The work presented here was conducted at three sites – the BCOB (22.17° N, 90.71° E), the MCOH (6.78° N, 73.18° E), and the MCOG (0.69° S, 73.15° E) – from early December 2017 to the end of March 2021. The BCOB is located on Bhola Island (also called Dakhin Shāhbāzpur Island) in the delta of the Bay of Bengal, about 300 km south of Dhaka, Bangladesh (Ahmed et al., 2018; Shohel et al., 2018; Dasari et al., 2019). The MCOH is located in the northern part of the island of Hanimaadhoo (Thiladhunmathi Atoll), which covers around 3.1 km² and has around 1800 inhabitants (Corrigan et al., 2006; Höpner et al., 2016; Budhavant et al., 2023). Measurements are taken from a tower platform at 15 m above sea level, from which air samples are directed to a ground-level, air-conditioned laboratory (Corrigan et al., 2006; Budhavant et al., 2018, 2023). The MCOG is located on the southernmost island of the Maldives, at the Equator, 500 km south of the capital city (Malé) and 800 km from the MCOH (Corrigan et al., 2006; Ramanathan et al., 2007). Detailed descriptions of these observatories are available in earlier publications (Corrigan et al., 2006; Stone et al., 2007; Dasari et al., 2019). Aerosol PM_2.5 samples were collected using 150 mm diameter quartz filters pre-combusted at 450 °C (Merck Millipore), employing high-volume samplers (DH-77, Digitel Elektronik AG) at 500 L min⁻¹. Blank filters were shipped, stored, and processed identically to the samples. Each of these three observatories was instrumented to record spectral data on aerosol optical depth (AOD) as part of the AErosol RObotic NETwork (AERONET) (Holben et al., 1998; Ramanathan et al., 2005; Nair et al., 2023).

The aerosols were analyzed for several carbonaceous components and major ions using standard protocols and suitable techniques (Dasari et al., 2019; Budhavant et al., 2023). The mass concentrations of elemental carbon (EC; here referred to as BC), OC, and total carbon (TC = BC + OC) were measured with a thermal–optical transmission analyzer (the Lab OC-EC Aerosol Analyzer from Sunset Laboratory) using the NIOSH 5040 method from the National Institute for Occupational Safety and Health (Birch and Cary, 1996; Budhavant et al., 2015, 2023). NIST-traceable laboratory standards (Reference Material 8785) were used to verify the accuracy of the OC, EC, and TC measurements. No detectable signal was observed for BC in field blanks. The OC concentration values were blank-corrected by subtracting an average field blank (5 % of sample signals). Following the established protocol, water-soluble organic carbon (WSOC) was measured using a Shimadzu TOC-VCPH analyzer (Kirillova et al., 2010, 2013; Budhavant et al., 2020).

Another portion of each aerosol filter was extracted with 18 MΩ cm Milli-Q water for the analysis of water-soluble inorganic ions using a Dionex Aquion ion chromatography (IC) system (Thermo Scientific). The system contains a guard column and an anion–cation separator column with a primary exchange resign and suppressor column (AERS 500/CERS 500). The quality of the data was tested with internal and external reference samples. The analytical error was less than 4 % for the anions and 5 % for the cations. A more detailed description can be found in Budhavant et al. (2023).

The relationship between atmospheric concentration and direct radiative forcing by BC is characterized by its mass absorption cross section (MAC). The laser beam (678 nm) of the Sunset Laboratory aerosol carbon analyzer was used to measure the light attenuation (ATN = -ln⁡(I/I0)) of the aerosols on the filter (Ram and Sarin, 2009). The MAC for BC (MAC_BC) is calculated as follows (Weingartner et al., 2003; Budhavant et al., 2020): 1MACBC=ATNBCloading⋅MS⋅R(ATN), where MS is an empirical multiple-scattering correction factor implemented in most filter-loading correction schemes. To account for the multiple-scattering effects, a factor of 4.5 was selected for estimation (Budhavant et al., 2020). Correction for non-linearity when measuring light absorption through a filter is denoted by R. 2R=11.114-1ln⁡ATN-ln(0.1)ln⁡0.5-ln(0.1)+1 The spectrophotometer measured the light absorption of the aerosol extracted from water. Subsequently, the MAC was calculated for water-soluble BrC (WS-BrC). 3MACWS-BrC=babs,365WSOC, where WSOC is the water-soluble organic carbon concentration and babs,365 is the absorption coefficient at 365 nm. The absorption Ångström exponent (AAE) was estimated as the slope in a linear regression of the logarithm of the absorption coefficient (babs) versus the logarithm of the wavelength (λ). 4lnbabs(λ)=-AAE⋅ln⁡λ+intercept The AAE was fitted between 330–400 nm to avoid interference from other light-absorbing solutes, such as ammonium nitrate, sodium nitrate, and nitrate ions (Cheng et al., 2011; Bosch et al., 2014).

The radiative forcing of aerosol particles is a major uncertainty factor in understanding the Earth's climate (Ramanathan et al., 2007; IPCC, 2021; Lu et al., 2023). The radiative implications of aerosols are quantified in terms of their direct aerosol radiative effects (DAREs). Spectral aerosol-optical-depth (AOD) data from three stations associated with the AErosol RObotic NETwork (AERONET) (Hess et al., 1998; Bedareva et al., 2014), ozone data from the Ozone Monitoring Instrument (OMI), water vapor and surface reflectance data from the Moderate Resolution Imaging Spectroradiometer (MODIS), and further surface reflectance data were used in this study. The aerosol optical model (version 3.1 of OPAC (Optical Properties of Aerosols and Clouds)), which works based on Mie scattering theory (Hess et al., 1998), was used to estimate the optical properties of newly defined aerosol mixtures (Hess et al., 1998). AOD measurements from sun photometers, single scattering albedo (SSA), and asymmetry parameters modeled using the Mie scattering model were used as inputs for the Santa Barbara DISORT Atmospheric Radiative Transfer (SBDART) model (Ricchiazzi et al., 1998; Lu et al., 2023). This model uses a complex discrete-ordinate method to numerically integrate the radiative-transfer equations (Stamnes et al., 1988). A detailed description of this model and approach is available elsewhere (Ramanathan et al., 2005; Satheesh, 2002; Nair et al., 2023).

This study examines air mass back trajectories to identify potential sources of BC and other aerosol components arriving at the MCOG station, which is even further away from the source regions than the MCOH. Given the long-distance travel involved, the focus on the dry season (increasing longevity), the understanding that BC has a longer lifetime than other aerosols, and the experiences from earlier studies (Budhavant et al., 2020, 2023), a back-trajectory (BT) time horizon of 10 days was selected as appropriate. The air mass back trajectories (AMBTs) were generated at an arrival height of 50m at all three sampling sites (Figs. S1–S4 in the Supplement) using the NOAA Hybrid Single-Particle Lagrangian Integrated Trajectory model (version 4) (Draxler and Hess, 1997; Draxler, 1999). This study's calculations were based on meteorological data from the Global Data Assimilation System (GDAS), which is run four times daily (at 00:00, 06:00, 12:00, and 18:00 UTC). These individual trajectories were clustered into different geographical regions (Fig. 1). The MODIS (Moderate Resolution Imaging Spectroradiometer) satellite-derived FIRMS (Fire Information for Resource Management System) fire-count data were combined with cluster analysis to understand the impact of biomass-burning emissions from potential source regions observed during the sampling period (Fig. 1).

The AMBTs, AOD measurements, and active fire data were used as parameters to study atmospheric transport and geographical source regions in the area. Based on atmospheric transport, we defined two temporal source domains: the influence of the heavily polluted Indo-Gangetic Plain (IGP) region (from 18 December 2017 to 8 February 2018) and the total period of the study. Measurements obtained at the BCOB represent an accumulation of IGP sources through air mass transport across northern Pakistan, northern India, and Bangladesh – a region containing many large cities and megacities, areas of heavy industrialization, and rural areas with extensive agricultural burning (Figs. 1 and S2). The MCOH and the MCOG are situated in the northern Indian Ocean and thus intercept long-range pollutant emissions from southern Asia, including the IGP, the western part of India, and the Indian Ocean (Figs. S3 and S4). Occasionally, winds in the IGP sector come from southern India or the Bay of Bengal. However, during winter, polluted winds from the IGP can reach the Bay of Bengal, leading to similar signals being detected over the MCOH and MCOG regions. This is particularly noticeable during synoptic observations. Cluster analysis of AMBTs, combined with AOD data, satellite measurements, and aerosol chemical composition, demonstrated that the wintertime northern Indian Ocean is greatly influenced by anthropogenic aerosols transported from source regions like the IGP and the western margin of India.

In general, varying primary and secondary sources, combined with the short atmospheric residence times of aerosol particles containing a high fraction of organic carbon, result in large regional differences in chemical composition, morphology, mixing state, size, and optical properties. OC was the main component of carbonaceous aerosols during the southern Asian winter, accounting for 85 ± 5 % of total carbon (TC) at the BCOB, 66 ± 9 % at the MCOH, and 67 ± 9 % at the MCOG. The OC contribution to TC was highest when the wind came from the IGP (Fig. 2) and lowest when the wind traveled over oceanic regions at all three sites. BC and OC are very well correlated (R>0.74) at all three sampling sites, indicating similar source emissions. However, the average ratio of OC to BC was 6.5 ± 2.1 at the BCOB, decreasing markedly to 2.2 ± 1.1 at the MCOH and 2.4 ± 1.8 at the MCOG. This large decrease from the exit of the IGP source region (BCOB) to the Indian Ocean receptor sites (MCOH and MCOG) demonstrates that OC/BC ratios were strongly affected by selective processing and/or washout of OC during long-range transport (LRT). The atmospheric lifetime of OC is typically shorter than that of BC in this region (Budhavant et al., 2020). Since OC represents a more complex mixture, it is subject to more atmospheric transformation than BC, as reflected in the comparatively large shift in the stable-isotope fingerprints of the OC component from the source to the receptor sites in this region (Dasari et al., 2019; Bosch et al., 2014; Kirillova et al., 2016). The highest concentrations of BC, OC, and nss-SO42- aerosols were associated with air masses from the IGP and the western margin of India.

The water-insoluble fraction of BrC exhibits a higher absorption rate per unit mass than the WS fraction (Liu et al., 2013; Cheng et al., 2016). We observed that the WSOC comprised a large but declining proportion of the overall OC (Fig. 3). The WSOC/OC ratio changed throughout the study, as shown in Fig. 4c. At the BCOB, the WSOC/OC ratio (0.35 ± 0.06) was less variable, indicating that the sources of both types of carbon were similar. Furthermore, we discovered a strong correlation between WSOC and OC concentrations at the BCOB (r=0.95; p<0.001). However, the lower WSOC/OC ratios at the MCOH (0.21 ± 0.1) and the BCOB (0.16 ± 0.1) suggest a higher contribution of water-insoluble OC at these locations and times. Carbonaceous aerosols derived from fossil fuel combustion may be less water-soluble (WSOC ≥ 20 %) due to the presence of fewer oxygenated organic moieties (Ruellan and Cachier, 2001). The mass fraction of WSOC relative to OC was observed as an indicator of aerosol photochemical processing in the atmosphere (Dasari et al., 2019). Overall, the decreasing OC/BC ratio from the IGP exit to after transportation over the ocean indicates that selective washout and bleaching reactions of organic carbon occurred.

The chemical composition of the aerosols changed both between sites and over time (Table 1; Fig. 2). Filter samples were characterized in terms of major anions (Cl⁻, NO3-, and SO42-) and major cations (Na⁺, K⁺, Mg²⁺, Ca²⁺, and NH4+) for the 4-month sampling period (Fig. 4). The highest concentrations of ions were recorded at the BCOB, which is expected as this site is situated at the outflow of the highly polluted IGP. However, the concentrations of SO42- and NH4+ followed a different pattern. Higher values for SO42- and NH4+ were observed at the MCOH. These MCOH-intercepted pollutants were traced to India's central and eastern regions and the IGP through AMBTs. The IGP region and its surrounding areas are hotspots for sulfur dioxide (SO₂) emissions, mainly due to the presence of multiple thermal power plants, construction industries, and petroleum refineries. These sources contribute to the region's SO₂ and nitrogen oxide (NO_x) levels (Guttikunda and Jawahar, 2014; Kuttippurath et al., 2022). Furthermore, a previous study conducted at the MCOH found that dimethyl sulfide (DMS) accounts for only up to 3 % of nss-SO42- in polluted air (Granat et al., 2010). The IGP is a hotspot of high anthropogenic aerosol loading due to intense agricultural crop residue burning, biomass burning, open waste burning, industrial activities, and high urbanization (Dasari et al., 2020; Ansari and Ramachandran, 2023; Nair et al., 2024).

To identify the effect of marine influences on aerosol composition, sea salt corrections were calculated using Na⁺ as the reference element (Keene et al., 1986). The nss-SO42- fraction relative to total sulfate was observed at the BCOB (99 ± 1 % (mean ± standard deviation)), the MCOH (98 ± 1 %), and the MCOG (86 ± 13 %), indicating significant contributions of SO42- and SO₂ from diesel combustion and coal-fired power plants in India and Bangladesh. Some nss-SO42- observed at the MCOH may have been due to ocean traffic over the northern Indian Ocean as the majority of shipping emissions result from fuel combustion that releases SO_x (sulfur oxide) and NO_x directly into the atmosphere (Corbett and Koehler, 2003; Gopikrishnan and Kuttippurath, 2021).

The BCOB, located near the IGP source region, was found to have a high Cl-/Na+ ratio (4.7 ± 3.5; Fig. 4) compared to the other two receptor sites (Fig. 4). This implies that, at the BCOB, a significant amount of total Cl⁻ comes from anthropogenic activities. The particulate Cl⁻ might come from the burning of plastics containing chlorine, such as polyvinyl chloride (PVC), during open waste burning (Pathak et al., 2023). This can lead to the formation of Cl⁻ radicals, which impact the oxidation capacity of polluted air. Marine aerosols often experience chlorine depletion, and releasing gas-phase HCl from particles can impact aerosol acidity and the concentration of water-soluble ions. However, once these particles enter the atmosphere, they become exposed to various pollutants, leading to the loss of particulate chlorine as it transitions into the gaseous phase. This loss of chlorine is typically attributed to ion exchange reactions with atmospheric acids like SO₂, H₂SO₄, and HNO₃, which result in the formation of sulfates and nitrates, as well as the degassing of HCl (Orsini et al., 1986; Brimblecombe and Clegg, 1988; Haslett et al., 2023). Other pathways lead to the loss of particulate chlorine, such as interactions with NO, N₂O₅, HOBr, and O₃, as well as the release of NOCl, HONO, CINO₂, Cl₂, and BrC (Vogt et al., 1996; Behnke and Zetzsch, 1989; Haslett et al., 2023). These pathways can have significant implications for marine tropospheric chemistry and the polluted coastal atmosphere due to the creation of photochemically active, halogenated gaseous compounds. This study found significant anthropogenic chloride emissions from human activities, which can affect the oxidation capacity of polluted air.

We observed a high correlation (r≥0.7) between BC, OC, and nss-K⁺ in aerosol samples collected at the BCOB (Table S1 in the Supplement) and the MCOH (Table S2). The nss-K⁺ fraction of total potassium was observed at the BCOB (97 ± 3 %), the MCOH (78 ± 13 %), and the MCOG (42 ± 33 %), indicating significant contributions from biomass burning at the BCOB and MCOH as nss-K⁺ is considered a proxy for identifying the regional impact of biomass-burning emissions (Andreae, 1983; Paris et al., 2010). High concentrations of nss-SO42-, nss-K⁺, and NH4+ in the measured ions and carbon aerosols indicate strong anthropogenic sources for the ambient aerosols over the northern Indian Ocean (Table 2).

Our observations show an increase in the SO42-/BC ratio when aerosols are transported from the IGP. Notably, this ratio is more pronounced at the MCOH than at the BCOB (Fig. 5). This shift in composition likely signifies the generation of secondary sulfate from anthropogenic SO₂ during extended transportation. It was observed that there was a lower SO42-/BC ratio at the MCOG than at the MCOH. This might be because SO42- is washed out more easily than BC in this region (Budhavant et al., 2020), and another factor is that the MCOG has slightly different AMBT paths than the MCOH (Fig. 1). It is worth noting that there may be some minor sources of emissions, such as ships and small islands, along the route to the receptor sites. However, their impact on the overall regional loading of BC is insignificant. Therefore, we can infer that most BC loading over the northern Indian Ocean originates from high-emission source areas in southern Asia. While there was an increase in the SO42-/BC ratio, two other important coating components, the WSOC/BC and WIOC/BC ratios (where WIOC denotes water-insoluble organic carbon), declined.

The impact of BC aerosols on air quality, boundary layer dynamics, and climate depends not only on BC concentration but also on the light absorption characteristics of BC. Moreover, MAC values are crucial for estimating radiative forcing accurately. The MAC of BC is here denoted as “BC-MAC₆₇₈”. During the SAPOEX-18 campaign, the calculated value of BC-MAC₆₇₈ was found to have a lower average value at the BCOB (4.4 ± 1.9 m² g⁻¹) and a higher average value at the most distant Indian Ocean receptor station, the MCOG (7.0 ± 1.9 m² g⁻¹), with the MCOH, at a shorter over-ocean transport distance, having a value of 6.1 ± 1.3 m² g⁻¹ (Fig. 2). A recent laboratory-based study of the coating enhancement of the BC MAC (E-MAC) from these observatories showed that the E-MAC value was about 1.6 at all three stations (indicating a 60 % enhancement in the net BC MAC due to coating effects). This constancy suggests that the coating-aging effect of BC was nearly complete upon arrival at the BCOB (Nair et al., 2024). The approximately 80 % increase in BC-MAC₆₇₈ during over-ocean transport must therefore signal another mechanism. It likely reflects a selective fractionation of the BC population, whereby larger and less hydrophobic BC is preferentially scavenged, while a finer pool with higher BC-MAC₆₇₈ becomes relatively more prevalent upon arrival at the distant Indian Ocean receptor observatories. This is consistent with an earlier finding obtained during the winter at the MCOH, where it was observed that BC-MAC₆₇₈ in the rain was lower than BC-MAC₆₇₈ measured for suspended aerosols collected simultaneously from the air (Budhavant et al., 2020). By shedding light on the aging effect of the optical properties of BC aerosols, the study results advance our understanding of this important topic.

These observational constraints in the southern Asian global hotspot region are consistent with global simulation models, which suggest that in ∼ 1–5 d, BC can internally mix with other aerosols (Jacobson et al., 2000). After mixing, the photochemical properties of pure BC are no longer retained due to the coating of other aerosols in the atmosphere, such as sulfates, nitrates, and organics. Observational data on BC MACs in an ambient atmosphere far from immediate sources are rare. Most models use laboratory-based or city-based measurements for MAC and E-MAC values (e.g., Bond and Bergstrom, 2006; Wang et al., 2016). This study can help bridge the gap between model underestimations and observational estimates of BC and absorption aerosol optical depth (AAOD) indicated for southern Asia (Gustafsson and Ramanathan, 2016). These findings can be utilized to refine model estimates of radiative forcing from both BC and BrC for the large-emission region of southern Asia. The severe air pollution from the IGP spreads over large regional scales across the Indian Ocean. Therefore, understanding the aging effect of the optical and chemical properties of aerosols is important, particularly for this region.

In addition to BC, BrC also affects radiative forcing at ultraviolet wavelengths, although its MAC is an order of magnitude lower than that of BC in the visible wavelength range (Bosch et al., 2014; Kirillova et al., 2013). During the campaign, measurements of WS extracts of BrC show significant differences in light absorption characteristics between the three sampling sites. The average MAC measured at 365 nm (BrC-MAC₃₆₅) at the BCOB (1.0 ± 0.3 m² g⁻¹) was 2 to 3 times higher than that measured at the MCOH (0.3 ± 0.3 m² g⁻¹) and that measured at the MCOG (0.6 ± 0.3 m² g⁻¹) (Fig. 4). BrC-MAC₃₆₅ measured during this study is broadly within the same range as that measured in earlier studies focusing on fewer locations in the same region (Bikkina and Sarin, 2014; Dasari et al., 2019). Primary BrC emitted from biomass burning appears to be more light-absorptive than secondary aerosols; MAC values obtained at 405 nm ranged from 0.2 to 1.5 m² g⁻¹ for humic and fulvic acids and from 0.001 to 0.09 m² g⁻¹ for secondary organic aerosols (Lambe et al., 2013), while Chen and Bond (2010) reported a range for primary aerosols of 0.1 to 1.1 m² g⁻¹. The average BrC-MAC₃₆₅ value measured during this study was lower than the values reported close to sources in megacities, such as 1.8 m² g⁻¹ for Beijing in winter (Cheng et al., 2011), 1.6 ± 0.5 m² g⁻¹ for Delhi (Kirillova et al., 2014), 1.6 ± 0.1 m² g⁻¹ for Kanpur (Choudhary et al., 2018), and 1.5 ± 0.2 m² g⁻¹ for Kathmandu (Chen et al., 2020). This indicates that the BrC MAC decreased by a factor of 3 from the IGP exit to the equatorial Indian Ocean (Fig. 3).

The AAE characterizes the spectral characteristic of BrC. Furthermore, the AAE is often used to characterize BrC from coal combustion, biomass burning, and biofuel burning (Chen and Bond, 2010; Rastogi et al., 2021). The AAE value of BrC is typically reported to be ∼1 for fossil fuel emissions, ∼7 for biomass burning, and 7–15 for laboratory-generated smoke and the smoldering of different types of woods (Hoffer et al., 2006; Chen and Bond, 2010). The average AAE values of WS-BrC intercepted at the southern Asian receptor observatories (in the 330–400 nm range) were 5.5 ± 2.7 at the BCOB, 6.5 ± 2.4 at the MCOH, and 4.1 ± 0.5 at the MCOG. These values can be compared with AAE values measured at the Nepal Climate Observatory – Pyramid station (4.9 ± 0.7; Kirillova et al., 2016) and in New Delhi in winter (5.1 ± 2.0; Kirillova et al., 2014). However, the AAE values in this study show differences that are not significant compared to those previously measured at the MCOH in winter (7.2 ± 0.7; Bosch et al., 2014) and at the IGP outflow over the Bay of Bengal (9.1 ± 2.5; Bikkina and Sarin, 2013). The results of this synoptic study, conducted on a large scale across southern Asia, are significant for understanding the aging effect of the optical and chemical properties of aerosols.

From December to March, the tropical Indian Ocean/atmosphere system provides a natural opportunity to study aerosol radiative forcing influenced by anthropogenic aerosols (Satheesh and Ramanathan, 2000; Nair et al., 2023). This is due to the fact that the Indian Ocean atmosphere receives polluted air that travels from the Indian subcontinent and surrounding regions (Fig. 1; Gustafsson et al., 2009; Budhavant et al., 2018).

Direct aerosol radiative forcing (DARF; cloud-free atmosphere) has been estimated over the BCOB, MCOH, and MCOG on a monthly basis using aerosol optical properties obtained from OPAC in the SBDART model (Fig. 5). The DARF at the top of the atmosphere (TOA; 3 km) and at the surface is calculated by estimating the differences in downward and upward fluxes simulated by the model under atmospheric conditions with and without aerosols for the three sites. The surface forcing and TOA DARF were both negative at all three sampling stations, indicating cooling effects. Meanwhile, the atmospheric column represented a warming effect. The negative sign depicts the dominant presence of scattering aerosols.

This study observed that the average atmospheric forcing at the BCOB was higher (10.5 ± 3.2 W m⁻²) than that at the MCOG (4.8 ± 2.1 W m⁻²) and that at the MCOH (8.0 ± 2.6 W m⁻²). This difference was attributed to anthropogenic aerosols, particularly BC (i.e., net-absorbing BC), which caused a warming effect in the atmosphere (Figs. 2 and 5). The BCOB experienced nearly double the atmospheric forcing compared to the remote equatorial ocean at the MCOG, likely due to its higher concentrations of anthropogenic aerosols, such as BC, NO3-, and nss-K⁺ (Table S4; Fig. 3). During winter, the IGP experiences a significant increase in aerosol loading, mainly due to carbon aerosols resulting from fossil fuel and biofuel combustion (Gustafsson et al., 2009; Kaskaoutis et al., 2014; Dasari et al., 2020). As spring progresses, dust becomes the dominant aerosol in the northwestern region of India and in arid areas of southwestern Asia (Kaskaoutis et al., 2014; Singh, 2014; Dumka et al., 2023). At the same time, significant agricultural burning in southeastern Asia results in significantly elevated concentrations of carbonaceous aerosols (Kaskaoutis et al., 2014; Budhavant et al., 2015; Bikkina et al., 2019). Additionally, Himalayan forest fires and wheat residue burning in the IGP contribute to the aerosol burden during spring (Gautam et al., 2007; Bikkina et al., 2019). The BCOB experienced high atmospheric forcing in March, particularly in the outflow region of the IGP (Fig. 5). In January, the MCOH experienced slightly higher atmospheric forcing (11.2 W m⁻²) than the BCOB (10.4 W m⁻²). These findings are consistent with those of our earlier study conducted at the MCOH, which showed that anthropogenic aerosols, such as BC, nss-K⁺, nss-SO42-, and NH4+, were predominantly in the fine mode (70 %–95 %) and were particularly observed in air masses coming from the IGP during this period (Budhavant et al., 2018). These findings suggest that the outflow region of the IGP is exposed to significant atmospheric-warming effects. Therefore, it is crucial to address this issue and take appropriate measures to reduce the amount of anthropogenic aerosol loading.