Assessment of Atmospheric Particles Flux Variation on the Different Underlying Surfaces (Grasslands and Forest) in the Lake Baikal Region

Abstract

In this study, the new data of experimental studies of the atmospheric particulate matter (PM) on the south-eastern coast of Lake Baikal (station Boyarsky) were analyzed in summer 2021. High-altitude measuring sites were arranged in the forest massif (mast, 16 m) and above the meadow vegetation (mast, 30 m). By the Giardina M. model and based on the measurements data the calculations of the deposition flux density of aerosol particles on forest and meadow vegetation were made. Our preliminary results of prediction obtained by Giardina M. model good agrees with measured dry deposition velocities across particle sizes. In the forest, the mass concentration of aerosol particles differs slightly from the mass concentrations in the grasslands and is equal on average 7.9 × 10⁻³ mg m⁻³ for the size particles below 200 nm (PM_0.2) and 6.7 × 10⁻⁴ mg m⁻³ for particles in the size range from 0.2 to 10 μm (PM_0.2–10). However, we found that mass flux density of aerosol particle is almost 4.8 times higher under forest canopy than in meadow vegetation. In addition, the leaf area index (LAI), which characterize the effective area of particle deposition, is also significantly higher in the tree canopy (5.6) compared to the grassland vegetation (2.4).

1. Introduction

Aerosol particles are among the key climate forcing factors and public health risks. The mechanisms and factors determining the processes of particle formation and their further transformation in the atmosphere are still not entirely clear [1,2]. In spite of the increased interest in experimental investigations of atmospheric particles in various regions, information on their spatial and temporal dynamics in the atmosphere is still very scarce. Such studies in the Lake Baikal region are particularly important and relevant, both in terms of assessing the state of the air environment over a unique natural object—Lake Baikal, and in general, for understanding the mechanisms of aerosol interaction with atmospheric processes, both in the drive layer and in the surface atmosphere, where particles are formed from different sources [3].

One of the major sources of ambient aerosol and gas emissions and simultaneously sinks for atmospheric impurities, including anthropogenic substances, are forest and grassland ecosystems. Without taking into account the natural sources of emissions/drainage of atmospheric impurities it is impossible to give an objective assessment of the state and forecast of atmospheric air quality in the Lake Baikal region. The systems of estimation and forecasting of emissions and especially flows of aerosol and gaseous impurities are currently based mainly on the methods of mathematical modeling, in particular [1,2,3].

Despite the fact that studies of atmospheric particulate matter (PM) are actively conducted in various regions of the world, there are still many problems that need to be solved in order to understand the processes that determine their variability. One of the main problems is to identify the features of the chemical and physical mechanisms of formation, transformation, and transfer of the fine aerosol fraction in various natural and climatic conditions, especially in those regions where such studies have not been conducted before. In this regard, the Baikal region is of particular interest as an ideal location characterized by a unique natural environment. Simultaneously, periods of anthropogenic impact are observed in the Baikal region due to the orographic isolation of the Baikal basin and the specific nature of air flow circulation over the lake. Forest and meadow ecosystems are one of the major sources of emissions of natural aerosols and gases and at the same time are runoffs for atmospheric impurities, including anthropogenic substances. Without taking into account natural sources of emission/runoff of atmospheric impurities, it is impossible to give an objective assessment of the state and forecast of atmospheric air quality in the Baikal region.

Current knowledge of fluctuations in aerosol parameters, including count and mass concentrations, is fragmentary in many respects.

The patterns of turbulent fluctuations in the components of wind velocity and air temperature in the surface and near-water layers cannot be considered fully studied, especially in the border zone of «water-land». The current situation determines the need to set up new special experiments to study fluctuations in atmospheric aerosol parameters, together with turbulent pulsations of wind velocity and air temperature components and their relationship with convective and advective processes in the atmosphere.

Such complex special experiments aimed at studying the spatial and temporal variability of atmospheric aerosols in interaction with atmospheric processes on Lake Baikal have not been held before.

From the increasing anthropogenic load, extreme natural phenomena such as wildfires an assessment of environmental damage to the ecosystem of Lake Baikal still remains an unsolved problem due to the lack of observational data.

There are insufficient data on observations of the composition of chemically active gas impurities, aerosols, and the inclusion of various factors affecting the composition and quality of atmospheric air, including runoff/emission of impurities in natural ecosystems (forest, soil, water).

A large number of models can be found in the literature [1,4,5,6,7,8,9,10]; however, none of them are able to deal exhaustively with most of the phenomenological phenomena associated with the deposition of contaminants due to the many complex processes involved. With regard to dry deposition of particles, the proposed calculation models for estimating dry deposition velocities and fluxes are far from complete due to the complex dependence of deposition on particle size, density, terrain, vegetation, meteorological conditions, and chemical species of aerosol particles. Giardina et al. [11] proposed a new model which is based on an electrical analogy to calculate the deposition velocity of particles to modify the laminar sublayer resistivity parameterization. New relations were obtained supposing that resistances which affect particle flow in quasilaminar sublayer can be combined taking into account local features of mutual in-fluence of inertial and turbulent interaction processes of particles.

A comparative study, which was carried out using significant experimental material [12,13,14,15,16,17], makes it possible to verify the correctness of the proposed approach. In this paper, the proposed approach is applied to quantify the deposition fluxes of atmospheric particulate matter on different types of underlying surface in Lake Baikal region. Based on our experimental data and Giardina et al.’s (2018) model [11] we calculated the deposition flux density of atmospheric particulate matter (PM0.2 and PM0.2–10 particles) on the underlying surface with forest and meadow vegetation.

The focus of this paper is to present preliminary data of deposition of atmospheric particulate matter in terrestrial ecosystems in the Baikal region and to make an attempt to apply the accumulated experience from such studies to other regions of the world.

2. Materials and Methods

2.1. Experimental Conditions and Measurement Methods

To investigate spatial and temporal peculiarities of aerosol admixtures transfer and distribution processes, in order to estimate particle deposition fluxes on the underlying surface, high-altitude measuring sites were arranged in the forest massif and above the soil vegetation in the coastal zone of Lake Baikal: (1) mast measurements (30 m) above the meadow vegetation, (2) mast measurements (16 m) in the deciduous forest. Measurements were taken from 27 July 2021 to 27 August 2021. The observation point is located on the territory of the research station “Boyarsky” at a distance of 500 m from the Baikal water surface and at a height of 60 m from the lake level. Station “Boyarsky” is remote from industrial centers and located near the coastal zone of Lake Baikal in a relatively clean area, close to “background” conditions. The masts are installed on concrete platforms. Figure 1 demonstrates the study area and measurement scheme. The forest, mixed on average, is 40–50 years old and consists of birch, poplar, larch, pine and cedar. At the measuring tower, mainly birch trees grow, with a tree crown height of h = 12 m with a peak leaf area index—LAI = 5.6 during the measurement period. LAI was determined by the results of the digital processing of hemispherical photos [18,19]. Air was continuously sampled at the height of 2 m on mast № 1, and at the height of 16 m on mast № 2, to determine counting concentration of particles in the range from 5 nm to 10 μm using diffusion aerosol spectrometer DAS 2702 M developed by company “AeroNanoTech” (Moscow) [20]. The limits of acceptable relative error of measurements of aerosol particle sizes are ±15%. Time of the complete measurement cycle is from 1 to 3 min. The spectrometer has 2 modes of operation: the mode of measurement of parameters of aerosol particles in the size range from 0.005 to 0.2 μm (40 ranges with a step of 5 nm) and the mode of measurement of parameters of aerosol particles in the size range from 0.2 to 10 μm (12 channels). In addition, measurements of meteorological and turbulent atmospheric parameters were carried out using ultrasonic weather stations EXMETEO, AMK-2 [21].

Based on the results of studies of daily variability of the total concentration of fine particulate matter with the size below 0.2 μm (PM_0.2) and the total concentration of particles in the size range from 0.2 to 10 μm (PM_0.2–10) at sites with differing types of underlying surface, calculations of the deposition flux density of atmospheric particulate matter on the underlying surface with forest and meadow vegetation were made.

2.2. Deposition Velocity Calculations

The fundamental concept under investigation in the study of the dry deposition process is the deposition velocity (V_d), expressed in meters per second (m s⁻¹). The deposition velocity at a specific height z plays a crucial role in establishing a relationship between the vertical flux of pollutants and the measured concentration of pollutants at height z in comparison to the concentration at ground level:

V_d = F/C(z)

(1)

where F (g (m² s)⁻³)—is the flux of pollutant removed from a unit area and C(z) (g m⁻³) is the concentration of pollutant at height z.

According to Stokes’ law, the gravitational settling velocity (Vs) of particles increases proportionally to the square of the particle diameter (dp). This law is applicable to particles with diameters up to 50 µm:

V_s = (d_p² × g × (p_p − p_a) × C_c)/18 × μ_a

(2)

where g is gravitational acceleration; p_p is particle density; p_a is air density; μ_a is dynamic viscosity of air; and C_c is Cunningham factor [22]. The C_c parameter can be represented as follows:

C_c = 1 + λₐ/dp × (2.514 + 0.8 × e ^{(−0.55 × dp)/λₐ})

(3)

where λₐ is the free path length of the molecule.

As highlighted in the work by Gallagher et al. [16], the dry deposition rate of particles, when estimated using the electrical analogy approach described earlier, exhibits inconsistencies with the principles of the mass conservation equation. To model the vertical transport of particles, it is proposed that turbulent transport and particle settling can be combined and summarized as follows, as suggested by Venkatram et al. [23]:

K_p × dC/dz + V_s × C = F

(4)

where Kp is the vortex diffusivity for mass transfer of species C and V_s is the settling velocity determined from Equation (2). Through the process of integration of the aforementioned equation, an expression for the settling velocity can be derived as follows:

V_d = Vs/(1 − e^{−(r(z) × Vs})

(5)

where r(z) is the total transport resistance, (r) can be computed as a function of the particle diameter (dp) and height (z). The novel model, proposed by Giardina et al. [11], is founded upon an electrical analogy. According to this model, the total resistance can be estimated using the following equation:

r(z) = r_a + r_ql

(6)

The resistance (r_a) can be ascertained by applying the Monin–Obukhov similarity theory, which establishes the dependence, as described by Seinfeld et al. [20], Venkatram et al. [23], Csanady [24] and Wesely et al. [25]:

r_a = 1/ku* × (ln((z − d)/z₀)) − ψ_h

(7)

z₀ = h (0.215 − LAI^0.25/10)

(8)

d = h (0.1 + LAI^0.2/2)

(9)

where u* is the friction velocity, which is the intensity of atmospheric turbulence; z₀ is the surface roughness height above the displacement plane; d is the height of displacement from the zero plane; and k is the von Karman constant (usually equal to 0.4). Roughness length (z₀) and displacement height (d) are calculated as a function of plant dome height (forest, grass) and LAI—leaf area index.

The following relations for the calculation of the parameter Ψ_h in Equation (7) were proposed in [3]:

ψ_h = −5 × (z − d)/L for (z − d)/L > 0 (stable atmospheric conditions)

(10)

ψ_h = e^{(0.5998 + 0.390 × Ln ((z − d)/L) − 0.09 × Ln (−((z − d)/L) × (z − d)/L))} for (z − d)/L < 0

(11)

where L is the Monin–Obukhov scale.

The resistance of the quasi-laminar sublayer r_ql is estimated by considering a parallel resistance chain r_db, r_ii and r_ti. Here, resistance r_db takes into account contribution of Brownian diffusion; resistance r_ii takes into account effect of inertial compaction for large particles; resistance r_ti takes into account effect of inertial compaction due to turbulent impact. Based on the aforementioned hypotheses, the resistances r_i (associated with inertial impact processes) and r_ti (related to turbulent impact) are interconnected in series to consider specific local characteristics of their mutual influence. Under these assumptions, the estimated resistance r_ql is given by the following equation:

1/r_ql = 1/(r_db + 1/r_ii + 1/(r_ii + r_ti))

(12)

In the scientific literature, different models propose a functional relationship between the resistance associated with Brownian diffusion phenomena (r_db) and the Schmidt number (S_c). The general expression for r_db can be represented as follows:

r_db = 1/u* × C(S_c)^p

(13)

where C and p are constants and S_c is the Schmidt number, which is estimated as follows:

S_c = v_a/D

(14)

where ν_a is the kinematic viscosity of air (m² s⁻¹) and D is the Brownian diffusion of a particle. The Brownian diffusion of air (m² s⁻¹⁾ determined from the Stokes–Einstein equation:

D = (K_b × T × C_c)/(3 × π × µ_a × d_p)

(15)

with K_b being Boltzmann constant (J K⁻¹⁾; T being absolute temperature; μ_a being the dynamic viscosity of air; and C_c being the Cunningham factor from Equation (3). The parameter p in Equation (13) typically falls within the range of 1/2 to 2/3, with higher values observed for rougher surfaces. For instance, Slinn et al. [2] proposed a value of 1/2 for water surfaces, while Peters et al. [26] suggested a value of 2/3 for vegetated surfaces. In the study by Zhang et al. [1], varying values of p were employed based on land-use categories. However, in the current study, a uniform relationship is adopted for all surface conditions, and it is represented as follows:

r_db = 1/u* × C(S_c)^−2/3

(16)

The transport of particles through Brownian diffusion, as described in Equation (16), has been widely recommended in various studies based on both theoretical and empirical findings [25,27,28]. The resistance associated with the inertial interaction process (r_ii) is estimated using a ratio that is applicable to both smooth and rough surfaces:

r_db = 1/u* × (S²_t/(S²_t + 400)) (for smooth surfaces)

(17)

r_ii = 1/u* × (S²_t/(S²_t + 1)) (for rough surfaces)

(18)

where S_t is the Stokes number, defined as follows:

S_t = (v_s × u*²)/g × v_a

(19)

The settling velocity (v_s) is determined from Equation (2). Several authors have suggested these or similar equations for compaction efficiency as a function of smooth and rough surfaces [26]. For instance, Zhang et al. [1] (2001) proposed a compaction efficiency that varies with the land use. The resistance associated with turbulence phenomena (r_ti) is predicted as a function of the dimensionless particle relaxation time, τ+, in the following manner:

r_ti = 1/(u* × m × τ₊ⁿ) (for all surfaces)

(20)

where τ₊ is estimated using the following relation:

τ₊ = τ × (u*²/v_a)

(21)

τ = (d_p² × p_p × C_c)/(18 × μ_a)

(22)

The constants m and n in Equation (20) were determined by fitting the experimental data from [29,30] to Equations (5), (12), (16), (17) and (20). The authors of these studies conducted experiments to investigate the deposition of aerosol particles under conditions with comparable values of friction velocities. By analyzing the experimental data and adapting the mentioned equations, they established suitable values for the constants m and n to achieve a good fit between the theoretical predictions and the observed data. In the light of the above considerations, Equation (5) can be written as follows:

V_d = v_s/(1 − e^{−vs × ((ra + 1)/(1/rdb + 1/rii + 1/(rii + rti))})

(23)

where r_a, r_db and r_ti are estimated using Equations (7), (13) and (20), which use constants m = 0.1 and n = 3. The parameter r_ii is estimated using Equation (17) or Equation (18) for smooth or rough surfaces, respectively.

3. Results and Discussions

The atmospheric particles flux variation is influenced by local meteorological parameters, affecting its residence time and chemical behavior in the lower atmosphere [31,32]. The complex interactions between PM deposition and meteorological parameters occur on both spatial and temporal scales. Many studies have highlighted the significant impacts of meteorological factors such as relative humidity, air temperature, wind speed, and precipitation [33,34,35,36,37,38,39].

Based on the equations from the model [11], we performed calculations of the flux density of the count concentration of particles and their mass concentration on meadow and forest vegetation during the experimental period. Meteorological characteristics of the atmosphere (air temperature (T), pressure (P) and wind speed (U)) and turbulence characteristics (Monin–Obukhov scale (L), surface friction velocity (u*)) from high-frequency acoustic meteorological complexes AMK were used as input parameters. Particle deposition velocities (v_di) were calculated for each size range (di) using Equations (7)–(23). Particle flux density (FN), mass flux density (FM) for each fraction (PM_0.2, PM_0.2–10) were determined by the following equations:

FN_PM0.2 = ∑⁴⁰_i=1 (v_di × C_1ni) FN_PM0.2–10 = ∑¹¹_j=1 (v_di × C_2ni)

(24)

FN_PM0.2 = ∑⁴⁰_i=1 (v_di × C_1mi) FN_PM0.2–10 = ∑¹¹_j=1 (v_di × C_2mi)

(25)

C_1mi = p_p × π/6 × d³_pi × C_1ni C_2mi = p_p × π/6 × d³_pj × C_2ni

(26)

where С_1ni, С2_nj are the counted concentrations of PM_0.2 and PM_0.2–10 particles, and С_1mi, С_2mj are the mass concentrations of PM_0.2 and PM_0.2–10 particles, respectively.

The calculations have been performed based on the VBA (Visual Basic for Application) macro programming language for working with spreadsheets in Microsoft Excel.

For calculations of particle dry deposition rates and fluxes on meadow and forest vegetation we chose days with approximately the same meteorological conditions for situations with the presence of smoke aerosol (11 August 2021) and with conditions close to background after rainy weather, which contributed to atmospheric washout (22 August 2021, 26 August 2021). The input data for calculations of dry deposition of aerosol particles are presented in

Table 1. Input data for calculating the flux and rate of dry deposition of particles on the underlying surface (grassland, forest).

Parameter	Unit	Parameter Values
		Grassland	Forest
Vegetation height (h)	m	0.3	12
Leaf area index (LAI)	-	2.4	5.6
Flat displacement (d)	m	0.03	9.7
Roughness length (z0)	m	0.2	0.7
Particle density (ρp)	kg m−3	1000	1000

.

Under smog conditions (smoke from forest fires) calculations of particle velocity and dry deposition fluxes were performed for 11 August 2021 (00:00–08:00) in forest environment and 11 August 2021 (12:30–24:00) for meadow vegetation.

Figure 2 shows the results of calculation of daily variability of counted and mass concentration, deposition fluxes (mass concentration) of PM_0.2 and PM_0.2–10 particles on 11 August 2021 (forest) and (grasslands). As can be seen from Figure 2a,b, abnormally high concentrations of smoke aerosol were observed on this day. As can be seen from Figure 2, high concentrations of atmospheric particulate matter, especially PM_0.2–10, are observed in smog conditions formed by large-scale forest fires.

Under conditions of smoke absence, calculations were made for the measurement 22 August 2021 and 26 August 2021.

Meteorological conditions have a significant effect on particle dispersion and deposition. In order to compare deposition velocities and fluxes of particles at different sites (grassland), (forest) days with synoptic and meteorological conditions close to each other were considered.

Table 2. Average meteorological parameters for calculating the flux and rate of dry deposition of particles on the underlying surface (22 August—grassland, 26 August—forest).

Parameter	Unit	Date
		22 August	26 August
Temperature	°C	12 ± 1.5	12.5 ± 1.4
Wind speed	m s−1	8.2 ± 3	8.1 ± 2.2
Relative humidity	%	90 ± 5	82 ± 8
Wind direction	deg	249.1 ± 5	248.4 ± 5

shows the average meteorological parameters for the measurement period of 22 August 2021 (grassland) and 26 August 2021 (forest).

Figure 3 shows results of calculation of daily variability of counted and mass concentration, deposition flux density (mass concentration) of PM_0.2 and PM_0.2–10 particles for the observation period of 22 August 2021 (grasslands) and 26 August 2021 (forest). On the grasslands, the mass concentration of aerosol particles varies in a range from 1.2 × 10⁻³ mg m⁻³ to 17 × 10⁻³ mg m⁻³ for particles PM_0.2–10 and in a range 1.1 × 10⁻⁴ mg m⁻³ to 9.1 × 10⁻⁴ mg m⁻³ for particles PM₁₀ with average values of 8.7 × 10⁻³ mg m⁻³ and 4.1 × 10⁻⁴ mg m⁻³, respectively. In the forest, the mass concentration differs slightly from the mass concentrations in the grasslands and is equal, on average, to 7.9 × 10⁻³ mg m⁻³ for PM_0.2–10 and 6.7 × 10⁻⁴ mg m⁻³ for PM₁₀. It should be noted that the mass flux density of aerosol particles (Figure 3e,f) is almost 4.8 times higher under the forest canopy than in the meadow vegetation.

Friction velocity (u*), which depends on wind speed (V) and roughness parameters (z₀, D), with equal wind speeds, is higher over the forest dome than over meadow vegetation and, accordingly, the rate of particle deposition in the forest is greater. In addition, the LAI, which characterizes the effective area of particle deposition, is also significantly higher in the tree canopy compared to the grassland vegetation. As an example, Figure 4 shows the calculated daytime particle deposition rates in the meadow vegetation (u* = 0.6 m s⁻¹, L = −300 for period 14:00–18:00, 22 August 2021) and in the forest (u* = 0.8 m s⁻¹, L = −400 for period 14:00–18:00, 26 August 2021) which indicate higher particle deposition rates in forest. Also, Figure 4 presents a comparison between our predicted deposition velocities (V_d) obtained from the proposed model and the experimental measurements reported in [12,30] for grassland and in [14,17,31] for forest, respectively. The model’s estimations of deposition velocities are plotted alongside the corresponding experimental data for aerosol particle deposition. This graphical representation allows for an assessment of the model’s performance in replicating the observed deposition velocities and provides insights into the model’s accuracy and agreement with experimental findings. Measured aerosol dry deposition velocities from the literature are represented as functions of particle size. Symbols represent the median values and the lines show predictions made by various models.

The size-resolved measurements apply to particles with optical diameters between 5 nm and 10 μm, and are the first ones to be reported above a varying type of underlying surface in the Lake Baikal region. For aerosol particles in the size range 0.005–10 μm, the average value of deposition velocity means of Vd = 0.3 cm s⁻¹ in the grassland and the forest Vd is equal to 1.1 cm s⁻¹.

Deposition velocities are clearly influenced by turbulence, typically characterized by friction velocity (u*), with more turbulent conditions resulting in a stronger flux [14,31,45,46]. The type of land used also affects deposition velocity, with more complex ecosystems having greater surface area, containing more collectors and facilitating more deposition through interception. Consequently, deposition velocities over forests are typically higher than those over grasslands, which are in turn higher than those over lakes or smooth aquatic surfaces. For larger particle sizes (greater than 10 μm in diameter), gravitational settling becomes the dominant factor, causing deposition rates to converge regardless of surface structure.

Confirmation by experimental measurement of the relevance of these Vd parameters remains an issue given the differences that are always found in the literature. For example, the averaged deposition velocity (Vd) in the Yatir forest in southern Israel was 0.38 ± 0.45 cm s⁻¹ for 0.25–0.28 μm particles, which is in agreement with the deposition velocities measured in mid and northern latitude coniferous forests, and is most heavily influenced by the atmospheric stability and turbulence conditions, and to a lesser degree by the particle size [41]. Above a temperate broadleaf forest in central Ontario, Canada by eddy-correlation method the median transfer velocity to size bins (68–292 nm) ranged from +0.13 to −0.27 cm s⁻¹ [42]. For the forest measurement point in Boyarsky, the calculated averaged deposition velocity for the particle concentrations of PM_0.2 and PM_0.2–10 are equal to 0.29 mm s⁻¹ and 5.38 mm s⁻¹, respectively.

Figure 4b compares the models for grasslands [40,41,44] to measurements. Although there are many measurements for grasslands, there seems to be much less consensus among them, even within the same studies, compared to forests. Due to the degree of scatter in the measurements, there is no clear guidance for parameter selection. While the rationale for microscale impaction may also apply to grass, given that grass leaves often have hairs or trichomes and serrated edges, the measurement evidence is less clear. The large scatter among the reported measurements for grasslands is likely due to the variety of grass species, which can have very different characteristics, including length. For example, the measurements reported by Pellerin et al. (2017) were made in grassland in the range of 0.2–0.5 m tall, while Connan et al. (2018) used artificial grass.

The Pleim et al. model generally fits the Giardina et al. model (present work) very well for grassland and the forest. The Pleim et al. model is a modification of the current model in the Community Multiscale Air Quality (CMAQ) model. The key innovation is the addition of a second inertial impaction term for microscale obstacles such as leaf hairs, microscale ridges, and needleleaf edge effects.

The predictions reported in Figure 4, obtained by the Giardina model [11], provide good agreement with the experimental data. There have been limited recent measurements of aerosol flux and subsequent observations of Vd over vegetated areas. This scarcity of data primarily stems from the difficulties associated with acquiring accurate aerosol flux measurements.

Confirmation of the significance of these Vd parameters through experimental measurements remains a challenge due to the discrepancies that are consistently found in the literature.

In

Table 3. Size-resolved particle flux observations, separated by underlying surface (grasslands and forest), method, size range and typical deposition velocity.

Land Use Type	Site Location and Details	Method	Size Range (µm)	Vd (cm s−1)	Reference
Grassland	Present study (research station “Boyarsky”, Lake Baikal, Russia)	Gradient	0.005–10	0.6	Giardina et al., 2018 [11]
	Grass and filter paper	Gradient	0.08–32	0.01–7.2	Chamberlain, 1967 [30]
	Moss (Hypnum cupressiforme) and Italian rye grass	Wind tunnel experiment	0.5	0.024	Clough, 1975 [47]
	Wood River refinery complex, Illinois, USA	Eddy covariance	0.05–0.1	0.6 ± 0.4	Wesely et al., 1977 [25]
	Grass	Gradient	0.05–1	0.525	Garland & Cox., 1982 [48]
	Mount St. Bernard Abbey near Coalville, Leicestershire, England	Gradient	5–30	2.4–7.0	Dollard & Unsworth, 1983 [49]
	Champaign, Illinois, USA	Eddy covariance	0.15–2.5	−0.05–−0.16	Katen & Hubbe, 1985 [50]
	Champaign, Illinois, USA	Eddy covariance	∼0.1–1	0.22 ± 0.06	Wesely et al., 1985 [5]
	South Charleston, Ohio, USA	Eddy covariance	<1	0.4–0.8	Hicks et al., 1986 [51]
	Moorland with Eriophorun and Juncus species, Great Dun Fell, England	Gradient	5–31	0.5–8.9	Gallagher et al., 1988 [52]
	Moorland with Eriophorun and Juncus species, Great Dun Fell, England	Gradient	2–30	2.1–3.9	Fowler et al., 1990 [53]
	Sports fields at the University of Essex, Colchester, England	Gradient	0.1–2	0.10 ± 0.03	Allen et al., 1991 [54]
	Transitional lowland raised bog, Sphagnum species, Auchencorth Moss field site, southeast Scotland	Eddy covariance	0.1–3	0.007–1.2	Nemitz et al., 2002 [55]
	Field of rye grass, Shedd, Oregon, USA	Eddy covariance	0.52	0.16–0.44	Vong et al., 2004 [56]
	Alfalfa (Medicago sativa) field, Southern Great Plains site, Lamont, Oklahoma, USA	Eddy covariance	0.07–0.6	0.03 ± 0.02	Emerson et al., 2018 [41]
	Grass cuttings and synthetic commercial grass	Gradient	0.24–7.8	0.046–2.3	Connan et al., 2018 [44]
Forest	Present study (research station “Boyarsky”, Lake Baikal, Russia)	Gradient	0.005–10	0.8	Giardina et al., 2018 [11]
	Solling forest (spruce and beech trees)	Gradient	0.26–2.4	0.7–1.8	Höfken & Gravenhorst, 1982 [13]
	Spruce forest	Gradient	0.5–10	0.8–1.6	Waraghai & Gravenhorst, 1989 [57]
	Pine plantation	Gradient	0.5–5	0.34–0.92	Lorenz & Murphy, 1989 [58]
	Douglas fir forest, Speulderbos, The Netherlands	Eddy covariance	0.1–3	0.02–11	Gallagher et al., 1997 [59]
	Scots pine forest (SMEAR II station), Hyytiäälä, Finland	Eddy covariance	0.012–1	NA	Buzorius et al., 1998 [60]
	Scots pine forest (SMEAR II station), Hyytiäälä, Finland	REA	0.05	0.43 ± 0.06	Gaman et al., 2004 [61]
	Beech forest (CarboEuroFlux experimental forest site), Sorø, Denmark	Eddy covariance	0.02–0.07	0.15–0.45	Pryor, 2006 [15]
	Scots pine forest (SMEAR II station), Hyytiäälä, Finland	REA	0.008–0.15	0.6–2.1	Grönholm et al., 2007 [43]
	Beech forest, Sorø, Denmark Scots pine forest (SMEAR II station), Hyytiälä, Finland	Eddy covariance and REA	0.01–0.1	0.2–0.5	Pryor et al., 2007 [14]
	Reserva Biológica do Cuieiras, Manaus, Brazil	Eddy covariance	0.01–0.1	NA	Ahlm et al., 2010 [45]
	Scots pine forest (SMEAR II station), Hyytiäälä, Finland	Eddy covariance	0.01–0.06	0.06–0.5	Grönholm et al., 2007 [43]
	Mixed deciduous forest: sugar maple, tulip poplar, sassafras, white oak, and black oak, Morgan-Monroe State Forest, Indiana, USA	Eddy covariance	0.008–0.1	0.06–0.3	Pryor et al., 2009 [27]
	Wet tropical rainforest, Amazonia, Brazil	Eddy covariance	0.25–2.5	NA	Ahlm et al., 2010 [45]
	Ponderosa pine plantation	Eddy covariance	0.25–1.0	0.2–0.6	Vong et al., 2010 [56]
	Mix of hardwood and coniferous trees, Borden Forest Research Station, Ontario, Canada	Eddy covariance	0.018–0.452	0.08–0.6	Gordon et al., 2011 [62]
	Scots pine forest (SMEAR II station), Hyytiäälä, Finland	Eddy covariance	0.01–0.3	0.07–0.4	Mammarella et al., 2011 [63]
	Aleppo pine trees (Yatir Forest Research Station), Israel	Eddy covariance	0.25–0.65	NA	Lavi et al., 2013 [64]
	Laboratory	Wind tunnel experiments	0.5–200	0.9–13	Zhang et al., 2014 [17]

, we have compiled a comprehensive list of dry deposition particle measurements, which includes our studies report on the deposition velocities in the research station “Boyarsky”. In Table 3, detailed information is provided about the measurement size range, method of measurement, and the specific locations where the experimental data for deposition velocities were obtained. Table 3 is categorized based on general land types, such as grassland and forest. Each category includes relevant data for the deposition velocities measured in various environmental settings corresponding to the specified land types. These tabulated data facilitate a systematic comparison and analysis of deposition velocities across different land-use categories, contributing to a better understanding of aerosol particle deposition behavior in various natural environments. The predictions reported in Figure 4, obtained by the Giardina model [11], applied with Equations (7)–(23), respectively, provide good agreement with the experimental data. It was noted that, while in smoke-free conditions the differences between the mass concentrations of PM_0.2 and PM_0.2–10 were of no more than one order of magnitude, in smoke-exposed conditions the particle concentrations of PM_0.2 and PM_0.2–10 differ by two orders of magnitude or more, the depositional flux densities differ by three orders of magnitude or more. Concentrations of particulate matter in forests are higher than in open meadow vegetation and, correspondingly, the densities of particle deposition fluxes are also higher.

4. Conclusions

This study contributed to our understanding of processes deposition of atmospheric particulate matter in terrestrial ecosystems in the Lake Baikal region. For the first time in this geographical region of Lake Baikal, we employed a diffusion aerosol spectrometer Model DAS 2702M to investigate the atmospheric aerosol. The new data of the temporal variations in aerosol distribution based on summer measurements in 2021 in the atmosphere of south-eastern coast of Lake Baikal at the Boyarsky station were received. To investigate the spatial and temporal peculiarities of aerosol admixtures transfer and distribution processes, in order to estimate particles deposition fluxes on the different underlying surface (grasslands and forest), high-altitude measuring sites were arranged in the forest massif and above the soil vegetation in the coastal zone of Lake Baikal: (1) mast measurements (30 m) above the meadow vegetation, (2) mast measurements (16 m) in the deciduous forest. Based on our experimental data and Giardina et al.’s model [11] we calculated the deposition flux density of atmospheric particulate matter (PM_0.2 and PM_0.2–10 particles) on the underlying surface with forest, meadow vegetation. We found that at relatively equal mass concentrations of particles, mass flux density is almost one order of magnitude higher under the forest canopy than in the meadow vegetation. We noticed that the friction velocity (u*), which depends on the wind speed (V) and roughness parameters (z₀, D), with equal wind speeds, is higher over the forest dome than over meadow vegetation and, accordingly, the rate of particle deposition in the forest is greater. In addition, the leaf area index (LAI), which characterized the effective area of particle deposition, is also significantly higher in the tree canopy compared to the grassland vegetation.

Indeed, size-resolved particle flux data over natural surfaces are relatively scarce. However, understanding the process of dry deposition to the surface is of utmost importance since it strongly depends on particle size and significantly influences the mass and number distributions of atmospheric particles. Consequently, this has far-reaching environmental implications, including both direct and indirect climate forcing, impacts on human health, and visibility degradation, especially under conditions of elevated particle concentrations.

Considering the significant environmental consequences, there is a critical need for observational studies that can collect data to aid in the development of accurate dry deposition algorithms. Additionally, these studies can play a crucial role in evaluating mechanistic models concerning size-resolved number fluxes, especially concerning ultrafine particles. Improved understanding of size-dependent deposition patterns will enhance our ability to assess particle transport, deposition, and potential environmental effects, leading to more effective strategies for managing and mitigating the impact of atmospheric particles.