Spectral Decomposition and Verification of NOGAPS 500MB Medium-Range Forecasts.
Abstract
A major goal in the atmospheric sciences is accurate and reliable numerical weather prediction. An important aspect of numerical model development is verification studies which can isolate model strengths or weaknesses and indicate errors which must be eliminated. This thesis investigates the Navy Operational Global Atmospheric Prediction System (NOGAPS) model forecasts of the position and movement of waves at the 500mB level. Twelve NOGAPS 500mB 5-day forecasts were spectrally decomposed into wavenumber groupings for verification purposes. Four forecasts were from the NOGAPS 2.0 ( 6-level) version and eight from the NOGAPS 2.1 (9-level) version. Wavenumber components of the forecast and observed waves were grouped into planetary (wavenumbers 1-3), long (wavenumbers 4-7) and medium (wavenumbers 8-12) to facilitate model comparison. Hovmoller (time-longitude) diagrams were used to analyze the observed and forecast fields. Two systematic errors emerged; amplitude smoothing of wave and trough features, and consistent positive error at high latitudes. NOGAPS 2.1 demonstrated modest improvement (over NOGAPS 2.0) as error magnitudes were reduced and initiation of error occurred later in the forecast.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1984
- Accession Number
- ADA152550
Entities
People
- R. C. Showalter
Organizations
- Naval Postgraduate School