The average acquisition frequency was 5 minutes for MNW (for both T and P), and 20 minutes and 3 hours for NOAA temperature and pressure data, respectively.

To ensure comparability, MNW observations were synchronized both temporally and altimetrically. Temperature values were adjusted to match the elevation of the corresponding NOAA station, while pressure data, already reduced to mean sea level in both networks, required no further correction. Temperature (ΔT) and pressure (ΔP) differences were computed for each station pair. Outliers were filtered using the Interquartile Range (IQR) method, and the 90th percentile of the RMSE of the filtered differences was adopted as the performance threshold for MNW stations.

The comparison yielded the following statistics for ΔP: mean = 0.34, standard deviation = 0.44, RMSE = 1.13, and 90th percentile of RMSE = 2.08 hPa; and for ΔT: mean = –0.33, standard deviation = 1.10, RMSE = 1.27, and 90th percentile of RMSE = 1.90 °C. These results demonstrate the stability and strong coherence between the two networks, highlighting MNW potential for meteorological applications that require dense and real-time observations.

This study represents a preliminary step toward evaluating the consistency of the main environmental sensor networks in Italy, with future analyses planned to include MISTRAl portal data.

The phenomenon of the urban heat island (UHI) has been observed since 1810 when a British scientist, Luke Howard, observed that the city of London is warmer than the rural surrounding. Since then, the concept of UHI has been well documented a, with many studies investigating the factors responsible for its formation and development, namely a decrease in vegetation and evapotranspiration, a rise in low-albedo, dark surfaces, and an increase in anthropogenic heat output in the urban cores. However, since the UHI is significantly affected by the geographic features and climatic conditions, the understanding of the topic remains quite limited especially in some areas. For instance, a significant gap remains in understanding the interaction of the UHI with heat waves (HW) conditions, with contrasting results in different areas. This study examines how UHI patterns change in space and time during extreme heat events in the city of Bologna, situated in the Po Valley in Italy, a well-known climate change and pollution hotspot. By combining an LCZ-based approach with the calculation of thermal comfort indices to assess the degree of human heat stress among different types of urban morphologies, the study aims to provide a better understanding of the UHI and HW interactions at the urban scale. The results indicate that the UHI in the area is significantly higher at nighttime, when the historic core of Bologna with dense LCZs (layers 2 and 3) experiences higher thermal stress. The level of heat stress reaches maximum levels during the HW period, in which greener or more open LCZs mitigate thermal discomfort. These results may be useful for informing targeted urban planning strategies for climate adaptation in the area.

A multi-scale observational analysis of a 1.6 km wide IF3 tornado in Northern Italy is conducted using radar and sounding data, ground weather stations, and damage surveys. The tornado occurred close to Alfonsine, along the Adriatic coast, on July 22, 2023, in one of the most tornado-prone regions of Europe. An initially hail-bearing supercell (which produced hailstones up to 10 cm in diameter) became tornadic as it approached a dryline bulge. During the transition from a hail-dominant to tornadic storm, the long-lived supercell generated a damaging Rear-Flank Downdraft (RFD) surge, with unusually cold wind gusts reaching 40 m/s. A dry and hot air mass from the southwest was partially ingested by the mesocyclone just before the tornadogenesis occurrence. At the same moment, the storm was also ingesting from the east a maritime air mass with very high values of equivalent potential temperature. A seamless wind damage pattern, transitioning from damage caused by straight-line wind gusts to tornadic damage, suggests that the tornado may have developed from the stretching of small-scale pre-tornadic vertical vorticity maxima within the RFD. As in other case of significant tornadoes in Northern Italy, the environment was characterized by strong deep layer shear and conditional instability, but weak low-level wind shear. However, numerical simulations indicate that along the dryline the low-level storm relative helicity and vertical vorticity were stronger, suggesting a higher tornado potential. The tornado resulted in only 14 injuries, likely because it impacted a sparsely populated area. Considering that past significant tornadoes in the region affected much more densely populated areas, and since no tornado warnings or shelters are currently in place, there are growing concerns about the potential catastrophic consequences of a future significant tornado in the highly populated areas of northeastern Italy.

On 29 October 2024, an intense supercell formed along the Valencian coastal area. The storm brought about 600 mm of rainfall with extensive damage and casualties. Although the large scale atmospheric patterns that trigger the event is quite marked and relevant, the role of the sea structure is not totally clear. WRF (Weather Research and Forecasting system) simulations were conducted to examine how oceanic factors – sea surface temperature (SST), SST anomalies, upper-ocean vertical stratification, and ocean heat content – influenced the development of the extreme weather event over Valencia area. The model was run at a resolution of 3-1 km, with the SLAB Ocean model activated in order to model the structure of the mixed layer depth consistently with the atmosphere. Simulations, under this configuration, is capable of represent the phenomenon very realistically, in space and time. The intensity reached by the control simulation is about 590 mm/12hr, close to the observed reality. We used observed SST, mixed layer depth produced by the CMEMS model reanalysis and anomalies are based on the CMEMS dataset 1987-2010. The results suggest that ocean–atmosphere interactions accounted for around 25% of the storm’s peak intensity, while storm genesis and track remained largely unaffected to ocean structure. Among the tested factors, pronounced upper-ocean stratification and vertical lapse rate enhanced surface latent heat fluxes and invigorated convection, thereby intensifying the storm. In contrast, SST anomalies had only a minor, spatially inconsistent influence due to their patchy distribution prior to the event. These findings underscore the importance of accurately representing upper- ocean structure and heat content in mesoscale models to improve forecasts of extreme Mediterranean weather events.

In recent decades, Europe has experienced a marked intensification of extreme weather events, with notable implications for population thermo-hygrometric well-being. This study investigates time series of atmospheric variables from the ERA5-Land and ERA5-Heat datasets, provided by the Copernicus Climate Change Service (C3S), spanning the period 1961–2024. Extreme events are identified with respect to the 1961–1990 climatological baseline. The analysis examines spatial and temporal variations in the frequency and intensity of heatwaves, identified using 2-m air temperature, and droughts, quantified through the Standardized Precipitation Evapotranspiration Index (SPEI), as well as their co-occurrence as compound events. The associated impacts on thermo-hygrometric stress are assessed using the Universal Thermal Climate Index (UTCI), derived from ERA5-Heat data. Four daily event-based scenarios, identified from the analysed datasets, are considered: (i) a reference case without extremes, (ii) heatwaves only, (iii) droughts only, and (iv) compound heatwave–drought events.

The results reveal a clear temporal evolution of these events, with significant increases in frequency and severity, particularly in recent decades, and demonstrate that the co-occurrence of heatwaves and droughts amplifies the risk of thermo-hygrometric stress compared to single events. Therefore, this study provides an integrated assessment of extreme climate risks, contributing to a better understanding of the thermo-hygrometric stress under extreme conditions.

The findings of this research are particularly relevant for urban planners and policymakers, as they can highlight the social implications of extreme climate events and can guide the design of tailored adaptation strategies. In particular, the results can support the development of early warning systems, inform the planning of resilient urban infrastructures, and provide actionable insights to enhance the resilience of European communities.

This study investigates outdoor thermal comfort conditions in a selected area of Lecce (southern Italy) through the Universal Thermal Climate Index (UTCI), employing a microclimate modeling approach. The analysis is conducted using the ENVI-met software, a three-dimensional Computational Fluid Dynamics (CFD) model designed for simulating surface–plant–air interactions at the microscale.

In the first phase, two scenarios are simulated: (A) the current configuration of the study area and (B) a redesigned future scenario developed within an urban redevelopment plan approved by the Municipality of Lecce, as part of the Ministerial Experimental Program of Interventions for Adaptation to Climate Change in Urban Areas (Ministero della Transizione Ecologica, 2021). Meteorological inputs are derived from ERA5 reanalysis data for the period 1991–2020, from which a representative summer day is applied to both scenarios. This phase aims to assess the microclimatic effects and improvements in outdoor thermal comfort associated with mitigation strategies, including the introduction of permeable surfaces, increased vegetation and shading and water features (fountain).

In the second phase, Scenario B is further analyzed under projected future climate conditions, using climate projections as meteorological inputs. The objective is to evaluate the effectiveness of the proposed mitigation measures in enhancing outdoor thermal comfort across three future time horizons: near-term (2021–2040), mid-term (2041–2060), and long-term (2081–2100) (Lee et al., 2021).

Overall, the study quantitatively highlights the potential benefits of climate-sensitive urban design strategies, showing how nature-based and structural interventions can mitigate urban heat stress and enhance thermal comfort under both current and future climatic conditions.

References

Lee, J.-Y., Marotzke, J., Bala, G., Cao, L., Corti, S., Dunne, J. P., Engelbrecht, F., Fischer, E., Fyfe, J. C., Jones, C., Maycock, A., Mutemi, O., Ndiaye, O., Panickal, S., Zhou, T., et al. (2021). Future global climate: Scenario-based projections and near-term information (Capitolo 4). In IPCC, Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press.

Ministero della Transizione Ecologica. Decreto direttoriale n. 117 del 15 aprile 2021, Programma sperimentale di interventi per l’adattamento ai cambiamenti climatici in ambito urbano — Gazzetta Ufficiale della Repubblica Italiana, Serie Generale n. 135, 8 giugno 2021.

The production and chemical composition of Secondary Organic Aerosols (SOA) are key factors in urban air quality. Analyzing data from a summer campaign in Beijing, conducted in 2017, we found nocturnal peaks of NOz (NOy - NOx) up to 40 ppb, correlated with nocturnal high concentrations of NO and NO2 and SOA formation the following day. We employed the Framework for 0-D Atmospheric Modeling (F0AM), based on the Master Chemical Mechanism (MCM), coupled with the Washington Aerosol Module, based on SIMPOL mechanism for the particle phase modeling, to run simulations investigating the speciation of ONs in both gas and particle phases anf to correlate the nocturnal NOz peaks, which we suggested are mainly in the gas phase, with the diurnal particle growth events observed in Beijing.

Il presente lavoro è incentrato sullo studio delle caratteristiche meteorologiche del ciclone simil-tropicale Qendresa, verificatosi nel Mediterraneo centrale tra il 6 e l'8 novembre 2014, utilizzando il modello WRF ad area limitata. Qendresa ha avuto origine nelle prime ore del 6 novembre attraverso il processo di “lee cyclogenesis” sul versante orientale della catena dell'Atlante. Successivamente il ciclone si è approfondito spostandosi verso l'isola di Pantelleria, per poi traslare verso sud-est, attraversando l'isola di Linosa, dove il suo nucleo ha raggiunto il minimo assoluto di pressione al livello del mare. Il sistema ha proseguito poi verso Malta e, già indebolito, si è spostato al largo della costa orientale della Sicilia, descrivendo una traiettoria ad anello durante la mattina dell'8 novembre, prima di spostarsi definitivamente verso est, in direzione della Grecia. Complessivamente, l'evento ha provocato tre vittime e ingenti danni strutturali nelle aree colpite.

L'obiettivo principale di questo lavoro è quello di riprodurre la struttura e l'evoluzione del ciclone, raggiungendo la migliore coerenza possibile con le osservazioni disponibili, durante il periodo di 36 ore dalle 00 UTC del 7 novembre alle 12 UTC dell'8 novembre 2014. Per migliorare l'identificazione dei processi fisici chiave e aumentare l'accuratezza della traccia simulata del ciclone, nel modello sono stati assimilati dati osservativi e di rianalisi a diversi livelli verticali e intervalli temporali. Sono stati testati due approcci di assimilazione dei dati: il filtro di Kalman d'insieme (EnKF) e l'assimilazione dei dati variazionali tridimensionali (3DVar).

Sulle Alpi nord-occidentali, alla quota di 3480 m s.l.m., presso la stazione di Plateau Rosa, la concentrazione di diossido di carbonio atmosferico viene misurata dal 1989, continuativamente dal 1993. L’altitudine a cui si trova la stazione e la distanza da fonti antropogeniche permettono di ottenere misure di background di concentrazione di gas serra e di inquinanti.

A partire dall’analisi della serie pluritrentennale di concentrazioni di CO₂, sono stati selezionati i valori di background, che costituiscono circa l’80% dell’intero dataset. Tali valori sono stati utilizzati per il calcolo del growth rate della CO₂ atmosferica, con risultati in accordo con quelli di stazioni di rilevanza globale.

L’intera serie di misure di CO₂ rivela, inoltre, l’influenza di masse d’aria con concentrazione variabile rispetto al background, come conseguenza della circolazione alla scala locale o alla mesoscala. In questo lavoro è stato sviluppato un metodo per l'identificazione degli eventi estremi e delle aree di provenienza delle particelle d’aria; tali episodi, numericamente limitati, sono stati quindi studiati applicando due diversi modelli di dispersione lagrangiani: MILORD e FLEXPART-WRF. Il primo è stato utilizzato per simulazioni long-range per tutti gli episodi di concentrazione classificata come estrema, mentre il secondo è stato applicato allo studio del trasporto alla mesoscala e alla scala regionale durante alcuni eventi.

I risultati rivelano come, durante gli episodi di picco, le principali aree di provenienza dei traccianti siano ravvisabili sulle zone fortemente industrializzate dell’Europa, mentre la circolazione atmosferica, durante i medesimi episodi, sia tipicamente ciclonica.

Climate change plays a fundamental role in the intensification of extreme weather events, such as heatwaves (HWs), droughts and floods. In the last year, intensity, frequency and duration of such events have increased and climate projections indicate further worsening in the coming years. As concerns health risks, extreme heat exposure can exacerbate cardiovascular and respiratory diseases. The negative effects of HWs can be intensified by concurrent deterioration of air quality. This connection is a widely researched topic, especially for specific pollutants, like tropospheric ozone. However, so far, few studies have investigated the association between particulate matter (PM) enhancements and HWs.

This study concerns the analysis of an intense compound event of HW and PM enhancement occurred in Bologna (Italy) during July 2023. The multidisciplinary approach implemented offers both an original method for investigating such events and a detailed description of the phenomenon by using ground-based sensors, remote-sensing measurements, satellite products and reanalysis data. The identification of compound events is conducted through extreme heat indices, i.e. the Excess Heat Factor and the Warm Spell Duration Index, and through novel index for PM based on the seasonal variability of PM. Seven compound events are found and cover about the 25% of the study period, 85 days out of 334 between January and November 2023. The analysis highlights the role of the African anticyclone in driving both the HW and the increase in PM concentrations. Besides the detailed analysis of the large-scale synoptic pattern, this is confirmed by measurements from a ground-based optical particle counter along with aerosol chemical speciation and satellite aerosol products.

These results can help policy makers in organizing more suitable responses to such compound events. Integrating Urban Heat Island analysis can offer further insights about the different negative impacts these events pose over people living in distinct urban areas.

The northeastern Italian plains are highly prone to severe convective storms, often producing large to giant hailstones due to specific orographic features. This study wants to numerically simulate and to analyze two such extreme events: a supercell on 1 August 2021 that generated hailstones up to 9 cm in diameter, and the 24 July 2023 outbreak, during which two supercells developed—one producing a European record-breaking 19 cm hailstone. Remarkably, both events produced their largest hail in the same area, near Azzano Decimo.

Numerical simulations were performed with the WRF model at 1 km resolution, coupled with the HAILCAST hail growth scheme. For the 2021 event, several configurations were tested, showing that initialization with IFS data provided the best performance. Realistic hail sizes and storm structure were reproduced only when radiosonde data from Udine Rivolto were assimilated through nudging.

The 2023 event was also simulated using the same configuration, but this setup yielded suboptimal results. Nudging degraded the simulation, as the radiosonde profile was substantially less unstable than the simulated surrounding environment, driving the model away from equilibrium and producing unrealistic convection. The most accurate results (with hailstones of sizes around 10 cm) were obtained using ERA5 initialization without data assimilation.

Comparison of the two events shows that, despite similar synoptic-scale conditions, they differ markedly in low-level moisture advection and instability, limiting dynamic and thermodynamic analogies. Analysis of the simulated updrafts—based on maximum vertical velocity, updraft area, updraft helicity, and liquid water content—shows that the 2023 event reached values nearly three times higher than those of 2021.

These results confirm, for simulations based on real events, a hypothesis previously supported only by idealized studies: hailstone size is not directly proportional to convective instability (e.g., CAPE), but depends primarily on the residence time of hail within the updraft, which can be effectively estimated through the analysis of the updraft morphology (i.e. updraft area).

The Bayesian yacht sank in Porticello, Sicily, at 0206 UTC on 19 August 2024 during a thunderstorm. Of the 22 people on board, 7 lost their lives. An in-depth analysis of available observations highlighted that the ship was likely struck by a quasi-linear convective system. Satellite images showed a Mesoscale Convective System over the Tyrrhenian Sea between 2300 UTC on 18 August 2024 and 0300 UTC on 19 August 2024, with convective cells that lasted less than 1h. The storm motion of the cell that hit Porticello was not consistent with that expected for a right mover supercell, suggesting that supercells were not present during the event. A few videos taken along the coast captured very intense northwesterly wind gusts, with no evidence of rotating winds or waterspouts. Before sinking, the yacht drifted southeastward, pushed by the northwesterly wind. Data from weather stations revealed classic downburst features, such as an increase in pressure and a drop in potential temperature corresponding to the strongest gusts. No signs of mesocyclones (e.g. sudden pressure drop) were detected.

The predictability of the event was also investigated. Operational simulations performed one day ahead with the ICON-2I model, running at 2.2 km horizontal resolution over a domain centred on Italy, pointed out that a convective wind gust hazard could have been expected over the southern Tyrrhenian Sea that night. Furthermore, the satellite analysis showed that the storm developed 3h before the accident and kept a coherent trajectory during its lifetime, suggesting that there may have been enough time to warn people. Lastly, we remark that radar data were unavailable in the area affected by the storm, which is a significant limitation for nowcasting, early warning systems, post-event analysis and research.

This study investigates the interaction between simplified urban-like obstacles and boundary layer flows under rotational effects, through laboratory experiments in a rotating water tank. Idealized building arrays were used to analyze how obstacle geometry influences flow separation, vortex formation, turbulence, and momentum transfer at the urban scale. The Rossby number (Ro) was varied to explore different regimes where rotational effects compete with inertial forces. Results, are presented in terms of flow and turbulence fields. Then, vertical profiles are analysed in order to identfy the effect of rotation both inside and out side the canions. The results contribute to the understanding of how rotation modifies boundary layer flow interactions with urban geometries, providing experimental insights relevant for urban flow modeling and environmental applications.

Thermals arise from differential heating of the Earth's surface, producing buoyant updrafts that are essential for sustaining altitude in non-motorized aviation. This study investigates the development of thermals within the atmospheric boundary layer (ABL) using a very high-resolution numerical model.

The Weather Research and Forecasting (WRF) model is employed to simulate the formation and dynamics of vertical velocities over a specific pre-alpine area in the north of Italy. The model performance is evaluated using Global Positioning System (GPS) flight data collected by paragliders during several summer days in July 2023.

The results show that the meteorological model can reasonably reproduce the spatial distribution of thermals, particularly in identifying areas where they are less likely to occur. These findings highlight the potential of high-resolution numerical models for improving thermal forecasting in recreational aviation, with implications for flight planning and safety in non-motorized flight activities.

Medicanes are tropical-like storms with growing impacts in the Mediterranean basin. This study examines the multiscale dynamics of two representative events, Qendresa (2014) and Ianos (2020), using high-resolution WRF simulations at 1 km and sensitivity tests on physical parameterizations.

Proper Orthogonal Decomposition (POD) is applied to temperature and wind fields to identify dominant modes, while Empirical Mode Decomposition (EMD) and Hilbert Spectral Analysis (HSA) capture temporal variability and multifractal features. Results show a vertically stratified energy distribution, with stronger coherence in the boundary layer and more isotropic structures in the upper troposphere.

This data-driven approach highlights key mechanisms of rapid intensification and improves understanding of cyclone predictability and mesoscale turbulence in the region.

Le grandinate costituiscono uno dei fenomeni meteorologici più impattanti per la società e le attività umane. I cambiamenti climatici stanno modificando i contesti ambientali in cui queste si sviluppano, influenzando fattori cruciali come l’umidità nei bassi strati, l’instabilità convettiva, i processi microfisici e il wind shear verticale. Ciò assume particolare rilevanza nel bacino del Mediterraneo, riconosciuto come un “hotspot climatico” e tra le aree più esposte al rischio di grandine a livello globale. Nonostante l’elevata frequenza e i rischi associati, resta ancora incompleta la conoscenza delle condizioni meteorologiche su scala sinottica che favoriscono lo sviluppo di questi eventi.

In questo studio, basato sulla climatologia satellitare giornaliera delle grandinate proposta da Laviola et al. (2022), sono stati analizzati i principali pattern spaziali estivi di grandine di grandi dimensioni (>2 cm) nell’Italia settentrionale (44.0-47.0°N, 6.0-14.0°E) ed i relativi schemi di circolazione atmosferica associati. Per l’analisi sono state impiegate la Principal Component Analysis e la Cluster Analysis, utilizzando diversi campi atmosferici derivati da rianalisi ERA5. Per garantire coerenza e omogeneità nei dati, l’indagine è stata condotta sul periodo 2014-2023.

I risultati evidenziano che la distribuzione spaziale degli eventi di grandine in Italia settentrionale si organizza in tre clusters principali, ben distinti fra loro. Il primo riguarda prevalentemente il Nord-Est, il secondo interessa soprattutto la Pianura Padana, mentre il terzo è localizzato nel settore nord-occidentale. Gli schemi circolatori associati ai primi due cluster mostrano la presenza di una saccatura sul Mediterraneo occidentale, che convoglia sull’Italia settentrionale un flusso caldo-umido da sud-ovest, accompagnato da un marcato gradiente termico a 850 hPa. Il terzo schema è invece legato a un’area ciclonica sull’Europa occidentale, caratterizzata da un forte contrasto termico sulla Francia e da un intenso flusso sud-occidentale di origine subtropicale verso l’Italia. Tra gli elementi comuni e determinanti in tutti i casi figurano il trasporto anomalo di vapore acqueo tra 2 e 5 km di quota, che favorisce un’elevata disponibilità di acqua liquida per la convezione, e la divergenza dei venti in alta troposfera.

Dal confronto tra i sottoperiodi 2014-2018 e 2019-2023, emerge un aumento significativo del contributo del terzo schema al numero totale di eventi, passato dal 19.6% al 31.8%. Ciò indica un incremento, sia in termini assoluti sia relativi, della frequenza delle grandinate nel Nord-Ovest italiano.

References

Laviola, S.; Monte, G.; Cattani, E.; Levizzani, V. Hail Climatology in the Mediterranean Basin Using the GPM Constellation (1999–2021). Remote Sens. 2022, 14, 4320. https://doi.org/10.3390/rs14174320.

Nitrogen-containing organic compounds (NOCs) represent key light-absorbing components of atmospheric PM₂.₅, yet the sources and formation mechanisms of nitrophenolic species remain unclear. Thirty-six PM₂.₅ samples collected during winter and summer from Yangon and Mandalay, Myanmar, were analyzed using UHPLC–Orbitrap MS. A total of 562–1318 organic compounds (average 1064) were identified in the ESI– mode, with NOCs accounting for 14–21% of molecular numbers and 13–35% of total concentrations.

Nitrophenolic compounds, defined by O/N ≥ 3 and AI > 0.5, were mainly distributed in zones C, F, and G of the Van Krevelen diagram and dominated the aromatic NOC fraction. Two ubiquitous nitrophenols—nitrocatechol (C₆H₅NO₄) and dimethylnitrocatechol (C₈H₉NO₄)—were detected in all samples and exhibited strong positive correlations, suggesting similar sources and transformation pathways. Their relative abundances showed distinct humidity dependence, with C₆H₅NO₄ favored under dry conditions (RH < 50%) and C₈H₉NO₄ under humid conditions (RH > 60%).

These findings highlight the significant role of nitrophenolic compounds in brown carbon formation and secondary processes in tropical aerosols, providing key mechanistic insights for subsequent modeling of their humidity-dependent formation pathways.