Climate Prediction. Part 1. Cyclone Frequency.

Climate prediction models based on multivariate analyses of cyclone frequencies are constructed from historical data (1885-1960) and evaluated for forecast skill on independent data (1960-1983). Cyclone frequencies are predicted for six-month duration seasons at 87 locations over eastern North Ameri...

Full description

Bibliographic Details
Main Author: Hayden,B P
Other Authors: VIRGINIA UNIV CHARLOTTESVILLE DEPT OF ENVIRONMENTAL SCIENCES
Format: Text
Language:English
Published: 1984
Subjects:
Meteorology
Statistics and Probability
*MATHEMATICAL MODELS
*MULTIVARIATE ANALYSIS
*WEATHER FORECASTING
*CYCLONES
FREQUENCY
TIME INTERVALS
SKILLS
PREDICTIONS
MODELS
GRAPHS
TEST METHODS
HISTORY
CLIMATE
SCORING
SEASONS
WUNR389170
North Atlantic
Online Access:http://www.dtic.mil/docs/citations/ADA143576
http://oai.dtic.mil/oai/oai?&verb=getRecord&metadataPrefix=html&identifier=ADA143576
Description
Summary:Climate prediction models based on multivariate analyses of cyclone frequencies are constructed from historical data (1885-1960) and evaluated for forecast skill on independent data (1960-1983). Cyclone frequencies are predicted for six-month duration seasons at 87 locations over eastern North America and the western North Atlantic from 27.5 deg to 55 deg. Three types of principal components models are constructed and tested. Model I uses unrotated principal component axes; Model II uses rigid rotation of the component axes; and, Model III uses oblique rotations of the component axes. Forecast skill averages 75% correct for 2 category measure of forecasts.

Similar Items