AN ARCTIC HURRICANE OVER THE BERING SEA

A mesoscale "arctic hurricane" developed over the western Bering Sea on 7 March 1977 and traveled eastward parallel to the ice edge along a zone of large sea surface temperature gradient. Satellite imagery reveals spiral cloud bands of unusual symmetry and mesoscale dimension associated with the mature stage of the low. The track of the low pressure center passed over the rawinsonde station at St. Paul Island where time series of surface data show a pronounced maximum in equivalent potential temperature at the core of the low. The storm made landfall with surface winds > 30 ms-1 at Cape Newenham, Alaska, on 9 March and rapidly dissipated thereafter. Synoptic analyses show that the arctic hurricane formed at the leading edge of an outflow of arctic air that originated over the ice and passed over the open water of the western Bering Sea. In the mid- and upper troposphere a large cold-core low dominated the Bering Sea region. Quasi-geostrophic analysis at 0000 UTC 7 March 1977 reveals conditions conducive to synoptic-scale ascent over the region of the incipient low, as a sharp upper-level short wave crosses the Siberian coast. Conversely, during its mature stage little quasi-geostrophic forcing is seen over the low. In order to investigate the ability of sea surface heat fluxes to develop and maintain the arctic hurricane, an analytical model based on the CArnot cycle, and an axisymmetric numerical model with the Kuo cumulus parameterization scheme are applied. The analytical calculation of the pressure drop from the outermost closed isobar to the storm center results in a central pressure of 973 mb, which agrees well with observation. When the initial environment of the numerical model is set to be similar to that observed with the arctic hurricane, the model correctly predicts the minimum sea-level pressure, strength of the wind circulation, and the magnitude of sensible heat fluxes observed with the storm. The dynamic and thermodynamic structures of the simulated storm are similar to those of tropical cyclones. The predicted development time of the storm is longer than observations suggest, and the diameter of the simulated anvil outflow is somewhat larger, pointing to the likely importance of baroclinic processes in the evolution of the disturbance, and the need for further numerical studies with mesoscale models that employ full three-dimensional primitive equations.