Cropping ARL Data#

arl-met supports two related workflows:

  • crop or subset lazily while reading into xarray

  • write a new ARL file that contains only the selected domain, levels, or variables

Lazy crop while reading#

Pass bbox= to arlmet.open_dataset() when you want an in-memory subset for analysis.

import arlmet

ds = arlmet.open_dataset(
    "input.arl",
    bbox=(-114.0, 39.0, -110.0, 42.0),
)

This path crops before unpacking, so it avoids reading the full grid when you only need a smaller region.

Select levels while reading#

ARL levels are selected by integer level index.

ds = arlmet.open_dataset(
    "input.arl",
    bbox=(-114.0, 39.0, -110.0, 42.0),
    levels=[0, 1, 2],
)

Write a smaller ARL file#

Use arlmet.extract_subset() when you want a new ARL file on disk.

import arlmet

arlmet.extract_subset(
    "input.arl",
    "cropped.arl",
    bbox=(-114.0, 39.0, -110.0, 42.0),
)

extract_subset() returns the new file opened in read mode, so you can chain straight into analysis. Use it as a context manager (or call .close()) when you keep the return value; ignore it if you only need the file on disk.

with arlmet.extract_subset(
    "input.arl", "cropped.arl", bbox=(-114.0, 39.0, -110.0, 42.0)
) as cropped:
    ds = cropped.to_dataset()

When you already hold an open arlmet.File, the same operation is available as a method:

with arlmet.File("input.arl") as met:
    with met.extract_subset("cropped.arl", bbox=(-114.0, 39.0, -110.0, 42.0)) as cropped:
        ds = cropped.to_dataset()

Keep only the levels and variables you need#

extract_subset() can keep only the levels and variables you need.

arlmet.extract_subset(
    "input.arl",
    "subset.arl",
    bbox=(-114.0, 39.0, -110.0, 42.0),
    levels=[0, 1, 2],
    variables=["UWND", "VWND", "TEMP"],
)

What changes in the output file#

When levels are selected, the output file is renumbered from zero upward while preserving the original level values in the vertical axis metadata.

When variables are selected, only those variables are written into each output time step. Forecast hours from the source index records are preserved.

Limitations#

  • bbox is always interpreted as (west, south, east, north) in degrees

  • projected-grid bounding boxes must not cross the dateline

  • the cropped grid must still be large enough to hold the ARL index record

If the selected domain is too small to hold the output index record, extract_subset() raises a ValueError explaining how many grid cells are required.