Automated Probability Forecasts of Ceiling and Visibility Based on Single-Station Data
Abstract
A set of computer programs was developed to produce multiple linear regression equations for predicting the probability of specified ceiling and visibility categories at air terminals. The equations were based upon weather observations made solely at the terminal and were derived with the REEP screening technique from 329 possible predictors. The program(s) accepted raw data in a standard National Climatic Center format, and a complete set of prediction equations for five time projections was produced for each of 50 stations in a single computer run. The accuracy of forecasts generated by the equations was evaluated for 20 terminals. Three measures of accuracy were used to compare the objective forecasts to persistence and climatology. Computer programs were developed to prepare forecasts on an operational basis for 20 terminals in the eastern U.S. and 23 terminals in Alaska.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1973
- Accession Number
- AD0762461
Entities
People
- Frank Lewis
- Richard L. Crisci
Organizations
- National Weather Service