Personal Sampler for Monitoring of Viable Viruses; Modeling of Indoor Sampling Conditions

Abstract

This article describes the development of a mathematical model for evaluation of particle concentration for indoor air conditions. The results of modeling could be used to predict particle concentration in an ambient air at different distances from the source, taking into account space geometry and ventilation/air conditioning created air streams. Two case studies were then undertaken to determine the particle concentration at hypothetical shopping center for the situations when: (1) the ambient air movement is created by a local ventilation system, and (2) besides the local ventilation, the outdoor air enters the space influencing particle distribution situation due to mixing with ventilation created air streams. The results of modeling were used to evaluate the minimal source concentration of virus containing aerosols measurable by our recently developed personal bioaerosol sampler moving along various routes during certain time intervals.

INTRODUCTION

A new personal sampler has been developed and verified to be a very efficient tool for monitoring of viable airborne bacteria, fungi, and viruses (Agranovski et al. 2002a, 2002b, 2005). The operation principle is based on aspiration of an air sample through peripheral holes located on the upper part of the sampler with following passing of air through porous media submerged into the layer of collecting liquid. As the result of passage through very narrow and tortuous channels inside the media, the air sample is split into very fine bubbles. The particulates are scavenged by these bubbles, and, thus, effectively removed before the effluent air leaves the vessel. The sampler was tested using viruses with different resistances to environmental physical and biological stresses including influenza, measles, mumps, vaccinia, and SARS (Agranovski et al. 2004a, 2004b, 2005), fungi and bacteria (Agranovski et al. 2002a).

During sampler operation, microorganism-containing particles are accumulated in the collecting liquid creating time-related capability of their identification by following microbiological procedures (Agranovski et al. 2006; Borodulin et al. 2006). The minimal detectable microbial concentrations in the collecting media have been mathematically evaluated for both, viable (Borodulin et al. 2006) and non-viable (Agranovski et al. 2006) bioaerosol particles.

Due to recent outbreaks of various infectious diseases, including SARS and Avian influenza, investigations into aerosol propagation for indoor and outdoor ambient air conditions attract substantial attention of scientists with broad range of interests in the areas of aerosol science, epidemiology, public health, and aerobiology. Besides experimental investigations of the bioaerosol behavior, a mathematical modeling is also becoming a very powerful tool capable to predict the bioaerosol propagation as a function of site geometry, direction of air streams, physical parameters of air carrier and physical and microbiological characteristics of airborne microorganisms of interest. However, considering tortuous and complicated geometry of indoor spaces and existence of ventilation and air conditioning created air streams of turbulent nature, accurate modelling of such processes is quite a challenging task based on utilization of fundamental fluid dynamics equations, which could only be solved numerically. A number of such models has been developed and described in details in the literature (see for example: Kuehn 1988; Kern 1989; Kuehn et al. 1991; Yamamoto et al. 1988). A very similar approach is usually employed for modelling of outdoor conditions, especially when detail consideration of rural or urban environment is required (Shlychkov et al. 2006).

Considering that for both conditions (indoor and outdoor) the aerosol propagation occurs in the air streams of turbulent nature, the most effective way to model the process is using of semi-empiric equation of turbulent diffusion earlier developed by Monin and Yaglom (1965).

In our previous work (Borodulin et al. 2006) we developed a mathematical model for performance evaluation of the personal bioaerosol sampler operating at outdoor air monitoring conditions. A hypothetical case of pathogenic viral particle spread across the central square of Novosibirsk, Russia, has been considered. It was found that the reliable results could be obtained even for sampling at remote distances from the source (∼600 m).

The current article evaluates the possibilities of using the sampler for monitoring of microbial aerosols under indoor sampling conditions. We propose a new aerosol propagation model for an indoor environment and apply it for two case studies. The results of calculations are, then, used for performance evaluation of the sampler operating inside the building.

MODEL DEVELOPMENT

The formula for the unbiased estimate of the countable concentration of virus-containing aerosols C _mes , based on the number of live viruses detected in the collecting liquid n, was derived in Borodulin et al. (2006):

where T is the sampling time; τ is the time of decrease of the virus activity by e ≈ 2.72 times; μ (T/τ) is the time related factor; κ is the coefficient taking into account the efficiency of aspiration and capture of virus-containing particles by collecting fluid (0 ≤ κ ≤ 1), Q is the airflow through the sampler. A formula for the dispersion of such estimate for the case of stationary concentration fields of virus-containing aerosols is also given in Borodulin et al. (2006) as:

where σ² is the dispersion of aerosol concentration and τ^(E) is the Euler time scale of turbulence. Using Equation (1), the estimate of the minimal concentration reliably measurable by the sampler was obtained in a form (Borodulin et al. 2006):

Equations (1) and (2) were derived in Borodulin et al. (2006) on the basis of experimentally measured concentrations using an integral transformation of the initial concentration field C(t) with a certain weighting function h(t) (Tikhonov 1982)

The instantaneous value of a countable concentration of virus-containing aerosols C(t) is the function of the time t and coordinates x, y, z of a point of space. As the sampler moves along a selected route, its coordinates are determinate functions of time x = x(t), y = y(t), z = z(t). As the result, Equation (4) with ⟨C[x(t), y(t), z(t), t)]⟩ is averaged over a statistical ensemble presenting the mathematical expectation of the unbiased estimate of the countable concentration of microbial aerosol for the sampler movement along the route specified by a parametric curve x = x(t), y = y(t), z = z(t).

Of practical interest is also the dispersion of the unbiased estimate of the countable bioaerosol concentration σ _mes ² measurable along sampler routes, which can be written in the following form:

In Equation (5), the value ⟨ _mes (t ₁) _mes (t ₂)⟩ presents a two-point correlation function of pulsations of bioaerosol concentration. To preset it, we firstly rewrite it in the form: B ₁₂(t ₁, t ₂) = ⟨ _mes (t ₁) _mes (t ₂)⟩ and then use an expression constructed on the basis of evident physical considerations given below.

First, for t ₁ = t ₂ = t, B ₁₂(t ₁, t ₂) is a single-point correlation function of concentration pulsationsB(ξ). In the case under study, concentration fields of atmospheric pollutants are not stationary. As the result, they require an introduction of a “quasi-stationary” correlation function of concentration pulsations (Byzova 1974):

Quasi-stationarity becomes apparent due to the fact that, along with slow changes of C(t) value along the interval(0, T), some quick pulsations with frequencies in order of 1/τ^(E) impose on the process of the slow concentration change. The argument of time t is responsible for “slow” changes, and ξ is responsible for “quick” changes. This assumption could be used when sampling times T are much larger than the Euler time scale τ^(E) (Monin and Yaglom 1965).

Second, if t ₂ − t ₁ = ξ is being increased, the correlation of concentration pulsations ⟨ _mes (t ₁) _mes (t ₂)⟩ tends to zero. Suppose that the estimate of a typical time at which correlations disappear coincides with τ^(E). Thus, the expression for evaluation of B ₁₂(t ₁, t ₂) can be written as:

In accordance with Equation (6), the expression for dispersion σ _mes ² takes the following form

It should be noted that the above expressions have been derived for the case when the sampler operates at a preset fixed point of space. They would allow us to determine ⟨C _mes ⟩ and σ _mes ² for non-stationary concentration fields of microbial aerosols.

Spatial distribution of velocity and temperature fields was determined using hydrothermodynamics equations in the Bussinesq approximation (Gutman 1969). In the indoor air temperature field T, we can distinguish a component, , reflecting the “basic” temperature stratification depending on the vertical coordinate z,

where Θ is the potential temperature. Equations for components of air velocity can be written in the following form

where u, v, w are mathematical expectations of components of the air velocity vector; ϑ are disturbances of the potential temperature; λ is the flotation parameter; S = is the temperature stratification; I _θ is the power of heat sources; τ _ij are components of the tensor of Reynolds turbulent tensions; τ _Ti are turbulent heat flows; ρ is air density, and p is the pressure disturbances.

It is supposed that τ _ij could be approximated with the gradient-diffusion closure in the form

where x _i = , u _i = ; K, K _H are the coefficients of turbulent exchange and thermal conductivity of “sub-grid” scales. A hypothesis about the isotropicity of small-scale turbulence (Monin and Yaglom 1965) served as a basis for Approximation 11.

The following equation was used to calculate the coefficients of turbulent exchange of sub-grid scales (Smagorinsky 1963):

where Def ² = (D ₁₁ ² + D ₂₂ ² + D ₃₃ ²) + D ₁₂ ² + D ₁₃ ² + D ₂₃ ²; Δ = (Δ x Δ y Δ z)^1/3; k = 0,21; D _ij = are the components of deformation tensor; Ri = λ S/ Def ² is the Richardson number, and Δ x,Δ y,Δ z are the steps of the calculation pattern along coordinate axes.

The above approach was tested and verified by using precise numerical solutions of two and three dimensional models of fluid flow, well described in the literature. The results of verification along with comprehensive discussion are given in Sarmanaev et al. (2003). In particular, the following cases were considered for verification: two and three dimensional fluid flow in caverns; three dimensional flow over an obstacle; stationary convective flow between parallel isothermic plates with different surface temperatures, and so on. The model has also been verified by using experimental results obtained by Davidson (1989) and Gudzovsky and Aksenov (1994). Generally, the agreement between the results was satisfactory for both theoretical and experimental verifications.

The semi-empiric equation of turbulent diffusion (Monin and Yaglom 1965) was used to calculate indoor bioaerosol concentration fields. The components of the tensor of turbulent diffusion coefficients preset in the calculations according to Borodulin (1996) on the basis of an algebraic model for turbulent flows and tensions (Teverovskii and Dmitriev 1988).

CASE STUDIES (METHODS)

As a practical example we considered the inner space of the shopping center (Figure 1a). Axes of Cartesian system of coordinates x, y, and z are directed along the length, the width of the room and upright respectively. There are two terraced halls with two staircases S ₁, S ₂ connecting them. The inserted floor, shown with translucent filling in Figure 1a, covers some part of the first floor and forms a sub-ceiling niche. A narrow band of the farthest segment of the inserted floor rests against the exit door D ₃. There are also pair doorways D ₁ and D ₂ on external end walls of the building. The doorways are shown in Figure 1a with black filling.

FIG. 1 The system of coordinates and the geometry of the room in the model calculations (a). Trajectories of particle movement under “no external wind” conditions at the section y = 9 m (b); trajectories of particle movement at the section y = 9 m in the presence of a pressure differential due to external wind presence (c).

Forced ventilation equipment is built in the ceiling of the second floor. It presents a system of forcing hatches located along the whole width of the room and oriented perpendicularly to the axis x. Air injection is performed over the whole ceiling area with the total flow rate of 17 m³/s. It was supposed that virus-containing aerosol was getting into the room through vent holes and doorways. The interaction between forced aspiration, free flows through the doors and vent holes causes the formation of airflow with a rather complex structure, which is the determinating factor of redistribution of indoor pollutant concentration. Aerosol, as a passive substance, repeats the trajectory of air particles forming an uneven field with the values of pollutant concentration distributed across the room.

Boundary conditions for Equation (10) were formulated as follows. On solid walls (with the exception of vent holes in the ceiling) adhesion conditions were specified as u = v = w = 0. At the places where vent holes are located at z = H, where H is the height of the ceiling, the components of vertical air velocity w = w _H were preset with the value of w _H = 0.02m/s. For the areas occupied by the doorways (D ₁,D ₂, D ₃), the distribution of atmospheric pressure p = p _i ; at (x,y,z) ∈ D _i i = was considered to be known. The components of air velocity at D ₁,D ₂, D ₃ were specified proceeding from conditions

If heat exchange with the environment, due to ventilation and thermal conductivity through the walls, is not taken into account, the indoor and outdoor temperatures could be considered as equal. In this case, we can preset the boundary condition ϑ = 0 for the temperature disturbances for all boundary surfaces. Boundary conditions for concentration ⟨ C ⟩, specifying the zero mass flow on solid surfaces, were preset in the form = 0, where n is the direction of the normal to the surface. In the boundary regions of external air inflow (vent panes and doors), the concentration of virus-containing aerosols was considered to be a known value ⟨ C ⟩ = C ₀. Aerosol was also considered to be inactivated during whirling process. As model calculations considered stationary conditions of bioaerosol spread, the integration of non-stationary Equations (10) by time was performed till a steady-state mode of the flow was obtained. The zero values of air velocity components were taken as initial conditions. The equations were solved with the method of finite differences. In a specified region, a regular grid was constructed with reference nodes (i, j, k) located evenly along the axes x, y, and, in the general case, unevenly along axis z. The digitization of Equations (10) was performed in terms of natural variables “velocity-pressure” using the main and half-integer nodes of the calculation grid. Thus, determined grid structure of the fields allows us to approximate the continuity equation with the second order of accuracy within the same unit cell as well as to construct the schemes with preservation of second moments. Finite difference analogues of initial systems were obtained from an equivalent conservative form of equations.

For time integration we used the implicit splitting method (Marchuk 1977) modified to increase the norm of the transition operator and to enhance the algorithm stability. Curvilinear boundaries were evaluated with the method of fictitious regions (Vabishevich 1991).

The mesh with 128 × 128 × 80 along x, y, and z axis, respectively, was used in our calculations. The time step was chosen to be 2 seconds, which was much shorter compared to selected averaging time and concentration pulsation time. Also, some additional calculations on the mesh 256 × 256 × 160 demonstrated that, compared to the results obtained for 128 × 128 × 80 mesh, the difference was negligible and did not exceed 4–6%. On this basis, using of finer meshes was considered as not reasonable for the model parameters used.

Let us consider two numeric experiments to evaluate the model outcomes: (1) spread of bioaerosol particles at the absence of external wind; (2) spread of bioaerosol particles with presence of wind pressure on the building.

CASE STUDIES (RESULTS)

For the first case, corresponding to “no external wind” ambient air conditions, the pressure parameter was preset as: p ₁ = p ₂ = p ₃ = 0. In this case, the behavior of the inner air circulation is determined by the intensity of forced air aspiration into the room. Microbial aerosol particles shift from the ceiling region to lower levels, accelerate horizontally near the floor surface and move towards the exit doors. The air flow streamlines are illustrated in Figure 1b. As is seen, the trajectories of particles, released from the vent holes located at y = 9 m, pass through the doorways D ₁, D ₂ (initial position of trajectories is shown in the figure with dark points). Space coordinates of trajectories in the nodes of the calculation pattern (x _k , k = ) were obtained according to the following equations:

The initial conditions we preset as y = y ₀ and z = z ₀ considering the point (x ₀, y ₀,z ₀) as a starting one for each calculated trajectory. Figure 1b shows that some streamlines are directed towards the doorwayD ₁, while the majority of trajectories are pointed towards the opposite end with the doorways D ₂ and D ₃ having a larger carrying capacity. Approaching the end sections, the air trajectories leave the calculation region with the largest values of horizontal velocity at the doors reaching the magnitude of 0.5–0.6 m/s.

Figure 2a shows the trajectories of particles projected to the horizontal plane. Their initial positions were specified at regularly located points of the ceiling plane. Intercepting of moving particle trajectories is inadmissible by virtue of the continuity equation (see the last expression of Equation [10]). However, the presence of interceptions is clearly seen in Figure 2b on the horizontal plane projection of particle trajectories. This phenomenon could simply be explained by the volumetric character of the gas flow (the particles are not actually intercepting; it is just visual effect due to particle motion at different heights above zero level with their trajectories projected to the same plane).

FIG. 2 Particles trajectories projected on the horizontal plane under “no external wind” conditions (a), and presence of the wind pressure on the building (b). Isolines of normalized concentration values C _n at the section y = 9 m under “no external wind” conditions (c).

For the second case of external wind presence, we considered a bioaerosol spread influenced by the wind pressure on the building caused by the horizontal gradient of atmospheric pressure. It was supposed that the wind direction was coincided with the positive direction of the x-axis in the accepted system of coordinates. A typical value of pressure differentialΔ p = 1 Pa at moderate wind parameters in the near-ground layer was preset for calculations. In this case, the pressure conditions were preset in the form p ₁ = 0 p ₂ = p ₃ = − Δ p. The ventilation air flow from the ceiling was preset to be the same as for the first case study. The air flow streamlines for this case scenario are illustrated in Figure 1c. As is seen, the trajectories of particles, released from the vent holes located at y = 9 m, pass through the doorways D ₂ and D ₃ only, with no particle escape through D ₁ expected (initial position of trajectories is shown in the figure with dark points).

The gradient of external pressure causes the appearance of a constant component of indoor velocity, u, and an additional transport of outdoor air through the door D ₁. The flow pattern along the vertical section, y = 9 m, is presented in Figure 2b. The longitudinal velocity, in this case, is positive everywhere and the flow lines deviate towards the main flow. On this basis, the rate of the air inflow through the doors D ₁ becomes 0.4 m/s and, correspondingly, the maximum velocity at exit doors increases to 1.5 m/s. As the result, a vortex with closed flow lines is formed behind the ledge and in the space under the staircase. Also, a direct flow is being developed between opposite doorways without considerable deviations from the axis x. Figure 2b demonstrates the projection of trajectories on the horizontal plane. Most trajectories are rectilinear and only at the exit doors peripheral jets turn abruptly, due to the wall influence and presence of pressure forces.

Let us pass on to the analysis of the countable concentration fields of microbial particles. Figure 2c demonstrates the concentration of bioaerosol field, C _n , calculated for “no external wind” conditions. Concentration values were normalized by the concentration in the outdoor air; C _n = ⟨ C ⟩/C ₀. As is clearly seen, for such conditions, the ventilation system creates practically even pollutant distribution with the majority of the sales area experiencing the concentration not strongly deviating from the values C _n = 0.9. Only in the niche under the ceiling a stagnant zone is formed where C _n goes down to 0.5. Noticeable pollutant accumulation could be observed near the exit doors: C _n ≈ 3 and C _n ≈ 2.6 for the doorsD ₁ and D ₂, respectively. This is explained by the convergence of pollutant flow at the exit sections due to limited carrying capacity of the doorways.

Figure 3a presents the concentration field of virus-containing aerosols for “no external wind” conditions in the horizontal section at the height of 1.5 m from the floor level of the 2nd floor (an incut rectangle in the lower right corner corresponds to the same height above the floor level of the 1st floor). Like in Figure 2c, regions with increased pollutant concentration at the doorways and a region with minimal concentrations at stair flights are clearly observed. Figure 3b shows the concentration distribution in the case when the pressure differential takes place. Considering the air enters the building from outside through the door D ₁, then the concentration at the inlet plane can be taken as C _n = 1.0 coinciding with the outdoor value. However, as the distance from the door increases, the indoor concentration decreases due to particle dispersion and dilution with clean indoor air. At the distances exceeding 10 m from D ₁, due to mixing with bioaerosol from the ceiling, the concentration begins to increase and towards the middle of the hall reaches the value of 0.9 averaged by the space. At the same time, the value of C _n considerably increases at the exit doors reaching 3 and 4.5 at the doorways D ₂ and D ₃ respectively (the corresponding values of 2.6 and 3 were obtained for “no external wind” conditions) and decreases to C _n = 0.7 in the sub-stairs dead end. Based on the above results, it can be stated that the wind presence causes more contrast distribution of aerosol particles across the shopping centre considered.

FIG. 3 Isolines of normalized values of virus-containing aerosol concentration C _n at the height of 1.5 m from the floor calculated for “no external wind” conditions (a); and the presence of the wind pressure on the building (b).

Now, let us estimate the number of particles collected by the personal sampler as it moves indoors along the specified routes. Three routes were considered. The first route, L ₁, corresponds to the passage from the door D ₁ to the door D ₂ directly through the staircase S ₁. The second route, L ₂, corresponds to the passage from the door D ₁ to the staircase S ₁. After reaching the lower level, turning left, proceeding towards the staircaseS ₂, moving upstairs and turning right to the building exit through the doorD ₃. And finally, the third route, L ₃, corresponds to the entering through the door D ₂, moving towards the staircase S ₁, going upstairs, turning right and moving towards the staircase S ₂. Then going down the stair S ₂ to leave the shopping center through the door D ₂. The rate of movement along each route was preset equal to V = 0.8 m/s. Based on this assumption, to cover distances associated with routes L ₁, L ₂, and L ₃ the following time intervals are correspondingly required:T ₁ = 70, T ₂ = 80, and T ₃ = 50 seconds.

Taking into account that we consider viral bioaerosols of nano-size range (largest viruses are around 400 nm diameter) which could remain in airborne form for extended time periods, we assume that these particles are weightless. Also, the collection efficiency of the sampler for particles of smaller than 300 nm diameter is above 95% (Agranovski et al. 2002a), thus, the efficiency κ of particle capture by the sampler was considered to be one. The averaged concentration value along the route L _k during time T _k is determined by the formula ⟨C _Lk ⟩ dl where dl is the increment of the way along the routeL _k . The integration was performed along the routing line. The results of calculations of the value ⟨C _Lk ⟩ for “no external wind” conditions and the presence of the wind pressure on the building are presented in Table 1. As is clearly seen, the concentration values obtained along the routes are very similar in magnitude. On the other hand, it is obviously possible to design some routes with the concentration values significantly altering from each other. For example, if the route trajectory would pass through the area below the stair (see Figure 2c) then, the number of collected viral particles ought to be much lower compared to the numbers given in Table 1.

TABLE 1 Values ⟨C _Lk ⟩ in units C ₀ calculated at the routes L _k for the first and the second examples of calculations

	Routes
	L 1	L 2	L 3
“No external wind” conditions	0.77	0.81	0.86
The presence of the wind pressure on the building	0.66	0.83	0.74

In our previous work (Borodulin et al. 2006) we found that τ (time of the virus inactivation by a factor of e ≈ 2.72) was within the range of 0.53 to 1.45 hours for a number of common viruses including measles, mumps, SARS, and influenza. Based on these data, the time of the decrease in the virus activity was preset within this interval at the level of τ ≈ 1 hour. Then the value of the fudge factor μ (T/τ) along the considered routes is becoming approximately μ ≈ 1. Table 2 presents the minimal reliably countable values of virus-containing aerosol concentration C _min in the environment calculated for all three routes for the first and the second examples of calculations. As is seen, the results obtained for C _min are large compared to the maximum possible background concentration of viral aerosol, which could occur in natural environments.

TABLE 2 Minimal reliably countable values of the virus-containing aerosol concentration C _min (per m³) in the environment calculated along the routes L _k for the first and the second examples of calculations

	Routes
	L 1	L 2	L 3
“No external wind” conditions	2.45 · 10^4	2.25 · 10^4	3.86 · 10^4
The presence of the wind pressure on the building	2.12 · 10^4	2.32 · 10^4	3.33 · 10^4

DISCUSSION AND CONCLUSIONS

The presented results allow us to conclude that the use of the sampler moving along the specified routes in the area under examination allows us to obtain reliable values of virus-containing aerosol concentration when these exceed the threshold value C _min≈ 2 · 10⁴–4 · 10⁴ units/m³. It should be noted that such values of virus-containing particle concentration at public places are practically unreal for natural conditions and could practically occur mainly as the result of some purposive spread (for example bioterrorist attack). However, these concentrations could be a case for agricultural facilities (for example, massive infection of birds or animals at animal houses), hospitals, biological production units, waste management facilities, and others.

Let us estimate the dispersion of concentration values obtained by the sampler for the above monitoring conditions. To determine σ _mes ² (T)/σ² according to Equation (2), the value τ^(E) could be taken approximately as 60 seconds; typical for the height z = 1.5 m in the near-ground atmospheric layer (Borodulin et al. 1992), acquiringτ/τ^(E)≈ 60. Taking time of movement along the routes T ≈ 60 seconds, the ratio T/τ ≈ 0.02 and, from Equation (2), σ _mes ² (T)/σ²≈ 0.7. Increasing the indoor sampling time up to several hours, the estimate σ _mes ² (T)/σ² practically tends to zero.

Thus, the calculations for the above monitoring conditions of indoor operation of the personal sampler showed that it can be used for mobile monitoring and detection of virus-containing aerosols performed along specified routes. To achieve the reliable estimates of countable particle concentrations averaged along the sampling routes, relatively high and practically unreal for natural appearance at public places concentrations of virus containing aerosols are required. However, these concentrations could naturally occur at a number of agricultural, health and research facilities. Also, it should be noted, that the sampling time taken as an example for the model execution was far too short for realistic bioaerosol monitoring procedures causing substantial decrease of required minimal bioaerosol concentration. Increasing of sampling time would cause microbial accumulation in the collecting liquid and corresponding decrease of minimal measurable concentration. Also, the worst case scenario microbial decay (for very stress-sensitive SARS virus) was used for the calculations. Much lower measurable concentrations would be required for robust microorganisms (e.g., robust Small Pox virus).

As discussed before, there are two issues which should be carefully considered to obtain reliable sampling results. The first issue is the necessity of introduction of the time related correcting term μ, which takes into account a rate of inactivation of virus in bubbling process. To minimize the influence of this factor, the sampling time period should be decreased. However, shortening of the sampling time leads to a corresponding reduction of collected viral material making the sensitivity of the following analytical procedure more crucial. The second issue is related to a lower detection limit of an analytical technique used for analysis of the collecting liquid. The suggested model helps to evaluate the influence of both factors on the monitoring results for particular geometry of indoor area under investigation and bioaerosol concentration. Also, we must admit that the results of calculations represent some ideal environmental situations and provide general insights into aerosol dispersion patterns in the large indoor space with active ventilation system. On the other hand, the outcomes of this project can be immediately used for modeling of particular sampling sites with minimal modifications required.

The models on aerosol propagation have been developed for indoor air conditions. They take into account the geometry of space and ventilation/air conditioning initiated air streams. Two case studies have been undertaken and the results of calculations were used for evaluation of the possibility of indoor bioaerosol monitoring by the new personal bioaerosol sampler. The main benefit of model utilization is the possibility of identification of strategic sampling points/routes to achieve the most efficient monitoring coverage of the space under investigation.