Genesis of a Severe Dust Storm Over the Indian Subcontinent: Dynamics and Impacts

Abstract

The genesis, dynamics, and impacts of a severe dust storm over the central Himalaya during June 13–17, 2018 have been investigated using in situ measurements, satellite data, and model reanalysis. A low‐pressure system over northern India and prevalence of strong winds (∼20 ms⁻¹) triggered the dust storm leading to poor visibility conditions and five‐fold enhancement in the fine particulate matter (PM_2.5) over the central Himalaya. Enhancements in Aerosol Optical Depth (AOD) were observed to be stronger over the Himalayan foothills site (Lumbini) than that over the Indo‐Gangetic Plain (IGP) site‐Gandhi College. The sharp reductions in Angstrom exponent (α) from about 1.2 to 0.3 indicated the dominance of coarse‐mode aerosols during the dust episode. Model results show an enhancement in the dust from 1.5 to 2.5 Tg (∼70%) over the northern Indian subcontinent, with about half of the contribution from the regional source (Thar Desert). Interestingly, dust storm also had significant impacts on turbulent kinetic energy (2.9–9.6 m² s⁻²), vertical momentum flux (0.9–3.3 Nm⁻²), and sensible heat flux (34.8 to −33.9 Wm⁻²), suggesting turbulent mixing of aerosols and cooling near the surface over the Himalayas. Our study highlights that the large‐scale dust storms exposed to additional dust and pollution from regional sources can profoundly impact the air quality, heat fluxes, and radiative balance over the northern Indian subcontinent. The study would also help in evaluating the results of climate models and to assess the impacts of dust on the hydrological processes and melting Himalayan glaciers.

Introduction

The strong near-surface flows inject massive mineral dust from the arid, semiarid, and desert regions over the globe into the atmosphere during the dust storm and act as a major contributor for natural aerosols and particulate matter (Chin et al., 2007; Ginoux et al., 2012; IPCC, 2013; Middleton, 1986; Sarkar et al., 2019; T. Wang et al., 2021). The particulate matter often expressed as PM_2.5 (diameter ≤2.5 µm) and PM₁₀ (diameter ≤10 µm), composed of dust, black carbon, and organic carbon, are produced via diverse natural and anthropogenic sources and influence human health as well as the climate, very significantly (e.g., Balakrishnan et al., 2019; Fuzzi et al., 2015; IPCC, 2013; Lelieveld et al., 2018; Ramanathan et al., 2001). Dust aerosols have been shown to significantly impact the atmospheric chemistry (Kumar, Barth, Madronich, et al., 2014; Kumar, Barth, Pfister, et al., 2014), radiation balance (e.g., Evan et al., 2014; N. Mahowald et al., 2014), cloud microphysics, and hydrological cycle (Ramanathan et al., 2001; Seinfeld et al., 2004; Vinoj et al., 2014; Yuan et al., 2021). The transported dust also reduces the albedo of the cryosphere and can enhance the retreating rate of glaciers (Gautam et al., 2013; Oerlemans et al., 2009; C. Sarangi et al., 2020; T. Wang et al., 2021; Wiscombe et al., 1980; Zhang et al., 2017). Dust aerosols over the South-Asian region are shown to affect the monsoonal circulation and modulate the rainfall pattern (Lau et al., 2006; Sanap & Pandithurai, 2015). Significant heating of the lower atmosphere has been observed over India, especially over the Indo-Gangetic Plain (IGP) in north India during dust episodes, occurring between March and June typically every year (Pandithurai et al., 2008; Sagar et al., 2014; Solanki & Singh, 2014; T. Wang et al., 2020). Ground-based and space-borne satellite-based observations have demonstrated that the long-range transported dust deteriorates the air quality and the surface albedo of snow cover over the high-altitude region and foothills of the Himalaya (Gautam et al., 2013; Hegde et al., 2007; T. Wang et al., 2021; Zhang et al., 2017).

In the Indian mainland scenario, primary sources of frequent dust episodes are linked with Thar and Margo deserts which influence most of northern India during premonsoon (April–June), supported by a low-pressure system over IGP (Banerjee et al., 2021; Hegde et al., 2007; Kumar, Barth, Madronich, et al., 2014; Kumar, Barth, Pfister, et al., 2014; Pandithurai et al., 2008). However, the analyses of chemical composition and air trajectories also show contributions of long-range transport from north African deserts, Europe, southwest Asian and Saudi Arabian basins (Chinnam et al., 2006; Hegde et al., 2007; Prasad & Singh, 2007; Solanki & Singh, 2014; Solanki et al., 2013; Srivastava et al., 2014). Notably, the IGP is considered to be the regional hotspot of the high-pollution loading with strong anthropogenic and biomass-burning influences (e.g., Dey & Di Girolamo, 2011; Dhaka et al., 2020; Di Girolamo et al., 2004; Jethva et al., 2005; Nair et al., 2007; Ojha et al., 2020; Zhang et al., 2017; T. Wang et al., 2021). The presence of diverse anthropogenic emissions converts the transported dust into polluted dust and, with strong convection and slope winds further transported to the Himalayan region (e.g., Hegde et al., 2007; Lau et al., 2006; Ningombam et al., 2014; Ojha et al., 2020; N. Singh et al., 2016; Zhang et al., 2017). The transport of pollutants to the central Himalaya have been studied earlier, revealing a crucial role of boundary layer dynamics and mountain-valley winds (Ojha et al., 2012; Pandey et al., 2020; T. Sarangi et al., 2014; N. Singh et al., 2016; Srivastava et al., 2015). Elevated terrain of the Himalaya and foothills can block wind flow leading to a trapping and more sedimentation of dust.

The prevailing meteorological conditions and micrometeorological measurements in the lowermost part of the boundary layer, also called the near-surface layer, are used to estimate the dust fluxes over different surfaces, for example, land-ocean (Niedermeier et al., 2014); however, such studies are limited over the Himalayan region (Solanki et al., 2016). The emissions and transport of dust can be mediated through land-atmosphere processes and near-surface dynamics (Evan et al., 2014; Marticorena & Bergametti, 1995). Surface layer characteristics, including stability, turbulent transport, mean flow, and other meteorological quantities, are also impacted by dust (e.g., Dumka et al., 2019; Niu et al., 2016; Zheng & Zhang, 2010). However, the micrometeorological observations for such estimations remain limited over the Himalayan region. Dust aerosols can be transported over regional to intercontinental scales, as shown in various modeling studies (Jing et al., 2017; N. M. Mahowald et al., 2005; Todd et al., 2008; Uno et al., 2006). Previous studies also reported dust fluxes, budgets, and radiative forcing using the model simulations and reanalysis over different parts of the world (Jing et al., 2017; Todd et al., 2008; Uno et al., 2006; Zender et al., 2004). The surface layer characteristics over the Himalayan region were investigated in few recent studies (e.g., Solanki et al., 2016; Solanki et al., 2019), however, the impacts of elevated dust loadings on surface layer remain unclear. Therefore, the study of the surface-layer characteristics, including turbulent transport of mass fluxes, turbulent kinetic energy, and surface winds are needed to evaluate the implications of dust episodes.

Here, we made an attempt to quantify the mixing of the dust and to study the impact of the dust storm on surface layer characteristics over the Himalayan region. Section 2 of this deals with sites, data, and methodology used. Section 3 contains results and discussion, the buildup mechanism, dynamics, and the enroute mixing of pollutants (that led to elevated levels of PM_2.5 over the central Himalaya). Impact analysis is carried out in Section 4, followed by the summary and conclusion in Section 5.

Site, Data Products, and Methodology Ground-Based Measurements

In order to analyze the dust storm, the measurements of PM_2.5 and micrometeorology from a high altitude site, Manora Peak (29.4°N, 79.6°E, amsl ∼1,940 m) in the central Himalaya region, aerosols characteristics from a foothill site Lumbini (27.49°N, 83.28°E) and IGP site Gandhi College (25.871°N, 84.128°E), and the radiosonde observations from Gorakhpur are utilized. The topography of this mountainous region and the locations of these observation sites are shown in Figure 1a.

Figure 1. (a) Topography of the northern Indian subcontinent based on the Shutter Radar Topography Mission (SRTM) with locations of the observational sites: Manora Peak for the micrometeorological measurements; Lumbini and Gandhi College for AERONET observations; and radiosonde station Gorakhpur. (b) True color composite image for June 14, 2018 from the Visible Infrared Imaging Radiometer Suite (VIIRS) Corrected Reflectance imagery (https://worldview.earthdata.nasa.gov).

Measurements of fine particle mass concentration (PM_2.5) have been made over Manora Peak, using an optical sensor based on the distribution of light scattering intensity up to the size particles about the diameter of ∼0.3 μm (Nakayama et al., 2018). Such sensors were earlier used for measurements of indoor and outdoor PM_2.5 concentrations (e.g., Niedermeier et al., 2014) and showed a good agreement (correlation coefficient ∼0.90) with standard techniques (Nakayama et al., 2018). Sharp gradients in the terrain height highlight the complexity of the central Himalaya and foothills region, which are also suggested to affect the wind flows and mixing of ozone and aerosols (e.g., Ojha et al., 2019; T. Sarangi et al., 2014; N. Singh et al., 2016; Solanki et al., 2019).

Along with the PM_2.5, micrometeorological measurements available from Manora Peak, carried out using the Ultrasonic anemometer (METEK, Germany, Model: USA-1 scientific), are utilized to understand the impact of aerosols on surface layer fluxes. Diurnal evolution of surface layer characteristics is generally expressed using sensible heat flux (SHF), vertical momentum flux (τ), turbulent kinetic energy (e), and associated wind components. These variables have been estimated using the following equations based on the eddy covariance method (e.g., Dyer, 1961; Solanki et al., 2016): [Image Omitted. See PDF] [Image Omitted. See PDF] [Image Omitted. See PDF]

Here $u\prime$, $v\prime$, $w\prime$, and $T\prime$ are deviation from corresponding 30-min mean values of wind components and temperature, respectively, which are known as turbulent fluctuations. ρ and C_p represent the atmospheric density and specific heat capacity at constant pressure. The preprocessing of the sonic anemometer data set, such as sector planar-fit methodology has been applied and discussed elsewhere (Solanki et al., 2016, 2019).

The spread of dust based on the true color composite image for June 14, 2018 is shown in Figure 1b. The dust storm is seen to be originated over 20–25°N 45–60°E. The dust is spread horizontally over ∼70–90°E and ∼25–30°N, which could evolve with time. Hence, the simultaneous measurements of aerosols characteristics are utilized from a foothill site Lumbini (27.49°N, 83.28°E) and IGP site Gandhi College (25.871°N, 84.128°E). The AErosols Robotic NETwork (AERONET) level 2.0 retrievals for Lumbini and the available level-1.5 cloud screened retrievals for the Gandhi College are used here. Details of the AERONET instrumentation and uncertainties can be found elsewhere (e.g., Dubovik et al., 2000; Dubovik & King, 2000; Holben et al., 2001).

Further, the radiosonde observations available from Gorakhpur, which is at a distance of ∼120 km in south of Lumbini and ∼150 km in the north-west of Gandhi College (Figure 1a), are utilized for analysis of the vertical profile of meteorological parameters and stability indices such as Total-Totals index (TTI), K-index (KI), Showalter index (SI), lifted index (LI), convective inhibition energy (CINE), and convective available potential energy (CAPE). The sounding data were available twice a day from the University of Wyoming (http://weather.uwyo.edu/upperair/sounding.html) at 5:30 and 17:30 Indian Standard Time (IST). IST is 5:30 ahead of the coordinated universal time (UTC). Further details of different instability indices can be found elsewhere (Bluestein, 1993; DeRuberts, 2006; Galway, 1956; Miller, 1972; Showalter, 1953).

Satellite-Based Measurements

The Cloud-Aerosol Lidar and Pathfinder Satellite Observations (CALIPSO) Level-2 profiles of aerosol and subtypes including dust (Powell et al., 2013) have been analyzed, and the evaluation of these datasets over this region is performed recently (Kumar et al., 2018). The Moderate Resolution Imaging Spectroradiometer (MODIS) dark blue target data on aerosol optical depth (AOD) level-2MYD04_L2 (Levy & Hsu, 2015), and Terra based MOD08_D3 (C6.1) AOD at 550 nm combined dark target (DT) and deep blue (DB) (Platnick et al., 2017) at a resolution of 1° × 1° are also used.

Reanalysis Products

Meteorological datasets of mean sea level pressure (MSLP), geo-potential height, zonal and meridional winds are utilized from ERA-Interim reanalysis, available from the European Center for Medium-Range Weather Forecasts (ECMWF) at the spatial resolution of 0.75° × 0.75° and temporal resolution of 6 h (Dee et al., 2011). In addition, the Modern-Era Retrospective Analysis for Research and Applications (MERRA-2) global reanalysis product is based on the Goddard Earth Observing System version-5 (GEOS-5) model, which has assimilation of space-based observations of aerosols, meteorological parameters, and clouds (Gelaro et al., 2017; Molod et al., 2015; Randles et al., 2017) is also utilized to estimate the dust budget. This basically uses the columnar dust mass density and transported mass flux available at a spatial resolution of 0.5° × 0.625° daily. The projected dust climatology carried out by Colarco et al. (2014) uses the GEOS-5 general circulation model output from 2011 to 2050 and estimated the global mean dust emission annually, ranging from 1,900 to 2,400 Tg yr⁻¹ with annual mean dust aerosol burden in a range of 20.8–23.1 Tg. MERRA-2 is shown to provide better estimates of the dust budget, including emissions, transport, and deposition (Buchard et al., 2017; Jing et al., 2017). The total mass of the dust (D_L) is calculated using the following relation. [Image Omitted. See PDF]where, S is the area of study region that encompasses north India (70–90°E and 20–35°N) and ρ_c the corresponding columnar dust mass density. The total dust transport (DT_tot), including the emission and deposition, is estimated using the following equation (Han et al., 2016; Jing et al., 2017): [Image Omitted. See PDF]where total dust transport (DT_tot) calculated using the mean dust transport fluxes (f_mean) is obtained from MERRA-2 for the given period (T). PM_2.5 and PM₁₀ data at the Manora Peak have also been used from the Copernicus Atmosphere Monitoring Service (CAMS) model reanalysis (Inness et al., 2019).

Results and Discussion Genesis, Evolution, and Dynamics of the Dust Storm

The first indicators of the Dust Storm (DS) at the mountain site were poor visibility and significant enhancement in the PM_2.5 concentrations. The evolution of the storm is investigated considering three distinct periods: pre-DS (June 8–12, 2018); DS (June 13–17, 2018); and post-DS (June 18–22, 2018). The variation in PM_2.5 during these days is given in Figure 2, where it is observed that PM_2.5 has peaked and prevailed over the site during the DS period. Hourly PM_2.5 levels are as high as 261.5 μg m⁻³ with the mean value during the DS period as 112.3 ± 69.9 μg m⁻³. This is an enhancement by a factor of 5, when compared to the pre-DS (24.3 ± 6.2 μg m⁻³). Notably, the mean value during DS is much higher than the annual mean 22 μg m⁻³. In the post-DS, PM_2.5 shows a decline to 31.4 ± 6.1 μg m⁻³; nevertheless, levels remain still higher than that during the pre-DS by ∼30%, indicating the partial settling of dust. The PM_2.5 measurements were found to be correlated well with the results of CAMS model reanalysis data (r = ∼0.8). Analysis of CAMS data revealed that PM₁₀ levels were enhanced substantially during the DS period, besides PM_2.5 (Figure S1). MODIS observations show higher columnar AOD₅₅₀ values (up to 1.8), associated with the simultaneously CALIPSO observed dust, in the vertical column distributed vertically up to an altitude of about 7 km during the pre-DS period over the Arabian Sea and Peninsular region (Figures 3a–3c). However, relatively lower AOD (0.2–1.2) on June 11 were observed over the IGP and the surrounding region, which increased to ≥2.0 on June 13 (Figure 3d) and June 15 (Figure 3g), and is attributed to the dust and the polluted dust (Figures 3f and 3i). This suggests that while passing over the densely populated and urbanized part of the western Indian region, the dust storm is exposed to anthropogenic emissions as well, and that contributes to the polluted dust as observed.

Figure 2. Variations in hourly PM2.5 and Moderate Resolution Imaging Spectroradiometer (MODIS) retrieved daily Aerosol Optical Depth (AOD)550 over Manora Peak in the central Himalaya. Shaded region marks peak dust impact period.

Figure 3. (a) Daily Aerosol Optical Depth (AOD)550 obtained from Moderate Resolution Imaging Spectroradiometer (MODIS)-Terra, (b) the vertical feature mask from Cloud-Aerosol Lidar and Pathfinder Satellite Observations (CALIPSO) (0-invalid, 1-clear air, 2-cloud, 3-aerosol, 4-stratospheric layer, 5-surface, 6-subsurface, 7-totally attenuated), and (c) the aerosols subtype from CALIPSO (0-not applicable, 1-clean marine, 2-dust, 3-polluted continental, 4-clean continental, 5-polluted dust, 6-smoke) for June 11 (in pre-Dust Storm (DS); first row), June 13 (in DS; second row: d–f) and June 15, 2018 (in DS; third row: g–i). The white line in (e, f, h, and i) shows the grid point of the CALIPSO pass closest to Manora Peak.

Figure 4 shows the distribution of column dust mass loading (ρ_c) from the MERRA-2 model and ERA-Interim wind fields over a large region covering northern India, the Arabic Peninsula, and the Thar Desert. Mean vertical velocity over the box shown in Figures 4b and 4d is found to be ascending (−0.03 Pas⁻¹) up to 750 hPa, indicating that the dust can be lifted up to ∼3.5 km altitude. This uplifting of dust is primarily maintained by strong convective flows observed over dust prone regions such as the Arabian Desert and north-eastern African coasts (Figures 4a and 4b) during June 9–12, 2018. This vertically transported dust is subsequently advected by the westerlies (about 10–15 m s⁻¹) toward northwest India with enroute influx from the Thar Desert. The dust mass density over IGP is additionally enhanced by obstruction to the synoptic-scale flow by the high-altitude mountain of the Himalaya. The geopotential, which is the measure of the approximate height of the isobaric surface from mean sea level, at the pressure level of 850 hPa, is shown in Figure 5a. A low geopotential region (trough) formed mainly over the land masses of Thar Desert, Arabian Peninsula and extended up to the northern Arabian sea that led to the convergence of air masses, which then propagated northward, spreading out over the IGP and Himalayan foothills. The low-pressure region (Figure 5b) having south to the north gradient of ∼25 hPa on June 13 drifts the westerly winds northward over the IGP and the Himalaya. Such airflow in front of the trough can bring dust to the IGP (T. Wang et al., 2021). The steepness on the gradient of MSLP is seen to decrease until June 16.

Figure 4. Distribution of the dust column mass density (ρc) over middle east and south Asian region, including northern India, the Arabic Peninsula, and the Thar Desert, based on the MERRA-2 reanalysis product. The mean wind from the Era-Interim at 850 hPa is also shown. (a–c) show the features for three individual days: June 9, 13, and 18, 2018, whereas the same for pre-Dust Storm (DS), DS, and post-DS periods is shown in (d–f). The mean vertical wind (ω, Pas−1) in the region covered by magenta-color box, over the Arabian sea and peninsula region, is shown to understand the vertical lift and subsequent transport of dust with westerly winds.

Figure 5. (a) Variation of the geopotential height at 850 hPa and (b) Mean sea level pressure during June 11–16, 2018 over the Indian subcontinent, Arabic Peninsula, and the Thar Desert from the ERA-Interim reanalysis.

As the low-pressure strengthens over the dust storm regions in the north-western part, the westerly winds drift northward over IGP, increasing ρ_c to as high as 2.5 gm⁻² during the DS. The ρ_c was less than 0.8 gm⁻² in IGP and adjacent Himalayan foothills, whereas it is estimated to be about 1.4 gm⁻² over the Thar region during pre-DS with the lower wind as compared to the DS. The strength of the low-pressure system over the IGP region is observed to decrease during post-DS along with the weakening of cyclonic flow and low pressure-induced change in wind direction from south-westerly to westerly (Figures 4c and 4d). The columnar dust loading, ρ_c values decrease up to 1.0 gm⁻² over the IGP and adjacent regions. The residual effects of dust transport are observed over IGP in the post-DS period, that is, ρ_c is ∼1.0 gm⁻², which was ≤0.8 gm⁻² in pre-DS. Summarily, the convective flows over the dust source regions and advection through westerlies in the presence of a low-pressure system over the northern Indian subcontinent resulted in favorable conditions for dust transport.

The stability indices such as convective available potential energy (CAPE) and convective Inhibition energy (CINE) have also been analyzed, as shown in Figure 6. These indices infer that the atmosphere can be more unstable and lead to severe thunderstorm provided sufficient moisture is available. The negative (positive) value of CINE (CAPE) represents the convection, and the value of CAPE higher than 2,500 Jkg⁻¹ can be considered to represent highly unstable and intense convective atmosphere (Bluestein, 1993). CAPE is seen to increase drastically from 75 to 2,283 Jkg⁻¹, from June 13 to 17, 2018 at 00:00 UTC, while CINE increased from −340 to −150 Jkg⁻¹. The CAPE and CINE are correspondingly high and low, respectively, even after the post-DS days, which depicted the unstable atmosphere. The TTI increases from 45.8 on June 12 to a maximum of 57.3 on June 16, 2018 at 12:00 UTC, which might have induced to the significant heating of the atmosphere from 850 hPa of the 500 hPa. Similarly, KI increases from 38 and remains higher than 40 during the DS period, with a maximum value of 47.7 on June 16. The SI and LI show decreasing trends during the DS period from −4.3 to −8.6 and from −1.0 to −10.5, respectively. Summarily, the aforementioned indices show that on June 13, 2018, on the arrival of the dust storm that is, the day of maximum aerosol loading, the atmospheric column was relatively stable compared to the previous days and the subsequent days when the dust starts settling over this region.

Figure 6. Variation in the stability indices: (a) convective inhibition energy (CINE) and convective available potential energy (CAPE), (b) Total-Totals index (TTI) and K-index (KI), and (c) Showalter index (SI) and lifted index (LI) for 00:00 UTC (left panel) and 12:00 UTC (right panel). The shaded area marks the Dust Storm (DS) period.

Columnar Meteorological Conditions

Atmospheric stability, a measure of the tendencies for vertical motions, is described by the vertical profiles of meteorological parameters such as temperature, water vapor, and winds. Here, the impacts of the dust storm on atmospheric stability and meteorological parameters are investigated by analyzing radiosonde observations (Figure 7). While the mean temperature between 900 and 800 hPa is observed to be ∼30°C in pre-DS days and 19.4°C–26.2°C during DS days indicating that the cooling near the surface could be due to the presence of dust (Saidou Chaibou et al., 2020) and other dynamical effects such as cold air advection (T. Wang et al., 2021). The distribution of the temperature field over the region is shown in Figure S2 depicts the temperature variation. The lower temperature at 700–500 hPa at 0:00 UTC, during DS days (−1.8°C–1.8°C) compared to pre-DS (6.2°C–7.8°C), indicates that the presence of the dust leads to the effect of more substantial attenuation of incoming solar radiation and absorption of the long-wave terrestrial radiation. Additionally, the water vapor mixing ratio exhibited a sharp decline at 700–500 hPa during DS (mean = 3.8 ± 1.0 g Kg⁻¹) than during pre-DS (8–9 g Kg⁻¹). Wind profiles show that wind speeds during pre-DS were about 5 ms⁻¹ up to 500 hPa (∼5.5 km amsl), and strong westerly winds (up to 20 ms⁻¹) are seen to have supported the transport of dust as reflected in synoptic-scale flow presented above. Further, the observations of aerosols from two AERONET sites are utilized to examine the aerosol characteristics and radiative impact in the following section.

Figure 7. Vertical profiles of temperature, water vapor mixing ratio, wind speed, and wind direction (a–d) for 00:00 UTC (5:30 IST) and (e–h) 12:00 UTC (17:30 IST) based on radiosonde observations from Gorakhpur, station in the Indo-Gangetic Plain (IGP) region.

Impact Analysis Aerosol Characteristics

Figure 8 shows the comparison of columnar aerosol characteristics such as AOD, Absorption AOD (AAOD), Single scattering albedo (SSA), and volume size distribution observed during the pre-DS, DS, and post-DS periods at Gandhi College and Lumbini stations. Mean AOD varied as 0.65 (0.41), 0.94 (1.5), and 0.72 (0.72) during pre-DS, DS, and post-DS periods respectively, as observed at Gandhi College (Lumbini). AOD is observed to be enhanced more strongly at Lumbini, located in the foothills of the central Himalaya. AOD and the AAOD values are observed to be the highest during the DS period at both stations, which indicates the enhanced dust loading over this region. AOD values are also higher during the post-DS period showing the residual effects of the dust storm due to the settling of the dust in the foothills and adjacent IGP region. Dust usually has a residence time of 5–6 days in the atmosphere.

Figure 8. Aerosol characteristics at Gandhi College (a–d) and Lumbini (e–h) based on AERONET observations. The value of angstrom exponent (AE) and absorption angstrom exponent (AAE) is quoted in Aerosol Optical Depth (AOD) (a, e) and Absorption AOD (AAOD) (d, h) plots.

The mean values of the angstrom exponent (α) were 1.17 (1.23), 0.28 (0.33), and 0.55 (0.60) during the pre-DS, DS, and post-DS periods respectively, at Gandhi College (Lumbini). Higher AOD and lower α values suggest the dominance of coarse-mode aerosols (dust) during the DS period. This also reflects in the bimodal volume size distribution (VSD) with the dominance of the coarse mode particles and relatively lesser abundance of smaller particles (Figures 8c and 8g). The peak in coarse mode particles is attributed to the dust-dominated environment (Dey et al., 2004; R. P. Singh et al., 2004; Srivastava et al., 2014). Observation sites are away from the dust sources; nevertheless, the large impact of the long-range transported dust is over the IGP and the Himalayan foothills, which is unraveled in the aerosol characteristics.

Further, enhancements in AOD and coarse mode particles are more substantial at Lumbini than at Gandhi College. This further highlights more impact of the dust storm over the Himalayan foothills than low altitude Gangetic basin. The values of SSA at 440 nm are found to be higher in the pre-DS period as compared to that during DS and post-DS (Figures 8b and 8f). SSA shows an increase with wavelength during the DS and post-DS days, indicating the dominance of coarse mode particles and stronger absorption of the radiation. AAOD and absorption angstrom exponent (AAE) show larger enhancements at Lumbini than those at Gandhi College. Higher AAE values (2.17 and 2.23) also reflect considerable dust loading (e.g., Bergstrom et al., 2007). AAE values in this region have been reported to be lower and impacted by anthropogenic emissions in earlier studies (e.g., Srivastava et al., 2011). AAE values observed during DS period over Gandhi College and Lumbini are about two times of that reported earlier over this region (e.g., Srivastava et al., 2011) which show much-intensified storms. Hence, aerosol characteristics are observed to be significantly affected by the dust storm and clearly show the predominance of coarse dust particles over IGP and the Himalayan region.

Additionally, the radiative impact of the dust storm is investigated, as shown in Figure S3 and the methodology of the calculation is described in the Supporting Information S1. The radiative flux at the top of the atmosphere (ΔF_TOA) increases by ∼34% and at the bottom of the atmosphere (ΔF_SUR) by ∼37% over Gandhi College, whereas, these values turned to be thrice for Lumbini in DS as compared to the pre-DS. The higher values of ΔF_TOA and ΔF_SUR (∼6%, 16% for Gandhi College and 84%, 50% for Lumbini) are also observed in post-DS, compared to the pre-DS period. The intense heating of the atmosphere leads to a rise in the heating rate (∼0.21 K day⁻¹: Gandhi College and 0.72 K day⁻¹: Lumbini) in DS as compared to pre-DS and remains about ∼0.16 K day⁻¹ in post-DS. Hence, a significant cooling is found at the top and bottom of the atmosphere, whereas heating of the atmosphere is evident from the calculations, more pronounced during the DS period. Therefore, the prominent dust storm can influence the energy budget over Himalaya, besides modifying the energy budget of snow surfaces through deposition. Potential reductions in the snow albedo could increase the snow melting rates in the region (Aoki et al., 2011; Gautam et al., 2013; Lau & Kim, 2010).

Estimation of Dust Budget

The transported dust loading (DT_tot) is calculated to quantify the contribution of on the way mixing of pollutants with the active storm. Dust transport was channeled through strong westerly/northwesterly winds up to 3 km. Relative contributions from the long-range transport and regional source (Thar Desert) in the dust budget are analyzed by considering the region of 70–90°E and 20–35°N shown in Figure 9. The Thar Desert is in the zonal coverage of 70–75°E, as shown by gray shade in Figure 9a while the dust mass fluxes entering from the west are estimated over 65–70°E, that is, the contributions from long-range transport and fluxes crossing 75°E are considered as the total fluxes of dust which includes the feedback from the Thar desert region. Day-to-day variation in total dust mass (D_L) over this region (Figure 9b) shows that the daily mean varies from 2.2 to 2.7 Tg, with the mean value of 2.5 Tg, during the entire DS period, about 90% enhancement as compared to the pre-DS days. D_L is observed to be less than 1.5 Tg over this region, with the mean value of 1.3 Tg, in pre-DS and post-DS. Enhancement in dust loading over the northern Indian subcontinent during the DS is shown to be linked with high dust loading over coastal regions of northern Africa and the Arabian Peninsula. The encounter of this flow with the Thar Desert adds up further regional dust into the air mass toward the northern India.

Figure 9. (a) Estimation of dust budget before during and after the dust episode (June 13–17, 2018), (b) Variation in transported dust (DTtot) and daily total dust mass (DL) loading over 70–90°E and 20–35°N during June 8–22, 2018 derived from the MERRA-2 reanalysis.

The quantitative analysis on long versus short-range transport of dust to IGP is carried out using mass fluxes from MERRA-2 data products. Total dust transported (DT_tot) to the IGP and downwind region is estimated to be ∼4.8 Tg during the DS period, out of which 2.3 Tg has been transported from the west as the long-range transport, and the remaining part is estimated to be contribution through the Thar desert. Dust emissions from the Thar further enriches the dust air, while mass passing through this region. Some uncertainties could also be involved in the emission of the dust and daily dust mass, as Jing et al. (2017) described. The subsequent section deals with the impact of dust on surface-layer characteristics are studied using micrometeorological measurements using Sonic Anemometer from Manora Peak in the Himalaya.

Impact on Surface Layer Characteristics

The surface layer is the lowest part of the troposphere, where the atmosphere interacts with the land surface. Surface layer characteristics strongly affect boundary layer evolution, vertical transport of moisture and pollution. In this subsection, the effects of the transported dust on surface layer characteristics in the Himalaya (Manora Peak) are investigated. Diurnal features of the wind (speed and direction), wind components (zonal, meridional, and vertical), and surface layer temperature are shown in Figures 10a–10f during the pre-DS, DS, and post-DS periods. In addition, the key factors responsible for the vertical mixing, such as turbulent kinetic energy, sensible heat fluxes, and momentum fluxes are analyzed (Figures 10g–10i).

Figure 10. Variations in wind components (a–c), wind direction (d) and speed, (e), and the temperature (f) observed using Sonic anemometer in the central Himalaya (Manora Peak) along with (g) sensible heat flux (SHF), (h) turbulent kinetic energy (e), and (i) momentum flux (τ) during June 8–12 (red), June 13–17(black) and June 18–22, 2018 (blue). The positive values of u components of wind show the eastward flow direction, and the negative value of v-component show the northerly wind direction. The local hours are representing the Indian standard time (IST).

The magnitudes of u and v components show larger values (∼3 ms⁻¹) in the daytime (09:00–19:00 IST) in pre-DS and post-DS. Slightly higher magnitude of u than v could be the effect of the mountain-valley wind circulation (e.g., Solanki et al., 2016). The mean zonal velocity (u) and meridional velocity (v) are observed to be 3.1 ms⁻¹ (range: 2.1–4.1 ms⁻¹) and −2.7 ms⁻¹ (range: −1.4–−2.7 ms⁻¹) with the higher magnitude during night (Figures 10b and 10c). Mean wind is mostly northwesterly in agreement with synoptic-scale pattern (Figure 4), which is a dominant wind flow at this site (Solanki et al., 2019). Wind speed shows significant diurnal variation during pre-DS and post-DS with the maxima in the daytime. Whereas during DS wind speed was minimum (2.6 ± 1.1 ms⁻¹) in the noon and maximum (5.6 ± 1.1 ms⁻¹) at night. Similar results were also reported for near-surface layer over Beijing during heavy pollution event (H. Wang et al., 2018). Since the observational site is surrounded by dense forest with minimal local dust sources, it is suggested that transport of dust can be more efficient with stronger winds during night time as the boundary layer is lower and non-turbulent than daytime. A slight lowering of daytime temperature (Figure 10f) is attributed to the attenuation of solar radiation in the daytime due to dust. In the nighttime, a higher temperature is due to more absorption of outgoing radiation.

Sensible heat flux is the measure of covariance of the vertical component of wind and temperature, which shows a diurnal variation similar to the temperature. Normally, the SHF starts increasing at about 07:00 IST attaining maximum value (179 Wm⁻²) during the noontime (12:00–15:00 IST). Whereas, SHF attains positive values with a delay of 3 h during the DS (Figure 10g). The SHF is observed to be lower during DS (−34.0 Wm⁻²) as compared to the pre-DS (34.8 Wm⁻²) and post-DS (60.7 Wm⁻²). The negative values of SHF (−10 to −270 Wm⁻²) show strong downward transport of heat during the night which is observed comparatively weaker in pre-DS (−10 to −62 Wm⁻²) and post-DS (−10 to −65 Wm⁻²). In contrast with pre-DS and post-DS periods, net diurnal mean SHF of −34 Wm⁻² shows net downward SHF from the atmosphere to surface. Significant diurnal variations are seen in TKE and momentum fluxes (Figures 10h and 10i) with higher values during daytime hours and lower in the afternoon hours. TKE is observed to be significantly enhanced during DS (19.7 m² s⁻²) as compared to pre-DS and post-DS (7 m² s⁻² and less). Higher TKE at night during DS is linked with stronger winds and corresponding high wind shear, leading to the more turbulent surface layer. The turbulent surface layer could lead to enhanced vertical momentum fluxes during the night resulting in the observed net cooling of ∼34 Wm⁻² of surface layer during the DS period.

Summary and Conclusions

The genesis, evolution, and impacts of a severe dust storm influencing the Indian region have been investigated using ground- and satellite-based observations and model results. Key findings of the study are given as follows:

1. Dust storm significantly enhanced the PM_2.5, PM₁₀, and AOD over the Himalayan region. In comparison to the annual mean and the pre-dust storm conditions, the PM_2.5 levels were higher by about a factor of 5, affecting air quality adversely.

2. Dust was uplifted from the Arabian Desert, and north-African eastern coasts, and then additional influx was from the Thar Desert during advection through westerlies. A low-pressure system over northern India favored the dust transport, and the synoptic-scale flow was subsequently obstructed by the Himalaya. Effects of anthropogenic emissions over the Indian subcontinent further explain the satellite-observed prevalence of polluted dust aerosols.

3. The analysis of the stability indices shows unstable atmosphere during the dust storm and residual effects persisting also during the post-DS.

4. The additional emission from the Thar Desert (2.5 Tg) worked as an on-the-way dust feeding system into the long-range transport from the Arabian Desert and north-African eastern coasts.

5. Dust storm resulted in cooling at the surface and top of the atmosphere but a significant heating of atmosphere (∼0.21 K day⁻¹: Gandhi College and 0.72 K day⁻¹: Lumbini) in DS as compared to pre-DS.

The analyses clearly highlight the importance of the long-range transport and regional dust sources in impacting the central Himalayan region. It has been suggested that dust has stronger role in the snow darkening and melt over high-altitude Asian mountains (C. Sarangi et al., 2020). Further studies are needed to identify any possible trends in dust transport and regional dust sources that could have a remarkable impact on the Himalayan air and climate.

Acknowledgments

This study is supported by ABLN & C: NOBLE project under ISRO-GBP. The authors thank to Director ARIES, Nainital, for valuable support. N. Ojha acknowledges the use of the computing resources (Vikram) at PRL Ahmedabad and valuable support from D. Pallamraju and Anil Bhardwaj. Constructive comments and suggestions from anonymous reviewers are acknowledged.

Conflict of Interest

The authors declare no conflict of interest.

Data Availability Statement

The subdaily wind, geo-potential height, and mean sea level pressure datasets are extracted from ECMWF (https://apps.ecmwf.int/datasets/data/interim-full-daily, last accessed: January 2021) at the spatial resolution of 0.75° × 0.75° and temporal resolution of 6 h (Dee et al., 2011) for the study region. The dust columnar mass density and fluxes are downloaded from MERRA-2 (https://disc.gsfc.nasa.gov/datasets/, last accessed: January 2021) at the spatial resolution of 0.5° × 0.625° (GMAO, 2015). The AOD from MODIS data is downloaded from https://ladsweb.modaps.eosdis.nasa.gov/search/order/2/MOD08_D3--61 (Platnick et al., 2017, last accessed: January, 2021). The aerosols feature flags and subclass level 2 datasets are downloaded from the CALIPSO archival (https://asdc.larc.nasa.gov/project/CALIPSO/CAL_LID_L2_05kmAPro-Standard-V4-20_V4-20 and https://asdc.larc.nasa.gov/project/CALIPSO/CAL_LID_L2_VFM-ValStage1-V3-40_V3-40; last accessed: January 2021; Powell et al., 2013). The surface-based aerosol measurements are used from the AERONET webpage (https://aeronet.gsfc.nasa.gov/, last accessed: January 2021). The radiosonde data over Gorakhpur station (station ID: 42379) is utilized from the Wyoming upper air sounding archival (http://weather.uwyo.edu/upperair/sounding.html; last accessed: January 2021). The PM_2.5 and PM₁₀ data over Manora Peak are utilized from the CAMS Global Reanalysis (https://ads.atmosphere.copernicus.eu/cdsapp#!/dataset/cams-global-reanalysis-eac4; last accessed: June 2021; Inness et al., 2019).