Sea Surface and High-Latitude Temperature Sensitivity to Radiative Forcing of Climate over Several Glacial Cycles

A compilation is presented of global sea surface temperature (SST) records that span around one glacial cycle or more, and it is compared with changes in the earth's radiative balance over the last 520 000 years, as determined from greenhouse gas concentrations, albedo changes related to ice sheet area and atmospheric dust fluctuations, and insolation changes. A first scenario uses global mean values for the radiative changes, and a second scenario uses zonal means for 10 latitude bands for a more regionally specific perspective. On the orbital time scales studied here, a smooth increase of SST response from the equator to high latitudes is found when comparison is made to global mean radiative forcing, but a sharply "stepped" increase at 20 -30 latitude when comparing with the more regionally specific forcings. The mean global SST sensitivities to radiative change are within similar limits for both scenarios, around 0.8 +/- 0.4 degrees C (W m(-2))(-1). Combined with previous estimates of 1.3-1.5 times stronger temperature sensitivity over land, this yields an estimate for global climate sensitivity of 0.85 (-0.4/+0.5)degrees C (W m(-2))(-1), close to previous estimates. If aerosol (dust) feedback were to be considered as a fast feedback, then the estimated central value for SST sensitivity would change to similar to 0.95 degrees C (W m(-2))(-1) and that for global climate sensitivity to similar to 1.05 degrees C (W m(-2))(-1). The zonal-mean scenario allows an assessment of (long-term) "normalized amplification" for Greenland and Antarctic temperature sensitivities, which is the ratio of temperature sensitivity for those sites relative to the global mean sensitivity, normalized per watt per meter squared of radiative change. This ratio is found to be 0.9 (-0.2/+ 0.6) and 1.4 (-0.4/+1.1) for Greenland and Antarctica, respectively. Given its value close to 1 for Greenland, but that larger Arctic amplification on shorter time scales due to fast sea ice albedo processes cannot be excluded, it is suggested that current high Arctic sensitivity is mainly due to sea ice albedo feedback processes and may decrease considerably if and when the Arctic sea ice cover has been eliminated. The normalized amplification value of 1.4 for Antarctica supports previous reconstructions of polar amplification in that region. The authors propose that this amplified response resulted from approximately threefold glacial-interglacial changes in the area of sea ice cover around Antarctica.