Published studies of ocean mesoscale processes in marginal ice zones (
MIZs) using numerical coupled ice-ocean models usually assume that the
surface wind speed is constant over the model domain and that wind st
ress variations are simply proportional to surface roughness variation
s. We show that this assumption is not realistic in most situations be
cause the surface wind stress is also significantly affected by mesosc
ale pressure variations, by changes in the surface wind vector, and by
changes in surface layer stability. Two numerical case studies, utili
zing detailed surface and atmospheric measurements, examine the factor
s affecting small-scale (<5 km) variations in wind stress within MIZs.
These case studies and surveyed observational and modeling results de
monstrate that wind stress fields are qualitatively different from the
surface roughness fields. A realistic wind stress scenario consists o
f a maximum just inside the ice edge and another maximum in the open o
cean. Stress minima occur within the pack ice region away from the MIZ
and over grease ice, if present. The effect of the rougher MIZ ice is
counteracted when wind stresses over the open ocean are enhanced by l
arge surface heat fluxes over the ocean, by a strong low level inversi
on over the ice, or by a sharp atmospheric front with surface winds pa
ralleling the ice edge. Such situations are common in MIZ regions. Som
e simple methods for including first-order atmospheric effects on wind
stress variations, which could be incorporated into current ice-ocean
mesoscale models of MIZ regions, are suggested.