Get AI Briefings, Transcripts & Alerts on Local & National Government Meetings — Forever.
Technical call: Officials warn NBMgrids can be biased and urged local NWS contact
Summary
Participants on the FEMs transition call discussed how the National Blend of Models (NBM) is often used as the forecasters' starting point and can run "too moist," creating delayed relative-humidity responses that affect fire-danger outputs; speakers urged local contact with National Weather Service offices and greater transparency about forecast inputs.
Speaker 2, an on-call weather-service forecaster, explained how FEMs is ingesting forecast products and the operational implications for fire-danger outputs. "The NDFD is the official forecast product from the National Weather Service," Speaker 2 said, and forecasters commonly begin with the National Blend of Models (NBM) as the base for edits. He added that the NBM "runs too moist" during summer in some regions and can be slow to respond to rapid changes, which can leave RH grids that do not reflect sudden dryings.
The host (Speaker 1) said the call aimed to increase transparency about how forecasts are derived so fire managers can anticipate directional bias. "When we know that these models are biased in one direction or another, we can have a much better discussion and anticipate these processes going forward," Speaker 1 said. The discussion centered on how a known bias (for example, a forecast that tends moist) can be handled operationally: managers can treat model outputs as directional guidance rather than an absolute trigger for staffing or resource decisions.
Several participants emphasized practical next steps. Speaker 2 urged local users to bring case examples to their local National Weather Service offices (screenshots or paired observed vs. forecast examples) so forecasters can examine and, where appropriate, edit local forecasts. "Bring them case examples," Speaker 2 said, "so that'll help them at least start digging in." Speaker 1 said the project team has documentation (prepared with NOAA research branch input) describing the forecast process and offered to share that document with call participants to clarify how files are produced and ingested.
Speakers also discussed the difference between the older WIMS workflow and the current FEMs approach: WIMS forced a human-edited text forecast that was reviewed at the station level, whereas FEMs is more gridded and largely hands-off after publication. While gridded inputs enable the potential for a true gridded fire-danger product, participants cautioned that gridded approaches bring trade-offs, including slower local responsiveness and the need for bias correction.
The call concluded with a recommendation that users contact local NWS offices with specific examples and that the FEMs team continue to document the forecast ingestion process. The group plans follow-up discussions on forecast handling at the next scheduled call.
