GribStream

National Blend of Models (NBM)

GribStream Code: nbm

Description

NBM (National Blend of Models) is NOAA’s flagship post-processing system for gridded forecasts across the United States. It combines a wide suite of numerical weather prediction models—including global, regional, deterministic, and ensemble systems—into a calibrated, high-resolution blend. By merging model guidance from the GFS, HRRR, RAP, GEFS, and international systems such as the ECMWF IFS, NBM creates a unified forecast that balances skill across timescales, geographic regions, and variables.

The blending is performed with statistically sophisticated techniques such as Model Output Statistics (MOS), quantile mapping, and ensemble weighting. Forecasts are bias-corrected against the URMA analysis of record, ensuring consistency and reliability. The NBM framework outputs both deterministic fields (e.g., temperature, wind, QPF) and probabilistic products (e.g., probability of precipitation, percentile guidance, probability of exceedance) that directly support risk-based decision making.

NBM is designed to deliver consistent forecasts nationwide: it provides seamless coverage from short-range (hours) through medium-range (days) horizons, updated on an hourly cycle. Its post-processed nature smooths out discontinuities between different model sources and offers a stable, trusted foundation for the National Weather Service digital forecast grids.

Detail

Parameters: