top

top

top

The NAS concludes that "... the warming trend in global mean surface temperature observations in the past 20 years is undoubtedly real and is substantially greater than the average rate of warming during the twentieth century. The disparity between surface and upper air trends in no way invalidates the conclusions that surface temperatures have been rising." The reason for this discrepancy was examined in a paper published in 2000 in the journal Science. The authors found that a combination of three factors were likely responsible. First, there is a degree of error in both the surface and the satellite data. Second, the areas covered by the two data sets do not overlap completely. Third, they showed how a more sophisticated climate model shows that such discrepancies are the expected result of the anthropogenic greenhouse effect.

Nevertheless, the unexpected anomaly poses a problem for the climate modellers, reducing the confidence in their predictions of future surface temperatures. As the IPCC states in their summary report (Page 9), "Such models cannot yet simulate all aspects of climate (e.g., they still cannot account fully for the observed trend in the surface-troposphere temperature difference since 1979)". See below for more information on the accuracy of climate models.

There are a variety of climate proxies that can be used to estimate past temperatures. These include tree rings, ice cores, corals and historical records. For the past 100 years or so, the instrument record provided by weather stations has overlapped with the record provided by climate proxies. This enables the climate proxies to be calibrated, and also allows scientists to check their accuracy. Although it has been argued that tree ring data could be susceptible to bias, for the periods where sufficient non-tree data are available to make a comparison,  the record reconstructed solely from the tree ring data is essentially the same as that constructed from non-tree ring data (see this link).

There are a number of different techniques used to recreate the paleoclimatic record, and the results of these different approaches are detailed in the links above. Although they differ in the specifics, all these reconstructions agree that the global mean temperatures have declined gradually over the past 1000 years. The trend in the record may partly be due to changes in solar activity, but is at least partly explained by deforestation, according to research from the Livermore National Laboratory reported in this news article.

The detailed causes of the recent warming trend have been investigated by the UK Meteorological Office using climate models, and are presented here. They found that about half of the warming is caused by solar variability but that, in the second half of the century, these effects have been countered by sulphate emissions from volcanoes (which act to cool the earth). The overall effect of all these natural causes (sun and volcanoes combined) has been quite small. Similarly, two recent studies of ocean temperatures have found that the observed increase is best explained by the effect of greenhouse gases.

According the the US EPA's page on the subject, "During the 20th century, precipitation increased by about 0.5-1 percent per decade over most middle and high latitudes of the Northern Hemisphere's continents. Rainfall over the sub-tropical Northern hemisphere declined about 0.3 percent per decade, while no significant change occurred over the tropics".

A basic overview of published research on mountain glaciers is provided by UCS: "Since 1850 the glaciers of the European Alps have lost about 30 to 40% of their surface area and about half of their volume (Haeberli and Beniston, 1998). Similarly, glaciers in the New Zealand Southern Alps have lost 25% of their area over the last 100 years (Chinn, 1996), and glaciers in several regions of central Asia have been retreating since the 1950s (Fitzharris, 1996; Meier, 1998). For three glaciers in the US Pacific Northwest, the seven-year average rate of ice loss was higher for the period since 1989 than for any other period studied (Hodge et al., 1998). Glaciers on Mt. Kenya and Kilimanjaro have lost over 60% of their area in the last century (Hastenrath, 1991; Hastenrath and Greischar, 1997), and accelerated retreat has been reported for the Peruvian Andes (Mosley-Thompson, 1997)". Here is a news report on the tropical and sub-tropical glaciers of Africa and South America.

A more sensitive measure of changes in Arctic ice is its thickness. The Polar Science Center concludes that "the mean ice draft at the end of the melt season has decreased by about 1.3 m in most of the deep water portion of the Arctic Ocean, from 3.1 m in 1958–1976 to 1.8 m in the 1990s ... Preliminary evidence is that the ice cover has continued to become thinner in some regions during the 1990s". These data are corroborated by an analysis of similar data from British submarines, reported in this news article. The cause of these changes is not known for certain. It is probably partly due to changes in surface temperature, but also partly due to changes in the North Atlantic Oscillation (see below).

"In recent years ... the NAO has been in a period of unprecedented activity. Its flips are bigger, and when the index is averaged over the winter it has usually been positive, often strongly so ... The last time the NAO was routinely positive was between 1900 and 1930. That period also coincided with a spell of planetary warming ... Rodwell <of the UK Meteorological Office> showed that a large, positive winter index was tied to changes in sea surface temperatures off the US coast the previous September ... Tim Osborn of the University of East Anglia recently ran a series of global climate models to simulate 1400 years of "natural" weather ... as soon as Osborn introduced global warming into the simulation, the index became strongly positive ... 'In effect, the only way you can get anything like the recent trend in the NAO is through global warming,' says Hurrell <of the National Center for Atmospheric Research in Boulder, Colorado>"

Nevertheless, there is still a great deal of uncertainty in predicting future climate. According to the IPCC report, Current estimates for the next 100 years range from 1.4–5.8°C, depending on the model used and the carbon production and carbon cycle model used. Although there is plenty of room for error, it is important to realise that the effect of any errors is unknown. That is, it could be that warming is less severe than predicted. However, it is just as likely that it will be more severe. A good article on the problems with models is available from the Geophysical Fluid Dynamics modelling at Princeton (although it is a little out of date since it was written in 1998).

• Firstly, it's not enough. Data from Mauna Loa show that carbon dioxide levels have steadily increased throughout the century. Studies have shown that the ability of trees to soak up excess carbon is limited by the availability of other nutrients, as discussed in this article from National Geographic.
• Secondly, there's more to greenhouse gas than CO2, as this link shows. Methane, a very important greenhouse gas, has doubled in concentration since the start of the industrial age.
• Secondly, we are actively reducing biomass by cutting down forests and draining marshland – deforestation in the 1980s was responsible for around 1.6 to 1.8 GtC emissions, and it seems likely that the 1990s will have been even worse, according a report from the Edinburgh Centre for Carbon Management.
• Thirdly, although forests contain more CO₂ than grassland, they are darker. This means that they absorb more heat. Scientists from the USA's Lawrence Livermore Laboratories report that deforestation likely played an important part in the global cooling that took part over the last millennium.
• Finally, and most importantly, climate change is likely to have important effects on the carbon cycle. For example, as the oceans warm they will release CO₂, and as the permafrost thaws it will start to decay, releasing large amounts of CO₂. Although the exact effects of these changes are hard to calculate, the current estimate from the HADCM3 model is that the net effect will be to increase global by a further 3ºC over the next 100 years – making the total increase up to 8ºC.