How Accurate are Future Projections of Climate Change? A Look at Past IPCC Reports Provides Some Answers

September 13, 2013
James J. McCarthy
Professor Biological Oceanography, Harvard University

The Intergovernmental Panel on Climate Change (IPCC) has established a strong track record for projecting the future consequences of climate change. Though the process for producing the report can seem opaque from the outside, we can put great confidence in the body’s findings.

ipcc-blog-series-imageThis post is part of a series on

The Intergovernmental Panel on Climate Change (IPCC)

Subscribe to the series RSS feed.

An inside view of the IPCC

As science historian Spencer Weart pointed out in a previous post, it was established jointly under the auspices of the World Meteorological Organization and the United Nations Environmental Program in 1988 to assess “the scientific information that is related to the various components of the climate change issue…needed to enable environmental and socio-economic consequences of climate change to be evaluated”.

I was a co-lead author of the final chapter in the first assessment science report (1990), a reviewer during the second assessment (1995) and co-chair of Working Group II, which addressed climate impacts, adaptation, and vulnerability for the third assessment (2001).

As is often pointed out, the IPCC does not conduct original research. Its mission is to assemble experts who are asked to use their best scientific judgment to assess the status of scientific understanding across multiple facets of climate change science—including social science, as this understanding relates to climate impacts and vulnerabilities—and to consider the costs and benefits of climate mitigation and adaptation strategies. These experts then write the reports.

Who writes the IPCC reports?

Who are these experts? They are nominated by participating nations (more than a hundred nations participate) and relevant UN bodies. In the case of the assessment report I headed, we had over 1000 nominations for authors with full resumes (I read them all), and from these we carefully selected authors (~180) who had published relevant papers in the scientific literature. We later drew from this list for chapter review editors (~30), and used many others (>400) as reviewers.

Assessing implies much more than simply reporting what is in the published literature. A new, interesting, and provocative finding might, for example, be given less weight than a somewhat older more established finding. This in part explains the range of confidence that will be given to a particular summary statement about the state of scientific understanding relating to some aspect of climate change or climate impacts.

How have climate models from previous reports performed?

Models are an important component of an assessment since they can be used to make projections based upon recent trends and a plausible understanding of what is driving the trends. Some critics of the IPCC approach say that such model projections are not to be trusted and thus should not be used in the policy making process. But it is illustrative to examine some of the projections made in the 1990 assessment. By 20th century standards the 1980s had been an exceptionally warm decade. If the scientific consensus that this warming was due to the accumulation of greenhouse gases in the atmosphere, and that human activities (e.g.. combustion of fossil fuels and land use practices) were responsible for the rising atmospheric concentrations of greenhouse gases, then a business as usual scenario could result in several discernable changes in climate. Among these were the following: 1) surface warming that would be greater on land than on the ocean; 2) the highest rates of warming would occur at high latitudes in the Northern Hemisphere; 3) an increase in anomalous high- and decrease in anomalous low- temperature events; 4) an increase in precipitation at high latitude in winter, and 5) intense precipitation events increasing in many regions. In 1990 none of these were yet apparent with statistical significance, and yet today, two decades later, data for each are now convincing and these trends are accepted within the climate science community as further indications of human-caused climate change.

Changes in temperature, sea level, and northern hemisphere snow cover. Figure courtesy of the IPCC.

Changes in temperature, sea level, and northern hemisphere snow cover. Figure courtesy of the IPCC (AR4).

The rate of change in some aspects of climate has exceeded expectations of what could occur in a decade or several decades into the future. Significant among these are the loss of summer sea ice in the central region of the Arctic Ocean and the loss of ice from Greenland. The latter has played into underestimates of sea level rise. In part this is because recently observed rates of Greenland melt are higher than expected, and in part because unlike the thermal expansion component of sea level rise (a warmer ocean has a larger ocean volume) the relationship between global warming and the rate of melt for Greenland’s glacial ice (2 – 3 km thick) is not as easily quantified. Thus it is more difficult to project the future contribution of Greenland (and Antarctic) ice loss in scenarios for sea level rise. At the time of the third IPCC assessment (2001) it was thought unlikely that Greenland could by this time be losing ice as rapidly as it is.

There remains much to be learned about how global climate and regional manifestations thereof will play out as we move further and further into uncharted territory with greenhouse gases in the atmosphere accumulating at higher rates and reaching higher levels than at any time in the last million or more years. How are internal cycles in the ocean, such as the El Niño – Southern Oscillation, or the North Atlantic Oscillation being affected by warming-induced changes in atmosphere-ocean interactions? Over the past decade precision in measurements of ocean heat content have increased substantially. It is now clear that the oceans are holding 90% of the heat that is accumulating in the climate system as a result of anthropogenic activities. Moreover, ocean heat content has shown an unequivocal upward trend over the last decade, even while the rate of increase in atmospheric temperature has not increased as rapidly as it did during the prior decade. A slowing in the rate of increase in atmospheric temperature may also be partly due to unusual solar activity at this time. Relative to the past three decades of solar observations, at this point in the nominal eleven-year solar cycle the sun would be expected to be more active. Instead we are seeing what some experts are suggesting what might be the smallest cycle in a century.

Looking ahead, biggest uncertainty is still how much carbon we will burn

Climate change projections made by the IPCC will always be limited by an incomplete understanding of Earth’s climate, by unanticipated events like major volcanoes, and by surprises like an unusually weak solar maximum. But the historically cautious nature of the IPCC assessment process has led to a remarkably robust set of projections.

Still though, one of the largest uncertainties that climate scientists must contend with is that of future societal decisions regarding greenhouse gas emissions. Were the peoples and nations of the world to embark upon an aggressive campaign to dramatically reduce these emissions, two things are certain: future climate projections would be more reliable, and we would have much greater confidence in projections that the most costly impacts of future climate change, which might otherwise be expected, could be avoided.