2.7
Changes in extremes in European records since the 18th century
Philip D. Jones, Climatic Research Unit, Norwich, Norfolk, United Kingdom
Instrumental records are generally about 100 years longer in Europe than in other parts of the world. Although a few have been published in monthly form and been used for many decades, daily records have generally been limited to the 20th century. The last 10-15 years, however, has seen a resurgence of interest in early instrumental data and many long records have had their long-term homogeneity reassessed. More recently, interest in changes in the frequency of extremes, has enabled eight of the long daily European sources to be digitized for temperature and pressure. Records are available from Uppsala (1739), Stockholm (1756), Central England (1772), Brussels (1833), St. Petersburg (1743), Milan (1763), Padua (1774) and Cadiz (1786). Homogenizing the records is a laborious exercise, involving many stages: (1) conversion to modern units, (2) correction for all known problems in the metadata due to site, screen and observation times and (3) adjustments for unrecorded problems clearly evident when comparisons with neighbours takes place.
The talk will address changes in the frequency of extremes, concentrating on the definition of extremes and the potential problems fo some measures because of uncorrected problems/issues with the early data. For example, day-to-day temperature variability is particularly sensitive to changes in observation times and correction appears impossible for some sites before about 1870. Despite this counts of extreme days above thresholds appear reliable back to the earliest years. Most of the variability in the daily series is apparent on the weather timescale (5-20 days) and this appears less susceptible to observation protocols. Variability on this timescale appears to strengthen during warmer periods and can be related to changes in the phase of the North Atlantic Oscillation.
Session 2, Observed Climate Variability and Change: Proxy Records (Parallel with Session 1)
Monday, 15 January 2001, 10:30 AM-2:45 PM
Previous paper Next paper