Skip to content

CatRaRE heavy-rainfall events — usage#

The catrare backend needs no credentials — the CatRaRE FileGDBs are public on opendata.dwd.de. A request selects a threshold and, optionally, a date window and/or a bounding box; the downloaded FileGDB is cached, so repeated requests reuse it.

For the concepts (what the data is, the thresholds, the CRS) see Introduction; the rendered API is the Reference page.

Quick start#

from earthlens.core import EarthLens

# Every T5 heavy-rainfall event (footprint polygons), reprojected to WGS84.
fc = EarthLens(
    "catrare",
    threshold="t5",
    path="catrare-data",
).download()

fc.head()  # a FeatureCollection: Event_ID, Date_START, …, Area, Eta, geometry

download() returns a pyramids FeatureCollection (a geopandas.GeoDataFrame subclass) in EPSG:4326 and also writes it to a GeoPackage under path.

Filtering by date and area#

# The July 2021 flood over west Germany (Ahr / NRW).
ahr = EarthLens(
    "catrare",
    threshold="t5",
    start="2021-07-01",
    end="2021-07-31",
    lat_lim=[50.0, 51.5],
    lon_lim=[6.0, 8.0],
).download()

start / end keep events whose [Date_START, Date_END] interval overlaps the window (either bound may be omitted). The bounding box keeps events whose geometry intersects it — the coordinates are ordinary WGS84 because the backend reprojects off the DWD RADOLAN grid.

Thresholds and geometry layers#

# The severity-weighted selection, as maximum-rainfall points instead of zones.
pts = EarthLens(
    "catrare",
    threshold="w3",
    geometry_layer="points",
).download()

threshold= is "t5" (return period ≥ 5 yr) or "w3" (severity-weighted); geometry_layer= is "zones" (event-footprint polygons, default) or "points" (one maximum-rainfall point per event).

Tabular output#

Pass geometry=False to drop the geometry and get a plain pandas.DataFrame of the events and their 13 core attributes:

table = EarthLens("catrare", threshold="t5", geometry=False).download()