A significant relationship has been found between interannual variations (correlation = -0.60) of the Northern Hemisphere snow-cover extent and land-surface air temperature in spring since the 1960s. However, the observed increase in temperature during the winter is not reflected in a reduced snow-cover extent. Reduced ice cover on the Northern Hemisphere lakes and rivers, primarily due to earlier onset in spring of ice-free conditions during the 20th century, is consistent with reduced snow cover extent in that season. Sea-ice retreat in the Arctic spring and summer is also consistent with an increase in spring, and to a lesser extent, summer temperatures in the high northern latitudes. Summer temperature increases have been less than in spring in nearby land areas, but Arctic sea-ice extent and especially thickness have markedly decreased. Nevertheless, there is only a small indication of reduced Arctic sea ice during winter when temperatures have also increased. Antarctic sea-ice extent has not decreased since the late 1970s, possibly related to recent indications of little change in Antarctic temperatures over much of the continent in that period . There is now ample evidence to support a major retreat of most mountain glaciers during the last 100 years in response to widespread increases in temperature. There has been especially fast glacial recession in the tropics in recent decades, although tropical temperatures in the free atmosphere near glacier levels have increased little since 1980 according to radiosonde and MSU data.

Global surface temperatures have increased between 0.4 and 0.8°C since the late 19th century, but most of this increase has occurred in two distinct periods, 1910 to 1945 and since 1976. The rate of temperature increase since 1976 has been over 0.15°C/decade. Our confidence in the rate of warming has increased since the SAR due to new analyses including: model simulations using observed SSTs with and without corrections for time-dependent biases, new studies of the effect of urbanisation on global land temperature trends, new evidence for mass ablation of glaciers, continued reductions in snow-cover extent, and a significant reduction in Arctic sea-ice extent in spring and summer, and in thickness. However, there is some disagreement between warming rates in the various land and ocean-based data sets in the 1990s, though all agree on appreciable warming.

New analyses of mean daily maximum and minimum temperatures continue to support a reduction in the diurnal temperature range with minimum temperatures increasing at about twice the rate of maximum temperatures over the second half of the 20th century. Seasonally, the greatest warming since 1976 over land has occurred during the Northern Hemisphere winter and spring, but significant warming has also occurred in the Northern Hemisphere summer. Southern Hemisphere warming has also been strongest during the winter over land, but little difference between the seasons is apparent when both land and oceans are considered. The largest rates of warming continue to be found in the mid- and high latitude continental regions of the Northern Hemisphere.

Analyses of overall temperature trends in the low to mid-troposphere and near the surface since 1958 are in good agreement, with a warming of about 0.1°C per decade. Since the beginning of the satellite record (1979), however, low to mid-troposphere temperatures have warmed in both satellite and weather balloon records at a global rate of only 0.04 and 0.03°C/decade respectively. This is about 0.12°C/decade less than the rate of temperature increase near the surface since 1979. About half of this difference in warming rate is very likely to be due to the combination of differences in spatial coverage and the real physical affects of volcanoes and ENSO (Santer et al., 2000), see also Chapter 12. The remaining difference remains unexplained, but is likely to be real. In the stratosphere, both satellites and weather balloons continue to show substantial cooling. The faster rate of recession of tropical mountain glaciers in the last twenty years than might have been expected from the MSU and radiosonde records remains unexplained, though some glaciers may still be responding to the warming indicated by radiosondes that occurred around 1976 to 1981.

IPCC Homepage