The Macroseismic Photographic Database (DFM) is a FAIR (Findable, Accessible, Interoperable, and Reusable) dataset developed and maintained by the Istituto Nazionale di Geofisica e Vulcanologia (INGV). It provides a structured archive of photographic evidence documenting the effects of moderate to strong earthquakes in Italy since the 1980s. The data collection is primarily carried out by the INGV's QUEST (QUick Earthquake Survey Team) during post-event macroseismic field campaigns. The database was initially conceived to preserve at-risk analogue photographic material but has evolved into a comprehensive digital resource where each image is catalogued with detailed metadata. The classification of building typologies and damage grades is standardised according to the principles of the European Macroseismic Scale 1998 (EMS-98). The DFM is designed for full interoperability within the INGV data ecosystem, linking each photograph to earthquakes, localities, and macroseismic observations contained in primary databases such as the Italian Seismological Instrumental and Parametric Database (ISIDe), the Parametric Catalogue of Italian Earthquakes (CPTI15), and the Italian Macroseismic Database (DBMI15). This paper describes the database structure, the data collection protocol, the metadata schema, and the technical solutions adopted to ensure data quality and accessibility. The DFM represents a valuable resource for scientific research in engineering seismology, historical seismology, and for training operators involved in damage assessment, providing crucial ground-truth data for seismic hazard studies and civil protection purposes. The dataset is publicly accessible at 10.13127/dfm (QUEST, 2023).
Introduction
Systematic documentation of earthquake-induced damage is fundamental to seismology and earthquake engineering. Macroseismic surveys assess seismic intensity based on the observed effects on buildings, people and the environment, and provide fundamental data for understanding seismic hazards (Tertulliani et al., 2011; Galli et al., 2009). The concept of macroseismic intensity and the scales used to measure it have evolved over two centuries of observations (Tertulliani, 2019; Musson et al., 2010). While written reports and intensity assignments are important, they are, by nature, the results of interpretations and therefore inherently abstract. In contrast, photographic documentation offers an unfiltered, objective record of damage, capturing nuances that are often lost in textual descriptions. This visual data is vital for calibrating macroseismic scales, studying the seismic performance of different building types, and validating damage scenarios (Rossi et al., 2019).
In Italy, the Istituto Nazionale di Geofisica e Vulcanologia (INGV) is mandated with the seismic surveillance of the national territory. This responsibility is carried out, in part, by the QUEST (QUick Earthquake Survey Team), an operational group active for over two decades (Tertulliani and QUEST Team, 2003). QUEST conducts rapid-response macroseismic surveys following significant earthquakes as an institutional task within the framework of the National Civil Protection System. Over several decades of activity, QUEST has amassed a vast collection of photographs, initially in analogue format and later in digital format. However, this collection was fragmented and at risk of degradation, lacking a standardised structure that limited its scientific usability. Further details on the group's activities are available on the official website (https://quest.ingv.it/, last access: 24 April 2026).
To address this gap, INGV developed the Macroseismic Photographic Database (in Italian “Database Fotografico Macrosismico”, DFM; QUEST, 2023). The recent formalisation of QUEST's operational structure, which supervises its contents, and the public launch of the DFM web portal after a period of development and testing, mark the maturation of this long-term effort. The DFM transforms this heterogeneous collection into a structured, curated, and openly accessible scientific dataset. Its primary objectives are to preserve a unique visual heritage of Italy's recent seismic history, to provide a standardised framework for damage classification based on the European Macroseismic Scale 1998 (EMS-98; Grünthal, 1998), and to create an interoperable resource integrated with Italy's national seismic databases.
While other valuable image datasets of earthquake damage exist, such as the EUCENTRE's IDEA database (Senaldi et al., 2025; Casarotti et al., 2025) and engineering-focused collections like EERI's Concrete Coalition Damaged Concrete Buildings Database (EERI, 2012), their primary purpose is typically to provide labelled data for training machine learning algorithms or to support detailed structural performance analysis. In contrast, the DFM is uniquely designed as a macroseismic tool. Each image is not just a depiction of damage, but a piece of evidence contextualised within the framework of the EMS-98 scale, intended to support expert-driven intensity assessment at the locality scale and to preserve a standardised visual record of earthquake effects on the built environment. This paper presents the DFM, detailing its design, architecture, contents, and potential applications for the earth science community.
Data collection and quality control
The scientific value of the DFM lies in its rigorous and consistent process of acquiring and validating data. This process is managed by the QUEST Operational Group. As INGV is a geographically distributed institution, QUEST involves personnel from various INGV sites. To ensure efficient and scientifically robust operations, QUEST is therefore structured with an overarching Coordination Committee and a Territorial Support Group.
Following a significant earthquake, QUEST is officially activated according to its operational protocol. The standard activation thresholds are for earthquakes with magnitude ≥5.0 in tectonic areas, although interventions can also be planned for lower magnitude events of particular scientific or training interest. Once activated, QUEST teams are deployed to the epicentral area to perform a systematic macroseismic survey, aiming to achieve a comprehensive overview of the earthquake's impact in the affected area by estimating macroseismic intensity and collecting data on building seismic vulnerability and damage severity. This is achieved by targeting a representative sample of buildings within each locality.
It is important to highlight the challenging conditions under which these surveys are conducted. For safety reasons, physical entry into damaged buildings is rarely possible, as they may be unsafe or officially interdicted. In the most heavily affected areas (the so-called “red zones”), access is often restricted, and QUEST personnel must be escorted by the fire brigade, with movements severely limited. Consequently, the photographic documentation is predominantly of building exteriors.
The data acquisition workflow starts in the field, where expert surveyors photograph significant damage patterns, building typologies and environmental effects. According to the standard protocol, a general overview of the building must be taken before capturing detailed photos of the damage, although this may not always be possible due to challenging field conditions or limitations of older analogue archives. Once collected, the photographic material enters a quality control pipeline. Building typology and damage grade are attributed by trained personnel and subsequently reviewed by at least one other expert to ensure consistency and minimise subjectivity. This cross-validation process, illustrated in the workflow diagram in Fig. 1, is fundamental to guarantee the reliability of the labels associated with each image. During this internal assessment phase at the office, experts routinely use online tools (e.g., Google Street View), when applicable and available, to evaluate the pre-existing conditions of buildings and resolve any uncertainties, keeping in mind the limitations of historical archives. This makes the DFM a high-quality dataset suitable for scientific analysis.
Flow chart of the process of collecting, entering and validating data in the DFM. The “Classification” step involves Typology, Vulnerability, and Damage Grading.
Finally, although QUEST has been officially operational since 2004, the DFM archive has also collected images from older macroseismic surveys, dating back to the 1980s. Regarding future field surveys, QUEST has recently updated its operational protocol to include the use of a dedicated mobile application (QUEST-DATA; Arcoraci et al., 2020). Among its features, it can feed photos and preliminary classifications directly into a centralised repository, significantly reducing the delay between field surveys and data becoming available in DFM.
Database architecture, content, and overviewTechnical architecture
The DFM is built on a relational database managed with MySQL, with a web interface developed using PHP and JavaScript. The system is designed to handle a large volume of high-resolution images and their metadata. The web portal is divided into two main sections: a public-facing area for data consultation and a restricted-access administration backend. This backend allows accredited contributors to upload and catalogue new photographic data, and enables coordinators to validate the content and metadata before final publication. The public architecture allows for multiple query modalities. Users can perform simple or advanced searches, filtering by earthquake or by subject type. A key feature of the web portal is the interactive, map-based search function, which enables users to explore the available data geographically. Furthermore, the DFM provides curated “thematic paths”, guiding users through collections of images related to specific earthquakes or notable damage typologies and offering a thematic approach to data consultation.
Data structure and metadata
The core of the DFM is the link between a photographic image and the subject it portrays. A key aspect of the database design is that each subject can be linked to multiple photographs, allowing for a comprehensive documentation from different angles. Behind the scenes, multiple pictures of the same building from different angles are linked together by a unique “Subject ID”. This approach is crucial for a reliable macroseismic assessment, as observing a single façade can often be misleading. Each photograph is enriched with a comprehensive set of metadata, which includes image-specific information such as the author, date of capture, original resolution, and a Creative Commons (CC BY 4.0) license. Crucially, full georeferencing is ensured by associating precise geospatial information with the photographed subject.
It is important to note that the DFM database currently does not record the precise geolocation of individual buildings, even for recent events. Given the available resources, grouping photos at the locality level was the most feasible approach. Therefore, the public portal only shows the locality to which the building belongs. To comply with GDPR (General Data Protection Regulation) and protect individual privacy, sensitive elements such as vehicle license plates, house numbers, and the faces of any individuals not belonging to the QUEST team are systematically blurred before publication. A detailed subject description further clarifies what is depicted in the image.
To ensure consistency, photographed subjects are classified into predefined categories such as building, special building (e.g., schools, hospitals, factories), monumental building (e.g., churches, towers), other structure (e.g., bridges, walls), domestic object, external object, overview, environmental effects, and miscellaneous.
For subjects classified as Building, the DFM implements a detailed cataloguing system based on the EMS-98 scale (Grünthal, 1998). This involves assigning a building type based on the primary structural material, such as masonry (with sub-categories like rubble stone, adobe, unreinforced, with manufactured stone units, or massive stone), reinforced concrete, steel, and wood, which directly corresponds to the EMS-98 vulnerability classes (A to F). The vulnerability assessment in DFM strictly adheres to the EMS-98 guidelines. For reinforced concrete structures, for instance, the scale implicitly considers the level of earthquake resistance in the design. During field surveys and subsequent photo cataloguing, expert surveyors assess vulnerability classes, considering building typologies and inferring likely earthquake-resistant design levels based on the construction period, field experience, and the municipality's seismic classification history. In parallel, a damage grade is assessed and classified into one of five grades, ranging from Grade 1 (negligible to slight damage) to Grade 5 (destruction), as defined by the EMS-98 guidelines. This structured approach allows users to perform granular queries, such as retrieving all images of unreinforced, with manufactured stone units, masonry buildings (vulnerability class B) that suffered Grade 4 damage during a specific earthquake.
Dataset overview
As of the current version, the DFM contains 1685 photographs. The dataset documents the effects of 11 significant seismic events that have occurred in Italy, with a temporal coverage extending from the 1980 Irpinia earthquake to the present day. The surveys have been conducted in 150 distinct localities across the national territory, providing a geographically diverse and extensive collection of macroseismic evidence, as shown in Fig. 2. It is important to note that these figures represent only the initial fraction of a much larger photographic archive collected over decades. The population of the DFM is a recent and ongoing long-term effort, carried out primarily on a best-effort basis by researchers alongside their other institutional duties. Our prioritisation scheme for adding new data generally follows three steps: (1) new events, to make data available quickly; (2) major past events, digitising analogue archives such as the 1980 Irpinia earthquake; and (3) progressively filling gaps for other recent events. The dataset is expected to grow substantially in the coming years.
Geographic coverage of DFM. The red stars represent the epicentres of earthquakes for which the QUEST group has taken action, and the blue dots represent inhabited locations for which photographs are available on the DFM.
Interoperability and programmatic access
A key feature of the DFM is its seamless integration into the INGV data infrastructure, achieved through web services and the use of unique, persistent identifiers that make it a fully interoperable component of a larger data ecosystem, as schematised in Fig. 3. Each earthquake is linked via a unique event ID to the Italian Seismological Instrumental and Parametric Database (ISIDe Working Group, 2007) for events since 1985, or to the Parametric Catalogue of Italian Earthquakes (CPTI15; Rovida et al., 2020, 2022) for historical events. Similarly, the localities where photographs are taken are linked to the Italian Macroseismic Database (DBMI15; Locati et al., 2022), allowing users to retrieve the complete seismic history and all macroseismic observations for a specific site. This connection is further enriched by links to the Italian Archive of Historical Earthquake Data (ASMI) (Rovida et al., 2025) and the Catalogue of Strong Italian Earthquakes (CFTI). Consistency across these resources is guaranteed by the adoption of the official INGV Gazetteer, a unified geographical reference system that facilitates robust cross-database queries and geospatial analysis. This deep integration allows researchers to move fluidly from a single photograph of a damaged building to the instrumental parameters of the causative earthquake, its complete macroseismic intensity map, and the full seismic history of the locality.
Diagram showing the interoperability of the DFM with other INGV seismological databases.
Data access via web services
To further enhance data accessibility and interoperability, the DFM provides a RESTful web service that allows for programmatic querying and downloading of the information associated with the archived photographs. This service is designed for researchers, developers, and specialists who need to integrate DFM data into their own systems, GIS applications, or analysis workflows. Access is provided via an HTTP GET request to the main endpoint (https://dfm.ingv.it/services/query, last access: 24 April 2026), which can be customised with a range of parameters to filter results based on geographic, temporal, or seismological criteria.
The service returns data in GeoJSON format, an open standard based on JSON for representing geographic features. Each feature in the response corresponds to a single photograph and includes a Point geometry with the coordinates of the photographed locality, along with a set of properties detailing the earthquake parameters, the locality, and the photograph's metadata, including the URL to the original image and the damage description according to the EMS-98 scale.
Use cases and future developments
The DFM is a versatile resource with multiple applications for research and operational activities. For scientific research, the georeferenced nature of each photograph, combined with the structured damage classification, enables detailed spatial analyses to correlate observed damage with local geological conditions, ground shaking parameters, or to study damage progression during seismic sequences (Graziani et al., 2019).
A crucial application is in the realm of training and education. As stated in the official QUEST operational protocol, the DFM is the reference tool for the online training of the group's surveyors. This creates a positive feedback loop: the high-quality data produced by experts are used to train new personnel, ensuring that the standards of data collection and classification remain consistent and robust over time.
The large corpus of images labelled with building structural typology and damage grade makes the DFM an ideal machine learning resource, providing a high-quality training set for developing algorithms for automated damage classification, such as the rapid damage mapping approaches described by Pittore et al. (2018). Finally, the database supports Civil Protection agencies in emergency planning and post-earthquake response by offering a visual record of damage patterns from past events in a given area.
Future developments will focus on continuously expanding the database with data from future seismic events and digitised historical archives. Furthermore, we plan to enhance the user interface with more advanced data visualisation tools, including making the date and time information of the captured photos prominently visible, although this level of detail is often unavailable for historical older data. Additionally, future releases of the web portal will make the grouping of multiple photographs depicting the same building via the unique “Subject ID” more evident to users. We will also explore future interoperability with engineering databases such as Da.D.O. (Dolce et al., 2019) and IDEA. While highly desirable for combining seismological and structural engineering data, direct cross-referencing on a building-by-building basis is currently limited by the lack of precise building coordinates in DFM and access-related constraints of external platforms. Finally, we aim to establish a formal link between the INGV Gazetteer and the official identifiers published by ISTAT (the Italian National Institute of Statistics) to facilitate socio-economic impact analysis.
Data availability
The DFM is an open-access resource, accessible via its public web portal at https://dfm.ingv.it/ (last access: 24 April 2026). The entire dataset, including images and metadata, is freely available, provided attribution is given. The data are published under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. The dataset is identified by the persistent Digital Object Identifier (DOI): 10.13127/dfm (QUEST, 2023). Detailed metadata are available through the INGV data portal at https://data.ingv.it/dataset/774 (last access: 24 April 2026). When using data or images from the DFM, users are required to cite this paper and the dataset itself (10.13127/dfm, QUEST, 2023).
Conclusions
The DFM provides a unique, structured, and open-access collection of photographic evidence of earthquake effects in Italy. By standardising damage and building types descriptions according to the EMS-98 scale and ensuring full interoperability with national seismic databases, the DFM represents more than a simple image archive: it is a scientific tool that connects direct field observations with parametric and historical data. The transparent data collection and quality control protocol, rooted in the institutional mandate of the INGV QUEST Operational Group, ensures the reliability of the dataset. It is a living dataset, continuously updated and improved, that serves as a fundamental resource for the seismological community, civil engineers, and Civil Protection authorities. Crucially, in a context where the Italian macroseismic community faces the risk of knowledge loss due to generational turnover, the DFM also acts as a vital tool for preserving and transferring decades of field experience to new generations of surveyors. We believe the DFM will foster new research on seismic vulnerability and contribute to a better understanding of seismic risk in Italy.
Author contributions
The DFM was conceived and developed by all the authors. AT and LG provided the main scientific supervision. MS and ML were responsible for the technological development of the database and the web portal, including both its public and private sections. MB coordinated the population of the database. CC provided overall support to all aspects of the project. ML wrote the main draft of this manuscript. All authors contributed to the final version with reviews and edits.
Competing interests
The contact author has declared that none of the authors has any competing interests.
Disclaimer
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. The authors bear the ultimate responsibility for providing appropriate place names. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.
Acknowledgements
The authors wish to thank all the participants in the QUEST (Quick Earthquake Survey Team) activities who, over the years, have taken and shared the photographs that constitute the core of this database; among them, we are especially grateful to Luca Arcoraci, Sergio Del Mese, Alessandra Maramai, Ilaria Leschiutta, Antonio Rossi and Maurizio Vecchi who participated in the early phases of DFM construction. Special thanks go to the members of the national Coordination Committee (Laura Graziani, Tiziana Tuvè, Salvatore D'Amico, Alessandra Maramai, Antonio Fodarella, Filippo Bernardini, Andrea Rovida, Andrea Tertulliani) and the Territorial Support Group (Corrado Castellano, Andrea Antonucci, Emanuela Ercolani, Elisabetta Giampiccolo) who oversee the entry and cataloguing of photographs and metadata in the DFM. The activities of the QUEST team and the development of the DFM are partially funded by the Italian Department of Civil Protection (Dipartimento della Protezione Civile) within the framework of the ongoing agreement with INGV. The authors acknowledge that they used Google's Gemini model (https://gemini.google.com/, last access: 24 April 2026) and DeepL (https://www.deepl.com/, last access: 24 April 2026) to refine the English text of this manuscript, which was then edited iteratively based on the authors' input and guidance. The authors retained full responsibility for the final content, scientific accuracy, and all conclusions presented in the paper.
Financial support
The study has benefited from the financial contribution of the Italian Presidenza del Consiglio dei ministri, Dipartimento della Protezione Civile. The publication does not necessarily reflect the position and policies of the Italian Presidenza del Consiglio dei ministri, Dipartimento della Protezione Civile.
Review statement
This paper was edited by Kirsten Elger and reviewed by Giacomo Buffarini and Massimiliano Pittore.
ReferencesArcoraci, L., Rossi, A., and Tertulliani, A.: QUEST-DATA raccolta digitale sul campo del dato macrosismico: una mappa dati in tempo reale, Rapporti Tecnici INGV, 415, 1–24, 10.13127/RPT/415, 2020.Casarotti, C., Senaldi, I. E., Mandirola, M., Cantoni, A., and European Centre for Training and Research in Earthquake Engineering: IDEA: Image Database for Earthquake damage Annotation (Version 1.0.0), EUCENTRE Foundation [data set], 10.5281/zenodo.15120522, 2025.Dolce, M., Speranza, E., Giordano, F., Borzi, B., Bocchi, F., Conte, C., Di Meo, A., Faravelli, M., and Pascale, V.: Observed damage database of past Italian earthquakes: the Da.D.O. WebGIS, B. Geofis. Teor. Appl., 60, 141–164, 10.4430/bgta0254, 2019.EERI – Earthquake Engineering Research Institute: Damaged concrete buildings database, Concrete Coalition, https://www.concretecoalition.org/concrete-building-reports/ (last access: 24 April 2026), 2012. Galli, P., Camassi, R., Azzaro, R., Bernardini, F., Castenetto, S., Ercolani, E., Molin, D., Peronace, E., Rossi, A., Vecchi, M., and Tertulliani, A.: Il terremoto aquilano del 6 aprile 2009: rilievo macrosismico, effetti di superficie ed implicazioni sismotettoniche, Il Quaternario, Italian Journal of Quaternary Sciences, 22, 235–246, 2009.Graziani, L., Del Mese, S., Tertulliani, A., Arcoraci, L., Maramai, A., and Rossi, A.: Investigation on damage progression during the 2016–2017 seismic sequence in Central Italy using the European Macroseismic Scale (EMS-98), B. Earthq. Eng., 17, 5535–5558, 10.1007/s10518-019-00645-w, 2019.Grünthal, G. (Ed.): European Macroseismic Scale 1998, Cahiers du Centre Européen de Géodynamique et de Séismologie, 15, Conseil de l'Europe, Luxembourg, 10.2312/EMS-98.full.en, 1998.ISIDe Working Group: Italian Seismological Instrumental and Parametric Database (ISIDe), Istituto Nazionale di Geofisica e Vulcanologia (INGV), 10.13127/ISIDE, 2007. Locati, M., Camassi, R., Rovida, A., Ercolani, E., Bernardini, F., Castelli, V., Caracciolo, C. H., Tertulliani, A., Rossi, A., Azzaro, R., D'Amico, S., and Antonucci, A.: Database Macrosismico Italiano (DBMI15), versione 4.0, Istituto Nazionale di Geofisica e Vulcanologia (INGV) [data set], 10.13127/DBMI/DBMI15.4, 2022.Musson, R. M. W., Grünthal, G., and Stucchi, M.: The comparison of macroseismic intensity scales, J. Seismol., 14, 413–428, 10.1007/s10950-009-9172-0, 2010.Pittore, M., Graziani, L., Maramai, A., Haas, M., Parolai, S., and Tertulliani, A.: Bayesian Estimation of Macroseismic Intensity from Post-Earthquake Rapid Damage Mapping, Earthq. Spectra, 34, 1853–1877, 10.1193/112517EQS241M, 2018.QUEST – Quick Earthquake Survey Team: Database Macrosismico Fotografico (DFM), INGV – Istituto Nazionale di Geofisica e Vulcanologia [data set], 10.13127/dfm, 2023.Rossi, A., Tertulliani, A., Azzaro, R., Graziani, L., Rovida, A., Maramai, A., Pessina, V., Hailemikael, S., Buffarini, G., Bernardini, F., Camassi, R., Del Mese, S., Ercolani, E., Fodarella, A., Locati, M., Martini, G., Paciello, A., Paolini, S., Arcoraci, L., Castellano, C., Verrubbi, V., and Stucchi, M.: The 2016–2017 earthquake sequence in central Italy: macroseismic survey and damage scenario through the EMS-98 intensity assessment, B. Earthq. Eng., 17, 2407–2431, 10.1007/s10518-019-00556-w, 2019.Rovida, A., Locati, M., Camassi, R., Lolli, B., and Gasperini, P.: The Italian earthquake catalogue CPTI15, B. Earthq. Eng., 18, 2953–2984, 10.1007/s10518-020-00818-y, 2020.Rovida, A., Locati, M., Camassi, R., Lolli, B., Gasperini, P., and Antonucci, A.: Catalogo Parametrico dei Terremoti Italiani (CPTI15), versione 4.0, Istituto Nazionale di Geofisica e Vulcanologia (INGV) [data set], 10.13127/CPTI/CPTI15.4, 2022.Rovida, A., Locati, M., Antonucci, A., and Camassi, R.: The Italian Archive of Historical Earthquake Data, ASMI, Earth Syst. Sci. Data, 17, 3109–3124, 10.5194/essd-17-3109-2025, 2025.Senaldi, I., Casarotti, C., Mandirola, M., and Cantoni, A.: IDEA: Image database for earthquake damage annotation, Data in Brief, 61, 111733, 10.1016/j.dib.2025.111733, 2025.Tertulliani, A.: Storia delle scale macrosismiche, Duecento anni di osservazione degli effetti del terremoto, Quaderni di Geofisica, 150, 10.13127/qdg/150, 2019 (in Italian).Tertulliani, A. and QUEST Team: Quick earthquake surveys for Italian recent events, EGS-AGU-EUG Joint Assembly, Abstracts from the meeting held in Nice, France, 6–11 April 2003, abstract #8909, https://ui.adsabs.harvard.edu/abs/2003EAEJA.....8909T/abstract (last access: 24 April 2026), 2003.Tertulliani, A., Arcoraci, L., Berardi, M., Bernardini, F., Camassi, R., Castellano, C., Del Mese, S., Ercolani, E., Graziani, L., Leschiutta, I., Rossi, A., and Vecchi, M.: An application of EMS98 in a medium-sized city: the case of L'Aquila (Central Italy) after the April 6, 2009 Mw 6.3 earthquake, B. Earthq. Eng., 9, 67–80, 10.1007/s10518-010-9188-4, 2011.