jgfs_atmos_prep ecf (v17) Analysis - TerrenceMcGuinness-NOAA/global-workflow GitHub Wiki
jgfs_atmos_prep.ecf (GFS v17)
Branch / tenant: dev/gfs.v17 (gw_v17)
Path: ecf/scripts/gfs/prep/atmos/jgfs_atmos_prep.ecf
Analysis date: 2026-06-11
Summary
jgfs_atmos_prep.ecf is the NCO production ecFlow task that performs atmospheric observation preparation (PREP) — decoding, quality-controlling, and formatting raw observational data (conventional, satellite, radar) into the prepBUFR/BUFR format needed by the GSI analysis. It calls the Rocoto job card prep.sh directly (not a J-Job in dev/jobs/).
1. Resources & Configuration
| Field | Value |
|---|---|
| PBS resources | 3 nodes, 5 mpiprocs/node, 480 GB/node, exclusive host, 30-min wall |
| Modules | load_modules.sh run |
| Job call | ${HOMEgfs}/dev/job_cards/rocoto/prep.sh (Rocoto card, NOT a standard J-Job) |
2. v17-specific notes
- Calls Rocoto job card directly: unlike most v17 ECF scripts that call
${HOMEgfs}/jobs/J*, this one invokesdev/job_cards/rocoto/prep.sh. This may indicate the prep system has not been fully migrated to the J-Job pattern. - High memory: 480 GB per node for observation processing (large BUFR datasets).
- 3 nodes: parallel observation decoding across data sources.
- Pre-analysis: PREP must complete before the GSI atmospheric analysis can run.
- CRITICAL on critical path: the analysis chain starts with prep.
3. Change-impact snapshot
| Symbol | Risk | Notes |
|---|---|---|
prep.sh |
HIGH | Upstream of all atmospheric data assimilation |
Sources
- On-disk v17:
ecf/scripts/gfs/prep/atmos/jgfs_atmos_prep.ecf.