The dominant mode of aerosol variability over West Asia (WA), a broad region extending from the Red Sea (RS) through the Arabian Peninsula (AP), Iranian Highlands, Arabian Sea (AS), and Afghanistan–Pakistan–Indian continent, is due to dust aerosols emitted from regional arid and semi-arid lands1,2. WA contains several prominent dust sources in AP, Iraq, Iran, Afghanistan, and the northwest Indian regions, among which, the AP deserts are the prime emission source3,4. Dust aerosols emitted from WA influence the regional climate by altering the net solar radiation, cloud properties, and precipitation5. During spring and early summer, elevated dust concentrations over AP and Iran establish an anomalous low-pressure system over the southern AP, AS, and Iran plateau through strong radiative feedback. This peculiar dynamical setup intensifies the low-level southwesterly monsoon wind, and particularly, favors a strong moisture convergence over the core monsoon zone of India and increases the mean seasonal monsoon rainfall5,6. Dust aerosols also alter the precipitation efficiency of clouds by actively contributing to the formation of ice nuclei7. Under high buoyant and moist conditions, dust aerosols facilitate the development of deep mixed/ice phase monsoon clouds through indirect effects and promote higher precipitation rates 8,9,10,11. The high dust concentrations at the cloud base cause the evaporation of cloud liquid water because of strong semi-direct effects and thus, inhibit the growth of oceanic monsoon clouds12. Moreover, dust deposition provides a surplus amount of soluble nutrients to the marine ecosystem, making it one of the key factors for ocean biological productivity13,14. Long-range transported dust can also neutralize the acidic nature of monsoon rains and provide large amounts of nutrients that may significantly impact the monsoon ecosystems15,16.

The strength of dust effects strongly depends on the vertical distribution of dust aerosols in the atmosphere17,18. Vertical variations in the dust aerosol concentrations alter the vertical profile of radiative heating rate, thereby influencing convection through its effect on the atmospheric stability19,20. The atmospheric lifetime of dust aerosols increases in the free troposphere21, allowing the transfer of dust particles over long distances from the source regions and, in turn, influencing the far-away cloud cover, rainfall, spatial features of the dry and wet dust deposition rates, and air quality22,23.

Dust particles injected into the atmosphere follow specific transport routes, termed dust transport pathways, and are an integral component of the dust life cycle24,25. Numerous studies have been performed to delineate the transport of dust aerosols over the north Atlantic and north Pacific oceans based on satellite observations and numerical model simulations26,27,28. However, to date, such efforts for the dust-rich WA region have been limited, and at best, remain at an exploratory stage. Yu et al.29 described the summer dust transport route and its association with the Shamal winds over the AP region using the Multiangle Imaging Spectro Radiometer (MISR) cloud motion vectors. Jin et al.30 indicated the establishment of low-level westerlies can transport of dust during the high dust accumulation days over the AS in summer. To the best of our knowledge, these are the only two reported seminal studies that were performed to investigate the dust transport corridors over WA. Although dust aerosols over the WA play a vital role in modulating the crucial facets of the regional climate, the three-dimensional (3-D) distribution of WA dust, its viable transport pathways, and the factors responsible for the respective distributions are yet to be described in detail. The present study was conducted to examine, these features using decade-long measurements (between 2007 and 2017) of dust aerosol properties obtained using the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP)17. In addition, the Modern-Era Retrospective analysis for Research and Applications, version 2 (MERRA-2) and Copernicus Atmosphere Monitoring Service (CAMS) reanalyses (see “Methods” section) were used to investigate the dust and meteorological parameters.

We investigate the spatiotemporal changes in the dust aerosol distribution over WA. We further analyze the variability in the meridional distribution of dust over RS, AP, AS, and IS with the season, and also, we discuss the vertical features of the zonal transport of Arabian dust during summer. Furthermore, we detail the seasonality in dust transport pathways, with a special emphasis on those present during summer.


Comparison of DAOD from CALIOP, MERRA-2, and CAMS vs AERONET

Accurate ground-based aerosol measurements are needed to evaluate the fidelity of satellite aerosol retrievals and aerosol model simulations. Here we use Level-2.0 quality-assured and cloud-screened coarse-mode optical depth retrievals as ground truth for the validation of satellite (CALIOP) and reanalyses (MERRA-2 and CAMS) dust datasets. We first perform one-to-one comparisons through a scatter plot analysis of the monthly mean coarse-mode optical depth at 500 nm from AERONET-Level-2.0 against the monthly mean Dust Aerosol Optical Depth (DAOD) from CALIOP (Fig. 1a), MERRA-2 (Fig. 1b) and CAMS (Fig. 1c), respectively. The selected AERONET sites (Fig. 1d) provided data for at least 5 years. To avoid the influence of sea-salt aerosols, AERONET sites close to the coastal regions were not considered. The scatter plot indicates that CALIOP is in good agreement with the AERONET coarse-mode values below 0.2; however, CALIOP overestimates the values above 0.3 (Fig. 1a). MERRA-2 slightly underestimates the DAOD values above 0.3 (Fig. 1b), whereas CAMS underestimates most of the DAOD values compared to AERONET (Fig. 1c). CALIOP exhibits a correlation of 0.76 and a root mean squared error (RMSE) of 0.16 with the AERONET observations, consistent with the results of earlier studies17,31,32. However, the presence of clouds, aerosol retrieval algorithm, footprint, and sampling intervals of the CALIOP reduced the one-to-one correlation with AERONET17,32. Despite these limitations, CALIOP is considered a prime dataset for studying and validating the global and regional model simulations of aerosols33,34. Furthermore, CALIOP provides the spatial and vertical distributions of the dust aerosols, which are crucial to understand their life cycle. MERRA-2, which assimilates the AERONET observations, shows better agreement with these measurements compared to CAMS, achieving a good correlation of 0.86, and a small RMSE of 0.07 (for CAMS: correlation: 0.83; RMSE: 0.1). All the aforementioned correlations above are statistically significant at 99% level. Even though both satellite and reanalyses dust datasets show a good correlation with the AERONET observations, the inherent biases in each dataset will result in uncertainties in the dust distributions. As discussed above, the cloud contamination and low sampling rates influence the CALIOP dust retrievals. During the summer season, the persistent cloud cover over the RS and AS may lead to important uncertainties in the CALIOP inferred dust distribution. Further, the imitations in dust aerosol schemes used in the aerosol reanalyses systems occasionally misrepresent the evolution of the dust events and their downwind transport.

Fig. 1: Comparison of satellite and reanalysis dust products with AERONET dust measurements.
figure 1

Evaluation of a CALIOP, b MERRA-2, and c CAMS DAOD against the AERONET coarse-mode optical depth for all sites across West Asia during the January 2003–December 2017. The black dashed line represents 1:1 slope line and the black solid line represents the actual slope. The locations of AERONET stations are indicated as blue colored markers in d.

Spatial distribution of DAOD

The spatial distributions of the seasonal mean DAODs over the WA region, obtained from CALIOP, MERRA-2, and CAMS, for the boreal winter (December–February, DJF), spring (March–May, MAM), summer (June–August, JJA), and autumn (September–November, SON) are displayed in Fig. 2a–i, respectively. The objective of this study was to delineate seasonal changes in the mean dust loading and transport.

Fig. 2: Spatial distribution of dust aerosol optical depth.
figure 2

Seasonal climatology of Dust aerosol optical depth from CALIOP (ad), MERRA-2 (eh), and CAMS (il) over the period 2007–2016.

Considerable growth in the dust loading is observed in the spring season (Fig. 2b) than in the preceding winter (Fig. 2a), as CALIOP measures high DAOD values (>0.5) concentrated over the Kingdom of Saudi Arabia, Arabian Gulf, Pakistan, and northwest Indian regions. Almazroui et al.35 noted that the passage of mid-latitude eastward propagating circumglobal wave train lowers (increases) the surface pressure (winds) over the northern AP during spring, which favors the extreme dust conditions. Similarly, the development of Sahara cyclones and Mediterranean low-pressure systems can pile up dust over the eastern AP and Arabian Gulf regions29. In the earlier studies based on the ground-based AERONET and satellite measurements, an enhancement in the dust activity during spring over central Saudi Arabia was reported36,37. Low-level easterlies (Supplementary Fig. 2b, f) over eastern Sudan and the Tokar Gap region support the westward displacement of the dust. A prominent eastward migration of the dust into the IS during spring is noticeable from the CALIOP observations (Fig. 2b). The strong low-level westerlies (Supplementary Fig. 2f) favor dust accumulation in this region. Frequent intrusions of mid-latitude airmasses amplify the spring season dust activity over northwest Indian regions38; previously, similar findings were reported using dust retrievals from the MODIS sensor4,39.

In the following summer (Fig. 2c), nearly two-third of the WA region gets covered by a heavy load of dust aerosols, with a prevalence of high DAOD (>0.7) values across the broad region of southern RS, Gulf of Aden, south–central AP, southern Iran, and Pakistan. The strengthening of the winds over the dust source regions of WA lifts the dust, leading to dust accumulation and downwind transport. The strong downwind transport from AP favors the maximum dust loading over the Arabian Gulf, Gulf of Aden, and AS. A significant west-east gradient of DAOD is noticeable over the AS, possibly related to the reduction in the dust concentrations (dust deposition) along the dust’s transit across the AS. The higher precipitation rates (Supplementary Fig. 1g) along the eastern boundary of the AS reduce the eastward migration of the dust to the Indian landmass and Bay of Bengal. Likewise, the onset of the monsoon low-level jet (MLLJ) suppresses the buildup of dust over the southern AS and northwest Indian ocean. The low DAOD values in autumn (Fig. 2d) and winter (Fig. 2a) can be attributed to the variations in the intensity and direction of the surface winds as well as the increase in the rainfall at the periphery of the dust sources40,41.

CAMS underrates the high dust loading as well as the eastward advection of dust over the IS during spring, relative to CALIOP observations. Pronounced differences are also noticeable between the MERRA-2 and CAMS observations in the summer season (Fig. 2g, k). Although a moderate to the high amount of dust load is seen in the CALIOP observations over the RS, Arabian Gulf, and Gulf of Aden owing to a strong downwind transport (Fig. 2c), the corresponding signal from the CAMS is comparatively weak (Fig. 2k). MERRA-2 adequately illustrates the downwind transport. We also provide a Supplementary Movie 1 (as a supplement to Fig. 2), which depicts the monthly changes in the dust accumulation and transport over WA. The dust fraction represents the fraction of dust optical depth present in the total optical depth17. As the DAOD, the dust fraction also exhibits substantial seasonal and spatial variations over the WA (Supplementary Fig. 3). Supplementary Fig. 3 indicates a higher dust fraction in the land regions, with persistent dust emissions, and nearby downwind regions during spring (Supplementary Fig. 3b) and summer (Supplementary Fig. 3c). The IS shows a higher dust fraction (>0.6) during spring, implying a dominant contribution of dust in the total aerosol loading. CAMS underestimates the dust fraction compared to CALIOP and MERRA-2, with higher anthropogenic species likely contributing more to the overall aerosol load42.

Vertical distribution of dust aerosols

In this section, we examine the profiles of dust aerosol extinction coefficient (DAEC), obtained from CALIOP, and the dust mixing ratios from MERRA-2 and CAMS, to elucidate the seasonality in the vertical distribution of dust aerosols. WA is divided into four regions based on the location of dust sources and sink areas: (a) RS (33°E–43°E), (b) AP (43°E–60°E), (c) AS (60°E–70°E), and (d) IS (73°E–90°E).

The RS is positioned at the center of the dust conveyor belt between the Saharan and Arabian deserts3. High mountain ridges on both sides of the RS influence the seasonally developing wind regimes43 and channelize the dust towards the RS44. Figure 3 displays the mean meridional distribution of DAEC and dust mixing ratios for different seasons over the RS. Overall, throughout the year, the magnitude of dust accumulation over the southern RS is larger than that in the northern RS, because of nearby dust sources and favorable wind patterns45,46. The dust accumulation observed in the CALIOP data varies rapidly from season to season, showing the maximum accumulation in summer (Fig. 3c) and the minimum during winter (Fig. 3a). A hitherto unrecorded distinct “tower-like” vertical distribution of the dust is identified during summer between 15 and 20°N latitudinal zones. Further, the same region also shows abundant dust aerosols in the atmosphere up to 500 hPa, and high DAEC values (>0.12 km−1) below 700 hPa. We attribute this low-level DAEC maximum to the occurrence of the Tokar Gap Jet around the 850 hPa level during summer. The Tokar Gap Jet blows almost all the days during summer over the elevated dust sources (e.g., Tokar Gap in Eritrea) located along the western boundary of the RS. The high jet winds uplift dry soil from the ephemeral sources and transfer it to the south–central RS47. Banks et al.48 also reported a significant amount of eastward dust flux below the 2 km altitude over the south–central RS. Due to the close proximity, the coastal plains of the RS also contribute to the near-surface dust loading over the southern RS36,49. The vertical decay in the DAEC is almost negligible from 15 to 20°N between the 700 and 500 hPa layer, suggesting a continuous supply of dust through the long-range transport. The summer wind climatology (Supplementary Fig. 2k) shows strong easterly and northeasterly winds at 700 hPa over the southern RS, contributing to the transport of the emitted dust from AP. The dust mixing ratio obtained from MERRA-2 (Fig. 3e–h) and CAMS (Fig. 3i–l) can reproduce the dust seasonality derived from CALIOP. MERRA-2 well reproduces the position of the summer dust-tower (Fig. 3g) seen in the CALIOP data. However, this marked feature is not prominent in the CAMS (Fig. 3k).

Fig. 3: Zonal climatology of dust aerosol distribution over Red Sea.
figure 3

Seasonal zonal climatology of Dust aerosol extinction from CALIOP (ad), and dust mixing ratios from MERRA-2 (eh) and CAMS (il) over the Red Sea (zonally averaged between 33°E and 43°E) during the period 2007–2016.

The AP is the second-largest dust source region in the northern hemisphere, behind North Africa, and is the major contributor of dust emissions over the WA3,4. Over the AP, the rising surface air temperatures between winter and summer (Supplementary Fig. 1a–d) and negligible annual rainfall (Supplementary Fig. 1e−h) increase the distribution of erodible soils, which favor high dust emissions. Figure 4 shows the latitude–pressure transect plots of the dust derived from CALIOP (Fig. 4a–d), MERRA-2, and CAMS (Fig. 4e–l). Evidently, dust uplift, lofting, and low–high loading persistently occur in all the seasons over the AP. Throughout the year, the maximum DAEC is confined between the latitudinal belt of 15−35°N, which coincides with the low-lying flat dust source areas of the eastern AP. Notable dust aloft at 600 hPa is noticeable during spring (Fig. 4b), compared to the preceding winter (Fig. 4a). During spring, the CALIOP reports high DAEC values (>0.16 km−1) below 700 hPa along with a negligible vertical gradient in the DAEC between the surface and 850 hPa level, suggesting an even distribution of dust throughout the arid boundary layer. However, this may also be an artifact from the averaging of the diurnal boundary processes. A similar pattern persists till the following summer (Fig. 4c). Dust is lifted vertically up to a level of 500 hPa (~5 km) in summer, forming a dust-filled layer over a wide region. The onset of low-level Shamal winds and associated dust emissions apparently result in the high DAEC values (>0.16 km−1). Similarly, a strong buoyancy due to the high surface temperatures further favors the uniform mixing of dust through the elevated boundary layer (>2.5 km) (Supplementary Fig. 1k). Harikishan et al.50 also identified the maximum CALIOP-derived extinction below 2 km, during the extreme dust events over the AP. Notably, we observe a slantwise advection of the dust between the 850 and 700 hPa level. A clear partition between the surface and the advected dust layer is also observed south of 18°N, and the rate of separation increases toward the south. The intensity of the advected dust layer gradually reduces between 10 and 18°N, after which a rapid decrease is observed. This type of phenomenon was reported earlier in the west Saharan region51. The physical mechanism underlying the aforementioned dust advection is analyzed in the later section. The dust uplift and lofting are less active during the autumn and winter periods (Fig. 4a, d), exhibiting a smaller vertical thickness relative to the other two seasons. Although both MERRA-2 (Fig. 4e–h) and CAMS (Fig. 4i–l) adequately describe the seasonal evolution of the dust aerosols over the AP, MERRA-2 captures the summer meridional dust features and slantwise dust advection more effectively than CAMS.

Fig. 4: Zonal climatology of dust aerosol distribution over Arabian Peninsula.
figure 4

Seasonal zonal climatology of Dust aerosol extinction from CALIOP (ad), and dust mixing ratios from MERRA-2 (eh) and CAMS (il) over the Arabian Peninsula (zonally averaged between 43°E and 60°E) during the period 2007–2016.

Geographically, the AS is bordered by hotspots of dust, which result in a high dust accumulation in this region. The immediate dust sources in the northern AS are topographically elevated (>1000 m) than those in the westside (<250 m). Figure 5 presents the seasonal characteristics of the meridional dust distribution over AS. As shown, AS is covered by dust throughout the year, although its intensity is characterized by strong seasonality. The dust is confined to the lower levels of the atmosphere (<1 km) during winter (Fig. 5a), whereas uplifting of the dust to the higher levels is observed in spring and summer (Fig. 5b, c). Over the AS, summer is the most active “dust” season, with a core (DAEC > 0.16 km−1) between 20 and 30°N. Furthermore, the low-level northerly winds from Iran52,53 and the north westerlies from AP30 further contribute to the higher DAEC values observed between 20 and 25°N. Tandule et al.54 identified the maximum dust activity over the AS during summer using the CALIOP observations. Similar to the AP, an advection of dust from the north favors the formation of an uplifted dust layer above the 850 hPa level between 10 and 20°N. The dust activity over the AS weakens toward the end of the year, and the dust concentration becomes insignificant in autumn and winter (Fig. 5a, d). Both MERRA-2 (Fig. 5e–h) and CAMS (Fig. 5i–l) data show good agreement with the seasonal characteristics of dust over the AS, observed using the CALIOP. Although the typical summer mid-tropospheric advection of dust is adequately reproduced in both the reanalyses, the CAMS reanalysis does not replicate the detected deep vertical development of dust between 20 and 35°N.

Fig. 5: Zonal climatology of dust aerosol distribution over Arabian Sea.
figure 5

Seasonal zonal climatology of dust aerosol extinction from CALIOP (ad), and dust mixing ratios from MERRA-2 (eh) and CAMS (il) over the Arabian Sea (zonally averaged between 60°E and 70°E) during the period 2007–2016.

The dust accumulated over the IS has a significant effect on the monsoon clouds and dynamics, snow albedo, and air quality 55,56,57,58. Figure 6 displays the seasonal evolution of the latitude–pressure cross-sections of DAEC and dust mixing ratio over the IS. A strong seasonality is evident in the meridional distribution of DAEC (Fig. 6a–d), with a very low dust activity in winter (Fig. 6a), peaking during the spring season (Fig. 6b). The maximum DAEC occurs in spring and is a result of the prevailing strong low westerlies, which enable the uplifting of dust over the natural and anthropogenic sources collocated in northwest India4. In addition, the convective down-drafts and northward shift of the inter-tropical discontinuity also initiate anomalous dust emissions59,60. The vertical stretch of the dust layer is also high during spring, as the dust reaches up to ~600 hPa (~3.5 km). A prominent boundary layer transport of the dust is visible from 18 to 10°N, mostly facilitated by the low-level northerlies from the higher latitudes (Supplementary Fig. 2f). The dust load north of 35°N is contributed by the Taklimakan desert, where the dust emissions are maximum during spring61. The depth of the dust layer between 20 and 30°N reduces during summer (Fig. 6c) owing to the wet scavenging favored by the onset of the Indian summer monsoon. A weekly intense mid-tropospheric dust transport is also noticeable above the 850 hPa level stretched between 10 and 20°N. Dust from the AP is transported to the IS between the top of the boundary layer and the mid-troposphere, as reported in a study based on continuous ground-based lidar measurements15. Notwithstanding the strong seasonality observed in the MERRA-2 (Fig. 6e–h) and CAMS (Fig. 6i–l), both these reanalyses underestimate the high boundary layer dust concentrations in the IS, measured by the CALIOP sensor during spring and summer.

Fig. 6: Zonal climatology of dust aerosol distribution over Indian Subcontinent.
figure 6

Seasonal zonal climatology of dust aerosol extinction from CALIOP (ad), and dust mixing ratios from MERRA-2 (eh) and CAMS (il) over the Indian Subcontinent (zonally averaged between 73°E and 90°E) during the period 2007–2016.

Zonal transport of dust over the Arabian Sea (AS) during summer

In the above section (see section Vertical distribution of Dust Aerosols), we described the features related to the slantwise advection of dust aerosols over the AP and AS during summer. In this section, we examine the zonal transport characteristics of the advected dust aerosols by analyzing the longitude–pressure cross-sections of the DAEC and dust mixing ratio distributions averaged between 5 and 20°N (Fig. 7a–c), the region over which the maximum amount of dust flux is observed during summer (Fig. 8c). In Fig. 7a, the high values of DAEC (>0.14 km−1) near the surface in the 40–50°E region reflect higher emissions of dust. The dust mobilization is confined to a single layer with lower and upper bounds at 850 and 500 hPa, respectively, and vertically located just above the top of the MLLJ core (Fig. 7d). Harikishan et al.12 reported the presence a of dust layer between 2 and 5 km over the AS. The results of the present investigation provide a better representation of the zonal dust transport over the AS and explain the underlying physical mechanism.

Fig. 7: Zonal dust transport over Arabian Sea during summer season.
figure 7

Summer climatology of the Longitude–Pressure cross-sections of a Dust aerosol extinction from CALIOP, Dust mixing ratios from b MERRA-2 and c CAMS, and d wind speed (shaded) and vectors from MERRA-2 averaged between 5°N and 20°N during the period 2007–2016.

Fig. 8: Seasonal changes in the dust transport pathways.
figure 8

Seasonal climatology of MERRA-2 column integrated horizontal dust mass flux (shaded) and vectors for a DJF, b MAM, c JJA, and d SON, during the period 2007–2016.

A plausible explanation of the formation and persistence of the mid-tropospheric dust transport route over the AS is suggested based on the aforementioned analysis. During summer, the dry Shamal winds over the AP favor the development of a deep vertical core of dust aerosols through a sturdy wind erosion of fertile soils. The intersection of the southwesterly monsoon winds with the dry and dusty Shamal winds at the lower levels develops strong-inversion layers above MLLJ62,63. The advection of dust into these inversion layers creates a thick dust blanket, which is later spread across the AS by the MLLJ westerlies present at that level. Zhu et al.64 identified the mid-tropospheric dust layer over the AS between 850 and 500 hPa and attributed its development to the presence of deep inversion layers above the MLLJ. As this dust layer resides above the marine monsoon clouds, wet scavenging of the dust is negligible, favoring a persistent mid-tropospheric dust passage. Cuesta et al.51 indicated that the dust particles with diameters larger than 30 µm settle down at a faster rate, and the long-range dust transport is mostly mediated by the smaller particles (diameter < 10 µm). This suggests that the dust transported in the middle levels of the troposphere is likely composed of fine dust particles, whose wet and dry deposition rates are relatively low. Thus, field campaigns are needed to examine the dust size distribution along the zonal dust transport pathway.

The zonal and vertical characteristics of the dust transport over the AS are well described by MERRA-2 and CAMS, although the magnitudes of the transferred dust derived from these reanalyses are different. MERRA-2 yields the highest mass fraction of emitted fine dust65, which may explain the higher amount of zonal dust transport shown by this reanalysis.

Dust transport pathways over West Asia (WA)

To investigate the dust transport pathways, we first analyzed the column integrated horizontal dust mass flux (DMF)66 obtained from the MERRA-2 reanalysis. Figure 8 outlines the seasonal variations in the DMF over WA. The DMF is low during winter (Fig. 8a) and autumn (Fig. 8d), moderate in spring (Fig. 8b), and highest in summer (Fig. 8c). In spring, the maximum influx of dust aerosols is observed over the north Indian region. The establishment of low-level subtropical westerlies accelerates the transfer of the dust emitted from local and remote regions to the north Indian landmass. The Himalayan Mountain range acts as a barrier to the north- and northeast-bound transport of dust aerosols, resulting in their accumulation over the foothills. However, in the following summer, the magnitude of DMF peaks over the AP and AS with the establishment of prominent transport pathways across these regions. Undoubtedly, the dust transport pathways identified over the AP and AS are primarily responsible for the rapid recirculation and further transboundary transport of the dust emissions in summer. The physical mechanism underlying the evolution of these preferential summer dust pathways is yet to be established.

Here, we analyze the diurnal changes in the spatial distribution of DMF over the AP, using MERRA-2. To generate a diurnal climatology, we take the mean of DMF values at 0000, 0600, 1200, and 1800 UTC hours during the summer months of 2007–2016. The diurnal progression of DMF over the AP (Fig. 9a–d) reveals the prominent spatial and diurnal features linked with dust transport. The anomalous patterns in DMF and wind diurnal variability are presented in Supplementary Fig. 4. Throughout the day, the northwestern DMF vectors are noticeable over the northern AP, changing to the northeast direction south of 25°N. There is significant diurnal variability in the DMF over the AP, with a low DMF at 0600 UTC (Fig. 9b) and a peak DMF between 1200 (Fig. 9c) and 1800 UTC (Fig. 9d). Figure 9d also shows the maximum spatial spread/extent of the DMF, with a core >0.006 kg m−1 s−1, located along the wide low-lying regions of the eastern AP and the Rub-Al-Khali desert. Anomalously, 1800 UTC shows the highest DMF over AP (Supplementary Fig. 4d). Both the Arabian Gulf and RS receive more influx of dust during the evening hours (i.e., 1800 UTC). Based on this, it can be stated that the out-of-phase relationship between the surface winds and the lower-level Shamal winds triggers the distinctive diurnal fluctuations in the DMF over the AP. To elucidate this, we depict the 10 m winds at 0000, 0600, 1200, and 1800 UTC during summer (Fig. 9e–h). In the morning hours, the surface winds are weak (Fig. 9f) and reach their maximum strength at 1200 UTC (Fig. 9g). These changes in the wind intensities favor the diurnal changes in the DMF spatial distributions, as shown, respectively, in Fig. 9b, c. The observed diurnal variability in the surface Shamal winds can be attributed to intraday variations in the pressure gradient between the eastern Mediterranean and Iranian heat low. The steep pressure gradient around the noontime between these two contrasting pressure centers29 results in strong surface winds (Fig. 9g) and the consequent dust transport at 1200 UTC (Fig. 9c). As the day progresses, the pressure gradient decreases, which weakens the strength of the surface winds.

Fig. 9: Diurnal distribution of dust transport and winds.
figure 9

Spatial patterns of 6 hourly summer season climatology of column integrated horizontal dust mass flux (shaded) and vectors (ad), 10 m winds (eh), and low-level winds averaged between 925 and 700 hPa levels (il) for the period 2007–2016. All the plots are based on the MERRA-2 reanalysis data.

However, a DMF peak is observed at 1800 UTC (Fig. 9d). The evolution of the low-level jet (LLJ) during the local nocturnal enhances the DMF. To explain this, we compute the vertical average of the winds between the 925 and 700 hPa levels, where the persisting low-level Shamal jets are observed67. At 1800 UTC, both the Shamal LLJ winds (Fig. 9l) and the DMF display their maximum spread and intensity. After the sunset, the mountain slopes of Zagros cool at a faster rate compared to that of the nearby valley region, thereby establishing the LLJ just above the soil surface, often at or above 100 m, across the eastern AP67. The formation of night-time LLJ over the orographic channels of the north African deserts has been well examined in the previous studies68. The LLJs can transport ample amounts of dust to distant regions because of their higher wind speeds. However, under certain conditions such as when the vertical wind shear underneath the LLJ core reaches a critical stage or under the influence of very large background pressure gradients, the LLJ momentum moves to the surface69. This enhances the surface dust emissions and the subsequent dust transport. Karam et al.70 also discussed the influence of out-of-phase diurnal evolution of surface and low-level Shamal winds on the AP dust emissions using Meso-NH dust model simulations. Geographically, we can identify the homogeneous intensification of the dust transport, which is relatively coupled to the homogeneous growth of the surface and LLJ Shamal winds, over the AP.

The DMF spatial distribution over the AS shows a complex diurnal variability (Fig. 9a–d). The peak DMF at 0000 UTC (Fig. 9a) extends from the Gulf of Aden to the west–central AS. As the day advances, the intensity of the DMF changes rapidly over the AS. This suggests that these diurnal fluctuations are linked with the diurnal variability of the MLLJ intensity, Shamal winds, and associated inversions. As discussed in an earlier section, dust transport over the AS occurs between 850 and 600 hPa, owing to the advection of the dust aerosols into the inversion layers, which are developed above the monsoon boundary layer. Consequently, the strength of these inversion layers influences the amount of dust advection. Dwivedi et al.63 reported that the magnitude of the inversions varies diurnally showing a peak at 0000 UTC, associated with the maximum wind shear between the Gulf to Aden and the western AS. The core of the DMF (Fig. 9a) is mostly collocated with a spatial pattern of high wind shear at 0000 UTC. Due to the maximum strength of inversion, the downward mixing of dust aerosols is almost negligible, supporting the sustenance of high DMF values over the AS. As the day progresses, the central core of the maximum DMF moves eastward with the monsoon westerlies at a reduced intensity and remains between 65 and 70°E at the end of the day. The rapid decay in the inversion strength between 0600 and 1200 UTC63 further weakens the DMF over the AS (Fig. 9b, c).


This study examined for the first time the seasonality in the 3-D structure and transport pathways of dust aerosols over the entire region of WA, which includes the RS, AP, Iranian Plateau, AS, and the IS. We analyzed the DAOD properties from CALIOP observations, MERRA-2, and CAMS aerosol reanalyses. The DAOD from CALIOP, MERRA-2, and CAMS showed a good correlation with the AERONET measurements. Furthermore, MERRA-2 mostly replicated the CALIOP identified dust pattern throughout the year and well captured the seasonal evolution and transport corridors of the dust aerosols. The key findings of this work are summarized as follows.

  • DAOD exhibits a well-defined seasonal cycle, with temporal and spatial distinctions. Our results showed distinctly different magnitudes of DAOD, in all the seasons. The southern RS, AP, AS, Iran–Afghanistan–Pakistan regions experience the maximum dust accumulation during summer, whereas the IS experiences the highest concentrations of dust during spring.

  • The latitude–height cross-sections of the dust aerosol distribution reveal a significant seasonal variation over the WA region. The maximum uplift of dust, between the surface and 500 hPa level over the west–central WA region is observed during summer. In this season, typical vertical distributions are detected over the RS, AP, and AS.

  • Throughout the year, the southern RS dominates the dust loading compared to the northern RS. A summer “dust-tower” forms over the southern RS because of persistent dust transport from the local dust sources, and incoming long-range transport from the eastern AP region. The onset of dust emitting Shamal winds, and the elevated boundary layer during summer allow the formation of a deep vertical core of dust over the AP.

  • The vertical distribution of the dust is significant during spring over the north Indian region. The boundary layer (mid-tropospheric) transport of dust is observed between 10 and 25°N latitudinal zones during spring (summer) over the IS.

  • During summer, the intrusion of moisture-laden monsoon air into a deep column of dry air facilitates the development of inversion layers. This further supports a slantwise advection of the dust aerosols into the inversion layer, between 850 and 700 hPa, simultaneously over the AP and AS.

  • Our results suggest an eastward zonal transport of the dust, within a layer between 850 and 500 hPa from the AP to the IS during summer. The core of the dust transport is situated just above the core of the MLLJ at around the 700 hPa level. The intensity of the transport gradually weakens toward the west coast of India, owing to both dry and wet depositions.

  • There is a strong seasonality in the dust transport routes over the WA. The dust transport is confined to the AP during winter and autumn. A strong transboundary transport, established from the northern AP to the east Indian region during spring, allows a high influx of dust aerosols into the Indian Subcontinent. The magnitude of DMF reaches its maximum during summer, because of the simultaneous establishment of prominent dust transport pathways over the AP and AS. These transport pathways play a prominent role in recirculating the highest dust emissions to the downwind areas.

  • The out-of-phase evolution of surface and low-level jets of Shamal winds induce a significant diurnal variability in terms of the magnitude and spatial spread of DMF over the AP. The diurnal peak of DMF occurs at noontime and is maintained by the strong surface winds. A secondary peak, appearing during the evening hours, is associated with the evolution of a LLJ over the AP. In contrast, the diurnal variability of DMF over the AS is altered by the inversion layers that are formed above the monsoon boundary layer. The reported DMF peak, appearing at the midnight, between the Gulf of Aden and the west–central AS is originated from the maximum strength of the inversions and the associated wind shear.

This paper describes the 3-D structure and transport pathways of dust aerosols based on satellite (CALIOP) and aerosol reanalyses (MERRA-2 and CAMS). These characteristics are crucial for understanding the impact of dust on the regional climate. The interannual variability of the dust emissions and transport pathways over studied regions will be investigated in the near future.


This section describes the datasets used in this study and outlined in (Supplementary Table 1).

CALIOP dust aerosol products

The CALIOP sensor onboard the CALIPSO satellite, with an equator crossing time of around 2 am and 2 pm local times, provides vertical profiles of the aerosols and clouds around the globe17. CALIOP measures the attenuated backscatter profiles from the surface of the earth to the top of the atmosphere every 333 m along the orbit track, which are stored in the Level 1B data product. Then, a Level-2 algorithm is implemented to detect the features, classification types (aerosol, cloud, surface), and retrieve extinction coefficients from the attenuated backscatter signals. Further, the algorithm assigns an aerosol subtype for each aerosol layer, based on a combination of information, such as the aerosol layer height, layer integrated attenuated backscatter, depolarization ratio at 532 nm, and surface type. The CALIOP-derived dust column properties have been extensively validated against the ground-based Aerosol Robotic Network (AERONET) measurements32,71,72 and passive spaceborne sensors73,74. When compared with the ground-based lidar observations, the dust profiles retrieved by CALIOP have been found to be robust and accurate 75,76,77. The CALIOP dust measurements are widely used for exploring the 3-D distribution and transport pathways of dust18,23,78,79 and for evaluating regional and global simulations of dust aerosols33,65,80. Further, Level-3 monthly mean globally gridded-aerosol profile products were developed by aggregating the quality-controlled level-2 aerosol products on a uniform 2° × 5° latitude–longitude grid. We used the CALIOP-Level-3 Version 4.20 all-sky day and night profiles, which include aerosols in clear skies, above clouds, and below the optically thin clouds available at 60 m, with a vertical resolution up to 12 km from January 2007 to December 2017. The details of the quality control methods used in the preparation of CALIOP−L3 data are provided by Winker et al.17 and Tackett et al.81.

MERRA-2 global aerosol reanalysis

MERRA-2, a reanalysis dataset developed using the Goddard Earth Observing System (GEOS–5.12.4) atmospheric model82,83, uses the radiatively coupled Goddard Chemistry, Aerosol, Radiation, and Transport (GOCART) model to simulate the aerosol properties around the globe. MERRA-2 directly assimilates the AERONET and MISR-derived aerosol optical depth, and the bias-corrected aerosol optical depth from the advanced very-high-resolution radiometer (AVHRR) and moderate resolution imagining spectroradiometer (MODIS). Dust emissions are mainly constrained by wind-driven erosion over the source locations and are identified based on the topographic depression map84. MERRA-2 uses the scheme proposed by Tegan and Lacis85 to allocate the size distribution for the airborne dust particles. Dust aerosols are represented in five non-interactive dry size bins of 0.1–1, 1–1.8, 1.8–3.0, 3–6, and 6–10 μm with corresponding effective radii of 0.73, 1.4, 2.4, 4.5, and 8 μm, respectively. MERRA-2 estimates the total dust emission/deposition rates by summing up those obtained from all the five bins. Here, we used the monthly means of several meteorological fields, aerosol, and dust properties available at a resolution of 0.625° × 0.5°, between January 2003 and December 2017.

CAMS global aerosol reanalysis

CAMS is the latest aerosol reanalysis (version 4)86 generated by the European Centre for Medium-Range Weather Forecasts (ECMWF). It provides information on aerosol optical properties around the globe. CAMS assimilates the aerosol optical depth measurement data obtained from the MODIS Terra/Aqua and advanced along-track scanning radiometer satellite sensors. The advection, transport, convection, and diffusion processes are included in the meteorological model of the ECMWF’s integrated forecast system. The reanalysis system comprises the CAMS bulk aerosol model combined with the integrated forecast system chemistry and carbon bond mechanism. In CAMS, the dust aerosol sizes are represented under three different size bins; 0.030–0.55, 0.55–0.9, and 0.9–20 μm. We used the monthly means of aerosol and dust properties from January 2003 to December 2017, available at a resolution of 0.75° × 0.75°.

AERONET observations

The Aerosol Robotic Network (AERONET) provides world-wide spectral observations of Aerosol optical depth and angular distribution of sky radiances at different wavelengths using the Cimel robotic sun photometers87. AERONET data are available at the high temporal resolution but with limited spatial coverage. AERONET retrieves the coarse-mode aerosol optical depth at 500 nm based on the spectral deconvolution algorithm. Here, monthly mean AERONET-Level-2.0 coarse-mode aerosol optical depth at 500 nm is used to validate the dust properties retrieved by the CALIOP and MERRA-2 and CAMS reanalyses. We select, only those stations (listed in Supplementary Table 2 and indicated in Fig. 1d) with enough long-term data to perform the validation analysis over the period of 2003–2017.

Other datasets

The tropical rainfall measuring mission derived monthly mean precipitation rate, available at a resolution of 0.5° × 0.5°, during the period 2007–201688, and the monthly means of surface air temperature from the Climate Research Unit (version 4), available at a resolution of 0.5° × 0.5°, over the period of 2007–201689 were also used to examine the seasonal changes in the air temperature and precipitation. A seasonal climatology of the 3-D dust distributions obtained from the CALIOP was generated over the reference period of January 2007–December 2016. We also considered the same period for the analysis of the seasonal climatology for dust mass flux (DMF), winds, sea-level pressure, geopotential height, surface air temperature, precipitation rate, and boundary layer height obtained from MERRA-2. Even though the AERONET data were available from the early twentieth century, we only used the data within the period 2003–2017, similar to that considered for the MERRA-2 and CAMS aerosol reanalyses.