The Detection and Attribution
of Climate Change

Humankind is presently involved in a climate change experiment that is global and, in the short term, irreversible. Greenhouse gases (GHGs), sulfate aerosol particles, landcover change due to urbanization and agriculture are all phenomena of human origin affecting our climate system.

The purpose of the climate change detection and attribution activity is to identify variability and trends in the climate system and to ascribe these changes to specific causative factors, whether natural or man-induced. Climate change can be the result of: 1) natural internal variability; 2) natural external factors (e.g., changes in solar input, eruption of volcanoes); or 3) man-made factors (e.g., injection of GHGs and sulfates into the atmosphere). Distinguishing between natural and anthropogenic influences in a quantitative way is an important, but difficult, task. Several combinations of natural and anthropogenic forcing could result in similar patterns of change. In addition, there are uncertainties in observations (forcings and responses) and the model physics in General Circulation Models that simulate the climate system.

The science of detection comprises several key elements: 1) understanding natural change through the paleo record and model simulations; 2) development and implementation of advanced statistical techniques for climate signal identification; 3) analysis of observations and model output to understand the limitations of both data sources (i.e., uncertainty estimates) and to validate model hindcasts of climate system response to natural and anthropogenic forcing; 4) the identification and calculation of appropriate multivariate indices (physical parameters) to further support statements about detection and attribution (e.g., changes in weather extremes, sea ice extent, glacier volumes, sea level rise, etc.).

To capture the full range of natural variability, recourse to the long paleoclimate record is necessary. Through analysis of paleo-proxy records (tree rings, ice cores, corals, etc.), past climate variations in the pre-industrial era can be described and used to provide a context for statements about present and future climate possibilities. The paleo record also informs us about possible extremes (e.g., extended droughts) and abrupt climate shifts, i.e., substantial changes occurring over several years or decades that may not be seen in the relatively short, modern, instrumental record. This information on background variability is essential to the separation of natural and man-made climate change. Figure 1 shows several paleo re-creations of the temperature record. Recent (late 20th century) temperature increases exceed reconstructed temperature rises over the last thousand years.

Figure 2 shows modeled natural and anthropogenic contributions to the total climate signal for surface temperature. GHGs alone do not explain the observed climate trends; a combination of GHGs, sulfate aerosols, and tropospheric ozone reproduce the observed trend with some fidelity.

Figure 2. Global mean near-surface temperature changes (ºC) for observations and simulations: (a) annual mean changes relative to 1880-1920 period with different forcings and individual realizations; (b) as in (a), but for decadal and ensemble averages. In both panels, black denotes observations. Red is GHG alone. Green is the full anthropogenic run forced by GHG, sulfate emissions (both direct and indirect effects), tropospheric and stratospheric ozone. Purple is the same as Green but with the stratospheric effect omitted. Blue is Natural, forced by changes in solar irradiance and volcanic aerosols. (Tett et al., Hadley Centre)

Figures 3-5 show evidence of a global warming signal in the cryospheric and oceanic environments. Arctic sea ice extent has been diminishing over the past 20 years in approximate agreement with model predictions and the change does not appear to be attributable solely to natural fluctuations (Figure 3).

Figure 4. Glacier mass balance for the years 1988-1998. Blue circles indicate positive, and red circles negative, balances. Area of the circles is proportional to the log of the rate of mass loss or gain, with the largest circles indicating rates of 2-3m/yr. (Broecker, Lamont-Doherty Earth Observatory)

An estimate of glacier change over the last decade based on variations of mountain glacier snowlines is depicted in Figure 4. On a global basis, the predominant trend has been retreat, with the exception of Norway where there have been compensating increases in precipitation.

Figure 5 shows the variation in the heat content of the world ocean from the surface to a depth of 3000m. The Atlantic and Pacific Oceans have undergone a net warming from the 1950s and the Indian Ocean has warmed since the mid-1960s. There is a decadal signature to the variability in many records that needs to be explained before more definitive conclusions can be drawn about cause-and-effect.

Our complex climate system (atmosphere-ocean-cryosphere-biosphere) sustains and enriches human life. We need to understand it more fully so that we can avoid surprises and intelligently react to changes.