Long-Term Variability of Dust Events in Southwestern Iran and Its Relationship with the Drought

: Dust storms represent a major environmental challenge in the Middle East. The southwest part of Iran is highly affected by dust events transported from neighboring desert regions, mostly from the Iraqi plains and Saudi Arabia, as well as from local dust storms. This study analyzes the spatio-temporal distribution of dust days at ﬁve meteorological stations located in southwestern Iran covering a period of 22 years (from 1997 to 2018). Dust codes (06, 07, 30 to 35) from meteorological observations are analyzed at each station, indicating that 84% of the dust events are not of local origin. The average number of dust days maximizes in June and July (188 and 193, respectively), while the dust activity weakens after August. The dust events exhibit large inter-annual variability, with statistically signiﬁcant increasing trends in all of ﬁve stations. Spatial distributions of the aerosol optical depth (AOD), dust loading, and surface dust concentrations from a moderate resolution imaging spectroradiometer (MODIS) and Modern-Era Retrospective analysis for Research and Applications (MERRA-2) retrievals reveal high dust accumulation over southwest Iran and surrounding regions. Furthermore, the spatial distribution of the (MODIS)-AOD trend (%) over southwest Iran indicates a large spatial heterogeneity during 2000–2018 with trends ranging mostly between − 9% and 9% (not statistically signiﬁcant). 2009 was the most active dust year, followed by 2011 and 2008, due to prolonged drought conditions in the fertile crescent and the enhanced dust emissions in the Iraqi plains during this period. In these years, the AOD was much higher than the 19-year average (2000 to 2018), while July 2009 was the dustiest month with about 25–30 dust days in each station. The years with highest dust activity were associated with less precipitation, negative anomalies of the vegetation health index (VHI) and normalized difference vegetation index (NDVI) over the Iraqi plains and southwest Iran, and favorable meteorological dynamics triggering stronger winds.


Introduction
Sand and dust storms (SDSs) are meteorological phenomena that are triggered due to turbulent wind enabling to remove loose sand and dust from exposed dry top soil in arid and semi-arid areas, causing movement of sand dunes or airborne dust [1][2][3][4][5]. SDSs have major and multiple economic, health, and environmental impacts causing respiratory infections and diseases [6][7][8][9][10]. The SDSs paths include trajectories of Aeolian dust particles transported to deposition areas from a primary source that is often far away, travelling over oceans and continents [11][12][13][14]. Removal of fertile top sediments may of dust days, seasonality, long-term variation, as well as the factors associated with shortand long-term trends in dust activity. Although previous studies have analyzed the long-term variability of dust events over SW Iran [52,[87][88][89], the current work provides a comprehensive analysis combining ground measurements at five stations with satellite and re-analysis observations and regional meteorology, and also differentiates the local and regional dust events.

Study Area and Dust Storms
The examined study area in SW part of Iran includes four Iranian provinces, namely Khuzestan, Bushehr, Ilam, and Kermanshah [87]. The five most dust-affected cities in these provinces were selected for the analysis, namely Ahvaz and Abadan in the Khuzestan province, Bushehr in the coastal plain of the Persian Gulf (located in the middle of the Bushehr province), and Ilam and Kermanshah located in the homonymous provinces ( Figure 1; Table 1). Ahvaz (48 • 68 E, 31 • 32 N; 18 m asl) is one of the largest cities in Iran with a population of about 1,112,000 habitats and an area of~530 km 2 [77]. It is one of the dustiest large cities in the world and the World Health Organization (WHO) ranked it as the world's most air-polluted city in 2011 in terms of PM 10 concentrations [90,91]. Maximum and minimum PM 10 concentrations in Ahvaz were detected in July (>400 µg m −3 ) and January (~150 µg m −3 ), respectively [74]. The weather in Ahvaz is very hot and humid, with temperatures up to 50 • C and relative humidity (RH) close to saturation point during summer. The city suffers from frequent and intense dust storms all year round with higher frequency in spring and early summer [52,92]. Abadan exhibits mostly similar meteorological conditions with Ahvaz and high dust loading throughout the year.  Kermanshah city is located near to Zagros mountainous range (34.3 • N, 47.5 • E and 1350 m asl) with population of about 1 million. Dust storms mostly from the Iraqi deserts affect the city and sometimes reduce visibility to less than 1 km [93,94]. The annual average PM 10 concentrations in Kermanshah is 229 µg m −3 and the city is among the most polluted in the Middle East [95]. In recent years, some dust sources were also activated near the city. Dust storms originated from the deserts in Iraq and Syria also affect Ilam city, located west of the Zagros Mountains [96,97]. Karimi et al. [95] also analyzed dust storms in Ilam and the performance of six numerical models for dust simulation during a severe dust storm over the area.
The city of Bushehr is located in the northern shore of the Persian Gulf in southwest part of Iran (29 • N and 51 • E). The annual rainfall in the city is low (about 268 mm) and its climate is warm and semi-desert. The prevailing wind direction is from south to north, thus facilitating dust plumes coming from the Arabian Peninsula. Keshavarzi et al. [98] examined the levels, source apportionment, and health risks of polycyclic aromatic hydrocarbons (PAHs) from street dust samples in Bushehr, indicating large effects from the deserts and considerable health risk.

Data Set and Methodology
In this study, the annual frequency of dust events is analyzed at five polluted cities in the southwestern part of Iran during the period 1997-2018. Table 1 includes the longitude, latitude, elevation, and the total number of dust days at the meteorological stations in each city. For the determination of the dust events/days, three-hour recordings (8 times a day) of the synoptic codes (06, 07, 30 to 35) related to dust were used at the meteorological stations. For the consideration of a dust day, at least one daily observation should include a dust code.
The synoptic code 06 shows widespread dust in suspension, which is not raised by wind at or near the station, indicating non-local dust. Code 07 indicates dust or sand raised by wind near the station, thus mostly corresponding to local dust events [99], while the codes 30-35 were very rare in all stations. Codes 30-32 indicate slight or moderate SDS, while intense or severe SDS correspond to codes 33-35 [89]. The analysis of the frequency and seasonality of the dust events was initially performed separately for the synoptic codes that represent different intensity of the dust events.
In addition, weekly vegetation health index (VHI) data from 2000 to 2016 were taken from the Center for Satellite Applications and Research (STAR; 4 km resolution) in the Ninawa (Nineveh) province of Iraq and the Dier-ez-Zur province of Syria (Figure 1), which are known for frequent dust emissions that strongly impact the west part of Iran [100,101].
VHI is an indicator for estimating the vegetation health and is provided from the advanced very high-resolution radiometer (AVHRR) NOAA sensor. This index is based on the properties of vegetation in absorbing and reflecting sunlight. In non-drought years, green plants reflect little visible light due to the absorption of sunlight by chlorophyll, and most of their radiation is due to the diffusion of infrared light by the inner fibers of the leaves and the water inside them. There is a big difference in reflectance between the visible and infrared spectrum, which indicates that the vegetation is green and abundant. In drought years, due to the decrease in chlorophyll and water content of plants, the reflection of visible light increases and the infrared light decreases [79]. Therefore, VHI is based on reflection and absorption of solar visible and infrared radiations and is calculated using the following formula: where: where a = 0.5 (combinations of vegetation condition index (VCI) and temperature condition index (TCI) as possible predictors of crop yield). BT is brightness temperature. All three indices are scaled to range from 0 (severe vegetation stress) to 100 (exceptionally favorable conditions). Furthermore, Terra Moderate Resolution Imaging Spectroradiometer (MODIS) for monthly normalized difference vegetation index (NDVI) with a spatial resolution of 0.05 degrees was analyzed over the study region from 2000 to 2018. NDVI is given by the formula: and corresponds to spectral reflectance measurements in Red (0.620-0.670 µm) and NIR (0.841-0.876 µm) wavelength bands. NDVI varies between −1 and 1. The higher vegetation corresponds to positive and closer to 1 NDVI values, since plants and vegetated areas reflect the visible light less than the infrared one.  [102,103]. Furthermore, spatial distribution and time series of several atmospheric parameters, such as AOD 550 , dust loading, dust surface concentration, and dry and wet dust deposition, were taken over the study region from the Modern-Era Retrospective analysis for Research and Applications (MERRA-2) re-analysis database, 5.12.4 model, at a horizontal resolution of 0.5 • × 0.625 • (latitude, longitude) [104,105]. MERRA-2 has been proved as an accurate database for monitoring of dust aerosols, spatial variability, and trends over the Middle East, exhibiting considerable agreement with MODIS observations [35,86]. Moreover, ERA-5 reanalysis [106], with a 31-km horizontal resolution and 137 levels spanning from the surface of the Earth to 0.01 hPa, was used to obtain meteorological fields of surface vector winds at 0.75 • × 0.75 • spatial resolution.

Evolution of Dust Days in SW Iran
This section analyzes the evolution of dust days at the five examined stations in SW Iran from 1997 to 2018. The dust events over the region are attributed to local emissions from the arid terrains, but mostly to transported dust storms of various intensity from the Mesopotamia plains and deserts in Iraq and Syria. Previous studies have shown that the dust particle size and composition play a major role in source apportionment of the dust storms, emission, and deposition rates [107][108][109][110][111]. The deposited dust in Ahvaz was found to be composed of~80% of carbonates, followed by quartz (~13%) [112], while a recent work justified this mineralogical composition [113]. The carbonates move faster and uplifted at higher heights than other minerals, such as quartz and feldspars, while rich carbonate soils prevail in the Khuzestan province and southern Iraqi plains [92], indicating dust of local/regional origin.
Analysis of the synoptic codes (06, 07, and 30 to 35) at the five stations revealed that dust episodes in Ahvaz observed at 10.6% of the total observations from 1997 to 2018, out of which 89% corresponds to code 6 and 10% to code 7, while codes 30-35 were very rare (1%). Dust in Bushehr occurred at 11% of the cases, while 94.8% of the meteorological observations in Kermanshah did not show dust presence; in Ilam, this fraction was 94.7% and in Abadan it was 90.7%. Furthermore, code 6 prevailed with fractions of 82% in Abadan, 73% in Bushehr, and 99% in Ilam and Kermanshah, indicating dominance of widespread dust in suspension from non-local sources, characterized as regional or longrange transported [89]. On average, 84% of dust events in all stations were non-local (code 06) and about 16% can be characterized as local dust (code 07). It should be noted that the mountainous terrains and forest cover in Ilam and Kermanshah provinces do not facilitate local dust sources and they are mostly affected by transported dust plumes originated from central-south Iraq and Syria [114].
On the other hand, local dust storms of higher intensity are more frequent in Ahvaz (10%), Abadan (18%), and Bushehr (26%) stations, which are located in plain and open areas directly impacted by the arid/desert surrounding terrains [52,84,115]. Furthermore, the desiccation of Hour Al-Azim wetland and Karkheh river in west Khuzestan contributes significantly to the local dust events in Ahvaz and Abadan [115][116][117]. After extensive field visits at 180 source points in Khuzestan plain, Heydarian et al. [118] reported that about 9% of the total area, equivalent to 349,254 hectares, is a potential source of dust storms. Dust sources have different characteristics in this area, including degraded pastures, abandoned rain-fed agricultural lands, uncovered lands, wetlands, and dried-up ponds and deserts [119]. Bushehr station is highly affected by transported dust plumes from Iraq and Saudi Arabia after crossing the Persian Gulf, while local dust events are also important originated from the arid/desert terrains in coastal southwest Iran, since~25% of the Bushehr province can be considered as an active dust-prone area [3]. A previous study reported two climatic dust peaks in Bushehr, i.e., between 1982 and 1990 and between 2005 and 2008 [120]. In agreement with our results, Arami et al. [121] reported that, in 25 stations in west and southwest Iran, the fractions of non-local (06 code) and local dust (code 07) were 74.9% and 25%, respectively, during a period of 20 years.
The annual evolution of the frequency of dust days at the five stations from 1997 to 2018 is shown in Figure 2. A remarkable annual variability is observed, while the stations present a general co-variability, indicating rather common years of low and high dust activity. Ahvaz and Abadan stations present the highest number of dust days, which increased dramatically after 2007 and till about 2012. This is the period of a drought shift in the fertile crescent in Iraq (Mesopotamia plains), which lasted for about 3-4 years [53,100] and highly increased the dust activity over Iraq, southwest Iran, Kuwait and northeast Saudi Arabia [32,69,100,122]. Increasing trends in dust days in almost all stations are also observed after 2014. Trend analysis in the number of annual dust days revealed statistically significant trends in all five stations at 95% confidence level, with Z values from the Mann-Kendall test above the critical of 1.645 (ranging from 1.90 in Ahvaz to 3.78 in Ilam). Ilam presents the highest increasing rate in the number of dust days (4.59 per year), while the increasing rates for the other stations range between 1.35 in Bushehr and 2.99 in Abadan ( Figure 2). This indicates an important increase in dust activity over southwest Iran, which was also found in the neighboring regions of the Middle East, such as Iraqi plains, Syrian desert, Persian Gulf, and nearly whole Saudi Arabia based on MODIS observations from 2000 to 2015 [32]. The peak in frequency of dust days occurred in 2009 at almost all the stations (Figure 2), similarly to previous works over the region [82,[123][124][125]. Figure 3a shows the annual variation of the common dust days in all stations, thus revealing dust plumes covering the whole region. The total number of common dust days was 810 during the period 1997-2018, which exhibit higher frequency in 2009 (104 days), following by 2008 with 89 days and 2011 with 83 days. It is characteristic that during 1999-2002, the dust days were less over SW Iran (except of 2000), while at the same time, east and southeast Iran faced the largest frequency of dust events due to prolonged drought conditions, driven by large-scale dynamics and the prolonged La-Nina phase [34,126,127]. This indicates that local, regional, and large-scale meteorological dynamics that modulate dust activity over the west-southwest and east parts of Iran are totally different, as also shown by recent works that examined the atmospheric circulation patterns and local winds over these areas [64,128].   [34,126,127]. This indicates that local, regional, and large-scale meteorological dynamics that modulate dust activity over the west-southwest and east parts of Iran are totally different, as also shown by recent works that examined the atmospheric circulation patterns and local winds over these areas [64,128].
The ratios of the common dust days per the number of dust days in each station show large differences between the stations (Figure 3b). In Bushehr, the ratio of 59% means that the city faces an important fraction of dust events (41%), which do not affect all other four stations. Furthermore, Kermanshah and Ilam exhibit large fractions of 93.8% and 85.4%, whereas in the Khuzestan stations the ratios are much lower, i.e., Abadan (62.8%) and Ahvaz (45.8%), indicating that a large fraction of the dust days is of local origin affecting only these cities.
The mean monthly variability of the dust days shows that the highest frequency occurred in July with 193 dust days, followed by June (188) and May with 179 dust days (Figure 3c). On the contrary, the average frequency of dust days is below 30 from November to January, which are mostly associated with frontal systems, i.e., pre-frontal and postfrontal dust storms [66,69]. Figure 3d shows the monthly variation of the dust days in each of the five stations throughout the 22-year period. In all stations, the highest frequency of dust days occurred in June and July (or May), indicating a rather similar annual pattern. Ahvaz is the most dust-affected site, especially in May, June, and July, while Ilam and Kermanshah exhibit very low number of dust days during the cold period (Figure 3d). Furthermore, Nabavi et al. [129] reported larger number of dust days in June and July in Khuzestan plain during a 16-year (2000-2015) period, while Zarasvandi et al. [92] also found an increasing number and duration of dust events in June and July over the same area. About 61% of the Khuzestan plain contains lowland and desert areas, and, apart from some mountainous areas in the northeast of the province, the wind erosion treats other areas [130,131]. Zagros   The ratios of the common dust days per the number of dust days in each station show large differences between the stations (Figure 3b). In Bushehr, the ratio of 59% means that the city faces an important fraction of dust events (41%), which do not affect all other four stations. Furthermore, Kermanshah and Ilam exhibit large fractions of 93.8% and 85.4%, whereas in the Khuzestan stations the ratios are much lower, i.e., Abadan (62.8%) and Ahvaz (45.8%), indicating that a large fraction of the dust days is of local origin affecting only these cities.
The mean monthly variability of the dust days shows that the highest frequency occurred in July with 193 dust days, followed by June (188) and May with 179 dust days (Figure 3c). On the contrary, the average frequency of dust days is below 30 from November to January, which are mostly associated with frontal systems, i.e., pre-frontal and post-frontal dust storms [66,69]. Figure 3d shows the monthly variation of the dust days in each of the five stations throughout the 22-year period. In all stations, the highest frequency of dust days occurred in June and July (or May), indicating a rather similar annual pattern. Ahvaz is the most dust-affected site, especially in May, June, and July, while Ilam and Kermanshah exhibit very low number of dust days during the cold period ( Figure 3d). Furthermore, Nabavi et al. [129] reported larger number of dust days in June and July in Khuzestan plain during a 16-year (2000-2015) period, while Zarasvandi et al. [92] also found an increasing number and duration of dust events in June and July over the same area. About 61% of the Khuzestan plain contains lowland and desert areas, and, apart from some mountainous areas in the northeast of the province, the wind erosion treats other areas [130,131]. Zagros Mountains also help in the accumulation and persistence of dust over the Khuzestan plain and prevents the higher-altitude stations of Ilam and Kermanshah from persistent dust plumes, thus contributing to the lower number of dust days [132].
As shown above, 2009 was the most dust-laden year in southwest Iran during the last decades, and more specifically, July 2009 was the dustiest month. Table 2  The results show that July 2009 was extremely dusty, while at most stations the dust days ranged from 25 to 30. Furthermore, the dust days in this month present a dramatic increase from the average number of dust days in July, which varied from 7 to 13 at most of the stations (Table 2), with serious respiratory, cardiovascular problems, and hospitalization of dozens of people [133]. Therefore, using of native plants and green belts in the upwind sides of urban centers as effective methods in trapping dust particles and reduce the movement of the sand dunes is highly recommended for such extreme dust cases [134,135].   Figure 4 shows the mean Terra-MODIS AOD 550 spatial distribution (resolution: 10 km) over southwest Iran and surrounding regions from 2000 to 2018 and the AOD 550 pattern over the same area during 2008-2009. In long-term basis, higher AODs are observed in vast areas of south Iraq, Kuwait, northeastern parts of Saudi Arabia, and southwest Iran, as previous studies also showed [32,136,137]. The AOD 550 was significantly higher in 2008-2009 indicating positive anomalies over nearly the whole study region. This was attributed to increasing dust emissions in Iraq (fertile crescent) after 2007 due to a shift to a drought period, influenced by La Nina and Pacific Decadal Oscillation teleconnection patterns that modulated sea surface temperature (SST); atmospheric circulation patterns; and rainfall over the northern Arabian Sea, the Arabian Peninsula, and the Middle East [53,100].  Figure 5b shows the spatial distribution of the MODIS-AOD trend (%) using high-resolution (10 km) data over the southwest part of Iran. The results indicate a large spatial heterogeneity in the AOD trends during 2000-2018 ranging mostly from −9% to 9%.  Table 1 from 2000 to 2016 shows that, in most years, VHI in Nineveh was higher than that in Dayr Az Zawr ( Figure 6). In general, VHI in the Dayr Az Zawr province was lower than 40, indicating a significant lack of vegetation in this area. Furthermore, a significant decrease was observed in VHI values in both regions from 2006 to 2009, followed by a slight recover afterwards, while this decrease was more intense in the Nineveh province. These provinces consist of arid/desert areas that highly affect the dust activity in the Middle East and southwest Iran. Therefore, the decrease in VHI after 2006 signaled the increase in dust activity over southwest Iran [52,84,100]. In the following, time series of columnar AOD, dust loading, surface dust concentration, NDVI, and wet and dry dust deposition, obtained from MODIS observations and MERRA-2 reanalysis, are analyzed over a specific domain including parts of southwest Iran and southeast Iraq (Figure 7a), in order to examine the variability and trends in dust activity.

Satellite Observations
The annual variation of surface dust concentration and dust load obtained from MERRA-2 reanalysis is shown in Figure 7b [64,80,140,141], the annual variability is mostly driven by the dust patterns in spring and summer, when dust activity highly maximizes. Karimi et al. [142] also showed that 2008 and 2009 were the dustiest years from 2005 to 2015m and July was the dustiest month during that time frame [143]. Dust concentration and load presented statistically significant (at 95% confidence level) increasing trends during 1997-2018, thus affecting the increasing trend in frequency of dust days over SW Iran.
The Terra-MODIS and MERRA-2 AOD annual variations (Figure 7c) exhibit considerable agreement with dust load (R 2 = 0.83, 0.94) and surface dust concentrations in a lower degree (R 2 = 0.63, 0.78). Both datasets present highest annual AODs in 2008 (MERRA-2) and 2009 (MODIS), while the strong correlation (R 2 = 0.81) indicated a great consistency between MODIS and MERRA-2 AODs over the Middle East and Iran [86,[144][145][146][147][148]. However, MERRA-2 generally exhibited lower annual AOD values, while the underestimation compared to MODIS increases in the high dust-laden years (e.g., 2008-2012), as has been documented from previous studies [35,137,148]. Both MODIS and MERRA-2 AODs exhibited an increasing trend during the study period, but without statistical significance at the 95% confidence level. Finally, the time series of dry and wet depositions, along with precipitation, obtained from MERRA-2, are shown in Figure 7e for the period 1997-2018. The highest amount of dry deposition occurred in 2008, while during this drought year, the amount of wet deposition and rainfall were significantly lower compared to the other years [100,150,151]. This indicates a higher possibility of dust to remain in the atmosphere for several days and/or weeks and to be transported at long-distances downwind [63,64,72], thus increasing the dust AODs [32,120]. This was also verified by the current analysis, indicating that apart from soil dryness and reduced vegetation cover, deficit of rainfall and lower wet deposition rates were responsible for enhanced accumulation of dust over the Middle East after 2007 and till about 2012. Figure 8 shows the spatial distributions of the dust concentration, dust load, and dry and wet dust deposition over SW Iran from MERRA-2 retrievals during the period 1997-2018. All patterns show a remarkable spatial heterogeneity of the dust properties even over a small geographic area, highlighting the difficulty in assessing common dust characteristics between the stations and the challenges and biases in simulating the dust plumes via numerical dust models [89,138,152]. The maximum surface dust concentrations and columnar dust load are observed in the southwestern part of the Khuzestan province and the northwest Persian Gulf, while high values of dust load also extended to the Bushehr province, in the northern shores of the Persian Gulf. The highest amount of dry deposition was over the Persian Gulf and the western half of Khuzestan plain in the borders with Iraq and Kuwait [153], thus contributing largely to the surface dust concentrations due to significant dust fallout rates. However, the spatial pattern of the wet deposition is completely different and the maximum values were observed over the northeast of Khuzestan plain and along the Zagros mountain range, where the rainfall is generally higher. Rainfall favors wet dust deposition and highly controlled its temporal variation over the study area (Figure 7e), while Broomandi et al. [123] found a strong negative correlation between annual rainfall and the number of dust storms in every year in the south and west parts of Iran.  Previous works and current analysis revealed that, over the last two decades, the Middle East faced an increasing trend in dust emissions and concentrations, especially over the Tigris-Euphrates basin, southwest Iran, and the eastern part of the Arabian Peninsula. Due to shift to a prolonged dry period after 2007, dust storms over the region have significantly increased, thus enhancing the dust-aerosol loading and concentrations. Current analysis highlighted July 2009 as the dustiest month over southwest Iran with several intense dust-storms like those on 5, 16, and 27 July 2009 [89]. Apart from the extensive drought over the fertile crescent after 2007 [100], the meteorological conditions prevailed in July 2009 facilitated enhanced dust emissions, due to an increase in pressure gradient between the Caspian Sea and low-pressure areas in Iraq and Saudi Arabia that triggered stronger Shamal winds over the Mesopotamia fluvial arid terrain. The synoptic meteorology prevailed during July 2009 was analyzed in a previous work [89], while here we present the spatial AOD distribution (from L3 MODIS observations) along with surface winds (from ERA-5 reanalysis; [64]) in July 2009 with respect to mean July (2000-2018) conditions ( Figure 9). The MODIS observations justify the large increase in AOD (mostly dust AOD) during July 2009 along the fertile crescent; Syrian desert; Iraqi plains; and in downwind regions, such as eastern Saudi Arabia. The southwest part of Iran also exhibited higher AODs, as the high-resolution (10 km) MODIS observations show ( Figure  9d). Furthermore, over these regions, an enhancement of the west/north-westerlies took place in July 2009 (Figure 9c), which, along with the drought conditions, favored dust emissions and an increase in dust days.

Conclusions
This study examined the long-term (for about two decades) variation and trends of dust days and dust characteristics (aerosol optical depth, surface concentration, dust load) over the southwest part of Iran, based on synergy of meteorological data at five cities in the region and satellite observations. In this respect, the dust presence was identified from meteorological observations of the dust synoptic codes (06, 07, and 30 to 35) at five stations (Ahvaz, Abadan, Bushehr, Ilam, and Kermanshah) highly impacted by local, regional, and long-range transported dust events throughout the year, but with higher frequency and intensity in spring and summer.
On average, 92% of the cases during the period 1997-2018 were dust-free, 7% were related to code 06 (raised or transported dust), and about 1% related to code 07 (local dust). In addition, 84% of the dust events were non-local and 16% of local origin, while these fractions presented notable differences between the stations. The analysis detected 810 dust days that affected concurrently all five stations, which exhibited highest frequency in 2009 (104 days), in 2008 (89 days), and in 2011 (83 days). On a monthly basis, the highest frequency was detected in July, June, and May months. Ahvaz and Abadan stations presented the highest number of dust days due to their location in the main downwind region (Khuzestan plain) affected by Iraqi dust events, whereas Ilam and Kermanshah exhibited the lowest number of dust events, as they are located at elevated areas in the Zagros Mountains. Trend analysis indicated a statistically significant (95% confidence level) increase in dust days in all stations during the period 1997-2018, which was driven by the large increase after the prolonged drought shift in 2007, which enhanced the dust activity over the Middle East. After 2012, the frequency of the dust days decreased but not at the levels during the beginning of the 2000s.
High-resolution Terra-MODIS observations over southwest Iran also justified the maximum annual-mean AODs in 2008-2009, in agreement with the peak in dust days. The drought conditions after 2007 and till about 2011/12 were also justified by the lowest NDVI values in southwest Iran and the lowest vegetation health index (VHI) values in two provinces that are prone to dust emissions in Iraq (the Nineveh province) and Syria (the Dayr Az Zawr province). This indicates that the regional shift to a drought period in the Middle East after 2007 had a dramatic effect on vegetation growth, land susceptibility to wind erosion, and an increase in dust activity over the region. However, long-term (2000-2018) trend analysis of level 2 MODIS AODs over the examined stations did not reveal a statistically significant trend in any of them, although the general increasing AODs.
July 2009 was found to be the dustiest month over the last few decades in southwest Iran. The extended drought in the fertile crescent in Iraq, along with favorable meteorological conditions of increased pressure gradients and winds and absence of rainfall, led to more than 25 dust days in this month in many stations in west and south Iran. MODIS observations also highlighted an abnormal increase in dust AODs over the region in July 2009, signaling high deterioration of air quality.