Elsevier

Atmospheric Environment

Volume 37, Issue 18, June 2003, Pages 2477-2484
Atmospheric Environment

Scavenging of aerosols and their chemical species by rain

https://doi.org/10.1016/S1352-2310(03)00162-6Get rights and content

Abstract

Washout or scavenging coefficients have been widely used to study the wet deposition processes quantitatively. In the present theoretical study, the washout coefficients are computed for the aerosols of diameters in the range of 0.02–10 μm having various densities in accordance with their chemical compositions for heavy rain regime. The theoretical scavenging rates are applied to the observed average particle size distributions of pre-monsoon months of the year 1998 and 1999 for Pune and 1999 for Himalayan regions. The evolution of particle size distributions at different time intervals for the non-hygroscopic particles of CaCO3, MgCO3, Zn and Mn indicates that the inertial impaction mechanism is the dominant one in removing particles of all sizes for the heavy rain regime. The size dependence of aerosols as a function of relative humidity is considered for the estimation of washout coefficients of hygroscopic particles such as NaCl and (NH4)2SO4. The washout coefficients are found to be highly dependent on relative humidity for hygroscopic particles. The rainwater concentrations are predicted as a function of rainfall depth and a comparison is made with the observed rainwater concentrations of sequential samples collected on 27 June 2001 in a single rain event to support the results of this theoretical work. The predicted rainwater concentrations for RH=50% are about two times larger than that for RH=95% in the case of hygroscopic particles.

Introduction

The removal of aerosol particles suspended in the atmosphere by precipitation is one of the major mechanisms for maintaining a balance between the sources and sinks of atmospheric aerosol particles. Atmospheric aerosols may be scavenged by precipitation due to Brownian diffusion, thermo- and diffusio-phoretic forces, inertial impaction, electrical forces and by serving as cloud condensation nuclei (nucleation). However, one or more of them may dominate for various regimes of particle/drop size, and type of the aerosols. Rain scavenging is generally classified as rainout—particles serving as cloud condensation nuclei or undergoing capture by cloud water and as washout—the removal of below-cloud particles by raindrops as they fall. The first in-cloud scavenging process is the activation of hygroscopic aerosols that serves as cloud condensation nuclei depending on the critical super-saturation and particles dry size. The second one is the attachment of the aerosol particles to the cloud droplets due to Brownian and turbulent diffusion. Finally, the removal of the aerosol containing cloud droplets produced by the above two processes by relatively large moving droplets inside the cloud. Below-cloud scavenging process depends on the characteristics of the rain, including raindrop size distributions and chemical nature of the particles and their concentration in the atmosphere. The field measurements of the below-cloud scavenging of atmospheric particulate concentrations during precipitation obtained apparent washout coefficient by taking into account the particle size distribution before and after a rainfall and the duration of the rain (Radke et al., 1980). However, raindrops falling through the cloud of atmospheric aerosols constituted of various metals, salts and chemical elements, collect the aerosol particles present in the cross-sectional area of their path. However, for the field measurements of the below-cloud scavenging, it is difficult to get particles size distribution and their chemical composition before and after a rainfall. Further, the rainwater chemistry field study data have not been supplemented with aerosol size distribution and their chemical composition, raindrop size spectra and the relative humidity profiles. The rainwater concentration at the ground is the quantity measured in rainwater chemistry field experiments. Therefore, it is advantageous to have scavenging rates for the particles of different chemical species and rainwater concentrations by theoretical modelling. Garcia et al. (1994) obtained the scavenging efficiencies of aerosols from the collision efficiency, terminal velocities, raindrop size distribution and particle size distribution using Slinn's (1983) model. Their results are applied to three-particle size distribution, characteristic of different atmospheric environments (clean, hazy and urban) and for different raindrop size distributions. They have not taken into account the collision efficiency of particles of various chemical compositions in computations of scavenging efficiency. The present theoretical model considers additional features such as hygroscopic aerosols which change sizes with relative humidity, non-hygroscopic aerosols of selected chemical species and output consisting of the prediction of the rainwater concentration as a function of rainfall amount. The theory describing the various scavenging mechanisms generally assumes flow around a rigid spherical droplet capturing spherical particles. The chemical nature of aerosol particles determines their inertia and thus their collection by raindrops especially by the process of inertial impaction should depend on the actual size of these aerosols attributed to their density. Grover and Beard (1975) and McGann and Jennings (1991) theoretically studied the effect of particle density on the collection efficiency of drops. However, the results of these theoretical models are not applicable to raindrops >1-mm diameter due to their deformed shape. Theoretical computations of washout coefficients reported by Pranesha and Kamra (1997) used the experimentally measured values of collection efficiencies under the controlled condition. They have computed washout coefficients for particles in coarse mode and raindrops >3.6 mm. The collection efficiency [E(Dp,D)] varies mainly with the size (Dp) of the aerosol particles and their chemical nature (Chate and Kamra, 1997). Using theoretical model of Slinn (1983) we have estimated the washout coefficients for various size particles of selected chemical constituents [NaCl, (NH4)2SO4, CaCO3, MgCO3 or CaCO3.MgCO3, and compounds of Mn and Zn]. For the sake of simplification in the calculation of collection efficiencies, it is assumed that the particles in the atmosphere are admixtures of these compounds. The present computations consider only below-cloud scavenging and hence the concentrations of the particles of these compounds in the raindrops below the cloud base are assumed to be almost negligible. These results are then applied to the average particle size distributions at Pune (semi-urban) and Himalayan (clean) regions to obtain the evolution of particle size distribution with the scavenging rates of heavy rain regime. Condensation of moisture on the particles, thereby increasing the particle size is another mechanism, which could possibly affect the particle removal below the cloud base during rain event. We have theoretically analysed the differences to be anticipated between non-hygroscopic and hygroscopic particles’ washout by rain at a high relative humidity (95%) using Gerber's (1988) parameterizations. Predicted rainwater concentrations as a function of rainfall depth are compared with the observed rainwater concentrations obtained from sequential rainwater samples. Mass size distribution of aerosols was measured twice a month for a period of 100 h each of the pre-monsoon months (February–May of 1998 and 1999) at the Pune and at the Himalayan region (May of 1999) using a 9-stage Andersen Sampler Mark-II (Andersen Inc., USA). Depositions of aerosols at all stages were collected on Whatman 41-filter paper (10-cm diameter) at a constant flow rate of 28.3 l min−1. The sampler was installed on the terrace of the Institute building at a height of about 12 m above ground level. The atmospheric aerosol samples obtained from the above location are representative of natural condition since the observational site is free from obstructions and sources of local pollution. The water-soluble extracts were analysed for Cl, SO4, NO3, NH4, Na, K, Ca, and Mg and the acid soluble extracts for Al, Fe, Mn, Cu and Zn. The chemical analyses were performed with Perkin-Elmer 373, double beam Atomic Absorption Spectrophotometer with air–acetylene flame. The detection limits for Na, Ca, Mg, Mn and Zn is 0.0002, 0.0005, 0.0001, 0.002 and 0.001 μm ml−1. The concentrations of SO4 and NH4 were determined with colorimeteric methods using UV/visible spectrophotometer. The detection limits for SO4 and NH4 are 0.1 and 0.01 μg ml−1, respectively. The analytical errors were nominal and varied within ±10%.

The parameters for lognormal distributions are given in Table 1, for the average mass size distributions at Pune and Himalayan region. A log probability graph is plotted with per cent less than the indicated size for cumulative mass distributions on the X-axis and particle diameter on the Y-axis for the observed data of Andersen sampler from Pune and Himalayan region. Mass median diameter (MMD=d50%) is obtained from log probability curve for both fine and coarse mode particle sizes. The geometric standard deviation is obtained byσg=d50%d16%=d84%d50%.The count mean diameter (CMD) and diameter of the particle with average mass size particle (dm̄) are determined from MMD and σg. The mass median diameters of the particle size distributions have been corrected by dividing them with the square root of density of aerosol material, i.e., 1.54, 1.74, 1.77, 2.16, 7.14 and 7.4 g cm−3 for the aerosols particles of CaCO3, MgCO3 or CaCO3.MgCO3, (NH4)2SO4, NaCl, Zn and Mn, respectively, to get their true size (Whitby, 1978). Pune is about 100 km away from the West Coast. The wind in the lower troposphere is an easterly flow from October to May and continental air masses, rich in particles of continental origin pass over the region. The air masses at Himalayan region are mostly of continental origin in the month of May; therefore, evolution of particle size distributions at both the places shows similar features.

Rainwater samples of constant volume (100 ml) amounting 0.63 mm rainfall, were collected at Pune at the terrace of the Institute building at a height of about 15 m above ground by using a stainless steel funnel of 45 cm diameter fitted on 1 l capacity polyethylene bottle which was previously rinsed with triple distilled water. The duration of each sample was recorded. The sample was transferred to another polyethylene bottle of 100-ml capacity. The procedure was repeated for every 100-ml rain sample until rain ceased (Naik et al., 1994). The single rain event of 27 June 2001 was assumed to be a convective shower. The average rainfall intensity for this rain episode was around 25 mm h−1 and did not vary much as the rain ceased within a short period. The relative humidity was found >95% throughout the sampling period. The rainwater samples were filtered through Whatman 41 filter papers and refrigerated at 4°C in the laboratory until all the major ionic components were analysed. The concentrations of Na+, Ca2+, and Mg2+ were determined by using a Perkin-Elmer (Model 373), Atomic Absorption Spectrophotometer with air–acetylene flame. The detection limits for Na+, Ca2+, and Mg2+ are 0.09, 0.125, and 0.04 μ mol l−1, respectively. The concentration of SO42− was determined by barium iodate method (Klockow and Ronicke, 1973). The detection limits for SO42− was 1 μ mol l−1.

Section snippets

Theory

Prediction of collision efficiency by solving the Navier–Stokes equations followed by solution of the particle's equation of motion in the flow field of drops of deformed shape is an extremely difficult undertaking. For a deformed drop, an alternative is to rely on dimensional analysis coupled with experimental data to compute collision efficiency. By considering the three most important scavenging mechanisms such as Brownian diffusion, inertial impaction and interception, a relationship

Results and discussion

We have obtained at different intervals the evolution of mass size distributions, in the semi-urban (Pune) and clean (Himalayan) environments, due to below-cloud scavenging of aerosol particles for heavy rain regime. Fig. 1 illustrates the evolution of Particle Size Distribution (PSD) at an interval of 36 and 60 min with a heavy rain of 25 mm h−1 (represents convective cloud characteristic). The evolution of PSD for Pune and Himalayan regions indicates similar trends as shown in Fig. 1, as their

Conclusions

Our results show that, aerosol mass removal by rain in below-cloud scavenging is efficient for particles of aerodynamic diameter (ρp=1.0 g cm−3) in the coarse mode at Pune and Himalayan regions. The evolutions of particle size distribution in the case of total suspended particulate are in good agreement with the earlier theoretical results of Garcia et al. (1994).

The contribution of inertial impaction mechanism is the dominant one in removing the aerosols by below-cloud scavenging for the smaller

Acknowledgements

The authors thank the Director, Indian Institute of Tropical Meteorology, Pune for giving the support and encouragement to conduct the field observations in the Himalayan region. Thanks are also due to the scientists of G.B. Pant Institute of Himalayan Environment and Development, Kullu for their help in carrying out these observations.

References (16)

There are more references available in the full text version of this article.

Cited by (115)

View all citing articles on Scopus
View full text