Empirical estimates of size-resolved precipitation scavenging coefficients for ultrafine particles

Below-cloud scavenging coefficients for ultrafine particles (UFP) exhibit comparatively large uncertainties in part because of the limited availability of observational data sets from which robust parameterizations can be derived or that can be used to evaluate output from numerical models. Long time series of measured near-surface UFP size distributions and precipitation intensity from the Midwestern USA are used here to explore uncertainties in scavenging coefficients and test both the generalizability of a previous empirical parameterization developed using similar data from a boreal forest in Finland (Laakso et al., 2003) and whether a more parsimonious formulation can be developed. Scavenging coefficients (lambda) over an ensemble of 95 rain events (with a median intensity of 1.56 mm h(-1)) and 104 particle diameter (D-p) classes (from 10 to 400 nm) indicate a mean value of 3.4 x 10(-5) s(-1) (with a standard error of 1.1 x 10(-6) s(-1)) and a median of 1.9 x 10(-5) s(-1) (interquartile range: -2.0 x 10(-5) to 7.5 x 10(-5) s(-1)). The median scavenging coefficients for D-p: 10-400 nm computed over all 95 rain events exhibit close agreement with the empirical parameterization proposed by (Laakso et al., 2003). They decline from -4.1 x 10(-5) s(-1) for D-p of 10-19 nm, to similar to 1.6 x 10(-5) s(-1) for D-p of 80-113 nm, and show an increasing tendency for D-p > 200 nm. (C) 2016 Elsevier Ltd. All rights reserved.