The prediction of January surface temperature : planetary wave climatology and a diagnostic case study of specification equation forecast errors
MetadataShow full item record
This study involved the analysis of the synoptic climatology of planetary wave flow relevant to the prediction of January surface temperature. Three statistical models were developed (Correlation, Anomalous Flow and Regression model) to aid in the understanding of how mid-tropospheric flow relates to surface air temperature. Linear steady state primitive equation models, such as the models of J. D. Opsteegh and H. M. Van den Dool described in 1980, predict patterns of geopotential height anomalies for very cold versus warm Januaries in the eastern half of the United States. These theoretical prediction models show that the Pacific North American teleconnection pattern (PNA) is the common pattern for height anomalies in the Northern Hemisphere for below normal Januaries. Test results from this study were consistent with the hypothetical height anomaly pattern of theoretical prediction models. A study was conducted to determine what determines the pattern of geopotential height anomalies for a given January. Specification equations were developed using a stepwise regression model. Test results from the Regression model determined that below normal Januaries are forecast with greater accuracy than above normal Januaries. A series of diagnostic test models was developed to search for reasons why colder than normal Januaries have greater forecast accuracy than above normal Januaries. The results from the statistical models yielded useful predictors which can be utilized in long-range temperature prediction for the eastern half of the United States.