Surface Fluxes and Boundary Layer Recovery in TOGA COARE: Sensitivity to Convective Organization

Abstract

Shipboard radar data collected during the Tropical Ocean Global Atmosphere Coupled Ocean‐Atmosphere Response Experiment (TOGA COARE) are used in conjunction with surface meteorological data from the Woods Hole Oceanographic Institute’s IMET buoy to describe in detail how three classifications of convective systems modify the surface fluxes of heat, moisture, and momentum. The classifications of convection were based on spatial-scale [sub‐mesoscale convective system (MCS) vs MCS scale], horizontal morphology (nonlinear vs linear organization), and the presence of stratiform precipitation as determined by quantitative radar data. Three types of convective organization were examined; sub-MCS-scale nonlinear events and MCS-scale linear events with and without significant stratiform precipitation. These three classifications were present about 90% of the time and produced over 90% of the rainfall during TOGA COARE, as determined by shipboard radar. Composite analyses of the surface fluxes along with the pertinent bulk variables have been constructed for each of these classes of convective organization. During the compositing process, the convectively active and recovery periods were separated, allowing these distinctly different phases to both be represented in the final composite analyses. The sensible and latent heat flux enhancements were decomposed through perturbation analyses and the relative importance of each term during the convectively active and recovery phases was also assessed. All three types of convective organization altered the surface fluxes in a similar manner, producing greatly enhanced surface fluxes during the convectively active phase with weaker enhancements during the recovery. However, the duration of the convectively active and recovery phases were highly dependent on the type of convective organization that was present. The average length of the convectively active phase ranged from approximately 45 min for MCS-scale linear events that had little stratiform precipitation to about 1.5 h for MCSscale linear events with extensive stratiform regions. The average length of the recovery phase was approximately 3.5 h for sub-MCS-scale nonlinear events, 2.5 h for MCS-scale linear events with little stratiform precipitation, and nearly 9.5 h for the MCS-scale linear events with extensive stratiform areas. The magnitudes of the surface fluxes were also highly dependent on the mode of convective organization. The MCS-scale linear systems that had extensive stratiform precipitation were indicative of highly organized and mature squall line systems and hence produced the greatest modulations of the surface fluxes. These events produced peak sensible and latent heat fluxes of about 60 W m 22 and 250 W m22, respectively. The sub-MCS-scale nonlinear events and the MCS-scale linear events with little stratiform precipitation produced much weaker peak sensible and latent heat fluxes (about 20 W m 22 and 150 W m22, respectively). For all three types of convective organization the enhanced sensible heat fluxes were due primarily to increased air‐sea temperature differences (i.e., decreased air temperature) and increased wind speeds. Approximately two-thirds of the total enhanced sensible heat transfer occurred during the recovery phase for each type of convective organization. For the sub-MCS-scale nonlinear events and MCS-scale linear events with little stratiform precipitation, the latent heat flux enhancements were due primarily to increased wind speeds. Increased wind speeds were also the primary contributor to the enhanced latent heat fluxes for the MCSscale linear events with extensive stratiform precipitation, but increases in the air‐sea humidity difference and the transfer coefficient for moisture also contributed. For the MCS-scale events, the enhanced latent heat transfer was split nearly evenly between the convectively active and recovery phases, whereas for the sub-MCS-scale nonlinear event, nearly 60% of the enhanced latent heat transfer occurred during the convectively active phase compared to about 40% during the recovery phase.