The IPCC WG1 Second Assessment Report (IPCC, 1996) (hereafter SAR) concluded, "the balance of evidence suggests that there is a discernible human influence on global climate". It noted that the detection and attribution of anthropogenic climate change signals can only be accomplished through a gradual accumulation of evidence. The SAR authors also noted uncertainties in a number of factors, including the magnitude and patterns of internal climate variability, external forcing and climate system response, which prevented them from drawing a stronger conclusion. The results of the research carried out since 1995 on these uncertainties and other aspects of detection and attribution are summarised below.
A longer and more closely scrutinised observational record
Three of the five years (1995, 1996 and 1998) added to the instrumental record since the SAR are the warmest in the instrumental record of global temperatures, consistent with the expectation that increases in greenhouse gases will lead to continued long-term warming. The impact of observational sampling errors has been estimated for the global and hemispheric mean surface temperature record and found to be small relative to the warming observed over the 20th century. Some sources of error and uncertainty in both the Microwave Sounding Unit (MSU) and radiosonde observations have been identified that largely resolve discrepancies between the two data sets. However, current climate models cannot fully account for the observed difference in the trend between the surface and lower-tropospheric temperatures over the last twenty years even when all known external influences are included. New reconstructions of the surface temperature record of the last 1,000 years indicate that the temperature changes over the last 100 years are unlikely to be entirely natural in origin, even taking into account the large uncertainties in palaeo-reconstructions.
New model estimates of internal variability
Since the SAR, more models have been used to estimate the magnitude of internal climate variability. Several of the models used for detection show similar or larger variability than observed on interannual to decadal time-scales, even in the absence of external forcing. The warming over the past 100 years is very unlikely to be due to internal variability alone as estimated by current models. Estimates of variability on the longer time-scales relevant to detection and attribution studies are uncertain. Nonetheless, conclusions on the detection of an anthropogenic signal are insensitive to the model used to estimate internal variability and recent changes cannot be accounted for as pure internal variability even if the amplitude of simulated internal variations is increased by a factor of two or more. In most recent studies, the residual variability that remains in the observations after removal of the estimated anthropogenic signals is consistent with model-simulated variability on the space- and time-scales used for detection and attribution. Note, however, that the power of the consistency test is limited. Detection studies to date have shown that the observed large-scale changes in surface temperature in recent decades are unlikely (bordering on very unlikely) to be entirely the result of internal variability.
New estimates of responses to natural forcing
Fully coupled ocean-atmosphere models have used reconstructions of solar and volcanic forcings over the last one to three centuries to estimate the contribution of natural forcing to climate variability and change. Including their effects produces an increase in variance on all time-scales and brings the low-frequency variability simulated by models closer to that deduced from palaeo-reconstructions. Assessments based on physical principles and model simulations indicate that natural forcing alone is unlikely to explain the increased rate of global warming since the middle of the 20th century or changes in vertical temperature structure. The reasons are that the trend in natural forcing has likely been negative over the last two decades and natural forcing alone is unlikely to account for the observed cooling of the stratosphere. However, there is evidence for a detectable volcanic influence on climate. The available evidence also suggests a solar influence in proxy records of the last few hundred years and also in the instrumental record of the early 20th century. Statistical assessments confirm that natural variability (the combination of internal and naturally forced) is unlikely to explain the warming in the latter half of the 20th century.
Continues on next page
Other reports in this collection