Geo-Disasters: geocoding climate-related events in the international disaster database EM-DAT
Khalil Teber, Mélanie Weynants, Fabian Gans, Miguel D. Mahecha · 2025
Read Geo-Disasters: geocoding climate-related events in the international disaster database EM-DAT by Khalil Teber, Mélanie Weynants, Fabian Gans, Miguel D. Mahecha (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-03797.txt.
Introduction
Khalil Teber and co-authors add precise locations to the world's main disaster database. Their 2025 paper works on EM-DAT, the free global record of humanitarian disasters, which lists events without exact geography. That gap blocks researchers who want to join disaster records with weather and population maps.
The team geocoded more than nine thousand climate-related events reported between 1990 and 2023. Their method stays accurate even when the source gives only a region name, and each record carries quality flags so users can judge trust. An open framework lets others update the coding as new disasters are logged. With locations attached, the records can feed studies of hazard intensity, exposure, and vulnerability together. The authors aim the resource at anticipatory action, helping agencies act before disasters peak.
Catalog record
| CREATOR | Khalil Teber, Mélanie Weynants, Fabian Gans, Miguel D. Mahecha |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-03797
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Geo-Disasters: geocoding climate-related events in the international disaster database EM-DAT Khalil Teber, Mélanie Weynants, Fabian Gans, Miguel D. Mahecha Abstract Climate hazards can escalate into humanitarian disasters. Understanding their trajectories—considering hazard intensity, human exposure, and societal vulnerability—is essential for effective anticipatory action. The International Disaster Database (EM-DAT) is the only freely available global resource of humanitarian disaster records. However, it lacks exact geospatial information, limiting its use for climate hazard impact research. Here, we provide geocoding of 9,217 climate-related disasters reported by EM-DAT from 1990 to 2023, along with an open, reproducible framework for updating. Our method remains accurate even when only region names are available and includes quality flags to assess reliability. The augmented EM-DAT enables integration with other geocoded data, supporting more accurate assessment of climate disaster impacts and adaptation deficits. Background & Summary Extreme climate events can negatively impact societies in various ways [ 1 ] and can escalate into humanitarian disasters with heavy loss and damage tolls. Understanding their impact dynamics across multiple events is crucial for future disaster risk reduction efforts. Quantifying hazard intensities as well as exposure and vulnerability to them within a consistent framework is important, as these factors co-determine expected impacts [ 2 ] . However, the primary challenges in this context are data quality and availability [ 3 ] . Specifically, there is a lack of geographically explicit datasets that allow integration of climate hazards, societal exposure and vulnerability, and observed humanitarian impacts. Among these gaps, the most critical is the absence of precise disaster-impact data. Geographically explicit impact data unlock the potential to link disaster observations with external datasets, enabling empirical investigation of all facets of disaster risk. Fig. 1 illustrates a core concept: once disaster events are geocoded, they become the spatial reference that allows linking impact records to gridded climate fields, and a wide range of ancillary socio-economic and other datasets. Climate reanalysis data such as ERA5 [ 4 ] can be integrated to quantify hazard intensity. Potential population and economic exposure can be derived from gridded population [ 5 ] and Gross Domestic Product (GDP) [ 6 ] . A wealth of newly available socio-economic and political datasets—e.g., critical infrastructure [ 7 ] , the Global Data Lab’s sub-national indicators [ 8 ] , and the V-Dem democracy indices [ 9 ] —offer insight into vulnerability. The biophysical context of impacted territories can be characterized with land cover products [ 10 ] , digital elevation models, and flood plain maps [ 11 ] . Yet, all these extensions hinge on accurately identifying the regions affected by each disaster. The Emergency Events Database (EM-DAT) [ 12 ] is the most widely used global data set on humanitarian disaster impacts [ 13 ] . It is freely available and documents human losses and economic damage resulting from both human-made and natural disasters. EM-DAT has become the key resource for analyzing global records of event occurrences and their corresponding impacts [ 14 , 13 ] . It is well known that the reporting is not bias free despite strict criteria to determine which disasters are recorded. Reasons can be found in the information management of local authorities and other issues [ 15 ] . Although reporting quality has improved over the past two decades, many impact numbers remain estimates and contain missing values [ 16 , 17 , 13 ] . However, it remains a very valuable global data resource and thus needs to be improved. One avenue is to overcome the lack of geospatially explicit data sources which constrains its usability for spatio-temporal analysis. Despite its well-known gaps, EM-DAT has proven to be a valuable resource for the study of the dynamics and impacts of different climate related hazards. For example, the better-geocoded portion of EM-DAT facilitated the spatially explicit analysis of flood impacts, revealing correlations between income inequality and flood fatalities [ 18 ] ; it has also served as a complement to the flood impact database FLODIS [ 19 ] , which was in turn used to examine global vulnerability to floods in the last two decades [ 20 ] . However, these examples underscore a broader point: when precise locations are available, analysts can quantify exposure and vulnerability directly; when they are not, substantial parts of EM-DAT remain out of reach. Recognizing this challenge, one initiative—the Geocoded Disasters (GDIS) extension [ 21 ] —has attempted to geocode the full EM-DAT catalogue. We turn next to that effort and evaluate how well it addresses the lack of precise geolocations. The Geocoded Disasters (GDIS) extension provides geographically explicit geocoding of EM-DAT disaster events between 1960 and 2018, which enabled the aforementioned studies [ 18 , 19 , 20 ] . However, we see several pitfalls that inform our new framework. First, because GDIS relies on the reference administrative database Global Administrative Areas (GADM) [ 22 ] rather than EM-DAT’s native Global Administrative Unit Layers (GAUL) [ 23 ] , name-based matching can be vulnerable to homonyms. For example, Hurricane Irma (2017) is mapped by GDIS even to California and Wisconsin–although EM-DAT reports landfall only in Florida (Fig. 2 a). Second, the underlying geocoding scripts are not publicly available, precluding replication or improvement. Third, other errors in the database, such as incorrect ISO-3 identifiers for certain countries, hinder data matching with EM-DAT and raise quality concerns. Fourth, the absence of quality flags leaves users unsure whether locations were assigned automatically or manually. Finally, using administrative level 3–unsupported in EM-DAT–introduces a scale mismatch whose effect on event footprints remains unclear. Together, these shortcomings highlight the need for a new, open, reproducible, and EM-DAT–compatible geocoding framework that builds on the lessons from GDIS and can be routinely updated and validated. In this paper, we introduce Geo-Disasters, an openly accessible database and its accompanying geocoding framework that locates climate-related EM-DAT disasters. We use the GAUL administrative reference database to leverage all available EM-DAT information and minimize uncertainty. We limit the geocoding to climate disasters recorded between 1990 and 2023 that include at least one impact metric. Each event is assigned a quality flag that grades the reliability of its geocoding. By expanding the pool of precisely geocoded events, Geo-Disasters and its transparent framework enable finer-scale analyses, essential for adaptation planning and disaster-risk reduction. Methods Geo-Disasters provides geocoding for 9,217 unique climate-disaster events reported by EM-DAT for the period 1990-2023, together with quality flags that grade location certainty. EM-DAT lists names of regions where disaster impacts were reported for a given event. Since 2000, EM-DAT includes GAUL identification codes for most reported regions. Therefore, initiatives that rely on GADM database—an alternative to GAUL—restrict geocoding to name-based matching and fail to use valuable information embedded in EM-DAT. Such approaches are error-prone and risk introducing bias to downstream analysis. These considerations led us to adopt GAUL as the reference framework for Geo-Disasters. Our geocoding workflow (Fig. 3 ) proceeds in three stages: (1) GAUL-ID matching, (2) GeoNames fallback, and (3) quality flagging. In the first step, we use the GAUL identification codes attached to most post-2000 events to assign each disaster to its correct sub-national region(s). (Fig. 3 , step 1). Because the GAUL identification encodes the relevant administrative level, no further action is required.