Q. J. R. Meteorol. Soc. (2002), 128, pp. 2281–2300 doi: 10.1256/qj.01.71 Micrometeorological measurements of particle deposition velocities to moorland vegetation By EIKO NEMITZ 1 ¤ , MARTIN W. GALLAGHER 2 , JAN H. DUYZER 3 and DAVID FOWLER 1 1 Centre for Ecology and Hydrology, UK 2 UMIST, UK 3 TNO-MEP, the Netherlands (Received 30 April 2001; revised 29 April 2002) SUMMARY Size-segregated particle number ux measurements using a micrometeorological technique, over the dia- meter (d p ) range 0.1 to 3 ¹m and total particle number uxes (d p > 11 nm), are reported for Scottish moorland vegetation. Mean particle deposition velocities (V d / range from 0.3 mm s ¡1 for 0.1 ¹m particles to more than 10 mm s ¡1 for 3 ¹m particles. On average, the measured size-dependence of V d is almost identical with the prediction by the widely used Slinn model, using the original parameters derived from wind-tunnel studies. Within each size-band V d increased linearly with the friction velocity (u ¤ /. Total particle number uxes measured with a condensation particle counter are dominated by the high concentration of small particles; therefore, this provides the rst estimate of V d for particles smaller than those that can be sized by optical techniques, for short vegetation. Values of V d ranged from 0.2 to 0.6 mm s ¡1 as u ¤ increased from 0.15 to 0.55 m s ¡1 . These deposition rates are one order of magnitude smaller than have been observed for a pine forest stand. The analysis shows that, at least for super-micron particles, the emission uxes that were frequently observed at the study site need to be included in the computation of average values. At this site, upward uxes are therefore more likely to be an artefact due to statistical constraints than a systematic physical phenomenon. KEYWORDS: Condensation particle counter Eddy-correlation technique Optical particle counter Particle number uxes Slinn model 1. I NTRODUCTION Aerosol particles are removed from the atmosphere by precipitation scavenging (wet deposition), as well as through turbulent vertical transport and interaction with natural surfaces (dry deposition). The quantication of these processes is necessary to establish the atmospheric residence time of atmospheric particles which governs their transport distance, trans-boundary uxes and potential climate effects. In addition, the deposition rate is needed to quantify the inputs of aerosol-bound pollutants, including toxic metals and the particulate forms of sulphur and nitrogen, into semi-natural and agricultural ecosystems. Dry deposition uxes of individual aerosol components (e.g. heavy metals, sul- phate, nitrate, ammonium, base cations) may be measured directly at a limited number of sites using micrometeorological techniques. Fast-response sensors, required for the application of the more direct eddy-correlation technique, are generally not available for these aerosol components. Fluxes are therefore derived from the long-term averages of concentration proles, for example using the aerodynamic gradient method (e.g. Wyers and Duyzer 1997), while the relaxed eddy accumulation (REA) technique is also being pioneered for aerosol particles (Schery et al. 1998; Nemitz et al. 2000). In the absence of direct measurements and to achieve spatial coverage the dry deposition ux (F Â / is frequently calculated from the elds of the concentration (Â ), either measured (inferen- tial approach) or predicted using atmospheric transport models, by multiplication with a dry deposition velocity (V d D¡F Â =Â ). Values of V d may be derived from speciated ux measurements (e.g. Wyers and Duyzer 1997) or, more generally, from the combined ¤ Corresponding author: Centre for Ecology and Hydrology, Edinburgh Research Station, Penicuik, Midlothian, EH26 0QB, UK. e-mail: en@ceh.ac.uk c ° Royal Meteorological Society, 2002. 2281