Monthly climate normals

Available since v0.11.0. The manifest requests 1991-2020 monthly normals of maximum temperature, minimum temperature, and precipitation at Will Rogers World Airport through noaa:climate-normals. Normals are 30-year averages, so the source needs no dates; period: monthly selects the monthly dataset and the whole year is requested.

In an activated Python 3.11+ virtual environment, install the published package and save the manifest as dataset.yaml in a working directory. Run the commands from that directory:

python -m pip install "usdata[pandas]"
usdata pull dataset.yaml
usdata verify dataset.yaml

Open the downloaded CSV with the generic pandas reader and compare an observed month against its normal. Run this with python in the same directory and environment:

from pathlib import Path

from usdata import pull, verify

manifest = Path("dataset.yaml")
(item,) = pull(manifest).fetched
normals = item.open(dtype={"DATE": "string"}).set_index("DATE")
print(normals[["MLY-TMAX-NORMAL", "MLY-TMIN-NORMAL", "MLY-PRCP-NORMAL"]])
may = normals.loc["05"]
print(
    f"May normal: high {may['MLY-TMAX-NORMAL']:.1f} C, "
    f"low {may['MLY-TMIN-NORMAL']:.1f} C, precipitation {may['MLY-PRCP-NORMAL']:.1f} mm"
)
assert verify(manifest) == []

For a source installation, run from examples/climate-normals/ and use uv run usdata and uv run python.

DATE is the two-digit month; explicit dtype keeps it as text. Metric units give degrees Celsius and millimeters. To compare with observations, pull the same station and month through noaa:gsom as in the monthly climate notebook and subtract the normal from the observed value to get the monthly anomaly.

For daily normals, set period: daily and optionally start and end to keep a month-day window; the year is ignored. See the NOAA access notes for variable codes, units, and limitations. Keep the manifest, lockfile, and cached bytes together: normals are corrected occasionally, and a checksum cannot recreate bytes that upstream no longer serves.