Velici Variation of Aerosol Concentration (2)

8/2/2019 Velici Variation of Aerosol Concentration (2)

1/32

REPORT SERIES IN

AEROSOL SCIENCE

N:o 46 (2000)

VARIATION OF AEROSOL CONCENTRATION

IN AMBIENT AIR

GINTAUTAS BUZORIUS

Department of PhysicsFaculty of Science

University of Helsinki

Helsinki, Finland

Academic dissertation

To be presented with permission of the Faculty of Science

of the University of Helsinki, for public criticism in the main Auditorium

of the Department of Physics, Siltavuorenpenger 20 D,May, 31-th, 2000, at 12 oclock noon.

Helsinki 2000


2/32

ISSN 0784-3496

ISBN 952-5027-22-8 (nid.)

ISBN 951-45-9428-2 (PDF version)

Helsingin yliopiston verkkojulkaisut

Helsinki 2000


3/32

1

Acknowledgements

This thesis was based on research work done at the Laboratory of Aerosol and

Environmental Physics at the University of Helsinki. I thank Prof. Juhani Keinonen, headof the Physics Department, for providing me the working facilities.

I would like to thank to Prof. Markku Kulmala, Doc. Jyrki M. Mkel, Doc. Timo Vesala

and Doc. Kaarle Hmeri for their invaluable support and guidance during the studies in

Helsinki. I also thank all the group members for the relaxed and supportive working

environment. I thank my friend and co-author llar Rannik for a friendship and

cooperation. I highly appreciate the collaboration with the people from the SMEAR II

station, especially Dr.Hc. Toivo Pohja for helping to construct and maintain the

measurement system what was used during the study.

My best memories stay with the aikido club Seitokai members where I had a lot of fun

enjoying the practises. I thank all the friends, only few I mentioned here, whom I met

during my study years in Finland, and shared good time. I heartily thank to Mika Tapio

Kleemola, Enzo Filomeno, Lasse Heikkanen and Meri Kukkavara for their invaluable

friendship and support.

Finally I thank to my family for their support and caring during my years away in

Finland.


4/32

2

Variation of aerosol concentration in ambient air

Gintautas Buzorius

University of Helsinki, 2000

Abstract

Experimental studies on atmospheric aerosol have been increasing over the last decade,

resulting in new findings about particle formation in the atmospheric surface layer and in

more intensive studies on public health effects in relation to aerosol number

concentration. This work consists of experimental results concerning aerosol number

concentration measured in urban and remote continental environment.

Studies on the aerosol spatial variability in urban areas showed that 10 min averages ofthe number concentration can be highly correlated (> 0.8) between places located more

than 1 km apart from each other. It suggested that aerosol number concentration in the

city is spatially rather homogeneous and simultaneously varying in time at different

places.

A Condensation Particle Counter was applied in an eddy correlation technique to measure

particle vertical fluxes. The method requires fast aerosol number concentration sampling.

The CPC performance under high frequency sampling conditions was analysed under

laboratory conditions where time constant studies were determined as being from 0.8 to

0.14 seconds, depending on the counter model. With the time resolution given by the

CPC it is possible to measure most of the fluctuations in the number concentration

contributing to the flux.

The flux measurements showed a prevailing depositing flux and a pollution source at

Southwest direction at the field measurement site (Hyytil). Fluxes were affected by

advection. There are episodes of upward fluxes, which were confirmed by the particle

gradient measurement. However these upward fluxes were associated with the local

pollution and are not characteristic features of the forest as a source of particles.

Keywords: turbulent flux, condensation particle counter, eddy covariance, spatial variation and deposition

velocity


5/32

3

Content

Acknowledgements 1

Abstract 2

List of publications 4

1. Introduction 5

2. Variation of aerosol number concentration 11

2.1. Spatial variation urban case 12

2.2. Remote continental environment 15

3. Flux measurements 21

4. Summary and conclusions 25

5. Review of publications 27

6. References 28


6/32

4

List of publications

1. G. Buzorius, K. Hmeri, J. Pekkanen and M. Kulmala, 1999: Spatial variation of

aerosol number concentration in Helsinki city,Atmospheric Environment, 33, 553

565.

2. G. Buzorius, 2000: Cut-off sizes and time constants of CPC TSI 3010 operating at

flow rates from 1 to 3 lpm, to be published inAerosol Science and Technology.

3. G. Buzorius, . Rannik, J. M. Mkel, T. Vesala and M. Kulmala, 1998: Vertical

aerosol particle fluxes measured by eddy covariance technique using

condensational particle counter,J. Aerosol Sc., 29, 157 171.

4. G. Buzorius, . Rannik, J. M. Mkel, P. Keronen, T. Vesala and M. Kulmala,

2000: Vertical aerosol fluxes measured by eddy covariance method and

deposition of nucleation mode particles above a Scots pine forest in Southern

Finland, to be published in Journal Geophysical Research Atmospheres.


7/32

5

1. Introduction

1.1 What are aerosols? Aerosols are two-phase systems, consisting of liquid or solid

particles and the gas in which they are suspended, For example, cloud particles, airborne

particulate matter, dust particles in the air, hair spray, etc. Additionally huge varieties of

aerosol are not visible to the human eye (photochemically formed particles, salt particles

from ocean spray). Secondary aerosol particles are formed by nucleation of gas

molecules. Particles grow in size by coagulation and condensation. From the atmosphere

particles are removed by deposition. The usefulness of aerosol physics and chemistry is

recognized by aerosol applications in industry (i. e. industry of thin layers, electronic) and

in daily life (air freshener, hair sprayer).

1.2 The size distribution of atmospheric aerosols. Measured aerosol size distribution in

ambient air covers a size range from a few nanometers to a hundreds of micrometers

depending on environment. New particles form by nucleation at sizes less than 3 nm,

afterwhich they rapidly grow onto ultra-fine mode particles of the order of 10 nm and

then into the fine mode (or Aitken mode) at 20 30 nm. A fraction of the Aitken mode

particles grow further into accumulation mode sizes of 100 300 nm. Cloud

Condensation Nuclei are particles what can become activated to grow to fog or cloud

droplet in presence of water vapor supersaturation. The minimum CCN particle diameter

is from 50 to 140 nm, and concentration of those could be from hundreds in remote

places to many thousands per cubic centimetre in urban environment (Seinfeld and

Pandis, 1998).

The number concentration of aerosol size distribution is typically dominated by fine and

accumulation mode, while surface area distribution is dominant by accumulation and

coarse mode, and mass/volume distribution of aerosol is dominated by the coarse mode.

The coarse mode consists of particles larger 1 micrometer and up to 200 300 m, for

example dust, pollen, fog, cloud droplets, etc (ODowd et al., 1997).

Two very important aspects of aerosols: impact to public health and particle removal

from the air historically have been studied for particles with sizes larger 100 nm. In


8/32

6

health studies impactors had been employed with cut-off size at 1 or 10 m. In deposition

studies instruments covered sizes down to 100 nm (Wesely et al., 1983, Neumann and

Hartog, 1985, Duan et al., 1988). Substantially, those aspects have not been studied

directly for particles sized below 100 nm where the highest number concentration occurs.

1.3 Source and production mechanisms of atmospheric aerosols. Aerosols primary

production mechanisms are combustion and mechanical generation. Typically

accumulation and coarse mode particles are produced in that way. Biomass burning,

anthropogenic combustion, dust from deserts, marine sea-salt contributes noticeable to

aerosol mass globally. Homogeneous and heterogeneous nucleation is a secondary

particle production mechanism. Globally this mechanism contributes remarkably to

number concentration of aerosol. Primary species for nucleation could be organic in the

forest, sulphates in marine and combustion gaseous products in urban environment.

Newly formed nucleus grow in size by condensation of condensable vapor or

coagulation.

1.4 Removal of particles from the atmosphere. The main aerosol removal processes are

dry and wet deposition. Deposited particles on the surface of the ecosystem can cause

chemical reactions. Additionally the impact of aerosol on acidification of the ecosystemand deforestation has been recognized (Erisman et. al., 1997). This caused the

establishment of regular estimation of aerosol loadings on the surface as well as

monitoring of aerosol particle deposition. Whether wet or dry deposition is more

important depends on uptake of particles by falling droplets, amount of precipitation in

region, the terrain and type of surface cover.

Factors influencing the dry deposition are the level of atmospheric turbulence, physical

and chemical properties of the deposited constituent and the nature of the surface itself:

physical properties such as particle size; chemical properties such as thermo chemical

reactions inside the particle, causing drift velocity of the particle in temperature gradient

field close to the depositing surface, and modifying deposition velocity. Particle dry

deposition governing processes are gravitational settling, interception, impaction and

Brownian diffusion. Dry deposition can be parameterised according to particle size,


9/32

7

roughness length, stability and friction velocity. Below deposition velocities to the forest

are presented.

Theoretical models predict higher deposition velocities (from 0.2 to 20 cm/s) for

nucleation mode particles (sized from 3 to 20 nm) due to higher diffusivity (i. e. Slanina,

1997) compared to Aitken mode particles. Accumulation mode particles, according

models, are deposited with a 0.04 0.4 cm/s velocity, whereas experimental data

systematically show ten times larger deposition velocity. Particles about 100 nm size

deposit at 0.08 0.4 cm/s according to measured and predicted values (Gallagher et al.,

1997). Before this thesis was published experimental results of the flux measurements

typically covered particle deposition for size ranges down to 100 nm (Wesely et al., 1983,

Neumann and Hartog, 1985, Duan et al., 1988), which is very often larger than the mean

or mode diameter of aerosol size distribution. The experimental works were verydifferent from modeling values, showing that deposition processes are poorly understood

(Ruigjrok et al., 1995). Most of the models do not include phoretic effects. Schery et al.

(1998) measured a deposition velocity of ~ 1 nm in diameter size particles using

unattached radon progeny as a tracer. Measured values were in the range of 5 35 cm

s-1.

A significant part of this work covered the development of a experimental system to

measure fluxes for particles with sizes down to 10 nm. Eddy correlation (EC) is widely

used to measure vertical gas and energy fluxes (Businger, 1986). For particle flux

measurements, for the first time a condensation particle counter (CPC) was applied to this

system. The thesis covers all the aspects of the application, including estimates of

different uncertainties, examples of sample data and measurement results. However, the

results opened a huge possibility of investigation of surface atmosphere interactions

with respect to aerosols, including removal and formation of new particles in or above the

forest. Therefore this research work should be continued in the future.

1.5 Aerosol and public health. Inhaled particles are deposited in the trachea and lungs

and may cause pathological effects. Established air quality standards or limit values for

particulate matter is based on mass concentration of particles smaller than 10 m and

shows critical air pollution levels (in Finland the level is 70 g m-3). Monthly averages of


10/32

8

PM10 measured at different places in Helsinki ranged from 20 to almost 60 g m-3 with a

peak up to 80 g m-3 in springtime, due to spring dust. While mass is dominant by larger

particles, small particles can penetrate deeper into the lungs due to lower inertia and

probably cause pathological effects. Therefore, PM10 standard is insufficient to

determine health effects of air quality. Experimental work in environmental

epidemiology during last decade suggested that adverse effect of pollution on public

health is number not mass concentration dependent (Dockery and Pope, 1994). It has

been hypothesized that the larger number of those particles in urban air may be the

explanation of the observed health effects of PM10 (Seaton et al., 1995). There is an

attempt to establish PM2.5 standard based on particulate smaller 2.5 m mass

concentration.

Instrument development allowed more intensive studies of aerosol deposition processes

and air pollution public health effects. Studies of pollution on public health require

aerosol loading per human per time during his normal daily activity. Therefore spatial

distribution of aerosol number concentration is needed. Part of this work analyses what is

a spatial correlation in aerosol number concentration between different places in Helsinki

city. Or in other words, it addresses the question I have been asked by epidemiological

scientists in how many places does aerosol concentration have to be measured in order

to representatively estimate the time variation of particulate pollution, in terms of numberconcentration, within a city?

1.6 Aerosols and Climate. Recent work addressed aerosol affects on the climate (e. g.

Charlson and Wigley, 1994). Aerosol affects climate indirectly by acting as a cloud

condensation nuclei and forming cloud and fog droplets, which reflect sunlight back to

space and infrared waves back to the Earth surface (Seinfeld and Pandis, 1998); and

directly by reflecting sunlight back to space and thereby cooling the atmosphere. The

radiative forcing is commonly defined as the instantaneous change in the irradiance, in W

m-2, at the tropopause due to change, for instance, in greenhouse gas (CO2, CH4, N2O and

CFCs) or aerosol concentration after allowing for stratospheric temperatures to re-adjust

to radiative equilibrium but with the surface and tropospheric temperature and

atmospheric moisture held fixed. Concentration in tropospheric aerosols from pre-


11/32

9

industrial times to the end of twentieth century has increased. An increase in CCN

number increases indirect cooling effect. An increase in greenhouse gases concentration

cause global warming effect. Overall, the global warming problem has not been fully

understood because there is a lack in knowledge of aerosols contributions to the cooling.

Estimated indirect effects of tropospheric aerosols on climate have large uncertainties.

Global forcing of sulphate aerosol was reported to be 0.9 W m -2 and attribute 0.45 W

m-2 of this to direct effect. About 25% of the direct radiative forcing effect is attributed to

scattering by aerosols in cloudy skies (IPCC, 1996).

1.7 Aims of this study. Objectives or aims of this research work were (1) to study spatial

variability of aerosol number concentration in Helsinki city; to analyse how much

concentration is correlated between different places in Helsinki downtown area. (2) Toinvestigate the time response limitations occurring during the aerosol sampling, including

frequency attenuation in the sampling line and the attenuation due to the slow response

counter. (3) To apply a particle counter with considerably lower cut-off size in order to

measure particle fluxes according to the eddy covariance method. The application

consisted of numerous test, FFT analysis of the measured scalar time series, estimation of

uncertainties and experiments in the field. (4) To characterize aerosol behaviour in a

boreal forest environment (Hyytil, SMEAR II station, Southern Finland) in terms of the

vertical flux measurements above the forest at the site, and to determine dry deposition

velocities at different size of particles.

1.8 Experimental tools used. A condensation particle counter is a single particle

counter. Because particles have too small size to be detected optically, the particles are

enlarged in their size by condensation of butanol in chamber supersaturated by the

alcohol. The inlet flow inside CPC passes saturator section, which is heated up and

contains liquid alcohol. The liquid evaporates and saturates the sheath airstream. The

airstream next enters the condenser, where temperature is lower compared to saturator.

The vapour cools, becomes supersaturated and condenses on particles growing them into

droplets. Later droplets are optically counted. Minimum detectable particle size is 10 nm

for the CPC model TSI 3010. The minimum size depends on supersaturation level, which


12/32

10

can be controlled by controlling the temperature difference between saturator and

condenser (Mertes et al., 1995). Increasing the difference allows one to increase the

supersaturation level and to decrease the minimum size at which particles are activated

and grow into droplets. Sampling flow rate is 1 lpm and the temperature difference is of

17 degrees for a CPC in standard modification. Another CPC, model TSI 3025 has cut-

off size at 3 nm.

Ultrasonic anemometer Gill Solent 1012R was used to sample three wind components

and air temperature at 20 Hz frequency. It is a widely used instrument.

Eddy covariance is a micrometeorological method to measure vertical fluxes of different

atmospheric constituents (Businger, 1986). In the method flux is denoted as a covariance

between the vertical wind speed, w, and the concentration of the constituent, c,

)()( ccwwF = .

Where, overbar represents an average. The method can be applied within the turbulent

boundary layer and assumes horizontal homogeneity of the constituents. The method is

widely used in measuring gas, momentum, and energy exchange between the surface and

atmosphere.

1.9 Contribution of this work of global aerosol issues. Results from the spatial

variability studies in Helsinki city are very important for health studies. Conclusionsdrawn during this study says that aerosol concentration can be monitored in one place in

downtown area representing the concentration changes in an area with spatial scale of

kilometer(s). It allows decreasing the number of monitoring stations saving costs of the

research projects.

Estimated time constants of the CPC and frequency attenuation allowed evaluation of

high speed aerosol sampling. The high speed sampling is needed in aerosol number flux

measurements, from which studies on particle dry deposition and possible sources of

aerosol particles can be conducted.

Obtained vertical particle fluxes allow estimation of the overall dry deposition velocities

for aerosols with sizes down to 10 nm. This data is required in aerosols transport models

to estimate aerosol exchange rates between atmosphere and ecosystems. Changes in


13/32

11

particle number lead to changes in CCN concentration, and thus, are very important in

modeling global climate change.

2. Variation of aerosol number concentration

Typically in the atmosphere, concentrations of long living species are relatively

homogenously spatially distributed due to turbulent mixing in the atmosphere. Spatial

variability of the concentration denotes how much correlation is between concentrations

changes measured at spatially different spots. The lifetime of different size particles is

different ranging from days for accumulation mode particles (larger 100 nm) to minutes

or hours for nucleation mode particles. Because of different age, particles are not mixed

in the same extend through the mixing layer. Therefore, the spatial variability of those

could be different. Typically accumulation mode particle concentration measured in

remote places is relatively stable and changes in time are rather slow. Nucleation mode

particles have rather shorter lifetimes due to their growth to Aitken mode by

condensation or coagulation (Juozaitis, et al., 1996, Kulmala et al., 1998). Fluctuations in

the concentration of those particles measured in one spot are relatively larger compared

to the fluctuations in concentration of the accumulation mode particles. Therefore,measured variability of aerosol number concentration depends on detected particle sizes.

Variation in the concentration measured by optical counters with a minimum detectable

size about 100 nm will be different from those measured by condensation particle

counters with cut-off size of 10 nm. Close to the pollutant source or sinks, perturbations

in the spatial homogeneity are created. For example, in the urban environment

anthropogenic aerosol sources cause a lot of variety in particle concentration. Measured

fluctuations depend on the location of the sampling site: if it is close to the source (busy

street, factory), and from the air mixing. In street canyon vertical profiles of

concentration can have large gradients (Vkev et al., 1999).

During the aerosol sampling, part of the fluctuations in concentration time series is lost.

Factors determining the loss are counter response time, averaging time period and

minimum detectable size.


14/32

12

2.1 Spatial variation in urban environment. There is a very little known about spatial

variability of aerosol number concentration in urban environment. Temporal variability is

rather high due to the large number of pollution sources in the city. Therefore spatial

variability is expected to be high too. Results from measurements with impactors in the

Washington D. C. showed a very high correlation in PM2.5 and PM10 between different

sites located downtown (Suh, et. al., 1997).

Typical concentrations in different modes in urban environment are larger compared to

remote environment. The Aitken mode has from 103 to 105 particles cm-3 concentration in

the urban areas, whereas in remote places the concentration hardly reaches 104 cm-3.

During particle bursts concentration in nucleation mode is present with concentration

from 103 to 106 particles cm-3 depending on the environment (Mkel et. al., 2000,ODowd, et al., 1999). Accumulation mode particles often are attribute to the long range

transport. The concentration in accumulation mode increases in polluted areas.

The spatial variation of the aerosol number concentration in Helsinki city was studied

(Paper 1). The purpose of the study was to find out how many measurement spots are

needed in the city in order to estimate a dose of inhaled particles per human per day. My

task was to find out in how large area concentration can be predicted by sampling it at

one spot. Measurement system and conditions are described in Paper 1. The minimum

detectable particle size of the used counter was 7 nm. The measured variability in

concentration time series depends on the sampling frequency. Figure 1 shows an example

of the same concentration plotted every minute, and averages every 10 min. There are

noticeable differences between 1 and 10 minute samples. The longer the period over

which mean values are calculated the smaller the variation in the averaged concentration

time series is seen.

Measurement results showed relatively large correlation between concentration changes

at different places (Figure 2). There are two structures in data behaviour versus time. First

structure is a slow simultaneous change in concentration correlating with traffic intensity

at time step of an hour. Studies on wind speed data neglected an explanation of

correlation due to advection. Interpretation of the high correlation is that in Helsinki city

the main air pollutant source is traffic. A relatively homogeneous network of streets with


15/32

13

traffic on them provide relatively homogeneous aerosol source. Second, at the smaller

time scales concentration time series exhibit individual behaviour. Occasionally, particle

concentration has larger values due to heavy traffic, or stopped cars, etc at the site closer

to the street.

0 4 8 12 16 20 241000

10000

100000

1 min raw data

10 min averaged data

Concentration,particlecm

-3

Time, hours, 13 March, 1997

Figure 1. Aerosol number concentration sampled by CPC TSI 3022A in

the downtown area in Helsinki every second. Minute averages are in

line, 10 minutes averages in scatter.

In case of higher wind speed air in the city is mixed relatively faster and with the

following dilution of the peaks. The results showed that the highest correlation occurs at

the higher wind speeds. The lowest correlations during the whole campaign occurredover weekends when traffic intensity was relatively small (Paper 1). The study showed

that, even though the temporal variability is relatively high, it is spatially correlated.


16/32

14

1000

10000

100000

KumpulaSiltavuori

Vallila2

Totalconcentration,pa

rticlecm

-3

030060090012001500

Siltavuori

Kumpula

Traffic,carshour

-1

21 22 23 24 25 26 27

Time, day, February, 1997

Figure2. Aerosol number concentration and traffic intensity in different Helsinki

downtown areas.

Figure 3 shows the correlation coefficient dependence from the averaging period T

(averaging window width) when the concentration was sampled every second at two

downtown places (Kumpula and Vallila) located at 1 km from each other (description of

the sites is in Paper 1). The concentration time series included several days data. The

longer averaging time period the smaller variability between the time series.

Concluding this section, a reminder is that the correlation between the aerosol number

concentration time series measured at different places depends on the used time

resolution. At time scales of minutes the time series show a lot of variability and are

rather different from place to place. At time scale of an hours (in epidemiological studies

4, 8 and 24 hour averages are used) temporal variability is averaged out, and the time

series in general correlates with the averaged traffic intensity. The main result of thisstudy is that for monitoring purpose of the time variation of aerosol concentration in the

downtown area probably it is enough to have a one sampling site, but the place for it must

be chosen carefully. It should not be in places where air pollution level is pronounceable

affected by local environment (not in closed yards or big parks).


17/32

15

0 5 1 0 15 20 25 30 35 4 0 4 5 50 55 6 0

0.78

0.80

0.82

0.84

0.86

0.88

0.90

Correlationc

oefficient

Time pe riod T, minutes

Figure 3. Correlation coefficient between aerosol number

concentration time series measured at different places versus

averaging window width T.

2.2 Remote continental environment. In contrast to urban environment where the

concentration time series exhibit a lot of variability at tens of seconds and minutes time

scales, in remote environment the concentration time series have smaller changes inabsolute number (Figure 4). In remote places comparable to urban places concentration

changes occur at slower time scales compared to the urban environment. Very often the

changes are originated from advection and typically, in absolute number, are smaller,

being of the order of 102 to 103 particles cm-3 in magnitude. In case of changes of air

masses at synoptic spatial scales the concentration can change in more drastic way

(Figure 4, morning of May 5).

In order to study variability of aerosol number concentration, turbulent mixing in the

atmospheric surface layer has to be taken into account. Turbulence frequency spectra of

interest covers from tens of minutes to tenths of second time scales. Analogous eddy

sizes are from centimeters up to hundreds of meters (Stull, 1999). There are experimental

difficulties in measuring aerosol number concentration at high frequencies. Therefore an

experimental investigation was carried out to study the CPC time response.


18/32

16

Figure 4. An example of aerosol number concentration sampled

above the forest with the CPC model TSI 3010, May, 1999.

2.2.1 Increase in CPC time resolution at high volume sampling flow rate. Fluctuation

frequency in atmospheric constituent concentrations due to turbulence covers time scales

from synoptic motion (hundreds of kilometres) to 0.1 sec variation where motion energy

is dissipated into heat (Kaimal and Finnigan, 1994). To resolve the fastest changes

measurement of fluctuations in concentration sampling should be performed at 10 Hz. It

is likely that not all situations require the sampling that fast. Modern aerosol

instrumentation allows measuring concentration of aerosol particles with sizes bigger 3

10 nm and with time resolution of 1 2 seconds. Most particle counters have too slow an

instrument time constant and are not able to resolve ambient fluctuations in concentration

at high frequency. Therefore I have done time constant studies on CPC model TSI 3010with an attempt to evaluate and increase the time resolution. This model was chosen due

to the higher flow rate. The high flow rate is needed to assure reliable counting statistics

and to reduce systematic uncertainties in flux measurements (Fairall, 1984). Paper 2

describes the performance of the CPC modified to sample at different flow rates ranging

from 1 to 3 lpm. The modification consisted of a removal of a critical orifice from the

10 0

1000

10000

Concentration,particles

cm

-3

1 2 3 4 5 6

Time, days


19/32

17

instrument and an installation of a valve before vacuum line plug to control the flow rate.

Changes in the cut-off size due to changes in the sampling flow rate were studied.

Increasing the sampling flow rate from 1 to 3 lpm allowed decreasing in the CPC time

constant from 0.8 to 0.4 sec.

The modified CPC was sampling aerosol under ambient conditions in remote continental

environments. In parallel, another CPC model TSI 3010 operating at standard conditions

was installed. Both CPCs had cut off sizes at 10 nm. Figure 5 shows the measured aerosol

concentration time series.

104.5 105.0 105.5 106.0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

Particleconcentration,particlescm-3

Time, Julian day, 1999

modified CPC

standard CPC

8 12 16 20 24 28 32 36 40 44 48

Hours, since begining of 14 April, 1999

Figure 5. Particle concentration measured by the modified

and standard CPC model TSI 3010.

Instantaneous differences in total concentration between the time series are negligible.

Every point on the plot is a half an hour average. Thus, the modified CPC provides

similar data compared to the standard one. Therefore, it can be concluded that the CPC

model TSI 3010 with 3 lpm sampling flow rate operates reliably.

Decreasing the instrument time constant improves the frequency transfer function of

concentration measurements. Thus, high frequency fluctuations are less attenuated during

the aerosol sampling (Paper 2).


20/32

18

2.2.2 Frequency attenuation. Mixing of the flow inside the counter volume (sampling

lines, condenser, saturator) distorts the profile of the signal. Those distortions can be

parameterized with a first order differential equation (Russell et al., 1995). The time

resolution of the instrument is limited due to them. High frequency fluctuations in the

concentration time series can be attenuated while aerosol is sampled through the

sampling line or due to relatively slow time response of the counter. The time response is

estimated using the time constant under assumption that the CPC is a first-order

instrument (Doebelin., 1990). The attenuation by counter is estimated from the

instrument frequency transfer function. To determine the function a time constant of the

instrument is needed. The time constant, , (for a step change) is the time required to

reach e-1 of its driving signal following formula

=

t

inp

outexp1 , where inp - is a signal

introduced to the instrument, out is the measured signal, t- time. The constant was

estimated from the CPC response to step and sinusoidal changes in the aerosol

concentration (Paper 2).

An additional study was conducted to evaluate frequency attenuation in the sampling

lines. The experimental setup was described in Paper 2 (setup Nr. 2). Generated step

changes in the aerosol concentration were measured by CPC model TSI 3025. The

experiment was repeated under slightly different experimental conditions. An additional

4.5 m long sampling line was installed before the CPC inlet. The experiment was done

several times using turbulent and laminar flow regimes in the sampling line. Later the

line was removed and the experiment without the line was repeated. Measurements were

repeated several times, and averages are shown in Figure 6. Scatter in the data between

different runs of the same experiment type was negligible. The response behaviour before

and after the experiment using the turbulent flow regime was the same (data points in

squares). The experimental system was stable and results were repeatable. Using a

turbulent flow regime (Re=2350) no difference in the response curve to the step changein the concentration was observed after the length of sampling line was increased from 10

to 460 centimetres. Therefore, there is no frequency attenuation during the turbulent

sampling for the time resolution given by the CPC model TSI 3025. The time constant of

this instrument is about 0.14 sec (Paper 2), which provides a better time response

compared to the model TSI 3010. If the carrier flow is in a laminar regime the response to


21/32

19

step change becomes slower (Figure 6, up-triangles). Increasing the length of the

sampling line causes a higher attenuation. Thus if turbulent flow is used in the sampling

line (adding carrier flow) the attenuation is caused mainly by the CPC.

For quality controls, the normalized (by variance) FFT spectra (Kaimal and Finnigan,

1994) of sampled concentration with the spectra of other well-known quantity, for

instance, temperature, can be compared (e. g. Laubach and Teichmann, 1996). Using

measured data taken over the forest I compared the CPC frequency spectra with the

spectra of temperature measurements taken from the same spot at 20 Hz frequency. Used

data were taken from periods between 13.00 and 13.30 on 4 April, 1999.

2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

MEAN before experiment with the line

MEAN after experiment with the line

MEAN with the line, flow 6.7 lpmMEAN with line flow 1.3 lpm cpc is 32 cm + 4.5 m lineMEAN no line flow 1.3 lpm cpc is 32 cm after dma

Normalisedconcentration,

CPCTSI3025

Relative time, sec

Figure 6. CPC model TSI 3025 responses to step change in

concentration. Aerosol was sampled through different length sampling

line with turbulent and laminar regime.

Simultaneously high frequency temperature measurements were performed by Sonic

anemometer. This is a one half an hour sample from flux measurements what were


22/32

20

described in Papers 3 and 4. Spectral densities of fluctuations in aerosol number

concentration and temperature were calculated using Fast-Fourier-Transformation, and

are plotted in Figure 7 (slope with sign 2/3 denotes energy dissipation slope). At

frequencies above 0.1 Hz slopes in the temperature and the concentration spectra decay

become different. The temperature spectral density decays with a slope similar to the

energy dissipation slope. Whereas the particle decay is faster, indicating higher variances

at seconds and tens of seconds time scales.

10-4

10-3

10-2

10-1

100

101

10-7

10-6

10-5

10-4

10-3

Frequency, f, Hz

S

pectraldensity,

f*S

circles-particle concentration, stars-temperature

C

T -2/3

Figure 7. Power spectra of temperature (stars) and particle number concentration

(circles) measured above the forest.


23/32

21

3. Flux measurements

After fluctuations of the aerosol number concentration are measured, vertical flux of

aerosol can be estimated by correlating the time series of the fluctuations with thefluctuations of the vertical wind speed data according to eddy correlation method. A

condensation particle counter model TSI 3010 was applied for the flux measurements.

The time constant of the CPC is less than a second (Paper 2) but still being higher than

time scale of the smallest eddies. The amount of those eddies is relatively small, and

underestimation under some atmospheric conditions is negligible. Nevertheless missing

flux can be estimated in two different ways. The first requires a prior knowledge of the

spectral characteristics of the variable being measured. Typically the spectral

characteristic is assumed to be similar to the characteristics of the other well measured

species (e.g. temperature). By comparing cospectra and after some integration, missing

flux can be evaluated (Shaw et. al., 1998). The second, way to estimate missing flux

requires knowledge of the instrument frequency transfer function or the time constant.

Simple formula determining missing flux using the time constant was derived by Horst

(1997). The application of the formula showed that the missing flux encounters about

15 % of measured value depending on atmospheric stability. Paper 4 describes examples

of the missing flux.Fluctuations in relative humidity cause artificial particle flux (Sievering, 1982, Fairall,

1982). However, this is relevant only for the detectors that have detection efficiency

depending on the particle sizes. The CPC model TSI 3010 detection efficiency is size

dependent for particles with sizes 4 14nm (Paper 2). During the events of high particle

concentration in nucleation mode (Mkel et al., 1997) the flux value can be distorted.

However, hygroscopisity measurements showed that the new particles have a small

change in their size when the relative humidity changes from 10 to 90 % (Hmeri et al.,

2000). Therefore, the fluxes most probably are not distorted. An example with sulphuric

acid particles is evaluated in the Paper 4.


24/32

22

Normalized flux versus deposition velocity. Measured vertical particle flux can be

expressed as

dvcF =

where: c is the concentration, vd deposition velocity. For the flux parameterization

purpose ratio of flux with concentration and reverse sign is used. The ratio has units of

velocity and has positive values for the downward direction. This normalized flux is

relevant to the surface deposition velocity. In the case of unstable stratification

normalized flux is close to the surface deposition velocity. Dry deposition velocity is the

quantity describing particle uptake by the surface. The deposition velocity is the function

of the particle size, and has a V shape with particle size. One of the commonly used

models (Slinn, 1982) includes processes like interception, impaction, rebound, diffusion

and sedimentation. The shape is a result of a combination of those processes. Diffusion is

a governing process for Aitken and nucleation mode particles, and gravitational settling

enhances deposition of coarse particles. Sizes in-between have a lower deposition

velocity due to relatively small diffusion coefficient and relatively slow settling velocity

(Hinds, 1982). However, experimental results show huge deviations in experimental data

from the predicted values for accumulation mode particles (Gallagher et al., 1997). There

seem to be other processes (for instance, phoretic effects, charging effects, artificial flux

due to saturation ration flux) influencing the measured vertical flux but are not includedin models.

Basically changes in concentration time series occur due to advection of different air

masses, or due to dilution/mixing among stratified vertical layers. The last is more likely

to occur during the morning when the convective boundary is growing and the surface

layer mixes with the residual layer (Nilsson et al., 2000). Predominantly the

concentration changes occurring over relatively slow time scales (hours, tens of minutes)

are usually addressed to the horizontal advection. The linear pattern in the changes of

concentration time series can be filtered out avoiding contamination of the flux data

(Lamaud et al., 1994). Under close to neutral or weak unstable conditions above very

rough surface like forests, episodes with upward particle flux can be measured. Figure 8

shows an example of the similar episode measured above the pine forest in Hyytl,

Southern Finland. The experimental setup was the same as described in Paper 3 and 4.


25/32

23

Additionally two particle counters CPC model TSI 3025 and two CPCs model 3010 were

sampling aerosol at 18 and 67 m (one 3010 and one 3025 at each level). The

concentration gradient was measured with a two independent measurement systems and

different only in CPC model. Turbulent mixing was generated by wind shear in night

time (higher on that night than average wind shear during the nights) and wind shear and

buoyancy forces in day time. Flux data at least gave right direction of the vertical

transport. Episodic fluxes upward (10.30 and 11.30) and systematic flux upward when

the wind comes from the Southwest direction (after 22.00) were observed (Figure 8).

Similar episodes are frequently observed above the forest at our measurement site.

From 5 a.m. to 7 a.m. the concentration rose probably due to advection of an air parcel

with higher concentration. At 67 m on that time, the concentration was higher than the

concentration at 18 m (see plot.). The eddy flux measurements at 23 m showed fluxdownward during the episode. Here the particle vertical gradient agrees with the

measured particle vertical transport. Air with the relatively higher aerosol concentration

is transported down, and the aerosol concentration inside the forest is increasing. During

the period from 7 a.m. to 8 a. m. the parcel passed away. The concentration then

decreased, first at higher levels, than in/close to the forest. The established vertical

gradient causes on upwards particle flux.

At 8.00 particle flux showed a small peak up. On that time the CPCs TSI 3025 showed

similar concentrations. Upward gradient was observed by CPCs TSI 3010 (not pictured),

but not visible in CPCs 3025 data. Probably because of the instabilities in sampling flow

rate of the CPC TSI 3025.

At 10.30 and 12.30 there are two episodes with fluxes and gradients directed upwards.

The ratio in concentration increased according the both gradient measurement systems

(see the Figure). Actually there were increases in the number concentration at lower

levels by order of magnitude at 10.28 and 12.29. The peaks lasted for a few minutes, and

are not explicitly visible in a an hour averages. At midday concentration started to rise

significantly due to a new particle formation observed at the site (Mkel et al., 1997,

Kulmala et al., 1998, Aalto et al., 2000). During the well-mixed conditions, concentration

gradients are very small and comparable to the CPC precision in measuring absolute

concentration. During the nucleation events, due to a large nucleation mode, the


26/32

24

downward flux is enhanced (13.00 17.00 hours in Figure 7) compared to periods when

nucleation mode is absent.

-40

-20

0

20

40

Flux,

106 particlem

-2s

-1

Time, hours

particle flux

103

104

Concentration,particlescm

-3

concentrationCPC 3025 at 18 m

0.7

0.8

0.9

1.0

1.1

1.2

1.3

Concen

trationratio18m/67m

conc(18 m)/conc(67 m)

0 2 4 6 8 10 12 14 16 18 20 22 24

Figure 8. Particle flux and concentrations during the April 8, 1999.

Later in the evening after 9 p.m. wind turned from the Southwest direction where aerosol

pollution is emitted from the local station buildings, and the vertical gradients and the EC

fluxes showed overall transport upwards. This upward flux was described in Paper 4.

From this example it is obvious that normalized flux is not the same as the deposition

velocity. Stratified layers or advection of air parcels with different aerosol concentration

causes vertical gradients above the rough surface. After consideration of possiblegradients due to spatial (vertical) variability in concentration, measured flux data can be

used for the estimation of surface deposition velocity. For estimation of the deposition

velocity it is recommended to choose data periods when the aerosol concentration was

stable.


27/32

25

Measured aerosol number concentration above the boreal forest varied from 103 to 104

particles cm-3. The particle fluxes varied from 15*107 to 5*107 particles m-2 s-1, where

negative fluxes mean a downward transport. The fluxes from the local pollution sources

were removed from the given flux value. If those were included the upward fluxes

reaches 108 particles m-2 s-1. In overall, the normalised flux varied from 20 to 30 mm s -1

(positive downward). During the periods when Aitken and the accumulation mode was

the dominant deposition velocity was from 0 to 10 mm s -1 and from 0 to 5 mm s-1

respectively. However, during the periods when the nucleation mode was dominant the

deposition velocity occasionally increased up to 30 mm s-1. In case of the pollution

episodes (local emissions and advected pollution from sources near by) the fluxes are

distorted, and the deposition velocities were ranging from 20 mm s-1 (local emissions) to

20 mm s-1. About 30% of the samples showed upward flux. Those attribute to thepollution from local sources and are not characteristic particle emissions from the forest.

4. Summary and conclusions

Measurements of the aerosol number concentration were conducted in the urban and

remote continental environment to study variation in the concentration. In the urbanenvironment the spatial variability of the aerosol number concentration was studied. In

the remote environment, particle vertical fluxes and deposition of particles above the

boreal forest were measured.

The aerosol number concentrations in Helsinki city varied from 103 to 105 particles cm-3,

and on average correlated with traffic intensity. The highest concentrations were

observed during the rush hours, the lowest in night time and during the weekends. The

spatial variability of the aerosol number concentration in Helsinki city was rather low,

with changes in the concentration time series at tens of minutes time scale occurring

almost simultaneously. The explanation of the correlation is that the pollution source is a

traffic, and the traffic is on the uniform network of the streets. The correlation is higher

on weekdays and during the higher wind speed conditions. The smallest correlation

occurred during the weekends when aerosol number concentrations were smaller. Results


28/32

26

suggested that in aerosol monitoring for epidemiological purposes, where loads of the

particle number per human per day are needed, probably it is enough to have one

monitoring station in the downtown area, but the place for it must be chosen carefully.

The particle number concentration in the boreal forest environment ranged from 103 to

104 particles cm-3, when in the urban environment measurements showed from 103 to 105

particles cm-3. Above the forest aerosol concentration was sampled at high speed (10 Hz),

in order to obtain vertical particle fluxes. Attenuation in the measured concentration time

series due to a slow response of CPC was evaluated under laboratory conditions. During

the studies, time constant of the CPC was decreased twice, maintaining the same cut-off

size and increasing the volume sampling flow rate. The estimated fraction of the flux that

is lost due to attenuation is about 15% in our measurement site.

Deposition velocities to the forest were derived from the flux measurements. The highestdeposition velocities (up to 30 mm s-1) were measured during the nucleation events.

During the background conditions the deposition velocities were from 0 to 5 mm s-1.

Results for dry deposition of the accumulation mode particles over the land are similar to

results from measurements in other places. However, experimental data for the Aitken

and nucleation mode from other places are not available. Compared to marine

environment where particle emissions were observed (Nilsson et al., 1999) the land

surface provide a particle sink. Forest acts as a sink for aerosols, decreasing number

concentration of particles in the atmosphere, especially in Aitken and nucleation mode.

Even thought the actual mass in those modes is negligible, the number flux is important

since the particle number influences radiative balance of Earth. They may be potentially

important as it is unknown how forest growth depends on nutrient of pollutant uptake via

particle deposition.

This work contributes significantly at studies of the spatial variability and at the general

removal of particulate from the atmosphere by ecosystems. Similar flux measurements as

conducted during the studies should be extended to different surface environments to

understand importance of different land-atmosphere interfaces in removal of particulates

from atmosphere.


29/32

27

5. Review of publications

This thesis consists of the experimental research work concerning the measurements of

variation in the aerosol number concentration conducted under laboratory conditions andat the field. All instrument calibration and test measurements, before and after the field

campaigns, were done in the laboratory. Some of the measurements were done in the

urban environment, (in Helsinki city) to investigate the spatial variation of the aerosol

number concentration. Other field measurements were conducted above the boreal forest

(Hyytil site, Southern Finland) using a novel set-up to measure particle fluxes.

Paper1 reports about the studies on the spatial variability of the aerosol number

concentration in Helsinki city. The aim of the paper was to find out in how many

locations in the city aerosol concentration has to be measured in order to have estimates

of the overall number concentration changes. This result is needed for epidemiologists to

study relationship between the air pollution levels and the health effects, respectively to

the number concentration of the particulates. The work was done within the framework of

the Exposure and risk assessment for fine and ultrafine particles in ambient air

ULTRA-project. The result showed the high spatial correlation of the aerosol number

concentration, due to homogeneous pollution in the whole downtown area. The pollution

mainly was dominated by emissions from the traffic. Very probably it is enough to haveone site to monitor aerosol concentration in Helsinki city, but the location of the site must

be chosen carefully.

Paper 2 contains the results from the laboratory studies on the operation of the modified

condensation particle counter TSI 3010. The CPC was modified to increase the sampling

flow rate in order to increase the sampling time resolution. Changes in the cut-off size

due to the increase in the flow rate were studied in respect of the temperature difference

between the saturator and the condenser.

Recent application of the CPC in fast aerosol size distribution and particle flux

measurements requires knowledge of the CPC time response or the frequency transfer

function. Therefore this study was designed and implemented. The result provided the

time constant for the CPC and the transfer function assuming that the CPC is the first-


30/32

28

order-response instrument. The assumption is reasonable and was confirmed by the

experimental results.

Paper 3 describes the application of the CPC model TSI 3010 in the eddy covariance

method to measure fluxes. State-of-the art experiments were conducted above the boreal

Scots pine forest. FFT analysis was used to study the fluctuations frequency range in the

aerosol number concentration time series. This paper is the first available report about the

aerosol number fluxes for particles with sizes down to 10 nm.

Paper 4 reports on the analysis of particle flux measurements above the Scots pine forest.

Continuous measurements at the forest site particle size distribution data were used to

characterize fluxes according to particle size. Periods with the prevailing nucleation mode

were analysed in particular and the deposition velocities for this size particles were

estimated. This is the first EC experimental measurement of the deposition velocity ofparticles with 10 20 nm sizes depositing to the forest canopy. Additionally, a pollution

source at Southwest direction was identified from the measurement results. This paper

describes how to deal with different systematic uncertainties in the particle EC

measurements.

6. References

Aalto, P., Hmeri, K., Becker, E., Weber, R., Salm, J., Mkel, J., Vkev, M., Koponen,I. and Hoell, C., 2000; Aerosol physical properties of aerosol particles duringnucleation events, to be submitted to Tellus.

Businger J. A., 1986: Evaluation of the Accuracy with Which Dry Deposition Can BeMeasured with Current Micrometeorological Techniques,J. Clim. Appl. Meteor., 25,1100 - 1124.

Charlson, R. J. and Wigley, T. M. L., 1994: Sulfate aerosol and climate change, Scientific

American, 23, 48 57.Dockery, D., W. and Pope, C., A., 1994: Acute respiratory effects of particulate airpollution.Annu. Rev. Public Health, 15, 107-132.

Doebelin, E. O., 1990: Measurement systems. McGraw-Hill Publishing Company, 104 194.

Duan, B., Fairall, C. W. and Thomson, D. W., 1988: Eddy Correlation Measurements ofthe Dry Deposition of Particles in Wintertime,J. Appl. Meteor., 27, 642-652.


31/32

29

Erisman, J. W., Draaijers, G., Duyzer, J., Hofschreuder, P., Leeuwen, N. V., Rmer, F.,Ruijgrok, P., Wyers, P. and Gallaghger, M., 1997: Particle deposition to forests summary of results and applications,Atmospheric Environment, 31, 321 332.

Fairall, C. W., 1984: Interpretation of Eddy-Correlation Measurements of ParticulateDeposition and Aerosol Flux,Atmospheric Environment, 18, 1329 - 1337.

Gallagher, M. W., Beswick, K. M., Duyzer, J., Westrate, H., Choularton, T. W. andHummelshoj P., 1997: Measurements of Aerosol Fluxes to Speulder Forest Using aMicrometeorological Technique,Atmospheric Environment, 31, 359 - 373.

Hinds, W. C., 1982: Aerosol technology,John Wiley & Sons, Inc.Horst, T. W., 1997: A simple formula for attenuation of eddy fluxes measured with first-

order-response scalar sensors,Boundary Layer Meteorology, 82, 219 233.Hmeri, K., Vkev, M., Aalto, P., Kulmala, M., Swietlicki, E., Seidl, W., Becker, E. and

ODowd, C., 2000: Hygroscopic and CNN properties of aerosol particles in borealforests, to be submitted to Tellus.

IPCC, 1996: Intergovernmental panel on climate change, Climate Change 1995, edited by

J. T. Houghton et al., Cambridge.Juozaitis, A., Trakumas, S., Sopauskiene, D., Mikelinskiene, A., Girgzdys, A. and

Ulevicius, V., 1996; Pecularities of atmospheric aerosol formation in marineenvironment, Proceed. of Nucleation and Atmospheric Aerosol 1996 conference.

Kaimal, J. C. and Finnigan, J. J., 1994: Atmospheric boundary layer flows, Oxford Univ.Press.

Lamaud, E., Brunet, Y., Labatut, A., Lopez, A., Fontan, J. and Druilhet, A., 1994:TheLandes Experiment: Biosphere - Atmosphere Exchanges of Ozone and AerosolParticles Above a Pine Forest,J. Geophys. Res., 99, 16511 - 16521.

Laubach, J. and Teichmann, U., 1996: Measuring energy budget components by eddy

correlation: data corrections and application over low vegetation,Beitr. Phys. Atmos.,69, 307 320.

Kulmala, M., Toivonen, A., Mkel, J. M. and Laaksonen, A., 1998: Analysis of growthof nucleation mode particles observed in Boreal forest, Tellus, 50B, 449 462.

Mertes S., Schroder F. and Wiedensohler A., 1995: The Particle Detection EfficiencyCurve of the TSI-3010 CPC as a Function of Temperature Difference BetweenSaturator and Condenser, Aerosol Sci. Technol., 23, 257 - 261.

Mkel, J.M., Koponen, I.K., Aalto, P. and Kulmala, M., 2000: One-year data ofsubmicron size modes of tropospheric background aerosol in southern Finland,J.Aerosol Sci. 31: 595 - 611.

Mkel, J. M., Aalto, P., Jokinen, V., Pohja, T., Nissinen, Palmroth, A. S., Markkanen,T., Seitsonen, K., Lihavainen, H., and Kulmala, M., 1997: Observations of UltrafineAerosol Particle Formation and Growth in Boreal Forest, Geophysical ResearchLetters, 24, 1219 - 1222.

Neumann, H. H. and den Hartog, G., 1985: Eddy correlation measurements ofatmospheric fluxes of ozone, sulphur, and particulates during the Champaignintercomparison study,J. Geophys. Res., 90, 2097 2110.


32/32

Nilsson, E. D., Rannik, ., Buzorius, G., Boy, M. and Laakso, L., 2000: Effects of thecontinental boundary layer evolution, convection, turbulence and entrainment onaerosol formation, submitted to Tellus.

Nilsson, E. D., Kulmala, M., Leck, C., Rannik, ., Swietlicki, E. and Bigg, K., 1999: Canaerosol particle fluxes from open sea and leads be a CCN source for Arctic Ocean

fogs?,J. Aerosol Sci., 30, S89 S90.ODowd, C., McFiggans, G., Creasey, D. J., Pirjola, L., Hoell, C., Smith, M. H., Allan,

B. J., Plane, J. M. C., Heard, D. E., Lee, J: D., Pilling, M. J. and Kulmala, M., 1999: Onthe photochemical production of new particles in the coastal boundary layer, Geoph.Res. Letters, 26, 1707 1710.

ODowd, C. D., Smith, M. H., Consterdine, I. E. and Lowe, J. A.: Marine aerosol, sea-salt, and the marine sulphur cycle: a short review, 1997: Atmospheric Environment, 31,73 80.

Russell, L. M., Flagan, R. C. and Seinfeld, J. H., 1995: Asymmetric instrument responseresulting from mixing effects in accelerated DMA-CPC measurements,Aerosol Science

and Technology,23, 491 509.Schery, S. D., Wasiolek, P. T., Nemetz, B. M. and Yarger, F. D., 1998: Relaxed eddy

accumulation for flux measurements of nanometer-size particles,Aerosol Sci. Techn.,28, 159 172.

Seaton, A., MacNee, W., Donaldson, K., Godden, D., 1995: Particulate air pollution andacute health effects.Lancet, 345, 176-78.

Seinfeld, J. H. and Pandis, S. N., 1998: Atmospheric chemistry and physics, from airpollution to climate change,John Wiley & Sons, Inc.

Shaw, W. J., Spicer, C. W. and Kenny, D. V., 1998: Eddy correlation fluxes of tracegases using a tandem mass spectrometer, 32, 2887 2898.

Sievering, H., Eastman, J. and Schmidt, J. A., 1982: Air-Sea Particle Exchange at aNearshore Oceanic Site,J. Geoph. Res., 87, 11027 11037.Slanina, S., 1997:Biosphere-atmosphere exchange of pollutants and trace substances,

Springer, 54 66.Slinn, W. G. N., 1982:Predictions for particle deposition to vegetation canopies,Atmospheric Environment, 16, 1785 1794.

Stull, R. B., 1999: Boundary layer meteorology, Kluwer academic, publishers.Suh, H. H., Nishioka, Y., Allen, G. A., Koutrakis, P. and Burton, R., M., 1997: The

Metropolitan Acid Aerosol Characterisation Study: Results from the summer 1994Washington, D. C. Field Study,Environmental Health Perspektives, 105, 826 834.

Vkev, M., Hmeri, K., Kulmala, M., Lahdes, R., Ruuskanen, J. and Laitinen, T., 1999:Street level versus rooftop concentration of submicron aerosol particles and gaseouspollutants in an urban street canyon,Atmospheric Environment, 33, 1385 1397.

Wesely, M. L., Cook, D. R. and Hart, R. L., 1983: Fluxes of gases and particles above adeciduous forest in wintertime,Boundary Layer Meteorology, 27, 237 255.

Velici Variation of Aerosol Concentration (2)

Documents

Transcript of Velici Variation of Aerosol Concentration (2)