Characteristics, Dynamics, and Impact of the Thar Desert Dust Storms on Air Quality over Northern India

doi:10.21203/rs.3.rs-3879277/v1

Download PDF

Research Article

Characteristics, Dynamics, and Impact of the Thar Desert Dust Storms on Air Quality over Northern India

https://doi.org/10.21203/rs.3.rs-3879277/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

In the present study, a pre-monsoon dust storm event (14 May 2018 ± 2 days), originating from Thar Desert (Rajasthan state), spread to north Indian states- Haryana, Delhi, and Uttar Pradesh is selected. Satellite observations, climate model reanalysis datasets, and ground monitoring station and meteorological parameters were used to study the dust storm event and its impact on air quality of the northern region of India against a non-dust storm background reference (viz., 25 May 2018 ± 1 day). We include dust column mass density (DCMD total and DCMD PM_2.5), and find 3-5 times higher air column dust loading (maximum 3.9 g/m²) during the event, against dust levels (~0.2-0.6 g/m²) during the background non-dust storm period. NOAA HYSPLIT 72-hour forward trajectories show air-mass transport from the Thar Desert towards IGP. Aerosol Optical Depth (AOD)> 2 in the affected region indicates a high residence time of dust aerosols. DCMD is used to show the 3-hourly progression of the dust storm (initiation, spread, and gradual dissipation). Moreover, hourly dust loading in the air (DCMD) and particulate matter (PM₁₀ and PM_2.5) ground-based observations over Jodhpur (Rajasthan), Rohtak (Haryana), Delhi (NCT of Delhi), and Lucknow (Uttar Pradesh) cities, show severe worsening of the already poor air quality of the region. Lastly, local spatial correlations are plotted to understand the spatial associations, variabilities, and local feedback between the variables- DCMD, AOD, 2m air temperature, precipitation, and wind speed.

Atmospheric Sciences

Dust storm

Northern India

Dust column mass density (DCMD)

Particulate Matter

Air Quality

NOAA HYSPLIT trajectories

Dust storms load the overlying troposphere with excessive quantities of aerosols, both fine and coarse, and can influence climate feedback mechanism at various spatiotemporal scales (Kumar et al., 2014; Kim et al., 2017; Chen at al., 2018; Jain, 2022a). High dust (aerosol) loading absorbs greater solar radiation along with affecting the radiative budget, cloud microphysics, cloud optical properties, precipitation patterns, surface level visibility, biogeochemistry of water bodies and even glacial characteristics (Zhuang et al., 2001; Toon, 2003; Seinfeld et al., 2004; Jickells et al., 2005; Slingo et al., 2006; Min et al., 2009; Di et al, 2016; Kedia et al., 2018; Sonwani and Kulshrestha, 2019). Radiative cooling or heating is dependent on the size and amount of the atmospheric dust, its chemical composition, and the atmospheric layers (Balkansi et al., 2007; Shen et al., 2013; Kok et al., 2017; Kedia et al., 2018; Tiwari et al., 2019). While aerosols such as mineral dust particles and black carbon lead to a warming effect, other aerosols such as sea salts and sulfates cause a cooling effect in the atmosphere (Ramachandran et al., 2012). Typically, anthropogenic aerosols are smaller in size as compared to natural aerosols and tend to have a warming effect; however quantifying individual contributions from aerosol optical depth (AOD) column datasets is highly challenging and full of uncertainty (Allen and Sherwood, 2011). Apart from a diverse range of climate and environmental effects, many adverse effects of aerosols on human health and socio-economic losses are also reported. Significantly high level of particulate matter (PM₁₀and PM_2.5) in the atmosphere increases the probability of traffic congestion and road accidents due to poor visibility (Wan et al., 2020). Severe respiratory and pulmonary illnesses such as asthma, chronic obstructive pulmonary disease, pneumonia, and rhinitis are the most reported health-related issues caused by high loading of atmospheric PMs, hence responsible for increasing hospitalizations and mortality rates (Wiggs et al., 2003; Prasad and Singh, 2007; Khaniabadi et al., 2017; Soleimani et al., 2020; Sonwani et al., 2022). Such harmful impacts are the point of concern in developing countries like India, which have a high population density, a high background level of air pollution, and a low per capita income.

Most trans-continental dust storms are sourced from major desert belts and permanent arid areas of the world (Jain, 2022a; Maji and Sonwani, 2022). The Sahara Desert in Africa, the Arabian Desert in the Middle East and the Taklimakan Desert in Central Asia, collectively account for more than 95% of global natural dust emissions (Chen et al., 2018). Although smaller deserts, semi-arid regions, agricultural areas, and bare lands contribute to dust aerosols lesser than the major dust sources, however, their importance in terms of air quality (local and regional) and human health remains much greater (Kim et at., 2017). This is due to their close proximity to population clusters. One such region that is immensely affected by dust storms, is the Indo-Gangetic Plains (IGP) situated in the northern Indian peninsula. The region’s significance lies in the fact that it has one of the largest and densest populations in the entire world (Lodhi et al., 2013, Pandey et al., 2020). In India, the IGP experiences the highest levels of AOD and PM concentration all around the year, but air quality severely deteriorates during the dusty days of the pre-monsoon season (David et al., 2018). Aerosol pollution has decreased human life expectancy in northern India (including IGP) by approximately seven years (Greenstone and Fan, 2018).In northern India, the Thar Desert is one of the prominent sources of dust aerosols (Kedia et al., 2018). In north-western India, the Thar desert is in the state of Rajasthan but extends into Pakistan too. Lower tropospheric winds largely flowing from the west or the southwest direction facilitate the transport of mineral and fine dust particles from the Thar Desert to northern India (Sikka, 1997). Depending on the wind speed and direction, dust aerosols are at times able to reach as far as West Bengal and Central India (Sarkar et al., 2019). Another source of dust aerosols in India is the long transboundary transport of dust originating in the Arabian Peninsula, flown across the Arabian Sea (Solmon et al., 2015).

Over northern India and the extended IGP region, dust storm occurrence is found to be the highest during the pre-monsoon season. Pre-monsoon months of March, April, and May show 1) higher warming of the continental landmass, 2) prevalence of strong pressure gradients before the advent of Indian Summer Monsoon (ISM) in late June, and 3) lower probability of precipitation (Sikka, 1997). Therefore, hot, and dry weather conditions in northern India, and a highly unstable adiabatic lapse rate in the Thar Desert, cause 1) the lifting of the dust particles (both fine and coarse) into the air, 2) an increase in local circulation and transboundary dust transport, and 3) a decrease in dust scavenging due to low precipitation (Sikka, 1997; Gautam et al., 2009; Kumar et al., 2014). As a result, ideal conditions for frequent dust storms are formed during pre-monsoon. Additionally, the magnitude and intensity of the dust storm is governed by the strength of the updrafts and the downdrafts, the creation of an intense low-pressure area, the existence of strong lofting winds, and strong advection (Seigel and van den Heever, 2012; Newman et al., 2022).It is noted that atmospheric dust loading in pre-monsoon greatly modulates the ISM variability and circulation patterns, especially during the early part of the monsoon (Pandey et al., 2017). The study further highlights that while over the last decade (2002-14) AOD has been increasing over northern India it seems to be declining during pre-monsoon by 10–20%. The AOD decline is more over the Thar Desert than over the eastern belt of the IGP (Pandey et al., 2017). Precipitation and wind circulation patterns over the dust source region are also important factors that govern dust storm occurrence. In the past two decades, a positive precipitation bias has been noted in the winter and pre-monsoon rainfall trends over the Thar Desert, implying an increase in wet scavenging of dust, and as a result fewer dust storms reach the IGP (Gautam et al., 2009; Pandey et al., 2017).

Ground-based observation network such as Aerosol Robotic NETwork (AERONET) has been widely used to gather aerosol properties relevant to dust storms (Dey et al., 2004; Prasad et al., 2007; Kumar et al., 2015). Besides such networks, remote sensing-based satellite observations e.g., Moderate Resolution Imaging Spectroradiometer (MODIS), reanalysis products e.g., Modern-Era Retrospective Analysis for Research and Applications version 2 (MERRA-2), and state-of-the-art climate models e.g., Weather Research and Forecasting Chemistry (WRF-Chem) are immensely useful in studying dust storm events. Using the WRF-Chem coupled model, Kumar et al. (2014) estimated that during a chosen pre-monsoon dust storm (17–22 April 2010) event, ~ 7.5 Tg total dust was emitted over the Thar Desert and northern India. The study further notes an over 50% increase in AOD during the dust storm event; and AOD > 2in some regions, indicative of dust lingering in the atmosphere in excessive amounts. In general, both satellite and ground-based observations show that during pre-monsoon dust storms, there is 1) a significant increase in AOD, 2) a significant decrease invisibility, and 3) a considerable effect on the radiation budget over the northern Indian region (Dey et al., 2004; Prasad and Singh, 2007; Gautam et al., 2009; Kedia et al., 2018; Tiwari et al., 2019). Forward and Backward trajectory analysis is the potential tool to identify long-range transport patterns with the use of archived meteorological data (Baker, 2010). The National Oceanic and Atmospheric Administration Hybrid Single-Particle Lagrangian Integrated Trajectory (NOAA HYSPLIT) models forward trajectory analysis is a widely used method for dust storm-related studies for the identification of dust storm starting point, source routing, and particle transport during and after dust storm events (Naimabadi et al. 2016).

In the present study, we make use of satellite observations, climate model reanalysis datasets, ground station air quality and meteorological data, and the NOAA HYSPLIT forward trajectories to understand the characteristics, dynamics, and progression of a pre-monsoon dust storm event in northern India. For this, a severe dust storm event that occurred over Northern India in the year 2018, during the pre-monsoon season is selected, and comprehensively analyzed against the non-dust storm period i.e., considered as a reference background. Aside from particulate matter (PM) data, we incorporate dust column mass density (DCMD) in our study, a parameter that has not yet been used in any dust storm study over the IGP or over northern India, and is found only in a handful of recent studies in the world (Jing et al., 2017; Gómez-Andújar et al., 2018; Sabetghadam et al., 2018; Hamidi, 2020; Yao et al., 2021).Due to many of the physical processes, atmospheric dynamics, and land-air interactions play an important role at local and regional scales. Thus, it is difficult to find significant linear statistical correlations, at each point, in the meteorological parameters associated with the dust storm event. As such, lags (either positive or negative) can exist with time as well as space. Thus, the present study understands the characteristics, dynamics, progressions, and impact of dust storm events on air quality parameters in the northern region of India during the pre-monsoon season.

Four northern Indian states- Rajasthan, Haryana, the National Capital Territory (NCT) of Delhi and Uttar Pradesh are considered in the present study. Based on existing literature, most dust storms in northern India, during the pre-monsoon, are found to originate from the Thar Desert in Rajasthan. These dust storms propagate north-eastward due to the prevailing wind direction in the region during this season. However, the Himalayas in the north act as a natural topographic barrier and force the dust-laden winds down towards the eastern IGP (i.e., Uttar Pradesh and sometimes to West Bengal). Based on this pre-monsoonal dominant wind pattern, we further selected four major cities- Jodhpur (in Rajasthan), Rohtak (in Haryana), Delhi (in NCT of Delhi), and Lucknow (in Uttar Pradesh), as a representative site for each state (Fig. 1). These sites were also preferred over other cities due to the availability of air quality and meteorological parameters data from the ground weather stations. The study region observes four different seasons in a year viz., winter (January to February; JF), pre-monsoon (March to May; MAM), monsoon (June toSeptember; JJAS), and post-monsoon (October toDecember; OND) (Krishnakumar et al., 2009). The focus remains on the pre-monsoon season in the present study. Extremely hot and dry weather conditions persist during pre-monsoon along with recurrent dust storm events that carry dust particles from the western Thar Desert over to northern India.

Table 1

Geographical location and population data (for the year 2011) for the representative cities of the four north Indian states considered in the present study.
City	State	Latitude	Longitude	Population
Jodhpur (UA)	Rajasthan	26.2389° N	73.0243° E	1,137,815
Rohtak (MCl)	Haryana	28.8955° N	76.6066° E	373,133
Delhi (UA)	NCT of Delhi	28.5863° N	77.2242° E	16,314,838
Lucknow (UA)	Uttar Pradesh	26.8669° N	80.9498° E	2,901,474
^{*Note− Urban agglomeration (UA) and Municipal Council (MCl) are types of urban clusters defined by the Census of India (2011) for each city. Latest city population data (Census of India, 2011) is available only till the year 2011 and is therefore considered for the study.}

Table 1 provides details on the geographical location and the population of the four considered cities. The area extent and the population statistics for each study site are extracted from the respective census district handbooks of each state (Census of India, 2011), and the climate data is sourced from India Meteorological Department (IMD) publications of climatological normal. Jodhpur district (area- 22850 sq. km), the largest in the state of Rajasthan, is characterized by sand dune formations in its western and north-western regions. Jodhpur UA (area- 78.60 sq. km) is a smaller urban agglomeration (UA) within this district and is an important north Indian city. The climate in Jodhpur is categorized as hot, dry, and arid, and the maximum temperature averages 37.7°C during MAM (IMD, 2011). Jodhpur receives significantly lower mean precipitation during MAM (7.8 mm) than the other three cities and shows a high inter annual ISM variability (IMD, 2011).

Both Rohtak and Delhi cities are part of the larger National Capital Region (NCR), but belong to different states. While Rohtak district occupies an area of 1745 sq. km, the urbanized Rohtak encompasses an area of only 72.18 sq. km. On the other hand, the NCT of Delhi has an administrative area of 1483 sq. km., and is interchangeably referred to as Delhi UA.A major difference between Rohtak and Delhi lies in their land use/ land cover patterns. In Rohtak, agriculture is still a dominant land use, while Delhi is almost entirely urban. As compared to Rohtak, the population and population density of Delhi, is also many times higher. The climate here is best described as hot and semi-arid, and most of the rainfall occurs between late June and August. The pre-monsoon average temperature is 35.5°C for Rohtak and 35.8°Cfor Delhi (IMD, 2015). The mean MAM precipitation is recorded as 19.9 mm and 20 mm for Rohtak and Delhi, respectively (IMD, 2015). The region also experiences highPM₁₀ and air pollution levels all around the year, due to crop residue burning (CRB) in Haryana, and vehicular emissions in Delhi, causing elevated levels of background AOD (Saxena et al., 2021). Lastly, Lucknow UA occupies an area of 348.80 sq. km, and falls within the larger Lucknow district (area- 2528 sq. km). Lucknow is a city of historical and cultural importance. For Lucknow, the average maximum temperature during MAM is 36.7°C, and the mean MAM precipitation is 12.9 mm (IMD, 2015).

3.1. Datasets

In the present study, we used multiple datasets (Table 2), including- satellite observations (e.g., AOD at 550µm, daily accumulated precipitation, etc.), climate model reanalysis products (e.g., dust column mass density (DCMD total and DCMD PM_2.5), 2m air temperature, etc.), and ground-based weather station observations (e.g., PM_2.5, PM₁₀, ambient temperature, wind speed, etc.). Satellite and climate model reanalysis gridded datasets used in the present study, have been extracted for the study region, from the NASA Giovanni web portal (https://giovanni.gsfc.nasa.gov/giovanni/). The MERRA-2data is a reanalysis product available from 1980 onwards and is based on the NASA Goddard Earth Observing System Model (GEOS-5) and the Atmospheric Data Assimilation System (ADAS). Variables such DCMD total and DCMD PM_2.5 are sourced from MERRA-2. The MODIS sensor on board the Terra sun-synchronous Earth observation satellite utilizes the deep-blue algorithm in order to estimate AOD (at 550 µm wavelength), a variable considered in the study. The Integrated Multi-satellite Retrievals for Global Precipitation Measurement (GPM-IMERG) is a unified algorithm, that provides a high-resolution precipitation product based on multiple satellites and is also used in the present study.

Table 2

Showing satellite observations, climate model reanalysis products, ground-based weather station observations and meteorological parameters.
S. No.	Variable	Source	Spatial Resolution	Temporal Resolution	Units
1.	Dust Column Mass Density (Total)	MERRA-2 Model (M2T1NXAER v5.12.4)	0.5° x 0.625°	Hourly	kg/m²
2.	Dust Column Mass Density (PM_2.5)	MERRA-2 Model (M2T1NXAER v5.12.4)	0.5° x 0.625°	Hourly	kg/m²
3.	Aerosol Optical Depth at 550 µm	MODIS Terra (MOD08_D3 v6.1)	1° x 1°	Daily	Unitless
4.	2m Air Temperature (Daily Mean)	MERRA-2 Model (M2SDNXSLV v5.12.4)	0.5° x 0.625°	Daily	°C
5.	Daily Accumulated Precipitation	GPM IMERG (GPM_3IMERGDF v06)	0.1° x 0.1°	Daily	Mm
6.	Surface Wind Speed	MERRA-2 Model (M2T1NXFLX v5.12.4)	0.5° x 0.625°	Hourly	m/s¹
7.	Particulate Matter (PM₁₀)	Central Pollution Control Board (CPCB)	Point data	Hourly	µg/m³
8.	Particulate Matter (PM_2.5)	Central Pollution Control Board (CPCB)	Point data	Hourly	µg/m³
9.	Ambient Temperature (AT)	India Meteorological Department (IMD)	Point data	Hourly	°C
10.	Relative Humidity (RH)	India Meteorological Department (IMD)	Point data	Hourly	%
11.	Wind Speed (WS)	India Meteorological Department (IMD)	Point data	Hourly	m/s
12.	Atmospheric Pressure (AP)	India Meteorological Department (IMD)	Point data	Hourly	mm Hg

Hourly concentrations of PM₁₀and PM_2.5 (air quality parameters; point dataset) for the four considered cities (Jodhpur, Rohtak, Delhi, and Lucknow) were downloaded from the data portal of Central Pollution Control Board (CPCB; http://cpcb.nic.in). Additionally, for these four cities, the hourly weather station (point) data for the meteorological parameters- ambient temperature (AT), relative humidity (RH), wind speed (WS) and atmospheric pressure (AP), was accessed from the Indian Meteorological Department(IMD)data portal (https://mausam.imd.gov.in).

Table 2. Details of the satellite-based observational climate model reanalysis and ground-based weather station datasets considered in the study. The CPCB and IMD data is for point locations (i.e., Jodhpur, Rohtak, Delhi, and Lucknow cities), while the gridded datasets are considered for the region encompassing the four north Indian states (i.e., Rajasthan, Haryana, NCT of Delhi, and Uttar Pradesh).

3.2. Dust storm event selection

Newspaper and IMD reports were referred to identify the major pre-monsoon dust storm events that occurred in the year 2018. In a preliminary analysis, the DCMD total for all MAM dust storm events of 2018 were plotted. From this analysis, the dust storm event of 14 May 2018, was found to be the most intense and widespread in the IGP region (northern India).The time duration from the onset of dust loading, to the occurrence of the dust storm event in the region, and finally the gradual dissipation of dust from the atmosphere were important factors in consideration of the total days that constituted the ‘dust storm event’.

From the MODIS satellite observations, and the MERRA-2 DCMD analysis of dust loading in the atmosphere, a time duration of ± 2 days from the maximum intensity of DCMD (total andPM_2.5) i.e., 14 May 2018, was selected as the ‘dust storm event’. Similarly, a continuous period of the minimum intensity of DCMD (days having zero or lowest background dust loading), within a span of 7–10 days from the occurrence of the selected dust storm event (i.e., 14 May 2018) was chosen. As no five continuous days, within this time period had considerably low or zero dust loading, a three-day period from 24–26 May 2018, having the lowest daily DCMD (total andPM_2.5; lowest observed for 25 May 2018)was selected as the background reference. The true color surface reflectance composite from the Aqua sensor onboard the MODIS satellite (https://worldview.earthdata.nasa.gov), for both 14 May 2018 (dust storm day; Fig. 2a) and 25 May 2018 (background; Fig. 2b) is shown.The MODIS imagery clearly shows increased dust loading over northern India on 14 May 2018, with its origin in the Thar Desert (circled in red), as well as the long-range trans-continental transport of Saharan dust(circled in black) over to the Arabian Sea and over western India. On the other hand, the background day shows no (or very little) dust loading from either of these sources.

3.3. Methodology

Daily average values of all the satellite observation and climate model reanalysis dataset variables(mentioned in Table 2, S. No. 1–6), for the time duration of the dust storm event and the background reference period, were downloaded over a bounding box covering entire northern India (latitude: 23°N to 33°N; longitude: 68°E to 88°E) in a gridded raster (Geotiff) format. All these gridded raster datasets were then clipped using the national boundary of India and thereafter plotted in ArcGIS ver. 10.6. Moreover, to map the spatiotemporal progression of the dust storm event (restricted to 14 May 2018), and to compare it against the background dust values (restricted to 25 May 2018), the hourly gridded data for the variable- DCMD total, was also downloaded over the northern Indian region. In addition, from this hourly DCMD total gridded data, the point location values (of DCMD total) for the four cites- Jodhpur, Rohtak, Delhi and Lucknow, was extracted in order to study the city-wise progression of the dust storm event. For these four cities, the ground station hourly distribution of particulate matter and meteorological parameters (mentioned in Table 2, S. No. 7–12),for the time duration of the dust storm event and the background reference period was also plotted. The calculated mean (± standard deviation) for the PM₁₀ and PM_2.5 levels in these four locations, during the dust storm event and the background period, was compared against the 24-h National Ambient Air Quality Standards (NAAQS).

Moreover, NOAA HYSPLIT forward trajectory analysis was performed over the Thar Desert; the region is considered the most active local source of dust and PM for the dust storms occurring in northern India. The NOAA HYSPLIT forward trajectory model was run for both- the dust storm event (14 May 2018) and the background period as a non-dust storm reference (25 May 2018), for a duration of 72 hours, in order to understand the different trajectories of air-mass movement from the dust source i.e., the Thar Desert. The air-mass trajectories were mapped at100m, 500m, and 1000m AGL.

Lastly, a local spatial correlation was computed for the gridded variables mentioned in Table 2 (S. No. 1–6), for the dust storm day (14 May 2018) and the background reference (25 May 2018) day. The days prior and after this was not considered for local spatial correlation analysis. Lag (positive or negative) can exist with time and space when analyzing meteorological parameters. Therefore, instead of simply correlating vis-à-vis singular pixels, the inclusion of locally weighted neighbourhood pixels reduces spatial lag error and makes the results more robust, especially for weather-based estimations (Rajagopalan and Lall, 1998). A maximum likelihood moving window-based non-stationary spatial correlation approach e.g., local spatial correlation is therefore more robust than linear statistical regression and correlation methods (Al-Ahmadi and Al-Ahmadi, 2013)and is therefore used in the present study. Prior to estimating the local spatial resolution between the variables mentioned in Table 2, all the raster datasets were resample to a common spatial resolution of 1°x1°. In order to compute the spatial correlation between any two variables, we defined a local window of size 3x3 pixels. Thus, a cluster of 9 pixels (or 3° x 3° spatial extent) is considered by the algorithm, each time, while calculating the correlation between any two rasters (e.g., DCMD vs AOD) (Rochette, 2018). The estimated correlation in the local window is recorded as the local spatial correlation for the central pixel. The process is repeated over the entire area (all pixel clusters), until the local spatial correlation values for all central pixels are estimated.

4.1. Dust storm event dynamics

Very few studies in the world have utilized DCMD, a parameter from the MERRA-2 gridded reanalysis dataset to study dust characteristics, transport, and the associated feedback (Jing et al., 2017; Gómez-Andújar et al., 2018; Sabetghadam et al., 2018; Yao et al., 2021). To the best of the authors’ knowledge, there exists no study on India that uses this dust dataset (aerosol column burden estimates) to characterize a dust storm event. Therefore, in this paper, we assess both, the total dust density in the air mass column and the fine dust (PM_2.5) component, referred to as DCMD total and DCMD PM_2.5, respectively. We estimate the changes in daily (average) dust quantity during the dust storm event (shown in Fig. 3a-e) against the background dust levels (shown in Fig. 3f-h). Similarly, Fig. 4a-e and Fig. 4f-h show the daily DCMD PM_2.5 average for the dust storm event and the background reference period, respectively. Northern India is found to be considerably dusty during the pre-monsoon season, due to the persistent hot and dry weather (especially in May), along with the significantly lower probability of large-scale convection. Although dust sources are largely local (Thar Desert), strong winds effectively increase the aerosol column burden (or DCMD) over a very large domain (Maharana et al., 2019). From 24–26 May 2018 (i.e., background period), we note that across the entire northern India DCMD total mostly ranges between 0-0.5 g/m² (Fig. 3f-h). However, some parts of Rajasthan and Uttar Pradesh show significantly higher DCMD total (0.5-1 g/m²).The background DCMD PM_2.5 remains < 0.2 g/m²over northern India (Fig. 4f-h). Considering the geographical and seasonal characteristics of the region, both total dust and fine dust column density during the considered background reference period remain substantially low. The day of 25 May 2018 is found to be the clearest in terms of dust loading in the air.

On the other hand, during the dust storm event, both DCMD total and DCMD PM_2.5 values are observed to be significantly elevated i.e., a high aerosol column burden is present over the region. Total and fine dust column mass density begins to elevate on 13 May 2018 in north-western parts of Rajasthan (Fig. 3b, 4b). A significant increase in dust loading is observed on the next day i.e., 14 May 2018 (Fig. 3c, 4c). During this peak time of the dust storm event, increased dust loading is spread over a significantly large portion of northern India. Compared to its background level (Fig. 3f-h), the DCMD total on 14 May 2018 (Fig. 3c) is noted to be around 3–5 times higher in northern India. Some regions in Rajasthan show DCMD total of close to 3 g/m², which is significantly high. Similarly, DCMD PM_2.5 is found to be at least 2–3 times higher during the dust storm event (Fig. 4c, 4f-h), and the maximum DCMD PM_2.5 is noted on 14 May 2018 (0.7 g/m²).Winds carry both the coarse and fine dust particles from the Thar Desert in Rajasthan (primary local source) and spread it over to the adjacent states- Haryana and Delhi, and then toward the farther state of Uttar Pradesh. Over the next two days (15–16 May 2018), the dust particles distribute and spread across the entire northern and central India (Fig. 3d,3e,4d,4e) and then dissipate entirely by 24 May 2018.

In regions of high dust storm activity, frequent atmospheric dust loading is the leading cause of significantly increased AOD (Alizadeh-Choobari et al., 2016; Sabetghadam et al., 2018). Other factors, such as, increased levels of air pollution also significantly contribute to the AOD increase. Typical values of AOD range between 0 (during clear skies) and 1 (during high aerosol loading).However, during dust storm events, AOD can cross its typical range of 0–1, due to an overabundance of coarse and fine dust in the overlying atmosphere. Previous studies suggest that during dust and sandstorm events, AOD is mostly found to exceed 1, but largely remains below 4 (Haywood et al., 2001; Alam et al., 2014; Basha et al., 2015). In the present case study, we observe that the daily average AOD is significantly higher during the dust storm event (Fig. 5a-e),as compared to the background reference period (Fig. 5f-h). A localized increase in AOD is noted over Rajasthan (including Jodhpur city) on 13 May 2018. As the dust laden winds carry the coarse and fine particles over to parts of northern India, an increase in AOD is also noted (Fig. 5c). The air quality over northern India is known to decline during days of high AOD (Dey et al., 2004; Sonwani and Kulshrestha, 2016). In the present study, we note high AOD during 14–16 May 2018 (Fig. 5c-d) over northern India, and an increased AOD over Uttar Pradesh till 24 May 2018 (Fig. 5f). During and after the dust storm event, AOD > 2 is also noted in some regions, implying a high residence time of dust aerosols and/ oran excessive amount of dust lingering in the air(Kumar et al., 2014).By 25 May 2018, the AOD over the state of Uttar Pradeshis noted to rapidly decline (Fig. 5g), showing dissipation of dust and PM. While AOD between the range of 0.2–0.6, is not indicative of very clear skies, it is still significantly low for a region like IGP, that has very high background levels of pollution.

Two important meteorological parameters- 2m temperature (Fig. 6) and accumulated precipitation (Fig. 7) are discussed in the following paragraph. Sarkar et al. (2019) studied three different pre-monsoon dust storm cases of 2018, over northern India, and highlighted a significant drop in surface temperature after each event i.e., a net radiative cooling was observed. Moreover, they noted that a higher residence time of dust in the atmosphere can lead to mid-tropospheric as well a supper-tropospheric heating. In the present study, daily mean 2m air temperature indicates warming (36–38°C) over Punjab, Haryana, Delhi, and some parts of Rajasthan, Uttar Pradesh, and Madhya Pradesh (Fig. 6a). Based on the weather reports (IMD, 2018; Outlook, 2018) for12-14 May 2018,the area experienced heat wave conditions in the preceding days, a high dust loading over Rajasthan (Fig. 3a-c), and heavy precipitation (in the form of squall and thunderstorm) due to the occurrence of a western disturbance (Fig. 7a,7b). This precipitation significantly cooled the atmosphere over northern India (Fig. 6b-e). On 14 May 2018, along with the widespread dust over northern India (Fig. 3c), the 2m air temperature (daily mean) is found to mostly remain lower than 34°C (Fig. 6c). In comparison, for the background reference period, the 2m air temperature remains higher(34–38°C), and moreover, some parts of IGP also note 2m air temperature over 42°C (Fig. 6f-h). The state of Rajasthan (including the Thar Desert) receives very less precipitation (IMD, 2011), and dust particles in sufficient quantity act as cloud condensation nuclei (CCN). Dust CCNs greatly affect cloud microphysics as well as rainfall patterns (Zhang et al., 2009). On 14 May 2018, with ‘high to excessive dust loading in the atmosphere, a greater number of CCNs are available, which could sufficiently aid in the occurrence of localized convection over south-eastern Rajasthan (shown in Fig. 7c). Even so, an overabundance of dust CCNs, over the remaining dust storm affected northern India, can, in turn, reduce the probability of precipitation. The local convective precipitation clouds can also dissipate due to strong winds, that were observed to blow over the region (IMD, 2018; Outlook, 2018) during this dust storm event. However, in the following two days of the dust storm event, scattered showers are observed (Fig. 7d,7e), indicating that a balance in CCN (not overabundance) and wind speed (not too strong) is important.

4.2. Dust storm progression

We plot the progression of the dust storm on 14 May 2018, on the basis ofthe 3-hourly average DCMD total over the region (Fig. 8). The 3-hourly average DCMD total is found to be many times higher (reaching over 3 g/m²) on the dust storm day (14 May 2018; Fig. 8), than on the dust background reference day (25 May 2018; Fig. 9), which mostly remains between 0.2–0.8 g/m². The dust storm begins to evolve between 0030–0330 UTC in the Thar Desert (Fig. 8a,8b),and an elevated level of total dust loading in the air mass column (1–2 g/m²) is observed mainly over north-western Rajasthan, including Jodhpur city. The dust load is observed to significantly increase with time, and it spreads eastward, following the prevailing wind direction (and with the Himalayas acting as the topographic barrier in the north). At 0930 UTC, a considerable increase in DCMD total over Rajasthan, Haryana and Delhi is observed, indicative of theintensification of the dust storm (Fig. 8d). The dust storm continues to spread over to Uttar Pradesh, and by 2130 UTC the entire northern India shows a 3-hourly average DCMD total of over 1.5 g/m² (Fig. 8d-h). Between 1530 UTC and 1830 UTC, DCMDtotal reaches a maximum of 3.9 g/m².

We also show the hourly variation of DCMD total for Jodhpur, Rohtak, Delhi, and Lucknow sites (Fig. 10). The increased dust loading on 14 May 2018 can be clearly noted, against that on 25 May 2018. Prior (i.e., from 0030–0830 UTC)to the peak dust loading, Jodhpur city shows the highest level of DCMD total (Fig. 10a). However, at 0930 UTC, a sharp spike in DCMD total is noted over both- Rohtak and Delhi cities, just as the dust laden air-mass from Jodhpur city is carried over to northern India. In comparison, the dust loading over Lucknow city remains low and increases gradually over time. Interestingly, as DCMD total levels gradually lower over Jodhpur city, they tend to keep increasing over Rohtak, Delhi, and Lucknow cities. This atmospheric dust loading pattern (using DCMD total) depicts the spatiotemporal progression of the regional mass transport of dust particles over northern India. As dust particles in the air column deplete from the dust storm source (i.e., Thar Desert),they go on to accumulate at far away sites across northern India. In contrast, the background level of DCMD total remains fairly constant throughout the reference day over all the four cities (Fig. 10b).

4.3. Distribution of particulate matter and meteorological parameters

We calculated the mean ± standard deviation values (shown in Table 3) of particulate matter (PM₁₀, and PM_2.5; Table 2, S. No. 7–8) as well as of the meteorological variables (AT, RH, WS, and AP; Table 2, S. No. 9–12),and plotted the real-time observations for these variables (shown in Fig. 11) to show their distribution during the dust storm event (12–16 May 2018) and the background reference (24–26 May 2018). The ground-based weather station data (used for Table 3 and Fig. 11) is obtained for the four cities- Jodhpur, Rohtak, Delhi, and Lucknow, which are representative of the states of Rajasthan, Haryana, NCT of Delhi, and Uttar Pradesh, respectively, in the larger north Indian region.

Table 3

Mean ± standard deviation calculated for particulate matter- PM₁₀and PM_2.5, and meteorological parameters-air temperature (AT), relative humidity (RH), wind speed (WS), and atmospheric pressure (AP) for the four cities- Jodhpur (Rajasthan), Rohtak (Haryana), Delhi (NCT of Delhi), and Lucknow (Uttar Pradesh), during the (a) dust storm event (12–16 May 2018),and (b) background reference period (24–26 May 2018).
(a) Dust storm event
	PM₁₀ (µg/m³)	PM_2.5 (µg/m³)	AT (^oC)	RH (%)	WS (m/s)	AP (mmHg)
Jodhpur	298.8 ± 74.1	232.8 ± 34.2	38.8 ± 2.9	35.7 ± 5.3	1.16 ± 0.4	755.5 ± 1.9
Rohtak	NA	98.1 ± 66.5	27.6 ± 5	58.6 ± 4	1.22 ± 0.5	978.9 ± 1.7
Delhi	250.9 ± 108.3	91.6 ± 45.1	32.7 ± 4.7	62.1 ± 6.1	1.37 ± 0.6	874.5 ± 36
Lucknow	NA	102.5 ± 71.9	32.1 ± 4.1	57.5 ± 10.7	0.86 ± 0.3	1003 ± 2.7
(b) Background reference period
Jodhpur	227.2 ± 41.1	117.5 ± 26.9	39.7 ± 3.5	28.4 ± 3.3	1 ± 0.5	760.5 ± 4.7
Rohtak	NA	59.6 ± 40	31.8 ± 5.2	53.1 ± 1.8	0.6 ± 0.3	976.1 ± 1.4
Delhi	341.8 ± 109.8	118 ± 52.7	34.8 ± 5.4	50.5 ± 4.3	7.4 ± 0.1	737.2 ± 0.3
Lucknow	NA	180.2 ± 77.8	36.5 ± 4.5	34.1 ± 20.1	0.8 ± 0.5	998.4 ± 1.8
Note- Data for PM₁₀ concentrations was unavailable for Rohtak and Lucknow cities during both- the dust storm event and the background reference period.*

Table 4

Daily average DCMD total and AOD over the four cities- Jodhpur (Rajasthan), Rohtak (Haryana), Delhi (NCT of Delhi), and Lucknow (Uttar Pradesh) for the dust storm day (14 May 2018) and the background reference day (25 May 2018).
	Dust storm day		Background reference day
	DCMD (g/m²)	AOD	DCMD (g/m²)	AOD
Jodhpur	1.85	0.69	0.44	0.29
Rohtak	2.27	1.34	0.38	0.34
Delhi	2.26	1.20	0.41	0.43
Lucknow	1.13	0.71	0.46	0.52

During the dust storm event (Table 3a),amongst all cities, Jodhpur (Rajasthan)was observed to have the highest average concentration of PM₁₀ (298.8 ± 74.1 µg/m³) and PM_2.5 (232.8 ± 34.2 µg/m³).This is similar to the elevated level of DCMD found at the onset of the dust storm event (Fig. 3, 4, 8, 10) in Rajasthan. The re-suspension of both- coarse and fine dust particles, from the Thar Desert (dust source), is attributed to the observed highest PM concentration in Jodhpur (Rajasthan). Interestingly, unlike the DCMD levels during the background reference day, which remained low for all four cities (Table 4; Fig. 10), the mean concentration of PM₁₀ (341.8 ± 109.8 µg/m³) and PM_2.5 (118 ± 52.7 µg/m³) is found to be the highest in Delhi, and is noted to even surpass the highest PM concentrations during the dust storm event (Table 3b). Simultaneously, the daily average DCMD total and daily average AOD, extracted over the four cities- Jodhpur, Rohtak, Delhi, and Lucknow (Table 4), show a multi-fold increase in both DCMD total and AOD on 14 May 2018 (i.e., dust storm day) as compared to the significantly low background values estimated on 25 May 2018. During the dust event, AOD is found to be higher in cities having maximum dust loading i.e., Rohtak and Delhi.

We highlight the role of extremely high local pollution in Delhi, from sources such as, industrial and vehicular emissions (Sonwani et al., 2021), as well as, emissions from urban built-up area construction and road/ rail infrastructure development-based activities (Jain et al., 2016; Jain, 2022b),in causing elevated PM at the ground measurement levels. In Table 3a, significantly higher temperature (38.8 ± 2.9°C) and low relative humidity (35.7 ± 5.3%),noted over Jodhpur (Rajasthan) than the other three cities, create favorable conditions (low-pressure zone and a less probability of local convection) for the lifting of dust from the Thar Desert and the occurrence of the dust storm in the larger north Indian region. The 24-h National Ambient Air Quality Standards (NAAQS) for PM₁₀, and PM_2.5 are prescribed as 100 and 60 µg/m³, respectively. Much like the significantly increased DCMD levels during the dust storm event, the PM₁₀ and PM_2.5 concentrations for all the four cities were found to exceed the prescribed NAAQS limits. For Jodhpur and Delhi, the PM₁₀ concentrations were found to be ~ 3 times, and ~ 2.5 times, higher than NAAQS limits, and in case of PM_2.5 concentrations, they were found to be ~ 4 times, and ~ 1.5 times, higher than NAAQS limits, respectively. Similarly, during the background reference period, all the cities were noted to cross the prescribed NAAQS limits for PM₁₀ and PM2.5 concentrations, indicating high levels of background pollution in the lower levels of the atmosphere in northern India and the IGP. We note high levels of the PM₁₀inthe afternoon and evening hours, as compared to night and early morning hours, in the four cities during the dust storm event (Fig. 11a).A peak in the PM₁₀concentrations was observed during the hours of the dust storm event, followed by a sharp decrease which could be attributed to precipitation led wet deposition. Similar observations were also reported by Sarkar et al. (2019). On the other hand, the PM_2.5 concentrations were found to be more during morning hours, coincident with the highest vehicular emissions arising during peak congested traffic hours (0700 to 1000 local time), a major contributor of PM_2.5 in cities. Crop residue burning in mid-May (after the rabi crop season harvest), in the northern Indian states of Punjab and Haryana, is another major source of PM_2.5 emissions that further degrades the air quality of the region during this time (Saxena et al., 2021). Overall, a significant worsening (DCMD total and AOD and observations over the air column) of the already poor air quality (background PM₁₀ and PM_2.5 observations) during the pre-monsoon dust storm event, at the four important cities of northern India (and the IGP) is observed in the present study.

4.4. NOAA HYSPLIT forward trajectories

Since the Thar Desert in India is spread into a large area of the state of Rajasthan, we selected three different locations- 1) Latitude/ Longitude: 27.4695°N, 70.6217°E, 2)Latitude/ Longitude: 26.9093°N, 70.9123°E, and 3)Latitude/ Longitude: 26.2389°N, 73.0243°E, asthe dust sources. The selection of these locations was based on the hotspots of the highest average DCMD total (Fig. 3c, 8) identified in the Thar Desert region, on the day of the dust storm event (14 May 2018). The NOAA HYSPLIT model was run, and we mapped the 72-hour forward propagation trajectories from these three selected dust source locations. For the case of the dust storm event, the 72-hour forward trajectories were set to begin from 14 May 2018 0000 UTC (Fig. 12a), and for the background reference period from 25 May 2018 0000 UTC (Fig. 12b). In Fig. 12, the air-mass trajectories for both events are shown at100m, 500m, and 1000m AGL.

A major difference in the air-mass movements in the case of the dust storm and the background reference period are noted. The 72-hour forward trajectories show that the dust laden air-masses originating from the Thar Desert on 14 May 2018, had the highest probability of reaching the neighbouring north Indian region and the IGP (Fig. 12a), causing a widespread dust storm event in the states of Haryana, Delhi, and Uttar Pradesh (Fig. 2–4). In contrast, most of the trajectories originating from the three dust sources on 25 May 2018 move towards central India, and do not reach over to the north Indian states of Haryana, Delhi or Uttar Pradesh. This is evident from the clear skies witnessed over northern India from the MODIS satellite image (Fig. 2b) and also from the low background dust levels from the DCMD plots (Fig. 3,4). The potential transport distance of dust particles increases with the height of the starting point (Aili et al., 2021).Our study also shows long-range transport of the air masses at 500m AGL and 1000m AGL, as compared to those at 100m AGL. This indicates a long-range, upper-level dust particle transport in northern India during the dust storm event (14 May 2018).These inferences are corroborated by the presence of high (low) air column DCMD total during the dust storm event (background reference period) over Jodhpur, Rohtak, Delhi, and Lucknow cities, as well as, the observed PM₁₀ and PM_2.5 concentrations at these monitoring ground level stations discussed in the previous subsection.

4.5. Local spatial correlation

Similar to linear statistical correlation, local spatial correlation values that are closer to 1 (-1) indicate a strong positive (negative) correlation between the considered variables and values closer to 0 are indicative of no relationship of statistical significance. Local spatial correlations need to be backed with proven theoretical causations in order to understand the dust storm dynamics. We note a significantly high local spatial correlation (> 0.5) between DCMD total and AOD for both- the dust storm day (14 May 2018) and the background reference day (25 May 2018). This implies that a high (low) column density of dust is associated with a high (low) AOD in the region (Fig. 13a, 13b). However, AOD is not solely dependent on dust loading, and other factors such as air pollution are also significant. This could be one of the reasons why some regions show a high AOD on the background reference day, despite low dust loading, thus resulting in a negative local spatial correlation over those regions(Fig. 13b).

In general, temperature and precipitation are inversely correlated with each other. The same is noted in Fig. 13 (c, d) for daily mean 2m air temperature and accumulated precipitation, respectively. In the case of accumulated precipitation and DCMD total (Fig. 13e, 13f), we find a significant negative local spatial correlation, especially for the dust storm event (14 May 2018). Simply put, if there is a precipitation event it will lead to significant scavenging of dust particles (i.e., lowering of DCMD total) present in the overlying atmosphere. On 14 May 2018, there was abundant dust availability in the air column and no major precipitation event occurred. It is noted that the algorithm used for computing local spatial correlations able to capture the intricate and complex spatial associations between dust loading and localized precipitation over the south-eastern part of Rajasthan. This positive and high correlation region is shown as the dark red region in Fig. 13e. Furthermore, we estimate the association between fine dust loading (DCMD PM_2.5) with wind speed. Fine dust (PM_2.5) has a higher residence time in the atmosphere and is there for preferred over DCMD (total dust loading in the air column), while computing the association with wind speeds. We note a high variability in the local spatial correlation of these two variables (Fig. 13g, 13h). Similarly, 2m air temperature and AOD also show a high local variability in the spatial correlation (Fig. 13i, 13j). This could be in part due to differences of local characteristics, air pollution emission differences, multi-factorial feedbacks of land air-interactions, and the other complex atmospheric dynamics at play. In general, local spatial correlation proved to be a very useful tool in assessing the local associations of meteorological parameters and helped to better understand the dust storm event.

An increasing number of severe dust storms with strong winds, in northern India, are being reported each year in the pre-monsoon months of April and May. Many of these cause human and animal deaths (and injuries), excessive damage to crops, electricity failures, broken telecommunication networks, tree uproots, damage to property and infrastructure, as well as flight diversions and cancellations. Thar Desert in the state of Rajasthan is the prominent local dust source for dust storms originating in the northern India, and both- coarse and fine dust particles, are transported via prevailing winds to a significantly large portion of the IGP and northern India. The IGP where over 700 million people (~ 9% of the global population) reside has very high background levels of PM that exceed health standards for humans (Greenstone and Fan, 2018; Mogno et al., 2021) as well as high year-around AOD, but more so during the pre-monsoon season (David et al., 2018). Increasingly frequent pre-monsoon dust storms further cause a severe worsening in the already poor air quality of this region, and therefore require urgent attention of the scientific community, government bodies, policymakers, health and emergency services. The present study shows that DCMD, a variable that has been used in only in a few dust storms studies across the world, and in none over the IGP or over the north Indian region, has a wide potential when studying the characteristics, dynamics, and progression of pre-monsoon dust storms. In the dearth of weather station measurements and/ or publicly available observational datasets, a common constraint over the Indian subcontinent, the inclusion of MERRA-2 climate model derived variables- DCMD total and DCMD PM_2.5, is recommended for dust storm studies.

Daily average of DCMD total, DCMD PM_2.5, and AOD are found to mostly remain between 0–1 g/m², 0-0.2g/m², and 0-0.6, respectively, over northern India and the IGP, during 24–26 May 2018, and provides a background reference of the air quality of the region during the period of relatively ‘clear’ skies. However, during the dust storm event, the daily average of DCMD total, DCMD PM_2.5, and AOD, is found to exceed 2.5g/m², 0.5g/m², and 1 over a significant portion of the affected region. Coarse and fine particulate matter (PM₁₀ and PM_2.5) ground measurements at monitoring stations in Jodhpur (Rajasthan), Rohtak (Haryana), Delhi (NCT of Delhi), and Lucknow (Uttar Pradesh) are found to exceed the 24-hour NAAQS permissible limits, during the dust storm event and the background reference period, due to other emission sources including- industrial and vehicular emissions, construction activities, CRB emissions, etc. Simultaneously, AOD > 2 along with significantly high DCMD total and DCMD PM_2.5 in the affected region (and over the four cities of Jodhpur, Rohtak, Delhi, and Lucknow) shows a high residence time of dust (coarse and fine) aerosols in the air or excessive dust lingering. The already poor background air quality of the region is shown to severely worsen during the pre-monsoon dust storm event due to the additional dust aerosol loading in the atmospheric column. Our study shows long-range transport of the air masses at 500m AGL and 1000m AGL, as deduced from the 72-hour forward propagation trajectories of the NOAA HYSPLIT model, indicating a lower-tropospheric level dust particle transport in northern India during the dust storm event. The study highlights the role of prevailing temperature (heat wave conditions in days preceding the dust storm event), and precipitation (a western disturbance occurrence) conditions in the region on the dust storm dynamics. Further, the feedback role of dust aerosol CCNs available in abundance in the overlying atmosphere, in lowering the regional temperature, and also in causing a localized convection over south-eastern Rajasthan (where generally probability of precipitation is low)is discussed.

To quantify the statistical significance of associations between the gridded data variables- DCMD, AOD, 2m air temperature, precipitation, and wind speed, local spatial correlations using a maximum likelihood moving window of 3x3 pixels were estimated. The present study highlights a- 1) a significantly high positive local spatial correlation (> 0.5) between DCMD total and AOD, 2) a significant negative local spatial correlation between 2m air temperature and accumulated precipitation in areas of occurrence of the western disturbance and the localized convection, 3) significant negative local spatial correlation, between DCMD total and accumulated precipitation, indicating wet scavenging of dust aerosol CCNs, and 4) high regional variability in the local spatial correlations between DCMD PM_2.5 and wind speed as well as between 2m air temperature and AOD, which could be due to the residence time properties of fine dust particles, presence of other air pollution emission sources, complex land-air interactions and other atmospheric dynamical systems involved. The use of local spatial correlation is suggested over linear statistical regression and correlation methods (Al-Ahmadi and Al-Ahmadi, 2013)as it is found to be robust in explaining the various spatial associations at play.

For better results, higher temporal resolution, such as hourly data should be considered. Hourly data is able to better capture the fine intricacies amongst all variables which otherwise might get averaged out in daily data estimates. However, since high temporal resolution data was unavailable for all the considered variables, the local spatial correlations were limited to daily means. Another limitation of the study is the quantification of contribution of dust aerosol versus the other air pollution emission sources in both- the AOD column estimates and the particulate matter (PM₁₀ and PM_2.5) ground weather station measurements. Such quantification still remains a challenging task and full of uncertainties (Allen and Sherwood, 2011). While this study adds to the existing scientific literature on dust storm characteristics and dynamics, it will also benefit from availability of more publicly accessible air quality and dust aerosol datasets at ground stations as well as for low-, and mid-troposphere in northern India or the larger IGP region. The results of this study could provide a basis for forecasting pre-monsoon dust storms in northern India and can be used for targeted mitigation and adaptation plans in dust storm-vulnerable areas. Implementation of such plans would help to reduce the negative impacts of dust storms on the regional environment, agriculture, economy, and on human health.

Acknowledgments

The authors acknowledge National Aeronautics and Space Administration (NASA) Giovanni portal for the access of satellite observation and climate model reanalysis datasets, Central Pollution Control Board (CPCB) portal and India Meteorological Department (IMD) portal for real time data of air quality and meteorological parameters at the ground weather monitoring stations. All these datasets were crucial in the present study. The authors acknowledge the role of independent reviewers and the editor in the improvement of this manuscript. There is no funding available for this research work.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Author Contributions

Madhavi Jain’s crediT roles: Conceptualization; data curation; formal analysis; investigation; methodology; software; validation; role/writing-original draft; writing, review & editing

Pallavi Saxena’s crediT roles: Conceptualization; formal analysis; formal analysis; investigation; methodology; supervision; validation; role/writing – review & editing writing-review.

Saurabh Sonwani’s crediT roles: Conceptualization; data curation; formal analysis; investigation; methodology; resources; software; visualization; writing-original draft; role/writing- review & editing

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Aili, A., Abuduwaili, J., Xu, H., Zhao, X., & Liu, X. (2021). A Cluster Analysis of Forward Trajectory to Identify the Transport Pathway of Salt-Dust Particles from Dried Bottom of Aral Sea, Central Asia. Atmosphere, 12(6), 764.
Al-Ahmadi, K. and Al-Ahmadi, S., 2013. Rainfall-altitude relationship in Saudi Arabia. Advances in Meteorology, 2013.
Alam, K., Trautmann, T., Blaschke, T. and Subhan, F., 2014. Changes in aerosol optical properties due to dust storms in the Middle East and Southwest Asia. Remote sensing of environment, 143, pp.216-227.
Alizadeh‐Choobari, O., Ghafarian, P. and Owlad, E., 2016. Temporal variations in the frequency and concentration of dust events over Iran based on surface observations. International Journal of Climatology, 36(4), pp.2050-2062.
Allen, R.J. and Sherwood, S.C., 2011. The impact of natural versus anthropogenic aerosols on atmospheric circulation in the Community Atmosphere Model. Climate dynamics, 36(9-10), pp.1959-1978.
Baker, J. (2010). A cluster analysis of long range air transport pathways and associated pollutant concentrations within the UK. Atmospheric Environment, 44(4), 563-571.
Balkanski, Y., Schulz, M., Claquin, T., & Guibert, S. (2007). Reevaluation of Mineral aerosol radiative forcings suggests a better agreement with satellite and AERONET data.
Basha, G., Phanikumar, D.V., Kumar, K.N., Ouarda, T.B. and Marpu, P.R., 2015. Investigation of aerosol optical, physical, and radiative characteristics of a severe dust storm observed over UAE. Remote Sensing of Environment, 169, pp.404-417.
Census of India (2011) https://censusindia.gov.in/ [Accessed on Aug 20, 2020]
Chen, S., Huang, J., Zhao, C., Qian, Y., Leung, L. R., & Yang, B. (2013). Modeling the transport and radiative forcing of Taklimakan dust over the Tibetan Plateau: A case study in the summer of 2006. Journal of Geophysical Research: Atmospheres, 118(2), 797-812.
Chen, S., Jiang, N., Huang, J., Xu, X., Zhang, H., Zang, Z., ... & Zhang, X. (2018). Quantifying contributions of natural and anthropogenic dust emission from different climatic regions. Atmospheric Environment, 191, 94-104.
David, L. M., Ravishankara, A. R., Kodros, J. K., Venkataraman, C., Sadavarte, P., Pierce, J. R., ... & Millet, D. B. (2018). Aerosol optical depth over India. Journal of Geophysical Research: Atmospheres, 123(7), 3688-3703.
Dey, S., Tripathi, S. N., Singh, R. P., & Holben, B. N. (2004). Influence of dust storms on the aerosol optical properties over the Indo-Gangetic basin. Journal of Geophysical Research: Atmospheres, 109(D20).
Di, A., Xue, Y., Yang, X., Leys, J., Guang, J., Mei, L., ... & Che, Y. (2016). Dust aerosol optical depth retrieval and dust storm detection for Xinjiang region using Indian National Satellite Observations. Remote Sensing, 8(9), 702.
Gautam, R., Liu, Z., Singh, R. P., & Hsu, N. C. (2009). Two contrasting dust‐dominant periods over India observed from MODIS and CALIPSO data. Geophysical Research Letters, 36(6).
Gómez-Andújar, N., Torres-Delgado, E. and Mayol-Bracero, O., 2018. Temporal Relationships Between African Dust and Chlorophyll-α in the Eastern Caribbean Basin.
Greenstone, M., & Fan, C. Q. (2018). Introducing the Air Quality Life Index. AQLI Annual Report.
Hamidi, M. (2020). The key role of water resources management in the Middle East dust events. Catena, 187, 104337.
Haywood, J.M., Francis, P.N., Geogdzhayev, I., Mishchenko, M. and Frey, R., 2001. Comparison of Saharan dust aerosol optical depths retrieved using aircraft mounted pyranometers and 2‐channel AVHRR algorithms. Geophysical research letters, 28(12), pp.2393-2396.
IMD (2011) India Meteorological Department: Climatological Normals 1971-2000 http://imdpune.gov.in/library/public/Climatological%20Normals%20%281971-2000%29.pdf [Accessed on Aug 20, 2020]
IMD (2015) India Meteorological Department: Climatological Normals 1981-2010 https://web.archive.org/web/20200205040301/http://imdpune.gov.in/library/public/1981-2010%20CLIM%20NORMALS%20%28STATWISE%29.pdf [Accessed on Aug 20, 2020]
IMD 2018. All India weekly weather report 10 – 16 May 2018. http://agricoop.nic.in/sites/default/files/CWWG_1852018.pdf Accessed 05 Sep 2020
Jain, M. (2022a). Increasing atmospheric extreme events and role of disaster risk management: Dimensions and approaches. In Extremes in Atmospheric Processes and Phenomenon: Assessment, Impacts and Mitigation (pp. 303-328). Singapore: Springer Nature Singapore.
Jain, M. (2022b). Mitigation of urbanization ill-effects through urban agriculture inclusion in cities. In New Forms of Urban Agriculture: An Urban Ecology Perspective (pp. 39-56). Singapore: Springer Nature Singapore.
Jain, M., Dawa, D., Mehta, R., Dimri, A. P., & Pandit, M. K. (2016). Monitoring land use change and its drivers in Delhi, India using multi-temporal satellite data. Modeling earth systems and environment, 2, 1-14.
Jickells, T. D., An, Z. S., Andersen, K. K., Baker, A. R., Bergametti, G., Brooks, N., ... & Kawahata, H. (2005). Global iron connections between desert dust, ocean biogeochemistry, and climate. science, 308(5718), 67-71.
Jing, Y., Zhang, P., Chen, L. and Xu, N., 2017. Integrated analysis of dust transport and budget in a severe Asian dust event. Aerosol and Air Quality Research, 17(10), pp.2390-2400.
Jing, Y., Zhang, P., Chen, L., & Xu, N. (2017). Integrated analysis of dust transport and budget in a severe Asian dust event. Aerosol and Air Quality Research, 17(10), 2390-2400.
Kedia, S., Kumar, R., Islam, S., Sathe, Y., & Kaginalkar, A. (2018). Radiative impact of a heavy dust storm over India and surrounding oceanic regions. Atmospheric Environment, 185, 109-120.
Khaniabadi, Y. O., Daryanoosh, S. M., Amrane, A., Polosa, R., Hopke, P. K., Goudarzi, G., ... & Armin, H. (2017). Impact of Middle Eastern Dust storms on human health. Atmospheric pollution research, 8(4), 606-613.
Kim, D., Chin, M., Kemp, E. M., Tao, Z., Peters-Lidard, C. D., & Ginoux, P. (2017). Development of high-resolution dynamic dust source function-A case study with a strong dust storm in a regional model. Atmospheric environment, 159, 11-25.
Kok, J. F., Ridley, D. A., Zhou, Q., Miller, R. L., Zhao, C., Heald, C. L., ... & Haustein, K. (2017). Smaller desert dust cooling effect estimated from analysis of dust size and abundance. Nature Geoscience, 10(4), 274-278.
Krishnakumar, K. N., Rao, G. P., & Gopakumar, C. S. (2009). Rainfall trends in twentieth century over Kerala, India. Atmospheric environment, 43(11), 1940-1944.
Kumar, R., Barth, M. C., Pfister, G. G., Naja, M., & Brasseur, G. P. (2014). WRF-Chem simulations of a typical pre-monsoon dust storm in northern India: influences on aerosol optical properties and radiation budget. Atmos. Chem. Phys, 14(5), 2431-2446.
Kumar, S., Kumar, S., Kaskaoutis, D. G., Singh, R. P., Singh, R. K., Mishra, A. K., ... & Singh, A. K. (2015). Meteorological, atmospheric and climatic perturbations during major dust storms over Indo-Gangetic Basin. Aeolian Research, 17, 15-31.
Lodhi, N. K., Beegum, S. N., Singh, S., & Kumar, K. (2013). Aerosol climatology at Delhi in the western Indo‐Gangetic Plain: Microphysics, long‐term trends, and source strengths. Journal of Geophysical Research: Atmospheres, 118(3), 1361-1375.
Maharana, P., Dimri, A.P. and Choudhary, A., 2019. Redistribution of Indian summer monsoon by dust aerosol forcing. Meteorological Applications, 26(4), pp.584-596.
Maji, S., & Sonwani, S. (2022). Nature of Sand and Dust Storm in South Asian Region: Extremities and Environmental Impacts. In Extremes in Atmospheric Processes and Phenomenon: Assessment, Impacts and Mitigation (pp. 113-139). Singapore: Springer Nature Singapore.
Min, Q. L., Li, R., Lin, B., Joseph, E., Wang, S., Hu, Y., ... & Chang, F. (2009). Evidence of mineral dust altering cloud microphysics and precipitation. Atmospheric Chemistry & Physics, 9(9).
Mogno, C., Palmer, P. I., Knote, C., Yao, F., & Wallington, T. J. (2021). Seasonal distribution and drivers of surface fine particulate matter and organic aerosol over the Indo-Gangetic Plain. Atmospheric Chemistry and Physics, 21(14), 10881-10909.
Naimabadi, A., Ghadiri, A., Idani, E., Babaei, A. A., Alavi, N., Shirmardi, M., ... & Goudarzi, G. (2016). Chemical composition of PM10 and its in vitro toxicological impacts on lung cells during the Middle Eastern Dust (MED) storms in Ahvaz, Iran. Environmental Pollution, 211, 316-324.
Newman, C. E., Hueso, R., Lemmon, M. T., Munguira, A., Vicente-Retortillo, Á., Apestigue, V., ... & Guzewich, S. D. (2022). The dynamic atmospheric and aeolian environment of Jezero crater, Mars. Science Advances, 8(21), eabn3783.
Oultook 2018. 53 Killed in storms across India, warning issued for today. https://www.outlookindia.com/website/story/41-killed-in-storms-across-india-warning-issued-for-today/311923. Accessed 13 Sep 2020
Pandey, S. K., Vinoj, V., & Panwar, A. (2020). The short-term variability of aerosols and their impact on cloud properties and radiative effect over the Indo-Gangetic Plain. Atmospheric Pollution Research, 11(3), 630-638.
Pandey, S. K., Vinoj, V., Landu, K., & Babu, S. S. (2017). Declining pre-monsoon dust loading over South Asia: Signature of a changing regional climate. Scientific reports, 7(1), 1-10.
Prasad, A. K., & Singh, R. P. (2007). Changes in aerosol parameters during major dust storm events (2001–2005) over the Indo‐Gangetic Plains using AERONET and MODIS data. Journal of Geophysical Research: Atmospheres, 112(D9).
Prasad, A. K., Singh, S., Chauhan, S. S., Srivastava, M. K., Singh, R. P., & Singh, R. (2007). Aerosol radiative forcing over the Indo-Gangetic plains during major dust storms. Atmospheric Environment, 41(29), 6289-6301.
Rajagopalan, B. and Lall, U., 1998. Locally weighted polynomial estimation of spatial precipitation. Journal of Geographic Information and Decision Analysis, 2(2), pp.44-51.
Ramachandran, S., Srivastava, R., Kedia, S. and Rajesh, T.A., 2012. Contribution of natural and anthropogenic aerosols to optical properties and radiative effects over an urban location. Environmental Research Letters, 7(3), p.034028.
Rochette S. (2018, Jan. 27). "Spatial correlation between rasters". Retrieved from https://statnmap.com/2018-01-27-spatial-correlation-between-rasters/. Accessed 02 Aug 2020
Sabetghadam, S., Khoshsima, M. and Alizadeh-Choobari, O., 2018. Spatial and temporal variations of satellite-based aerosol optical depth over Iran in Southwest Asia: Identification of a regional aerosol hot spot. Atmospheric Pollution Research, 9(5), pp.849-856.
Sarkar, S., Chauhan, A., Kumar, R. and Singh, R.P., 2019. Impact of deadly dust storms (May 2018) on air quality, meteorological, and atmospheric parameters over the northern parts of India. GeoHealth, 3(3), pp.67-80.
Sikka, D. R. (1997). Desert climate and its dynamics. Current Science, 35-46.
Slingo, A., Ackerman, T. P., Allan, R. P., Kassianov, E. I., McFarlane, S. A., Robinson, G. J., ... & Dewitte, S. (2006). Observations of the impact of a major Saharan dust storm on the atmospheric radiation balance. Geophysical Research Letters, 33(24).
Soleimani, Z., Teymouri, P., Boloorani, A. D., Mesdaghinia, A., Middleton, N., & Griffin, D. W. (2020). An overview of bioaerosol load and health impacts associated with dust storms: A focus on the Middle East. Atmospheric Environment, 223, 117187.
Solmon, F., Nair, V. S., & Mallet, M. (2015). Increasing Arabian dust activity and the Indian summer monsoon. Atmospheric Chemistry and Physics, 15(14), 8051.
Sonwani, S., & Kulshreshtha, U. (2016). Particulate matter levels and it’s associated health risks in East Delhi. In Proceedings of Indian aerosol science and technology association conference on aerosol and climate change: insight and challenges. IASTA Bull (Vol. 22, No. 1-2).
Sonwani, S., & Kulshrestha, U. C. (2019). PM 10 carbonaceous aerosols and their real-time wet scavenging during monsoon and non-monsoon seasons at Delhi, India. Journal of Atmospheric Chemistry, 76(3), 171-200.
Sonwani, S., Saxena, P., & Khillare, P. S. (2022). Profile of atmospheric particulate PAHs near busy roadway in tropical megacity, India. Inhalation Toxicology, 34(1-2), 39-50.
Sonwani, S., Saxena, P., & Shukla, A. (2021). Carbonaceous aerosol characterization and their relationship with meteorological parameters during summer monsoon and winter monsoon at an industrial region in Delhi, India. Earth and Space Science, 8(4), e2020EA001303.
Stein, A. F., Draxler, R. R., Rolph, G. D., Stunder, B. J., Cohen, M. D., & Ngan, F. (2015). NOAA’s HYSPLIT atmospheric transport and dispersion modeling system. Bulletin of the American Meteorological Society, 96(12), 2059-2077.
Tiwari, S., Kumar, A., Pratap, V., & Singh, A. K. (2019). Assessment of two intense dust storm characteristics over Indo–Gangetic basin and their radiative impacts: A case study. Atmospheric Research, 228, 23-40.
Toon, O. B. (2003). African dust in Florida clouds. Nature, 424(6949), 623-624.
Wan, Y., Li, Y., Liu, C., & Li, Z. (2020). Is traffic accident related to air pollution? A case report from an island of Taihu Lake, China. Atmospheric Pollution Research, 11(5), 1028-1033.
Wiggs, G. F., O'hara, S. L., Wegerdt, J., Van Der Meer, J., Small, I., & Hubbard, R. (2003). The dynamics and characteristics of aeolian dust in dryland Central Asia: possible impacts on human exposure and respiratory health in the Aral Sea basin. Geographical Journal, 169(2), 142-157.
Yao, W., Gui, K., Wang, Y., Che, H., & Zhang, X. (2021). Identifying the dominant local factors of 2000–2019 changes in dust loading over East Asia. Science of The Total Environment, 777, 146064.
Zhang, H., McFarquhar, G.M., Cotton, W.R. and Deng, Y., 2009. Direct and indirect impacts of Saharan dust acting as cloud condensation nuclei on tropical cyclone eyewall development. Geophysical Research Letters, 36(6).
Zhuang, G., Guo, J., Yuan, H., & Zhao, C. (2001). The compositions, sources, and size distribution of the dust storm from China in spring of 2000 and its impact on the global environment. Chinese Science Bulletin, 46(11), 895-900.

The authors declare no competing interests.

GraphicalAbstract.jpg

Download PDF

Version 1

posted

You are reading this latest preprint version

Characteristics, Dynamics, and Impact of the Thar Desert Dust Storms on Air Quality over Northern India

Status:

Version 1

Abstract

Figures

1. Introduction

2. Study Area

3. Data and Methods

3.1. Datasets

3.2. Dust storm event selection

3.3. Methodology

4. Results and Discussion

4.1. Dust storm event dynamics

4.2. Dust storm progression

4.3. Distribution of particulate matter and meteorological parameters

4.4. NOAA HYSPLIT forward trajectories

4.5. Local spatial correlation

5. Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1