Dry deposition of particles is an important way of aerosol removal from the atmosphere and a key process in surface-atmosphere exchanges. The deposition velocities, Vd, are often parameterised in air quality and climate modelling as function of the friction velocity, u*, atmospheric stability, and particle size (if size-segregated information is available). In this work, a study of the correlation between Vd and u* over different surfaces is presented for both PM2.5 and particle number fluxes. Results indicate an almost linear increase of Vd with u* with slopes similar for PM2.5 fluxes and particle number fluxes over the different surfaces analysed. This means that the ratios Vd/u* tend to collapse over similar values even if Vd and u* are significantly different because u* take into account most of the surface effects. There is a limited difference between stable cases and unstable/neutral cases with slightly lower deposition velocities in stable cases for fixed values of u*. The average value of Vd/u* is 0.010 ± 0.0017 (median 0.0062 ± 0.0015) (considering all stabilities) and 0.0097 ± 0.002 (median 0.005 ± 0.001) for stable cases. This could be the base for an empirical parameterisation of deposition velocities in air quality models.
1. Introduction
Atmospheric aerosol particles are generated by both anthropogenic and natural sources and through chemical and physical processes in the atmosphere. The dynamics of atmospheric aerosols is highly complex, involving particle formation, growth, and surface exchange processes [1]. Effects of aerosols include direct and indirect climate forcing through the absorption and scattering of incoming solar radiation and the formation of clouds by condensation nuclei activation [2, 3], reduction of visibility [4], and impact on human health [5–8]. Dry deposition of particles is a key process in atmosphere-surface exchange. It is a continuous process that gives a significant contribution to atmospheric particles removal in most environments. The analysis and the parameterization of the processes affecting vertical transport and exchange of particles are a relevant research topic for air quality and climate modelling [9–11]. There are several possible methods to investigate dry deposition of aerosol; however, in the last several years the eddy-covariance method (EC) became widely used to investigate dry deposition velocities over several typologies of surfaces [12, 13]. EC has been used to characterise deposition velocities in rural sites and over forests [12, 14, 15], over ice and snow [16–19], and in urban environments to characterise emission velocities (i.e., upward fluxes) [20–25].
There are several factors influencing dry deposition of aerosol, mainly the friction velocity, the particle size, boundary layer conditions (turbulence intensity), atmospheric stability, and collecting properties of the surface. There are many models that try to explain and parameterise the deposition process in complex surfaces by taking into account several mechanisms characterising turbulent fluxes [26–28]. In other cases, the simpler parameterisations of dry deposition velocities Vd as function of friction velocity u* and of stability (through the Monin-Obukhov length, L) are used [12, 13, 15]. These parameterisations could be used in transport and dispersion models to take into account the transfer of materials from atmosphere to the surface. In general, the nondimensional ratio Vd/u* appears to be smaller in stable atmospheric conditions with respect to unstable conditions [29].
In this work an analysis of the correlation between dry deposition velocities, taken over different surfaces, and friction velocities as function of stability is discussed. Measurements refer to both PM2.5 mass fluxes and particle number fluxes.
2. Measurement Sites
The datasets analyzed in this work were taken in different experimental campaigns over a wide range of surface roughness conditions, from almost smooth surfaces (i.e., iced surfaces in Antarctica) to surfaces with different degrees of complexity: urban background, urban canopy, and industrial district (in Italy) or patchy Venice lagoon surface (Italy). In the following sections each experimental site and campaign have been briefly described and the details of the sites characteristics are summarised in Table 1 and shown in Figure 1. The micrometeorological measurements were used to evaluate the displacement height d and the roughness height z0 for each site (Table 1) using the method reported in [30], which uses similarity relationship for sonic temperature and vertical wind component.
Summary of experimental sites and instruments used in aerosol sampling. The table includes the details of the set-up, such as measurement height (z), displacement height (d), and roughness length (z0).
2.1. Venice Lagoon (Mazzorbo Island, North-Eastern Italy)
Measurements of PM2.5 concentrations and fluxes were performed at a background site placed on the island of Mazzorbo, in the Venice lagoon at 10 m above the ground. The measurement site (45°29′09.5′′N, 12°24′12.7′′E) was a field located at about 8 km NE of the Venice town. This site was located very close (about 5 m) to the water lagoon at the W-SW side, while, in the other directions (north, east, and south), it was characterised by land for about 1-2 km with short vegetation, some small trees, and one or two-floor houses, although channels and water were also present in this area (Figure 1(a)). Three measurement campaigns were performed: the first measurement campaign (summer) in July 2004 (2 to 18), the second campaign (winter) in February and March 2005 (16 February to 15 March), and the third campaign (spring) in May 2006 (5 to 23). These campaigns are analysed together in this work. More details on the site can be found in [31].
2.2. Lecce Urban Background Site (South-Eastern Italy)
A measurement campaign was performed during spring/summer 2005 form April until June relative to PM2.5 concentrations and fluxes. A second campaign was performed between 12 and 30 July, 2010 for measurements of particle number concentrations and fluxes. The site was the experimental field of the Lecce Unit of ISAC-CNR placed inside the University Campus (40°20′10.8′′N, 18°07′21.0′′E) and located at about 3.5 km SW from the town of Lecce. The site is a rectangular field with a major side of about 200 m characterized by short vegetation, with two contiguous sides surrounded by small trees (Figure 1(b)). The urban background area is characterized for at least 1 km in all directions by the presence of patches of trees (8–10 m tall) and small two-storey buildings and some roads with no industrial releases nearby. Due to the proximity of urban areas, the site can be categorized as an urban background area. Measurements were taken at 10 m above the ground. More details on this site can be found in [32] or [33].
2.3. Bologna Industrial District (Central Italy)
Measurements of particle number concentration and fluxes were taken between 6 June and 22 July 2008 (summer campaign) and from 20 January to 10 March 2009 (winter campaign). Nearby measurement site, on left of Mobile Laboratory in Figure 1(c), there was the incinerator plant for the city of Bologna (44°31′17.59′′N, 11°25′53.48′′E). The two campaigns are analysed together in this work.
2.4. Antarctica Remote Site
The data, relative to particle number concentrations and fluxes, were collected over ice/snow surface in Antarctica during austral summer in 2006 in the framework of the Italian National Research Program in Antarctica (PNRA). Measurements were performed on the Nansen Ice Sheet (NIS), a coastal region of the Northern Victoria Land (Antarctica). The NIS is a permanently frozen branch of the Ross Sea that penetrates into a region of about 35 × 70 km2, surrounded by complex topography. Because of its remote inland location, the site appears ideal for sampling unperturbed atmospheric aerosol characteristics (Figure 1(d)). The campaign was performed throughout the period from 8 to 31 December 2006; the micrometeorological tower (12 m height) was located at a distance of about 50 km from the open sea (74°30′02.0′′S, 163°27′30.0′′E). More details on the site can be found in [19].
2.5. Maglie Urban Site (South-Eastern Italy)
The measurement site was located in the NE boundary of the town of Maglie in the Apulia regions in the SE of Italy (40°07′38.39′′N, 18°17′59.50′′E). The site could be considered an urban background site influenced by an industrial area. The town is extending mainly in the sector of wind direction between SE and SW and the countryside is in the sector between NNO and E. In the town direction the site is characterized by the presence of small buildings (1-2 storeys) and roads with relatively high traffic volume (Figure 1(f)). Five measurement campaigns have been performed (January 2004, December 2004, December 2006, December 2007, and September 2008) for a total of 101 measurement days. These datasets are analysed together in this work. More information can be found in [34].
3. Instrument Setup
In all the experimental sites, micrometeorological flux systems based on the eddy-covariance (EC) technique were used to measure vertical turbulent fluxes of momentum, tracers, and energy. The measuring station was based on a three-dimensional ultrasonic anemometer (R3 Gill Instruments Ltd., Lymington, UK), operating at 100 Hz in calibrated mode. A slow-response thermohygrometer (Rotronic MP100A) was installed in order to measure temperature and relative humidity during the campaigns.
An infrared optical sensor pDR-1200 (Personal Data logging Real Time Aerosol Monitor by Thermo Electron, Mie Corp.) was used to measure PM2.5 concentrations and fluxes during the field campaign in Lecce (2005), Venice lagoon, and Maglie, as reported in Table 1. The pDR-1200 was operating at 1 Hz in active sampling (4 L/min) and it was equipped with a cyclone (2.5 μm cut-off at the 4 L/min flow rate used, model GK2.05) for PM2.5 selection [35]. It was verified that exists a delay t0, about 2 s, between change in mass concentration and the effective measure of pDR-1200. This delay has been also verified by searching the maximum of the absolute value of the covariance between the vertical wind velocity and the concentration time-series and it has been taken into account in the evaluation of the turbulent fluxes. Atmospheric aerosol can be highly hygroscopic and it can absorb water vapour at high relative humidity changing dimension, density, and optical properties; this process modifies the scattering and absorption coefficients of particles and then it modifies the response of the optical detector used [36]. Therefore, measured concentrations were corrected, using the procedure described in [33], to take into account the role of relative humidity.
In Antarctica, Bologna, and Lecce (2010) sites (Table 1), instrumental setup included a Condensation Particle Counter (CPC-Grimm Aerosol, model 5.403) that measured the total particle number concentration (PNC) with a sampling frequency of 1 Hz. The performances of this CPC are analyzed by [37]. The CPC output was connected to the analog inputs of the anemometer by means of a digital-to-analog conversion with a simple two-channel interface. More information about the used instruments configuration is reported in [19, 25]. The particle losses for the inlet system were calculated according to the formulation of [38] for the laminar flow inside the last part of the inlet and according to [39] for turbulent flow in the large section tube. The results show that the cut-off diameter (at 50% efficiency), D50, is about 9 nm. Therefore, the system used was able to detect particles of between 9 and 1000 nm (i.e., the upper limit of the CPC). Like for the pDR-1200, the delay in the inlet tube between concentration and velocity fluctuations was taken into account in the evaluation of the eddy-covariance.
4. Method and Data Processing
All datasets have been reduced in the streamlines reference system [40] with three rotations using linear detrending of time-histories in order to remove variations related to synoptic time scales [41] and an averaging time of 30 min. Before the computation of turbulent fluxes, the basic instrumental and physical corrections have been applied to the measured time series. Spikes as well as runs with wind directions contaminated by tower/obstacles distortions were discarded. A stationary test has been performed for data series, after the process of detrending [42], in order to individuate nonstationary cases. The nonstationary cases have been eliminated from successive data analysis and the percentages of occurrence of these cases are reported in Table 2.
Summary of experimental sites with indication of the typology of measurements. The table includes the percentages associated with the different corrections such as density fluctuation correction, high frequency loss correction, and nonstationary data removed.
Site
Measurements
Density fluctuationcorrection (%)
Nonstationarydata (%)
High frequency losscorrection (%)
Antarctica
Particle number
3.4
15
43
Venice lagoon
PM2.5
Not applied
6
27
Bologna
Particle number
Not applied
18
30
Lecce 2005
PM2.5
Not applied
5
20
Lecce 2010
Particle number
0.2
6
23
Maglie
PM2.5
Not applied
8
24
Fast measurements allowed us to use eddy-covariance technique separating the aerosol concentration and the vertical wind component into mean values and turbulent fluctuations [43]. It is useful to normalize aerosol fluxes using the aerosol concentrations obtaining the deposition velocity:
(1)Vd=-w′C′¯C¯,
where w′ are the fluctuations of the vertical wind velocity, C′ the fluctuations in aerosol concentrations, and C- the average aerosol concentration. The averaging period for application of the eddy-covariance was 30 minutes for all the measurement campaigns. In the measurement campaigns analysing PM2.5 fluxes, C was a mass concentration of PM2.5; in cases in which the particle number fluxes were analysed, C was the particle number concentration. EC measurements with wind velocities lower than 0.5 m/s (wind calm) were removed as they are considered unreliable for calculation of fluxes due to low turbulent mixing.
In the campaigns in which the measurements of latent heat fluxes were available the aerosol fluxes were corrected for variation in air density due to the water vapour fluxes following [44]. No correction was made for variation in density due to heat flux, because heat fluctuations are assumed to be dissipated in the inlet tube [45]. The amounts of these corrections are reported in Table 2.
Measured aerosol fluxes were also corrected for the high frequency losses due to the limited time response of the instruments used. The first-order time response of the pDR-1200 used for measurements of PM2.5 fluxes was 1.1 s and that of the CPC used for measurements of particle number fluxes was 1.3 s. The correction of high frequency losses was performed following the method proposed in [46]. However, in the Antarctica dataset, this method of correction appears to give an overestimation of the correction likely due to the strongly stable conditions. Thus an alternative method was developed that used a low-pass digital filter (first-order Butterworth) approximating the CPC response to a concentration step measured in laboratory as discussed in [19]. The strengths of the high frequency loss corrections are reported in Table 2.
5. Results
Dry deposition velocities have been analysed selecting downward fluxes for the different datasets to separate emission (upward fluxes often associated with local sources) from deposition processes [47, 48]. In general, under turbulence conditions, especially during daytime, dry deposition is controlled by the settling velocity, aerodynamic resistance, turbulent diffusion of the particles (Brownian motion), and their impaction and interception [26]. In general terms the deposition velocity is often parameterised as function of the friction velocity and of atmospheric stability [12, 13, 15]. Specifically, it parameterised the ratio Vd/u* as a function of L. In Figure 2, the dependence of Vd on friction velocity u* for each dataset referring to PM2.5 mass fluxes is reported. Results in Figures 2 and 3 are obtained segregating the data in intervals of u*. Different intervals of u* were selected to optimize the number of data points within each interval and, in each interval, the average and the standard error of Vd were calculated. In Figures 2 and 3 the horizontal bars represent the intervals in friction velocity and the vertical bars represent the standard error of the average deposition velocity within the specific interval of u*. In Figure 2, four cases have been separated: all stabilities, only cases with L > 0, only cases with L < 0, and only cases in strictly stable conditions (z/L > 0.1 with z indicating the measurement height). In Figure 3 the same analysis is reported for particle number fluxes. These figures show that even if there is some scatter in the data, deposition velocity grows with friction velocity both for PM2.5 and for particle number fluxes, even if they are measured with different instruments and over different surfaces. This growth is almost up to a friction velocity of u* = 1 m/s. Other studies also displayed a linear or close to linear dependence of Vd on u* for particles in the accumulation mode [15, 22, 43]. Results in Figure 2(b) show that at the urban site the increase of Vd at low u* (lower than 0.2-0.3 m/s) is quite limited especially in stable cases. This could be due to a larger influence of urban obstacles and differences of roughness with wind direction considering that low u* are generally associated with low wind velocities with larger fluctuations in wind direction. Our datasets cannot characterize definitively the dependence of particle fluxes on stability conditions and eventually this dependence may be indirect and expressed by the dependence of u* on atmospheric stratification, although there are some evidences for an increase in particle Vd in unstable conditions and a reduction with stable or neutral atmospheric stability. As reported in [13], just few studies have been able to quantify the influence of stability with high degree of statistical certainty.
Functional dependence of deposition velocity (Vd) from friction velocity (u*) for different measurement datasets (as reported in the title of each graph) for PM2.5 mass concentration. Error bars represent standard errors.
Functional dependence of deposition velocity (Vd) from friction velocity (u*) for different measurement datasets (as reported in the title of each graph) for particle number concentration (PNC). Error bars represent standard errors.
In Table 3, the average (with standard deviation) and median (with 25th and 75th quartiles) values are reported for deposition velocities, fluxes, concentrations, and normalized deposition velocities (Vd/u*) considering whole dataset, that is, data in every atmospheric stability conditions. In Table 4 the same variables are reported for a selection of cases in unstable and quasineutral atmospheric conditions (L < 0). In Table 5 the same variables are reported for a selection of cases in stable and quasineutral atmospheric conditions (L > 0). Finally, in Table 6 all these variables are reported in conditions of strictly stable atmosphere (z/L > 0.1). Results show minimal differences in the ratio Vd/u* measured with different instruments over grass, water, iced land, or built and patched surfaces even if the actual values of Vd and u* are significantly different. This probably is due to the fact that friction velocity carries most of the information regarding the surface effects. Further, a significant difference between average and median values that is likely associated with the nonsymmetrical distributions of Vd/u* and with the sensitivity of average values to outliers is observed. The effects of postprocessing and the detailed response of the instruments could be further analysed and results show a certain scatter in Vd/u* values, as observed also in [29]. The results seem to indicate that a first parameterisation of Vd, for example, to be used in pollution transport and dispersion modelling, could be based on using the Vd/u* ratio with a constant value or differentiating two values: one for stable conditions and the other for unstable/neutral conditions. Considering together all the datasets, an average value of Vd/u* equal to 0.010±0.0017 (median value 0.0062±0.0015) represents the cases for all stabilities. This value is reduced to 0.0097±0.002 (median value 0.005±0.001) considering cases with L>0.
Average (with standard deviation) and median (with 25th and 75th quartiles) values are reported for deposition velocities, fluxes, concentrations, and normalized deposition velocities for each site, separating PM2.5 from particle number data. In this table all data is considered, without any selection involving stability conditions.
All stability conditions
PM2.5
Vd (mm/s)
Flux (μg/m2s)
Conc. (μg/m3)
Vd/u*
Venice
Average
2.81
−0.040
26.4
0.0108
Std. dev.
4.70
0.064
25.5
0.0155
Median
1.19
−0.017
18.0
0.0048
25th quartile
0.38
−0.044
8.6
0.0019
75th quartile
3.11
−0.006
35.4
0.0140
Lecce 2005
Average
4.32
−0.031
10.2
0.0103
Std. dev.
6.16
0.047
5.9
0.0131
Median
2.04
−0.019
9.8
0.0054
25th quartile
0.85
−0.039
5.4
0.0025
75th quartile
4.99
−0.006
14.5
0.0124
Maglie
Average
4.18
−0.035
13.4
0.0077
Std. dev.
4.68
0.034
14.7
0.0091
Median
2.49
−0.023
8.8
0.0055
25th quartile
0.87
−0.052
5.7
0.0024
75th quartile
5.78
−0.008
14.2
0.0096
PNC
Vd (mm/s)
Flux (#/cm2s)
Conc. (#/cm3)
Vd/u*
Antarctica
Average
1.17
−89.2
1233.5
0.0095
Std. dev.
1.55
97.6
1029.7
0.0107
Median
0.62
−56.7
960.7
0.0059
25th quartile
0.25
−119.6
423.3
0.0028
75th quartile
1.51
−23.9
1762.5
0.0121
Bologna
Average
1.85
−2312.3
12599.3
0.0124
Std. dev.
2.43
3387.9
5798.3
0.0208
Median
1.03
−1200.0
11813.0
0.0063
25th quartile
0.42
−2872.1
8400.4
0.0027
75th quartile
2.25
−459.9
15638.0
0.0133
Lecce 2010
Average
5.77
−9718.7
12203.1
0.0122
Std. dev.
5.74
18611.6
10821.5
0.0113
Median
4.01
−3352.9
7266.9
0.0091
25th quartile
1.47
−7648.1
5397.4
0.0038
75th quartile
7.87
−1074.2
16887.1
0.0174
Average (with standard deviation) and median (with 25th and 75th quartiles) values are reported for deposition velocities, fluxes, concentrations, and normalized deposition velocities for each site, separating PM2.5 from particle number data. In this table data is selected for unstable atmospheric stability cases L<0 (including quasi-neutral cases).
Unstable and neutral conditions (L<0)
PM2.5
Vd (mm/s)
Flux (μg/m2s)
Conc. (μg/m3)
Vd/u*
Venice
Average
3.15
−0.050
23.6
0.0118
Std. dev.
4.04
0.072
23.2
0.0152
Median
1.60
−0.023
16.9
0.0059
25th quartile
0.57
−0.059
8.5
0.0023
75th quartile
3.93
−0.009
30.0
0.0144
Lecce 2005
Average
5.10
−0.040
10.6
0.0102
Std. dev.
6.61
0.054
5.8
0.0132
Median
2.83
−0.026
10.3
0.0053
25th quartile
1.31
−0.048
6.2
0.0027
75th quartile
6.07
−0.013
14.7
0.0123
Maglie
Average
3.77
−0.039
14.3
0.0075
Std. dev.
3.45
0.035
12.1
0.0075
Median
2.32
−0.025
12.5
0.0054
25th quartile
1.22
−0.061
7.4
0.0019
75th quartile
5.67
−0.011
15.8
0.0098
PNC
Vd (mm/s)
Flux (#/cm2s)
Conc. (#/cm3)
Vd/u*
Antarctica
Average
0.73
−97.4
1762.7
0.0108
Std. dev.
0.83
118.4
1182.5
0.0122
Median
0.41
−57.3
1593
0.0078
25th quartile
0.20
−111.1
815.7
0.0031
75th quartile
0.83
−27.3
1958.6
0.0128
Bologna
Average
2.41
−2883.7
11992.1
0.0111
Std. dev.
2.86
3842.4
5889.4
0.0130
Median
1.57
−1690
11059.2
0.0070
25th quartile
0.61
−3787.0
7501.3
0.0028
75th quartile
3.15
−530
14783.4
0.0143
Lecce 2010
Average
6.17
−10502.5
12672.3
0.0127
Std. dev.
5.81
19241.5
11169.5
0.0114
Median
4.60
−3886.2
7382.4
0.0095
25th quartile
1.88
−9008.2
5385.1
0.0043
75th quartile
8.27
−1420.3
17209.7
0.0178
Average (with standard deviation) and median (with 25th and 75th quartiles) values are reported for deposition velocities, fluxes, concentrations, and normalized deposition velocities for each site, separating PM2.5 from particle number data. In this table data is selected for stable atmospheric stability cases L>0 (including quasi-neutral cases).
Stable and neutral conditions (L>0)
PM2.5
Vd (mm/s)
Flux (μg/m2s)
Conc. (μg/m3)
Vd/u*
Venice
Average
2.42
−0.028
29.9
0.0096
Std. dev.
5.43
0.048
27.8
0.0157
Median
0.64
−0.012
22.0
0.0037
25th quartile
0.26
−0.027
8.7
0.0015
75th quartile
1.99
−0.004
41.6
0.0102
Lecce 2005
Average
2.79
−0.015
9.4
0.0105
Std. dev.
4.83
0.024
6.2
0.0121
Median
1.04
−0.008
8.5
0.0056
25th quartile
0.34
−0.019
4.3
0.0022
75th quartile
2.90
−0.003
13.8
0.0122
Maglie
Average
4.15
−0.034
13.3
0.0075
Std. dev.
4.55
0.034
15.3
0.0095
Median
2.55
−0.023
7.9
0.0055
25th quartile
0.83
−0.051
5.5
0.0025
75th quartile
5.72
−0.007
13.5
0.0094
PNC
Vd (mm/s)
Flux (#/cm2s)
Conc. (#/cm3)
Vd/u*
Antarctica
Average
1.32
−86.5
1062.4
0.0091
Std. dev.
1.70
90.2
916.6
0.0102
Median
0.75
−56.5
753.8
0.0059
25th quartile
0.33
−119.1
398.5
0.0028
75th quartile
1.63
−22.7
1538.9
0.0116
Bologna
Average
1.43
−1872.9
13067.3
0.0134
Std. dev.
1.95
2920.5
5688.3
0.0251
Median
0.79
−951.2
12365.6
0.0059
25th quartile
0.32
−2090.0
9051.0
0.0027
75th quartile
1.72
−388.7
16199.3
0.0125
Lecce 2010
Average
1.27
−1030.9
7002.6
0.0073
Std. dev.
1.28
1420.4
1935.5
0.0084
Median
0.89
−594.8
6877.1
0.0038
25th quartile
0.54
−923.9
5912.9
0.0020
75th quartile
1.56
−278.3
7469.7
0.0086
Average (with standard deviation) and median (with 25th and 75th quartiles) values are reported for deposition velocities, fluxes, concentrations, and normalized deposition velocities for each site, separating PM2.5 from particle number data. In this table data is selected for strictly stable atmospheric stability cases z/L>0.1.
Strictly stable condition (z/L>0.1)
PM2.5
Vd (mm/s)
Flux (μg/m2s)
Conc. (μg/m3)
Vd/u*
Venice
Average
1.16
−0.023
36.9
0.0105
Std. dev.
2.28
0.033
28.9
0.0176
Median
0.47
−0.011
32.9
0.0041
25th quartile
0.19
−0.021
13.3
0.0018
75th quartile
1.21
−0.004
53.3
0.0108
Lecce 2005
Average
1.22
−0.012
11.5
0.0102
Std. dev.
1.88
0.028
6.4
0.0111
Median
0.52
−0.006
11.3
0.0056
25th quartile
0.24
−0.011
5.6
0.0020
75th quartile
1.36
−0.002
15.8
0.0128
Maglie
Average
1.08
−0.030
34.6
0.0104
Std. dev.
1.52
0.043
24.7
0.0182
Median
0.69
−0.011
30.7
0.0046
25th quartile
0.18
−0.037
15.6
0.0016
75th quartile
1.09
−0.005
53.7
0.00944
PNC
Vd (mm/s)
Flux (#/cm2s)
Conc. (#/cm3)
Vd/u*
Antarctica
Average
1.15
−81.2
1101.0
0.0091
Std. dev.
1.56
93.3
861.8
0.0109
Median
0.63
−48.0
882.0
0.0055
25th quartile
0.25
−114.1
401.0
0.0027
75th quartile
1.34
−19.8
1633.9
0.0104
Bologna
Average
1.24
−1802.3
13989.4
0.0160
Std. dev.
1.69
2970.2
5685.7
0.0288
Median
0.71
−929.3
13383.9
0.0070
25th quartile
0.31
−1928.9
10140.0
0.0030
75th quartile
1.42
−380.0
16541.0
0.0147
Lecce 2010
Average
1.36
−1241.9
7811.3
0.0057
Std. dev.
1.56
1723.4
1861.4
0.0087
Median
0.86
−608.8
7178.1
0.0024
25th quartile
0.08
−1385.9
6878.6
0.0001
75th quartile
1.98
−55.0
8149.4
0.0069
6. Conclusions
Dry deposition of particles is a key process in atmosphere-surface exchange. The analysis and the parameterization of the processes affecting vertical transport and exchange of particles are a relevant research topic for air quality dispersion modelling and for climate modelling. There are several factors influencing dry deposition of aerosol, mainly the friction velocity, the particle size, boundary layer conditions (turbulence intensity), atmospheric stability, and collecting properties of the surface. In this work, an analysis of the correlation between dry deposition velocities, taken over different surfaces, and friction velocities as function of stability is discussed for both PM2.5 mass fluxes and particle number fluxes. Results indicate that deposition velocity increases almost linearly with the increase of u*, up to a friction velocity of around 1 m/s. This happens with similar slopes for PM2.5 fluxes and for particle number fluxes measured with different instruments over the different surfaces. This means that the average ratio Vd/u* tends to collapse towards a constant value even if the absolute values of fluxes and concentrations are significantly different. This probably is due to the fact that the friction velocity carries most of the information regarding the surface effects. Only limited effect of stability is observed with a slight reduction of the deposition velocities at fixed u* in stable conditions. Considering together all the datasets, an average value of Vd/u* equal to 0.010±0.0017 (median value 0.0062±0.0015) represents the cases for all stabilities. This value is reduced to 0.0097±0.002 (median value 0.005±0.001) considering cases with L > 0. This could be a relatively simple parameterisation to be used in transport and dispersion modelling for simulations over different surfaces.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
The authors wish to thank Ing. F. M. Grasso (ISAC-CNR) and Ing. C. Elefante (University of Salento) for their help in performing the measurements. Further, the authors wish to thank Dr. F. Belosi (ISAC-CNR) for the useful discussion in the interpretation of results.
SeinfeldJ. H.PandisS. N.2006New York, NY, USAJohn Wiley & SonsSolomonS.QinD.ManningM.ChenZ.MarquisM.AverytK. B.TigoraM.MillerH. L.2007Cambridge, UKCambridge University PressIsaksenI. S. A.GranierC.MyhreG.BerntsenT. K.DalsørenS. B.GaussM.KlimontZ.BenestadR.BousquetP.CollinsW.CoxT.EyringV.FowlerD.FuzziS.JöckelP.LajP.LohmannU.MaioneM.MonksP.PrevotA. S. H.RaesF.RichterA.RognerudB.SchulzM.ShindellD.StevensonD. S.StorelvmoT.WangW.-C.van WeeleM.WildM.WuebblesD.Atmospheric composition change: climate-chemistry interactions20094333513851922-s2.0-7035003195210.1016/j.atmosenv.2009.08.003MalmW. C.PotterD. T.ColmanB. R.Fundamentals of visibility2003Hoboken, NJ, USAJohn Wiley & Sons285329DominiciF.PengR. D.BellM. L.PhamL.McDermottA.ZegerS. L.SametJ. M.Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases200629510112711342-s2.0-3364480743710.1001/jama.295.10.1127BrookR. D.RajagopalanS.PopeC. A.BrookJ. R.BhatnagarA.Diez-RouxA. V.HolguinF.HongY.LuepkerR. V.MittlemanM. A.PetersA.SiscovickD.SmithS. C.WhitselL.KaufmanJ. D.Particulate matter air pollution and cardiovascular disease: an update to the scientific statement from the American Heart Association201012121233123782-s2.0-7795322921610.1161/CIR.0b013e3181dbece1CorbettJ. J.WinebrakeJ. J.GreenE. H.KasibhatlaP.EyringV.LauerA.Mortality from ship emissions: a global assessment20074124851285182-s2.0-3724902923710.1021/es071686zDelfinoR. J.SioutasC.MalikS.Potential role of ultrafine particles in associations between airborne particle mass and cardiovascular health200511389349462-s2.0-2384449669110.1289/ehp.7938WeselyM. L.HicksB. B.A review of the current status of knowledge on dry deposition20003412–14226122822-s2.0-003410247910.1016/S1352-2310(99)00467-7TextorC.SchulzM.GuibertS.KinneS.BalkanskiY.BauerS.BerntsenT.BerglenT.BoucherO.ChinM.DentenerF.DiehlT.EasterR.FeichterH.FillmoreD.GhanS.GinouxP.GongS.GriniA.HendricksJ.HorowitzL.HuangP.IsaksenI.IversenT.KlosterS.KochD.KirkevågA.KristjanssonJ. E.KrolM.LauerA.LamarqueJ. F.LiuX.MontanaroV.MyhreG.PennerJ.PitariG.ReddyS.SelandØ.StierP.TakemuraT.TieX.Analysis and quantification of the diversities of aerosol life cycles within AeroCom200667177718132-s2.0-33746911201KanakidouM.SeinfeldJ. H.PandisS. N.BarnesI.DentenerF. J.FacchiniM. C.Van DingenenR.ErvensB.NenesA.NielsenC. J.SwietlickiE.PutaudJ. P.BalkanskiY.FuzziS.HorthJ.MoortgatG. K.WinterhalterR.MyhreC. E. L.TsigaridisK.VignatiE.StephanouE. G.WilsonJ.Organic aerosol and global climate modelling: a review200554105311232-s2.0-17644411455GallagherM. W.NemitzE.DorseyJ. R.FowlerD.SuttonM. A.FlynnM.DuyzerJ.Measurements and parameterizations of small aerosol deposition velocities to grassland, arable crops, and forest: influence of surface roughness length on deposition2002107121102-s2.0-17044453966PryorS. C.GallagherM.SieveringH.LarsenS. E.BarthelmieR. J.BirsanF.NemitzE.RinneJ.KulmalaM.GrönholmT.TaipaleR.VesalaT.A review of measurement and modelling results of particle atmosphere-surface exchange200860142752-s2.0-3834918010710.1111/j.1600-0889.2007.00298.xBuzoriusG.RannikÜ.MäkeläJ. M.VesalaT.KulmalaM.Vertical aerosol particle fluxes measured by eddy covariance technique using condensational particle counter1998291-21571712-s2.0-003197607710.1016/S0021-8502(97)00458-8GallagherM. W.BeswickK. M.DuyzerJ.WestrateH.ChoulartonT. W.HummelshøjP.Measurements of aerosol fluxes to speulderforest using a micrometeorological technique19973133593732-s2.0-003061740610.1016/S1352-2310(96)00057-XDuanB.FairallW.ThomsonD. W.Eddy correlation measurements of the dry deposition of particles in wintertime199827642652GrönlundA.NilssonD.KoponenI. K.VirkkulaA.HanssonM. E.Aerosol dry deposition measured with eddy-covariance technique at Wasa and Aboa, Dronning Maud Land, Antarctica2002353553612-s2.0-0036968631NilssonE. D.RannikÜ.Turbulent aerosol fluxes over the Arctic Ocean 1. Dry deposition over sea and pack ice20011062332125321372-s2.0-0035726396ContiniD.DonateoA.BelosiF.GrassoF. M.SantachiaraG.ProdiF.Deposition velocity of ultrafine particles measured with the Eddy-Correlation method over the Nansen Ice Sheet (Antarctica)2010115162-s2.0-7795621621110.1029/2009JD013600D16202DorseyJ. R.NemitzE.GallagherM. W.FowlerD.WilliamsP. I.BowerK. N.BeswickK. M.Direct measurements and parameterisation of aerosol flux, concentration and emission velocity above a city20023657918002-s2.0-003647248910.1016/S1352-2310(01)00526-XMårtenssonE. M.NilssonE. D.BuzoriusG.JohanssonC.Eddy covariance measurements and parameterisation of traffic related particle emissions in an urban environment2006637697852-s2.0-33645813593NemitzE.JimenezJ. L.HuffmanJ. A.UlbrichI. M.CanagaratnaM. R.WorsnopD. R.GuentherA. B.An eddy-covariance system for the measurement of surface/atmosphere exchange fluxes of submicron aerosol chemical species—first application above an urban area20084286366572-s2.0-4214916158310.1080/02786820802227352SchmidtA.KlemmO.Direct determination of highly size-resolved turbulent particle fluxes with the disjunct eddy covariance method and a 12 stage electrical low pressure impactor2008824740574172-s2.0-77956320410JärviL.RannikÜ.MammarellaI.SogachevA.AaltoP. P.KeronenP.SiivolaE.KulmalaM.VesalaT.Annual particle flux observations over a heterogeneous urban area2009920784778562-s2.0-77950387405ContiniD.DonateoA.ElefanteC.GrassoF. M.Analysis of particles and carbon dioxide concentrations and fluxes in an urban area: correlation with traffic rate and local micrometeorology20124625352-s2.0-8305518427310.1016/j.atmosenv.2011.10.039SlinnW. G. N.Predictions for particle deposition to vegetative canopies1982167178517942-s2.0-0020001955PetroffA.MailliatA.AmielhM.AnselmetF.Aerosol dry deposition on vegetative canopies. Part I: review of present knowledge20084216362536532-s2.0-4304911864710.1016/j.atmosenv.2007.09.043PetroffA.ZhangL.Development and validation of a size-resolved particle dry deposition scheme for application in aerosol transport models2010375376910.5194/gmd-3-753-2010LaviA.FarmerD. K.SegreE.MoiseT.RotenbergE.JimenezJ. L.RudichY.Fluxes of fine particles over a semi-arid pine forest: possible effects of a complex terrain20134789069152-s2.0-8487911469010.1080/02786826.2013.800940TodaM.SugitaM.Single level turbulence measurements to determine roughness parameters of complex terrain2003108122-s2.0-0942267371DonateoA.ContiniD.BelosiF.GambaroA.SantachiaraG.CesariD.ProdiF.Characterisation of PM2.5 concentrations and turbulent fluxes on a island of the Venice lagoon using high temporal resolution measurements20122143853982-s2.0-8487461454610.1127/0941-2948/2012/0354CavaD.ContiniD.DonateoA.MartanoP.Analysis of short-term closure of the surface energy balance above short vegetation2008148182932-s2.0-3754905624910.1016/j.agrformet.2007.09.003DonateoA.ContiniD.BelosiF.Real time measurements of PM2.5 concentrations and vertical turbulent fluxes using an optical detector2006407134613602-s2.0-3104444648410.1016/j.atmosenv.2005.10.026ContiniD.DonateoA.CesariD.BelosiF.FranciosoS.Identification and characterisation of local aerosol sources using high temporal resolution measurements2010129170917212-s2.0-7795633711410.1039/c000642dKennyL. C.GussmanR. A.Characterization and modelling of a family of cyclone aerosol preseparators19972846776882-s2.0-003117205210.1016/S0021-8502(96)00455-7ChakrabartiB.FineP. M.DelfinoR.SioutasC.Performance evaluation of the active-flow personal DataRAM PM2.5 mass monitor (Thermo Anderson pDR-1200) designed for continuous personal exposure measurements20043820332933402-s2.0-244242831010.1016/j.atmosenv.2004.03.007HeimM.KasperG.ReischlG. P.GerhartC.Performance of a new commercial electrical mobility spectrometer200438supplement 23142-s2.0-1114421998710.1080/02786820490519252BaronP. A.WillekeK.20012ndJohn Wiley & SonsHindsW. C.19992ndJohn Wiley & SonsMcMillenR. T.An eddy correlation technique with extended applicability to non-simple terrain19884332312452-s2.0-002382351810.1007/BF00128405RannikÜ.VesalaT.Autoregressive filtering versus linear detrending in estimation of fluxes by the eddy covariance method19999122592802-s2.0-003275186810.1023/A:1001840416858MahrtL.Flux sampling errors for aircraft and towers19981524164292-s2.0-11644323234LamaudE.ChapuisA.FontanJ.SerieE.Measurements and parameterization of aerosol dry deposition in a semi-arid area19942815246124712-s2.0-002848630810.1016/1352-2310(94)90397-2WebbE. K.PearmanG. I.LeuningR.Correction of flux measurements for density effects due to heat and water vapour transfer1980106447851002-s2.0-0018923326RannikÜ.VesalaT.KeskinenR.On the damping of temperature fluctuations in a circular tube relevant to the eddy covariance measurement technique19971021112789127942-s2.0-16444362333HorstT. W.A simple formula for attenuation of eddy fluxes measured with first-order-response scalar sensors19978222192332-s2.0-0031464861DamayP. E.MaroD.CoppalleA.LamaudE.ConnanO.HébertD.TalbautM.IrvineM.Size-resolved eddy covariance measurements of fine particle vertical fluxes20094012105010582-s2.0-7044971327910.1016/j.jaerosci.2009.09.010RoupsardP.AmielhM.MaroD.CoppalleA.BrangerH.ConnanO.LaguionieP.HebertD.TalbautM.Measurement in a wind tunnel of dry deposition velocities of submicron aerosol with associated turbulence onto rough and smooth urban surfaces20135512242-s2.0-8486576589410.1016/j.jaerosci.2012.07.006