APPLICATION OF RADAR DATA TO FORECASTING OF TERMINAL WEATHER CONDITIONS
Abstract
A screening and regression technique was applied to 60 predictands and 120 possible predictors in more or less random positions within a range of 60 miles. Each station consisted of a five by five mile square. The investigation was based on records of selected storms which contained large precipitation areas and lasted several hours. It was found that for prediction times ranging from 10 to 240 minutes the use of 4 to 11 predictors selected by the screening process provided a mean reduction in variance which was considerably larger than that obtained simply by persistence or advection. The linear predictability of precipitation was shown to be relatively insensitive to reasonable variations in spatial density of data, but the spatial extent may be a critical parameter. The space resolution with which a forecast is made appears to influence to some extent the level of predictability. Comparison was made of the variations in precipitation rate as measured by radar with fluctuations in ceiling and visibility, recorded by rotating-beam ceilometers and transmissometers respectively at two airports which are located within a few miles of the radar site. It was found that the time scale of the variations is approximately the same for the three quantities, but the relationships between them differ for different storm types and geographical locations. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1962
- Accession Number
- AD0293421
Entities
People
- Pauline M. Austin
Organizations
- Air Force Cambridge Research Laboratories