PHYSICOCHEMICAL MODELING OF THE FIRST AEROSOL CHARACTERIZATION EXPERIMENT (ACE-1) LAGRANGIAN-B - 1 - A MOVING COLUMN APPROACH

During Lagrangian experiment B (LB in the following) of the First Aero sol Characterization Experiment (ACE 1): a clean maritime air mass was followed over a period of 28 hours. During that time span, the vertic al distribution of aerosols and their gas phase precursors were charac terized by a total of nine aircraft soundings which were performed dur ing three research flights that followed the trajectory of a set of ma rked tetroons. The objective of this paper is to study the time evolut ion of gas phase photochemistry in this Lagrangian framework. A box mo del approach to the wind shear driven and vertically stratified bounda ry layer is questionable, since its basic assumption of instantaneous turbulent mixing of the entire air column is not satisfied here. To ov ercome this obstacle, a one-dimensional Lagrangian boundary layer mete orological model with coupled gas phase photochemistry is used. To our knowledge, this is the first time that such a model is applied to a L agrangian experiment and that enough measurements are available to ful ly constrain the simulations. A major part of this paper is devoted to the question of to what degree our model is able to reproduce the tim e evolution and the vertical distribution of the observed species. Com parison with observations of O-3, OH, H2O2, CH3OOH, DMS, and CH3I, mad e on the nine Lagrangian aircraft soundings shows that this is in gene ral the case, although the dynamical simulation started to deviate fro m the observations on the last Lagrangian flight. In agreement with ex perimental findings reported by Q.Wang et al. (unpublished manuscript, 1998b), generation of turbulence in the model appears to be most sens itive to the imposed sea surface temperature. Concerning the different modeled and observed chemical species, a number of conclusions are dr awn: (1) Ozone, having a relatively long photochemical lifetime in the clean marine boundary layer, is found to be controlled by vertical tr ansport processes, in particular synoptic-scale subsidence or ascent. (2) Starting with initally constant vertical profiles, the model is ab le to ''create'' qualitatively the vertical structure of the observed peroxides. (3) OH concentrations are in agreement with observations, b oth on cloudy and noncloudy days. On the first flight, a layer of dry ozone rich air topped the boundary layer. The model predicts a minimum in OH and peroxides at that altitude consistent with observations. (4 ) Atmospheric DMS concentrations are modeled correctly only when using the Liss and Merlivat [1986] flux parameterization, the Wanninkhof [1 992] flux parameterization giving values twice those observed. To arri ve at this conclusion, OH is assumed to be the major DMS oxidant, but no assumptions about mixing heights or entrainment rates are necessary in this type of model. DMS seawater concentrations are constrained by observations.