A simple carbon cycle representation for economic and policy analyses

Integrated Assessment Models (IAMs) that couple the climate system and the economy require a representation of ocean CO2 uptake to translate human-produced emissions to atmospheric concentrations and in turn to climate change. The simple linear carbon cycle representations in most IAMs are not however physical at long timescales, since ocean carbonate chemistry makes CO2 uptake highly nonlinear. No linearized representation can capture the ocean’s dual-mode behavior, with initial rapid uptake and then slow equilibration over ∽10,000 years. In a business-as-usual scenario followed by cessation of emissions, the carbon cycle in the 2007 version of the most widely used IAM, DICE (Dynamic Integrated model of Climate and the Economy), produces errors of ∽2∘C by the year 2300 and ∽6∘C by the year 3500. We suggest here a simple alternative representation that captures the relevant physics and show that it reproduces carbon uptake in several more complex models to within the inter-model spread. The scheme involves little additional complexity over the DICE model, making it a useful tool for economic and policy analyses.