The factors causing apparent trends or discontinuities in long-term climato
logical observations are studied using urban records as examples. In additi
on to instrumental and methodological inhomogeneities urban records also in
clude various progressive changes in the surroundings of the observation st
ation. The effects of site changes on a record are also found alongside urb
anisation effects. The apparent urban temperature trend is shown to be arou
nd 0.1 degrees C per decade. Precipitation is generally greater at urban th
an at adjacent rural sites, but the differences are mostly due to the openn
ess of neighbouring sites. Urbanisation tends to lower humidity and wind sp
eed, but any effect on cloudiness are contradictory. (C) 1999 Elsevier Scie
nce Ltd. All rights reserved.