Climate Change Science


Recent observations of climate

The year since Kyoto has witnessed the largest El Nino, the warming of the eastern tropical Pacific, on record. The effects of this on global communities, from forest fires in Indonesia to torrential rains in Peru and East Africa, have been well publicised. We cannot say yet whether global warming has led to more frequent or larger El Ninos. The figure alongside shows an underlying trend of rising temperatures, but also shows that large El Ninos were seen at the end of the last century. The global effect of El Nino, however, demonstrates how vulnerable society is to changes in climate.

1997 was the warmest year in the global instrumental record going back to 1860. Due in part, but certainly not wholly, to the large El Nino, global surface temperatures have been at a consistently high level in 1998. Each of the first eight months of the year has been the warmest, or equal warmest, such month on record. Despite the fact that El Nino is now moving into a colder La Nino phase, which will have a corresponding cooling effect on global climate, the annual temperature for 1998 is almost certain to be warmer than in 1997, and hence the warmest on record. However, due to natural climate variability, temperatures are not expected to rise successively each year, and it seems likely that 1999 will be cooler than 1998.

[figure 1]

Changes in sea-surface temperature 1871-1998, relative to the 1961-90 average, for the eastern tropical Pacific off Peru.

[figure 2]

Changes in global anual mean surface temperature, relative to that at the end of the last century, are shown by blue bars, with a smoothed curve in red. The value for 1998, shown in green, includes observations up to the end of September.


Predictions from the new Hadley Centre climate model

A climate model aims to represent all the main components of the climate system which can affect change in the future. Up to now, coupling together models of two main components of the climate system, the atmosphere and the ocean, has caused the climate model's own simulated climate to drift away from reality. To prevent this, corrections --so-called flux-adjustments-- were introduced, and these raised issues about the confidence we had in using such models for predictions. In the third Hadley Centre coupled climate model, the ocean is represented at a much higher resolution than previously, 1.25o latitude x 1.25o longitude, which gives a greatly improved representation of ocean currents such as the Gulf Stream. This, together with improvements in the representation of processes in the atmosphere and on land, has allowed the model's climate to remain stable without the need for flux-adjustments.

[figure 3]

Changes in land, sea and global mean surface temperature from the new Hadley Centre model driven with changes in greenhouse gas concentrations (without sulphate aerosols): observed to 1995 and IPCC 'business-as-usual' to 2100. The black line shows the low climate drift of the uperturbed model.

The new model has been used to make simulations and predictions of climate over the period 1860-2100, and some of the results are shown on this page. When driven by observed increases in greenhouse gases to the present, and IPCC-projected increases in the future, the model predicts that a rise in global temperature of about 1oC should have occurred already, and a further rise of about 3oC is to be expected by 2100; rises over land (where the main impacts are expected) will be almost twice as fast as those over sea.

The regional patterns of change are also shown on this page; high-latitude winters will warm faster due to feedback from the melting of sea-ice, and there will be some ocean areas where rises are quite limited. Rainfall changes will be most marked in tropical regions. These patterns, and the global mean rises, are not very different from those previously predicted with flux-adjusted models. Using the new model we see relatively large decreases in rainfall over Amazonia and parts of Africa, but this may be due to changes in the model other than omitting flux-adjustment. Predictions from the new model have been used to investigate impacts on a number of socio-economic areas, and these are described in later sections of this report.

The new climate model has, very recently, also been run including changes in sulphate aerosols from revised projections of SO2 emissions about a half of those used in IPCC 1995. Although the direct cooling effect of sulphate aerosols is much reduced, we also now include cooling from the indirect effect via changes in cloud brightness. The diagram alongside shows how the simulation of change to date is in broad agreement with the observed temperature rise. Other factors, such as changes in solar output and volcanic activity, will be included in future simulations, and are expected to give better agreement in mid-20th century. However, natural variability means that simulated and observed temperatures will not always agree. Because total global sulphur emissions are not expected to change substantially in future, the warming over the next 100 years is much the same as without sulphate aerosols about 3oC (see figure below). There are some differences in regional detail, however.


[figure 4]

The patterns of change in northern winter temperature (top) and precipitation (bottom) for the 2050's compared to the present day, when the climate model is driven with IPCC 'business-as-usual' changes in greenhouse gases only.

[figure 5]

The global mean surface temperature change when the model includes the effect of greenhouse gases (red), abd also including the direct and indirect effects of sulphate aerosols (blue). Observations are shown in green.


Uncertainty in climate change predictions

Recent work at the Hadley Centre has shown that, when driven by all the observed changes in climate forcing factors (natural and human-made), the climate model does reasonably well at simulating changes in climate over the last 140 years, taking into account the vagaries of natural internal climate variability. It is also able to mimic changes in the more distant past, for example during the warm period some 6,000 years ago. This limited model validation gives increased confidence in predictions of future climate change.


Rapid changes in ocean circulation

There is considerable interest in how stable ocean thermohaline circulation (the 'conveyor belt') will remain in the face of rises in greenhouse gas concentrations. The possibility has been raised of a southwards movement in the Gulf Stream leaving Europe exposed to a colder climate. The improved representation of the ocean currents in the new Hadley Centre model has allowed us to investigate this with more realism. We find that, even when CO2 is increased at an unrealistically rapid rate of 2% per year, and then stabilised at four times the present concentration, the strength of the Atlantic ocean circulation decreases by about 25% (see diagram alongside). This slowdown does decrease the amount of heat transported into north-west Europe, but this is more than offset by direct greenhouse warming, so that temperatures in Europe still rise. However, changes in ocean circulation do alter warming patterns in the North Atlantic, and hence will contribute to changes in European weather, including, for example, storminess. Some other climate models show a more drastic reduction in ocean circulation, and the reason for this range of responses is not fuly understood.


Identifying the causes of recent climate change

The change seen over the past 140 years, shown earlier, amounts to some 0.6 C. Is this due to human activities such as the burning of fossil fuels? There are many factors that influence climate, and distinguishing the human-made signal from background natural climate variability is a challenge. To do this we use advanced statistical techniques which look at changes in patterns of temperature, both at the surface of the earth and through the depth of the atmosphere, giving greater importance to those areas where natural variability is low and vice versa. This statistical analysis indicates that, over the past 50 years, human-made greenhouse gases have contributed substantially to global warming. There is still considerable uncertainty in this estimate, but this recent study supports and strengthens the IPCC 95 statement that the balance of evidence suggests a discernible climate change due to human activities. Although it is still not possible to unambiguously assign recent change to human activity, we are working towards a more robust assessment through better model simulations, better observations, and better estimates, from models and observations, of the natural variability of climate.


Making better estimates of climate change

The reason why climate predictions are so uncertain is that, once climate change begins, consequential changes will feed back, either positively or negatively, on the original warming. These feedbacks are poorly understood. Although we believe the most important feedbacks in the atmosphere, ocean, land surface and sea-ice are already included (albeit imperfectly) in the climate model, there are others which should be taken into account, and this can only be done by including all components of the climate system in a fully interactive model. Most recently we have included a sulphur cycle which creates sulphate aerosol from natural and industrial SO2 emissions. The next step is to add the carbon cycle and chemistry; climate change has the potential to disturb the natural carbon cycle in such a way as to alter atmospheric concentration of CO2, and to disturb the chemistry of the atmosphere so as to alter concentrations of other greenhouse gases such as ozone and methane. Sub-models currently under development will be incorporated (over the next few years) in the main model to form the first Hadley Centre Earth Systems Model. Ultimately we plan to include feedbacks from socio-economic sectors such as agriculture and energy use.


[figure 6]

The strength of the Atlantic circulation is seen to decrease substantially as greenhouse gas concentrations rise at 2% per year to four times the pre-industrial value (green). The black line shows the unperturbed circulation, and the red line the change when greenhouse gases follow the IPCC 'businesss-as-usual' emissions scenario.


Contributors: Chris Folland, David Parker, Briony Horton, John Mitchell, Tim Johns, Christine Coghlan, Anne Keen, Nick Rayner, David Roberts, Andy Jones, Paul Jacobs, Simon Tett, and Geoff Jenkins at the Hadley Centre for Climate Prediction and Research, The Met Office; Keith Briffa, Phil Jones, Mike Hulme and David Viner at the Climatic Research Unit, University of East Anglia