Key points are not available for this paper at this time.
Statistical prediction of July–August Central England Temperature (CET) using January–February sea surface temperature anomalies (SSTA) in the North Atlantic was described by Colman (1997), Int. J. Climatol., 17, 1285–1300. In this paper, the method is extended to examine the predictability of rainfall, surface temperature and pressure in Europe. Using a January–February North Atlantic SSTA pattern as predictor, it is found that mean, minimum and maximum July–August temperatures over much of northwest Europe are predictable using linear regression with correlation skills in the range 0.4–0.7, while July–August rainfall and surface pressures are less predictable, with correlation up to 0.4. July–August temperatures in southeast Canada and the northeast USA are also predictable, with correlation up to 0.3. The predictability of other seasons and timescales is also investigated. For UK temperature, half-month to 2-monthly ranges were considered; predictive correlation skills are best for the period from mid July to the end of August. The highest correlations are from predictions of August temperature in eastern France, where correlation over 50 years exceeds 0.7. For other seasons, long-lead predictability from North Atlantic SSTA was found to be much less than for the summer season. Analysis of composite cases shows an association of predictable warm summers with movement of anomalously warm SSTA across the North Atlantic from the East Coast of the USA to the northwest European coast during spring months. Predictability of cold summers seems to be related to persistence of anomalously cold sea surface temperatures (SST) near, or to the east of, the UK throughout spring. Coherence between the winter SST and the summer temperatures is strongest at 7–8 years. Other studies have revealed North Atlantic ocean variability at this timescale. Copyright © 1999 Controller, Her Majesty's Stationery Office, Norwich, England
Colman et al. (1999) studied this question.