Short-range streamflow (chrtout)¶

A live pull from the public, unsigned noaa-nwm-pds bucket — no credentials. We fetch one recent short_range channel_rt file: the whole-CONUS channel-routing output (~14 MB) holding streamflow for all ~2.7 M NHDPlus reaches at one forecast hour.

We pick a cycle date a few days back so it is reliably published, and a whole-Earth bbox so the request is a plain whole-file download (a narrower bbox or sites= would need the pyramids PY-G reader).

Setup¶

The imports: EarthLens is the unified entry point; tempfile / pathlib give us a scratch directory for the downloaded NetCDF, and datetime is used to compute a recent cycle date.

In [ ]:

import datetime as dt
import tempfile
from pathlib import Path

from earthlens import EarthLens

import datetime as dt import tempfile from pathlib import Path from earthlens import EarthLens

Pick a cycle date and a scratch directory¶

We target a cycle three days back so it is reliably published, and download into a fresh temporary directory.

In [ ]:

cycle_date = (dt.datetime.now() - dt.timedelta(days=3)).strftime('%Y-%m-%d')
out_dir = Path(tempfile.mkdtemp(prefix='nwm_'))
print('cycle date:', cycle_date)

cycle_date = (dt.datetime.now() - dt.timedelta(days=3)).strftime('%Y-%m-%d') out_dir = Path(tempfile.mkdtemp(prefix='nwm_')) print('cycle date:', cycle_date)

Fetch the channel_rt file¶

We build the request first: the nwm backend, the short_range configuration, cycle 0 and forecast step 1, and a whole-Earth aoi so the request is a plain whole-file download.

In [ ]:

nwm = EarthLens(
    data_source="nwm",
    start=cycle_date,
    end=cycle_date,
    dataset='chrtout',
    variables=['streamflow'],
    aoi=[-180, -90, 180, 90],
    configuration='short_range',
    cycles=[0],
    steps=[1],
    path=str(out_dir),
)

nwm = EarthLens( data_source="nwm", start=cycle_date, end=cycle_date, dataset='chrtout', variables=['streamflow'], aoi=[-180, -90, 180, 90], configuration='short_range', cycles=[0], steps=[1], path=str(out_dir), )

Download¶

download() returns the list of written paths; we print each file's name and size.

In [ ]:

paths = nwm.download(progress_bar=False)
for p in paths:
    print(p.name, f'{p.stat().st_size / 1e6:.1f} MB')

paths = nwm.download(progress_bar=False) for p in paths: print(p.name, f'{p.stat().st_size / 1e6:.1f} MB')

The downloaded file is the native NWM NetCDF. Decoding channel_rt into a tidy feature_id × time streamflow table — and subsetting it to specific reaches or a bounding box — is a read, which is the pyramids PY-G capability (unreleased). Until then earthlens fetches the whole file and leaves the decode to you / pyramids; it never imports xarray/zarr itself.

Confirm the download¶

A quick sanity check that we got the expected file, and a peek at the backend's OUTPUT_KIND (tabular).

In [ ]:

assert paths, 'expected one channel_rt file for a recent short_range cycle'
print('OUTPUT_KIND:', nwm.OUTPUT_KIND)  # tabular

assert paths, 'expected one channel_rt file for a recent short_range cycle' print('OUTPUT_KIND:', nwm.OUTPUT_KIND) # tabular