Looking across these different reasons for concern, what can we conclude about what change in global average temperature is "dangerous"? A few general caveats apply:

• In spite of many studies on climate change impacts, there is still substantial uncertainty about how effective adaptation will be (and could be) in ameliorating negative effects of climate change and taking advantage of positive effects.
  
• The effect of changes in baseline conditions, such as economic growth and development of new technologies, that could reduce vulnerability has not been adequately considered in most impact studies.
  
• Most impact studies assess the effects of a stable climate, so our understanding of what rates of change may be dangerous is limited.

It does not appear to be possible—or perhaps even appropriate—to combine the different reasons for concern into a unified reason for concern that has meaning and is credible. However, we can review the relationship between impacts and temperature over the 21st century for each reason for concern and draw some preliminary conclusions about what change may be dangerous for each reason for concern. Note that the following findings do not incorporate the costs of limiting climate change to these levels. Also note that there is substantial uncertainty regarding the temperatures mentioned below. These magnitudes of change in global mean temperature should be taken as an approximate indicator of when various categories of impacts might happen; they are not intended to define absolute thresholds.

For simplification, we group different levels of global mean temperature increase into "small," "medium," and "large." "Small" denotes a global mean temperature increase of up to approximately 2°C;⁴ "medium" denotes a global mean temperature increase of approximately 2-3°C; and "large" denotes a global mean temperature increase of more than approximately 3°C. In addition, changes in global mean temperature do not describe all relevant aspects of climate-change impacts, such as rates and patterns of change and changes in precipitation, extreme climate events, or lagged (or latent) effects such as rising sea levels.

19.8.2.1. Unique and Threatened Systems

Tropical glaciers, coral reefs, mangroves, biodiversity "hot spots," and ecotones are examples of unique and threatened entities that are confined to narrow geographical ranges and are very sensitive to climate change. However, their degradation or loss could affect regions outside their range. There is medium confidence that many of these unique and threatened systems will be affected by a small temperature increase. For example, coral reefs will bleach and glaciers will recede; at higher magnitudes of temperature increase, other and more numerous unique and threatened systems would become adversely affected.

19.8.2.2. Distributional Impacts

The impact of climate change will not be evenly distributed among the peoples of the world. There is high confidence that developing countries tend to be more vulnerable to climate change than developed countries, and there is medium confidence that climate change would exacerbate income inequalities between and within countries. There also is medium confidence that a small temperature increase would have net negative impacts on market sectors in many developing countries and net positive impacts on market sectors in many developed countries. However, there is high confidence that with medium to high increases in temperature, net positive impacts would start to decline and eventually turn negative, and negative impacts would be exacerbated. Estimates of distributional effects are uncertain because of aggregation and comparison methods, assumptions about climate variability, adaptation, levels of development, and other factors. In addition, impacts are likely to vary between and within countries. Thus, not all developing or developed countries will necessarily have benefits or damages in unison.

19.8.2.3. Aggregate Impacts

Figure 19-7: Impacts of or risks from climate change, by reason for concern. Each row corresponds to a reason for concern; shades correspond to severity of impact or risk. White means no or virtually neutral impact or risk, light gray means somewhat negative impacts or low risks, and dark gray means more negative impacts or higher risks. Global average temperatures in the 20th century increased by 0.6°C and led to some impacts. Impacts are plotted against increases in global mean temperature after 1990. This figure addresses only how impacts or risks change as thresholds of increase in global mean temperature are crossed, not how impacts or risks change at different rates of change in climate. Temperatures should be taken as approximate indications of impacts, not as absolute thresholds.

With a small temperature increase, there is medium confidence that aggregate market sector impacts would amount to plus or minus a few percent of world GDP; there is low confidence that aggregate nonmarket impacts would be negative. Some studies find a potential for small net positive market impacts under a small to medium temperature increase. However, given the uncertainties about aggregate estimates, the possibility of negative effects cannot be excluded. In addition, most people in the world would be negatively affected by a small to medium temperature increase. Most studies of aggregate impacts find that there are net damages at the global scale beyond a medium temperature increase and that damages increase from there with further temperature increases. The important qualifications raised regarding distributional analysis also apply to aggregate analysis. By its nature, aggregate analysis masks potentially serious equity differences. Estimates of aggregate impacts are controversial because they treat gains for some as cancelling out losses for others and because weights that are used to aggregate over individuals are necessarily subjective.

19.8.2.4. Extreme Climate Effects

The frequency and magnitude of many extreme climate events increase even with a small temperature increase and will become greater at higher temperatures (high confidence). Extreme events include, for example, floods, soil moisture deficits, tropical and other storms, anomalous temperatures, and fires. The impacts of extreme events often are large locally and could strongly affect specific sectors and regions. Increases in extreme events can cause critical design or natural thresholds to be exceeded, beyond which the magnitude of impacts increases rapidly (high confidence).

19.8.2.5. Large-Scale Singularities

Large-scale singularities in the response of the climate system to external forcing, such as shutdown of the North Atlantic THC or collapse of the WAIS, have occurred in the past as a result of complex forcings. Similar events in the future could have substantial impacts on natural and socioeconomic systems, but the implications have not been well studied. Determining the timing and probability of occurrence of large-scale singularities is difficult because these events are triggered by complex interactions between components of the climate system. The actual impact could lag the climate change cause (involving the magnitude and the rate of climate change) by decades to millenia. There is low to medium confidence that rapid and large temperature increases would exceed thresholds that would lead to large-scale singularities in the climate system.

Figure 19-7 sums up the reasons for concern regarding impacts relative to change in temperature. Each row corresponds to a reason for concern, and the shades correspond to the severity of impact or risk. White means no or virtually neutral impact or risk, light gray means somewhat negative impacts or low risks, and dark gray means more negative impacts or higher risks. The period 1850-1990 warmed by 0.6°C and led to some impacts. Unique and threatened systems were affected, and the magnitude and frequency of some extreme events have changed. Future impacts are plotted against increases in global mean temperature after 1990.

Adverse impacts are estimated to occur in three reasons for concern even at a small increase in temperature: unique and threatened systems, extreme weather events, and distributional impacts. For the other two reasons for concern—adverse impacts and large-scale discontinuities—adverse impacts begin at the medium level of temperature increase for the former and a large temperature increase for the latter.

IPCC Homepage