We conclude that the 2◦ C global warming “guardrail”, affirmed in the Copenhagen Accord (2009), does not provide safety, as such warming would likely yield sea level rise of several meters along with numerous other severely disruptive consequences for human society and ecosystems. The Eemian, less than 2 ◦ C warmer than pre-industrial Earth, itself provides a clear indication of the danger, even though the orbital drive for Eemian warming differed from today’s human-made climate forcing. Ongoing changes in the Southern Ocean, while global warming is less than 1◦ C, provide a strong warning, as observed changes tend to confirm the mechanisms amplifying change. Predicted effects, such as cooling of the surface ocean around Antarctica, are occurring even faster than modeled.

Our finding of global cooling from ice melt calls into question whether global temperature is the most fundamental metric for global climate in the 21st century. The first order requirement to stabilize climate is to remove Earth’s energy imbalance, which is now about +0.6 Wm^−2 , more energy coming in than going out. If other forcings are unchanged, removing this imbalance requires reducing atmospheric CO_2 from ∼400 to ∼350ppm (Hansen et al., 2008, 2013a).

The message that the climate science delivers to policymakers, instead of defining a safe “guardrail”, is that fossil fuel CO_2 emissions must be reduced as rapidly as practical. Hansen et al. (2013a) conclude that this implies a need for a rising carbon fee or tax, an approach that has the potential to be near-global, as opposed to national caps or goals for emission reductions. Although a carbon fee is the sine qua non for phasing out emissions, the urgency of slowing emissions also implies other needs including widespread technical cooperation in clean energy technologies (Hansen et al., 2013a).

The task of achieving a reduction of atmospheric CO_2 is formidable, but not impossible. Rapid transition to abundant affordable carbon-free electricity is the core requirement, as that would also permit production of net-zero-carbon liquid fuels from electricity. The rate at which CO_2 emissions must be reduced is about 6%yr^−1 to reach 350ppm atmospheric CO_2 by about 2100, under the assumption that improved agricultural and forestry practices could sequester 100 GtC (Hansen et al., 2013a). The amount of CO_2 fossil fuel emissions taken up by the ocean, soil and biosphere has continued to increase (Fig. S23), thus providing hope that it may be possible to sequester more than 100GtC. Improved understanding of the carbon cycle and non-CO_2 forcings are needed, but it is clear that the essential requirement is to begin to phase down fossil fuel CO_2 emissions rapidly. It is also clear that continued high emissions are likely to lock-in continued global energy imbalance, ocean warming, ice sheet disintegration, and large sea level rise, which young people and future generations would not be able to avoid. Given the inertia of the climate and energy systems, and the grave threat posed by continued high emissions, the matter is urgent and calls for emergency cooperation among nations.