Comparison of Methods for Estimating Mixing Height Used during the 1992 Atlanta Field Intensive

Abstract

During the summer of 1992, measurements of the boundary layer mixing height were conducted at five locations around the city of Atlanta, Georgia, as part of the 1992 Atlanta Field Intensive of the Southern Oxidants Research Program on Ozone Non-Attainment. These measurements were made during a series of ?high-ozone-event days? for the purpose of acquiring information about the temporal evolution of the convective mixed layer. The information acquired from these systems was included in a database of meteorological variables for use in the photochemical modeling efforts associated with the study. The following measurement systems were selected for use in this study by organizers of the 1992 Atlanta Field Intensive: one rawinsonde system, four radar wind profiler?RASS (radar acoustic sounding system) systems, and two lidar systems. A comparison of the mixing-height estimates from each of the measurement systems used during the 1992 Atlanta Field Intensive was performed in an effort to evaluate the consistency of the estimates between the different systems and, further, to evaluate the relative performance of each system during the study period. Statistical analyses were performed on the dataset, with in-depth statistical analyses presented for two specific days: 30 July and 4 August 1992. Results indicate that there is often disagreement in the mixing-height estimates between the various systems, particularly during the early morning and late afternoon. It is believed that the differences between estimates are the result of 1) the physical limitations of the different instrument system 2) the assumptions used with each system as to which tracer most accurately defines the structure of the convective boundary layer, and 3) the spatial inhomogeneity of convective boundary layer structure across the region studied. In general, the rawinsonde system appeared to give the most accurate mixing-height estimates under the meteorological conditions studied. The lidar estimates were comparable to the rawinsonde estimates, except during the early morning hours, when some estimates were erroneously high. This overestimation occurred when the return signal from aerosols suspended aloft during the previous day's mixing exceeded that from new aerosols being mixed into the developing convective boundary layer. Radar wind profiler estimates were also erroneously high during the morning hours, due mainly to the use of a setup configuration that precluded detection of the newly developing mixed layer below 400?600 m.