Limiting global average temperature increase to 3.6°F (2°C) will require a major reduction in emissions.
Projections of future changes in climate are based on scenarios of greenhouse gas emissions and other pollutants from human activities. The primary scenarios used in this assessment are called Representative Concentration Pathways (RCPs)219 and are numbered according to changes in radiative forcing (a measure of the influence that a factor, such as greenhouse gas emissions, has in changing the global balance of incoming and outgoing energy) in 2100 relative to preindustrial conditions: +2.6 (very low), +4.5 (lower), +6.0 (mid-high) and +8.5 (higher) watts per square meter (W/m2). Some scenarios are consistent with increasing dependence on fossil fuels, while others could only be achieved by deliberate actions to reduce emissions (see Section 4.2 in Hayhoe et al. 201724 for more details). The resulting range in forcing scenarios reflects the uncertainty inherent in quantifying human activities and their influence on climate (e.g., Hawkins and Sutton 2009, 201123,220).
Which scenario is more likely? The observed acceleration in carbon emissions over the past 15–20 years has been consistent with the higher future scenarios (such as RCP8.5) considered in this assessment.221,222,223 Since 2014, however, the growth in emission rates of carbon dioxide has begun to slow as economic growth has become less carbon-intensive224,225,226 with the trend in 2016 estimated at near zero.227,228 Preliminary data for 2017, however, indicate growth in carbon emissions once again.228 These latest results highlight how separating systemic change due to decarbonization from short-term variability that is often affected by economic changes remains difficult.
To stabilize the global temperature at any level requires that emission rates decrease eventually to zero. To stabilize global average temperature at or below specific long-term warming targets such as 3.6°F (2°C), or the more ambitious target of 2.7°F (1.5°C), would require substantial reductions in net global carbon emissions relative to present-day values well before 2040, and likely would require net emissions to become zero or possibly negative later in the century. Accounting for emissions of carbon as well as other greenhouse gases and particles that remain in the atmosphere from weeks to centuries, cumulative human-caused carbon emissions since the beginning of the industrial era would likely need to stay below about 800 GtC in order to provide a two-thirds likelihood of preventing 3.6°F (2°C) of warming, implying that approximately only 230 GtC more could be emitted globally in order to meet that target.27 Several recent studies specifically examine remaining emissions commensurate with 3.6°F (2°C) warming. They show estimates of cumulative emissions that are both smaller and larger due to a range of factors and differences in underlying assumptions (e.g., Millar et al. 2017 and correction, Rogelj et al. 2018229,230,231).
If global emissions are consistent with a pathway that lies between the higher RCP8.5 and lower RCP4.5 scenarios, emissions could continue for only about two decades before this cumulative carbon threshold is exceeded. Any further emissions beyond these thresholds would cause global average temperature to overshoot the 2°C warming target. At current emission rates, unless there is a very rapid decarbonization of the world’s energy systems over the next few decades, stabilization at neither target would be remotely possible.27,229,232,233
In addition, the warming and associated climate effects from carbon emissions will persist for decades to millennia.234,235 Climate intervention or geoengineering strategies, such as solar radiation management, are measures that attempt to limit the increase in or reduce global temperature. For many of these proposed strategies, however, the technical feasibilities, costs, risks, co-benefits, and governance challenges remain unproven. It would be necessary to comprehensively assess these strategies before their benefits and risks can be confidently judged.27