Studies of the economic impact of carbon emissions usually approximate uncertainty in climate via the long-term equilibrium response. This representation of climate uncertainty is potentially misleading, argued the new study, since focusing on so-called equilibrium climate sensitivity can result in estimated costs being entirely dominated by low-probability outcomes that are unlikely to emerge until the distant future, making the whole problem seem more remote and uncertain than it actually is.

The Oxford team, led by Alex Otto, looked at a simple climate model with systematically varied parameters of long-term (equilibrium) climate sensitivity (ECS) and transient climate response parameters. This approach allowed the researchers to calculate the cumulative damage from an additional tonne of carbon emissions above a baseline. Uncertainty in the climate ensemble was translated into a range of possible values for the cost of carbon. The study showed that understanding the "real" value of the transient climate response reduces uncertainty in the social cost of carbon far more effectively than does understanding the "real" value of climate sensitivity.

This result depends on the assumptions about the form that climate-induced damages take and our attitude towards the future. The result fails, for example, if damage occurs far more rapidly than that predicted by most studies to date. But in all other cases, i.e. with more gradually increasing damage as temperatures rise, the transient climate response captures the essence of climate uncertainty much better than climate sensitivity.

The finding highlights how important it is to put forward a more realistic representation of climate uncertainty in economic assessments of climate change.

Related link