1.3 The Development and Success of NCEP's Global Forecast System

Monday, 7 January 2019: 9:00 AM
North 222C (Phoenix Convention Center - West and North Buildings)
Environmental Modeling Center, EMC, College Park, MD; and K. Campana, P. Caplan, D. J. Halperin, B. Lapenta, S. Lilly, Y. Lin, A. B. Penny, S. Saha, V. Tallapragada, G. H. White, J. S. Woollen, and F. Yang
Manuscript (2.7 MB)

Handout (909.2 kB)

The invention of the first electronic computer in 1946 enabled a continuing revolution in weather forecasting driven by analyses and forecasts produced by integrating the equations of motion of the atmosphere. In the United States public forecasts have become increasingly based on computer analyses and forecasts produced by the National Centers for Environmental Prediction (NCEP) Production Suite, the foundation of which is the Global Forecast System (GFS). Most of the other NCEP forecast systems use fields from the GFS.

Following the introduction in the 1970s of spectral techniques that produced forecasts as skillful as and more computationally efficient than finite differencing methods, Joseph Sela began to develop spectral modeling at NCEP (then the National Meteorological Center (NMC)) in 1975. The first global NMC model with a hydrostatic spectral dynamic core became operational in Aug. 1980 with a limited physics package, 12 layers in the vertical and a horizontal resolution of 30 waves, equivalent to 375 km.

Over the years vertical and horizontal resolution have increased as has the complexity of the physics and the dynamic core. Its current resolution is 64 layers in the vertical and 1534 waves, equivalent to 13 km. Of at least equal importance have been advances in data assimilation techniques and observations. This paper reviews the evolution and increasing skill of the GFS over its 39-year history.

The GFS’s resolution now approaches the atmospheric scales where non-hydrostatic effects are important. The American meteorological community joined with NCEP to test several non-hydrostatic dynamic cores and choose a new dynamic core for NCEP’s GFS.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner