Inhomogeneities and trends in the surface ozone record (1988-1996) at Jungfraujoch in the Swiss Alps

Citation
P. Zanis et al., Inhomogeneities and trends in the surface ozone record (1988-1996) at Jungfraujoch in the Swiss Alps, ATMOS ENVIR, 33(23), 1999, pp. 3777-3786
Citations number
22
Categorie Soggetti
Environment/Ecology,"Earth Sciences
Journal title
ATMOSPHERIC ENVIRONMENT
ISSN journal
13522310 → ACNP
Volume
33
Issue
23
Year of publication
1999
Pages
3777 - 3786
Database
ISI
SICI code
1352-2310(199910)33:23<3777:IATITS>2.0.ZU;2-O
Abstract
Ozone data at the Jungfraujoch Observatory (3580 m asl) in the Swiss Alps h ave been recorded continuously since 1986 in the framework of the Swiss Nat ional Air Pollution Monitoring Network (NABEL), operated by the Swiss Feder al Laboratories for Materials Testing and Research (EMPA). The long-term oz one record (1988-1996) was examined with regard to potential inhomogeneitie s using Zugspitze (2960 m asl), a neighbouring alpine site in Germany, as a reference site. The Alexandersson test (for shift) and the Easterling-Pete rson test (for trend and shift) were applied to a difference times series, calculated from monthly ozone means at Jungfraujoch and Zugspitze. The Alex andersson test revealed a significant shift discontinuity in November 1989 and the Easterling-Peterson test a trend discontinuity in September 1991. I t is assumed that the shift discontinuity may be related to an instrument c hange, but there is not firm evidence due to lack of detailed information f rom the instrumental history in the earlier years of the ozone record at Ju ngfraujoch. Monthly ozone means at Davos (1640 m asl) in eastern Switzerlan d indicate that part of the trend discontinuity in September 1991 may be of natural origin. Adjustment of the monthly ozone means at Jungfraujoch for the observed shift discontinuity removed the inhomogeneity in November 1989 , and reduced the trend discontinuity by a factor of 4.8. A trend analysis on both adjusted (homogenised) and unadjusted monthly ozone means (1988-199 6) at Jungfraujoch showed no statistically significant linear trend. The bo undaries encompassing the true linear trend are indicated by two linear reg ression calculations on both adjusted and unadjusted data. Trends in the co ld season are positive albeit not significant. The only statistically signi ficant trends are found in October, December, and February for the adjusted data. (C) 1999 Elsevier Science Ltd. All rights reserved.