RELATIONSHIP BETWEEN SEA ICE FREEBOARD AND DRAFT IN THE ARCTIC BASIN, AND IMPLICATIONS FOR ICE THICKNESS MONITORING

We have confirmed our earlier finding that the probability density function (pdf) of ice freeboard in the Arctic Ocean can be converted to a pdf of ice draft by applying a simple coordinate transformation based on the measured mean draft and mean elevation. This applies in each of six 50-km sections (north of Greenland) of joint airborne laser and submarine sonar profile obtained along nearly coincident tracks from the Arctic Basin north of Greenland and tested for this study. Detailed differences in the shape of the pdf can be explained on the basis of snow load and can, in principle, be compensated by the use of-a more sophisticated freeboard-dependent transformation. The measured ''density ratio'' R (actually mean draft/mean elevation ratio) for each section was found to be consistent over all sections tested, despite differences in the ice regime, indicating that a single value of R might be used for measurements done in this season of the year. The mean value [R] from all six sections is 7.89; on the assumption that all six values are drawn from the same population, the standard deviation is 0.55 for a single 50-km section, and thus 0.22 for 300 km of track. In attempting to infer ice draft from laser-measured freeboard, we would therefore expect an accuracy of about +/- 28 cm in 50 km of track (if mean draft is about 4 m) and about +/- 11 cm in 300 km of track; these accuracies are compatible with the resolution of predictions from numerical models. A simple model for the variability of R with season and with mean ice thickness gives results in reasonable agreement with observations. They show that although there is a large seasonal variability due to snow load, there is a stable period from November to April when the variability is chiefly dependent on the mean ice thickness alone. Thus, in principle, R can be mapped over the Arctic Ocean as a basis for interpreting survey data. Better field data are needed on the seasonal and spatial variability of three key quantities: area-averaged snow load, mean density of first-year and multiyear ice (including the effect of ridging with these two ice regimes), and density of near-surface water.