Changes of Permanent Lake Surfaces, and Their Consequences for Dust Aerosols and Air Quality: The Hamoun Lakes of the Sistan Area, Iran

respectively). The correlation between dusty days and percentage of dried Hamoun beds (100-percent of water surface) shows high correlation coefficient values Hamoun Saberi and Baringak (r=0.88 and r=0.82 respectively) and lower correlation for Hamoun Hirmand (r= 0.63). The high correlation for the Hamoun Saberi and Baringak indicates that Sistan dust storms are directly affected by the north and


Introduction
Changes in the frequency and extent of natural inundation occurring on large permanent and ephemeral lake systems may lead to significant fluctuations in regional dust loading on both a seasonal and an inter-annual basis [1]. As surface water diversion increases, arid-land surfaces that were previously wet or stabilized by vegetation are increasingly susceptible to deflation by wind, resulting in desertification and increase in dust outbreaks [2][3][4]. Desiccation of lake beds, whether due to drought or to water diversion schemes, as in the Aral Sea in Turkmenistan [5,6], Owens lake in California [7,8], lake Eyre in Australia, Hamoun lakes in Iran [9][10][11][12], can lead to increased dust storm activity. Thus, some dust may be derived from dried lake beds and can be highly saline, while the finest aerosols can be injurious to health. Anthropogenic sources were previously considered as important dust contributors [13], but more recent estimates of only 5-7% of total mineral dust from such sources gives major importance to natural sources [14]. Each year, several billion tons of soil-dust are entrained into the atmosphere playing a vital role in solar irradiance attenuation, and affects marine environments, atmospheric dynamics and weather [15][16][17][18][19][20].
Atmospheric aerosols affect the global climatic system in many ways, i.e. by attenuating the solar radiation reaching the ground, modifying the solar spectrum, re-distributing the earthatmosphere energy budget and influencing cloud microphysics and the hydrological cycle. Mineral dust plays an important role in the optical, physical and chemical processes in the atmosphere, while dust deposition adds exogenous mineral and organic material to terrestrial surfaces, having a significant impact on the Earth's ecosystems and biogeochemical cycles. The impact of dust aerosols in the Earth's system depends mainly on particle characteristics such as size, shape and mineralogy [21], which are initially determined by the terrestrial sources from which the soil sediments are entrained and from their chemical composition [22]. Size distribution is a key parameter to characterize the aerosol chemical, physical, optical properties and their effects on health. The lower and upper size limits of dust aerosols are from a few nanometers to ~100μm and aerosol properties change substantially over this size range. Several studies demonstrated that airborne Particulate Matter (PM) has an impact on climate [23], biogeochemical cycling in ecosystems [24], visibility [25] and human health [26][27][28]. Over recent years in the public health domain the PM concentration has become a topic of considerable importance, since epidemiological studies have shown that exposure to particulates with aerodynamic diameters of < 10 μm (PM10) and especially < 2.5 μm (PM2.5) induces an increase of lung cancer, morbidity and cardiopulmonary mortality [e.g. [29][30][31][32][33][34][35]. Thus air pollution appears to have an adverse effect on respiratory and cardiovascular systems [36], which might result in an acute reduction of lung function, aggravation of asthma, increased risk of pneumonia in the elderly, low birth weight and high death rates in newborns [37].
Some dust contaminants (soluble and chelatable metallic salts, pesticides, etc) affect human health when they are transported over densely populated areas [38], retained in residences and other occupied structures [39], and they also impact the nutrient loading of waters flowing from adjacent watersheds [40] and terminal bodies of water by direct and indirect deposition [41][42]. PM is a complex mixture of substances suspended in to the atmosphere in solid or liquid state with different properties (e.g. variable size distribution or chemical composition amongst others) and origins (anthropogenic and natural). Owing to this mixture of substances, the chemical composition of PM may vary widely as a function of emission sources and the subsequent chemical reactions which take place in the atmosphere [43][44][45].The chemical mass balance is the most commonly used method for assessing PM source contributions [46], while statistical methods, such as factor analysis and multi-linear regression [47], have also produced interesting results regarding dust source identification. Elemental and mineralogical analyses have also been used to identify the source regions of dust deposited in Arctic ice caps [48] and on other depositional surfaces [49,42].
One spectacular example of such dust effects is the Hamoun (dry) lakes in the eastern part of Iran that has attracted scientific interest during recent years, since it constitutes a major dust source region in southwest Asia, often producing intense dust storms that cover the Sistan region in eastern of Iran and the southwest of Afghanistan and Pakistan and influencing the air quality, human health and ecosystems as well as aerosol loading and climate from local to regional scales [50][51][52]12]. Particles from the Hamoun dust storms might also cover farm and grasslands to result in damage to crops and fill the rivers and water channels with aeolian material. Over recent years, tens of thousands of people have suffered from respiratory diseases and asthma during months of devastating dust storms in the Sistan basin, especially in the cities of Zabol and Zahak and the surrounding villages [53]. According to the Asthma Mortality Map of Iran, the rate of asthma in Sistan is, in general, higher than in other regions [54].
The aim of this Chapter is to: (1) assess the seasonal and annual variability of dust storms that originate from the Hamoun Basin, (2) identify the amount of dust loading from the dry lakes, (3) assess the contribution of the lake beds to regional dust emissions for specific locations, (4) assess the dust chemical and mineralogical composition to provide useful In 1949 the United States initiated a new program for the improvement of underdeveloped areas of the world. The damming of the Helmand river in southern Afghanistan became one of the showcase projects of U.S. foreign aid in the "Third World" after World War II. Dams were built on the Helmand river and its main tributary (Arghandab river) during the 1950s. The main project goals were to provide hydroelectric power and increase agricultural productivity through irrigation and land reclamation. The Arghandab dam, located northwest of the city of Kandahar, was completed in 1952 with a height of 145 feet (44.2 meters) and storage capacity of 388,000 acre-feet (478.6 million cubic meters). The larger Kajakai dam on the Helmand was completed a year later with a height of 300 feet (91.4 meters) and length of 919 feet (280 meters) and storage capacity of 1,495,000 acre-feet (1,844 million cubic meters). About 300 miles (482.8 kilometers) of concrete-lined canals were built to distribute the reservoir waters [9]. The Hirmand river is the longest river in Afghanistan (ca. 1150 km; catchment > 160,000 km 2 ) and the main watershed for the Sistan basin, finally draining into the natural swamp of the Hamoun lakes complex [10].
Severe dust plumes usually extend from Sistan into southern Afghanistan and southwestern Pakistan obscuring the surface over much of the region (Fig 2). Severe droughts during the past decades, especially after 1999, have caused desiccation of the Hamoun lakes leaving a fine layer of sediment that is easily lifted by the wind [55], thus modifying the basin to one of the most active sources of dust in southwest Asia [56,50,57]. Therefore, the Hamoun dry-lake beds exhibit large similarities with the other two major dust source regions of the world that comprise of dried lakes and topographic lows, i.e. Bodele depression in Chad [3] and lake Eyre in Australia [4]. The strong winds blow fine sand off the exposed Hamoun lake beds and deposit it to form huge dunes that may cover a hundred or more villages along the former lakeshore. As a consequence, the wildlife around the lake has been negatively impacted and fisheries have been brought to a halt. The intense dust plumes form a giant U-shape extending from Sistan into southern Afghanistan and southwestern Pakistan that obscures the surface over much of the region. The pale color of the dust plume is consistent with that of dried wetland soils. The dust is blowing off the dry lake beds that become the Hamoun wetlands during wet years.

Hamoun lakes
The Hamoun lakes are situated roughly at the termination of the Hirmand river's inland delta. The Hamoun lakes complex (Hamoun-e-Puzak, Hamoun-e-Sabori and Hamoun-e-Hirmand and Baringak) are located in the north of the Sistan region, which is also the largest fresh water ecosystem of the Iranian Plateau and one of the first wetlands in the Ramsar Convention [58]. Water in the Hamoun lakes is rarely more than 3 meters deep, while the size of the lakes varies both seasonally and from year to year. Maximum expansion takes place in late spring, following snowmelt and spring precipitation in the mountains. In years of exceptionally high runoff, the Hamoun lakes overflow their low divides and create one large lake that is approximately 160 kilometers long and 8-25 kilometers wide with nearly 4,500 square kilometres surface area. Overflow from this lake is carried southward into the normally dry Gaud-i Zirreh (Fig. 1), the lowest playa (463-meter altitude) in the Sistan depression. Furthermore, mountain runoff varies considerably from year to year; in fact, the Hamoun lakes have completely dried up at least three times in the 20th century [9]. The maximum extent of the Hamoun lakes following large floods is shown in Fig. 4, where a continuous large lake has been created covering an extended area of ~4,500 km 2 with a volume of 13000 million m 3 in Sistan and southwestern Afghanistan. This figure corresponds to spring of 1998 after snowmelt in the Afghanistan mountains that transferred large quantities of water into the Hamoun Basin. As a consequence, livelihoods in the Sistan region are strongly interlinked with and dependent on the wetland products and services, as well as on agricultural activities in the Sistan plain. Fishing and hunting represent an important source of income for many households and, therefore, the local and regional economy is strongly dependent on weather conditions, precipitation and land use -land cover changes. The political boundary between Iran and Afghanistan splits the Hamoun system, further complicating management possibilities in the area. Most (90%) of the watershed is located in Afghanistan and practically all of the wetlands' water sources originate there. The Iranian part is desert, and produces runoff only in rare cases of significant local rainfall [10].
In view of the Hirmand and the surrounding rivers that supply most of the sediments to the Hamoun lakes, a brief encapsulation of the relevant geology of the catchment area in Afghanistan is given. Afghanistan has a very complex geology, encompassing two major relatively young orogenies, Triassic and subsequent Himalayan, resulting in amalgamation of crustal blocks and formation of concomitant ophiolites and younger clastic and carbonate sedimentary rocks as well as basaltic lavas and, more recently, extensive alluvial and eolian detritus [59] . The Sistan region and Hamoun dry lake beds are mainly composed of Quaternary lacustrine silt and clay material as well as Holocene fluvial sand, silt and clay (Fig. 3). These materials have been carried to the basin by the rivers, while along their courses Neogene fluvial sand, eolian sand, silt and clay are the main constituents. Note also the difference in the soil-dust composition between two major desert areas, Registan and Dasht-i-Margo in Afghanistan. The former is composed of Neogene coarse gravels and the latter of Quaternary eolian sand. More details about the geology in the Sistan region can be found in [60].

Droughts in the Sistan
The Sistan Basin has recently experienced an unusually long 10-year drought starting in 2000 [10]. Combined with war and severe political disruption over the past 2 decades, the 10-year drought has created conditions of widespread famine that affected many people in eastern Iran and southwestern Afghanistan. A suggested, climatic forcing mechanism has been proposed for the recent drought by Barlow and others (2002). A prolonged ENSO (El Niño-Southern oscillation) cold phase (known as La Niña) from 1998 to 2001 and unusually warm ocean waters in the western Pacific appear to have contributed to the prolonged drought. The unusually warm waters (warm pool) resulted in positive precipitation anomalies in the Indian Ocean and negative anomalies over central Afghanistan [64], thus contributing to the drying of the Hamoun Basin. The contrast between a relatively wet year in 1976 and the nearly dry Hamoun lakes in 2001 is shown in Figs. 2 and 4. Millions of fish and untold numbers of wildlife and cattle died. Agricultural fields and approximately 100 villages were abandoned, and many succumbed to blowing sand and moving dunes [65].
Most of the Sistan population lives near the Hamoun lakes and is employed in agricultural, fishery, handicrafts and other jobs. To counter the effects of droughts, the Iranian government prepares facilities such as food and flour supplies, medicine and health services and employment in the region to prevent the forced emigration of people, but the continuous and extreme droughts have forced some people to leave the Sistan region. Long droughts at the end of the 1960s, middle of the 1980s, and from 1999 to 2010 affected the Sistan region significantly and resulted in desiccation of the Hamoun lakes, making the surrounding lands saline and disturbing their soil fertility, while some places became barren (see  to black indicates deeper waters, which, however, do not exceed four meters. (2) By 2001, the Hamoun lakes had vanished since central and southwest Asia were hit by the largest persistent drought anywhere in the world. The only sign of water in this scorched landscape of extensive salt flats (white) is the Chah Nimeh reservoir in the southern part of Sistan (not shown on the satellite image), which is now only used for drinking water. Degraded reed stands in muddy soil are visible as dark green hues at the southern end of Hamoun-i Puzak. In 2003 the Hamoun Basin was covered with water again, but with significantly lower coverage than in the mid-1970s [65].

Data set and experimental methods
In order to address the scientific topics mentioned above multiple ground-based instrumentation and several sets of satellite images (mostly Landsat and MODIS) were used for illustration purposes and for detection of water level in Hamoun and for monitoring of land use -land cover changes, seasonality of dust storms and associated sediments, air quality perspectives, chemical and mineralogical composition of dust over the Sistan region and Hamoun Basin. Within this framework, satellite images from different years were used to identify changes in the lake's surface. Information about dust storm occurrence was obtained from Zabol meteorological station, 5 km from the Hamoun lakes (see Fig. 4). The ground-based measurements were primarily used to compare effects of the Hamoun surface on dust aerosols.
More specifically, the amount of dust loading during dust storms was measured using passive dust traps fixed at two monitoring towers (respectively, at four and eight meters above ground level in altitude), with one meter distance between the adjacent individual traps; the 4 m tower had four traps and the 8 m tower had 8 traps (Fig. 6). The two towers were established in open sites (station A and station B, denoted by red stars in Fig. 2 right panel) close to the dry-bed lake dust source region during the period August 2009 to July 2010. The dust sampler used in the campaign was developed by the Agricultural and Natural Research Center of Sistan (Fig. 6), and is a modified version of the SSDS sampler [66][67]. At the observation sites, the samplers collect airborne dust sediment. The traps were mounted on a stable bracket parallel to the wind direction. The samplers consist of a tube with a diameter of 12 cm. The sediment-laden air passes through a vertical 2.5 cm x 6 cm sampler opening in the middle. Inside the sampler, air speed is reduced and the particles settle in a collection pan at the bottom, while the air discharges through an outlet with a U shape. After each measurement, the samplers were evacuated to make them ready for measuring the following dust events. The collected samples were oven dried at 105 o C for 24 hours, and then, dried samples were weighed using an electronic scale in order to obtain total mass quantities at each sampling height and for each dust storm. The samples were also transported to a laboratory for chemical and mineralogical analysis.
Furthermore, soil samples were collected from topsoil (0-5 cm depth) at several locations in the dry-bed Hamoun lakes and downwind areas. These samples were analyzed to investigate the chemical and mineralogical characteristics of dust, relevance of inferred sources and contributions to air pollution. These samples were analyzed for major and trace elements and for minerals by applying X-Ray Fluorescence (XRF) and X-Ray Diffraction (XRD) techniques, respectively. The samples were prepared for XRD analysis using a back loading preparation method. They were analyzed using a PANalytical X'Pert Pro powder diffractometer with X'Celerator detector and variable divergence and receiving slits with Fe filtered Co-Kα radiation. The phases were identified using X'Pert High score plus software. The relative phase amounts (weights %) were estimated using the Rietveld method (Autoquan Program). Mineral analysis by XRD is the single most important non-destructive technique for the characterization of minerals such as quartz, feldspars, calcite, dolomite, clay, silt and iron oxides in fine dust. Mineral phase analysis by XRD is one of the few techniques that are phase sensitive, rather than chemically sensitive, as is the case with XRF spectrometry. Quantitative mineralogical analyses using the XRD technique have been performed by a number of scientists over the globe [e.g., 68-70, 44, 71-72].
The sample preparation for XRF is made up of two methods, pressed powders and fusions. The former samples were prepared for trace element analyses and the latter for major

Dust trap Dust trap
Dust Air Flow element analyses. Each milled sample (<75μm) was combined with a polyvinyl alcohol, transferred into an aluminum cup and manually pressed to ten tons. The pressed powders were dried at 100°C for at least 30 minutes and stored in a desiccator before analyses were conducted. For the fusion method, each milled sample (<75μm) was weighed out in a 1/6 sample to flux (Lithium tetraborate) ratio. These samples were then transferred into mouldable Pt/Au crucibles and fused at 1050°C in a muffle furnace. Aluminum cooling caps were treated with an iodine-ethanol mixture (releasing agent) and placed on top of the crucibles as they cooled. Some samples needed to be treated with an extra 3g of flux if they continued to crack. Finally, all geochemical samples were analyzed using the Thermo Fisher ARL 9400 XP+ Sequential XRF. The Quantas software package was used for the major element analyses and the WinXRF software package was used for the trace element analyses. The concentrations of the major elements are reported as oxides in weight percentages, while the trace element concentrations are reported as elements in parts per million (ppm).
Furthermore, in order to provide analysis of the air quality, PM10 concentration measurements were obtained by using an automatic Met One BAM 1020 beta gauge monitor (Met One, Inc.,) over Zabol. The instrument measures PM10 concentrations (in μg.m -3 ) with a temporal resolution of one hour. The measurements were carried out at the Environmental Institute of Sistan located at the outskirts of Zabol during the period September 2010 to September 2011 (total of 373 days). The recording station is close to the Hamoun basin and is placed in the main pathway of the dust storms of the Sistan region. The hourly measured PM10 data were daily-averaged, from which the monthly values and seasonal variations were obtained. For further assessing the air quality over Zabol, the PM10 concentrations were used to calculate an Air Quality Index (AQI) [52].

Meteorology and climatology over Sistan
The climate over Sistan is arid, with low annual average precipitation of ~55 mm occurring mainly in the winter (December to February) and evaporation exceeding ~4000 mmyear -1 [58]. During summer, the area is under the influence of a low pressure system attributed to the Indian thermal low that extends further to the west as a consequence of the south Asian monsoon system. These low pressure conditions are the trigger for the development of the Levar northerly wind, commonly known as the "120-day wind" [73], causing frequent dust and sand storms, especially during summer (June to August) [74,56] and contributing to the deterioration of air quality [52]. Therefore, one of the main factors affecting the weather conditions over the region is the strong winds rendering Sistan as one of the windiest deserts in the world. These winds blow continuously in spring and summer (from May to September), and on some days during winter, and have significant impacts on the landscape and the lives of the local inhabitants.
The annual variation of mean Temperature (T), Relative Humidity (RH) and atmospheric Pressure (P) over Zabol (a large city in the Sistan region) during the period 1963 to 2010 is shown in Fig 7. The monthly mean T exhibits a clear annual pattern with low values in the winter (9 to 12 o C) and high (~35 o C) in summer, following the common pattern found in the northern Mid-latitudes. During the summer period the maximum T often goes up to 46 or 48 o C causing an extremely large diurnal variation, which is a characteristic of many arid environments. RH illustrates an inverse annual variation with larger values in winter (50 to 57%) and very low values in summer (~25%), which are about 10 to 15% during daytime. During the period October to April P values are generally high (1020 to 1024 hPa in winter), which is above the standard mean sea level value of 1013.25 hPa. P values decrease during summer (~ 996 hPa in July) as a result of the Indian thermal low that develops over the entire south Asia during summer monsoon months. This has a direct impact on the intensity of the winds over the region, which has a monthly mean of as high as 12 m.s -1 during June and July with frequent gusts of above 20 to 25 m.s -1 . In contrast, during late autumn and winter months the wind speed is confined to ~3 to 4 m.s -1 [12].

Temporal changes of Hamoun dry lake beds and dust storms
The dust storms over a region cause several climatic implications, environmental and human concerns [16,75,18,76,36] and can be examined via multiple instrumentation and techniques. Among others, the analysis of the visibility records can constitute a powerful tool for monitoring of the seasonal and inter-annual variation of the dust storms, since the main result of such phenomena is the limitation of visibility and deterioration of air quality. The annual variation of visibility (as the main indicator for the dust storms) over Sistan follows a clear annual pattern, with large values in winter, usually above 10 km, and very low in summer (< 4 km on average) as analyzed from meteorological observations taken at Zabol (Fig. 8). A power-decreasing curve relation associated with 93% of the variance was observed between wind speed and visibility [12]. This inverse relation indicates that the wind speed does not act as a ventilation phenomenon over Zabol, as usually occurs in coastal urban environments with local sea-breeze cells [77], but rather as a factor responsible for the deterioration of visibility, since the intense Levar winds are the cause of the dust outbreaks over Sistan. Therefore, the major dust storms over the region are associated with intense winds of northwesterly direction that are responsible for the deterioration of visibility to lower than 100 m in many cases. Although the visibility exhibits a clear annual pattern (Fig. 8) suggesting that the summer season is the favourable period for the occurrence of frequent and intense dust storms, longterm data series over Zabol  show that it contains considerable year-to-year variations (Fig. 9b). Focusing on recent years, the days with visibility <= 2 km have been dramatically increased from about 20 during 1995 to 1999 to >100 during 2000 to 2001. This is attributed to a severe drought period that dried the largest part of the Hamoun wetlands (see Fig. 5) and favored the alluvial uplift, as well as the frequency and mass intensity of dust storms that affected the visibility over Sistan (Fig. 9b). However, in the 2000s the days with very low visibility seem to have a decreasing trend, but remaining above the standards of the climatological mean. It is, therefore, concluded that the regional and synoptic meteorology (mainly precipitation) is strongly linked to land use -land cover changes over Hamoun and then, to dust outbreaks over Sistan region.
In contrast, the annual variation of the wind speed (Fig. 9a) exhibits an opposite pattern with higher intensities during summer (June and July) and lower in winter. As far as the wind direction is concerned, it is found from the Zabol data series that the northwestern direction clearly dominates, being more apparent in summer, while high percentages for intense winds are also associated with a northwesterly flow (Fig. 9a). The probability for intense winds to blow from other directions is low; summer winds are much more intense with ~27% of wind speeds above 11 ms -1 , while calm conditions are limited to 3% against 19% and 20% for autumn and winter, respectively. The higher frequency and intensity of northwestern winds is the reason for the frequent dust storms that affect Zabol. It is to be noted that Zabol is located at the downwind direction of dust storms that normally originate from Hamoun (Fig. 2). The water levels in the Hamoun lakes change considerably from year to year as has been discussed above.

Dust loading measurements
Dust activity is a function of several parameters, such as topography, rainfall, soil moisture, surface winds, regional meteorology, boundary layer height and convective activity [78-79. Data on dust loading are available at only a few places around the world [e.g. [80][81][82][83][84] and those presented here are the first for the Sistan region. Hence, obtaining measurements of horizontal dust flux will significantly increase our understanding of wind erosion and dust influences. Apart from the natural emissions of dust, [85] identified two ways in which human activities can influence dust emissions: (a) by changes in land use, which alter the potential for dust emission, and (b) by perturbing local climate that, in turn, alter dust emissions. As has been discussed above, both ways are considerably active over the Sistan region and Hamoun Basin.
The dust loading measured at the two stations close to the Hamoun basin for several dust events during the period August 2009 to July 2010 is plotted in Fig. 11. In the same graph, meteorological data from the Zabol station that give information about the duration of dust events (for the examined days as well as on the preceding or succeeding days, i.e. about 2-3 days before the peak-day of the dust storm) and daily mean and maximum wind speeds, are also plotted. The results of the average dust loading measured at eight heights at station B and at four heights at station A reveal considerable variation, ranging from ~0.10 to ~2.5 kgm -2 . In general, the highest dust loading is observed for dust events occurring in summer, but intense dust storms can also take place in winter, since the Hamoun basin is an active dust source region throughout the year. The dust loading is highly correlated with the duration of the dust storms, as shown from their correlation, with the linear regressions being statistically significant at the 0.99% confidence level (Fig. 12). Apart from the strong linkage to the duration of dust storms, the dust loading at both stations also seems to have a dependence on the daily mean and maximum wind speeds (not presented). However, this dependence was found to be more intense and statistically significant (at the 95% confidence level) at station B, which is located closer to the dust source, whereas for station A the correlation was not found to be statistically significant. This finding emphasizes the strong effect of the wind speed on dust erosion and transportation, as well as on dust loading, at least for areas close to dust sources. However, the results show that the main factor that controls the dust loading at both stations is the duration of the dust storms, and secondly the wind speed. The role of the wind might have been found to be more critical if measurements were taken at the sampler stations instead of using the meteorological data from Zabol. The analysis showed that the total dust loading for the 19 events of measurements at station A is 16.9 kg m -2 corresponding to 0.88 kg m -2 per event, whereas at station B the measurements yielded 15.8 kg m -2 (17 events), corresponding to 0.93 kg m -2 per event. The larger dust loading at station B is attributed to the smaller distance from the source region. Figure 11. Average dust loading (kgm -2 ) during various dust events in the Sistan region as measured at the 4m (station A) and 8m (station B) monitoring towers. The duration of dust events (hours), as well as the mean and maximum wind speeds on the dusty days were obtained from the Zabol meteorological station. Figure 13 illustrates the height variation in dust loading during the dust storms measured at station A (19 days, up to 4m in height) and station B (17 days, up to 8m in height). Contrasting height variations measured during intense dust storms occurred between the two stations, while similar variations correspond to moderate and low dust storm events. More specifically, the dust loading shows an increase (decrease) with height in station A (station B), revealing a difference in the dust transport mechanisms. This finding can be explained by considering the fact that station B is located closer to the Hamoun dust source region, meaning that uplift and newly transported dust concentration is higher near the surface. On the other hand, at station A that is located about 20 km away, the dust loading presents larger values up to 3 m since the near-ground dust particles have already been deposited near the source, and as the distance increases so does the dust-plume height. The diurnal variability of the dust loading at the two stations (not presented) showed increased mass concentrations during daytime that can be explained by enhanced convection and turbulent mixing in a deepened boundary layer. Furthermore, the local winds are stronger during daytime due to thermal convections.

PM10 measurements
In order to provide a first ever in-situ analysis of the air quality over Sistan, PM10 concentration measurements were obtained by using an automatic Met One BAM 1020 beta gauge monitor (Met One, Inc.,) at Zabol [12]. The instrument measures PM10 concentrations (in μg.m -3 ) with a temporal resolution of one hour. The measurements were carried out at the Environmental Institute in Sistan located at the outskirts of Zabol during the period September 2010 to August 2011. The recording station is close to the Hamoun basin and is placed in the main pathway of the dust storms in the Sistan region. The hourly measured PM10 data were daily-averaged, from which the monthly values and seasonal variations were obtained (  The results show extremely large PM10 concentrations at Zabol (see Fig. 14). Even the mean values are much higher than the most risky and dangerous maximum levels provided by the U.S. Environmental Protection Agency (397 g.m -3 ). Throughout the year, and especially during the period June to October, the area suffers from severe pollution since even the lower PM10 values are above 100 μg.m -3 , while the maximum ones are usually above 1000 μg.m -3 . On the other hand, extreme PM10 measurements associated with severe dust events may also occur in other months, for example like December. Daily PM10 concentrations during major dust storms are about 10 to 20 times above the standard levels. Regarding the monthly mean PM10 concentrations, the results show extremely large values (>500 μg.m -3 ) during the period June to October, reaching up to 847 μg.m -3 in July.
Changes of Permanent Lake Surfaces, and Their Consequences for Dust Aerosols and Air Quality: The Hamoun Lakes of the Sistan Area, Iran 181 The frequency of occurrence of PM10 concentrations for each season over Zabol is depicted in Fig. 15. In summer ~60% of the PM10 values were higher than 425 μg.m -3 , while the lower PM10 values occur in winter and spring with larger frequency in the 55-154 μg.m -3 interval. A very significant finding is the very low frequency for PM10 concentrations below ~400 μg.m -3 in summer, suggesting an extremely turbid atmosphere with frequent dust storms and near absence of clear or relatively clear conditions over Sistan during summer. Autumn also presents high frequency in the >425 μg.m -3 interval that might be due to continuation of the Levar winds in September favouring the dust storms over Sistan.

Air quality index (AQI)
In order to identify the impact of air pollution on human health, air pollution indices are commonly used, of which the AQI is the most well known [86][87][88]. As a consequence, the AQI is a powerful prenatainarry tool to ensure public health protection [86].
The AQI is divided into six categories, varying from 0 to 500, with different health impacts [86] as listed in Based on the technological rules related to AQI, the following formula was used to derive the PM10 concentration from AQI [90,88]: Where I is the concentration of PM10, Ilow and Ihigh are AQI grading limited values that are lower and larger than I (AQI index), respectively, and Chigh and Clow denote the PM10 concentrations corresponding to Ihigh and Ilow, respectively.
Provisional studies focusing on air quality and dust over Iran have already been carried out. For example, amongst others, [91] performed a comparative study of air quality in Tehran during the period 1997 to 1998. The results revealed that in 1997 the air quality on 32% of the days was unhealthy, and on 5% of the days it could be regarded as very unhealthy, whereas in 1998 the unhealthy and very unhealthy days increased to 34% and 6%, respectively. [92] studied the air quality in Tehran and Isfahan and offered solutions for its improvement using the AQI. It was found that on 329 days of the year in Tehran, and on 34 of the days in Isfahan, the AQI departed beyond 100. [93] also studied AQI in Tehran reporting that on 273 days in 2001 the values were higher than those set for the air quality standards; 13% of the days were considered as very unhealthy and 0.27% were classified as dangerous. [52] found that 15 % of the days were unhealthy for sensitive people in the city of Zahedan thatwas affected by Sistan dust storms, while 2 % were associated with a high health risk or were even hazardous.
Comparing the present results with those of the above-mentioned studies, it is concluded that the Sistan region experiences much higher PM concentration levels. Assessment of air quality in Zabol shows that 243 days out of 370 (65%) exhibit air pollution levels of above the air quality standards (>155 μg.m -3 ), a fraction that is much higher than that (26.5%) reported for Zahedan city located also in Sistan about 200 km south of Zabol [52]. The most significant finding is the 129 days (34.9%) that are characterized as hazardous (Table 4), which in combination with the adverse effects on human health, make it clear that environmental conditions in the Sistan region are rather poor for human well-being. On the other hand, only 5.7% of the days are associated with low pollution levels when the air quality is considered satisfactory and air pollution poses little or no risk. Several studies have shown that ambient air pollution is highly correlated with respiratory morbidity, mainly amongst children [94,36,95]. The results gathered from hospitals in the Sistan region showed that during dust storms respiratory patients increased significantly, especially those affected by chronic obstructive pulmonary disease and asthma. The percentage of these diseases increases in summer (June and July) [53]. Apart from the dust storms, re-suspended dust within the urban environment is a strong source of PM10 concentrations, while urban-anthropogenic and industrial activities are considered to have a much lower effect on the air pollution over Zabol.
The mean diurnal variation of PM10 concentrations for each season in Zabol indicates a clear pattern for all seasons except winter, with the maximum of the diurnal variation being observed in the middle of the day (~08:00-11 LST) while in winter PM10 values reach a maximum in the afternoon hours to early morning (~16:00 -02:00 LST). In general, solar heating and vertical mixing of pollutants are the main factors for the reduction of PM10 levels at local noon to early afternoon hours. However, the maximum PM10 concentrations normally occur between 08:00 and 11:00 (LST) over Sistan. The diurnal PM10 variability in all seasons, except winter, is closely associated with the intensity of the wind speed measured at the Zabol meteorological station (see Fig. 16). This wind, being northerly in direction, carries large quantities of dust from the Hamoun dry lake bed. The mean diurnal wind speed variation is similar for all seasons; however, the wind favors the increase of aerosol load in summer and autumn (maximum PM10 for higher wind speeds). Note that the Hirmand River and some other ephemeral channels provide little water in winter and spring to the Hamoun lake beds. Therefore, in early summer, the Hamoun lakes are wet and at the end of summer and early autumn are always completely dried out. On the other hand, the Levar winds continue also in September and so high wind speeds cause huge dust storms.

Mineralogical characteristics of dust
In order to understand the influence of dust on the atmospheric environment, climatic system and health and to establish effective remedial policies and strategies, it is regarded as necessary to investigate the compositional (chemical and mineralogy) characteristics of airborne and soil dust over Sistan. To the best of our knowledge there are currently no published studies about the geochemical characteristics and dust mineralogy in this region. Moreover, nearby locations, Bagram and Khowst in Afghanistan, were selected for analyzing the mineralogical dust composition, major and trace elements within the framework of the Enhanced Particulate Matter Surveillance Program (EPMSP) campaign [44]. Furthermore, mineralogical and geochemical characteristics of dust were recently examined at Khuzestan province in southwestern Iran [72].
In this Chapter, an overview of the geological-geochemical characteristics of airborne and soil dust in the Sistan region is given for airborne and soil samples collected during the period August 2009 to August 2010. The chemical constituents during major dust storms over the region are analyzed at two locations (Fig. 2), also investigating the relationship between the chemical constituents of the dust storms and those of the inferred (Hamoun) source soils. The mineralogy percentage composition averaged at all heights for each day is shown in Fig. 17a, b for stations A and B, respectively. The chemical formulas of the main mineralogical components are given in [96], as well as the chemical reactions of dust with atmospheric constituents and trace gases during the dust life cycle. The mineralogical composition corresponds to screened samples with diameter <75 μm and can constitute an indication of both regional geology and wind transported dust that is deposited in local soils [44].
Emphasizing the dust mineralogy at station A, it is seen that the airborne dust is mainly composed of quartz, which is the dominant component (26-40%) for all the days of observations. Calcareous particles, mainly consisting of calcite, are the second dominant mineralogical component over the site with average mass percentage of 22%, while micas (muscovite) contribute 13% and plagioclase (albite), 11%. The remaining components contribute much less to the dust mass, while chlorite (6.3%) is apparent in all dust samples for all days. The others, i.e. dolomite, enstatite, gypsum, halite, etc are present only in some samples with various percentages. It is quite interesting to note that quartz is much more common over Sistan than the feldspars (plagioclase, microcline and orthoclase).  The mineralogical analysis for the 9-days recorded data at station B (Fig. 16b) shows more or less similar results to those obtained for station A and, therefore, any discussion will be given on their comparison (Fig. 19). The mineralogical composition has the same descending order as in station A, i.e. quartz (39.8±4.4%), calcite (18.8±3.5%), plagioclase (albite) (12.7±1.4%) and muscovite (10.1±3.2%). On the other hand, dust deposition may influence biogeochemical cycling in terrestrial ecosystems, while dust accumulation in soils can influence texture, element composition and acid neutralizing capacity [97][98]. Furthermore, the chemical and mineralogical composition of soil dust provides useful information about its provenance [99], radiative forcing implications [100] and human health effects [101]. For these reasons, in addition to the airborne dust samples, soil samples were collected at 16 locations around Sistan and Hamoun, at depths ranging from 0 to 5 cm from the soil crust.
The results of soil sample mineralogy are summarized in Fig. 18. From an initial consideration of these results, it is established that the soil samples exhibit similar mineralogy to the airborne dust at both stations, thus suggesting similarity in sources for both airborne and soil dust. On the other hand, some soils in the Sistan region have been primarily formed from dust transported from the Hamoun lakes, presenting large similarities in mineralogy and chemical composition to airborne dust. However, atmospheric chemical reactions involving dust and aerosols of other types can alter the chemical characteristics of dust before its deposition [102]. Therefore, the mineralogy of the soil samples may differ significantly in comparison to the results obtained for airborne dust at stations A and B, since some of the soil samples (11 samples) were collected in the Hamoun dried lakes and others (five samples) around stations A and B.  Figure 19 summarizes the results from the mineralogical analysis of samples taken at the two stations and from the soil samples, allowing a quantitative comparison between them. The vertical bars correspond to one standard deviation from the mean for both airborne and soil samples. The distance from the source region from whence dust is deposited also influences the particle size distribution, mineralogy and chemical composition of dust. Therefore, generally speaking, at local scales quartz clearly dominates with fractions up to ~50%, while as the distance from the dust source increases, feldspars (plagioclase, microcline) and phyllosilicate minerals (illite and kaolinite) present increased fractions [103,42]. However, in our study the dust samples were all obtained within the same area and, therefore, are mineralogically similar. Nevertheless, station B, which is located closer to the Hamoun basin, the source of dust exposures, exhibits higher percentages of quartz, while station A (near to Zabol city) exhibits higher concentrations of calcite and muscovite compared to station B. On the other hand, the soil samples exhibit a lower mean percentage for quartz (27.7±4.7) and higher percentages for calcite, chlorite, halite and muscovite compared to the airborne samples. These mineralogical airborne dust and soil compositions, derived essentially from the Hamoun source region, reflect the composition of the material available from this provenance as well as the relevant grain size characteristics, enabling the wind storms to entrain this material into the lower atmosphere. While most of the minerals (quartz, feldspars of various types, muscovite) can easily be tied to basement-type lithology of generally gneissic-granitic character, others (chlorite, pyroxenes and hornblende) rather suggest mafic parent rocks, as can be inferred from basic mineralogical analysis [e.g., 104] . However, the calcite, dolomite, halite and gypsum suggest evaporate minerals, although both calcite and dolomite can also reflect alteration products of primary acid or mafic rock constituents. The inferred evaporate minerals reflect local derivation of salt from desiccating water bodies in the Hamoun lakes, originally formed from altered transported components via the Hirmand river system. Thus, the semi-quantitative mineral determinations for the airborne dust over the Sistan region support derivation of the particles from well weathered and well eroded (transported) argillaceous alluvium from the extensive Hirmand river system draining Afghanistan and terminating in the Hamoun Basin. The general geology of Afghanistan encompasses extensive terrains of both acidic and mafic rocks, while similar mineralogical composition of dust (i.e. dominance of quartz, but lower percentage of calcite) was found at the Bagram and Khowst sites located in eastern Afghanistan [44]. More specifically, they found that these sites are underlain by loess (wind deposited silt), sand, clay and alluvium containing gravel. As shown in Fig. 2, as well as in other studies [44,51,12], nearly the whole of Afghanistan is affected by the dust storms originating from Hamoun, since the dust plume usually follows a counter-clockwise direction, carrying wind-blown dust towards eastern Afghanistan. Similarly to our findings, the airborne dust at selected locations in southwestern Iran was found to be composed mainly from quartz and calcite, suggesting detritus sedimentary origin, followed by kaolinite and a minor percentage of gypsum [72]. Furthermore, [44] found that airborne dust samples derived from poorly drained rivers and lakes in central and southern Iraq contain substantial calcite (33-48%), quartz and feldspar with minor chlorite and clay minerals. Previous studies [105][106], have shown that silicate minerals (quartz, feldspars) and phyllosilicates (illite, kaolinite, smectite/montmorillonite clays, chlorite) dominate aeolian dust. Dust samples may also contain substantial amounts of carbonates, oxides, gypsum, halite and soluble salts, but the quantity and percentage of these minerals are quite variable from site to site.

Elemental composition of dust
Knowledge of the chemical composition of airborne dust is necessary for clarifying the likely source regions and is important for quantitative climate modelling, in understanding possible effects on human health, precipitation, ocean biogeochemistry and weathering phenomena [50]. Chemical analysis of dust provides valuable information about potentially harmful trace elements such as lead, arsenic and heavy metals (Co, Cr, Cu, Ni, Pb). On the other hand, the major-element and ion-chemistry analyses provide estimates of mineral components, which themselves may be hazardous to human health and ecosystems and which can act as carriers of other toxic substances. The chemical analysis of dust samples at both stations was performed via XRF analysis for the major oxides (Figs. 20a, b).
In general, the analysis reveals that all samples at both stations contain major amounts of SiO2, mainly in the mineral quartz, variable amounts of CaO in the mineral calcite, plagioclase feldspar and to a limited extent in dolomite, as well as substantial Al2O3 concentrations. More specifically, average major elements of airborne dust at both stations indicate a predominant SiO2 mass component (46.8 -47.8%) with significant CaO (12-12.2%)  in SiO2, Al2O3, K2O and total Fe and significantly enriched in CaO, Na2O and MgO. The MgO is largely contained in dolomite and, to a lesser extent, in clay minerals such as palygorskite and montmorillonite [78,44]. These components can be ascribed to the importance of evaporate minerals such as calcite, dolomite, halite and gypsum (as also suggested by the mineralogical analysis) inferred to have come from the desiccation taking place in the Hamoun dust source region. Furthermore, the elevated values for the trace elements Cl, F and S ( Hamoun dried lake beds, while others had more silica, reflecting weathered rock detritus from the Hirmand river and Afghanistan mountains. An explanation of these variable chemical compositions of dust samples is a real challenge, but it is postulated here that they may reflect local desiccation cycles and, possibly, even micro-climatic changes in the Hamoun-lakes dust source region. Excessive desiccation of the lakes would enhance potential evaporate minerals for deflation in drier periods, while in wetter periods, airborne dust would logically have been derived more from weathered fluvial detritus rich in SiO2. Figure 21 summarizes the results of the elemental compositions determined by XRF analysis at both stations. For comparison reasons, the mean elemental composition found for several sites in southwestern Iran (Khuzestan province) [71][72] is also shown. The vertical bars express one standard deviation from the mean. Concerning the major elemental oxides over Sistan, both stations exhibit similar results, well within the standard deviations, suggesting that the transported dust over Sistan is locally or regionally produced with similarity in source region. In contrast, the mean elemental composition of airborne dust over Khuzestan province exhibits remarkable differences from that over Sistan, revealing various source regions and dust mineralogy. More specifically, the SiO2 percentage is significantly lower and highly variable over Khuzestan, which is also characterized by higher contributions of Na2O, MgO and K2O compared to Sistan. The dust storms over southwestern Iran may originate from local sources as well as being transported over medium-and long-ranges from different sources located in Iraq as well as in the Arabian Peninsula. A comparative study of the mineralogy and elemental composition of airborne dust at several locations in Iraq, Kuwait and the Arabian Peninsula [44] has shown significantly variable contributions, suggesting differences in overall geology, lithology and mineralogy of these regions. In further contrast, airborne dust over Sistan seems to have its individual characteristics originating from local and well-defined sources. The Earth's crust is dominated by silicon and aluminum oxides. Numerous studies [78, 50 and references therein] reviewing the elemental composition of airborne dust over the globe report that mineral dust is composed of ~60% SiO2 and 10-15% Al2O3. The contribution of other oxides, i.e. Fe2O3 (~7%), MgO (~2.5%) and CaO (~4%), are, in general, more variable depending on source location. Furthermore, the review study of [96] showed that airborne dust samples collected over the globe have fairly small variations in elemental composition. The CaO concentrations over Sistan are found to be much higher than those (5.5%) summarized in [96].

Trace elements
The average concentrations of trace elements (in ppm) in dust samples collected during major dust storms at stations A and B are summarized in Table 5 (Fig. 20a). The dominance of chlorine indicates soil salinization in the Hamoun basin and along the Hirmand river and its tributaries. Furthermore, S exhibits higher concentration at station A, while for the other elements the concentrations between the two stations are more or less similar. The concentrations of potentially harmful and toxic elements, like Cs, Pb and As are, in general, low at both stations; however, Ba, Cr and Zn present moderate concentrations.
On the other hand, the analysis of the major element ratios provides essential knowledge of the dust chemical composition and source region. In contrast, the Fe/Al ratio is low over Sistan and is nearly half of that found for airborne dust over southwestern Iran and several locations over the globe [72], but is comparable to that found over central Asia [111]. It should be noted that this ratio remains nearly invariant, ranging from 0.47 to 0.54, for all the collected dust samples at both stations and can be a good surrogate for the dust source region, since any variation in Fe/Al mainly corresponds to variations in clay minerals and not to coating during dust transportation [50]. In contrast, the Ca/Al ratio exhibits the highest variations from sample to sample (2.80-3.46), since it is influenced by particle size, with higher values as particle-size increases [72]. Synoptically, all the ratio values and the low standard deviations suggest similarity in geochemical characteristics over Sistan and a uniform source of airborne dust. Cl  28670  15047  F  13938  13456  S  4445  2506  Ba  210  253  Sr  154  125  Zr  83  76  Cr  70  84  V  69  69  Zn  57  51  La  30  32  Rb  24  19  Ni  18  16  Ce  17  16  Cs  14  13  Sc  11  11  Cu  11  11  Pb 10 10

Conclusions
The present Chapter focused on shedding light on the dust loading, PM concentrations, physical and chemical composition of dust in the Sistan region, southeastern Iran, which constitutes a major dust source region in south west Asia. Sistan region is a closed topographic low basin surrounded by arid and rocky mountains, while its northern part drains the Hilmand river, thus constituting a wetland area known as Hamoun. Hamoun lakes complex have an area about 4500 Km 2 with water volumes of 13025 million m 3 and play the role of a "water cooler" for the region when they are full of water as the severe winds blow across the lakes.
Severe droughts over the past decades, especially after 1999, have caused desiccation of the Hamoun lakes, leaving a fine layer of sediment that is easily lifted by the wind and therefore making the basin one of the most active sources of dust in south-west Asia [56,50]. The strong "Levar", especially during the summer season, blows fine sands off the exposed lake bed and deposits this detritus within huge dune bed forms that may cover a hundred or more villages along the former lakeshore. As a consequence, the wildlife around the lake has been negatively impacted and fisheries have been brought to a halt, which also implies an impact on society. The drainage of the Hamoun wetlands, in association with the intense Levar winds in summer, is the main factor responsible for the frequent and massive dust storms over the Sistan region. Analysis of water surface in combination with dust storms showed that the Hamoun dried beds, particularly Hamoun Saberi and Baringak, have a dramatic effect on dust storms as sources of aerosols.
Systematic PM concentrations were measured in Zabol city, affected by the Sistan dust storms, covering the period September 2010 to August 2011. The results show that the PM10 concentrations were considerably higher than the corresponding European Union air quality annual standard. The analysis of the daily PM concentrations showed that the air quality is affected by dust storms from the Sistan desert, which may be very intense during summer. Hamoun, as an intense dust source region, caused a dramatic increase in PM10 concentrations and a deterioration of air quality (65% of the days were considered unhealthy for sensitive people and 34.9% as hazardous).
Dust loading from the Hamoun basin appears to have a significant contributing influence on the development of extreme dust storms, especially during the summer days. This influence firstly seems to depend on the intensity and duration of dust storms, and secondly, on the distance from the source region, the wind speed and altitude. The grain-size distribution of the dust loading was strongly influenced by the distance from the dust source, since grain sizes shifted to larger values towards station B that is closer to the Hamoun basin. Furthermore, the particle size distribution exhibited a shift towards lower values as the altitude increases, with this feature seen to be more obvious amongst larger size particles, while the frequency of particles below 2.5 μm seemed not to be affected by altitude. In general, the regional dust loading and characteristics are subject to significant spatiotemporal variability. This finding necessitates more systematic observations at as many locations as possible around the Hamoun basin in order to improve the understanding of forcing dynamics, transport mechanisms as well as to quantify the dust amounts emitted from the Hamoun basin.
To fully understand mineral dust characteristics and the potential impact on human health, dust mineralogy and geochemical properties were examined in the Sistan region by collecting airborne samples at two stations and soil samples from several locations over Sistan and the Hamoun basin. The Sistan region is an ideal site to study the nature of dust storms as it receives large amounts of fine alluvial material from the extended Hirmand river system draining much of the Afghanistan highlands, which comprise crystalline basement rocks, Phanerozoic sediments and extensive flood basalts. As a result, large quantities of quartz-rich, feldspar-and mica-bearing silt, as well as mafic material from flood basalt sources and carbonate minerals from dolomites, are transported to the Hamoun wetlands in northern Sistan. Due to droughts at Hamoun and large irrigation projects upstream on the river catchment, extensive desiccation has occurred in the wetlands resulting in large dry lake environments. These have produced large quantities of evaporate minerals to add to the alluvial silts, and the combination of these materials provides the provenance for the airborne dust.
Dust aerosol characterization included chemical analysis of major and trace elements by XRF and mineral analysis by XRD. The results showed that quartz, calcite, muscovite, plagioclase and chlorite are the main mineralogical components of the dust, in descending order, over Sistan, and were present in all the selected airborne dust samples. In contrast, significantly lower percentages for enstatite, halite, dolomite, microcline, gypsum, diopside, orthoclase and hornblende were found, since these minerals occurred only in some of the samples at both stations. On the other hand, SiO2, CaO, Al2O3, Na2O, MgO and Fe2O3 were the major elements characterising the dust, while large amounts of F, Cl and S were also found as trace elements. The mineralogy and chemical composition of airborne dust at both stations were nearly the same and quite similar to the soil samples collected at several locations downwind. This suggests that the dust over Sistan is locally emitted, i.e. from the Hamoun basin, and in a few cases can also be long-range transported to distant regions. On the other hand, individual dust storms showed significant differences between either evaporite-dominated aerosols or those characterized by deflation from alluvial silts. These possibly reflect either localized climatic cyclicity or desiccation cycles. However, in some cases the soil samples showed poor comparisons with aerosol compositions, suggesting that dynamic sorting, soil-forming processes and climatic influences, such as rainfall, altered the mineralogy and chemistry in these partially eolian deposits. Sistan is also an ideal site for studying dust storms and enrichment factors relative to crustal norms; the latter factors suggest that the dust is essentially of crustal rather than anthropogenic origin. SEM analyses of the samples indicated that airborne dust has rounded irregular, prismatic and rhombic shapes, with only the finer particles and a few examples of the coarser dust being spherical.