Meteorologists who work in the energy commodities market continue to investigate ways to enhance
predictions of seasonal temperature anomalies using oceanic/atmospheric indices. This study examines
the relationship of three climate indices - ENSO (El Niño/Southern Oscillation), PNA (Pacific North
American) and NAO (North Atlantic Oscillation) - to heating degree day (HDD) totals accumulated
in 11 cities in the Midwest and northeastern United States, to determine which, if any, has predictive
power. The data covers the 48-year period between 1951/52 and 1998/99, and focuses on two periods
either side of 1 January (i.e. the winter months of October-December and January-April). The index
most strongly related to the HDD anomalies during both winter periods was NAO. NAO values were
negative for cold (above-average HDD) anomalies occurring prior to and after 1 January, while the
NAO values were generally positive during warm (below-average HDD) anomalies. During cold
anomalies, the PNA values were generally positive in the three months before 1 January and negative
afterwards, indicating that different atmospheric teleconnection patterns cause similar temperature
anomalies in these regions. The relationship between the equatorial Pacific sea-surface temperatures
(SST) data and temperature anomalies was the weakest. Confidence in these relationships increased
when the extreme HDD anomaly years were examined. These results indicated that the relationships of
climate indices to HDD anomalies exist and that these would be useful in developing and improving
seasonal predictions for business applications.