Global Temperature Plot

This shows the change in global average temperature from 1750 to 2300 (in expert level, north/south land/ocean temperatures are also shown). Historical measured and proxy temperatures are shown for comparison.
  • Java source code

    Curves

    (All from Climate Module)

    Model Prediction

  • Brown: Global average
  • Red, Yellow: North, South Land
  • Blue, Cyan: North, South Ocean

    Historical measurements

  • Dk Green: Surface thermometer data (1860-2001, Jones et al)
  • Grey: Proxy data from tree rings, corals, sediments (1750-1990, Mann et al)
  • Light Green: Trend of above (7 year moving average)

    Units & Baseline

    Degrees Celcius, relative to the calculated temperature in the baseline year (adjustable parameter). This is actually averaged over five years (chosen year +/- 2) to smooth odd peaks.

    It is recommended to set the baseline to 1765, for understanding the effect of science model parameters, and to 1990 for comparison with IPCC predictions.Note this only affects the plot and the stabilise temperature control (see below), not the climate model.

    Controls and Options

  • Baseline year -see above:

    Climate Model parameters

    See Climate module and discussion below.
  • Climate Sensitivity, Eddy Diffusivity,
  • (expert level only) Land-Ocean Temperature Ratio, Polar Sink Temperature Ratio, North-South Conductivity, Land-Ocean Conductivity, Sea-Ice parameter, Mixed Layer Depth, Temperature-Upwelling Feedback option.
  • GCM-fit menu, adjusts above parameters to match GCM results

    Stabilise Temperature

    (Available if "stabilise temperature" is selected from the mitigation menu)
    See Mitigation Module
  • 4-pointed brown arrow sets stabilisation level (relative to baseline -see above) and year.
  • (expert level only) Stabilise by iteration option, damping option.

    Discussion

  • See also Ocean temperature plot

    Fitting to GCM predictions

    The credibility of a simple climate model depends on fitting the parameters to predictions from more sophisticated global climate models (GCMs). In IPCC-TAR seven GCMs were used for this purpose. If you select the "expert" complexity level, you can see all the parameters changing together, as you choose different GCMs from the menu.

    This parameter fitting was carried out by Sarah Raper et al, as described in IPCC-TAR WG1 Appx9.1. See also Correspondence to IPCC calculations.

    Climate Model parameters

    You can also adjust these parameters individually, to understand the effect of each one (see also Climate Module -How it works)

    The most important uncertainty is the "climate sensitivity". This takes into account the fast, physical feedback processes in the climate system, such as changes in water vapour and clouds. This parameter is defined as the global average temperature rise caused by a doubling of CO2 concentration.

    The land-ocean temperature ratio refers to the equilibrium change and is used to calculate the internal parameters which determine heat fluxes to space. The land-ocean conductivity and north-south conductivity affect the fluxes between the four surface boxes, their effect is fairly intuitive.

    The eddy diffusivity affects the flux of heat from the surface to the deep ocean. Increasing this will cool the surface slightly, but warm the deep ocean, and consequently increase the sea-level rise due to thermal expansion. See also

  • Ocean temperature plot
  • Sealevel plot.

    The cold water which sinks around Greenland and Antarctica also has a major influence on the deep ocean temperature. Since most of this cold salty water is formed near freezing ice, its temperature is always close to zero. In this model the warming of this water, compared to the average, is determined by a prescribed "polar sink temperature ratio". The rate of this sinking may also decrease as the surface temperature rises (temperature-upwelling feedback), due to stratification of the water column which weakens the thermohaline circulation. Sea-ice also introduces a difference between air and water temperatures in high latitudes.

    Matching to historical data

    You can see how well the climate model predictions match the historical data, as you adjust the balance of radiative forcing from greenhouse gases, sulphate aerosols, and solar variability, which have different patterns over time (see Radiative Forcing Plot).

    You can also observe that some forcings are distributed unevenly between the four surface boxes (expert level). For example, most of the sulphate is in the northen land box.

    Timescales of response

    The surface temperature follows changes in radiative forcing fairly rapidly, with some delay due to exchange of heat with the deeper ocean. Note that short spikes in forcing (e.g. from volcanos) affect the land temperatures more than the ocean. The deep ocean, and hence sea-level rise, responds much more slowly. See also Timescales of response.

    Stabilise temperature

    If you choose "stabilise temperature" from the mitigation menu, a control appears which sets a target stabilisation year and level, relative to the baseline temperature (see above). For example, the European Union proposed a maximum safe limit of 2C above the preindustrial level.

    The emissions (of CO2, or all gases, depending on the "other gas" menu) are adjusted to attempt to reach this target level. The default method works by guessing the CO2 stabilisation level and correcting this guess iteratively (i.e. running the model several times). An alternative method (expert level) works by "fuzzy-control" correcting the emissions in each timestep according to the deviation from the target curve -this tends to produce oscillations. For further explanaton see the Mitigation module.

    Regional climate change

    Beware that local temperature changes can be much greater than the global averages shown here. See Regional climate map.