In situ measurement of surface turbulent exchange over Arctic sea ice

Bulk flux algorithms are routinely used in climate and numerical weather prediction models to estimate the surface exchange of momentum, heat, and moisture. However, these algorithms often perform poorly over sea ice where the effect of variable ice conditions on the exchange coefficient is poorly understood. This is largely due to the paucity of direct surface flux measurements in polar regions which are needed to develop better algorithms. Some recent state of the art parameterizations attempt to address this issue by accounting for both the spatial and temporal heterogeneity of sea ice due to, for example, the appearance of leads and melt ponds in summer, and the dependence of ice surface roughness on morphology and age. These parameterizations are derived largely from theoretical considerations and have yet to be thoroughly tested against observations. Here, we present direct measurements of surface stress and heat fluxes obtained during a 3 month cruise in the Arctic Ocean during summer 2014. Measurements were made via the eddy covariance technique from a mast over the bow of the icebreaker Oden. Flux estimates were obtained over a wide range of surface conditions, ranging from open water to multi-year pack ice. Digital imagery is used to classify the surface through estimation of local ice properties, including the fractional area of multiple ice categories, melt ponds, and lead. Variability in transfer coefficient estimates with surface conditions is investigated, and several parameterizations of differing complexity are tested using the direct estimates