Recent Developments in Automated Prediction of Ceiling and Visibility

Joseph R. Bocchieri Techniques Development Laboratory, National Weather Service, NOAA, Silver Spring, Md. 20910

Search for other papers by Joseph R. Bocchieri in
Current site
Google Scholar
PubMed
Close
,
Richard L. Crisci Techniques Development Laboratory, National Weather Service, NOAA, Silver Spring, Md. 20910

Search for other papers by Richard L. Crisci in
Current site
Google Scholar
PubMed
Close
,
Harry R. Glahn Techniques Development Laboratory, National Weather Service, NOAA, Silver Spring, Md. 20910

Search for other papers by Harry R. Glahn in
Current site
Google Scholar
PubMed
Close
,
Frank Lewis Techniques Development Laboratory, National Weather Service, NOAA, Silver Spring, Md. 20910

Search for other papers by Frank Lewis in
Current site
Google Scholar
PubMed
Close
, and
Frank T. Globokar Techniques Development Laboratory, National Weather Service, NOAA, Silver Spring, Md. 20910

Search for other papers by Frank T. Globokar in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

The history of the National Weather Service's development efforts in terminal weather prediction for aviation is discussed and results from recent experiments involving three approaches are presented.

In one approach, single-station equations for predicting the probability of specific ceiling and visibility categories are developed. The equations are based upon weather observations at the local terminal only and are derived by using the Regression Estimation of Event Probabilities screening technique.

In another approach, Model Output Statistics (MOS) is used to develop probability forecast equations for ceiling and visibility. MOS consists of determining a statistical relationship between a predictand and the forecast output of numerical prediction models. The statistical relationship is determined by screening regression in this paper. The two numerical models used are the National Meteorological Center's (NMC) primitive equation (PE) model and the Techniques Development Laboratory's Subsynoptic Advection Model (SAM). A forecast system developed by MOS is shown to be useful as guidance to the Aviation Forecast Branch at NMC.

In a third approach, another probability forecast system, called SINGMOS (SINGle station and Model Output Statistics), is developed. SINGMOS is a combination of the single-station and MOS systems and includes observed surface data, forecast output from SAM and PE, and forecasts from single-station prediction equations as predictors. From comparison with official terminal forecasts (FT's) on independent data, it is concluded that SINGMOS is better than the FT's for the long-range (8–16 hr) forecasts; however, for the short-range (4 hr) forecasts, the FT's are better than SINGMOS.

Abstract

The history of the National Weather Service's development efforts in terminal weather prediction for aviation is discussed and results from recent experiments involving three approaches are presented.

In one approach, single-station equations for predicting the probability of specific ceiling and visibility categories are developed. The equations are based upon weather observations at the local terminal only and are derived by using the Regression Estimation of Event Probabilities screening technique.

In another approach, Model Output Statistics (MOS) is used to develop probability forecast equations for ceiling and visibility. MOS consists of determining a statistical relationship between a predictand and the forecast output of numerical prediction models. The statistical relationship is determined by screening regression in this paper. The two numerical models used are the National Meteorological Center's (NMC) primitive equation (PE) model and the Techniques Development Laboratory's Subsynoptic Advection Model (SAM). A forecast system developed by MOS is shown to be useful as guidance to the Aviation Forecast Branch at NMC.

In a third approach, another probability forecast system, called SINGMOS (SINGle station and Model Output Statistics), is developed. SINGMOS is a combination of the single-station and MOS systems and includes observed surface data, forecast output from SAM and PE, and forecasts from single-station prediction equations as predictors. From comparison with official terminal forecasts (FT's) on independent data, it is concluded that SINGMOS is better than the FT's for the long-range (8–16 hr) forecasts; however, for the short-range (4 hr) forecasts, the FT's are better than SINGMOS.

Save