What We Publish

Estimating the contribution of sea ice response to climate sensitivity in a climate model

Ken Caldeira and Ivana Cvijanovic · Journal of Climate 27, 8597-8607 · 2014

Key finding. In a configuration of CESM 1.0.4 with a slab ocean and a thermodynamic-dynamic sea ice model, the climate feedback parameter averages about 1.31 watts per square metre per kelvin when sea ice is absent and about 1.05 with active sea ice, so sea ice response accounts for about 20% of this model's climate sensitivity to an imposed change in radiative forcing.

Four scatter plots, labelled 2xCO2, 4xCO2, 6xCO2 and 8xCO2, of top-of-atmosphere radiation imbalance in watts per square metre against global mean temperature change in kelvin. Each panel shows three clouds of points in black, red and blue for the Active, Zero and Prescribed sea ice treatments, with fitted regression lines. Annotations give the radiative forcing and the climate feedback parameter for each treatment; in the 4xCO2 panel these are 1.02 for Active and 1.33 for Zero watts per square metre per kelvin.
Gregory regressions for three sea ice treatments at four CO2 levels. The slope is the climate feedback parameter, and it is consistently steeper without sea ice — a steeper slope meaning a less sensitive climate, which is how the roughly 20% contribution of sea ice is measured. Figure 2 from Caldeira and Cvijanovic (2014), Journal of Climate 27, 8597-8607. Reproduced under author reuse rights. Extracted from the published PDF and resized for web display.

What question did this research address?

Sea ice loss is known to amplify Arctic warming through surface albedo and through altered heat and moisture exchange between ocean and atmosphere. It is usually discussed as a regional amplifier.

But a feedback that changes the planet's absorbed sunlight also changes global climate sensitivity. This paper asked how much of a model's total sensitivity is attributable to sea ice.

What did we find?

The experiment removes sea ice rather than perturbing it: simulations are run in which sea ice cannot form and ocean temperatures are allowed to fall below freezing, and compared with controls that have active sea ice.

Without sea ice the feedback parameter averages about 1.31 watts per square metre per kelvin; with it, about 1.05. A smaller feedback parameter means a more sensitive climate, so sea ice contributes roughly 20% of sensitivity.

The effect is not purely a sunlight-reflection story. Sea ice response changes the longwave feedback parameter nearly half as much as it changes the shortwave one — the insulating effect of ice matters alongside its brightness.

The total feedback decomposes into two measurable sensitivities: how much sea ice area changes per degree of warming, and how much radiative forcing changes per unit of sea ice area.

An alternative method of disabling the sea ice response gives similar conclusions, which guards against the result being an artefact of one experimental design.

Framed as a forcing rather than a feedback, the sea ice present in the preindustrial control simulation has a climate effect equivalent to about 3 watts per square metre.

Why does it matter?

It converts a regional process into a number in the global sensitivity budget. Twenty percent of climate sensitivity resting on sea ice means that as sea ice disappears, the remaining climate becomes less sensitive to further forcing.

The longwave result corrects a common simplification. Treating sea ice feedback as an albedo effect alone misses nearly a third of what it does.

The decomposition into ice-area-per-degree and forcing-per-unit-area gives a diagnostic that can be measured in any model, which is what makes the result comparable across models rather than particular to CESM.

Citation

Ken Caldeira and Ivana Cvijanovic (2014). Estimating the contribution of sea ice response to climate sensitivity in a climate model. Journal of Climate 27, 8597-8607.

Related