"The mechanism is that emission of other greenhouse gases leads to warming, which in turn leads to emissions of carbon dioxide from soil and from the ocean," Nathan Gillett of Environment Canada told environmentalresearchweb. "This additional carbon dioxide leads to additional warming, thereby amplifying the effect of the non-carbon-dioxide greenhouse gases over and above the warming that would have occurred without this carbon-cycle feedback."

According to Gillett, the results suggest that the warming effects of methane and nitrous oxide are underestimated in emissions management policies. The Kyoto Protocol uses Global Warming Potentials as a measure of the relative effects of greenhouse gases on climate, but the metric does not include enhanced warming due to carbon-cycle feedbacks.

Greenhouse gases other than carbon dioxide are estimated to have caused about 37% of the total greenhouse-gas forcing. And mitigation strategies that include a range of gases are likely to be 30–40% cheaper than those focusing on carbon dioxide alone.

Together with Damon Matthews of the University of Concordia, Gillett assessed the response of a coupled carbon-climate model to a pulse of carbon dioxide, methane or nitrous oxide emissions equal to 50 times their 1990 emission levels. The researchers found that the Mean Global Temperature Change Potential – a measure they prefer to the Global Warming Potential – of methane and nitrous oxide were roughly 20% higher than expected, because of a reduction in the efficiency of carbon sinks as temperatures rise.

Since the size of the enhanced-warming effect is likely to depend on the exact model used, the researchers "do not advocate that metrics from a single carbon-climate model are directly used in climate policy". Instead, writing in ERL, they recommend that carbon-climate feedbacks "should at least be considered when such metrics are calculated and used".

Global Warming Potential is defined as the ratio of radiative forcing of a particular gas integrated over a 100-year period following a pulse emission, to the same measure for carbon dioxide. To compare gas effects on climate more directly, scientists have turned to Global Temperature Change Potential – the ratio of the temperature response per unit mass emission of a given greenhouse gas to that for carbon dioxide.

Now Gillett and Matthews argue that Mean Global Temperature Change Potential – a version of Global Temperature Change Potential integrated over time – has advantages over both Global Warming Potential and Global Temperature Change Potential. For example, the temperature response to a methane pulse peaks after 5–15 years, so the Mean Global Temperature Change Potential gives the gas more weight than the Global Temperature Change Potential, and increases the incentive to reduce emissions of short-lived gases.

Over a century, the Mean Global Temperature Change Potential is numerically similar to the Global Warming Potential, say the researchers, because temperature response is proportional to radiative forcing over a long enough period.

That said, the pair note that metrics such as Global Warming Potential and Global Temperature Change Potential are time-dependent, scenario-dependent and subject to large uncertainties.

The metrics also miss another carbon-cycle effect, this time on carbon dioxide levels; Gillett and Matthews found that the fraction of a pulse emission of carbon dioxide remaining in the atmosphere increased after two to three decades of decline. That means that carbon dioxide emissions today may lead to a larger increase in atmospheric carbon dioxide levels in 2100 than in 2030, they say.

"A study led by Damon Matthews of Concordia University had previously demonstrated that global mean warming is closely proportional to the cumulative mass of carbon dioxide emitted," added Gillett. "We wanted to examine how other greenhouse gases – methane and nitrous oxide – might be accounted for in such a cumulative-emissions framework."

Now the researchers plan to "return to the climate effects of carbon dioxide and other greenhouse gases" and carry out "a careful comparison of the temperature response to cumulative carbon dioxide emissions in a new generation of coupled carbon-climate models with the observed temperature response to carbon dioxide emissions".

The researchers reported their work in ERL.