Introduction

India experiences severe air pollution tied to anthropogenic emissions1,2,3. Among these, anthropogenic chlorine emissions are an important yet underexplored factor influencing regional air pollution. Continental chlorine chemistry has gained significant importance in recent years due to its broad implications in tropospheric chemistry and air quality4,5,6. In polluted environments, the photolysis of reactive chlorine species (Cl2, HCl, HOCl, ClONO, ClNO2, CHCl3, CH2Cl2) releases highly reactive Cl radicals1,5,7,8. These radicals drive VOC oxidation, promoting secondary organic aerosol (SOA) formation, modifying hydrocarbon, and particulate matter (PM) levels, the lifetime of trace species, and thus altering atmospheric composition and oxidative capacity8,9,10,11. In urban environments, nighttime chlorine chemistry plays a crucial role in modulating surface ozone (O3) and fine particulate PM2.5 concentrations during morning hours3,12,13. \({NO}_{3}^{\cdot }\), a key night-time oxidant, exists in equilibrium with dinitrogen pentoxide (N2O5) which undergoes heterogeneous reactions with chloride-containing aerosols to form nitryl chloride (ClNO2) ((1) and (2)). Both \({NO}_{3}^{\cdot }\) and N2O5 accumulate overnight and enhance ClNO2 production. Acting as a nighttime reservoir, ClNO2 photolyzes at sunrise to release Cl radicals (3) and recycles NO2, influencing the early NOx-O3 cycle6,14.

$${\text{NO}}_{3}+{{\text{NO}}}_{2}\leftrightarrow {{\text{N}}}_{2}{{\text{O}}}_{5}$$
(1)
$${{\text{N}}}_{2}{{\text{O}}}_{5}+{{\text{Cl}}}^{-}\to {{\text{ClNO}}}_{2}+{{\text{NO}}}_{3}^{-}$$
(2)
$${\text{ClNO}}_{2}+{\text{h}}\nu \to {\text{Cl}}\cdot +{{\text{NO}}}_{2}$$
(3)

Abundant Cl ⋅ radicals initiate VOC oxidation, generating peroxy radicals (RO2 ⋅), surface O3, and SOA ((4)–(9)). This nocturnal chlorine chemistry is particularly important in polluted regions15,16,17.

$${\text{Cl}}\cdot +{\text{RH}}\to {\text{HCl}}+{\text{R}}\cdot$$
(4)
$${\text{R}}\cdot +{{\text{O}}}_{2}\to {{\text{RO}}}_{2}\cdot$$
(5)
$${\text{RO}}_{2}\cdot +{\text{NO}}\to {\text{RO}}\cdot +{{\text{NO}}}_{2}$$
(6)
$${\text{NO}}_{2}+{\text{h}}\nu \to {\text{NO}}+{\text{O}}$$
(7)
$${\text{O}}+{{\text{O}}}_{2}\to {{\text{O}}}_{3}$$
(8)
$${\text{RO}}\cdot +{{\text{O}}}_{2}\to {\text{Carbonyls}}+{\text{Peroxides}}+{\text{SOA}}$$
(9)

India is a global hotspot for chlorine emissions, including high concentrations of gaseous hydrochloric acid (HCl) and particulate chloride (pCl−)18,19. An assessment of chlorine emissions and their impact on Indian air pollution, however, remains limited. Sources include indoor biofuel combustion, waste incineration, industrial processes, crop residue and biomass burning, coal combustion, municipal solid waste, and brick kilns1,20,21,22. Chloride concentrations in Delhi and across the Indo-Gangetic Plain (IGP) frequently exceed those observed in other major global cities, particularly during winter and post-monsoon seasons1,18,22. Unlike Europe and North America, where continental chloride largely originates from marine sources via long-range transport, India’s atmospheric chloride is predominantly derived from inland anthropogenic HCl and pCl− emissions20,21,23,24. HCl emissions over India range from 40-220 Gg a−1 due to waste burning, having a mean of approximately 120 Gg a−125 whereas pCl− is estimated to be 250 Gg a−1, primarily from biomass (~ 68%) and waste burning (~ 21%)20. In comparison, China’s anthropogenic HCl and pCl− emissions were 458 and 486 Gg a−1, respectively, in 201426, increasing to 756 Gg a−1 in 201827. Combined HCl and pCl− emissions from India are the second-largest after China (~ 913 Gg a−1)21. For comparison, the United States emitted less than 100 Gg in 201428. Furthermore, India has high ammonia (NH3) rich atmosphere due to animal husbandry29,30, creating favorable conditions of rapid particulate ammonium chloride (NH4Cl) formation through HCl. This process enhances PM2.5 pollution and contributes substantially to urban haze and secondary aerosol growth, sometimes accounting for 40-50% of aerosol liquid water content19,31. Even so, chlorine chemistry remains poorly investigated as a component of India’s air pollution problem.

The objective of this study is to evaluate the impact of anthropogenic chlorine emissions on regional air pollution over India. To achieve this, we have used GEOS-Chem Chemical Transport Model in a high-resolution (0.25° × 0.3125°, ~ 25 km) configuration with hourly outputs for the full year of 2018. The details about the incorporated HCl and pCl− emission inventory in model (Fig. 1, Figs. S1–S2), detailed model setup and modeled pCl− validation (Fig. 2) are outlined in methods section. The analysis examines the role of chlorine emissions in regulating the concentrations of pCl−, PM2.5, maximum daily 8-hour average of ozone (MDA8 O3), and ClNO2. The presented findings show that anthropogenic chlorine emissions and multiphase chlorine chemistry have important roles in modulating regional air quality, particularly over densely populated and industrialized areas. From these results, the recommendation is to incorporate chlorine chemistry and region-specific emission inventories into air quality models of India to better assess the impact of pollution on climate and human health.

Fig. 1: Spatial distribution and sectoral contributions of anthropogenic chlorine emissions and sea-salt aerosol chloride flux over India.
Fig. 1: Spatial distribution and sectoral contributions of anthropogenic chlorine emissions and sea-salt aerosol chloride flux over India.
Full size image

a Annual anthropogenic chlorine emissions from GT-Chlorine Emission inventory and (b) pCl− emission flux from offline sea-salt aerosol extension in GEOS-Chem71. c Anthropogenic chlorine emissions (HCl + pCl−) are used from21 which includes contributions from agricultural residual burning, residential, energy, industry, and open waste burning, open biomass burning. Note the difference in scales between (a), (b). Pie charts and stacked bar plots represent the percentage (%) and absolute (Gg a−1) contribution of each sector over each region (Fig. S1) of India. Sub-sectoral percentage distribution of HCl and pCl− is shown in Fig. S2.

Fig. 2: Annually averaged modeled pCl− distribution over India and its validation with different observation datasets.
Fig. 2: Annually averaged modeled pCl− distribution over India and its validation with different observation datasets.
Full size image

Spatial distribution of annual average particulate chloride (pCl−, μg m−3) over the Indian region. Bar plots show the comparison between observed and model-simulated (Wi-AnthroHCl and Wo-AnthroHCl) pCl− concentrations at six monitoring stations, indicated by black arrows on the map. Delhi and Kanpur do not show a Wo-AnthroHCl bar because they are deeply inland sites, with no contribution from marine transport in the Wo-AnthroHCl simulation. These six sites, representing distinct regional markers, were used to validate the modeled pCl−. Campaign-averaged observations are compared against the corresponding modeled pCl− concentrations for the same periods.

Results

Model simulated HCl and pCl− over India

The spatial distribution of annual average HCl and particulate chloride (pCl−) concentrations, along with their relative changes between the sensitivity simulations of GEOS-Chem model, one including (Wi-AnthroHCl) and the other excluding anthropogenic HCl + pCl− emissions (Wo-AnthroHCl), reveals strong spatial heterogeneity across India (Fig. 3). The Indo-Gangetic Plain (IGP) emerges as the most prominent hotspot of HCl, followed by parts of western and southern India, as well as several other urban and industrial clusters (Fig. 3a). High HCl concentrations are largely associated with densely populated regions and intensive anthropogenic activities. In contrast, the spatial pattern of pCl− (Fig. 3d) shows weaker correspondence with that of HCl, although elevated levels persist over the IGP. Unlike HCl, pCl− does not exhibit the distinct hotspots observed over western and central India.

Fig. 3: Annual mean HCl and pCl− and their absolute relative differences between Wi-AnthroHCl and Wo-AnthroHCl simulations.
Fig. 3: Annual mean HCl and pCl− and their absolute relative differences between Wi-AnthroHCl and Wo-AnthroHCl simulations.
Full size image

Spatial distribution of annually averaged HCl (a, b) and pCl− (d, e) concentrations from Wi-AnthroHCl and Wo-AnthroHCl model simulations. c, f represent the annual averaged absolute difference for both HCl and pCl− when taking the difference of Wi-AnthroHCl and Wo-AnthroHCl model simulations for year the 2018.

In the absence of anthropogenic chlorine emissions, as represented by the Wo-AnthroHCl simulation (Fig. 3b, e), both HCl and pCl− concentrations become negligible over inland continental regions, with minor residual levels near coastal zones. These concentrations primarily result from marine influences, where HCl is formed through acid displacement reactions or dechlorination of sea-salt aerosols and ocean spray11,32. Depending on the seasons, the resulting HCl is transported inland, contributing to small but detectable levels of pCl− near coastal regions. Figure 3c and f, together with the relative changes shown in Fig. S3, highlight the substantial enhancement of both HCl and pCl− due to anthropogenic emissions, particularly over the IGP. The sea-salt-derived contribution, though evident, remains limited to the southern coastal regions of India, consistent with the spatial patterns shown in Fig. 1b and Fig. S4.

The Indo-Gangetic Plain (IGP) exhibits the highest concentrations of particulate chloride (pCl−), particularly over Delhi and West Bengal, where distinct hotspots are evident (Fig. 3d). The enhancement in pCl− arises from gas-to-particle partitioning of HCl in the model, which is highly sensitive to anthropogenic HCl emissions33. Seasonal variations Fig. (S5) indicate that pCl− concentrations are considerably higher in winter and autumn than in summer and spring, reaching a maximum seasonal average of 5.1 μg m−3 during winter. Seasonal anthropogenic chlorine emissions (Fig. S6) reveal relatively weak variability in total emissions across this region, with only modest sector-specific variations. These emission differences are small in magnitude and do not explain the strong seasonal cycle observed in pCl−. Instead, the pronounced wintertime enhancement of pCl− primarily reflects the combined effects of lower temperatures, higher relative humidity (RH), and greater aerosol liquid water availability, which shift the HCl partitioning equilibrium toward the particle phase, enhancing the conversion of gaseous HCl to pCl−34,35. Additionally, cooler winter and autumn conditions suppress the reverse conversion of pCl− to HCl, a process that becomes more efficient at the elevated temperatures typical of spring and summer12.

The elevated pCl− levels over the IGP can also be attributed to high chlorine emissions and ammonia-rich environments (Fig. S7), in which HCl readily dissolves into aerosol liquid water. This thermodynamically favors gas-to-particle partitioning through ammonium (NH4+) and pCl− formation1,19,36. In contrast, regions such as Punjab, Jammu & Kashmir, and the western coastal states of Gujarat and Maharashtra exhibit relatively high HCl concentrations (Fig. 3a) yet do not show correspondingly elevated pCl− levels (Fig. 3d). This is likely attributable to reduced partitioning efficiency, with pCl−/(HCl+pCl−) ratios below 0.4 (40%) and RH values ≤50% (Fig. S8). Under such drier conditions, limited aerosol liquid water content (ALWC) restricts the aqueous dissolution and ionization processes that drive NH4+ and pCl− formation1,31,36.

Interestingly, despite relatively lower HCl and NH3 concentrations, the southwestern coastal regions, particularly Kerala and coastal Karnataka, exhibit notably high pCl− abundance. This area also shows a significantly elevated pCl−/(HCl+pCl−) ratio (Fig. S8), indicating efficient partitioning into the particulate phase. In addition to moderate chlorine emissions relative to adjacent regions (Fig. 1), persistently high RH (70–80% throughout most of the year) strongly promotes such a chemical regime and behavior. Under such humid conditions, chloride partitioning increases sharply from approximately 0.4 at RH ≤50% to about 0.95 at RH ≥50%, enabling even low HCl concentrations to efficiently transition to the particle phase36. Moreover, the marine boundary layer in these regions contains substantial sea-salt aerosol contributions that supply additional chloride sources37,38. Elevated ALWC under coastal humid conditions enhances heterogeneous chemistry and aerosol growth31,39. This mechanism contrasts with the ammonia-driven partitioning prevalent in the IGP, demonstrating that high humidity and marine influences can sustain significant pCl− formation even in environments with limited NH3 availability19,37.

Chlorine emission-driven perturbations in key pollutants

Enhancement of ClNO2 via chlorine emissions

The spatial distribution of annually-averaged nighttime maximum ClNO2 concentrations highlights the strong influence of anthropogenic chlorine emissions on nighttime chemistry across India (Fig. 4). Here, ’nighttime maximum’ refers to the highest ClNO2 values obtained within the 21:00–09:00 local time window, which represents the period of ClNO2 accumulation during the night and early morning before rapid photolysis after sunrise. The inclusion of anthropogenic HCl emissions in the Wi-AnthroHCl simulation results in an annual mean nighttime ClNO2 enhancement exceeding 700 ppt over high-emission continental regions, whereas the eastern coastal areas remain dominated by marine chlorine sources even in the absence of anthropogenic chlorine inputs (Fig. 4b).

Fig. 4: Anthropogenic influence on nighttime ClNO2 concentrations over India.
Fig. 4: Anthropogenic influence on nighttime ClNO2 concentrations over India.
Full size image

Spatial distribution of annually averaged nighttime ClNO2 concentrations from GEOS-Chem simulations using an accumulation window from 21:00 to 09:00 local time. a shows the Wi-AnthroHCl simulation, highlighting the contribution of anthropogenic sources. b presents the Wo-AnthroHCl simulation, representing the background levels of ClNO2 driven mainly by natural or marine-influenced sources. c illustrates the difference (ΔClNO2) between Wi-AnthroHCl and Wo-AnthroHCl (a–b), isolating the net effect of anthropogenic HCl emissions on nighttime ClNO2. This comparison clearly reveals the regions where anthropogenic chlorine plays a dominant role in modulating nocturnal chemistry.

The primary nighttime formation pathway for ClNO2 involves the heterogeneous uptake of N2O5 on chloride-containing aerosols under NOx-rich conditions (Reactions (2) and (3)), which converts otherwise inert particulate chloride into highly reactive ClNO2, establishing an important nocturnal reservoir of chlorine. Its subsequent photolysis increases chlorine radicals (Cl ⋅) in the morning hours, which enhances the oxidation of VOCs and carbon monoxide (CO), producing peroxy radicals (RO2 and HO2) that accelerate photochemical ozone production6,40. Consistent with this mechanism, our global sensitivity simulations indicate that the inclusion of anthropogenic HCl emissions leads to an annual mean enhancement in OH of up to ~ 8% in some regions, as reported by previous studies11.

The highest concentrations and the largest spatial extent of ClNO2 are observed in winter and autumn, whereas lower levels are found in the spring and summer seasons (Fig. S9). Elevated ClNO2 concentrations during winter and autumn are primarily associated with higher levels of particulate chloride (pCl−) and N2O5 during these seasons. The aqueous-phase chloride concentration plays a crucial role in determining ClNO2 yield41. Additionally, the eastern coastal regions consistently exhibit higher ClNO2 concentrations than the western coast, underscoring the enhanced influence of chlorine chemistry along the coastal Bay of Bengal. The most significant positive changes are concentrated over the Indo-Gangetic Plain (IGP), coinciding with regions of high anthropogenic chlorine emissions.

By contrast, summer exhibits the lowest ClNO2 concentrations, primarily due to reduced availability of N2O5 and pCl−. These reductions occur because of unfavorable conditions for N2O5 formation during this season42, as rapid photolysis of the NO3 radical under extended daylight hours and elevated temperatures shifts the NO3-N2O5 equilibrium toward \({NO}_{3}^{-}\). Overall, the simulations reveal substantial enhancements in ClNO2 across much of the Indian subcontinent, highlighting the potentially significant role of anthropogenic chlorine emissions in modifying nocturnal oxidation chemistry. However, the lack of observational data in these regions underscores the urgent need for targeted field measurements to validate model predictions, reduce current uncertainties8, and better quantify the impacts of chlorine species on regional atmospheric composition.

Influence on surface ozone variability

The impact of anthropogenic chlorine emissions is reflected in the spatial patterns and seasonal variations of changes in maximum daily 8-hour average ozone (ΔMDA8 O3) across India. Figure 5 presents (a) the annually averaged MDA8 O3 in Wi-AnthroHCl simulation, (b) its absolute changes (ΔMDA8 O3) between the Wi-AnthroHCl and Wo-AnthroHCl simulations, and (c) the corresponding regional and seasonal relative changes. MDA8 O3 exhibits pronounced spatial heterogeneity and strong seasonal variability across the Indian subcontinent, with lower concentrations during autumn and winter and higher levels during spring and summer. This pattern reflects the strong dependence of ozone formation on solar radiation, as also shown by the absolute seasonal changes in Fig. S10.

Fig. 5: Annual and seasonal impacts on MDA8 O3 across India.
Fig. 5: Annual and seasonal impacts on MDA8 O3 across India.
Full size image

Spatial distribution of annually averaged (a) MDA8 O3 of model simulation having anthropogenic HCl emissions and (b) changes in ΔMDA8 O3 by taking the difference of Wi-AnthroHCl and Wo-AnthroHCl simulations, (c) Seasonal distribution of ΔMDA8 O3 across Indian regions. Each box represents the interquartile range (IQR) with the median line, whiskers extending to 1.5 × IQR. Seasonal variations are shown for all four seasons for each region of India as divided in Fig. S1.

Previous studies over China have reported annual mean ozone increases of 1.9 ppb (3.2%), 2.5 ppb (5.3%), and 2.0 ppb (4.1%) in response to anthropogenic chlorine emissions11,33,43. Over Europe, monthly mean ozone increases by 0.22 and 0.41 (1.2%) ppb in summer, with a 0.2–1.6 ppb increase over the Northern Hemisphere24,44,45, while in the United States, a 1–2 ppb increase is observed (in February and September), with even smaller changes (< 0.3 ppb, July) over New York46,47. However, the reported studies over Europe and the United States exclude detailed anthropogenic chlorine sources, and these impacts are mostly dominated by long-range transport of SSA23,24. Moreover, these studies primarily focus on the overall effects of chlorine-halogen chemistry, rather than explicitly isolating the contribution from anthropogenic chlorine emissions. Consequently, these ozone responses over Europe and the United States cannot be directly compared with the ozone responses to anthropogenic chlorine emissions, as investigated in the present study.

Our simulations reveal both positive and negative changes in ΔMDA8 O3, with substantial spatial and seasonal variability. Although the overall magnitude of the annual mean change remains modest in our analysis, distinct regional and seasonal features clearly emerge. The largest increases in ΔMDA8 O3 occur over the Indo-Gangetic Plain (IGP) during winter and autumn, which is consistent with the seasonal behavior reported for the Yangtze River Delta Region in China12. These increases are mainly driven by enhanced biomass and residential burning emissions of chlorine and the longer nighttime durations prevalent in high-emission regions11 combined with lower boundary layer mixing heights48. In contrast, southern India (SI) shows no significant positive change in any season. Spatially, the positive ozone changes correspond well with regions exhibiting higher ΔHCl and ΔpCl− values (Fig. 3), suggesting that elevated anthropogenic chlorine emissions enhance ozone formation; however, the change remains minimal. The seasonally averaged ΔMDA8 O3 across India is estimated to be −0.19% (−0.14 ppb) in spring, –0.47% (−0.23 ppb) in summer, –0.15% (−0.09 ppb) in autumn, and +0.28% (+0.17 ppb) in winter. The IGP shows the highest relative increases among all regions, with a maximum of +0.76% (+0.44 ppb) during winter (Fig. 5c).

Notably, regions with positive ΔClNO2 (Fig. 4c) exhibit concurrent increases in ΔMDA8 O3 (Fig. 5b), suggesting a strong spatial correlation between chlorine activation and ozone enhancement. Conversely, during summer, most regions including the IGP, exhibit negative ΔMDA8 O3 values, indicating that chlorine chemistry suppresses surface ozone. Similar behavior has been noted in previous studies12,43, where chlorine reactions were found to reduce ozone during the warm season. The reduced effectiveness of ClNO2 chemistry in summer arises from higher temperatures, shorter nights, and lower nighttime N2O5 levels, potentially making the ClNO2 pathway less significant23. Furthermore, under intense solar radiation, ClNO2 photolysis fails to efficiently recycle NO to NO2, instead enhancing NOx scavenging (via NO2 + OH → HNO3) and ozone titration, thereby lowering ozone production efficiency. Increased HO2 and Cl radical reactions also contribute to catalytic ozone loss by reacting with O3 to form HOCl, which photolyzes rapidly, leading to further ozone depletion13,23.

The effect of anthropogenic chlorine emissions on the diurnal cycle of surface ozone is illustrated in Fig. 6, which presents the averaged diurnal variations of O3 from Wi-AnthroHCl and Wo-AnthroHCl simulations, along with their differences (ΔO3) mapped over Central Pollution Control Board (CPCB) monitoring stations across India. The coefficients of determination (r2) between modeled and observed O3 are included for each diurnal cycle. Winter is the only season showing positive ΔO3, with the strongest enhancement over the IGP. The ΔO3 patterns indicate that anthropogenic chlorine emissions significantly increase morning-time O3 concentrations (09:00-11:00 local time), corresponding to the photolysis of nighttime-accumulated ClNO2 and subsequent release of Cl radicals. These radicals accelerate VOC oxidation, leading to early morning ozone peaks (reactions (5)-(9)). Consequently, under strong chlorine emission episodes, the daily ozone maximum may shift toward earlier hours, or an additional morning ozone peak may emerge alongside the typical afternoon maximum. Such temporal shifts in O3 peaks could have adverse health implications and alter the photochemistry of the environment during winter mornings over densely populated regions.

Fig. 6: Impact on wintertime diurnal variation in ozone.
Fig. 6: Impact on wintertime diurnal variation in ozone.
Full size image

Averaged diurnal variations in O3 and ΔO3 during the winter season from Wi-AnthroHCl and Wo-AnthroHCl simulations where ΔO3 is calculated as a difference between Wi-AnthroHCl and Wo-AnthroHCl simulations. The r2 represents the coefficient of determination when the modeled O3 is compared with surface observation of O3 provided by CPCB at each station.

Overall, our results indicate that the magnitude of tropospheric ΔO3 over India is smaller that reported for China, reflecting a weaker sensitivity to anthropogenic chlorine emissions (Fig. S11), where ozone exhibits a substantially stronger response to chlorine perturbations. This contrast suggests that India and China may operate under distinct chemical regimes of ozone formation with respect to chlorine chemistry. Such a complex and nonlinear spatiotemporal response of ozone to HCl emissions requires further investigations using chemical box models and will be covered in follow-up studies. This will help to elucidate as to why, despite the IGP being one of the most polluted places across the globe, it exhibits relatively lower average ozone responses to chlorine emissions as compared to similar polluted sites in China11. Therefore, further high-resolution modeling and observational studies are needed to elucidate the underlying chemical mechanisms governing these regional differences.

Impact on PM2.5 and its components

The influence of anthropogenic chlorine emissions on fine particulate matter over India is evident from the annually averaged differences in PM2.5 concentrations between the Wi-AnthroHCl and Wo-AnthroHCl simulations (Fig. 7). The spatial distribution of ΔPM2.5, along with the box plots representing respective changes in each region across India (regional divisions shown in Fig. S1), illustrates consistent increases in PM2.5 concentrations following the inclusion of chlorine emissions in all four seasons. The most affected regions include the Indo-Gangetic Plain (IGP), parts of northern India (NI), and Kerala in southern India (SI).

Fig. 7: Regional and seasonal sensitivity of ΔPM2.5 to anthropogenic chlorine emissions across India.
Fig. 7: Regional and seasonal sensitivity of ΔPM2.5 to anthropogenic chlorine emissions across India.
Full size image

Spatial distribution of annual averaged ΔPM2.5 and seasonal variability in the sensitivity of ΔPM2.5 concentrations to chlorine emissions. Box plots show the regional changes in ΔPM2.5 (μg m−3) derived from the difference between Wi-AnthroHCl and Wo-AnthroHCl simulations. Each box represents the ΔPM2.5 within a specific region for each season, highlighting both the magnitude and spread of the response to anthropogenic chlorine emissions. Note the difference in scales across box plot panels to better capture the variability in regions with smaller ΔPM2.5, as using a uniform scale would mask these variations.

Seasonal variations in ΔPM2.5 (Fig. S12) reveal that the largest increases in PM2.5 occur during winter (+ 2.09 μg m−3, +2.71%), followed by autumn (+ 1.96 μg m−3, +2.67%). Spring exhibits the lowest enhancement (+ 0.07 μg m−3, +0.28%), primarily affecting northern India. The annual mean PM2.5 rises by + 1.18 μg m−3 (+1.64 %) nationally, with the greatest increase over IGP (+ 2.9 μg m−3, +2.6%), particularly around New Delhi and West Bengal, followed by NI (+ 1.2 μg m−3, +4%). Although these values may appear modest when averaged annually, the box plots in Fig. 7 indicate strong seasonal variability. During winter, the enhancement reaches + 5.1 μg m−3 (+4%) over the IGP and + 1.75 μg m−3 (+7.3%) over NI. The spatial distribution of annual relative changes is further presented in Fig. S13.

To better understand the composition driving these increases, Fig. 8 shows the spatial distribution of annually averaged changes in individual PM2.5 components (pCl−, NH4+, \({{SO}_{4}}^{2-}\), \({NO}_{3}^{-}\)) derived from the difference between the Wi-AnthroHCl and Wo-AnthroHCl simulations. The results indicate that pCl− and \({NH}_{4}^{+}\) are the primary contributors to the overall PM2.5 enhancement, with annual average increases of + 0.91 μg m−3 and + 0.41 μg m−3, respectively, across the Indian region. These species also exhibit notable seasonal dependence (Figs. S14–S16).

Fig. 8: Impact of anthropogenic chlorine emissions on the chemical components of PM2.5.
Fig. 8: Impact of anthropogenic chlorine emissions on the chemical components of PM2.5.
Full size image

Annual changes in the chemical components of PM2.5 arising from the influence of anthropogenic chlorine emissions. The panels present the differences (Δ) between Wi-AnthroHCl and Wo-AnthroHCl simulations for particulate chloride (pCl−), sulfate (\({{SO}_{4}}^{2-}\)), ammonium (NH4+), and nitrate (\({NO}_{3}^{-}\)).

Enhancement in \({NH}_{4}^{+}\) concentrations over the region is due to reactions of excess NH329 with HCl. Conversely, nitrate (\({NO}_{3}^{-}\)) concentrations decrease in all seasons except summer (Fig. S14), primarily due to the suppression of N2O5 hydrolysis (reaction 1, in supplement) by elevated pCl− levels. Instead of hydrolysis, N2O5 reacts with chloride to form ClNO2 (reaction 2, in supplement), thereby reducing the overall efficiency of nocturnal nitrate formation49,50,51. During summer, however, \(\Delta {NO}_{3}^{-}\) becomes positive (Fig. S15) because ClNO2 chemistry is less active, allowing nitrate to form via heterogeneous N2O5 uptake52.

In contrast, sulfate (\({SO}_{4}^{2-}\)) concentrations show relatively greater changes than nitrate, with a decrease observed across most regions of India (Fig. 5, Fig. S16). This reduction in \({SO}_{4}^{2-}\) arises from the competition between HCl and H2SO4 for ammonia in aerosol thermodynamic system, whereby the presence of HCl drives ammonia preferentially towards NH4Cl formation19 (reaction 3-4, in supplement). Consequently, less NH4+ is available for sulfate neutralization, leading to lower \({SO}_{4}^{2-}\) concentrations.

Discussion

This study provides a comprehensive assessment of anthropogenic chlorine emissions and their impacts on regional air chemistry over India, demonstrating that chlorine is a key, non-linear, and previously underrepresented driver of regional air quality. Over the IGP, home to ~ 700 million people, anthropogenic chlorine increases wintertime PM2.5 by up to + 5.1 μg m−3 (+4%), aggravating exposure to fine particulates associated with respiratory and cardiovascular diseases and premature mortality53,54. Chlorine-rich aerosols also contribute to visibility degradation by enhancing aerosol water uptake and hygroscopic growth. Future studies should examine the effects of chlorine-containing particles on aerosol optical properties and their role in wintertime haze formation over the region.

The comparatively weaker and non-linear ozone response to chlorine emissions compared to China indicates distinct chemical regimes across India and the need for high-resolution regional and box-model studies to elucidate chlorine-driven ozone production and loss. This also implies that findings from East Asian studies cannot be directly applied to South Asia, underscoring the importance of region-specific assessments. Enhanced nighttime formation of ClNO2 shifts peak ozone to early morning hours, coinciding with commuting periods and shallow boundary layers, potentially increasing human exposure.

Current emission inventories strongly underestimate inland chlorine sources and poorly constrained emissions such as Cl2, HOCl, chlorinated VOCs, plastic burning, brick kilns, and steel pickling create large uncertainties in chlorine budgets. Given strong spatial heterogeneity and chemical regimes, region-specific inventories and improved representation of primary and multiphase secondary chlorine-VOC chemistry are urgently required to better assess oxidation capacity and secondary organic aerosol formation.

Future efforts must explicitly address chlorine emission control. Reducing NH3 emissions to limit pCl− production is unrealistic in agriculturally dominated regions; direct control of primary HCl sources is therefore more feasible. Effective mitigation requires stricter enforcement against open burning, improved waste management, and regulation of industrial emissions. Experience from China shows that cleaner fuels, stronger industrial standards, and the adoption of HCl scrubbers and flue-gas desulfurization systems ( > 95% removal efficiency) can substantially reduce reactive chlorine emissions. Future studies should prioritize quantifying HCl emissions from informal sectors, waste management, and plastic burning to support mitigation strategies.

Looking ahead, targeted observations of HCl, pCl−, Cl2, N2O5, and ClNO2 using chemical ionization mass spectrometry (CIMS) and high-resolution aerosol mass spectrometry (AMS) across inland, coastal and urban regions, especially in winter, are critically important to constrain nocturnal processes, morning photochemical transitions and model uncertainties.

Methods

GEOS-chem model setup

We used the GEOS-Chem chemical transport model (v14.4.2; https://zenodo.org/records/12807579) which contains a detailed NOx-Ox-VOC-PM-halogen (Br-Cl-I) chemistry55. GEOS-Chem was driven by offline assimilated meteorological fields from GEOS-FP (Goddard Earth Observing System- forward processing) having 47 vertical hybrid pressure-sigma levels up to 0.01 hPa with native horizontal resolution of 0.25° × 0.3125° provided by NASA’s Global Modeling and Assimilation Office (GMAO). For the given study, we run the GEOS-Chem model both on a global scale and then at a nested domain over India using its standard full chemistry simulation with a simple SOA (secondary organic aerosol) scheme which follows fixed yield approach for SOA formation in the model56,57. The dynamical boundary conditions generated every 3 h from a global simulation (2° × 2. 5°) were used to conduct nested-grid simulations (0.25° × 0.3125°, ~ 25km) at every hour over the Indian region (5° S-45° N, 55°-105° E) for one full year of 2018, following 6 month spin-up time. GEOS-Chem has been previously used and showed a reasonable reproducibility of pollutants over Indian region and provides a wide range of research topics in atmospheric chemistry and air quality58,59,60,61.

GEOS-Chem has a detailed chlorine chemistry mechanism developed by refs. 11,16 and heterogeneous reactions of N2O5 − ClNO2 mechanism based on ref. 62. Tropospheric halogen chemistry, as simulated in previous studies16,63, was extended by ref. 11 to include chloride mobilization from sea-salt aerosol (SSA) through HCl acid displacement and other heterogeneous processes. The mechanism treats both fine-mode ( < 0.5 μm) and coarse-mode ( > 0.5 μm) pCl− and detailed gas-phase inorganic chlorine species16,27. In the default GEOS-Chem configuration, the primary source of tropospheric chlorine is governed by pCl− release from SSA in fine and coarse modes11,27. Chlorine Emissions are added as HCl and ISORROPIA-II repartitions between particle and gas phases are included as part of the H2SO4 − HCl − HNO3 − NH3 − NVCs thermodynamic system11,34.

Emissions

Emissions in the GEOS-Chem model are implemented through the HEMCO (Harvard-NASA Emissions Component) module, which provides flexible configurations for different emission types64. The global anthropogenic emissions of SO2, NOx, CO, NMVOC and NH3 were taken from the Hemispheric Transport of Air Pollution (HTAP_v3-2018) inventory (https://edgar.jrc.ec.europa.eu/dataset_htap_v3). Biogenic emissions of VOC were taken from MEGAN v2.165 and Global Fire Emissions Database version 4 (GFED4) was implemented for biomass burning emisisons (ref. 66). Continental inorganic chlorine emissions were taken from a recently developed global HCl emission inventory (GT-Emission Inventory) gridded at a spatial resolution of 0. 1° × 0. 1° by21. This inventory provides continental chlorine emissions in the form of HCl and pCl−for 6 different sectors as mentioned in Fig. 1. India contributes 380 Gg a−1 to the total global HCl emissions of 4731 Gg a−1 with 213 Gg a−1 from the residential sector, 76 Gg a−1 from open waste burning, 49 Gg a−1 from energy, 25 Gg a−1 from industry, 13 Gg a−1 from agricultural residue burning, and 8 Gg a−1 from open biomass burning. These sectors consider sources of HCl and pCl− from coal combustion, solid waste and biomass combustion, indoor biofuel, coke and brick production. Sub-sector wise percentage distribution of HCl and pCl− over the Indian region is shown in Fig. S2, more details can be found in21. For this study, we performed two simulations using the GEOS-Chem model, one including anthropogenic HCl + pCl− emissions (Wi-AnthroHCl) and another excluding anthropogenic HCl + pCl− emissions (Wo-AnthroHCl).

Observations and model evaluation

We used multiple data sets of observations for the GEOS-Chem performance evaluation. We compared the modeled pCl− with observations for six different locations in India, each from different campaigns. These locations are Delhi (28.61 °N, 77.23 °E), Chennai (13.08 °N, 80.27 °E), Ahmedabad (23.03 °N, 72.58 °E), Kanpur (26.45 °N, 80.35 °E), Mahabaleshwar (17.92 °N, 73.65 °E) and Munnar (10.09 °N, 77.06 °E). Out of six locations, pCl− for four are obtained from published sources1,22,67; (Table S1). It should be acknowledged that, due to the limited availability of chloride measurements in the Indian region, some of the observational datasets used for comparison are drawn from different years, ignoring the potential uncertainties related to interannual variability. In India, ClNO2 had been measured only in Delhi68, our model results were evaluated against its published measurements (Fig. S17). We used hourly data of O3 and PM2.5 provided by the Central Pollution Control Board (CPCB, https://airquality.cpcb.gov.in/ccr/#/caaqm-dashboard-all/caaqm-landing/, last access on 20 June 2025). Model-simulated mean surface concentrations from the nearest grid cell are compared with observations (Table S2).

Figure 2 presents a comparison between simulated pCl− concentrations and observations from six distinct field campaigns across different regions of the Indian subcontinent. The model successfully captures the spatial variability and magnitude ranges of pCl− at these sites (Table S3), reflecting diverse regional characteristics. However, pCl− is overestimated at western coastal sites, where strong marine influence already leads to high background chloride levels. The addition of anthropogenic HCl, together with high relative humidity and potential overestimation of sea-salt aerosol emissions in GEOS-Chem under high-wind coastal conditions69,70, likely enhances chloride retention in the particle phase. In contrast, the model tends to underestimate pCl− concentrations at high-emission locations like Delhi and Kanpur. This discrepancy may arise because the model configuration does not include emissions from certain chlorine-containing species such as Cl2, CH2Cl2, and CHCl3, which are predominantly emitted from industrial activities dominant in these regions. Additionally, the current model simulation lacks chlorine emissions from other key sources such as plastic burning and steel pickling industries and the GT-Emission inventory reportedly underestimates biomass burning emissions in India by approximately an order of magnitude ( ~ 10 ×) as documented by ref. 3. To compensate for these underestimations, we applied the scaling factor to anthropogenic HCl emissions, which significantly improved the agreement between the modeled and observed pCl− concentrations.