ne
Israe
14 July 2009
Accepted 16 July 2009
Keywords:
PM10
PM2.5
Pollution level
Dust increase
Chemical and mineralogical composition
Tel-Aviv Israel
health standard for airborne particles is given in weight in micro-
grams per cubic meter. The Israeli PM10 standard is 60 mgm3
year1 and 150 mgm3/24-h. The Israeli recommended PM2.5
standard is 15 mgm3 year1 and 65 mgm3/24-h. The European
Commission goal for 2010 (EC Directive, 1999) is PM10 of 20 mgm3
2006 for Tel-Aviv, Israel, in the same range as that found for other
cities in the region.
Natural mineral dust storms (DS) from the Arabo-African region
blow over the Mediterranean and reach Israel, Lebanon, Syria,
Turkey, Cyprus, Greece, Italy,Malta, France, Spain, can reach England
and have even been reported in satellite images over Norway and
Odessa (Barkan, 2009). DS reach Israel mostly during the transition
and winter seasons (Mamane et al., 1980). The mineral dust adds
directly to the anthropogenic particulate pollution, and the mineral
* Corresponding author. Tel./fax: þ972 3 6405205.
E-mail addresses: ganor@post.tau.ac.il (E. Ganor), amnons@post.tau.ac.il
Contents lists availab
Atmospheric E
lse
Atmospheric Environment 43 (2009) 5463–5468(A. Stupp).Particulate air pollution has increased over the past decades in
many cities of the world. Particles with aerodynamic diameter less
than 10 mm (PM10) are considered inhalable particles, which can
reach the lungs. Particles above 2.5 mm and below 10 mm penetrate
to the bronchi, while particles below 2.5 mm (PM2.5) are respirable
particles and penetrate to the alveoli (Heyder, 1986). Inhalable
particles are considered amajor health hazard (Dockery et al., 1993;
Schwartz, 1994; Prospero, 1999; Annesi-Maesano et al., 2007). The
Previous studies have shown annual PM10 concentrations of
44–53 mgm3 for major Italian cities (Galassi et al., 2000), 18–
41 mgm3 depending on station in Berlin (Lenschow et al., 2001),
51 mgm3 at Heraklion, Crete and only 28 mgm3 at Finokalia,
Crete (Gerasopoulos et al., 2006), and upto 75 mgm3 at Athens
(Chaloulakou et al., 2003). In Beirut, Lebanon, monthly average of
76 mgm3 was measured (Shaka and Saliba, 2004). In this study we
find PM10 annual averages of 47–68 mgm3 for the years 1995–1. Introduction1352-2310/$ – see front matter  2009 Elsevier Ltd.
doi:10.1016/j.atmosenv.2009.07.028require any other inputs such as satellite observations, model back-trajectories, dust forecast models, or
mineralogical analyses. The method employs an automatic algorithm with three thresholds: the half-
hour PM10 average must be above 100, this level is maintained for at least 3 h, and the maximum
concentration recorded is above 180 mgm3. The algorithm was designed for Israel, but can be adapted
for other locations.
The contribution of DS caused PM10 values to exceed the Israeli annual standard of 60 mgm3 year1
in 6 of the 12 years examined. The DS contribution to PM10 annual average ranged from 9.4% to 29.5%.
The level recommended by WHO, 20 mgm3 year1, was exceeded every year even without the DS
contribution. The number of days in which the daily Israeli standard (150 mgm3) was exceeded during
the 12 years was 6–20 days per year. The number of days in which the daily standard was exceeded
shows an increasing trend of 7 days per decade.
PM2.5 in Israel is in the range 40–56% of PM10. PM2.5 values were over the recommended standard
with and without DS. The contribution of DS to annual average of PM2.5 ranged from 3.6% to 19.1%.
The automatic algorithmwas calibrated with a list of Dust Storms identified by visual means supported
by mineralogical analysis. Mineralogical analyses of single particles were performed using Environ-
mental Scanning Electron Microscope (ESEM). Two representative samples are given. The main
difference is that the particles of the Saudi-Arabian storm had much more palygorskite, while the North-
African storm had more sea-salt and organic particles. The mineral composition differences indicate that
analysis can differentiate between sources.
 2009 Elsevier Ltd. All rights reserved.
annual mean with natural sources excluded. The level recom-
mended by WHO for PM10 is 20 mgm3 year1 (WHO, 2006).Received 6 May 2009
Received in revised formIsrael, and add to the anthropogenic particulate pollution. The effects of mineral dust on air quality in
Israel were investigated using only PM10 and PM2.5 automatic measurements. The method does notArticle history: Natural mineral dust storms (DS) from the Arabo-African region blow over the Mediterranean, reachA method to determine the effect of mi
Eliezer Ganor, Amnon Stupp*, Pinhas Alpert
Department of Geophysics and Planetary Sciences, Tel-Aviv University, Tel-Aviv 69978,
a r t i c l e i n f o a b s t r a c t
journal homepage: www.eAll rights reserved.ral dust aerosols on air quality
l
le at ScienceDirect
nvironment
vier .com/locate/atmosenv

dust particles can also adsorb and become coated by anthropogenic
pollutants such as sulfates, nitrates, pesticides, PAH and heavy
metals (Levin et al., 1996; Falkovich et al., 2001, 2004).
The Israeli standard does not allow subtraction of the natural
contribution, as does the EU standard. However, it is important to
differentiate between the natural and anthropogenic contributions,
measurements from automatic stations. Once calibrated, the
method does not require any other input, either data or analysis.
natural mineral dust from other aerosol sources, using only auto-
matic PM10 concentration measurements and no other input.
The algorithm was designed based on our experience with DS in
Israel for over 50 years, leading to knowledge such as that a DS is
characterised by a minimal duration, large minimum concentration,
and large maximum concentration (Ganor et al., 2000, 1998; Ganor
mpo
d on
rsion
on
E. Ganor et al. / Atmospheric Environment 43 (2009) 5463–54685464The method produces a list of DS with beginning and ending times
at a half-hourly temporal resolution.
The usefulness of the method is demonstrated through the
investigation of the period 1995–2006, with 384 DS identified in
Tel-Aviv, and the implications of DS to air quality in Israel analyzed.
2. Methodology and measurements
Table 1 shows a qualitative division of levels and sources of
particulate pollution in Tel-Aviv, Israel. The table is based on daily
observations of visibility and weather conditions, and on
measurements at the Tel-Aviv university Israel Ministry of Envi-
ronment station, including chemical and mineralogical analyses
(Ganor et al., 2000). The table shows that during DS pollution
increases to above 150 mgm3 while the background is about
50 mgm3, and therefore distinguishing between DS and other
sources of PM is possible.
DS were identified manually at Tel-Aviv, and a list of DS for the
years 1958–2006 was prepared (Ganor et al., 2007). Our decades-
long knowledge of DS in Israel and the region enabled us to identify
DS from visual indicators – visibility, sky color, sample color, and
dry and wet deposition. Our confidence in this ability relies on
electron microscope analyses and mineralogical analyses which
were performed over decades, and which showed that indeed the
majority of PM10 during such episodes is from natural dust.
An automatic algorithm was developed, which for the first time
makes it possible to distinguish the contribution to PM due to
Table 1
Daily average concentration (mgm3), annual frequency, weather conditions, and co
Pollution level Concentration [mgm3] Frequency Weather condition
Low pollution 5–40 28% During and after Rain, an
Medium pollution 35–65 45% Summer smog, high inve
High pollution 65–150 24% Winter smog, low inversifor health considerations, and in order to determine the reduction
needed in the anthropogenic emissions.
The usual way of determining the natural contribution is via
chemical analysis of the aerosols, where particles with large frac-
tions of Al, Ca and other crustal components are attributed to
natural mineral dust. This method requires collection of aerosols on
filters, frequent replacement of the filters, at best daily, and costly
analysis of tens or hundreds of filters. Other methods involve back-
trajectories and satellite images to track dust plumes from the
sources. Recently Escudero et al. (2007) and Mitsakou et al. (2008)
have quantified the contribution of mineral dust to air quality in
Spain and in Greece, using a methodology different from the one
presented here. Escudero’s method relies on back-trajectories and
other data (satellites, synoptic) to identify DS, and also chemical
analysis to verify this identification. Then the measured daily
average PM10 for the days without DS is used to estimate PM10
background, and from that the contribution of DS is calculated.
Mitsakou’s method is model based, using the SKIRON (DREAM)
model to estimate contribution of DS to daily average PM10
measurements.
We present a method for identification of DS using only PM10Very high pollution 100–3000 3% Dust Stormsand Foner, 2001; Falkovich et al., 2001, 2004; Ganor, 1999; Pardess
et al.,1992; Levin et al.,1990). The algorithmwas calibrated using the
list ofmanually identifiedDS for one particular year. After calibration
the algorithm was used on a different year and the results were
compared with the manual identification for that year, with very
goodagreement. AllDSon themanual list appearedon the automatic
list,withsomeadditionalDS identifiedwith theautomaticalgorithm.
The automatic algorithm uses three thresholds. An episode is
identified as DS only if: the half-hour PM10 average is above 100,
this level is maintained for at least 3 h, and the maximum
concentration recorded is above 180 mgm3.
PM10 was measured from 1995 to 2006 by the TEOM series
1400 in the Tel-Aviv automatic stations of the Israel Electric
Company. The record of the measurements is of half-hourly aver-
ages. Using the algorithm on the record produces a list of DS with
a half-hour resolution of the beginning and end of each DS. The list
of DS produced by the algorithm makes it possible to analyze the
contribution of DS to air quality in Tel-Aviv over this period.
The contribution of DS to PM10 for any time period is calculated
as follows:
A time period is defined by the number of half-hour records
within it nT.
For every half-hour i there is a PM10 concentration
PM10i
The total PM10 during a time period is therefore
T ¼
XnT
i¼1
PM10i
The average PM10 for a time period is
avg ¼ T=nT
During a time period one or more DS might be identified. The
number of records during the DS or DS-s in a time period is nDS.
The total PM10 during DS which occurred during a time period
is therefore
TDS ¼
XnDS
i¼1
PM10i
The average PM10 without DS is
avgWithoutDS ¼ T  TDS
nT  nDS
Therefore the contribution of DS to the average PM10 during this
time period is
DS contribution ¼ avg  avgWithoutDS
The automatic method makes it possible to analyze a large
number of DS over a longmeasurement period, but themethod also
sition of inhalable particles, PM10, in Tel-Aviv (Ganor et al., 2000).
Composition
Yom-Kipur Black carbon, Minerals
Sulphates, Black Carbon, Sea Spray, Nitrates, Minerals
Black Carbon, Salts of Sulphates and Nitrates, Heavy Metals, Minerals
Pollen, Minerals, Sea Spray