How the numbers are made
Everything Terra shows is a research estimate. This page explains every number: what a score is, which dataset feeds which pillar, which climate models Terra uses, which pathways and horizons it projects, and what to cite if you use any of it. Terra turns public scientific datasets and climate-model experiments into one 0–100 score per hazard, per ~10 km cell.
Terra scores ~10 km hexagonal cells from public global datasets — it is area analysis, not a property assessment. Tornado history covers the United States only; flood hazard is modeled at ~1 km and ignores local defenses; hurricane records before 1966 and wildfire records before 2001 are incomplete. Where data is missing or coarse for your area, Terra says so instead of guessing.
What a score is
A score is the share of the world's land cells that a cell is safer than, for one hazard, from one dataset, on a 0–100 scale. A score of 70 means lower hazard than 70% of the land on Earth. The only reading in words is "lower hazard than N% of land".
The combined score mixes two per-pillar roll-ups, a life mix and a property mix. The life share is 0.6 and the property share is 0.4 (api/scoring.py). The per-pillar weights are:
| Pillar | Life | Property |
|---|---|---|
| Heat | 0.18 | 0.06 |
| Flood | 0.14 | 0.24 |
| Wind | 0.10 | 0.16 |
| Earthquake | 0.10 | 0.16 |
| Wildfire | 0.10 | 0.12 |
| Climate velocity | 0.06 | 0.04 |
| Air | 0.14 | 0.02 |
| Water | 0.08 | 0.06 |
| Coast | 0.05 | 0.08 |
| Ground | 0.05 | 0.06 |
A score moves when a dataset is rebuilt: a new vintage changes the global distribution, so the share-safer-than ranking changes too. The vintage of every number in a response is named in the payload.
Each dataset
Every section below is a dataset Terra reads, with its publisher, vintage, licence and citation when the section carries one. Grouped under the pillar it feeds, where the attribution text names one, otherwise under "Base map and reference". The full text is tracked in ATTRIBUTION.md at the repository root.
The committed seed (`data/cells.parquet` and `data/sources/*.parquet`) contains only the derived per-cell indices computed from the sources listed below. No raw data from any upstream dataset is redistributed.
GEM's Global Earthquake Model Foundation global seismic hazard map was evaluated and excluded because its CC BY-NC-SA licence bars commercial reuse.
Heat
Copernicus / ERA5
Generated using Copernicus Climate Change Service information [2026]. Contains modified Copernicus Climate Change Service information; neither the European Commission nor ECMWF is responsible for any use that may be made of it.
Licence: Copernicus Climate Change Service Licence
Citation: Hersbach, H., Bell, B., Berrisford, P., et al. (2020). "The ERA5 global reanalysis." Quarterly Journal of the Royal Meteorological Society, 146(730), 1999–2049. doi:10.1002/qj.3803. Dataset DOI (CDS): doi:10.24381/cds.adbb2d47.
NASA GISTEMP v4
GISTEMP Team, 2026: GISS Surface Temperature Analysis (GISTEMP), version 4. NASA Goddard Institute for Space Studies.
Licence: Public domain (U.S. Government work)
Citation: Lenssen, N., Schmidt, G., Hansen, J., Menne, M., Persin, A., Ruedy, R., & Zyss, D. (2019). "Improvements in the GISTEMP uncertainty model." Journal of Geophysical Research: Atmospheres, 124(12), 6307–6326. doi:10.1029/2018JD029522.
Flood
WRI Aqueduct Floods
WRI Aqueduct Floods (Ward et al. 2020), World Resources Institute, CC BY 4.0.
Licence: Creative Commons Attribution 4.0 International (CC BY 4.0)
Citation: Ward, P. J., Winsemius, H. C., Kuzma, S., Bierkens, M. F. P., Bouwman, A., Moel, H. de, Loaiza, A. D., Eilander, D., Englhardt, J., Erkens, G., Gebremedhin, E. T., Iceland, C., Kooi, H., Ligtvoet, W., Luo, X., Perez-Peña, M. V., Beek, R. van, Vliet, M. T. H. van, Wada, Y. (2020). "Aqueduct Floods methodology." Technical Note, World Resources Institute. See publisher page https://www.wri.org/research/aqueduct-floods-methodology.
Wind
NOAA NCEI IBTrACS
Knapp et al., International Best Track Archive for Climate Stewardship (IBTrACS) v4, NOAA National Centers for Environmental Information.
Licence: Public domain
Citation: Knapp, K. R., Kruk, M. C., Levinson, D. H., Diamond, H. J., & Neumann, C. J. (2010). "The International Best Track Archive for Climate Stewardship (IBTrACS): Unifying tropical cyclone data." Bulletin of the American Meteorological Society, 91(3), 363–376. doi:10.1175/2009BAMS2755.1.
NOAA SPC SVRGIS
NOAA/NWS Storm Prediction Center, Severe Weather GIS (SVRGIS) tornado database.
Licence: Public domain
Citation: see publisher page https://www.spc.noaa.gov/wcm/
Earthquake
USGS ComCat
U.S. Geological Survey, ANSS Comprehensive Earthquake Catalog (ComCat).
Licence: Public domain
Citation: see publisher page https://earthquake.usgs.gov/data/comcat/
AWS Terrain Tiles (Mapzen terrarium)
Terrain tiles hosted on AWS Open Data. Elevation data derived from multiple sources listed on the AWS Open Data Terrain Tiles page, including USGS, NASA, and NGA data. Tile format: Mapzen terrarium.
Licence: Varies by source — see AWS Open Data Terrain Tiles. All underlying sources are public domain or CC BY.
Citation: see publisher page https://registry.opendata.aws/terrain-tiles/
Global Tsunami Model 2017
Global Tsunami Model 2017 (Davies et al. 2018). © Commonwealth of Australia (Geoscience Australia) and NGI 2017. Earthquake-generated tsunamis only.
Licence: Creative Commons Attribution 4.0 International (CC BY 4.0)
Citation: Davies, G., Griffin, J., Løvholt, F., Glimsdal, S., Harbitz, C., Thio, H. K., Lorito, S., Basili, R., Selva, J., Geist, E., & Baptista, M. A. (2018). "A global probabilistic tsunami hazard assessment from earthquake sources." Geological Society, London, Special Publications, 456(1), 219–244. doi:10.1144/SP456.5.
USGS 2023 National Seismic Hazard Model
USGS 2023 National Seismic Hazard Model, revised release CONUS.2023.R2 (doi:10.5066/P14VGAV4). Evidence only; not part of the earthquake score.
Licence: Public domain (U.S. Government work)
Citation: Petersen, M. D., Shumway, A. M., Powers, P. M., Field, E. H., Moschetti, M. P., Jaiswal, K. S., Milner, K. R., Rezaeian, S., Frankel, A. D., Llenos, A. L., Michael, A. J., Altekruse, J. M., Ahdi, S. K., Withers, K. B., Mueller, C. S., Zeng, Y., Chase, R. E., Salditch, L. M., Luco, N., Rukstales, K. S., Herrick, J. A., Girot, D. L., Aagaard, B. T., Bender, A. M., Blanpied, M. L., Briggs, R. W., Boyd, O. S., Clayton, B. S., DuRoss, C. B., Evans, E. L., Haeussler, P. J., Hatem, A. E., Haynie, K. L., Hearn, E. H., Johnson, K. M., Kortum, Z. A., Kwong, N. S., Makdisi, A. J., Mason, H. B., McNamara, D. E., McPhillips, D. F., Okubo, P. G., Page, M. T., Pollitz, F. F., Rubinstein, J. L., Shaw, B. E., Shen, Z.-K., Shiro, B. R., Smith, J. A., Stephenson, W. J., Thompson, E. M., Thompson Jobe, J. A., Wirth, E. A., & Witter, R. C. (2024). "The 2023 US 50-State National Seismic Hazard Model: Overview and implications." Earthquake Spectra, 40(1), 5–88. doi:10.1177/87552930231215428.
ESHM20
ESHM20, the 2020 European Seismic Hazard Model (Danciu et al. 2021), European Facilities for Earthquake Hazard and Risk, doi:10.12686/a15. Evidence only; not part of the earthquake score.
Licence: Creative Commons Attribution 4.0 International (CC BY 4.0)
Citation: Danciu, L., Nandan, S., Reyes, C., Basili, R., Weatherill, G., Beauval, C., Rovida, A., Vilanova, S., Sesetyan, K., Bard, P.-Y., Cotton, F., Wiemer, S., & Giardini, D. (2021). "The 2020 update of the European Seismic Hazard Model: Model Overview." EFEHR Technical Report 001, v1.0.0. doi:10.12686/a15.
Wildfire
NASA FIRMS
NASA FIRMS (Fire Information for Resource Management System), MODIS Collection 6.1 active fire data (yearly country archives, 2001–2024).
Licence: NASA Open Data
Citation: see publisher page https://firms.modaps.eosdis.nasa.gov/
Climate velocity
NASA NEX-GDDP-CMIP6
NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP-CMIP6), Thrasher et al. 2022; CMIP6 model output courtesy of the WCRP modelling groups.
Licence: NASA data use policy (public domain, attribution required)
Citation: Thrasher, B., Wang, W., Michaelis, A., Melton, F., Lee, T., & Nemani, R. (2022). "NASA Global Daily Downscaled Projections, CMIP6." Scientific Data, 9, 262. doi:10.1038/s41597-022-01393-4.
Air
Washington University ACAG SatPM2.5
Atmospheric Composition Analysis Group, Washington University in St. Louis, SatPM2.5 V6.GL.03 (Shen et al. 2024), annual 0.1° surface PM2.5, 2020–2024.
Licence: Creative Commons Attribution 4.0 International (CC BY 4.0)
Citation: van Donkelaar, A., Hammer, M. S., Bindle, L., Brauer, M., Brook, J. R., Garay, M. J., Hsu, N. C., Kalashnikova, O. V., Kahn, R. A., Lee, C., Levy, R. C., Lyapustin, A., Sayer, A. M., & Martin, R. V. (2021). "Monthly global estimates of fine particulate matter and their uncertainty." Environmental Science & Technology, 55(22), 15287–15300. doi:10.1021/acs.est.1c05309.
Water
WRI Aqueduct 4.0
WRI Aqueduct 4.0 (Kuzma et al. 2023), World Resources Institute.
Licence: Creative Commons Attribution 4.0 International (CC BY 4.0)
Citation: Kuzma, S., Bierkens, M. F. P., Lakshman, S., Luo, T., Saccoccia, L., Sutanudjaja, E. H., & Van Beek, R. (2023). "Aqueduct 4.0: Updated decision-relevant global water risk indicators." Technical Note, World Resources Institute. See publisher page https://www.wri.org/research/aqueduct-40-updated-decision-relevant-global-water-risk-indicators.
Ground
NASA Global Landslide Susceptibility Map
NASA Global Landslide Susceptibility Map (Stanley and Kirschbaum 2017; 2023 update).
Licence: Public domain (U.S. Government work); source acknowledgement requested
Citation: Stanley, T., & Kirschbaum, D. B. (2017). "A heuristic approach to global landslide susceptibility mapping." Natural Hazards, 87, 145–164. doi:10.1007/s11069-017-2757-y.
NOAA NCEI Global Volcano Locations Database
NOAA National Centers for Environmental Information, Global Volcano Locations Database, doi:10.7289/V5JW8BSH.
Licence: Not subject to copyright protection within the United States
Citation: NOAA National Centers for Environmental Information. Global Volcano Locations Database. doi:10.7289/V5JW8BSH. See publisher page https://www.ncei.noaa.gov/products/natural-hazards/volcanoes.
Base map and reference
OpenFreeMap / OpenStreetMap contributors
Map data © OpenStreetMap contributors. Base map tiles provided by OpenFreeMap.
Licence: Open Database Licence (ODbL) 1.0
Citation: see publisher page https://openfreemap.org/ and https://www.openstreetmap.org/copyright
H3
H3: A Hexagonal Hierarchical Geospatial Indexing System, Uber Technologies, Inc.
Licence: Apache License 2.0
Citation: see publisher page https://h3geo.org/
Natural Earth
Natural Earth (public domain) via nvkelso/natural-earth-vector
Licence: Public domain
Citation: see publisher page https://www.naturalearthdata.com/
GeoNames
GeoNames, geonames.org
Licence: Creative Commons Attribution 4.0 International (CC BY 4.0) Used for place names and search only (`data/places.parquet`). No hazard value is derived from GeoNames.
Citation: see publisher page https://www.geonames.org/
Cite our sources
Terra by SVRGN Inc., 2026. Area analysis of public hazard datasets at H3 resolution-5 hexagonal cells (~10 km). Suggested citation: "Terra (SVRGN Inc., 2026), terra.svrgninc.com," followed by the upstream sources listed above for every pillar used.
Licence: Terra's derived indices are released under CC BY 4.0; upstream data remains under its own licence as listed in the sections above.
Coverage, pillar by pillar
- Heat: Daily temperatures from ERA5, projected with NEX-GDDP-CMIP6.
- Flood: Modelled at about 1 km resolution and ignores local defences.
- Wind: Tornado history covers the United States only; hurricane records before 1966 are incomplete.
- Earthquake: Instrumental catalogue only; large pre-instrumental events are missing from the counts.
- Wildfire: A presence signal from satellite fire detections, not a measured intensity.
- Climate velocity: Warming rate from the NEX-GDDP-CMIP6 ensemble, held static where projections stop.
- Air: Five-year mean of annual surface PM2.5; no coverage north of 70°N or south of 60°S.
- Water: WRI Aqueduct 4.0 water stress; no SSP2-4.5 projection, so that pathway is held at present-day values.
- Coast: Coastal exposure comes from the tsunami runup layer only until the elevation layer lands.
- Ground: Landslide susceptibility and volcano exposure, each a class, not a rate.
The climate models
The model list is being added.
The scenarios
Future scores come from NASA NEX-GDDP-CMIP6, three climate models, three pathways and two horizons (2050 and 2085). The pathways are SSP2-4.5 (likely trajectory), SSP3-7.0 (middle-high path) and SSP5-8.5 (stress test).
Four pillars shift under a future scenario: heat, water, coast and climate velocity, each by its own delta column. Wildfire shifts by a fire-weather scenario proxy. The other five — flood, wind, earthquake, air and ground — are held at present-day values, because the underlying datasets publish no projection; the response carries a sentence that names this (api/scenario_registry.py: held at present-day values under future scenarios).
Terra does not show a model spread on the globe. A spread across three models reads as a measurement when it is not: the global reorder from one rebuild to the next moves cells more than the inter-model difference, so a ± number would mis-weight it. The three-model median is the one number shown; the spread is in the API.
The AMOC collapse switch
This switch is being built; this section describes it so the explanation is ready when it ships.
The AMOC — the Atlantic Meridional Overturning Circulation — is the ocean current that carries warm water into the North Atlantic. A collapse would cool north-western Europe, shift the tropical rainbelt and change where rain falls across the Atlantic basin. The IPCC judges a collapse before 2100 unlikely and scientists disagree on the century after. Terra attaches no date and no probability to the scenario.
The switch is built from the ClimTip experiment: three climate models (MPI-ESM1-2-HR, HadGEM3-GC31-MM and CESM1-CAM5), each held at a +2 °C warming above pre-industrial, forced with 0.3 Sv of fresh water into the North Atlantic over one hundred years. Terra takes the monthly difference between the forced run and its unforced control and translates it into a change in its daily counts — hot days, warm nights, extreme rain.
The switch touches four pillars: heat, warming, pluvial flood and water. It holds the other six at their non-switch values. Each shifted number carries the experiment and the model used to make it.
Feedback and briefs
The written brief under each score is a template filled from the numbers above. No sentence is generated by a language model. Every figure in a brief traces to a field in the /score payload for that cell — the component value, the pillar score, the combined roll-up.
Cite our sources
If you use anything from Terra in research, writing or a product, cite the sources. Every dataset and every climate model Terra uses is listed above with its licence and its citation. The licences — CC-BY 4.0 and public domain among them — require credit to the publishers, and so do we.
A suggested citation for Terra itself: Sovereign Terra, SVRGN Inc., <vintage>, terra.svrgninc.com, with the underlying datasets as listed.
Report a number that looks wrong
If a cell reads differently from what you measure on the ground, tell us on the corrections page. Every report is checked against the source data and the result is published.