ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-16-5063-2016Aerosol optical depth trend over the Middle EastKlingmüllerKlausk.klingmueller@mpic.dehttps://orcid.org/0000-0002-8425-8150PozzerAndreahttps://orcid.org/0000-0003-2440-6104MetzgerSwenhttps://orcid.org/0000-0002-3623-7160StenchikovGeorgiy L.LelieveldJoshttps://orcid.org/0000-0001-6307-3846Max Planck Institute for Chemistry, P.O. Box 3060, 55020 Mainz,
GermanyThe Cyprus Institute, P.O. Box 27456, 1645 Nicosia, CyprusKing Abdullah University of Science and Technology, Thuwal 23955-6900,
Saudi ArabiaKlaus Klingmüller (k.klingmueller@mpic.de)22April20161685063507316October201515January201612April201614April2016This work is licensed under a Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/This article is available from https://acp.copernicus.org/articles/16/5063/2016/acp-16-5063-2016.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/16/5063/2016/acp-16-5063-2016.pdf
We use the combined Dark Target/Deep Blue aerosol optical depth (AOD)
satellite product of the moderate-resolution imaging spectroradiometer
(MODIS) collection 6 to study trends over the Middle East between 2000 and
2015. Our analysis corroborates a previously identified positive AOD trend
over large parts of the Middle East during the period 2001 to 2012.
We relate the annual AOD to precipitation, soil moisture and surface
winds to identify regions where these attributes are directly related to the AOD
over Saudi Arabia, Iraq and Iran. Regarding precipitation and soil moisture,
a relatively small area in and surrounding Iraq turns out to be of prime
importance for the AOD over these countries. Regarding surface wind speed,
the African Red Sea coastal area is relevant for the Saudi Arabian AOD.
Using multiple linear regression we show that AOD trends and interannual
variability can be attributed to soil moisture, precipitation and surface
winds, being the main factors controlling the dust cycle. Our results confirm
the dust driven AOD trends and variability, supported by a decreasing
MODIS-derived Ångström exponent and a decreasing AERONET-derived fine
mode fraction that accompany the AOD increase over Saudi Arabia. The positive
AOD trend relates to a negative soil moisture trend. As a lower soil moisture
translates into enhanced dust emissions, it is not needed to assume growing
anthropogenic aerosol and aerosol precursor emissions to explain the
observations. Instead, our results suggest that increasing temperature and
decreasing relative humidity in the last decade have promoted soil drying,
leading to increased dust emissions and AOD; consequently an AOD increase is
expected due to climate change.
Introduction
The Middle East and the adjacent Mediterranean region have been identified as
a hot-spot of climate change . These studies indicate a relatively
strong precipitation decrease and temperature increase in the near future, which
would be especially important for an arid and desert dominated region such as the
Middle East. Since the middle of the 20th century, the annual number of
unusually hot days and nights in the Middle East has already increased whereas
the number of cool days and nights has decreased significantly
.
Beside the climatic importance of the region, the Middle East is also
outstanding from an atmospheric chemistry point of view. High levels of ozone
are expected due to the strong insolation acting on local pollution emissions,
in addition to long-range transport and stratospheric ozone intrusions
. Model projections of future
anthropogenic emissions suggest that this situation will likely exacerbate in
the near future although in the past few years pollution
emissions have decreased in the aftermath of economic crises and conflicts in
the region . Nevertheless, pollution levels are very
high, including aerosols that add to a high background of natural particles,
leading to more than 100 000 premature deaths per year in the Middle East
.
The Middle East is centrally located in the so-called dust-belt
and strongly influenced by natural sources. The high
atmospheric dust concentrations near the Earth's surface are reflected in
high aerosol optical depth (AOD) levels e.g.,.
Increasingly, the natural dust is supplemented by local anthropogenic
emissions associated with the economic and population growth on the Arabian
Peninsula which adds to pollution originating from Europe, Asia and Africa.
Polluted dust can significantly enhance atmospheric heating
, especially by intense incoming solar radiation
that is typical in the Middle East. On the other hand, chemical ageing in
polluted dust air masses transported over long distances can increase
deposition rates of dust particles . Various studies
show a strong increase in the AOD over the Middle East during the last
decade. These studies were performed with numerical models
and with different remotely sensed observational
data, e.g., from SeaWiFS , MODIS, MISR and
AERONET .
Different from previous studies, in this work we analyse trends using the
most recent (as of 2015) data collection 6 of the Moderate-resolution Imaging
Spectroradiometer (MODIS) which includes AOD data based on refined retrieval algorithms
, in particular the expanded Deep Blue algorithm. The
latter is of special importance for our study because it is suited for
retrievals over bright surfaces such as the deserts and semi-deserts covering
large parts of the Middle East. Among other improvements, the new data
version provides extended spatiotemporal coverage. In addition, it introduces
a merged AOD product, combining retrievals from the Deep Blue and Dark Target
algorithms to produce a consistent data set covering a multitude of surface
types ranging from oceans to bright deserts. A crucial aspect is the problem
of spurious AOD trends due to instrumental drift found in earlier MODIS
collections , which
has been addressed recently . Another difference to
previous studies is the extended time period considered: to verify the
persistence of the AOD increase, we take MODIS data up to 2015 into account.
Global AOD trends between March 2000 and February 2015, based on the
Dark Target/Deep Blue 550 nm AOD from MODIS Terra, collection 6.
Regions with significant trends (p value <0.01) are dotted, regions
with incomplete time series are coloured white. Despite being less
pronounced than for the period January 2001 to December 2010 considered
by (see also Fig. S1 in
the Supplement), the Middle East is a hot spot of AOD increase.
Comparable trends are only found in China. An exceptional trend over
the Aral Sea region defines the upper limit of the colour scale, but is
likely predominantly attributed to retrieval artefacts (see the
discussion in the Supplement).
A second objective of this study is to shed light on the causes of the
observed trends. In view of changing anthropogenic aerosol emissions,
a crucial question is whether these are responsible for the AOD trends or if
they are related to natural dust. The observed aerosol properties and the
number of dust events seem to favour the latter. Based on station data,
even define a regime shift around 2006 from an
inactive to an active dust period which they attribute to synergistic
interactions between the El Niño Southern Oscillation (ENSO) and the
Pacific Decadal Oscillation (PDO), superimposed onto a long-term drying trend
in the Fertile Crescent. To explore further the link between AOD trend and
increased dust activity, we relate the MODIS AOD to observations of major
parameters controlling the dust cycle: precipitation, surface wind and soil
moisture.
This article is structured as follows: in Sect. the
observed aerosol trends are presented and discussed. The AOD is related to
precipitation, soil moisture and surface wind speed individually in
Sects. , and . In addition,
trends of these variables are discussed. Finally, in Sect.
a multivariable linear model based on all three observables is used to
reproduce the MODIS AOD trends, leading to our conclusions in
Sect. .
Pattern of Middle East AOD trends between March 2000 and
February 2015 (left) and between January 2001 and December 2012 (right),
based on the Dark Target/Deep Blue 550 nm AOD from MODIS Terra,
collection 6. Pixels with significant trend (p value <0.01) are marked
with a dot. The same plot for the period January 2001 to December 2010 is
shown in Fig. S2 in the Supplement.
Aerosol trends
Figure presents a global overview of AOD trends over
15 years between March 2000 and February 2015. It has been computed from
the monthly values of the combined Dark Target/Deep Blue 550 nm AOD
from MODIS Terra, collection 6 . For
deseasonalisation, the annual cycle obtained by harmonic regression of sixth
order has been subtracted. The trend of the deseasonalised AOD has been
calculated by fitting a linear model using generalised least squares
. Consistent with ,
, and , the
1-month lag autocorrelation was used to account for the autocorrelation
structure of the time series, see Figs. S31 to S38 in the Supplement.
Strong and significant positive trends extending over large areas are found
in the Middle East, in particular the Arabian Peninsula. Comparably strong
trends are only found in China. Our analysis identifies even stronger trends
over the Aral Sea region. However, the shrinking Aral Sea not only exposes
new dust sources but also implies a constantly changing
surface reflectance making a consistent AOD retrieval over this region
extremely challenging and likely contributing a large spurious component to
the MODIS trend (A. M. Sayer, personal communication, 2016, see also the discussion in the Supplement). While the 10-year
trend from January 2001 to December 2010 considered by
(see also Fig. S1 in the Supplement) clearly
identifies the Middle East as having the strongest AOD increase worldwide, it
is surpassed by the corresponding trend over China when the period up to 2015
is also included.
Nevertheless, the Middle East is clearly a hot spot of AOD increase.
Magnifying this region, the left panel of Fig. reveals
large areas with an average increase in excess of 0.01 year-1, and
regional increase rates higher than 0.02 year-1. Over major parts
of Saudi Arabia the positive AOD trend is significant with probability (p)
values of the observations (using the trendless case as null hypothesis)
below the significance level of 1 %. Compared to the 10-year trend
from 2001 to 2010 (Fig. S2 in the Supplement), in particular over Iraq,
northern Saudi Arabia and the Persian Gulf the AOD increase is less
significant with p values exceeding 1 %.
In the following we mostly consider AOD values spatially averaged over the
territory of selected countries. The advantage is that boundaries of countries
are well defined, whereas regions of strong AOD trends, for example, might vary
over time. Country level trends might also be most relevant for policy makers,
e.g., in view of air quality standards.
Figure shows the evolution of the MODIS AOD averaged
over Saudi Arabia. The figure reveals that the AOD increase is limited to the
period between 2001 and 2012. The AOD levels in the year 2000 are high
compared to the subsequent years, and the levels in 2013 and especially in
2014 are low compared to the preceding years. Systematically varying length
and centre of the time interval used for the trend analysis (see Fig. S4 in
the Supplement) further supports that the most significant AOD increase is
observed for time intervals centred around the year 2007, which coincides
with the regime shift from an inactive to an active dust period around 2006
defined by . The same applies to the MODIS AOD
trends over Iraq and Iran (Figs. S6 and S8 in the Supplement), so that our trend analysis primarily considers the 12-year period
between 2001 and 2012. Note that in particular the end of the period is not
distinctly defined. For example, over Iraq a strong increase until 2008 is
observed, whereas from 2008 to 2012 the deseasonalised AOD remains
approximately constant on a high level (Fig. S5 in the Supplement). The right
panel of Fig. displays the Middle Eastern pattern of
the 12-year trends, which is very similar to the 10-year pattern shown
in Fig. S1.
Evolution of the average 550 nm AOD over Saudi Arabia from
March 2000 to February 2015. Monthly values from MODIS Terra Dark Target/Deep
Blue, collection 6, are shown. A seasonal cycle (pink) is obtained by
harmonic regression. Subtracting the seasonal cycle yields the deseasonalised
AOD (lower panel), the trend of which is quantified using linear regression
taking the 1-month lag autocorrelation of the time series into account
(green). The autocorrelation structure of the time series is studied in
Figs. S31 and S32 in the Supplement.
Evolution of the AOD measured by the AERONET station Solar Village.
The measurements have been linearly interpolated to 550 nm
wavelength. Deseasonalisation and trend analysis are performed as for
Fig. ; see also Figs. S33 and S34 in the Supplement.
For the 2001 to 2012 period, our trend analysis yields an average annual AOD
increase of 0.0135±0.0020 over Saudi Arabia (see
Fig. ). Consistent with , the winter AOD
levels are relatively invariant; exceptions are the very low levels during
the winters 2000/2001 and 2001/2002. This hints at dust being the aerosol
component mainly responsible for the perennial AOD variability.
There are numerous AERONET stations in the Middle East
, however only few have data records
extending over 10 years or more (Fig. S9 in the Supplement). Only one,
Solar Village, is located in a region of, according to the MODIS data,
strong and significant positive AOD trend.
From April 2000 to April 2013 the Solar Village data overlaps with MODIS
data. The AERONET AOD values, shown in Fig. ,
exhibit a similar trend as the MODIS AOD in Saudi Arabia
(Fig. ). The fluctuations of the two time series share
common features and both, the annual AOD increase and the corresponding
p value, are of comparable magnitude. Interestingly, the two closest
AERONET stations with long-term data records, the Israeli stations at Nes
Ziona and Sde Boker do not observe a significant AOD trend, see
Figs. S11–S14 in the Supplement.
Evolution of the average Saudi Arabian Ångström exponent
from March 2000 to February 2015. Monthly values from MODIS Terra Deep Blue,
collection 6, are shown. Deseasonalisation and trend analysis are performed
as for Fig. ; see also Figs. S35 and S36 in the
Supplement.
AERONET fine mode fraction at the station Solar Village.
Deseasonalisation and trend analysis is performed as for
Fig. ; see also Figs. S37 and S38 in the Supplement.
Correlation of precipitation (top row) and soil moisture (bottom
row) with the AOD over Saudi Arabia, Iraq and Iran. For soil moisture and
AOD, annual averages are used where the averaging period starts with the AOD
season on 1 December, for the precipitation the averaging period starts with
the precipitation season on 1 September to take effects on vegetation and
soils into account. Pixels with significant correlation
(p value < 0.01) are marked with a dot. Pixels with correlation
coefficients below -0.8 are outlined in yellow. MODIS Terra collection 6
combined Dark Target/Deep Blue 550 nm AOD data are used, precipitation
data are taken from TRMM 3B42, soil moisture from the European Space Agency
Climate Change Initiative (ESA-CCI).
The increase of AOD in Saudi Arabia during the last decade is accompanied by
a decrease of the Ångström exponent, which is inversely related to
the size of the particles. The MODIS Ångström exponent shown in
Fig. is noticeably anti-correlated to the AOD:
during the dusty high AOD periods in spring and summer, the Ångström
exponent reaches annual minima. Moreover, the maxima in the deseasonalised
AOD of Fig. (2001, 2004, 2008 to 2009, 2011 to 2012)
correspond to minima in the deseasonalised Ångström exponent. The
correlation coefficient of the deseasonalised monthly values between
March 2000 and February 2015 is -0.79. The decreasing Ångström
exponent clearly indicates a trend from 2001 to 2012 towards relatively large
aerosol particles, suggesting an increased amount of coarse mode dust
aerosols and generally a dominant role of dust regarding the AOD variability.
The same conclusion can be drawn from the AERONET fine mode fraction shown in
Fig. which exhibits a significant negative trend (in
contrast to the fine mode fractions at the Israeli stations, Figs. S16–S19
in the Supplement).
To further investigate the potentially dominant role of dust and possible
reasons for the AOD increase, in the following we study major parameters that
control the dust cycle: precipitation, surface wind and soil conditions, the
latter represented by the surface soil moisture. In this study, we rely
solely on correlation analysis. Deeper insight into the causality between the
observables as well as the role of atmospheric transport will be provided by
future modelling studies.
Precipitation
Precipitation can affect the aerosol and especially the dust load via several
mechanisms: precipitation scavenging can remove aerosol particles from the
atmosphere, increased soil moisture reduces wind-induced dust emissions and
additionally fosters vegetation growth, which further inhibits dust
emissions. Hence, a trend in the precipitation rate could help explain the
AOD increase in the Middle East.
We use the TRMM 3B42 precipitation data
to study the relation between precipitation and AOD. Annual average values of
the precipitation rate are used, where the averaging period starts with the
precipitation season of the region on 1 September, and the precipitation data
are re-gridded to a coarser 2∘ grid. The annual average AOD values are
calculated with an averaging period starting with the regional AOD season on
1 December. The time shift between the averaging periods used for
precipitation and AOD allows for effects of vegetation growth and soil
moistening during autumn. For Saudi Arabia, Iraq and Iran the correlation of
the individual precipitation pixels with the AOD, spatially averaged over the
country territories, is computed. The resulting three correlation maps are
shown in the top row of Fig. . For all three countries
the AOD is significantly anticorrelated to the precipitation in Iraq and
northern Saudi Arabia with absolute values of the correlation coefficients
regionally exceeding 0.8.
For the precipitation trend analysis, in addition to the TRMM 3B42 data, we
consider TRMM 3B31 data , the three
CMORPH variants RAW, ADJ, BLD (beta) and the
enhanced version of the CMAP data . We use the
same analysis as for the MODIS AOD (Sect. ). For each
data set, Fig. S22 in the Supplement shows the resulting trend pattern over
the Middle East. None individually shows a significant trend, neither
positive nor negative, extending over a larger region. Regionally, p values
below 0.01 are found, but the affected regions are not consistent for the
different data sets and it is evident that the precipitation does not exhibit
the same distinct trend as the AOD, cf. Fig. . This is
confirmed by Fig. S23 in the Supplement, where annual averaged instead of
deseasonalised precipitation values are considered which is expected to yield
more robust results in regions of sporadic precipitation. We conclude that
even though the AOD in the Middle East can be strongly linked to regional
precipitation, changes in precipitation alone are unlikely to be the reason
for the positive AOD trend.
Correlation of surface wind with AOD over Saudi Arabia, Iraq and
Iran. For both, wind and AOD, annual averages are used where the averaging
period starts with the enhanced AOD season on 1 December. Pixels with
significant correlation (p value < 0.01) are marked with a dot. Pixels
with correlation coefficients above 0.8 are outlined in yellow. MODIS Terra
collection 6 combined Dark Target/Deep Blue 550 nm AOD data are used,
for the surface wind we use the ERA-Interim wind at 10 m altitude.
Soil moisture
A long-term global surface soil moisture (SSM) data record based on satellite
mounted active and passive microwave sensors is provided by the European
Space Agency Climate Initiative (ESA-CCI) . We use the COMBINED data set which
covers a time period up to and including the year 2013.
We perform an AOD–soil-moisture correlation analysis analogously to the
AOD-precipitation analysis above. For both AOD and soil moisture, we use
annual average values where the averaging period starts with the AOD season
on 1 December. The soil-moisture data are re-gridded to the coarser 2∘
grid. For Saudi Arabia, Iraq and Iran the correlation of the individual
soil-moisture pixels with the AOD, spatially averaged over the national
territory, is computed. The resulting three correlation maps are shown in the
bottom row of Fig. . In all three cases we find strong
anti-correlations of the AOD to the soil moisture in Iraq and surrounding
areas, with absolute values of the correlation coefficients above 0.8.
Again we use the same analysis as for the MODIS AOD
(Sect. ) to analyse trends in the soil moisture data.
Figure S24 in the Supplement shows significant negative trends in large areas
of Syria, Iraq and Saudi Arabia. As the AOD is anti-correlated with soil
moisture, this could translate into the observed positive AOD trend. Since we
do not observe an accompanying trend with comparable significance in the
precipitation data, we assume that the soil moisture trend is predominantly
caused by increased evaporation due to increasing temperatures (Fig. S26 in
the Supplement) and a resulting decrease of the relative humidity (Fig. S27
in the Supplement), but possibly also due to changing land use, and that soil
moisture reflects the increasing drought conditions in the Fertile Crescent
during the last decade more distinctively than the precipitation rate.
Surface wind
Another important factor impacting dust emissions is the surface wind which
drives the saltation bombardment mechanism . To study
this we use the ERA-Interim data for wind speed at 10 m altitude
and perform the same analysis as for the soil
moisture. Figure shows the correlation of the AOD
over Saudi Arabia and Iraq with the wind speed over coastal regions of the
Red Sea, and the AOD over Iran with the wind speed over coastal regions of
the Persian Gulf. The regions with significant correlation are smaller than
the corresponding regions for soil moisture and precipitation. However, the
region around the Hala'ib Triangle on the African Red Sea coast, where the
wind speed is correlated to the AOD over Saudi Arabia and the Iraq, coincides
with a region of positive surface wind speed trend (see Fig. S28 in the
Supplement) so that it contributes to the observed AOD increase, e.g., over
the Red Sea. In contrast, the surface wind speed in the Fertile Crescent did
not increase and regionally even decreased. Therefore, in agreement with
, we conclude that it does not notably contribute to
increased dust activity.
Small-sample-size corrected Akaike information criterion (AICc) and
R2 values for linear AOD models using different predictors. For each
country, lower AICc values indicate preferable models. For Iran,
precipitation is a more relevant predictor than soil moisture,
suggesting greater relevance for atmospheric aerosol transport than in
Saudi Arabia and Iraq. Only for Saudi Arabia, the model can be improved
by including surface wind as predictor (last row). Corresponding
scatter plots are shown in Fig. S30 in the
Supplement.
Saudi Arabia Iraq Iran PredictorsAICcR2AICcR2AICcR2soil moisture-54.00.82-49.40.89-82.40.84precipitation-45.40.66-34.40.67-88.70.90soil moisture + precipitation-52.00.85-52.10.94-90.50.94soil moisture + precipitation + wind-55.20.93-47.60.94-86.00.94Multiple linear regression analysis
From the previous sections we conclude that the AOD is correlated to all
three variables considered, precipitation, soil moisture and surface wind.
Especially soil moisture exhibits a significant trend which could explain the
AOD increase, possibly supported by the regional positive surface wind trend.
For further analysis regarding the impact of each of these factors, we employ
multivariable linear models for the annual AOD.
The annual values averaged over the yellow marked pixels in
Figs. and , for which the
correlation indices are below -0.8 or above 0.8, respectively, are used as
predictors representing precipitation, soil moisture and surface wind; the
response variable is the annual AOD over Saudi Arabia, Iraq and Iran. The
same averaging periods as above are used (for precipitation starting on
1 September, for the other variables starting on 1 December). For each
country, the most general model we consider takes the form
τ=β0+βθθ+βPP+βww,
where τ is the AOD, θ the volumetric water content of the surface
soil, P the precipitation rate and w the surface wind speed. The
regression coefficients β0, βθ, βP and
βw are determined by the least squares method. A similar model has
been used by for seasonal dust prediction based on
precipitation rates and sea surface temperatures.
In Table we compare four variants of the model. The
corresponding scatter plots are shown in Fig. S30 in the Supplement. We
initially focus on the inhibiting factors precipitation and soil moisture and
omit the wind speed term βww in Eq. (). The resulting
model and the two possible submodels using only one predictor, either soil
moisture or precipitation, are represented by the first three rows of the
table. Overall, the very good agreement between the modelled values and the
observations, reflected by high R2 values, is striking. For all three
countries, the model using both predictors yields the highest R2 values.
For Saudi Arabia however, using the soil moisture alone yields a comparable
R2 value. The small-sample-size corrected Akaike information criterion
(AICc) value, which includes a penalty for additional estimated parameters,
suggests that in Saudi Arabia soil moisture is the dominant factor and even
that precipitation could be omitted from the linear model. Also for Iraq soil
moisture is a more relevant predictor than precipitation, but taking into
account that precipitation improves the model significantly using both
predictors is justified. For Iran we find the reverse relationship, where
precipitation is the more relevant predictor, but also here using both
variables yields the optimal model. The importance of precipitation for the
Iranian AOD indicates a more important role of aerosol transport and
precipitation scavenging along the way.
The relevance of each predictor for the different countries becomes even more
apparent in Fig. . It compares the AOD observed by MODIS
with the values resulting from the linear model using soil moisture and
precipitation as predictors (third row in Table ), breaking
down the variability contributions of the soil moisture term βθθ and the precipitation term βPP. The corresponding
regression coefficients are presented in Table . Again, for
all three countries, the agreement of model and observation is apparent,
suggesting that the interannual variability as well as the observed AOD
trends, in particular the increase from 2001 to 2009, can be attributed to
soil properties and precipitation that control the dust emissions and
removal. Moreover, the individual contributions of the soil moisture term
βθθ and the precipitation term βPP indicate
that the decreasing soil moisture from 2001 to 2009 is the driving force
behind the observed AOD increase during the same period, especially in Saudi
Arabia and Iraq. For Iran, the precipitation term contributes more to the AOD
variability than soil moisture which, as mentioned above, could be explained
by a more important role of aerosol transport and associated precipitation
scavenging.
Comparison of MODIS observations (black) and linearly modelled AOD
(grey). The brown (blue) bars represent the contribution of the soil moisture
(precipitation) term to the interannual variability. For Saudi Arabia and
Iraq, the soil moisture term is most strongly related to the AOD increase
during the last decade, whereas the precipitation term adds variability. For
Iran, the precipitation term is much more relevant, indicating that aerosol
transport and associated precipitation scavenging dominates the aerosol
variability.
Even though the model using soil moisture and precipitation as predictors
performs well for Saudi Arabia with an R2 value of 0.85, it performs less
well than for Iraq and Iran for which it yields R2 values of 0.94.
A considerable improvement can be achieved by taking into account the wind
speed as a third predictor. As displayed in the last row of
Table the R2 value is enhanced to 0.93 and the AICc value
is reduced. This corroborates the impact of dust emissions from the African
Red Sea coast on the Saudi Arabian AOD . It has
been reported that dust generated and channelled through the Tokar Gap on the
Red Sea coast of Sudan is frequently crossing the Red Sea and reaches Saudi
Arabia . The region identified by our method is
located slightly further north along the coast. The relevance of this region
for the AOD averaged over the whole Saudi Arabian territory might be related
to its location north of the highest peaks of the Asir Mountains which are
a natural barrier for atmospheric transport, extending along the
southwestern coast of Saudi Arabia. An analysis of wind patterns at
700 hPa altitude that corroborates the importance of westerly winds
for Saudi Arabia, in particular during winter, is shown in Fig. S29 in the
Supplement. A more detailed analysis of the atmospheric transport involved
and the response of the African dust emissions to the surface wind trends
remains to be conducted in a future modelling study using the atmospheric
chemistry-climate model EMAC as used by which
includes online dust emission schemes and the option to assimilate meteorological parameters, e.g.,
from the ERA-Interim re-analyses. Unlike for Saudi Arabia, for Iraq and Iran
including the surface wind as predictor does not improve the model.
Conclusions
Using the combined Dark Target/Deep Blue AOD products of the recently
available MODIS collection 6, our study corroborates the positive AOD trend
over the Middle East during the 10 year period from 2001 to 2010, which has
been reported previously based on other satellite products. Extending the
trend analysis to a time period of 15 years from 2000 to 2015 shows that
this trend persists until approximately the year 2012. Since then it is
interrupted by lower AOD values in subsequent years. Considering that the AOD
increase over Saudi Arabia is accompanied by a decreasing MODIS
Ångström exponent and a decreasing AERONET fine mode fraction at the
Solar Village station, coarse mode dust particles are likely the main
contributors to the positive AOD trend.
We have identified regions where precipitation, soil moisture and surface winds
are correlated with the AOD over Saudi Arabia, Iraq and Iran. Regarding
precipitation and soil moisture, the relevant regions are located in Iraq and
adjacent areas, regardless over which country the AOD is considered, identifying
this central region, largely coincident with the Fertile Crescent, as crucial
for the Middle Eastern AOD. Owing to transport by northwesterly Shamal winds its
dust activity affects the whole Arabian Peninsula .
Our multivariable regression analysis shows that the AOD trend and interannual
variability can be related to variations of soil moisture, precipitation and
surface winds in the particular regions. The positive AOD trend over Saudi
Arabia and Iraq relates to a negative soil moisture trend. The AOD
over Iran appears to be more strongly affected by precipitation, which points to
a more important role by aerosol transported into the country crossing areas
with significant precipitation, e.g., over the Zagros Mountains. For Saudi Arabia
the increasing surface wind near the African coast of the Red Sea
seems to be additionally involved.
Regression coefficients for the linear AOD model and their standard
errors.
We conclude that our results confirm a dust driven AOD trend and variability,
and that additional anthropogenic aerosol emissions are likely less relevant for
the AOD increase observed in the last decade. However, as the increased dust
emissions are likely related to drought conditions due to anomalously high
temperatures, a contribution by climate change is expected. It seems that dust
emissions sensitively respond to increasing temperature, which reduces the
relative humidity, thus enhancing evaporation and promoting soil drying. We
note that changing industrial emissions can alter aerosol properties which are
not considered in this study. For example, mixing with black carbon decreases
the single scattering albedo without notably affecting the AOD
and the interaction of dust and air pollution can
enhance hygroscopic growth , increasing the particle
size and thereby the AOD, similar to enhanced dust activity. These effects
remain to be analysed in future studies. We plan a modelling study to further
interpret the identified correlations in terms of causal relationships.
The Supplement related to this article is available online at doi:10.5194/acp-16-5063-2016-supplement.
Acknowledgements
The research reported in this publication has received funding from King
Abdullah University of Science and Technology (KAUST). For computer time, the
resources of the KAUST Supercomputing Laboratory were used by the KAUST
researchers. K. Klingmüller is supported by the KAUST CRG3 grant
URF/1/2180-01 Combined Radiative and Air Quality Effects of Anthropogenic Air
Pollution and Dust over the Arabian Peninsula.
The article processing charges for this open-access
publication were covered by the Max Planck Society. Edited by: M. Dameris
ReferencesAbdelkader, M., Metzger, S., Mamouri, R. E., Astitha, M., Barrie, L.,
Levin, Z., and Lelieveld, J.: Dust–air pollution dynamics over the eastern
Mediterranean, Atmos. Chem. Phys., 15, 9173–9189,
doi:10.5194/acp-15-9173-2015,
2015.AERONET: available at: http://aeronet.gsfc.nasa.gov, last access: 31 August 2015.Astitha, M., Lelieveld, J., Abdel Kader, M., Pozzer, A., and de Meij, A.:
Parameterization of dust emissions in the global atmospheric
chemistry-climate model EMAC: impact of nudging and soil properties, Atmos.
Chem. Phys., 12, 11057–11083,
doi:10.5194/acp-12-11057-2012,
2012.Balkanski, Y., Schulz, M., Claquin, T., Moulin, C., and Ginoux, P.: Global
emissions of mineral aerosol: formulation and validation using satellite
imagery, in: Emissions of Atmospheric Trace Compounds, edited by:
Granier, C., Artaxo, P., and Reeves, C., Vol. 18 of Adv. Glob. Change Res.,
Springer, the Netherlands, 239–267,
doi:10.1007/978-1-4020-2167-1_6,
2004.CMAP: available at: ftp://ftp.cdc.noaa.gov/Datasets/cmap/, last access:
31 August 2015.CMORPH V1.0: available at:
ftp://ftp.cpc.ncep.noaa.gov/precip/CMORPH_V1.0/, last access: 31 August
2015.
de Meij, A. and Lelieveld, J.: Evaluating aerosol optical properties observed
by ground-based and satellite remote sensing over the Mediterranean and the
Middle East in 2006, Atmos. Res., 99, 415–433, 2011.
de Meij, A., Pozzer, A., and Lelieveld, J.: Trend analysis in aerosol optical
depths and pollutant emission estimates between 2000 and 2009, Atmos.
Environ., 51, 75–85, 2012.Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P.,
Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P.,
Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N.,
Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L.,
Healy, S. B., Hersbach, H., Hólm, E. V., Isaksen, L., Kållberg, P.,
Köhler, M., Matricardi, M., McNally, A. P., Monge-Sanz, B. M.,
Morcrette, J.-J., Park, B.-K., Peubey, C., de Rosnay, P., Tavolato, C.,
Thépaut, J.-N., and Vitart, F.: The ERA-Interim reanalysis: configuration
and performance of the data assimilation system, Q. J. Roy. Meteor. Soc.,
137, 553–597, doi:10.1002/qj.828,
2011.ESA-CCI soil moisture: available at:
http://www.esa-soilmoisture-cci.org, last access: 31 August 2015.Giannadaki, D., Pozzer, A., and Lelieveld, J.: Modeled global effects of
airborne desert dust on air quality and premature mortality, Atmos. Chem.
Phys., 14, 957–968,
doi:10.5194/acp-14-957-2014,
2014.Giorgi, F. and Lionello, P.: Climate change projections for the Mediterranean
region, Global Planet. Change, 63, 90–104,
doi:10.1016/j.gloplacha.2007.09.005,
2008.
Hemming, D., Buontempo, C., Burke, E., Collins, M., and Kaye, N.: How
uncertain are climate model projections of water availability indicators
across the Middle East?, Philos. T. R. Soc. A, 368, 5117–5135, 2010.
Hoerling, M., Eischeid, J., Perlwitz, J., Quan, X., Zhang, T., and
Pegion, P.: On the increased frequency of Mediterranean drought, J. Climate,
25, 2146–2161, 2012.Holben, B., Eck, T., Slutsker, I., Tanré, D., Buis, J., Setzer, A.,
Vermote, E., Reagan, J., Kaufman, Y., Nakajima, T., Lavenu, F.,
Jankowiak, I., and Smirnov, A.: AERONET A Federated Instrument Network and
Data Archive for Aerosol Characterization, Remote Sens. Environ., 66, 1–16,
doi:10.1016/S0034-4257(98)00031-5,
1998.Hsu, N. C., Gautam, R., Sayer, A. M., Bettenhausen,
C., Li, C., Jeong, M. J., Tsay, S.-C., and Holben, B. N.: Global and regional
trends of aerosol optical depth over land and ocean using SeaWiFS
measurements from 1997 to 2010, Atmos. Chem. Phys., 12, 8037–8053,
doi:10.5194/acp-12-8037-2012,
2012.
Hubanks, P., Platnick, S., King, M., and Ridgway, B.: MODIS Atmosphere L3
Gridded Product Algorithm Theoretical Basis Document (ATBD) and Users Guide,
2015.Huffman, G. J., Bolvin, D. T., Nelkin, E. J., Wolff, D. B., Adler, R. F.,
Gu, G., Hong, Y., Bowman, K. P., and Stocker, E. F.: The TRMM Multisatellite
Precipitation Analysis (TMPA): quasi-global, multiyear, combined-sensor
precipitation estimates at fine scales, J. Hydrometeorol., 8, 38–55,
doi:10.1175/JHM560.1, 2007.IPCC: Climate Change 2013 – The Physical Science Basis, Cambridge University
Press,
doi:10.1017/CBO9781107415324,
Cambridge Books Online, 2014.Jiang, H., Farrar, J. T., Beardsley, R. C., Chen, R., and Chen, C.: Zonal
surface wind jets across the Red Sea due to mountain gap forcing along both
sides of the Red Sea, Geophys. Res. Lett., 36, L19605,
doi:10.1029/2009GL040008,
2009.Joyce, R. J., Janowiak, J. E., Arkin, P. A., and Xie, P.: CMORPH: a method
that produces global precipitation estimates from passive microwave and
infrared data at high spatial and temporal resolution, J. Hydrometeorol., 5,
487–503,
doi:10.1175/1525-7541(2004)005<0487:CAMTPG>2.0.CO;2,
2004.Kalenderski, S., Stenchikov, G., and Zhao, C.: Modeling a typical winter-time
dust event over the Arabian Peninsula and the Red Sea, Atmos. Chem. Phys.,
13, 1999–2014,
doi:10.5194/acp-13-1999-2013,
2013.Klingmüller, K., Steil, B., Brühl, C., Tost, H., and Lelieveld, J.:
Sensitivity of aerosol radiative effects to different mixing assumptions in
the AEROPT 1.0 submodel of the EMAC atmospheric-chemistry–climate model,
Geosci. Model Dev., 7, 2503–2516,
doi:10.5194/gmd-7-2503-2014,
2014.Lelieveld, J., Hoor, P.,
Jöckel, P., Pozzer, A., Hadjinicolaou, P., Cammas, J.-P., and Beirle, S.:
Severe ozone air pollution in the Persian Gulf region, Atmos. Chem. Phys., 9,
1393–1406,
doi:10.5194/acp-9-1393-2009,
2009.Lelieveld, J., Hadjinicolaou, P., Kostopoulou, E., Chenoweth, J.,
El Maayar, M., Giannakopoulos, C., Hannides, C., Lange, M. A., Tanarhte, M.,
Tyrlis, E., and Xoplaki, E.: Climate change and impacts in the Eastern
Mediterranean and the Middle East, Climatic Change, 114, 667–687,
doi:10.1007/s10584-012-0418-4,
2012.Lelieveld, J., Beirle, S., Hormann, C., Stenchikov, G., and Wagner, T.:
Abrupt recent trend changes in atmospheric nitrogen dioxide over the Middle
East, Science Advances, 1, e1500498,
doi:10.1126/sciadv.1500498,
2015.Levy, R. C., Remer, L. A., Kleidman, R. G., Mattoo, S., Ichoku, C., Kahn, R.,
and Eck, T. F.: Global evaluation of the Collection 5 MODIS dark-target
aerosol products over land, Atmos. Chem. Phys., 10, 10399–10420,
doi:10.5194/acp-10-10399-2010,
2010.Levy, R. C., Mattoo, S., Munchak, L. A., Remer, L. A., Sayer, A. M.,
Patadia, F., and Hsu, N. C.: The Collection 6 MODIS aerosol products over
land and ocean, Atmos. Meas. Tech., 6, 2989–3034,
doi:10.5194/amt-6-2989-2013,
2013.Liu, Y. Y., Parinussa, R. M., Dorigo, W. A., De Jeu, R. A. M., Wagner, W.,
van Dijk, A. I. J. M., McCabe, M. F., and Evans, J. P.: Developing an
improved soil moisture dataset by blending passive and active microwave
satellite-based retrievals, Hydrol. Earth Syst. Sci., 15, 425–436,
doi:10.5194/hess-15-425-2011,
2011.Liu, Y. Y., Dorigo, W. A., Parinussa, R. M., de Jeu, R. A. M., Wagner, W.,
McCabe, M. F., Evans, J. P., and van Dijk, A. I. J. M.: Trend-preserving
blending of passive and active microwave soil moisture retrievals, Remote
Sens. Environ., 123, 280–297,
doi:10.1016/j.rse.2012.03.014,
2012.Mishra, A. K., Klingmueller, K., Fredj, E., Lelieveld, J., Rudich, Y., and
Koren, I.: Radiative signature of absorbing aerosol over the eastern
Mediterranean basin, Atmos. Chem. Phys., 14, 7213–7231,
doi:10.5194/acp-14-7213-2014,
2014.MODIS MOD08 M3: available at:
ftp://ladsweb.nascom.nasa.gov/allData/6/MOD08_M3/, last access: 31 August 2015.Notaro, M., Alkolibi, F., Fadda, E., and Bakhrjy, F.: Trajectory analysis of
Saudi Arabian dust storms, J. Geophys. Res.-Atmos., 118, 6028–6043,
doi:10.1002/jgrd.50346, 2013.Notaro, M., Yu, Y., and Kalashnikova, O. V.: Regime shift in Arabian dust
activity, triggered by persistent fertile crescent drought, J. Geophys.
Res.-Atmos., 120, 10229–10249,
doi:10.1002/2015JD023855,
2015.Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., and R Core Team: nlme:
Linear and Nonlinear Mixed Effects Models, available at:
http://CRAN.R-project.org/package=nlme (R package version 3.1-122),
last access: 17 September 2015.Pozzer, A., Zimmermann, P., Doering, U.M., van Aardenne, J., Tost, H.,
Dentener, F., Janssens-Maenhout, G., and Lelieveld, J.: Effects of
business-as-usual anthropogenic emissions on air quality, Atmos. Chem. Phys.,
12, 6915–6937,
doi:10.5194/acp-12-6915-2012,
2012.Pozzer, A., de Meij, A., Yoon, J., Tost, H., Georgoulias, A. K., and
Astitha, M.: AOD trends during 2001–2010 from observations and model
simulations, Atmos. Chem. Phys., 15, 5521–5535,
doi:10.5194/acp-15-5521-2015,
2015.Sayer, A. M., Munchak, L. A., Hsu, N. C., Levy, R. C., Bettenhausen, C., and
Jeong, M.-J.: MODIS Collection 6 aerosol products: comparison between aqua's
e-deep blue, dark target, and “merged” data sets, and usage
recommendations, J. Geophys. Res.-Atmos., 119, 13965–13989,
doi:10.1002/2014JD022453,
2014.Shao, Y., Raupach, M. R., and Findlater, P. A.: Effect of saltation
bombardment on the entrainment of dust by wind, J. Geophys. Res., 98, 12719,
doi:10.1029/93JD00396, 1993.Stevens, B. and Schwartz, S. E.: Observing and modeling Earth's energy flows,
Surv. Geophys., 33, 779–816,
doi:10.1007/s10712-012-9184-0,
2012.Tegen, I.: Impact of vegetation and preferential source areas on global dust
aerosol: results from a model study, J. Geophys. Res., 107, 4576,
doi:10.1029/2001JD000963,
2002.TRMM 3B31: available at:
ftp://disc2.nascom.nasa.gov/ftp/data/s4pa/TRMM_L3/TRMM_3B31/, last
access: 31 August 2015.TRMM 3B42: available at:
ftp://disc2.nascom.nasa.gov/ftp/data/s4pa/TRMM_L3/TRMM_3B42_daily/,
last access: 31 August 2015.Wagner, W., Dorigo, W., de Jeu, R., Fernandez, D., Benveniste, J., Haas, E.,
and Ertl, M.: Fusion of active and passive microwave observations to create
an essential climate variable data record on soil moisture, ISPRS Annals of
Photogrammetry, Remote Sensing and Spatial Information Sciences, I7,
315–321,
doi:10.5194/isprsannals-I-7-315-2012,
2012.Wang, D., Morton, D., Masek, J., Wu, A., Nagol, J., Xiong, X., Levy, R.,
Vermote, E., and Wolfe, R.: Impact of sensor degradation on the MODIS NDVI
time series, Remote Sens. Environ., 119, 55–61,
doi:10.1016/j.rse.2011.12.001,
2012.Weatherhead, E. C., Reinsel, G. C., Tiao, G. C., Meng, X.-L., Choi, D.,
Cheang, W.-K., Keller, T., DeLuisi, J., Wuebbles, D. J., Kerr, J. B., Miller,
A. J., Oltmans, S. J., and Frederick, J. E.: Factors affecting the detection
of trends: Statistical considerations and applications to environmental data,
J. Geophys. Res.-Atmos., 103, 17149–17161,
10.1029/98JD00995, 1998.Wiggs, G. F. S., O'hara, S. L., Wegerdt, J., Van Der Meer, J., Small, I., and
Hubbard, R.: The dynamics and characteristics of aeolian dust in dryland
Central Asia: possible impacts on human exposure and respiratory health in
the Aral Sea basin, Geogr. J., 169, 142–157,
10.1111/1475-4959.04976, 2003.Xie, P. and Arkin, P. A.: Global precipitation: a 17-year monthly analysis
based on gauge observations, satellite estimates, and numerical model
outputs, B. Am. Meteorol. Soc., 78, 2539–2558,
doi:10.1175/1520-0477(1997)078<2539:GPAYMA>2.0.CO;2,
1997.Yu, Y., Notaro, M., Liu, Z., Wang, F., Alkolibi, F., Fadda, E., and
Bakhrjy, F.: Climatic controls on the interannual to decadal variability in
Saudi Arabian dust activity: toward the development of a seasonal dust
prediction model, J. Geophys. Res.-Atmos., 120, 1739–1758,
doi:10.1002/2014JD022611,
2015.Zanis, P., Hadjinicolaou, P., Pozzer, A., Tyrlis, E., Dafka, S.,
Mihalopoulos, N., and Lelieveld, J.: Summertime free-tropospheric ozone pool
over the eastern Mediterranean/Middle East, Atmos. Chem. Phys., 14, 115–132,
doi:10.5194/acp-14-115-2014,
2014.
Zhang, J. and Reid, J. S.: A decadal regional and global trend analysis of
the aerosol optical depth using a data-assimilation grade over-water MODIS
and Level 2 MISR aerosol products, Atmos. Chem. Phys., 10, 10949–10963,
doi:10.5194/acp-10-10949-2010,
2010.