Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
Atmospheric
aerosol
contribution
to
visible
light absorption and scattering in Mexico City
and simple tools to determine refractive indices
and size distributions
S. Eidels-Dubovoi
Institute Mexicano del Petroleo, Mexico
Abstract
Mie theory is used to calculate diurnal aerosol visible light (X = 0.50pm.)
absorption and scattering patterns from particle size distributions measured in the
diameter range 0.006 - 1.0pm., during February 16 - March 1, 1991, at three
different sites in the Mexico City Valley. The calculated patterns showed
variations even from one day to another at the same place but reaching the
highest values in the morning between 9:07 hrs. and 10:40 hrs. and the lowest
around 13:40 hrs. or 16:40 hrs. was the predominating trend. The degree at
which aerosol light absorption contributes to the total aerosol light extinction is
found to be of the order of that measured in highly polluted cities while the
scattering contribution is found to be great enough to produce a cooling effect on
the atmosphere. Further the influence of size distribution and refractive indices
on the diurnal aerosol visible light absorption and scattering patterns is examined
and a method for determining both parameters is suggested.
1
Introduction
The loss of visibility is due to the scattering and absorption of light by pollutants.
Particulate matter suspended in air is generally responsible for the majority of
light scattering and absorption and hence for visibility reduction associated with
air pollution. Diurnal aerosol visible light scattering and absorption patterns then
become a necessary tool in characterizing polluted atmospheres and aerosols
themselves. Indeed these patterns are closely related to particle size distribution
Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
710 Air Pollution VIII
and refractive indices and can be obtained either from field observations or from
calculations of light scattering, bsp, and absorption, bap, coefficients.
In this paper Mie theory is used to calculate diurnal aerosol visible light (X =
0.50pm.) absorption and scattering patterns from particle size distributions
measured, in the diameter range 0.006 - 1.0pm., during February 16 - March 1,
1991, at three different sites in the Mexico City Valley. Further, the size
distribution and refractive indices influence on the diurnal aerosol visible light
absorption and scattering patterns is examined. It is found that their structure, i.e.
the frequency at which maximum and minimum values occur during the day,
depends on thefirstparameter while their peak heights or magnitudes depend on
the second one. It follows then, that particle size distribution and refractive
indices can be determined by fitting the calculated bap and bsp patterns to the
measured ones. The structure of the patterns obtained for ten days of the
observational period went from simple i.e., few peaks separated by long time
intervals (Figs, la, b), to complex i.e., lot of sharp peaks separated by short time
intervals (Fig. Ic), being the simple structure with maximum values in the
morning the predominating one.
2
Data
The UNAM Geophysics Institute collected the radiometric and atmospheric
aerosol particle data [1] for the Mexico City Air Quality Research Initiative
during a major measurement campaign of two weeks from February to March
1991.
Measurements were performed approximately every half an hour from 9:00 hrs.
to 18:00 hrs. local time, at the following sites and dates:
a) "Valle de Mexico" Thermoelectric Power Plant, Estado de Mexico, (19.62N,
98.97W), Feb. 16-17. The power plant is located northeast of the city center
at a side of a high transited highway which crosses from east to west. To the
north of this site there are big depopulated zones while to the south there are
urban and industrial zones.
b) Wilfrido Massieu Stadium, IPN, Zacatenco, D.F. (19.50N, 99.20W),
Feb. 19-23. The Wilfrido Massieu Stadium belongs to the Institute
Politecnico Nacional (IPN) devoted to superior technical education and is
located north of the city center within a major industrial area. Northeast of
this site, a zone with very heavy traffic can be found.
c) Xochimilco Stadium, Xochimilco, D.F. (19.25N, 99.11W), Feb.26-Marchl.
The Xochimilco Stadium is located south of the city center and is
surrounded by a very heavy traffic road at the north; an agricultural terrain at
the south; a residential zone at the southwest; popular colonies, restaurants
and a commercial zone at the southeast.
The distribution of particle number concentration relative to size for ten intervals
in the diameter range 0.0032 - l.Ojum. was recorded with a TSI model 3030
electric analyzer. Due to the inherent limitations of this equipment, only data in
the range 0.0056 - 1.0pm. comprising nine intervals centered at: 0.0078pm.,
0.0138pm., 0.0248pm., 0.044pm., 0.078pm., 0.138pm., 0.248pm., 0.44pm. and
Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
Air Pollution VIII 711
0.78p.m., were taken into account. The initial and final data were recorded at a
different time each day, being 9:06 hrs. the earliest and 17:33 hrs. the latest one.
3
Procedure
The extinction bep and scattering bsp coefficients were calculated using Mie
theory in a computational program based on the Lentz algorithm to estimate
Bessel functions and the absorption coefficient bap was obtained as a difference
of the former two. These calculations were performed using the size distribution
data of ten days of the observational period and a complex refractive index m =
1.65 - 0.017i. The real part, mr, of this index is that of a typical tropospheric
aerosol [2] while the imaginary part, mi, is the average often mi values obtained
for the ten studied days with an original method [3]. Since measurements were
performed during a dry season, no relative humidity effects were considered for
these calculations. Indeed, the value 0.017 obtained for mi was close to the 0.02
typical value associated to a dry atmospheric aerosol [4].
In order to examine the influence of the refractive index, m, on the diurnal
aerosol visible light (1 = 0.50pm.) absorption and scattering patterns, Mie
calculations were performed for different mr and mi values keeping fixed the
particle size distribution. The dependence of these patterns on the particle size
distribution was studied using the same refractive indices and different size
distribution data in the Mie calculations.
4
Results and discussion
Diurnal average of the recorded particle number concentration per unit size
interval peaked sharply at the lowest diameter interval centered at 0.0078 jim.
The overall ten days average concentration of particles within this diameter
interval was very high: 1.21x10^ particles/cm^. In some of the ten studied days,
either the third or the third and/or fifth diameter intervals, centered respectively
at 0.0248pm. and at 0.078pm., exhibited smaller concentration peaks. All these
three diameters are lower than 0.08(im. i.e., than the upper limit of the so called
"nucleation" range which is associated with particles that are emitted directly
from combustion sources and are detected only when fresh emissions sources are
close to the measurement site [5].
Three of the ten calculated bsp and bap diurnal patterns, each of them for one of
the three sites, are shown in figure 1. The absorption features were very similar
to those of the scattering features in all of the ten days patterns. The scattering
and absorption coefficients reached their maximum values 1.307 km"* and 0.145
km"* respectively the 23 of February at 10:40 hrs. and their minimum values
0.031 km"' and 0.003 km ' respectively on the 27 of February at 16:36 hrs.
The February 16-17 patterns of the Thermoelectric Power Plant site showed
opposite features, i.e.; the maximum/minimum values of the February 16
patterns appeared at approximately the same hour as the minimum/maximum
values of the February 17 patterns. The variation in the patterns from one day to
Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
712 Air Pollution VIII
bap( km'')
bsp (km"')
0.8.
0.12
16/2/91
0.10
0.6-
0.08
0.06
0.4.
0.04
0.2.
0.02
0.0,
10:00
1.4
11:00
12:00 13:00 14:00 15:00 16:00
LOCAL TIME ( hrs.)
17:00
bap (km-*)
bsp (km
0.18
23/2/91
0.16
1.2.
(b)
0.14
l.O.
0.12
0.8-
0.10
0.6.
0.08
0.40.06
0.2.
0.04
'lOOO
11:00
12:00 13:00 14:00 15:00
LOCAL TIME ( hrs. )
16:00
Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
Air Pollution VIII 713
bsp (km"*)
27/2/91
0.120.10-
bap (km"*)
(c)
0.020
0.018
0.016
0.014
0.080.012
0.06-
0.010
0.04-
0.008
0.02-
0.006
0.004
0.00
10:00
11:00 12:00 13:00 14:00 15:00 16:00
LOCAL TIME ( hrs.)
17:00
Figure 1: Diurnal aerosol visible light (X = O.SO^im.) scattering and absorption
patterns at the:
a) Valle de Mexico Thermoelectric Power Plant.
b) Wilfrido Massieu Stadium.
c) Xochimilco Stadium.
another can be attributed to a wind effect but these results have to be handled
with care because of the limited number of data and the difference in time of the
last records; 16:38 hrs. for the first day and 13:36 hrs. for the second day.
A general trend of very high bsp and bap values in the morning decreasing at
midday and fluctuating in the afternoon was appreciated in the diurnal variation
of the four days (February 19-23) patterns, obtained for the Wilfrido Massieu
Stadium site. Indeed, the maximum values were achieved between 9:07 hrs. and
10:40 hrs. while the minimum values were reached approximately either
between 12:36 hrs. and 13:42 hrs. or between 15:36 hrs. and 16:38 hrs. Other
prominent peaks appeared at 11:08 hrs. and 14:36 hrs. in the February 21
patterns and at 15:06 hrs. in the February 23 patterns.
The four days (February 26-March 1) patterns, calculated for the Xochimilco
Stadium site, could be classified as simple and complex. The February 26 and
the March 1 patterns were of the first category, attaining maximum values at
10:00 hrs. and 9:52 hrs., respectively, and minimum values at 16:36 hrs. and
12:36 hrs., respectively, while the February 27 and 28 patterns were of the
second category, showing maximum values at 12:36 hrs. and 11:06 hrs.
respectively and minimum values at 16:36 hrs. and 13:36 hrs. respectively.
Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
714 Air Pollution VIII
During the sampling period at this site, strong local dust storms were observed
[1] which could explain the diversity encountered on the four days patterns.
Indeed, winds with velocities greater than 5 m/s, starting in the afternoon, were
recorded during these four days by the automated monitoring network (RAMA)
at the Pedregal and Cerro de la Estrella stations located respectively 12.92 km.
northwest and 9.99 km. northeast of the Xochimilco Stadium site. Even more,
those winds of February 27, the day which had the most complex diurnal
patterns of all the four days, lasted a longer time within the afternoon particle
size distribution measurement period and reached velocities up to 12.8 m/s at the
Cerro de la Estrella station.
It is well known that elemental carbon is essentially responsible for aerosol
visible light absorption [6-9]. Elemental carbon aerosol is a primary emission
from combustion sources dominated in the Mexico basin by motor vehicles.
Thus, the high bap values obtained in the morning may be attributed to the onset
of the morning rush hour. On the other hand, aerosol light scattering, the
dominant contributor to total light extinction, is caused principally by inorganic
and organic aerosols which are formed photochemically, such that the magnitude
of aerosol light scattering is much more dependent on meteorological conditions
[10]. Hence, the high bsp values obtained around 13:00 hrs. may be attributed to
the onset of much photochemistry.
The appearance of prominent peaks at different times in the ten studied diurnal
patterns could mean that primary as well as secondary aerosols which depend
more on meteorological conditions, contribute significantly to the Mexico City
Valley pollution at any time of the day. Indeed, the contribution of aerosol light
absorption to the total aerosol light extinction defined as bap/bep, varied from
0.064 at 13:37 hrs. on February 16 to 0.126 at 10:06 hrs. on February 27, being
the average value of the ten days 0.099 comparable to the one 0.11 ± 0.04
measured in the Los Angeles basin and similar to most data for other highly
polluted urban areas [10]. The bap/bep diurnal average reached the minimum
value of 0.098 on February 21 at the Wilfrido Massieu Stadium and the
maximum value of 0.103 on February 17 at the Thermoelectric Power Plant. On
the other hand, the contribution of aerosol light scattering to the total aerosol
light extinction or single scattering albedo defined as bsp/bep, varied from 0.874
at 10:06 hrs. on February 27 to 0.936 at 13:37 hrs. on February 16 being the
average value of the ten days 0.90 greater than the 0.85 value beyond which
cooling effects are expected to occur [11]. The bsp/bep diurnal average reached
the minimum value of 0.897 on February 17 at the Thermoelectric Power Plant
and the maximum value of 0.902 on February 21 at the Wilfrido Massieu
Stadium. Visibility, as given by the relationship 3.912/bep, varied from 2.69 km.
at 10:40 hrs. on February 23 to 113.53 km. at 16:36 hrs. on February 27 being
the average value of the ten days 19.37 km. indicative of clear conditions
according to the International Visibility Code [12-13]. The Visibility diurnal
average reached the minimum value of 6.13 km. on February 23 at the Wilfrido
Massieu Stadium and the maximum value of 50.01 km. on February 27 at the
Xochimilco Stadium.
Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
Air Pollution VIII 715
bsp
1.421/2/91
m = mr- 0.015 i
1.2-
mr=1.7
mr=1.3
1.0.
0.80.60.40.2-
0.0
O9':00 10':00 ll':00 12':00 13':00 l4:00 l£o6 l&OO l7:00
LOCAL TIME (hrs. )
bap (km"')
0.14
21/2/91
0.12.
0.10-
mi-0.015
mi = 0.005
0.0&
0.06
0.04
0.02
0.00
09 :00 ld:00 ll'rOO 1J:00 13':00 l4:00 1^:00 1(4:00
LOCAL TIME (hrs.)
Figure 2: a) Visible light (1 = 0.50|im.) scattering pattern variation with the real
refractive index part, mr.
b) Visible light (X = 0.50p.m.) absorption pattern variation with the
imaginary refractive index part, mi.
Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
716 Air Pollution Vlll
Figures 2a and 2b illustrate, respectively, the increase in magnitude of the bsp
pattern with the mr increment and that of the bap pattern with the mi increment,
while the pattern structures remain almost the same in both cases.
Different bap and bsp structures can be obtained using different size
distributions and the same refractive index in the calculations as shown in Fig. 3.
It is then concluded that the refractive index determine the peak heights of the
visible light scattering and absorption patterns while the size distribution
determine their structure. Since the same size distribution was used in the
calculations of the bsp and bap patterns i.e., the one measured for each day, it is
not surprising that both patterns showed similar structures. Hence, time delays of
bsp peaks relative to bap peaks as those observed infieldstudies [ 10] cannot be
obtained from this theoretical study and the primary or secondary nature of the
aerosols can hardly be inferred. However, when data of bsp and bap are
available, which is not our case, it is possible to determine the refractive index
and the particle size distribution by fitting the calculated bap and bsp patterns to
the measured ones. Even more, a simple comparison between the structure of the
observed bsp and bap patterns can reveal differences between the size
distributions of the scattering and absorbing particles. It follows then, that the
analysis of both theoretical and experimental diurnal visible light bsp and bap
patterns, become a valuable tool for determining aerosol properties as well as
their nature and impact on visibility.
bap (km *)
0.12.
bsp (km
—.— bsp 28/2/91
—o— bap 21/2/91
0.35
0.30
0.100.25
0.0&
0.20
0.06
0.15
0.040.10
0.02.
0.05
09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00
LOCAL TIME ( hrs.)
Figure 3: Diurnal aerosol visible light (k = 0.50fj,m.) scattering and absorption
patterns calculated, respectively, for the February 28 and 21 particle
size distribution data.
Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
Air Pollution VIII 717
5
Summary and conclusions
Ten diurnal aerosol visible light (X = O.SOjim.) absorption and scattering patterns
using particle size distributions measured in the diameter range 0.006 - l.Ofim.,
during February 16 - March 1, 1991, at three different sites in the Mexico City
Valley were calculated from Mie theory. These patterns showed simple and
complex structures and varied even from one day to another at the same place.
Simple structures, with maximum values appearing at the first or second record
in the morning when primary vehicle emissions are elevated, predominated. The
more complex structures corresponding to the February 17, 27 and 28 patterns
exhibited their maximum values when much photochemistry onsets at 13:36 hrs.,
12:36 hrs. and 11:06 hrs., respectively. Atmospheric aerosol contribution to
visible light absorption and scattering was found to be maximum (minimum
visibility) on February 23 at 10:40 hrs. in the Wilfrido Massieu Stadium site and
minimum (maximum visibility) on February 27 at 16:36 hrs. in the Xochimilco
Stadium site. Even the overall ten days visibility average revealed clear
conditions; the overall ten days bap/bep average reflected the fact that pollution
due to atmospheric aerosol particles is significant in the Mexico City basin and
the overall ten days bsp/bep average indicated that cooling effects took place.
It was also shown that according to Mie theory, the refractive index determine
the peak heights or magnitudes of the visible light scattering and absorption
patterns while the size distribution determine their structure. This result
suggested in turn that both parameters can be obtained by fitting the calculated
bap and bsp patterns to the measured ones. The combination of observed and
calculated bsp and bap patterns then emerged as a simple powerful tool for
investigating aerosol properties as well as their nature and impact on visibility.
Acknowledgments
This work was supported by Petroleos Mexicanos (PEMEX) under Mexican
Petroleum Institute (IMP) research project: "Investigacion sobre Materia
Particulada y Deterioro Atmosferico" (IMADA).
References
[1] Muhlia, A. Reporte Tecnico del Monitoreo de Pardmetros Radiacionales y
Muestreo de Aerosoles en Cuatro Puntos del Area Metropolitana de la
Ciudad de Mexico, Institute de Geoflsica, UN AM: Mexico, D. F., 1991.
[2] Ivlev, L. S., and Popova. S. I. The complex refractive indices of substances
in the atmospheric-aerosol dispersed phase. Izv. Atmos. Oceanic Phys. 10,
pp. 587-591, 1973.
[3] Eidels-Dubovoi, S. Solar radiation attenuation by atmospheric aerosol
particles at different sites in the Mexico City Valley. International
Symposium on Heat and Mass Transfer in Energy Systems and
Environmental Effects, International Center for Heat and Mass Transfer,
Air Pollution VIII, C.A. Brebbia, H. Power & J.W.S Longhurst (Editors)
© 2000 WIT Press, www.witpress.com, ISBN 1-85312-822-8
718 Air Pollution VIII
Universidad Nacional Autonoma de Mexico, Institute de Ingenieria, UNAM,
pp. 140-142, 1993.
[4] Finlayson-Pitts B. J., and Pitts J. N., Jr. Atmospheric Chemistry:
Fundamentals and Experimental Techniques, John Wiley & Sons, Inc.: New
York, Chichester, Brisbane, Toronto, Singapore, p. 762, 1986.
[5] Watson J.G., and Chow J.C. Clear sky visibility as a challenge for society.
Annu. Rev. Energy Environ., 19, pp. 241-66, 1994.
[6] Rosen H., Hansen A. D. A., Gundel L and Novakov T. Identification of the
optically absorbing component in urban aerosols. Appl Opt., 17, pp. 38593861, 1978.
[7] Yasa Z., Amer N. M., Rosen H., Hansen A. D. A. and Novakov T.
Photoacustic investigation of urban aerosol particles. Appl Opt., 18, pp.
2528-2530, 1979.
[8] Japar S. M., Brachaczek W. W., Gorse R. A.,Jr., Norbeck J. M. and Pierson
W. R. The contribution of elemental carbon to the optical properties of rural
atmospheric aerosols. Atmos. Environ., 20, pp. 1281-1289, 1986.
[9] Adams, K. M., Davis L. I.,Jr, Japar S.M., Finley D.R., and Cary R.A.
Measurement of Atmospheric Elemental Carbon: Real - Time Data for Los
Angeles during Summer 1987. Atmos. Environ., pp. 597-604, 1990.
[10] Adams, K. M., Davis L. I.,Jr., Japar S.M., and Finley D.R. Real - time, in
situ measurements of atmospheric optical absorption in the visible via
photoacustic spectroscopy - IV. Visibility degradation and aerosol optical
properties in Los Angeles. Atmos. Environ., 24A, pp. 605-610, 1990.
[11] Shettle, E. P., and Fenn R. W. Models for the aerosols of the lower
troposphere and the effect of humidity variations on their optical
properties, Report AFCRL TR 79 0214, Air Force Cambridge Lab.,
Hambscome A. F. B., 1979.
[12] Hulburt, E.G. Optics of atmospheric haze. J. Opt. Soc. Atm., 31, pp.467476, 1941.
[13] McCartney, E. J. Optics of the Atmosphere, John Wiley & Sons: New York,
London and Toronto, p. 43, 1976.