NWM 1km LDAS rechunked dataset

recipes/NWM-1km/meta.yaml

@@ -0,0 +1,22 @@
+title: 'NWM-2.1-grid1km-LDAS-rechunked'
+description: LDAS variables at 3-hourly intervals from the US National Water Model Version 2.1 Retrospective (1980-). There are 21 LDAS variables including soil moisture, rain, snow, evapotranspiration.  This dataset has been rechunked from the original chunk pattern to better facilitate time series access.
+pangeo_forge_version: '0.9.2'
+pangeo_notebook_version: '2022.07.17'
+recipes:
+  - id: NWM-2.1-grid1km-LDAS-rechunked
+    object: 'recipe:recipe'
+provenance:
+  providers:
+    - name: 'NOAA Office of Water Prediction'
+      description: 'US NOAA OWP'
+      roles:
+        - provider
+        - licensor
+      url: https://water.noaa.gov/about/nwm
+  license: 'CC0'
+maintainers:
+  - name: 'Rich Signell'
+    orcid: '0000-0003-0682-9613'
+    github: rsignell
+bakery:
+  id: 'pangeo-ldeo-nsf-earthcube'

recipes/NWM-1km/recipe.py

@@ -0,0 +1,29 @@
+import pandas as pd
+
+from pangeo_forge_recipes.patterns import ConcatDim, FilePattern
+from pangeo_forge_recipes.recipes import XarrayZarrRecipe
+
+dates = pd.date_range('1979-02-01 03:00', '2020-12-31 21:00', freq='3H')
+
+time_concat_dim = ConcatDim('time', dates, nitems_per_file=1)
+
+URL_FORMAT = (
+    's3://noaa-nwm-retrospective-2-1-pds/model_output/'
+    '{time:%Y}/{time:%Y%m%d%H}00.LDASOUT_DOMAIN1.comp'
+)
+
+
+def make_url(time):
+    return URL_FORMAT.format(time=time)
+
+
+pattern = FilePattern(make_url, time_concat_dim)
+
+target_chunks = {'time': 72, 'x': 512, 'y': 512}
+
+
+def preprocess(ds):
+    return ds.drop('reference_time')
+
+
+recipe = XarrayZarrRecipe(pattern, target_chunks=target_chunks, process_chunk=preprocess)