Determination of size-dependent dry particle deposition velocities with multiple intrinsic elemental tracers

Dry deposition flux and aerosol size distribution measurements were made concurrently aboard the RV Lake Guardian 19 km east of the Chicago shoreline during summer 1994 to assess atmospheric inputs of minor and trace elements to southern Lake Michigan. Size-segregated aerosol measurements were made over consecutive 12-h periods with Micro-Orifice and Noll Rotary impactors (MOI and NRI), and depositing-particulate collections to aerodynamically smooth airfoils were made over periods of 3-4 days. The combination of the MOI and NRI provided size-segregated particulate samples in 12 discrete intervals between 0.059 and, nominally, 100 μm. The samples were analyzed for As, Ca, Mg, Se, Sb, V, and Zn by instrumental neutron activation analysis and for S by X-ray fluorescence. Aerosol and deposition data for individual elemental constituents were fit with a chemical mass balance deposition (CMBD) model in which a set of particle-size-specific deposition velocities (V(d)), best reconciling the deposition data, were determined by iterative (constrained) solution of a series of six linear equations using the Levenberg-Marquardt method. Under stable conditions and mean wind speed of 4.0 m s^-1, minimum V(d) values for particles with physical diameters between 0.09 and 0.53 μm averaged 0.006 ± 0.005 cm s^-1, wherein uncertainties were determined by Monte Carlo analysis. This agrees favorably with values determined by microscopy for which uncertainties were much larger and with those predicted by the Williams model for the same period. The results suggest that physically significant V(d) values are obtainable from a constrained CMBD model.