jgfs_marine_bmat_init ecf (v17) Analysis - TerrenceMcGuinness-NOAA/global-workflow GitHub Wiki
jgfs_marine_bmat_init.ecf (GFS v17)
Branch / tenant: dev/gfs.v17 (gw_v17)
Path: ecf/scripts/gfs/analysis/marine/jgfs_marine_bmat_init.ecf
Analysis date: 2026-06-11
Method: AWS Bedrock AgentCore agentcore-mcp-rag MCP cross-checked against the on-disk v17 worktree.
Summary
jgfs_marine_bmat_init.ecf is the NCO production ecFlow task that initializes the B-matrix computation for the marine data assimilation. It stages ensemble member data and prepares the working environment for the subsequent marine_bmat task. It calls JGLOBAL_MARINE_BMAT_INITIALIZE.
1. The script itself
| Section | Detail |
|---|---|
| PBS resources | 1 node, 128 cpus, 24 GB, 20-min wall, shared |
| Modules | load_modules.sh ufsda (JEDI/SOCA stack) |
| J-Job | JGLOBAL_MARINE_BMAT_INITIALIZE |
| Rocoto breadcrumb | marinebmatinit.sh |
2. Execution chain
jgfs_marine_bmat_init.ecf
+-- JGLOBAL_MARINE_BMAT_INITIALIZE
+-- exglobal_marine_bmat_initialize.py (Python/pygfs)
+-- Stages ensemble ocean/ice backgrounds
+-- Prepares BUMP operator configuration
3. v17-specific notes
- Lightweight setup for B-matrix computation (128 cpus but only 24 GB).
- Python-based driver using
pygfs. - JEDI/SOCA: loads
ufsdamodules. - Part of marine DA chain: runs before
marine_bmat, stages ensemble data. - 20-min walltime — data staging operation.
4. Change-impact snapshot
| Symbol | Risk | Notes |
|---|---|---|
JGLOBAL_MARINE_BMAT_INITIALIZE |
MEDIUM | Upstream of B-matrix computation |
Sources
- On-disk v17 worktree:
ecf/scripts/gfs/analysis/marine/jgfs_marine_bmat_init.ecf. - MCP (tenant
gw_v17): on-disk analysis.