> For the complete documentation index, see [llms.txt](https://api-v2-docs.disasteraware.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://api-v2-docs.disasteraware.com/hazards/requesting-historical-hazards.md).

# Requesting Historical Hazards

The historical archive holds every hazard the platform has tracked, in the same V2 record shape as active hazards, and can be queried by type, time, geography, name, and quantitative thresholds. To query it, use the route defined here:

{% openapi src="/files/wYQaKpPZEdPb68ZWOF6w" path="/historical-hazards-query" method="get" %}
[openapi.yaml](https://3664004773-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbwQ2AvqpbIy1hD3QFTQG%2Fuploads%2Fgit-blob-70df7a05b771cedf034307d9d49bc1c4e0436fe0%2Fopenapi.yaml?alt=media)
{% endopenapi %}

{% hint style="info" %}
See [Authorizing Requests](https://api-v2-docs.disasteraware.com/authentication/authorizing-requests.md#get-accesstoken-and-refreshtoken) page to get your accessToken.
{% endhint %}

{% hint style="warning" %}
**This endpoint requires a specific role for historical hazard access on your credentials.** It is not enabled by default. If your token works against other endpoints but this one returns `403`, contact your sales contact or <support@disasteraware.com> to have the role added.
{% endhint %}

The previous `/historical_hazards` endpoint (V1 record shape) is deprecated and no longer documented here; see the V1 API documentation if you must remain on it. Its companion `/hazards_count` and alert-geography endpoints are likewise retired: use `count_only=true` for counts, and read `alertGeometry` directly from the V2 record, which carries geometry inline.

## Coverage

The archive is a periodic build, not a live mirror. Records run from 1999 to the date of the most recent build — as of the build published on 2026-10-13, hazards started up to 2026-10-01. Hazards newer than the latest build are not in the archive yet; they are available from [`/active-hazards-query`](https://api-v2-docs.disasteraware.com/hazards/requesting-active-hazards.md) while active and from [`/hazards/{uuid}`](https://api-v2-docs.disasteraware.com/hazards/requesting-a-hazard.md) at any time. A `count` of `0` for a recent hazard usually means it post-dates the build rather than that it does not exist.

Two exclusions are deliberate. A hazard that was **still active on the platform** when the build was cut is held out of the archive until it has ended, so a long-running event can be absent here while present on the active feed. **Exercise and test records are withdrawn** entirely.

## Response Format

{% hint style="warning" %}
**The response is a redirect, not inline JSON.** A successful query returns **HTTP 302** with a `Location` header pointing at a pre-signed S3 URL. The object at that URL is **gzipped NDJSON**. Your client must follow the redirect and decompress the body. The pre-signed URL is valid for 300 seconds.
{% endhint %}

The decompressed body has one JSON object per line. **The first line is metadata**, and every following line is one hazard record:

```
{"count": 2, "limit": 2, "next_cursor": "WzE3ODUwNTU1MjEsICIwZGM1MGJlZi1lYzM5LTQ1NTQtOTNlNC1mOTc3ZDM3NDFiMzYiXQ=="}
{"uuid": "c554507b-c769-43ab-8c67-07c116d4aa8c", "type": "EARTHQUAKE", ...}
{"uuid": "0dc50bef-ec39-4554-93e4-f977d3741b36", "type": "EARTHQUAKE", ...}
```

| Metadata field   | Meaning                                                                             |
| ---------------- | ----------------------------------------------------------------------------------- |
| **count**        | Number of records in this page.                                                     |
| **limit**        | The page size that was applied.                                                     |
| **next\_cursor** | Opaque cursor for the next page, or `null` on the last page. See [Paging](#paging). |

The curl pattern that handles all of this is:

```bash
curl -sL -H "Authorization: Bearer $TOKEN" \
  'https://api-v2.disasteraware.com/historical-hazards-query?type=EARTHQUAKE&year_gte=2020&country_code=US' \
  | gunzip
```

`-L` follows the redirect; `gunzip` decompresses the body. The only request that does **not** redirect is `count_only=true`, which returns a direct `200 {"count": n}`.

## Record Formats

The `format` parameter selects what each record line contains.

| `format`           | Contents                                                                                                                                                                                                                                                                                                                                                                                                       | Use it for                                                                      |
| ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------- |
| **full** (default) | The complete V2 hazard record — the same camelCase shape as [`/hazards/{uuid}`](https://api-v2-docs.disasteraware.com/hazards/requesting-a-hazard.md) and the [hazard sample pages](https://api-v2-docs.disasteraware.com/hazards/hazard-samples/earthquake-sample.md), with `alertGeometry` inline, plus `legacy_hazard_id`, `location`, and any enhancement blocks. See [The full record](#the-full-record). | Retrieving complete hazards for display, mapping, or archival.                  |
| **list**           | A flat, snake\_case analytical projection: identity, timestamps, admin codes, nearest place, magnitude, enhancement block names, `legacy_hazard_id`, and the quantitative columns listed below.                                                                                                                                                                                                                | Filtering and bulk analysis. Rows are small, so large result sets page quickly. |
| **id\_only**       | `{"uuid": "..."}` per line.                                                                                                                                                                                                                                                                                                                                                                                    | Building an id list to fetch individually.                                      |

The `list` projection carries these quantitative columns, null where they do not apply to the hazard type: `magnitude` and `magnitude_kind`, `mmi_max`, `area_km2`, `casualties`, `damage_usd`, `ef_rating`, `path_length_km`, `hail_max_in`, `wind_gust_mph`, `saffir_cat`, `max_wind_kt`, `min_pressure_mb`, `wind_kt_at_landfall`, `landfalls`, `landfall_admin0s`, `landfall_admin1s`, `basin`, and `exposed_population`. Location comes as `admin0_iso2`, `admin1_code`, `admin2`, and — for events at sea — `nearest_admin0_iso2`, `nearest_admin1_code`, `nearest_place`, and `nearest_distance_km`.

## Paging

Results are capped per request by `limit`: the default is **100** and the hard maximum is **1000**. Values above 1000 are clamped to 1000. For larger sets, walk the cursor:

1. Make the query. Read `next_cursor` from the metadata line.
2. Repeat the **same query** with `cursor=<next_cursor>` appended.
3. Stop when `next_cursor` is `null`.

The cursor is opaque base64 and must be passed back unchanged. Results are stable-ordered — the sort key with `uuid` as a tiebreak — so pages do not overlap or skip records while you walk them. Paging works identically for all three formats.

### Page size caps on full records

`limit` is capped per hazard type when `format=full`, because those records are large: **300 for FLOOD** and **500 for WILDFIRE**. Asking for more is not an error — the response's first line reports the limit that was actually applied, and `next_cursor` continues from there:

```
?type=FLOOD&limit=400&format=full
```

```json
{"count": 300, "limit": 300, "next_cursor": "WzE3OTAyNDQyMTIsICI3ODEwYzg3ZC0xNTk1LTRjNWQtYTY5Mi1lNmJiM2Y1NGEwM2UiXQ=="}
```

Read `limit` from that line rather than assuming you received what you asked for. Very large `format=full` pages can also exceed the gateway timeout before the cap is reached, so stay well inside it for bulk pulls.

## Performance

{% hint style="warning" %}
**Always include a selective filter** — at minimum `type`, or a `year_gte`/time bound. The archive holds several hundred thousand records, and a broad scan (for example `min_ef=3&state_code=OK` with no `type`) can exceed the API gateway's request timeout, returning `{"message": "Endpoint request timed out"}` rather than results. Adding `type=TORNADO` to that same query returns in well under a second.
{% endhint %}

`count_only=true` is inexpensive even without filters and is the right way to size a result set before fetching it.

## Query Parameters

All parameters are optional and combine with AND. Comma-separated lists mean "any of". Timestamps accept **epoch seconds or ISO-8601** (for example `2024-01-01T00:00:00Z`).

### Identifiers

| Parameter              | Description                                                                                                                                                                                                                        |
| ---------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **uuid**               | Point lookup by V2 hazard `uuid`.                                                                                                                                                                                                  |
| **legacy\_hazard\_id** | Look up by the numeric V1 `hazard_ID`. A legacy construct, supported for migration. **`uuid` is the stable identifier** — V1 ids were reused over the system's life, so treat the result as a set and handle more than one record. |

### Type and class

| Parameter     | Description                                                                                                                                                                                |
| ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **type**      | Hazard type, e.g. `TORNADO,HAIL`. Comma-separated list. See [Requesting Hazard Types](https://api-v2-docs.disasteraware.com/hazards/requesting-active-hazards/requesting-hazard-types.md). |
| **sub\_type** | V2 hazard sub-type, e.g. `SMOKE`. Comma-separated list, case-insensitive. See [Sub-types](#sub-types).                                                                                     |
| **category**  | Hazard category. Comma-separated list. See [Requesting Hazard Categories](https://api-v2-docs.disasteraware.com/hazards/requesting-active-hazards/requesting-hazard-categories.md).        |
| **severity**  | Hazard severity. See [Requesting Hazard Severities](https://api-v2-docs.disasteraware.com/hazards/requesting-active-hazards/requesting-hazard-severities.md).                              |

#### Sub-types

`sub_type` is the record's V2 sub-type, and the **one queryable column taken from `incident.properties`** — it is copied from `incident.properties.subType`. It is returned in the `list` projection as well as being filterable.

Most records have no sub-type and carry `null`. The sub-type in use across the archive today is `SMOKE` on BIOMEDICAL records:

```
?type=BIOMEDICAL&sub_type=SMOKE&count_only=true
```

```json
{"count": 34793}
```

Of 36,064 BIOMEDICAL records, 34,793 are smoke events — named `Smoke — <place>` and carrying `smokeAssessment` and `smokeDetails` — leaving 1,271 with no sub-type. Matching is case-insensitive, and an unrecognised value simply matches nothing rather than returning an error, so check your spelling against the values you see in the `sub_type` column.

### Time

| Parameter                                   | Description                                                                                                      |
| ------------------------------------------- | ---------------------------------------------------------------------------------------------------------------- |
| **started\_at\_gte** / **started\_at\_lte** | Event start at or after / at or before.                                                                          |
| **ended\_at\_gte** / **ended\_at\_lte**     | Event end at or after / at or before.                                                                            |
| **created\_at\_gte** / **created\_at\_lte** | Record creation at or after / at or before.                                                                      |
| **updated\_at\_gte** / **updated\_at\_lte** | Record update at or after / at or before.                                                                        |
| **since**                                   | Alias for `updated_at_gte`. Store the highest `updated_at` you have seen and pass it back for incremental pulls. |
| **year\_gte** / **year\_lte**               | Event year bounds.                                                                                               |
| **month\_in**                               | Event month is one of these values, 1–12. Comma-separated list, e.g. `4,5`.                                      |

### Geography

| Parameter                          | Description                                                                                                                                                                                                               |
| ---------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **bbox**                           | Bounding box as `minLon,minLat,maxLon,maxLat` (WGS84). Matches on the record's geometry, not its centre point. A box whose west value exceeds its east value crosses the antimeridian. See [Bounding box](#bounding-box). |
| **contains**                       | Exact point-in-outline test as `lat,lon`. Only records whose `alertGeometry` outline contains the point. See [Is a location inside the event?](#is-a-location-inside-the-event).                                          |
| **exclude\_global**                | `true` drops records whose outline spans the world.                                                                                                                                                                       |
| **near**                           | Point-radius filter as `lat,lon,radius_km`. The radius must be **2000 km or less**.                                                                                                                                       |
| **country\_code**                  | ISO 3166-1 alpha-2 code, e.g. `US`. Matches events **in or affecting** that country, so it includes offshore events whose nearest country is the one given. Accepts a comma-separated list.                               |
| **country\_code\_3**               | ISO 3166-1 alpha-3 code, e.g. `USA`. Equivalent to `country_code` — see [Filtering by country](#filtering-by-country).                                                                                                    |
| **state\_code** / **admin1\_code** | Admin-1 code, e.g. `US-OK`. Synonyms.                                                                                                                                                                                     |
| **admin2**                         | Admin-2 name (county, district, municipality).                                                                                                                                                                            |

{% hint style="info" %}
**"In or affecting" is wider than "inside".** `country_code` matches an event that reached a country as well as one inside it, which for offshore-heavy types is a large difference: on 2026-10-09, `?type=TSUNAMI&country_code=US` returned 1,385 records, of which a sample of twelve held only two inside the United States — the other ten were at sea with the US as their nearest country. For hazards that stay inside one country the distinction does not arise.

Use `near` or `contains` instead when you need a geographic test rather than a political one, and read the `nearest_*` columns to tell which case a given record is.
{% endhint %}

Admin codes describe **where the event happened**, derived from its coordinates. Events at sea have no `admin*` values; use `near` or `bbox` for those, or read the `nearest_*` columns.

#### Filtering by country

`country_code` and `country_code_3` are the same filter in two notations: the alpha-3 code is resolved to its alpha-2 equivalent and matched identically, both mean "in or affecting", and both accept comma-separated lists. `country_code=US,JP` and `country_code_3=USA,JPN` return the same records. Both are case-insensitive.

Giving both parameters combines them — a record matching any of the codes is returned. An unrecognised code returns `400` naming the parameter and the value, rather than quietly matching nothing.

### Name

| Parameter                                       | Description                                                    |
| ----------------------------------------------- | -------------------------------------------------------------- |
| **name\_contains** / **hazard\_name\_contains** | Case-insensitive substring match on the hazard name. Synonyms. |

### Enhancement

| Parameter            | Description                                                                                                                                                                                                         |
| -------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **has\_enhancement** | Only records carrying these enhancement blocks. Comma-separated list of block names: `earthquake`, `cyclone`, `flood`, `tornado`, `hail`, `wind`, `scs`, `exposure`. See [Enhancement blocks](#enhancement-blocks). |

### Quantitative thresholds

Each of these is a `min_` / `max_` pair, e.g. `min_magnitude=6.0` or `max_casualties=10`. A record whose value is null for the column is excluded.

| Column                                                                                                                                                                                                           | Applies to                        |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------- |
| **magnitude**, **mmi**, **mmi\_max**, **shakemap\_mag**                                                                                                                                                          | Earthquake                        |
| **saffir\_cat**, **max\_wind\_kt** (as `min_max_wind_kt` / `max_max_wind_kt`), **wind\_kt**, **wind\_kt\_at\_landfall**, **min\_pressure\_mb** (as `min_min_pressure_mb` / `max_min_pressure_mb`), **landfalls** | Cyclone                           |
| **wind\_gust\_mph**, **hail\_in**, **hail\_max\_in**, **path\_length\_km**                                                                                                                                       | Severe convective storms, tornado |
| **area\_km2**, **casualties**, **damage\_usd**, **exposed\_population**, **exposed\_econ\_usd**, **nearest\_distance\_km**                                                                                       | Any                               |
| **final\_acres**                                                                                                                                                                                                 | Wildfire                          |
| **track\_length\_km**                                                                                                                                                                                            | Cyclone                           |
| **max\_above\_flood\_stage\_ft**                                                                                                                                                                                 | Flood                             |
| **damage\_estimated\_wind\_mph**                                                                                                                                                                                 | Severe convective storms, tornado |

Two shorthands are also accepted: **min\_ef** for tornadoes, which accepts `3` or `EF3`; and **min\_category** for cyclones, equivalent to `min_saffir_cat`.

{% hint style="warning" %}
**Combine a size threshold with a time bound.** `min_hail_in`, `min_wind_gust_mph` and the other `min_`/`max_` pairs are not selective on their own: with no date range the query scans the whole archive. The API gateway cuts a request off at **29 seconds** and returns **`504`** with `{"message": "Endpoint request timed out"}` rather than partial results.

The lake is partitioned by hazard type and year and ordered by creation time, so a time bound limits how much of a partition is read; a size threshold cannot. A hazard type is **not** enough on its own, and a time bound is **not a guarantee** — it narrows the scan rather than removing the limit. Measured on 2026-11-13, `?type=STORM&min_hail_in=0.75&limit=300` returned `504` with no bound; with `created_at_gte` six weeks back it returned in 20 s; with a four-month window it timed out again. **Keep the window to weeks, not months.** Any date-range parameter works — `year_gte`, `created_at_gte`, `started_at_gte`, `updated_at_gte`/`since`.

If a query returns `504`, narrow the window and retry, and page with `next_cursor` rather than asking for a large `limit`. A window that succeeds today may not as the archive grows.

This is a requirement of the query, not a suggestion: without a time bound a size-filtered query will time out as the archive grows, even where it succeeds today. Use `count_only=true` to size the result set first — it is inexpensive even without filters.
{% endhint %}

### Type-specific and provenance

| Parameter                                    | Description                                                                                                                                                                                                   |
| -------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **basin**                                    | Cyclone basin: `NA` (North Atlantic), `EP` (East Pacific), `WP` (West Pacific), `NI` (North Indian), `SI` (South Indian), `SP` (South Pacific).                                                               |
| **landfall\_country** / **landfall\_admin1** | Cyclones that made landfall in the given country / admin-1. `landfall_country` takes ISO 3166 country codes, which treat US territories as their own countries — see [Cyclone landfalls](#cyclone-landfalls). |
| **ef\_rating**                               | Exact tornado EF rating, e.g. `EF5`.                                                                                                                                                                          |
| **fire\_basis**                              | Wildfire evidence basis, e.g. `confirmed_incident`. Case-insensitive. See [Enhancement blocks](#enhancement-blocks).                                                                                          |
| **magnitude\_source**                        | Provenance of the magnitude value.                                                                                                                                                                            |
| **fire\_basis**                              | Wildfire evidence basis, e.g. `confirmed_incident`. Case-insensitive. See [Enhancement blocks](#enhancement-blocks).                                                                                          |
| **is\_scs**                                  | `true` for records classified as a severe convective storm. Null is "not classified", so `is_scs=true` is not the complement of `is_scs=false`.                                                               |
| **scs\_peril**                               | Observed severe-convective sub-perils, any of `hail`, `wind`, `tornado`.                                                                                                                                      |
| **casualties\_source** / **damage\_source**  | Restrict to records whose casualty or damage figure came from a given source. Pair with the matching `min_` filter when summing.                                                                              |
| **hail\_size\_kind** / **wind\_gust\_kind**  | Restrict to a given kind of hail size or gust value — a measured report, an estimated report, or a radar-derived value.                                                                                       |
| **max\_gauge\_category**                     | Highest NWS flood category a gauge reached.                                                                                                                                                                   |

### Output control

| Parameter       | Description                                                                                                                                                                                                                                                                                                                                                            |
| --------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **format**      | `full` (default), `list`, or `id_only`. See [Record Formats](#record-formats).                                                                                                                                                                                                                                                                                         |
| **limit**       | Page size. Default 100, maximum 1000.                                                                                                                                                                                                                                                                                                                                  |
| **cursor**      | `next_cursor` from the previous page.                                                                                                                                                                                                                                                                                                                                  |
| **count\_only** | `true` returns `200 {"count": n}` directly, with no redirect.                                                                                                                                                                                                                                                                                                          |
| **sort**        | `created_at_desc` (default), `created_at_asc`, `updated_at_desc`, `started_at_desc`, `started_at_asc`.                                                                                                                                                                                                                                                                 |
| **order\_by**   | Order by a numeric column, as `column` or `column desc`. Columns: `area_km2`, `casualties`, `damage_usd`, `exposed_econ_usd`, `exposed_population`, `hail_max_in`, `landfalls`, `magnitude`, `max_wind_kt`, `min_pressure_mb`, `mmi_max`, `nearest_distance_km`, `path_length_km`, `saffir_cat`, `shakemap_mag`, `started_at`, `wind_gust_mph`, `wind_kt_at_landfall`. |

When ordering by a column, records with a null value for that column sort first under `desc`. Pair `order_by` with the matching `min_` filter (for example `min_casualties=1`) to exclude them.

### Cyclone landfalls

`landfall_country` takes **ISO 3166 country codes**, and ISO gives Puerto Rico, Guam, the Northern Mariana Islands, the US Virgin Islands and American Samoa their own codes. They are therefore not matched by `US`.

Hurricane Maria in 2017 is the clearest case. Its landfall columns read:

```json
{ "landfalls": 2, "landfall_admin0s": ["DM", "PR"], "landfall_admin1s": ["DM-03", "US-PR"], "wind_kt_at_landfall": 145.0, "max_wind_kt": 150.0 }
```

So `?type=CYCLONE&name_contains=MARIA&year_gte=2017&year_lte=2017&landfall_country=US` returns nothing, while `landfall_country=PR` returns the storm, and so does `landfall_admin1=US-PR` — the admin-1 code is namespaced under `US` even though the country code is not.

To mean "the United States including its territories", list the codes:

```
?type=CYCLONE&landfall_country=US,PR,GU,MP,VI,AS&count_only=true
```

On 2026-10-09 that returned 108 records against 103 for `landfall_country=US` alone.

{% hint style="info" %}
This is not the same rule as the general [`country_code`](#geography) location filter, which means "in or affecting" and matches offshore records by nearest country. The landfall columns describe where the storm's centre actually crossed a coast.
{% endhint %}

**Reading the landfall columns.** Two things follow from how the track is sampled, and both can surprise:

* **`wind_kt_at_landfall` is the wind at the nearest track position, not the storm's peak.** Track positions are mostly three-hourly, so the landfall figure is often below `max_wind_kt` — 145 kt against a 150 kt peak for Maria above.
* **Landfalls on very small islands can be missed**, for the same reason: the storm can pass over between two sampled positions.

Only **tropical** landfalls are counted, and the landfall columns come from the cyclone pipeline alone. A post-tropical crossing is still listed on the full record, tagged with its `landfall_type`, but is not included in `landfalls` or the landfall admin arrays.

## Errors

**400 — validation.** Any bad parameter, value, or cursor returns a body that says what was wrong and lists every parameter the endpoint accepts. The `allowed_parameters` list is generated from the same table the API validates against, so it is always current:

```json
{
  "error": "Query validation failed",
  "details": ["near must be lat,lon,radius_km (radius <= 2000)"],
  "allowed_parameters": ["admin1_code", "admin2", "basin", "bbox", "category", "count_only", "..."]
}
```

A malformed cursor returns `{"error": "bad cursor"}`. Unknown values for `type`, `has_enhancement`, or `month_in` are not errors; they simply match nothing.

**403** — the token is invalid or expired, or the credentials lack the role required for historical hazard access.

## The full record

With `format=full` (the default), and from [`/hazards/{uuid}`](https://api-v2-docs.disasteraware.com/hazards/requesting-a-hazard.md), each record is the complete V2 hazard shape documented in [Requesting Active Hazards](https://api-v2-docs.disasteraware.com/hazards/requesting-active-hazards.md#response-fields), with three additions that only historical records carry.

**`legacy_hazard_id`** — the hazard's V1 numeric `hazard_ID`. Use it to map existing V1 integrations onto V2 uuids.

**`alertGeometry`** — always GeoJSON, never WKT, and `null` when the hazard has no geometry. Two GeoJSON shapes appear in the archive today: records produced by the platform carry a **FeatureCollection** with one Feature per area, while older converted records carry a **bare `Polygon` or `MultiPolygon`**. Both shapes are present within a single hazard type and month — of 25 STORM records sampled from March 2026, 5 were FeatureCollections and 20 bare geometries, while a September 2026 sample of 25 was entirely FeatureCollections.

Handle both: unwrap a FeatureCollection to its Features' geometries, and treat a bare geometry as a single area. A monthly delivery wraps the remaining bare geometries, after which the FeatureCollection shape is the only one returned — so code that handles both today keeps working afterwards.

**`location`** — a reverse-geocoded administrative block. Events on land carry the admin hierarchy; events at sea carry the nearest place instead:

```json
{ "admin0": "United States of America", "admin0_iso2": "US", "admin0_iso3": "USA", "admin1": "Missouri", "admin1_code": "US-MO", "admin2": "Ste. Genevieve", "admin2_code": "29186", "continent": "North America", "subregion": "Northern America", "location_type": "land", "source": "arcgis_reverse_geocoder" }
```

```json
{ "location_type": "ocean", "nearest_country_code": "TO", "nearest_admin1": "'Eua", "nearest_place": "`Ohonua", "source": "arcgis_reverse_geocoder" }
```

### Incident properties are source detail

`incident.properties` is **exactly what the reporting source delivered**, carried through unchanged. It is not normalised across sources, and the same hazard type can arrive with different key sets depending on which pipeline produced the record. Nothing is renamed to a common vocabulary and nothing is moved into a wrapper.

That means the key set varies by record, not by date. Two broad families appear in the archive:

* **Native V2 CAP vocabulary** — `sentAt`, `expiredAt`, `areaDescription`, `senderName`, `sourceMessageIdentifiers`, `summaryStatus`, usually with `creator` `RICHTER` and a FeatureCollection `alertGeometry`.
* **Converted vocabulary** — `createDate`, `eventDate`, `hazardUUID`, `sncUrl`, often with a bare `Polygon` or `MultiPolygon` `alertGeometry` and `creator` `DisasterAWARE (Automated)` or `DataFeeds Enhancement Pipeline`. Older analyst-entered records carry no `alertGeometry` at all.

Both families appear within the same hazard type and the same month: of 289 STORM records sampled across 2026, 140 carried the native set and 149 the converted one. Nor does type decide it — among records starting after 2026-08-01, STORM, SEVEREWEATHER, HIGHWIND, MARINE, FLOOD and TORNADO carried the native set, EARTHQUAKE still carried `createDate` and `eventDate`, and WILDFIRE and CYCLONE carried neither, having their own type-specific property sets.

{% hint style="info" %}
**Do not key your integration on `incident.properties`.** Read `incident.sourceId` and `creator` to identify a record's source, as described under [Source Identifiers](https://api-v2-docs.disasteraware.com/hazards/requesting-active-hazards.md#source-identifiers), then interpret the properties for that source — the [hazard sample pages](https://api-v2-docs.disasteraware.com/hazards/hazard-samples/earthquake-sample.md) document them per type. Code that requires a particular key should treat its absence as "this record came from a different source", not as an error.

Everything you can **filter, sort or group on** comes from the record's top-level fields or its [enhancement blocks](#enhancement-blocks), never from `incident.properties`. Top-level fields are consistent on every record: `startedAt`, `endedAt`, `severity`, `type` and `location`.
{% endhint %}

### Enhancement blocks

Historical records may carry post-event analysis blocks that active records never have. The `list` format names them in `enhancement_blocks`; `has_enhancement` filters on them; and the `full` format includes them as top-level objects.

| `has_enhancement` value | Blocks on the full record             | Source and contents                                                                                                                                                                                                                                                                                                                                                                            |
| ----------------------- | ------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **earthquake**          | `earthquakeAssessment`                | USGS ShakeMap summary: `usgs_event_id`, `magnitude`, `depth_km`, `max_mmi`, `max_pga_g`, `max_pga_pctg`, `pga_contour_count`, `pga_contours`, `shakemap_version`.                                                                                                                                                                                                                              |
| **cyclone**             | `cycloneAssessment`, `cycloneDetails` | NOAA IBTrACS best-track archive. The assessment summarises the storm (`storm_name`, `season`, `basin`, `max_category`, `max_wind_kts`, `min_pressure_mb`, `duration_hours`, `hours_at_hurricane_force`, `hours_at_major_hurricane`, `landfalls`, `track_length_km`, `track_points_count`); the details block carries the full observed `track`.                                                |
| **flood**               | `floodAssessment`, `floodDetails`     | NWS flood post-event analysis. The assessment carries `flood_class`, `triggering_event`, `max_gauge_category`, `max_above_flood_stage_ft`, `depth_method`, and counts of `gauges`, `road_closures`, `storm_reports`, and `survey_points`; the details block carries the underlying `alert` and `alert_polygon`, `gauges`, `depth`, `impacts`, `road_closures`, `storm_reports`, and `sources`. |
| **tornado**             | `tornadoAssessment`, `tornadoDetails` | NWS post-event analysis: Damage Assessment Toolkit paths and points, Storm Prediction Center reports, and Local Storm Reports. The assessment carries `max_ef_rating`, `total_fatalities`, `total_injuries`, and per-source counts; the details block carries the underlying `dat_paths`, `dat_points`, `local_storm_reports`, and `spc_reports`.                                              |
| **hail**                | `hailAssessment`, `hailDetails`       | Hail reports and MESH radar swaths: `best_estimate_in`, `max_measured_in`, `max_reported_in`, `mesh_max_swath_in`, `size_description`, with `hail_reports`, `mesh_swaths`, and `svr_warnings` in the details.                                                                                                                                                                                  |
| **wind**                | `windAssessment`, `windDetails`       | Wind damage and gust reports, with `wind_reports` in the details.                                                                                                                                                                                                                                                                                                                              |
| **scs**                 | `scsClassification`                   | Severe convective storm classification across the tornado, hail, and wind sub-perils: `is_scs`, `sub_perils_present`, `max_ef_rating`, `max_hail_size_in`, and the source of each.                                                                                                                                                                                                             |
| **exposure**            | `exposureAssessment`                  | The standard population and infrastructure exposure block, as on active hazards.                                                                                                                                                                                                                                                                                                               |

A tornado record from a surveyed outbreak typically carries `tornado`, `hail`, `wind`, and `scs` together. Field-level detail for the earthquake, cyclone, and tornado blocks is on the enhanced historical sample pages ([earthquake](https://api-v2-docs.disasteraware.com/hazards/hazard-samples/enhanced-historical-earthquake-sample.md), [cyclone](https://api-v2-docs.disasteraware.com/hazards/hazard-samples/enhanced-historical-cyclone-sample.md), [tornado](https://api-v2-docs.disasteraware.com/hazards/hazard-samples/enhanced-historical-tornado-sample.md)); note that those pages show the blocks inside the legacy V1 wrapper, whereas this endpoint returns them on the V2 record.

**WILDFIRE records carry `fireAssessment`.** It is present on every wildfire record in the archive, and its `fire_basis` says what kind of evidence the record rests on — which matters because the archive mixes official incident records with satellite detections:

| `fire_basis`                | Meaning                                         |
| --------------------------- | ----------------------------------------------- |
| **confirmed\_incident**     | An official incident record from a fire agency. |
| **satellite\_detection**    | A satellite thermal detection.                  |
| **satellite\_burned\_area** | Derived from a satellite burned-area product.   |
| **unconfirmed\_report**     | Reported but not confirmed by an authority.     |
| **prescribed\_burn**        | A planned burn rather than a wildfire.          |
| **not\_a\_fire**            | Classified as not a fire.                       |

`confirmed_incident` and `satellite_detection` account for the great majority of records; the other four are each a small fraction. Records that did not come through the wildfire pipeline carry a minimal `fireAssessment` with `fire_basis`, `fire_basis_source`, `enhancedAt`, `enhancer` and `source`, and no `match_status`.

`fire_basis` is also a query parameter, matched case-insensitively, so you can count the officially confirmed fires directly:

```
?type=WILDFIRE&fire_basis=confirmed_incident&count_only=true
```

`fire_basis` is the value to filter and read; a human-readable label for each value is available through the MCP tools rather than as a column on this endpoint.

Historical full records also have `source` set to `null` and `features` set to `null` in the general case; `alertGeometry`, `exposureAreas`, and `exposureAssessment` are populated as on active hazards.

## What each column measures

The `list` columns and the `min_`/`max_` filters are copied or derived from the record, never computed independently of it. Most are self-explanatory; the ones below are not, and reading them the obvious way gives the wrong answer.

Where a measure can come from more than one place, a companion column says which: `magnitude_kind`, `hail_size_kind`, `wind_gust_kind`, `casualties_source`, `damage_source`, `damage_estimated_wind_source`, `fire_basis_source`, `final_acres_source` and `landfall_source`. Read the companion before comparing or summing values across records.

| Column                                                    | What it actually is                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
| --------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **magnitude**                                             | **Type-dependent.** For EARTHQUAKE it is the moment magnitude. For TORNADO it is the EF number. For CYCLONE it is the Saffir-Simpson category. Null for every other type. `magnitude_kind` names which. Never compare `magnitude` across types.                                                                                                                                                                                                                                            |
| **area\_km2**                                             | The area of the **alert polygon**, not the burned, flooded or damaged area. A large `area_km2` means a large warning footprint.                                                                                                                                                                                                                                                                                                                                                            |
| **saffir\_cat**, **max\_wind\_kt**, **min\_pressure\_mb** | The storm's **peak** over its whole life, not its state at landfall. For landfall intensity use `wind_kt_at_landfall`.                                                                                                                                                                                                                                                                                                                                                                     |
| **wind\_kt\_at\_landfall**                                | The wind at the track position nearest the coast crossing. Track positions are roughly three-hourly, so this often sits below the peak. See [Cyclone landfalls](#cyclone-landfalls).                                                                                                                                                                                                                                                                                                       |
| **hail\_max\_in**                                         | The best available size estimate, which may be a **radar estimate rather than an observed report**. `hail_size_kind` distinguishes a measured report, an estimated report, and a radar-derived value; a substantial share are radar.                                                                                                                                                                                                                                                       |
| **wind\_gust\_mph**                                       | A gust value, measured or estimated — `wind_gust_kind` says which.                                                                                                                                                                                                                                                                                                                                                                                                                         |
| **damage\_estimated\_wind\_mph**                          | A wind speed **inferred from damage** during a survey, not a measured gust. Do not mix it with `wind_gust_mph` in the same series.                                                                                                                                                                                                                                                                                                                                                         |
| **path\_length\_km**                                      | Tornado path length, and **tornadoes only**. A cyclone's track length is `track_length_km`. See the note below if you previously read cyclone track lengths from this column.                                                                                                                                                                                                                                                                                                              |
| **casualties**                                            | Deaths, from whichever source reported them; `casualties_source` names it — `tornado_dat_owned` on tornado records, for example. Each record carries the deaths **it owns**, so summing across records is safe and does not double-count. Records with no tornado enhancement have **no casualty value at all** — null, not `0` — and are excluded from sums rather than counted as zero. Sources differ in what they measure, so filter or group by `casualties_source` before comparing. |
| **damage\_usd**                                           | Economic loss or reported property and crop damage depending on `damage_source` — **the sources measure different things.** Filter or group by `damage_source` before summing.                                                                                                                                                                                                                                                                                                             |
| **exposed\_population**, **exposed\_econ\_usd**           | Modelled exposure inside the hazard's footprint, from `exposureAssessment`. Not an observed impact count.                                                                                                                                                                                                                                                                                                                                                                                  |
| **is\_scs**                                               | True when a severe convective sub-peril was observed. **Null means not classified, not false** — many records carry no classification at all.                                                                                                                                                                                                                                                                                                                                              |
| **ended\_at** / **no\_scheduled\_end**                    | `ended_at` is null for a hazard with no scheduled end, and `no_scheduled_end` is `true` on those records. Treat the pair together rather than reading a null end date as missing data.                                                                                                                                                                                                                                                                                                     |
| **legacy\_hazard\_id**                                    | The numeric V1 id. A small number of records have a non-numeric legacy id, which is null in this column and present verbatim on the full record.                                                                                                                                                                                                                                                                                                                                           |

{% hint style="warning" %}
**`path_length_km` changed with the 2026-10 rebuild.** It was previously populated for cyclones with the storm's track length and was empty on tornado records. It is now tornado path length only, and cyclone track length has moved to its own `track_length_km` column.

If you filtered or ordered cyclones on `path_length_km` — or on `min_path_length_km` — switch to `track_length_km`. Against the earlier build, `?type=CYCLONE&min_saffir_cat=4` returned track lengths of several thousand kilometres in `path_length_km`, so a cyclone query against that column will now come back empty rather than wrong, and a tornado query that previously returned nothing will start returning values.
{% endhint %}

{% hint style="info" %}
**Null means "not reported", never zero.** An unreported damage figure, a blank wind radius or a tropical depression's absent Saffir-Simpson category are all null rather than `0`, and a tropical depression carries `max_category_name` "Tropical Depression" instead. Treating null as zero will understate averages and invent zero-damage events. The `min_`/`max_` filters exclude records whose value is null for the column being filtered.
{% endhint %}

{% hint style="info" %}
The enhancement blocks carry far more fields than there are columns — the columns are the filterable and sortable subset. Everything else reaches you only inside the full record, via `format=full` or [`/hazards/{uuid}`](https://api-v2-docs.disasteraware.com/hazards/requesting-a-hazard.md), and cannot be filtered or ordered on. See [Enhancement blocks](#enhancement-blocks).
{% endhint %}

## Worked Examples

Every example below was run against the production API — the parameter-syntax examples on 2026-09-18, the rest on 2026-09-11. Counts will have grown since. `$TOKEN` is the access token from `/authorize`; the pattern is always:

```bash
curl -sL -H "Authorization: Bearer $TOKEN" 'https://api-v2.disasteraware.com/historical-hazards-query?...' | gunzip
```

### Combining parameters

Two rules cover the whole query syntax:

* **Different parameters are combined with AND.** `?type=TORNADO&state_code=OK` returns tornadoes that are also in Oklahoma.
* **Multiple values within one parameter are comma-separated and combined with OR.** `?type=TORNADO,HAIL` returns tornadoes and hail events.

Anything containing a space must be URL-encoded — `order_by=casualties%20desc`, `+` also works — and the URL should be single-quoted in `curl` so the shell does not interpret the `&` characters. A misspelled parameter returns `400` with the full list of accepted parameters (see [Errors](#errors)).

Two hazard types at once, narrowed by severity:

```
?type=TORNADO,HAIL&severity=WARNING&count_only=true
```

```json
{"count": 101337}
```

Add a time bound and a seasonal window to cut that down — April and May of 2024 onward:

```
?type=TORNADO,HAIL&severity=WARNING&year_gte=2024&month_in=4,5&count_only=true
```

```json
{"count": 12262}
```

Every additional parameter narrows the result, so build a query up one filter at a time with `count_only=true` before fetching the records themselves (see [Sizing a query before fetching it](#sizing-a-query-before-fetching-it)).

### Bounding box

`bbox` takes four comma-separated numbers in the order `minLon,minLat,maxLon,maxLat`. Note that longitude comes first, which is the GeoJSON convention and the reverse of the `near` parameter's `lat,lon,radius_km`.

```
?type=WILDFIRE&bbox=-125,32,-114,42&year_gte=2023&limit=2&format=list
```

```
{"count": 2, "limit": 2, "next_cursor": "WzE3ODU4MzU2ODUsICJjOGYxODkyOC02YjM0LTRhY2QtODI4Ny1mNjg4N2ZiZTE2OTIiXQ=="}
{"uuid": "439f4a96-0b5e-4ffa-ba63-b1be646c8c0e", "type": "WILDFIRE", "severity": "ADVISORY", "name_en": "Wildfire - ANDERSON, California, United States", "latitude": 39.78707, "longitude": -123.24822, "admin1_code": "US-CA", "admin2": "Mendocino", "area_km2": 1191.9804052226245, ...}
{"uuid": "c8f18928-6b34-4acd-8287-f6887fbe1692", "type": "WILDFIRE", "severity": "WARNING", "name_en": "Wildfire - Mexicali, Baja California, Mexico", "latitude": 32.54095, "longitude": -115.1077, "admin1_code": "MX-BCN", "admin2": null, ...}
```

415 records matched in total. The box does not respect national borders — the second record is in Mexico, inside the box. Add `country_code=US` to restrict it.

**What the box is tested against.** A record matches when the bounding box of its geometry overlaps the box you gave, not when its centre point falls inside it. A hazard whose point sits well outside your box still matches if its footprint reaches in. Typhoon Halong, for example, has a centre point at longitude 149.9 and a footprint reaching to 166.9, so it matches `bbox=160,33,166,35` even though its point is nowhere near it. Records that carry no geometry — a little under a third of the archive — have no footprint to test, so for those the filter does behave as a point filter.

`bbox` follows RFC 7946: a box whose west value is greater than its east value crosses the antimeridian and is tested as such. `bbox=170,-30,-170,-10` is the twenty-degree strip over the dateline; writing the same numbers the other way round, `bbox=-170,-30,170,-10`, is the 340-degree box that is everything except that strip.

{% hint style="info" %}
`bbox` is a coarse filter. A large outline has a large bounding box, so a box over Denver overlaps the bounding boxes of Pacific-wide tsunami zones: `?type=TSUNAMI&bbox=-105.2,39.5,-104.7,40.0` matches 590 records, none of which reach Colorado. Use `contains` when you need containment, and `exclude_global=true` to drop world-spanning outlines.
{% endhint %}

### Is a location inside the event?

`contains=lat,lon` is an exact point-in-outline test against each record's `alertGeometry`. Holes in the outline are respected, and outlines that span the antimeridian are handled whether or not they are split at 180 degrees.

```
?type=TSUNAMI&contains=21.31,-157.86&count_only=true
```

```json
{"count": 266}
```

266 tsunami events have an alert outline covering Honolulu. The same test for Denver returns none:

```
?type=TSUNAMI&contains=39.74,-104.99&count_only=true
```

```json
{"count": 0}
```

Compare that with the 590 records a small `bbox` over Denver matches, and the difference between overlapping a bounding box and falling inside an outline is the whole point of the parameter.

Two things to know. Records that carry no geometry never match `contains`; reach those with `near` instead. And scope the query with `type`: a type-scoped `contains` returns in a few seconds, while an unscoped one takes around twenty.

```
?type=EARTHQUAKE&contains=35.68,139.69&count_only=true
```

```json
{"count": 194}
```

### Dropping world-spanning outlines

Some events carry an outline that spans the world, which is rarely what you want in a regional query. `exclude_global=true` removes them:

```
?type=TSUNAMI&exclude_global=true&count_only=true
```

```json
{"count": 2137}
```

Against 2,368 tsunami records in total, so 231 carry a world-spanning outline.

### Threshold plus time bound

```
?type=EARTHQUAKE&min_magnitude=7.0&year_gte=2015&count_only=true
```

```json
{"count": 179}
```

Every quantitative column has a `min_` / `max_` pair that can be combined the same way; `?type=EARTHQUAKE&min_magnitude=7.0&max_magnitude=7.9` brackets a band.

### Sizing a query before fetching it

```
?type=EARTHQUAKE&year_gte=2020&country_code=US&count_only=true
```

```json
{"count": 490}
```

This returns `200` directly, with no redirect.

### Basic filter

```
?type=EARTHQUAKE&year_gte=2020&country_code=US&limit=2&format=list
```

```
{"count": 2, "limit": 2, "next_cursor": "WzE3ODUwNTU1MjEsICIwZGM1MGJlZi1lYzM5LTQ1NTQtOTNlNC1mOTc3ZDM3NDFiMzYiXQ=="}
{"uuid": "c554507b-c769-43ab-8c67-07c116d4aa8c", "type": "EARTHQUAKE", "severity": "ADVISORY", "category": "EVENT", "name_en": "Earthquake - 5.0 - 51 km E of Denali National Park, Alaska", "latitude": 63.611, "longitude": -150.673, "created_at": 1785272838, "started_at": 1785271764, "ended_at": 1785358164, "no_scheduled_end": false, "admin0_iso2": "US", "admin1_code": "US-AK", "admin2": "Denali", "nearest_admin0_iso2": null, "nearest_admin1_code": null, "nearest_place": null, "nearest_distance_km": null, "magnitude": 5.0, "magnitude_kind": "mw", "enhancement_blocks": ["earthquake", "exposure"], "legacy_hazard_id": 471892, "hail_max_in": null, "wind_gust_mph": null, "saffir_cat": null, "max_wind_kt": null, "min_pressure_mb": null, "basin": null, "landfalls": null, "landfall_admin0s": null, "landfall_admin1s": null, "wind_kt_at_landfall": null, "mmi_max": 5.23, "ef_rating": null, "casualties": null, "damage_usd": null, "area_km2": 98101.71167124063, "path_length_km": null, "exposed_population": 0.0}
{"uuid": "0dc50bef-ec39-4554-93e4-f977d3741b36", "type": "EARTHQUAKE", "severity": "WATCH", ...}
```

### Spatial: tornadoes within 100 km of Denver

```
?near=39.74,-104.99,100&type=TORNADO&count_only=true
```

```json
{"count": 617}
```

### Threshold: Category 4+ North Atlantic hurricanes

```
?type=CYCLONE&min_saffir_cat=4&basin=NA&limit=2&format=list
```

```
{"count": 2, "limit": 2, "next_cursor": "..."}
{"uuid": "1e4d762a-c03e-4ea5-bbed-8001827b1045", "type": "CYCLONE", "name_en": "Major Hurricane - MELISSA", "nearest_admin0_iso2": "JM", ...}
{"uuid": "67da3838-25f6-4713-bd80-6e3e4fdc0e1b", "type": "CYCLONE", "name_en": "Major Hurricane - HUMBERTO", "nearest_admin0_iso2": "AI", ...}
```

Sixty-six records matched in total.

### Ordering: strongest Oklahoma tornadoes by casualties

```
?type=TORNADO&min_ef=3&state_code=OK&order_by=casualties%20desc&limit=3&format=list
```

```
{"count": 3, "limit": 3, "next_cursor": "..."}
{"name_en": "Tornado - Norman, OK WFO Region, US", "ef_rating": "EF3", "casualties": 16.0, ...}
{"name_en": "Tornado — Comanche [OK]", "ef_rating": "EF3", "casualties": 12.0, ...}
{"name_en": "Tornado — Bryan [OK]", "ef_rating": "EF3", "casualties": 12.0, ...}
```

Note the `type=TORNADO` filter. Without it, this query scans the whole archive and times out at the gateway. The space in `order_by` must be URL-encoded as `%20` (or `+`).

### Walking the cursor

Page 1:

```
?type=EARTHQUAKE&year_gte=2020&country_code=US&limit=2&format=list
```

```
{"count": 2, "limit": 2, "next_cursor": "WzE3ODUwNTU1MjEsICIwZGM1MGJlZi1lYzM5LTQ1NTQtOTNlNC1mOTc3ZDM3NDFiMzYiXQ=="}
{"uuid": "c554507b-c769-43ab-8c67-07c116d4aa8c", "created_at": 1785272838, ...}
{"uuid": "0dc50bef-ec39-4554-93e4-f977d3741b36", "created_at": 1785055521, ...}
```

Page 2 — the same query with `cursor` appended:

```
?type=EARTHQUAKE&year_gte=2020&country_code=US&limit=2&format=list&cursor=WzE3ODUwNTU1MjEsICIwZGM1MGJlZi1lYzM5LTQ1NTQtOTNlNC1mOTc3ZDM3NDFiMzYiXQ==
```

```
{"count": 2, "limit": 2, "next_cursor": "WzE3ODQ4NTU2NjUsICJjNTNjOTQxYS0zOWFmLTQ0NDMtYjU4NC1hYzc2ZWQ4NTA2ZTAiXQ=="}
{"uuid": "dd242b92-0731-4c4b-80aa-b612bb1c106a", "created_at": 1785014061, ...}
{"uuid": "c53c941a-39af-4443-b584-ac76ed8506e0", "created_at": 1784855665, ...}
```

Continue until `next_cursor` is `null`. Note that `created_at` descends across the page boundary with no gap or overlap.

### Migration: looking up a V1 hazard by its numeric id

```
?legacy_hazard_id=471892&format=full
```

Returns the V2 record for the hazard that was `hazard_ID` 471892 in the V1 API — the Denali earthquake above. The record's own `legacy_hazard_id` field carries the same value, so a one-time pass over your stored V1 ids will resolve them to uuids.

{% hint style="warning" %}
**A V1 id can match more than one record.** V1 minted ids from ranges and reused them across runs, so the same numeric id can appear on two unrelated hazards. `?legacy_hazard_id=950210` returns two: a severe thunderstorm in Giles County, Tennessee and a tornado in St. Tammany Parish, Louisiana. `906288` likewise returns a storm in Covington County, Alabama and a tornado in Franklin County, Massachusetts.

Write your migration to expect a **set** of records per V1 id and to disambiguate on `type`, `started_at` or location — not to assume a single hit. Once resolved, key everything on `uuid`, which is unique and stable.
{% endhint %}

A `count` of `0` means that V1 id is not in the historical archive — most often because the hazard was created after the archive's most recent build (see [Coverage](#coverage)). For those, [`/hazards/{id}`](https://api-v2-docs.disasteraware.com/hazards/requesting-a-hazard.md) with the numeric id still returns the V1-shaped record.
