On the Arctic Ocean ice thickness response to changes in the external forcing

Abstract

Submarine and satellite observations show that the Arctic Ocean ice cover has undergone a large thickness reduction and a decrease in the areal extent during the last decades. Here the response of the Arctic Ocean ice cover to changes in the poleward atmospheric energy transport, F _wall, is investigated using coupled atmosphere-ice-ocean column models. Two models with highly different complexity are used in order to illustrate the importance of different internal processes and the results highlight the dramatic effects of the negative ice thickness—ice volume export feedback and the positive surface albedo feedback. The steady state ice thickness as a function of F _wall is determined for various model setups and defines what we call ice thickness response curves. When a variable surface albedo and snow precipitation is included, a complex response curve appears with two distinct regimes: a perennial ice cover regime with a fairly linear response and a less responsive seasonal ice cover regime. The two regimes are separated by a steep transition associated with surface albedo feedback. The associated hysteresis is however small, indicating that the Arctic climate system does not have an irreversible tipping point behaviour related to the surface albedo feedback. The results are discussed in the context of the recent reduction of the Arctic sea ice cover. A new mechanism related to regional and temporal variations of the ice divergence within the Arctic Ocean is presented as an explanation for the observed regional variation of the ice thickness reduction. Our results further suggest that the recent reduction in areal ice extent and loss of multiyear ice is related to the albedo dependent transition between seasonal and perennial ice i.e. large areas of the Arctic Ocean that has previously been dominated by multiyear ice might have been pushed below a critical mean ice thickness, corresponding to the above mentioned transition, and into a state dominated by seasonal ice.

Appendix

We base this divergence formulation (DF₃) on an approximate version of the ice export—ice thickness relation, Q′ _ice = f(H _ice), (Fig. 6a), presented by Hibler et al. (2006, their Fig. 7a). The observed annual cycle of the divergence as reported by Kwok and Rothrock (1999) is, at least to the first order, related to the prevailing wind systems over Fram Strait. Thus we keep the seasonal cycle of the divergence by using a normalized ice thickness factor, K _ice in this formulation. K _ice is obtained by dividing Q′ _ice by the ice thickness Eq. 8 and then normalizing so that an ice thickness of 2.5 m has K _ice equal to one Eq. 9, (Fig. 6b). The ice thickness factor is also set to one for ice thinner than 2.5 m assuming that the ice export is purely wind driven and thus unaffected by the internal ice stress.

$$ A^{\prime}_{\exp } = \frac{{f(H_{\text{ice}} )}}{{H_{\text{ice}} }} $$

(8)

$$ K_{\text{ice}} = \frac{{A^{\prime}_{\exp } }}{{A^{\prime}_{\exp } (H_{\text{ice}} = 2.5)}} $$

(9)

The ice volume export in the DF₃ case is then given by

$$ Q_{\text{ice}} = K_{\text{ice}} DA_{B} H_{\text{ice}} $$

(10)

Fig. 6

a Net Arctic Ocean ice volume export, Q′ _ice, as a function of ice thickness, approximately reproduced from Hibler et al. (2006). b Normalized ice thickness factor, K _ice used in the DF₃ case, see Eq. 10