Caldeira Lab

Policy thresholds in mitigation

Katharine L. Ricke, Juan B. Moreno-Cruz, Jacob Schewe, Anders Levermann & Ken Caldeira

Ricke, K. L., J. B. Moreno-Cruz, J. Schewe, A. Levermann, and K. Caldeira, 2015: Policy thresholds in mitigation. Nature Geosci, advance online publication, doi:10.1038/ngeo2607.



Some climate change impacts rise fast with little warming, and then taper off. To avoid diminishing incentives to reduce emissions and inadvertently slipping into a lower-welfare world, mitigation policy needs to be ambitious early on.

Climate impacts are often assumed to increase steadily with global mean temperature for the foreseeable future. However, for a range of systems — such as croplands for staple foods1, coral reefs2 or UNESCO world heritage sites that are affected by sea-level rise3 — impacts are high for relatively modest warming, and then begin to saturate. Indeed, sectorspecific simulations of the impacts of climate change on agriculture, ecosystems, freshwater resources, and other parts of the human environment suggest that the scaling of many environmental impacts with temperature approximates a sigmoidal — S-shaped — pattern1–8 (Fig. 1). For these sectors, the pace of the impact increases drastically at or before 2 °C of warming over a pre-industrial baseline and a large fraction of the impacts is manifested by the point that global temperature reaches this level. Impacts then taper off as warming continues...

Figure 1. Climate change impacts for various sectors. For climate impacts such as coral reef degradation2, river floods1, changes in terrestrial vegetation3, yield losses in staple cropland5, UNESCO world heritage sites in danger from sea-level rise4, population threatened by sea-level rise4 and population with scarce access to freshwater6, impacts rise steeply relative to the total expected impact at low levels of warming and then begin to saturate, usually as a result of either an approach towards the total potential loss or adaptation. The impacts displayed represent the median of an ensemble of simulations; lines are based on a sigmoidal fit. Impacts are normalized by the total sector size (left) and the maximum potential impacts (right) as determined by the sigmoidal fit, ranging from 10 to 100% of the sector for these examples.