FEMA county hazard data:
free Excel and CSV.
Download free FEMA county hazard data for 5 counties and 27 reviewed metrics. CSV and JSON retain the source fields. The records were extracted September 6, 2026. Use the sample to inspect ratings, annualized loss, exposure and resilience before scoping a portfolio comparison.
Chart loss and population in the five-county comparison. The downloads retain source precision.
| County / FIPS | Coastal flood | Riverine flood | Hurricane | Wildfire | Earthquake | Tornado | Heat wave |
|---|---|---|---|---|---|---|---|
| Los Angeles, CA06037 | Relatively High | Very High | No Rating | Very High | Very High | Relatively High | Relatively High |
| Denver, CO08031 | Not Applicable | Relatively High | Not Applicable | Very Low | Relatively Moderate | Relatively High | Relatively Low |
| Miami-Dade, FL12086 | Very High | Very High | Very High | Relatively Moderate | Relatively Low | Relatively High | Relatively High |
| Harris, TX48201 | Relatively High | Very High | Very High | Relatively Low | Relatively Moderate | Very High | Very High |
| Chittenden, VT50007 | Not Applicable | Relatively Moderate | Relatively Low | Very Low | Relatively Low | Relatively Low | Relatively Low |
How do the sample counties differ by hazard?
Miami-Dade’s hurricane rating is Very High; Denver’s is Not Applicable; Los Angeles’s is No Rating. Unrated statuses indicate that no ordered category is supplied; they do not establish low risk.
Harris and Los Angeles both have Very High riverine flooding ratings. Individual properties can have different flood probabilities, losses and exposure. Compare counties first, then check individual locations.
How should I use the county sample?
- Select a county and review each hazard separately.
- Keep original FEMA field names and five-digit county codes (FIPS) in the workbook.
- List hazards needing a location check; keep their categories separate.
- Check address matches and map coverage before drawing property conclusions.
Which questions require additional location evidence?
Flood zones, future heat exposure, repair costs and insurance availability need separate location evidence or appropriately licensed models.
See the proposed portfolio screening scope, including the planned workbook and review queue.
How should I read the source labels?
We preserve Very Low, Relatively Low, Relatively Moderate, Relatively High and Very High exactly as supplied. Not Applicable, No Rating and Insufficient Data remain distinct.
Source: FEMA National Risk Index (December 2025 release). Extracted September 6, 2026. The filename identifies the dataset release; individual fields describe different observation periods.
The original source column names are retained. No custom composite score is added.
Read the methodology ↗What is in the FEMA NRI county sample?
The workbook has County context and Read me sheets. The CSV holds the same five source rows. Keep the source and methodology notes with your files. The FEMA source data is public-domain US government data. Our compiled files are for internal use unless redistribution is agreed in writing.
What does each field mean?
| Field | Label | How to use it |
|---|---|---|
| STCOFIPS | County identifier | Keep five digits as text, including leading zeros. |
| COUNTY / STATEABBRV | County name / state abbreviation | Use with the county code (FIPS); names alone may not match reliably. |
| CFLD_RISKR | Coastal flooding risk rating | Keep separate from riverine flooding. |
| IFLD_RISKR | Riverine flooding risk rating | A county label, not a property flood-zone code. |
| WFIR_RISKR | Wildfire risk rating | County context, not a structure inspection. |
| HWAV_RISKR | Heat wave risk rating | Not a count of hot days or a future temperature forecast. |
| HRCN_RISKR | Hurricane risk rating | Retain unrated and not-applicable labels instead of replacing them with numeric values. |
| EAL_VALT | Total expected annual loss | US dollars per year across the county; not a building-loss estimate. |
| EAL_VALB | Building expected annual loss | Building component of county expected annual loss, in US dollars per year. |
| POPULATION | Population | Persons in the source county record; keep source vintage attached. |
What should you check next?
Harris and Los Angeles share the Very High riverine category. Check each property separately; map zones and expected losses may differ. Record questions, needed sources and unresolved status in the blank review template.
These five counties are a selected comparison sample, not a random national sample. The homepage distribution uses the full workbook.
What else should I know before downloading?
Does the download contain every U.S. county?
No. It contains Los Angeles, Denver, Miami-Dade, Harris and Chittenden. The homepage county comparison uses the same five; its separate distribution chart counts all 3,232 source records.
How do I preserve FIPS in Excel?
The XLSX stores county codes (FIPS) as text. For CSV, set STCOFIPS to Text in Excel’s importer. Check Los Angeles stays 06037 and Denver stays 08031.
Use the XLSX when you want county identifiers to arrive as text. For a CSV import, set STCOFIPS to Text before loading the file. A leading zero is part of the identifier. Formatting a number after import does not restore a lost leading zero or repair an unsuccessful match.
Keep the identifier beside the county name and state abbreviation. Use it to match records between the sample, your shortlist and later source extracts. Check for unmatched rows before interpreting any comparison. Preserve the original identifier in a separate column if your workflow creates a cleaned version.
The JSON also retains the county identifier as text. Treat ratings as category labels and loss fields as numeric values with their units attached. Do not let a spreadsheet turn a missing category into a zero.
Can I average the five hazard ratings?
No. Keep the source categories separate. See how to interpret the ratings for a worked comparison.
How recent is the sample?
Extracted September 6, 2026 from the FEMA National Risk Index (December 2025 release). The release name does not date every observation to December 2025. See source dates and update limits.
Can I use the data in my report?
Keep attribution and release notes with your internal report. Discuss redistribution before including our compiled files in a client package or a public download.
Cite this dataset: Climate Atlas, FEMA county hazard sample; FEMA National Risk Index, December 2025 release; extracted September 6, 2026; https://climateatlas.us/sample/.
Keep this citation beside the figures you use, rather than only in an email accompanying the workbook. The extraction date records when the sample was prepared. It does not make every underlying observation equally recent. See the methodology for periods, checks and corrections.
How can I request broader coverage or programmatic access?
Review data access and the field registry for scoped bulk and API discussions. The free sample is a way to inspect the output before requesting coverage. It does not establish that every requested county or delivery method is available.
Tell us the geography, fields and file format your team needs. Keep any unmatched locations in the exceptions queue until coverage is confirmed. Read the ratings guide before using categories in a comparison, and see pricing for paid scope.