Abstract
An examination is made of the current National Meteorological Center (NMC) operational models’ ability to forecast surface anticyclones. A study of the 1981–82 cold season reveals systematic underprediction of the phenomenon on the part of both the Limited Area Fine Mesh (LFM) and spectral models. However, the LFM forecasts weaker anticyclones than does the spectral model. This difference is apparent in the region of eastern North America and the western Atlantic Ocean. The systematic underprediction found in this study is as great as Colucci and Bosart found for NMC's six-layer primitive equation model.
No overall systematic forecast bias is found for the 1000–500 mb mean temperatures over the surface anti-cyclones. However, excessively warm temperatures are forecast in the Pacific northwest region of both models, and the LFM forecasts erroneously cold temperatures in the western Atlantic basin south of 40°N. The spectral model shows a significant improvement over the LFM in this latter region.
The mean anticyclone displacement error for both models at 48-h range is about 500 km. There is also a tendency for both models to place anticyclones erroneously to the south and east of their observed positions, suggesting the models' translation of these anticyclones to be too fast. Colucci and Bosart also found a fast bias, but this study suggests an overall improvement in anticyclone placement.
Finally, a case of a recent poorly forecasted anticyclone-cyclone complex illustrates the deleterious effects those forecasts can have in the attempt to correctly forecast significant precipitation events. Our study shows an unforecasted precipitation event to have occurred in a lower troposphere warm advection region associated with a poorly forecasted surface anticyclone.
