Roll and Cell Convection in Wintertime Arctic Cold-Air Outbreaks

Abstract

Cold-air outbreaks from the polar ice caps or winterly continents over the open ocean lead to organized convection that typically starts as longitudinal roll patterns and changes to cellular patterns in downstream direction. During the field experiments ARKTIS 1991 and ARKTIS 1993, aircraft missions were conducted in 13 cold-air outbreak events over the Greenland and Barents Seas to determine the characteristic parameters of both the mean (primary) flow and the superimposed organized convection (secondary flow). The measurements are classified into four categories with respect to the convective pattern form: longitudinal rolls with small and wider horizontal wavelengths, transitional forms between rolls and cells, and cells. Rolls were observed for boundary layer depths h < 1 km with horizontal wavelengths ? < 5 km and aspect ratios ?/h between 2.6 and 6.5. Distinct cellular structures occurred for h > 1.4 km with ? > 8 km and ?/h between 4 and 12. The amplitudes of the secondary flow-scale variations of the temperature ?R, moisture mR, and the longitudinal, uR; transversal, ?R; and vertical, wR, wind components were on the order of 0.1?0.4 K, 0.03?0.30 g kg?1, 0.6?2.5 m s?1, 0.8?2.5 m s?1, and 0.4?1.8 m s?1, respectively, generally increasing from the roll to the cell region. The same is true for the ratio uR/?R (from about 0.6 to nearly 1) and for the ratio LmR/cp?R (from 0.7 to more than 2), hinting at increasing importance of moisture processes in the cell compared to the roll region. The importance of the secondary-flow transports of heat and momentum in relation to the total vertical transports increases with height and from rolls to cells. Particularly clear is the vertical profile of the vertical moisture transport mRwR, which exhibits a maximum around cloud base and is on the average related to the surface moisture flux as (mRwR)max = 0.35(m?w?)o. The thermodynamic conditions of the basic flow are characterized by the Rayleigh number Ra, the stability of the capping inversion, and the net condensation rate in the cloud layer. Here Ra is clearly overcritical in the whole cold-air outbreak region; it is around 105 in the roll region and around 2 ? 106 in the cell region. The Monin?Obukhov stability parameter does not appear to be suitable measure to distinguish between roll and cell convection. The stability above the boundary layer is about two to three times larger for rolls than for cells. The net condensation in clouds is three times larger in cell than in roll regions and the resulting heating of the boundary layer exceeds that of the surface heat flux in the cell region. The kinematic conditions of the basic flow are characterized by a larger shear of the longitudinal wind component u in the roll than in the cell region. The curvature of the u profile is mostly overcritical in rolls and always subcritical in cells. The secondary flow-scale kinetic energy Ekin,R is related to Ra. The best least squares fit is given by Ekin,R = 3.7Ra0.4.