ABSTRACT: The impacts of 2 greenhouse gas emissions scenarios (one from the IPCC and one from Greenpeace International) on the occurrence of extreme daily temperature events are considered at several sites in the UK. For each site, a number of probability distributions were tested for goodness-of-fit to 1961-87 observed daily maximum and minimum temperature data using the Kolmogorov-Smirnov test. The parameters of the best-fitting distributions were then perturbed to take into account climate change, both mean and variability. Probabilities of the occurrence of particular temperature threshold events were calculated for both present and future climates. Changes in climate variability were considered in 3 ways: (1) by assuming the present variance stays the same in the future; (2) by imposing standardised percent changes in variance; and (3) by imposing variance changes derived from the UK Meteorological Office high resolution GCM equilibrium climate change experiment. Results presented for 2 contrasting sites illustrate the importance of including changes in variability in climate change studies. Specific results depend on the site and threshold temperature chosen and on the distribution characteristics. However, for example, at Fortrose the 1961-87 mean maximum temperature in July is below 20*C. With increases in global-mean temperature, the probability of this threshold being exceeded increases, although the rate of increase depends on the variance change being considered. The largest rate of increase in probability occurs with a 20% per *C increase in variance. The approach described here has been used in one component of a climate change scenario generator for the UK developed for the UK Ministry of Agriculture, Fisheries and Food.

KEY WORDS: Extreme events . Climate change . Climate variability

Published in CR Vol. 6, No. 1 (1996) on January 30
ISSN: 0936-577X. Copyright © Inter-Research, Oldendorf/Luhe, 1996