---
license: apache-2.0
pretty_name: NEXRAD Storm Catalog
tags:
  - weather
  - radar
  - nexrad
  - storm-events
  - meteorology
  - geospatial
task_categories:
  - tabular-classification
  - time-series-forecasting
size_categories:
  - 100K<n<1M
language:
  - en
---

# NEXRAD Storm Catalog

A metadata index pairing storm events from the [NOAA Storm Events Database](https://www.ncei.noaa.gov/stormevents/)
with nearby NEXRAD WSR-88D weather-radar stations. Each row is one
`(event, station)` pair where the station was within **150 km** of the event
location, scoped to **2015–2024**.

This catalog is **metadata only**. It does **not** contain reflectivity,
velocity, or any other radar payloads — it points at radar windows so a
downstream pipeline can fetch the matching Level-2 volumes from the
[Unidata NEXRAD Level-2 archive on AWS Open Data](https://registry.opendata.aws/noaa-nexrad/)
(`s3://unidata-nexrad-level2/`).

## Files

| File | Rows | Size | Description |
|---|---|---|---|
| `nexrad_event_catalog.parquet` | 583,674 | ~70 MB | Storm event ↔ station pairs |
| `nexrad_stations.csv` | 158 | ~7 KB | WSR-88D station metadata |
| `LICENSE` | — | — | Apache License 2.0 |
| `README.md` | — | — | This file |

## Schema — `nexrad_event_catalog.parquet`

| Column | Type | Description |
|---|---|---|
| `station_id` | `str` | WSR-88D station identifier (e.g. `KTLX`, `KABR`) |
| `event_id` | `int` | NOAA Storm Events identifier |
| `event_type` | `str` | One of 17 storm types (Thunderstorm Wind, Hail, Flash Flood, Flood, Winter Storm, Heavy Snow, Tornado, Heavy Rain, Blizzard, Tropical Storm, Lightning, Funnel Cloud, Ice Storm, Waterspout, Lake-Effect Snow, Dust Storm, Hurricane (Typhoon)) |
| `category` | `str` | Coarse 6-way grouping: `severe_convective`, `heavy_rain_flood`, `winter_weather`, `tropical`, `other_severe`, `general` |
| `ef_scale` | `str` \| `null` | EF rating for tornado events |
| `begin_datetime_utc` | `datetime[UTC]` | Event start time |
| `end_datetime_utc` | `datetime[UTC]` | Event end time |
| `distance_km` | `float` | Great-circle distance from station to event location |
| `begin_lat` | `float` | Event latitude (degrees) |
| `begin_lon` | `float` | Event longitude (degrees) |
| `state` | `str` | US state / territory code |
| `magnitude` | `float` \| `null` | Event magnitude (units depend on `magnitude_type`) |
| `magnitude_type` | `str` \| `null` | Magnitude units / scale |
| `episode_narrative` | `str` | Free-text NWS narrative for the surrounding episode |

### Category mapping

| Category | Event types | Row count |
|---|---|---|
| `severe_convective` | Tornado, Hail, Thunderstorm Wind | 383,134 |
| `heavy_rain_flood` | Heavy Rain, Flash Flood, Flood | 114,761 |
| `winter_weather` | Winter Storm, Ice Storm, Blizzard, Heavy Snow | 67,299 |
| `general` | Lightning, Waterspout, Lake-Effect Snow (fallback bucket) | 8,427 |
| `tropical` | Tropical Storm, Hurricane (Typhoon) | 5,455 |
| `other_severe` | Funnel Cloud, Dust Storm | 4,598 |

## Schema — `nexrad_stations.csv`

| Column | Type | Description |
|---|---|---|
| `station_id` | `str` | WSR-88D station identifier — matches `station_id` in the events table |
| `name` | `str` | Human-readable site name (e.g. *Lake Charles*, *El Paso*) |
| `lat` | `float` | Station latitude (degrees) |
| `lon` | `float` | Station longitude (degrees) |
| `elevation_m` | `float` | Antenna elevation above sea level (metres) |
| `state` | `str` \| `null` | US state / territory code, where available |
| `rda_type` | `str` \| `null` | Radar Data Acquisition unit type, where available |

Coverage: 158 NEXRAD WSR-88D sites across the United States and territories.

## Quick start

```python
import pandas as pd

events = pd.read_parquet(
    "hf://datasets/Silurian/nexrad-storm-catalog/nexrad_event_catalog.parquet"
)
stations = pd.read_csv(
    "hf://datasets/Silurian/nexrad-storm-catalog/nexrad_stations.csv"
)

print(events.shape)               # (583674, 14)
print(events["category"].value_counts())
```

Filter to a single category, station, and year:

```python
ktlx_tornadoes_2019 = events[
    (events["station_id"] == "KTLX")
    & (events["event_type"] == "Tornado")
    & (events["begin_datetime_utc"].dt.year == 2019)
]
```

## How the catalog was built

1. Download the NOAA Storm Events Database bulk CSVs (2015–2024) from
   <https://www.ncei.noaa.gov/pub/data/swdi/stormevents/csvfiles/>.
2. Join each storm event to all WSR-88D stations within 150 km, using a
   great-circle distance computed against the station coordinates in
   `nexrad_stations.csv`.
3. Map each `event_type` into the 6-way `category` field per the table above
   (events whose type is not in any explicit bucket are tagged `general`).

## Provenance & credits

- **Storm event records** — [NOAA Storm Events Database](https://www.ncei.noaa.gov/stormevents/),
  US-government public-domain data published by NOAA NCEI.
- **WSR-88D station metadata** — [NOAA National Weather Service](https://www.weather.gov/),
  Radar Operations Center.
- **Underlying radar volumes** (not redistributed here) — NOAA NEXRAD Level-2
  on AWS Open Data, hosted by Unidata: `s3://unidata-nexrad-level2/`.
- **Catalog assembly, categorisation, and packaging** — Silurian AI.

## Citation

If you use this catalog, please cite both the upstream NOAA source and this
packaged release.

```bibtex
@misc{silurian_nexrad_storm_catalog_2026,
  title  = {NEXRAD Storm Catalog},
  author = {{Silurian AI}},
  year   = {2026},
  url    = {https://huggingface.co/datasets/Silurian/nexrad-storm-catalog}
}

@misc{noaa_storm_events,
  title  = {NOAA Storm Events Database},
  author = {{National Oceanic and Atmospheric Administration, National Centers for Environmental Information}},
  url    = {https://www.ncei.noaa.gov/stormevents/}
}
```

## License

This catalog is released under the **Apache License 2.0** — see the bundled
`LICENSE` file for the full text.

The upstream NOAA Storm Events records are US-government public-domain data;
the Apache 2.0 license here covers the value-add layer (the join to
WSR-88D stations, the category mapping, and the packaged distribution).