Introduction

Exposure to higher concentrations of different trace gases and particulate matter leads to a deteriorating impact on human health (Lelieveld et al. 2013) and crops (Lal et al. 2017). This is of major concern in the developing countries where anthropogenic emissions are shown to be increasing (Akimoto 2003; Li et al. 2017a, b) in the want of economic developments. The hazardous air pollutants cause premature mortality, with the highest mortality rate in the Western Pacific region (~ 46%) and Southeast Asia (~ 25%) (Lelieveld et al. 2013). It has been shown that industrial coal-burning, including thermal power plants, is one of the major sources of air pollution in developing countries like India, apart from biomass burning (Venkataraman et al. 2018; Park and Choi 2016). In addition to the higher pollutant levels, abundant water vapor and the intense solar radiation in the tropical developing countries further contribute to more OH radicals, hence complex photochemistry, involving several trace gases.

India, the rapidly developing country, is increasing its urbanization and industrialization at a faster rate and adding more anthropogenic emissions (Kurokawa et al. 2013, Kurokawa and Ohara 2020; Kumar et al. 2018; Yarragunta et al. 2021). This is of serious concern in the northern region, where emissions from coal-based power plants, automobiles, industries, extensive agriculture practices, etc. are very significant (Singh and Kaskaoutis 2014; Cusworth et al. 2018; Bhardwaj et al. 2016; Srivastava and Naja 2021). However, there is a slightly lower emission and larger washout of pollutants in India’s southern part (Viswanadham and Santosh 1989; Rosenfeld et al. 2002).

Due to the recent pandemic, lockdowns over different countries have resulted in reductions in emissions of different atmospheric pollutants. The reduction was very significant over China (Xu et al. 2020; Silver et al. 2020) and somewhat moderate over Europe (Ropkins and Tate 2020) and the USA (Shakoor et al. 2020). Similarly, over India, the surface and space-borne observations of different air pollutants showed the highest decrease for PM2.5 and PM10 (40–60%) followed by NO2 (30–70%) and CO (20–40%) (Singh et al. 2020; Mahato et al. 2020; Biswal et al. 2020; Ravindra, et al. 2021). However, to our knowledge, no study examined the effect of lockdown in the vertical profiles of trace gases over India.

This work presents the changes in ozone and CO at different altitudes along with the columnar study of ozone, CO, and NO2 during 24 March to 20 April (lockdown period) in 2018, 2019, and 2020 over India. The CO and NO2, the primary air pollutants, are emitted from both anthropogenic and natural activities. Their major emissions sources are combustion processes, including those from industrial/vehicular/biomass/biofuel, and found in higher abundances around highly populated areas (Lamsal et al. 2013). Atmospheric ozone, a useful stratospheric gas for absorbing harmful UV radiation (Madronich 1993; Gao et al. 2012), is a greenhouse gas (Fishman et al. 1979) and a secondary air pollutant in the troposphere (Desqueyroux et al. 2002; Kim et al. 2013). Apart from photochemical production, ozone concentrations are also altered via meteorology and dynamics, including downward transport of ozone-rich air from the stratosphere (Holton et al. 1995; Park et al. 2012). Hence, ozone concentrations and variabilities are complex and largely depend upon local, regional, and seasonal factors. Furthermore, it has been shown that the ozone variations near the surface are more complex than the primary pollutants during the lockdown period over different regions of the world (Shi and Brasseur 2020; Sicard et al. 2020; Ordóñez et al. 2020; Peralta et al. 2020).

Here, we utilized multi-satellite observations, namely OMI on the Aura satellite (Schoeberl et al. 2006), TROPOMI on the Sentinel-5 Precursor satellite (Veefkind et al. 2012), IASI (Klaes et al. 2007), and IASI + GOME-2 on the MetOp satellite, for tracing pollutant levels vertically and horizontally. The synergic IASI + GOME-2 observation of ozone has a very high sensitivity for the lowermost tropospheric ozone (Cuesta et al. 2013, 2018) and offers a unique capability for observing tropospheric ozone profiles during lockdown. The NO2 columnar observation from OMI and TROPOMI and the subsequent differences of their retrieval over the Indian region are also presented. Additionally, we have employed the data from the year 2018, apart from the year 2019 that has shown new results in terms of changes in concentrations over different parts of India and minimized the conclusions’ biases with respect to a particular year.

Study region, dataset, and methodology

Study region

Figure 1 shows the present study domain (5–35° N and 65–95° E) and the marked locations corresponding to selected twelve cities and four coal-based thermal power plants of India. Figure 1 also demonstrates the National Centers for Environmental Prediction (NCEP) reanalysis-based average wind speed and wind direction for 06 UTC (11:30 IST) at 850 hPa over the study region during the period 24 March to 20 April in 2018, 2019, and 2020. NCEP-based wind analysis shows no significant changes in wind patterns among 2018, 2019, and 2020 with mostly regional localized and low-magnitude winds near the surface.

Fig. 1
figure 1

Average wind speed and direction at 850 hPa during 24 March to 20 April in a 2018, b 2019, and c 2020 based on 06 UTC NCEP-FNL reanalysis data. The numbers indicate cities and thermal power plants (TPPs) namely, (1) Nainital, (2) Delhi, (3) Kanpur, (4) Jaipur, (5) Ahmedabad, (6) Mumbai, (7) Kolkata, (8) Hyderabad, (9) Bangalore, (10) Chennai, (11) Thiruvananthapuram, (12) Visakhapatnam, (13) Vindhyachal TPP, (14) Mundra TPP, (15) Chandrapur TPP, and (16) Guru Gobind TPP

Full size image

IASI CO observations

IASI is a nadir-viewing Michelson interferometer onboard MetOP satellite (Klaes et al. 2007). MetOP is a polar sun-synchronous satellite with a descending node around 09:30 local time. The IASI nadir-viewing instrument senses IR spectrum (3.7 to 15.4 µm, total 8461 channels) and has a horizontal resolution of 12 km at nadir up to 40 km over a swath width of about 2,200 km (Clerbaux et al. 2009; Turquety, et al. 2009). The nadir spectral radiance in the range of 2143 to 2181.25 cm−1 is utilized to retrieve CO. Here, we have utilized the IASI level2 CO profiles. The retrieval of the CO profile is based on the O3M-SAF ULB/LATMOS Fast Optimal Retrievals on Layers for IASI CO (FORLI-CO algorithm (Hurtmans et al. 2012). The CO profile is provided in 19 fixed layers up to 60 km, corresponding to 18 layers from 0 to 18 km of 1 km width, and a top layer between 18 and 60 km. The IASI CO FORLI retrieval has higher sensitivity of the retrieved CO around 500 hPa with peaking averaging kernel at this altitude and having retrieval error of less than 10% and 21% in the upper troposphere and lower troposphere, respectively (Wachter et al. 2012). We have used the best quality level 2 CO profiles with a degree of freedom for signal (DOFS) more than 0.5376 (George et al. 2009). Here, the FORLI-CO partial profile (molecules/cm2) of IASI is constructed as follows:

$$\mathrm{CO}\_\mathrm{CP}\_{\mathrm{CO}}_{\mathrm i}=\mathrm{CO}\_\mathrm{CP}\_\mathrm{CO}\_{\mathrm A}_{\mathrm i}\times\mathrm{CO}\_\mathrm X\_{\mathrm{CO}}_{\mathrm i}$$

which is further converted to volume mixing ratio (VMR) profile as:

$$\mathrm{CO}\_\mathrm{VMR}\_{\mathrm{CO}}_{\mathrm i}=\mathrm{CO}\_\mathrm{CP}\_{\mathrm{CO}}_{\mathrm i}\div\mathrm{CO}\_\mathrm{CP}\_{\mathrm{AIR}}_{\mathrm i}$$

where i stands for different altitude layers, CO_CP_CO_A is a priori partial columns for CO on each retrieved layer (molecules/cm2), CO_X_CO is scaling vector used for multiplying the a-priori CO vector in order to find the retrieved CO partial profile, and CO_CP_AIR is air partial column on each retrieved layer (molecules/cm2).

OMI NO2column observations

The ozone monitoring instrument (OMI), onboard the sun-synchronous polar-orbiting Aura satellite, is a push broom scanning instrument. OMI has an ultraviolet–visible (UV–VIS) spectrometer that measures solar irradiance and the earth’s backscattered radiances at the top of the atmosphere. Its swath width is of 2600 km, enabling global daily coverage with a nadir field-of-view (FOV) of up to 13 km × 24 km. The OMI NO2 Standard Product (SP) algorithm has utilized the Differential Optical Absorption Spectroscopy (DOAS) (Marchenko et al. 2015; Krotkov et al. 2017; Lamsal et al. 2021). The most recent OMI NO2 SP, Version 4.0, has been found more reliable and consistent with ground-based and model-based observations (Lamsal et al. 2021), and is minimally affected by modeling artifacts and reduced negative tropospheric NO2 values (Bucsela et al. 2013). Its validation studies show a general good agreement of its column with lower and higher values to within ± 20% relative to in situ measurements with no consistent seasonal biases (Lamsal et al. 2014, 2021). However, slightly lower columns in urban regions and somewhat higher in the remote areas are observed (Lamsal et al. 2021). We have used the best quality data (VcdQualityFlags = 0) of the OMI tropospheric NO2 column as in other studies (Goldberg et al. 2017; de Foy et al. 2015).

TROPOMI NO2,CO, and HCHO column observations

TROPOMI onboard Sentinel-5 Precursor (S-5P) is a more recent and advanced satellite in near-polar sun-synchronous orbit, launched on 13 October 2017 (Veefkind et al. 2012). It is a passive-sensing hyperspectral nadir-viewing imager, flying at an altitude of 817 km, with an ascending node of 13:30 local time and a repeat cycle of 17 days. It has four separate spectrometers, which measure in the ultraviolet (UV), UV–VIS, near-infrared (NIR), and short-wavelength infrared (SWIR) spectral bands. TROPOMI provides both offline and near real-time products for vertical columns of NO2, HCHO, CO, and other trace gases with a higher spatial resolution around 5.5 to 7.5 km before August 2019 and 3.5 to 5.5 km afterward (Veefkind et al. 2012; Griffin et al. 2019). The retrieval of column densities from TROPOMI L1 UV–VIS radiance and solar irradiance observation is done using the DOAS method (Platt and Stutz 2008). The column L2 offline data of NO2, CO, and HCHO of better quality (qa_value > 0.7) is used in the present study (Zhao et al. 2020). The TROPOMI retrieved offline product is available after May 2018; thereby, we could utilize TROPOMI observations in 2019 and 2020 only.

Synergetic IASI + GOME-2 tropospheric ozone observations

Satellite remote sensors utilize spectral response of atmospheric ozone in UV (240–340 nm) and IR (9.6 μm) to observe ozone from space (Worden et al. 2007; Landgraf and Hasekamp 2007). The UV-based satellite measurements are more sensitive for the upper tropospheric and stratospheric ozone; however, the IR-based satellite retrievals relying on the emission/absorption of terrestrial and atmospheric emitted 9.6-μm radiation are more sensitive for the tropospheric ozone. Hence, their synergic observations are used for improved ozone profile retrievals, particularly in the troposphere (Sato et al. 2018; Cuesta et al. 2013).

Here, we have utilized the synergic observation of the IASI and GOME-2 in thermal infrared and ultraviolet, respectively. The co-located morning overpasses of IASI and GOME-2 measurements are used to derive ozone profiles. This unique retrieval technique uses Tikhonov–Phillips-type altitude-dependent constraints, which optimize sensitivity to lower tropospheric ozone. The linearized pseudo-spherical vector discrete ordinate radiative transfer (VLIDORT) and Karlsruhe Optimized and Precise Radiative transfer Algorithm (KOPRA) radiative transfer models are used for simulating UV reflectance and TIR radiance, and the successive retrieval is accompanied by a priori based cost-minimization of solutions (Rodgers 1976, 1990) as discussed in details in Cuesta et al. (2013). The novelty of this synergic IASI + GOME-2 observation is that it is capable of observing the horizontal distribution of ozone pollution plumes located below 3 km (Fig. 2) of altitude in the lower troposphere with a mean bias of 1% and a precision of 16% (Cuesta et al. 2013). IASI + GOME-2 retrieval output shows a higher degree of freedom of signal (DOFS) over the Indian region during 2018, 2019, and 2020, of average values 4.87, 4.85, and 4.88, respectively, with peaking averaging kernels, even in lower troposphere (Fig. 2). The tropospheric ozone column is also derived utilizing IASI and GOME-2 observations by integrating ozone profiles from surface to tropopause, where the thermal tropopause height was taken from IASI observations (Barret et al. 2020).

Fig. 2
figure 2

Typical averaging kernel and corresponding degree of freedom of signal (DOFS) for IASI + GOME-2 vertical ozone profile on 25 March 2020. Higher sensitivity or notable a priori independent ozone retrieval can be seen even at the lower troposphere

Full size image

Results and discussions

Impact on CO over Indian regions: observations from IASI and TROPOMI

The CO total column observations from TROPOMI re-gridded to 1° × 1° resolution were analyzed for the Indian lockdown period (24 March to 20 April) in 2019 and 2020 (Fig. 3). Higher concentrations of CO (> 3.0e18 molecule/cm2) were observed over central-eastern India, apart from western coastal regions. During the lockdown, a reduction of CO total column is seen over the Indo-Gangetic plain (IGP) and some parts of southern India and the southern Arabian Sea by more than 5%. In contrast, an increase (more than 20%) in CO total column is observed over the central and western regions, including the northern Arabian Sea (Fig. 3c). The increase in CO over central and west India seems to be contributed more from the coal-based thermal power plants (TPPs) around these regions; however, higher lifetime and influence from transport cannot be neglected. Surface-based observation also reported a relatively higher CO mixing ratio (141%) over a western site Dewas during the Indian lockdown (Navinya et al. 2020). Later, it will be discussed that there has been an increase in power generation and production capacities in the central-western thermal power plants during the year 2020.

Fig. 3
figure 3

TROPOMI total CO observed in a 2019 and b 2020 during lockdown period (24 March to 20 April) in India. c The percentage difference between 2020 and 2019 is also shown

Full size image

Figure 4 shows IASI CO vertical profiles over twelve cities and the four largest thermal power plants of India during 24 March to 20 April in 2018, 2019, and 2020. CO mixing ratios are generally found to be lesser (as low as 20%) in the boundary layer during 2020, while they are greater (more than 20%) above the boundary layer when compared with 2019 and 2018. This feature is more prominent in Western India (Ahmedabad, Mumbai, and Mundra TPP) apart from Hyderabad. Increase (5–15%) in CO is also observed over the northern regions, up to the middle troposphere, which reduces in the upper troposphere (Fig. 4 top panel). Near the surface, the maximum mixing ratio of CO (> 300ppbv) was observed over Hyderabad, Visakhapatnam, and most of the TPP. Similar to the profile observations, the CO total column obtained from TROPOMI (Fig. 3) shows an increase over western and central India and a decrease over northern India as compared to 2019.

Fig. 4
figure 4

IASI CO profiles over twelve Indian cities and four major thermal power plants during March 24 to April 20 that correspond to the lockdown period of 2020. Corresponding percentage differences with respect to 2019 in dotted blue and with 2018 dotted cyan lines are also shown. The number in the bracket shows the cities/TPPs location over India based on Fig. 1

Full size image

The increased CO mixing ratio at various altitudes and the increased CO total column over western and central India are suggested to be due to the increased production capacity of western-central TPPs. For example, the Chandrapur TPP (biggest TPP in Mumbai with a capacity of 3340 MW) had increased its Plant Load Factor (PLF, a measure of plant capacity utilization) by more than 3% during March 2020 compared to 2019 (https://npp.gov.in/dailyCoalReports); besides, fire counts were also higher (Fig. S3) over western regions (Fig. S3). It is observed that the fire counts over India (Fig. S3) have been greater during 2018 (2406), followed by the year 2019 (2172) and 2020 (2047). Furthermore, surface winds (Fig. 1) and HYSPLIT air trajectory (Fig. S1) show low-magnitude winds and dominant local circulation with little signatures of Indo-Gangetic Plain outflows that could further confine these emissions to the source regions of TPPs. Nevertheless, a persistent yearly increase of vertical CO profiles needs more detailed analysis.

Figure 5 shows relative changes in CO in four height layers, namely boundary layer (0–2 km), lower troposphere (2–5 km), middle troposphere (5–10 km), and upper troposphere (10–16 km). In general, CO shows a reduction in the range of 2–21% in the boundary layer (lower panel), except in western regions. However, CO shows a consistent increase (as high as 31%) in the lower troposphere, middle troposphere, and upper troposphere over the majority of Indian regions, including over the thermal power plants regions.

Fig. 5
figure 5

IASI CO percentage difference [(2020—ref_year)/ref_year]*100 at different altitudes with respect to 2019 and 2018. We have divided the atmosphere in boundary layer (0–2 km), lower troposphere (2–5 km), middle troposphere (5–10 km), and upper troposphere (10–16 km)

Full size image

Impact on the tropospheric NO2column over Indian regions

The average spatial distributions of the OMI tropospheric NO2 column during 24 March to 20 April in the years 2018, 2019, and 2020 are shown in Fig. 6. Tropospheric NO2 columns were seen to be significantly lower (more than 20%) in 2020 when compared with the same period in 2018 and 2019 over Indo-Gangetic plain, northern, central, and eastern India. However, significant increase (> 20%) over western India during 2020 was seen distinctly with respect to 2019.

Fig. 6
figure 6

Tropospheric OMI NO2 spatial distribution (re-gridded to coarser 1° × 1° resolution) observed over India during 24 March to 20 April in 2018, 2019, and 2020. Percentage changes in NO2 during 2020, with respect to 2018 and 2019, are also shown

Full size image

In addition to OMI, the spatial distribution of tropospheric NO2 columns was also studied using TROPOMI data (Fig. 7), the more recent and advanced instrument onboard Sentinel-5p with higher ground resolution. TROPOMI shows higher NO2 concentration over the central, eastern, and some parts of northern India during 2019, similar to OMI with very sharp source distributions, and exhibits the role of higher spatial resolution over OMI. TROPOMI also shows signatures of increase over the western region (~ 15%) (Fig. 7), which is also the downwind region of some of the major TPPs with slow winds and dominant local circulation (Fig. S1). Furthermore, the tropospheric NO2 column is also studied over the selected cities and TPPs (Fig. S2). The largest reduction was seen over the northern and southern cities (20–52%), with the highest reduction over metropolitan city Delhi, about 52% with respect to 2018, as observed by OMI. This reduction is lower when compared with respect to 2019 as observed by OMI (48%) and TROPOMI (31%).

Fig. 7
figure 7

Tropospheric NO2 spatial distribution observed from TROPOMI (re-gridded to coarser 1° × 1° resolution) over India for 2019 and 2020 lockdown time and their percentage difference

Full size image

During the lockdown, higher reduction of NO2 (> 20%) in comparison to 2018 was observed over most of the Indian region (Fig. 6). However, in comparison to 2019, the reduction was mostly over northern, eastern, and southern India. As discussed in the “Impact on CO over Indian regions: observations from IASI and TROPOMI” section, the fire counts have been higher in 2018 followed by in 2019 and 2020. Therefore, biomass burning could have contributed to higher NO2 in 2018, leading to display of greater reduction in NO2 during 2020.

Furthermore, higher tropospheric NO2 columns were observed over central and eastern India during 2018, which subsequently reduced in 2019 and 2020. The central-eastern region is home to various coal-based TPPs (i.e., Rihand TPP, Sipat TPP, Mouda TPP, Farakka TPP, Kahalgaon TPP), including India’s largest coal-based TPP (Vindhyachal super TPP, installed capacity 4760 MW). An annual report of National Thermal Power Corporation (NTPC) (https://www.ntpc.co.in/annual-reports/8842/download-complete-annual-report-2018-19, page no. 21 and https://www.ntpc.co.in/sites/default/files/downloads/44-final-NTPC-AR-30082020.pdf, page no. 35) shows higher gross power generation (megawatt) by more than 5% for 2018 as compared to 2019 for most of these central-eastern TPPs, which could be a possible reason for higher NO2 in central and eastern India during 2018.

Difference between OMI and TROPOMI over the Indian regions

Figure 8 shows the percentage difference between TROPOMI and OMI tropospheric NO2 columns over the Indian region. In general higher, differences in NO2 retrieval are observed over the remote location, e.g., the Himalayas, the Arabian Sea, the Bay of Bengal, and coastal areas, apart from central India. The spatial distribution of NO2 anomalies between both OMI and TROPOMI is more or less consistent in 2019 and 2020. The TROPOMI columnar NO2 is lower by more than 30% over the Himalayas, the Tibetan plateau, and central India. However, the TROPOMI NO2 column is higher by more than 40% over the Arabian Sea, the Bay of Bengal, and the coastal region apart from the Himalayan–Tibetan transition belt. A significant difference of more than 50% is also observed between OMI and TROPOMI over remote locations like Nainital in the central Himalaya and coastal cities during 2019 and 2020 (Fig. S2). Such differences between the TROPOMI and OMI over different parts of the world are also discussed in other studies (Bauwens et al. 2020; Wang et al. 2020); however, over India are discussed here.

Fig. 8
figure 8

Percentage difference of tropospheric NO2 column between OMI and TROPOMI (re-gridded to coarser 1° × 1° resolution) observations [((TROPOMI—OMI)/OMI)*100] during 24 March to 20 April in the year 2019 and 2020. Their frequency distribution is also shown in the right panel

Full size image

The possible reason for the difference between the two sensors could be the complex topography over the Himalaya, higher reflectivity of the Tibetan plateau, and cloudy condition over the oceanic regions, which induces notable differences in the retrieval outputs along with retrieval a priories and input parameters (Zhou et al. 2009; Lin et al. 2015). Additionally, space-based sensors are indirect and their retrieval relies on certain algorithms that reflect the higher contribution from retrieval a priories and possess larger errors over the remote locations having lower concentrations. Although OMI and TROPOMI have similar measurement times, the difference among the two sensors might be arising from different a priories (a priories NO2 profile based on TM5-MP in TROPOMI and GOES-Chem in OMI), other auxiliary information (e.g., surface reflectivity and cloud parameters) as well as differences in cloud masking and other sources of uncertainties in air mass factor calculations. Algorithms for OMI and TROPOMI also differ in the separation of the stratospheric and tropospheric components, which is a significant source of uncertainty in clean background areas. Furthermore, the frequency distribution of TROPOMI and OMI NO2 shows populated columns around 0.5e15 (molecule/cm2), and the values of populated columns are slightly higher for TROPOMI (Fig. 8, right panel). Furthermore, it can be seen that the number of populated columns (frequency) of about 0.5e15 (molecule/cm2) is higher during the lockdown period as expected in both TROPOMI and OMI observations.

Temporal variations in the tropospheric NO2and total CO column

The temporal variations in the percentage changes in the tropospheric NO2 column and total CO column over twelve cities and four thermal power plants are shown in Fig. 9. The tropospheric NO2 column decreased after lockdown over most of the cities and as high as 45% over Delhi. Similar reductions in NO2 were also reported worldwide, e.g., a reduction over China up to 54% (Xu et al. 2020), over Europe in the range 30–50% (Ropkins and Tate 2020), over the USA up to 36% (Shakoor et al. 2020), and in India in the range 30–70% (Singh et al. 2020; Biswal et al. 2020). Nevertheless, few Indian cities and TPP regions show some tendency of increase during the lockdown (Fig. 9). Furthermore, over some seaport cities like Chennai, Thiruvananthapuram, and Mumbai, where the significant anthropogenic contribution of NO2 is also coming from the international and domestic shipping, it shows lower NO2 concentration before the Indian lockdown and probably reflects the effect of restriction in international shipping before the lockdown began in India.

Fig. 9
figure 9

Temporal variations in percentage changes [(2020—2019)/2019)*100] in total CO column and NO2 tropospheric column for selected locations. The shaded gray area shows the strict lockdown period in India (24 March to 20 April 2020). To smoothen the time series, a three day’s moving average is used. The number in the bracket shows the cities/TPPs location over India based on Fig. 1

Full size image

The total CO column decreased (~ 5%) over northern regions immediately after lockdown and increased over Ahmedabad, Mumbai, Hyderabad, Mundra TPP, and Chandrapur TPP by more than 10%. However, no significant change was also seen over Kanpur, Jaipur, Kolkata, Visakhapatnam, and Vindhyachal TPP. Southern cities Bangalore, Chennai, and Thiruvananthapuram showed sharp reduction immediately after lockdown, but both CO and NO2 columns increased afterward. Such a feature is also seen at Nainital, a high-altitude place in the central Himalayas. Figure 10 shows a time series of tropospheric NO2 column and total CO column as observed by TROPOMI along with in situ measurement of surface ozone (Sarangi et al. 2014) in 2020. All these gases show a slight reduction just after lockdown but a systematic increase afterward. This increase has been very significant in NO2 (77%), ozone (67%), and CO (17%) during the lockdown period. Hence, it seems air pollutants in remote regions show a tendency to increase.

Fig. 10
figure 10

Time series of tropospheric NO2 column, total CO column from TROPOMI along with in situ measurement of surface ozone over Nainital in the central Himalayas. Red solid line shows daily average data and the light red dot shows 15-min data of surface ozone

Full size image

Impact on the tropospheric ozone profiles: synergic observation of IASI and GOME-2

The average spatial distribution of ozone during 24 March to 20 April in the years 2018, 2019, and 2020 utilizing synergic observation of IASI and GOME-2 at seven consecutive altitudes (2, 3, 4, 5, 6, 7, and 8 km) is shown in Fig. 11. In general, higher ozone (> 75 ppb) was observed over central India, extending to slightly western and southern regions (Fig. 11r, 11w, 11b1) in the 3 to 5 km altitudes during 2020. This feature is more prominent in 2018, which also extends to the north and northeast regions. The higher ozone concentration over central India decreased systematically with higher altitudes and diminished around 8 km with slightly lower ozone concentration (< 45 ppb), particularly in 2020. Over the Arabian Sea, high ozone concentrations of more than 75 ppb were observed, particularly at 4 km altitude during 2019 and 2020, which also decreased with higher altitudes. In the 2 km altitude, the ozone concentrations were lower than the adjacent 3 km altitude. The possible region of lower concentrations could be less reliable satellite observations near the surface.

Fig. 11
figure 11

Spatial distribution of tropospheric ozone at 2, 3, 4, 5, 6, 7, and 8 km altitude from IASI + GOME-2 observations during 24 March to 20 April in years 2018, 2019, and 2020 and their percentage differences are also shown at the two right columns

Full size image

Figure 11 also shows the percentage difference in ozone during 2020 with respect to 2018 and 2019. In comparison to 2018, a decrease of ozone by more than 30% was seen over the northern and northeastern regions, including Indo-Gangetic plain, but an increase (> 15%) is observed over the Arabian Sea and central-western India at all altitudes up to 8 km (second right panel). However, in comparison to 2019, an increase in ozone concentration was observed over central, western, and northern regions by more than 20% and a significant decrease (~ 20%) over the Indian Ocean (rightmost panel). The increase has been most notable in northern India, and the reduction of ozone over the Indian Ocean is more significant above 5 km altitude. The increased ozone concentration over central and western India in 2020, with respect to 2019, is lower at higher altitudes and it diminishes around 8 km. At the lower altitude region, an increase in ozone is in agreement with an increase in CO (Fig. 4). The increased ozone over the northern region, even at the higher altitudes, exhibits the role of dynamics. The average eternal potential vorticity (EPV) distributions at 600 hPa obtained from MERRA-2 reanalysis data showed a significant increase (> 30%) of EPV for 2020 as compared to 2019 (Fig. S4) over northern India and suggest the contribution of downward transport of ozone in 2020. In addition to satellite observations, balloon-borne observation of ozone over the central Himalayan site shows ozone enhancement episodes even at 3 km altitude during spring 2020, which is discussed in the “Ozonesonde observations over the central Himalayas” section.

Furthermore, the spatial distribution of average ozone, particularly at 3, 4, 5, and 6 km altitudes in 2018, 2019, and 2020, is more or less consistent with NO2 concentrations over India (Fig. 5). The ratio between tropospheric columns of formaldehyde (HCHO) and NO2 (Fig. 12) over India is studied to assess the ozone production regime. The ratios are mostly above 2.3 in 2019 and 2020. Other studies (Chang et al. 2016; Peralta et al. 2020) have shown that it is a VOC limited regime if HCHO/NO2 ratio is less than 1.5, while it is NOx limited regime if HCHO/NO2 ratio is greater than 2.3. Therefore, there have been strongly NO2 limited regimes over India during the lockdown period. A signature of VOC limited or transition regimes near the Vindhyachal coal-based thermal power plant and its surroundings is evident. It could also be noted that the ratio is further increased in 2020 as compared to 2019 (Fig. 12). Previous model-based studies (Kumar et al. 2012; Ojha et al. 2012) also showed the NO2 limited behavior of ozone during spring over the Indian regions.

Fig. 12
figure 12

HCHO/NO2 ratio from TROPOMI column observations over India. It shows a greater tendency of the NO2 limited regime during 2020 than in 2019

Full size image

Additionally, the average ozone profile up to 500 hPa (~ 6 km) over twelve populated cities and India’s four largest thermal power plants during 24 March to 20 April in 2018, 2019, and 2020 is also studied (Fig. S5). In comparison to 2018, a general decrease in ozone was observed except over Mumbai and Guru Gobind TPP. With respect to 2019, an increase of ozone up to 30% was observed over northern cities and a decrease over the southern cities. Further percentage changes of ozone concentrations in 2020 were studied with respect to 2018 and 2019 for four height layers (Fig. S6). In comparison to 2018, ozone concentration mostly decreased over all selected regions (Table S1) in all layers with relatively higher reduction over the northern cities and slightly lower reduction over the southern cities, except in the upper troposphere. In contrast to 2019, an increase of ozone over the northern, western regions, and thermal power plants (mainly in BL and LT) (Table S1) was seen, apart from a notable decrease (> 5%) in the upper troposphere and a reduction of ozone over the southern region is seen in all layers.

Figure 13 shows the average tropospheric ozone column over India during 24 March to 20 April in 2018, 2019, and 2020. In 2020, a uniformly distributed tropospheric ozone column of about 40 DU was seen over India; however, in 2018 and 2019, a slightly higher column (> 50 DU) over the northern Indo-Gangetic Plain and the northeast region was seen, which is more prominent in 2018. A considerable reduction of tropospheric ozone columns was seen in 2020 over the northern and northeast regions, which is significant (> 15%) compared with 2018. In contrast, an increase of more than 20% is observed over the northern region when compared with respect to 2019. An analysis of the tropospheric ozone column over twelve populated cities and the four biggest TPPs shows the highest column over Vindhyachal TPP (48 DU), Kanpur (46 DU), and Chandrapur TPP (43 DU) (Fig. S7) in the years 2018, 2019, and 2020, respectively. Mostly decrease of the tropospheric ozone column was seen over the Indian regions similar to the NO2 except for Nainital and Guru Gobind TPP, which are located in northern India, where apart from precursor’s photochemistry, downward transport significantly alters the ozone concentrations (Bhardwaj et al. 2018; Rawat et al. 2020).

Fig. 13
figure 13

Tropospheric ozone column over India during 24 March to 20 April in 2018, 2019, and 2020 from IASI + GOME-2 synergic observations. Changes in the tropospheric ozone column are also shown

Full size image

Ozonesonde observations over the central Himalayas

Figure 14 shows the average vertical ozone distribution (Ojha et al. 2014; Rawat et al. 2020) observed during March, April, and May over the central Himalayas. We could not conduct ozone soundings in April 2020 due to the nationwide lockdown. April 2020 could have shown the best effects of lockdown on ozone profiles, still the less restricted lockdown ozone profiles in May 2020 are studied by making four ozonesoundings (6, 13, 20, 27 May 2020). A successive increase in average ozone could be seen from March to May month in the average profiles (Fig. 14a). In general, during all 4 days in May, ozone distributions are found to be more or less within one sigma variation of May month’s average. However, ozone increase is seen on 27 May at around 3 km and 9 km and on 20 May at 5–7 km. Tropospheric NO2 showed a slight but insignificant decrease on 27 May. The influence of westerly wind seems to play an important role, particularly for 27 May 2020. Strong westerly winds (> 8 m/s and > 15 m/s) were observed near the high ozone altitudes (3 and 9 km) on 27 May 2020 (Fig. 14b). The HYSPLIT trajectory (Fig. S8) analysis also confirms the long-range transport of upper tropospheric air for this high ozone event. In all four ozone profiles of May 2020, higher ozone above 8 km was observed, which could be due to the contribution from the active downward transport over northern regions during spring (Ojha et al. 2014; Naja et al. 2016). Hence, dynamics is likely playing an important role in ozone enhancements, even at lower altitude (~ 3 km) regions.

Fig. 14
figure 14

Average vertical ozone distribution over the central Himalayas during March, April, and May (2011–2019). a Four-day ozone profiles during May 2020 and b wind observations are also shown. Here, wind speeds are shown by the color bar of the quiver and quiver arrows show wind direction

Full size image

Summary and conclusions

In 2020, a short-term improvement in the air quality near the surface had been reported from different countries due to the nationwide lockdown. Here, we have analyzed data from space-borne sensors (IASI, OMI, TROPOMI, and GOME-2) to study changes in vertical distribution and columnar values of ozone, CO, and NO2 during the nationwide lockdown in India that show some contrary results. Furthermore, we have used data of the year 2018, in addition to 2019, which has shown new results while minimizing the conclusions’ biases with respect to a particular year. Twelve populated cities and India’s largest coal-based thermal power plants were further selected to quantify the effects.

Overall significant reduction of NO2 (up to 52%) is observed over India, with some enhancement over the downwind regions of coal-based thermal power plants. Unlike NO2, ozone and CO concentrations show vastly varying influences of lockdown over India’s different areas and in the vertical distributions. In 2020, vertical ozone distribution showed an explicit increase (> 20%) over western and central India, including the Arabian Sea, and a decrease (~ 20%) over the southern coastal areas compared to both 2018 and 2019. Furthermore, a reduction (> 30%) compared to 2018 and an increase (as high as 110%) compared to 2019 are observed over northern India. Similar to the profile, tropospheric ozone columns also show a considerable reduction (> 15%) over the northern and northeast regions when compared with 2018 and an increase of more than 20% with respect to 2019.

An increase (~ 20%) in total CO is observed over central and western India and a decrease (~ 5%) over northern India during the lockdown period, which was consistent with changes seen in ozone. The vertical distribution in CO shows a systematic increase (as high as 31%) in the lower troposphere, middle troposphere, and upper troposphere over the majority of Indian regions, including over the thermal power plants regions. This feature was more prominent over the western regions.

The tropospheric NO2 columns significantly reduced over most of the Indian regions except the western part, where a notable increase was seen (~ 20%) during the lockdown. An increase in CO and NO2 during the lockdown period is suggested to be largely due to increased power production in coal-based thermal power plants (TPPs) that has also induced an increase in ozone. Furthermore, an analysis on the identification of ozone production regime showed an increased tendency of NO2 limited regime over major parts of India, while VOC limited regime over thermal power plants regions. The columnar changes in NO2 were more or less agreeable in both OMI and TROPOMI observations; however, notable differences (> 30%) are also seen over the remote areas, e.g., Tibetan plateau, the central Himalayas, and coastal regions.

Unlike surface-based studies, those have shown a clear decrease in pollutant levels, the present study shows an increase in CO, NO2, and ozone (decrease) at several locations and at different altitude regions. Lockdown had provided a good opportunity to assess the impact of the coal-based thermal power plants in India, when the majority of other emissions were at nominals. Furthermore, it has been shown that apart from the lockdown impact, the ozone concentration is also altered via dynamical variability, mainly over northern India. This study has used polar-orbiting satellite sensors that have limited temporal coverage over a specific region. It is expected that the recently launched GEMS mission onboard GEO-KOMPSAT-2 satellite will provide proficient observations with the diurnal variations of several trace gases and aerosols over the different Asian countries, including large part of India (Kim et al. 2020) that will lead to better assessments of emissions of many harmful atmospheric pollutants.