Contribution of Aerosol Species to the 2019 Smoke Episodes over the East Coast of Peninsular Malaysia

Received: January 3, 2022
Revised: April 5, 2022
Accepted: May 12, 2022

 Copyright The Author's institutions. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited. 

Cite this article:

Mohd Napi, N.N.L., Chel Gee Ooi, M., Talib, M.L., Juneng, L., Mohd Nadzir, M.S., Chan, A., Li, L., Abdullah, S. (2022). Contribution of Aerosol Species to the 2019 Smoke Episodes over the East Coast of Peninsular Malaysia. Aerosol Air Qual. Res. 22, 210393. https://doi.org/10.4209/aaqr.210393

HIGHLIGHTS

• High/low PM_2.5 and AOD level periods are identified in East Coast Peninsular Malaysia.
• High PM_2.5 and AOD related to the biomass burning in Indochina and Indonesia.
• Black and organic carbon have strong and positive influence on high PM_2.5 and AOD.

Keywords: Biomass burning smoke, Long-range transport, AOD, Aerosol species, East coast Peninsular Malaysia

1 INTRODUCTION

Biomass burning (BB) aerosols have a longstanding history of creating societal and environmental issues, particularly in the Maritime Continent (MC) (Tangang et al., 2010; Latif et al., 2018). BB aerosols released into the atmosphere contain large amounts of hazardous particulate matter and trace gases that can disrupt the Earth's energy balance by directly absorbing solar radiation and indirectly disrupting the cloud properties through aerosol-cloud interactions (Hatzianastassiou et al., 2007; Forster et al., 2007; Hatzianastassiou, 2009). The continual release of black carbon (BC) and organic carbon (OC) from BB aerosols which can minimizes the penetration of solar radiation reaching ground level. Consequently, it interrupts the local and regional circulation and degrades the air quality of the entire MC through transboundary smoke events (Morgan et al., 2020; Tang et al., 2020). BB aerosols also introduce indirect effects which are known as cloud albedo and cloud lifetime by modulating the cloud properties by increasing and decreasing the droplet concentrations and influencing the amounts of precipitation in local and regional areas (Huang et al., 2019; Liu et al., 2019; Jacobson, 2014). Hence, the continual uncontrolled emission of BB aerosols can cause unhealthy air quality as well as lead to the climate change problem (IPCC, 2013).

Peninsular Malaysia has unique geographical features. It is located at the middle of the MC and is surrounded by multiple countries that experience recurring large-scale BB events, hence making it susceptible to transboundary smoke events derived from BB activities. The wide South China Sea (SCS) is situated off the east coast of Peninsular Malaysia (ECPM). The coastal region, ECPM is often known be affected by air-sea interaction processes which influence convection and the air quality of local and regional areas (He et al., 2018). Hence, this study focused more on the ECPM region since it has complex weather conditions and its air quality is susceptible to the influence of large-scale BB activities that occur in Southeast Asia (Rahim et al., 2021; Latif et al., 2018). Similar to most coastal cities, the ECPM is influenced by synoptic and local (land-sea breeze) weather features. The main synoptic factor is the northeast monsoon (NEM) that occurs during the boreal winter. When the surface temperatures decrease in continental Asia during the winter, the outgoing cold air masses from Siberian-Mongolian highs form wet northeasterlies over the ECPM (Mohyeddin et al., 2020; Ashfold et al., 2017; Kok et al., 2015; Ooi et al., 2011). The northeasterly winds were found to introduce the BB smoke from Indochina’s BB emissions to the ECPM via the SCS (Chenoli et al., 2018). On the other hand, the southwest monsoon (SWM) which occurs during the boreal summer is often closely linked with severe transboundary smoke events in Indonesia. However, very limited studies have been conducted on how synoptic conditions in the ECPM region influence the long-range transport of transboundary BB smoke through the SCS, affecting air quality.

The collective analysis of backward trajectory analyses, model data, and satellite data can aid in identifying the source of emissions and dominant types of aerosols that influence the air quality during transboundary smoke events in regions of the ECPM. The aerosol product data are available from satellites, such as from the Moderate Resolution Spectroradiometer (MODIS) and Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) while modelling data can be obtained through Modern-Era Retrospective Analysis for Research and Application (MERRA) reanalysis data. The term "aerosol optical depth" (AOD), sometimes known as "aerosol optical thickness" (AOT), was used in the model and satellite product, as it refers to how much the aerosol loading in the form of particulate matter can prevent sunlight from reaching the ground level in a column. Numerous previous studies have shown the success of those satellite and model data in simulating the relationship of aerosols with their species and the vertical profiling of aerosol species on a spatial and temporal scale (Zhang et al., 2021; Rizza et al., 2019; Liu et al., 2018). However, in the ECPM region, MERRA and CALIOP data have rarely been used to distinguish between the dominant aerosol species from local and non-local emissions.

The main objectives of this work were to identify the high pollution periods and the sources of emissions that influence the high PM_2.5 concentrations and AOD levels over the ECPM region. The study also attempted to identify the role of synoptic meteorological conditions that can influence long-range transport of emissions from their burning sources. The burning aerosol species in the ECPM region were also determined through the MERRA-2 reanalysis dataset during different high pollution periods. This study could provide a thorough understanding of the weather factors and aerosol species that influence the degradation of the air quality over the ECPM, thus providing important information for developing mitigation measures to improve air quality conditions.

 
2 METHODS

 
2.1 Study Period and Location

The year 2019 was selected as the study period for air quality study as it experienced two major high pollution episodes of ambient PM_2.5 concentrations on the east coast of Malaysia (Fig. 1) and west Malaysia (Fig. S1). The high pollution periods were characterized by pollution values well above the New Malaysia Ambient Air Quality Standard (NMAAQS) limit, 35 µg m^–3 for 24 hours (DOE, 2021). These major pollution periods coincided with BB seasons both in Southeast Asia during the boreal spring and the MC during the boreal summer (shown in Fig. 2). Hence, 2019 can serve as an exemplary case study period to determine the potential sources of PM_2.5 from burning, as well as distinguishing the difference between the PM_2.5 species from the two pollution episodes.

Fig. 1. The study area showing (a) the Maritime Continent, (b) the ECPM regions with the CAQMS involved in this study (shaded in red) and the MERRA-2 reanalysis data extracted for correlation analyses in the ECPM region (dotted orange rectangle).

Malaysia recorded higher than usual average annual temperature of around 32.7°C in 2019, while the total rainfall was between 1800 mm and 3900 mm and the highest average daily wind speed was 6.6 m s^–1 (MET, 2019). It experienced a weak El-Niño phase from early in the year until June 2019 and continued with a neutral El-Niño Southern Oscillation (ENSO) event until December 2019 (MET, 2019). The regions experienced a SWM, inter-monsoon, and NEM with occasional tropical cyclones in 2019 with distinctive typhoons in August (Lekima) and November (Nakri) across the SCS.

This study covered three states along the ECPM, namely Terengganu, Kelantan, and Pahang. The location of the ECPM increases its likelihood of receiving air pollution from various sources including long-range transport, local emissions, and natural sources. The region also faces the SCS, where the local and long-range air pollution could mix with clean marine air, making it hard to differentiate between the two. The main economic activities of these three regions include the oil and gas and petrochemical industries, and agriculture, education, tourism, and manufacturing sectors (Latif et al., 2018). One Continuous Air Quality Monitoring Station (CAQMS) was selected for each state with the highest daily PM_2.5 reading, as marked in Fig. 1: the CAQMS in SMK. Bukit Kuang (CA42T) with a latitude of 4.261° and longitude of 103.426°, SMK Tanah Merah (CA46D) with a latitude of 5.811° and longitude of 102.135°, and SK Indera Mahkota (CA40C) with a latitude of 3.818° and longitude of 103.297°, to represent Terengganu, Kelantan, and Pahang, respectively.

 
2.2 Data

2.2.1 Ambient PM_2.5 concentrations data

The one-year data of PM_2.5 (µg m^–3) from January until December of 2019 were acquired from the Department of Environment (DOE) Malaysia. The hourly PM_2.5 measurement data were measured using a particle analyzer of a Continuous Dichotomous Ambient Air Monitor 1405-DF Tapered Element Oscillating Microbalance (TEOMTM) (Azhari et al., 2021). The quality of assurance and quality control of the data were conducted with a manual calibration using gas detection, to ensure the missing data resulting from technical problems were checked and verified before data was published (PSTW, 2018a, 2018b).

 
2.2.2 MERRA-2 reanalysis data

The AOD (550 nm) and other five aerosol species of BC, OC, sulfate (SO₄^2–), dust, and sea salt, as well as wind speed, wind direction, and sea level pressure (SLP) in 2019 were obtained through the through the MERRA-2 reanalysis based on the Goddard Earth Observing System Model, version 5 (GEOS-5), and together with Atmospheric Data Assimilation System (ADAS) through the open-access NASA Earth web-based data at https://disc.gsfc.nasa.gov/ with monthly and hourly distribution data (M2IMNXGAS and M2T1NXAER) and six-hourly intervals of meteorological data (M2I6NPANA). The resolution of the data was about 0.5° × 0.65° latitude and longitude, with 72 hybrid-eta levels from the surface to 0.01 hPa (Penna et al., 2018). The Spearman bivariate analysis in the SPSS® version 25 was used to analyze the relationship of AOD with BC, OC, SO₄^2–, dust, and sea salt.

The hourly near-surface PM_2.5 of MERRA-2 was verified with the ground-based data model for the three stations based on the mean fractional bias (MFB), mean fraction error (MFE), and correlation coefficient (R), the equations are shown in the supplementary Fig. S2. The model agreed well with the observation data, with the MFB was slightly underestimated (around –0.02 to 0.09), MFE (up to 0.15), and R (around 0.38 to 0.72) for all three locations in the ECPM region. The ground based AOD measurements were unavailable at these stations, while standard errors of 15% were unexpected according to several previous verification studies (Yang et al., 2021; Rizza et al., 2019).

 
2.2.3 CALIOP data

The LIDAR level 1 version 4.1 product of CALIOP with day and night 532 nm backscatter parameters (for aerosol vertical feature masks and their subtype) were obtained from the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) satellite based on the period that synced with high aerosol emission detected through backward trajectory analyses of the studied ECPM regions and assessed in https://subset.larc.nasa.gov/calipso/ (Nowottnick et al., 2015; Winker et al., 2010). The incorporation of the MERRA version 2 in the CALIOP validation showed an increase of the accuracy and reliability of the output. However, a small percentage of bias needed to be accounted for since the stratospheric aerosol background has a greater tendency in the tropics compared to higher atmosphere’s altitudes especially in the daytime (Vaughan et al., 2012).

 
2.3 Backward Trajectory Analysis

The Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model through the web-based version (https://www.ready.noaa.gov/hypub-bin/trajtype.pl) provided by the National Oceanic and Atmospheric Administration (NOAA) Air Resource Laboratory’s (ARL) Real-Time Environmental Application and Display System (READY) (NOAA, 2021a) was used in this study to determine the sources of emissions and their transport medium to the ECPM. The meteorological dataset of Global Data Assimilation System (GDAS) data with a 1° × 1° horizontal resolution with different horizontal and vertical resolutions over 72 hours was applied to drive the backward trajectory HYSPLIT model in order to trace the source of higher PM_2.5 concentrations that arrived at the three selected CAQMS stations in the ECPM region (NOAA, 2021a).

Five different above-ground level (AGL) heights at 10 m (near-surface layer), between 500 m to 1500 m (boundary layer), and between 1500 m to 3000 m (free lower troposphere layer) were selected to observe the atmospheric stability and determine in which AGL level the aerosols were transported and deposited in Malaysia (Hee et al., 2016). Considering some limitations of HYSPLIT in processing the quality of complex terrains and meteorological datasets, the simulation and episodic data interpretation had to account for an estimated error of about 15–30% (NOAA, 2021b; Watt et al., 2017).

 
3 RESULTS AND DISCUSSION

 
3.1 Distinct Periods of Aerosol Loading along the ECPM

In this study, the time series of ambient PM_2.5 concentrations in the year 2019 for each station along the ECPM shows a fluctuating trend as shown in Figs. 2(a–c). The hourly PM_2.5 concentrations in Terengganu, Kelantan, and Pahang were almost constantly higher than NMAAQS permissible limit throughout with two significant periods of hourly PM_2.5 concentrations in each station. The daily trends of average PM_2.5 concentrations were plotted for each station based on the CAQMS and MERRA-2 reanalysis data (Fig. S3). The AOD also showed a fluctuating trend throughout the year as shown in Fig. 2(d). Hence, four different periods were identified in this study for the maximum and minimum PM_2.5 concentrations and AOD: (i) March–April (high pollution period 1, HP1), (ii) May (low pollution period 1, LP1), (iii) August–September (high pollution period 2, HP2), and (iv) December (low pollution period 2, LP2). The minimum and maximum values of PM_2.5 concentrations and AOD are also shown in the time-series graph in Fig. 2. The average PM_2.5 concentrations at Terengganu during HP1, LP1, HP2, and LP2 were 20.75 µg m^–3, 16.33 µg m^–3, 39.02 µg m^–3, and 10.92 µg m^–3, respectively. In Kelantan, the average PM_2.5 concentrations were 28.34 µg m^–3 (HP1), 14.57 µg m^–3 (LP1), 26.39 µg m^–3 (HP2), and 16.82 µg m^–3 (LP2). Meanwhile in Pahang, average PM_2.5 concentrations of 16.38 µg m^–3, 20.10 µg m^–3, 42.54 µg m^–3, and 10.04 µg m^–3 were recorded during HP1, LP1, LP2, and LP1, respectively. The ECPM regions had average AOD levels of about 0.22, 0.16, 0.51, and 0.16 during HP1, LP1, HP2, and LP2. This statistical analysis describes the selection criteria for high and low pollution periods at ECPM and used the DOE permissible limit of PM_2.5 concentration as benchmarks. Comparisons with Figs. 2(a–c) and 2(d) show that the PM_2.5 concentrations and AOD levels exhibited a simultaneous trend of high and low aerosol loadings during HP1, LP1, HP2, and LP2. Hence, the aerosol particles could be found in form of small and fine particulate matter (i.e., PM_2.5) in the atmosphere. The ECPM has a lower GDP (gross domestic product) than the main economic hub in west Peninsular Malaysia. Indeed, the ECPM region usually records low PM_2.5 levels compared to west Peninsular Malaysia, but the values of both increased tremendously during BB activities especially in March and September in year 2019 (Fig. 2 and Fig. S1 in the supplementary materials). This observation is also supported by findings by Samsuddin et al. (2018) who noted that the PM_2.5 values exhibited a two-fold increase in semi-urban and urban areas in Peninsular Malaysia during hazy days due to BB.

Fig. 2. Hourly PM_2.5 concentrations in (a) Terengganu, (b) Kelantan, (c) Pahang, and (d) AOD levels in the ECPM region. The periods of high pollution (HP) are marked by blue rectangles and the low pollution (LP) periods are marked by black dotted rectangles.

Figs. 3(a–l) illustrate the monthly spatial MERRA-2 AOD levels from January until December of 2019. High AOD levels were identified over two main regimes during different periods of time. From January until May, the AOD levels were higher over the Indochina region (Figs. 3(c–e)) which exerted a significant influence on the AOD levels in Malaysia from March to May when HP1 was identified. Meanwhile, from July until November, the AOD levels were higher in Malaysia and Indonesia (Figs. 3(g–k)), concurrent with the HP2 period. The AOD hotspots show that the origins and sources of the aerosols were located in Kalimantan, Sumatra, and Java.

Fig. 3. Monthly AOD distributions extracted from MERRA-2 for the year 2019.

In this study, the meteorological parameters played an important role in triggering the high AOD levels and transboundary smoke events that affected the ECPM region as shown in Fig. 4. The wind speed, wind direction, and SLP versus AOD levels at 925 hPa pressure, an ideal condition to present the influence of meteorological factors at the east coast areas (Ashfold et al., 2017) were plotted to observe the long-range route of transboundary smoke to the ECPM. The transboundary smoke events in this study were defined when the level of AOD at the sources of emissions exceeded 0.5. The AOD threshold value of 0.5 in the sources of emissions was chosen as a benchmark because the levels of AOD at the origins of emissions must be higher than in the receptor regions. The minimum AOD levels during high pollution periods along the ECPM ranged from 0.01 to 0.3 as shown in Fig. 2. This is also supported by findings from Pani et al. (2018) which revealed that the mean AOD levels during low BB to extreme BB burning events in Thailand (closest country to the ECPM region) were between 0.6 to 2.50 when severe smoke episodes occurred in the Northern Peninsular Southeast Asia. Moreover, Yin et al. (2020) also stated that the AOD levels were below 0.3 when there were no BB activities occurring in 2015. Strong winds exceeding 10 m s^–1, supported by findings from Mohyeddin et al. (2020), either from the northeast or southwest of the ECPM with a high SLP at the upper and lower latitudes of the ECPM can push significant amounts of BB emissions towards the ECPM and decline its air quality. Additional information on the meteorological characteristics are provided in the supplementary Fig. S4. However, further studies are required to determine the extents to which local and non-local pollution can affect the ECPM region.

Fig. 4. The meteorological conditions of wind speed, wind direction, and sea level pressure versus AOD levels during (a) HP1, (b) HP2, (c) LP1, and (d) LP2.

 
3.2 Identification of the Origzin of Aerosols through the HYSPLIT Model

Trajectory analyses were used to identify the possible emission sources which influenced the PM_2.5 concentrations and AOD levels along the ECPM in 2019. The trajectory analyses for HP1 and HP2 were selected based on the time that the PM_2.5 concentrations exceeded the permissible limit of the NMAAQS as listed in Table 1. The LP1 and LP2 periods were selected based on th lower values of PM_2.5 concentrations and those within the permissible limit of the NMAAQS to act as a background for the local emissions indicators. The study further determined the types of aerosols suspended at different heights above the Earth’s surface using CALIOP data. The identification of the types of aerosols within the atmosphere can help distinguish the dominant potential sources of air pollutants that influenced the air quality of the ECPM region during HP1, HP2, LP1, and LP2. The vertical feature mask attenuated backscatter at 532 nm was plotted to differentiate the aerosol types, followed by the profiling of tropospheric aerosol subtypes during HP1, HP2, LP1, and LP2 along the ECPM for daytime and nighttime.

 
3.2.1 High pollution period 1 (HP1)

The HP1 period coincided with the inter-monsoon season from the 19^th of March until the 1^st of May 2019 (MET, 2019). During HP1, the backward trajectory analysis showed that the aerosols were derived from the SCS and Indochina at all heights (Fig. 5 and Fig. S5 in the supplementary materials). On the 24^th of March 2019, the analysis showed that the pollutant sources that influenced the ECPM region likely originated from Indochina, as supported by the high AOD values (Figs. 5(a) and 5(b)) located approximately at a 10°N–15°N latitude. The higher pressure in the northern part of the hemisphere, also known as the Siberian high associated with cold air masses, was still active towards the end of the boreal winter and managed to push the airmasses from the northern to the low-pressure southern part of Southeast Asia which is closer to the equator. Although the northeasterly winds during the inter-monsoon are not as strong as during the northeast monsoon, they can still help carry BB pollutants from upwind Indochina to the ECPM region (Mohyeddin et al., 2020; Akhir et al., 2014).

Fig. 5. Backward trajectories analyses at (a, c) 0–1000 m and (b, d) 1500–2000 m for HP1 (a, b) and HP2 (c, d) on three consecutive days with high PM_2.5 concentrations in the ECPM region. The black dotted lines A (daytime) and A’ (nighttime) are represent as a swath of CALIOP.

Figs. 5(a–d) illustrates the vertical feature mask and aerosol subtypes in the ECPM region (marked by red boxes) in the daytime and nighttime during HP1 and HP2 from CALIOP. The tropospheric aerosols detected were scattered at altitudes below 5 km throughout the day, with an additional layer at altitudes above 10 to 15 km in the nighttime during HP1, corresponding to the backward trajectory results shown in Fig. 5(b) since the emissions can only reach a height of 2.5 km. The tropospheric aerosols from polluted dust, dust, and polluted continental/haze were found to be scattered below 5 km in the daytime; during the nighttime, marine, polluted continental/haze, and elevated haze were found suspended below 5 km altitudes while the dust and elevated haze were scattered at altitudes within 10–15 km as shown in Figs. 6(b) and 6(d). The CALIPSO swath showed that the polluted and elevated smoke suspended at 5 km altitudes in the atmosphere were being transported from the northeast of the ECPM (Indochina) and degraded the air quality of the ECPM region.

Fig. 6. CALIOP data of backscatter 532 nm cloud and aerosols in the ECPM regions during (a–d) HP1 and (e–h) HP2 for day and nighttime. The red dotted boxes show the areas that cover the ECPM regions.

3.2.2 High pollution period 2 (HP2)

The HP2 period coincided with the southeast monsoon period which occurred from the 2^nd of May until the 24^th of September 2019, the ENSO phenomenon (El-Niño) as it reported by MET, (2019), and with a drier condition with less precipitation, less cloud, prevailing westerly wind, and high outgoing long-wave radiation (Che et al., 2019). The drier conditions provided conducive conditions for burning activities and increased the photochemical reactions of the pollutant for secondary aerosol formation in the atmosphere (Chenoli et al., 2018; Wang et al., 2016). Kanniah and Yaso (2010) concluded that low levels of seasonal rainfall, high relative ambient temperatures, together with southwesterly winds create a medium for the advection of air masses which contain BB-derived aerosols, especially from Kalimantan and Sumatra in Indonesia, triggering long-range pollutant transport to Malaysia’s atmosphere.

For HP2, Fig. 5(c) shows that the emissions from the Karimata Strait (marked in Fig. 1(a)) and Kalimantan scattered at high altitudes before reaching the near-surface level of the ECPM region. Meanwhile, at altitudes of 1500 to 2000 m AGL, the pollutants were generated in Sumatra (21^st and 22^nd September 2019) and Kalimantan (23^rd September 2019) (Fig. 5(c)). The sources were concurrent with significant AOD levels defined from MERRA-2 data at 5°S to 5°N and 100°E to 115°E. The presence of low precipitation, less clouds, high outgoing long-wave radiation. and southerly winds during the southwest monsoon initiated the horizontal transport of air masses that contained air pollutants from Sumatra and Kalimantan across the Malacca and Karimata straits to Peninsular Malaysia where they affected the air quality (Bucci et al., 2020; Oozeer et al., 2016). For more information on the sources of emissions during HP2, see Fig. S5.

The tropospheric aerosols were suspended at atmospheric altitudes below 5 km which agrees with the findings of the backward trajectory analyses as shown in Fig. 5(d), since the highest emissions were detected at 3.5 km altitudes; by having aerosol subtypes from marine, polluted continental/haze, and polluted dust sources observed in the daytime as shown in Figs. 6(e–f). In the nighttime, the troposphere aerosols were observed it scatter at below 5 km AGL together with marine, polluted continental/haze, elevated haze, and dust. Meanwhile, sulfate was found it scattered between 18 to 20 km AGL and considered a stratospheric aerosol (Figs. 6(g–h)). Hence, the CALIOP swaths obtained during the daytime and nighttime showed that the emissions of polluted smoke and elevated smoke suspended in the vertical aerosol profile originated from lower latitudes of the ECPM region and were travelling and engulfing the atmosphere of the ECPM region.

3.2.3 Low pollution period 1 (LP1)

The LP1 period was the transitional period of monsoonal changes or the end of the northeast monsoon period. For LP1, Figs. 7(a–b) show that the air masses passed through Singapore, the Strait of Johor, Strait of Malacca, and Sumatra. The emissions from the Strait of Johor, Singapore, and Strait of Malacca were significant at the near-surface level as shown in the supplementary Figs. S6(a–d). However, Fig. 7(b) shows that at 1500 to 2000 m AGL the emissions were originating from neighboring Sumatra, Indonesia. The ECPM region is considered to receive inputs from mixed sources of pollution, from local and non-local sources, that can influence the degradation of the air quality. The local emissions had a more significant effect on air quality than non-local emissions during this time. During LP1, the northern inter-tropical convergence zone (ITCZ) started to shift towards the south, and the low-polluted air masses carried from Sumatra by weak winds blew in emissions from BB (Dahari et al., 2019; Andronache et al., 2002).

Fig. 7. Same as Fig. 5 but for the (a–b) LP1 and (c–d) LP2 periods.

Fig. 8 shows the vertical distribution of aerosols during LP1 and LP2. These periods were considered as having a relatively clean atmosphere compared to HP1 and HP2. Some tropospheric aerosols that consisted of marine, dusty marine, and smoke pollutants were traced to altitudes below 2 km in the daytime and nighttime during LP1. The aerosols from dust and polluted dust were scattered at altitudes above 2 km at both times as shown in Figs. 8(a–d). The swath of CALIOP also shows that the clouds and suspended aerosols during LP1 were coming from the south of the ECPM region.

Fig. 8. Same as Fig. 6 but for the (a–d) LP1 and (e–h) LP2 periods.

3.2.4 Low pollution period 2 (LP2)

For LP2, the trajectory analyses showed that the air masses were arriving from 500 m to 2000 m AGL from Indochina, China, and the SCS as shown in Fig. 6. The strong wind speeds with northeasterly directions and high SLP pushed the BB air masses to the ECPM region as shown in Fig. 4. Although the air masses were observed to originate from Indochina during LP2, the ECPM region experienced extremely heavy rainfall during the upcoming northeast monsoon season during LP2 with maximum values of rainfall of about 400 mm in December based on previous findings by Fakaruddin et al. (2017). This condition prevented the air pollution from Indochina from reaching the ECPM region since the heavy rainfall had a negative correlation with the air pollutants, especially with particulate matter, because the pollutants were washed out along the transport pathway. This has been supported by previous findings from Yassen and Jahi (2007) which showed that the concentrations of total suspended solids in the air were decreasing as they were removed and washed out by the wet conditions during NEMs. Therefore, during the LP2 period, concurrent with the NEM, the atmosphere in November was generally clean. Fig. S6 provides more information on the trajectories analyses one day before and after the LP2. As expected, Figs. 8(e–h) show a relatively low significance of suspended tropospheric aerosols in the daytime and nighttime. Hence, the LP2 period showed the lowest significance of tropospheric aerosols in the atmosphere relative to HP1, HP2, and LP1, as it was correlated with the weather conditions experienced by the ECPM region as discussed earlier. Meanwhile, during LP2, the CALIOP swath also showed that clouds were dominating the atmosphere over suspended aerosols, as detected in the vertical aerosol profiling from the upper latitudes of the ECPM region.

 
3.3 Aerosol Species Contribution

The monthly aerosol species contribution to the ECPM region during the HP1, HP2, LP1, and LP2 periods were obtained from MERRA-2 model data. Figs. 8(a–g) show the distribution of aerosol species during HP1 and HP2. The results show that BC and OC had dominant contributions to the aerosol species during these two periods of time. During HP1, BC and OC were simultaneously present in significant levels in Indochina and the ECPM region. This is supported by the results of Tsai et al. (2013), since the HP1 period also experienced the ‘Asian Brown Cloud’ that occurred between March and April in Indochina. The overflow of BB emissions during the Asian Brown Cloud events released large amounts of BC and OC to the SCS and the wind directions during the inter-monsoon may have had a great influence on the atmosphere of the ECPM.

Figs. 9(h–n) also show that the dominant aerosol species consisted of BC and OC which influenced the air quality in the ECPM region during HP2. It occurs simultaneously with the BB emissions from Sumatra and Kalimantan. The BB activities in Indonesia involved carbonaceous matter from slash and burning, open burning, wildfires, and others. The results agree with previous findings by Latif et al. (2018) and Lee et al. (2018) which showed that the BB that occurred in Kalimantan and Sumatra was dominated by emissions of OC and BC (estimated at about 50% for BC and 73% OC) in the form of particulate matter relative to local emissions. However, during both HP1 and HP2, the SO₄^2– values were also higher but most significant during LP1. Emissions of SO₄^2– are categorized as secondary aerosols which commonly relate with local emissions, specifically from natural and anthropogenic sources such as sea-spray, volcanoes, industrial activities, and transportation. Hence, further correlation analyses are needed to support the hypothesis that SO₄^2– was most significant due to local emissions during LP1.

Fig. 9. The sources of contaminants that influenced the higher AOD levels in the ECPM in (a–g) HP1 and (h–n) HP2 MERRA-2 reanalysis data, with information on (a, h) the source of air masses, (b, i) AOD levels, (c, j) BC, (d, k) SO⁴_2-, (e, l) dust, (f, m) OC, and (g, n) sea salt.

The transitional monsoon period influenced the meteorological conditions during the LP1 period. The effects of the NEM were weaker than during HP1 as conditions had started to shift towards the SWM in May as mentioned in an annual report of 2019 from MET Malaysia. Hence, during this period, the weather conditions began to introduce drier ambient temperatures, low precipitation, and a shift of the winds blowing from northernly to western directions. According to Figs. 10(a–g), the high AOD levels were influenced by the mixed aerosol species which originated from local and non-local emissions from natural and anthropogenic activities; these were located in the south and north of Peninsular Malaysia which affected the air quality especially in the ECPM region during this time. Meanwhile during LP2, the air quality was influenced by the NEM weather conditions which were characterized by a high intensity of precipitation, high relative humidity, low temperatures, and cold winds blowing from the Siberian high, promoting wet conditions which lowered the AOD levels and all their species as shown in Figs. 10(h–n). Although the trajectory analysis showed that some emissions originated from Indochina during these periods, these did not influence the air quality in the ECPM region since all the pollutants from local and non-local sources had been washed out and deposited away by the heavy rains. Hence, this period can be considered as representing a clean atmospheric environment. As a result, this study focused on the HP1 and HP2 periods because the PM_2.5 concentrations and AOD levels were severe and exceeded the permissible limit defined by the DOE of Malaysia. An attempt was thus made at determining their sources, the long-range transport route of smoke events, and the types of aerosol species that dominated and degraded the air quality in the ECPM region due to its synoptic weather patterns.

Fig. 10. The sources of contaminants that influenced the higher AOD levels in the ECPM in (a–g) LP1 and (h–n) LP2 MERRA-2 reanalysis data, with information on (a, h) the source of air masses, (b, i) AOD levels, (c, j) BC, (d, k) SO₄^2–, (e, l) dust, (f, m) OC, and (g, n) sea salt.

 
3.3.1 Correlation analyses between AOD and aerosol species

The study continued to determine the relationship between AOD levels and the related aerosol species. The hourly AOD data from MERRA-2 analyses are not normally distributed as the Kolmogorov-Smirnov and Shapiro-Wilk tests have P(0.00) ≤ 0.05 and the normal distribution graph also skews to the left. In addition, the dataset does not have an equal variance as the Levene’s test (homogeneous test) result is P(0.00) ≤ 0.05. Since the data did not fulfill the requirements of both tests, it proceeded to the Kruskal-Wallis test. As a result, P(0.00) ≤ 0.05 shows that the data have a significant difference of the median. Hence, because the data was non-parametric and not normally distributed, a Spearman bivariate correlation analysis was selected to determine the relationship between AOD and the aerosol species during two different periods in the ECPM region.

Table 2 shows the Spearman bivariate correlation results for the four pollution periods. The relationships of AOD with BC, SO₄^2–, dust, OC, and sea salt were determined. The relationship of AOD with the other aerosol species in the ECPM region during different periods can be summarized as BC > OC > SO₄^2– > sea salt > dust during HP1, BC > OC > SO₄^2– > sea salt > dust during HP2, and SO₄^2– > BC > OC > dust > sea salt during LP1. The LP2 period is not discussed since this period had low pollution levels with a minimal influence from BB emissions because it experienced the NEM associated with wet conditions as explained in Section 3.2.4. Aerosol species, especially from BC, OC, and SO₄^2–, influenced the AOD levels in the same way during both HP1 and HP2. Indeed, the BC, OC, and SO₄^2– pollutants had strong and positive relationship with AOD during HP1 and HP2: BC (r = 0.908, p < 0.01), OC (r = 0.899, p < 0.01), and SO₄^2– (r = 0.793, p < 0.01) during HP1, and BC (r = 0.963, p < 0.01), OC (r = 0.935, p < 0.01), and SO₄^2– (r = 0.614, p < 0.01) during HP2. Meanwhile, during LP1 the SO₄^2– (r = 0.942, p < 0.01), BC (r = 0.902, p < 0.01), OC (r = 0.833, p < 0.01), and dust (r = 0.574, p < 0.574) levels showed strong positive correlation, and sea salt (r = 0.417, p < 0.417) a moderately positive relationship with AOD levels.

The correlation analysis results support the findings presented in Section 3.3, showing significant levels of BC and OC during HP1 and HP2. Due to favorable weather conditions, the air quality was degraded by dominant sources of aerosol species from large amounts of BB emissions from Indochina (during HP1) and Indonesia (during HP2). Khoir et al. (2021) estimated an average of more than 25000 fire hotspots in Indochina during March and April and more than 2000 from August to September of 2015 based on data from 2001 to 2021. These findings support the hypothesis that BB had a greater impact than local emissions on the air quality of the ECPM region based on the large number of fire hotspots detected. Black carbon, also known as elemental carbon, and OC, was released from BB, fossil fuels, and secondary organic aerosol residues. Both were released into the atmosphere as fine and coarse particulate matter. The combustion of carbonaceous material during BB resulted in significant levels of BC and OC. The results reported by Engling et al. (2014) proved that BC and OC comprised about 50% in total suspended particles during hazy days compared to 25% during non-hazy days. The BC released by BB was coated with organic materials. Meanwhile, local emissions such as those from urban activities coated the BC primarily with SO₄^2–, nitrate, and primary and secondary OC (Qi and Wang, 2019). During the LP1 period, when non-local and local emissions were mixed, SO₄^2– had the strongest correlation with AOD levels. The formation of SO₄^2– involves photochemical and oxidation reactions which mainly take place in the industrial sector, and in biogenic and natural sources. However, BB emission has a minor influence on secondary inorganic carbon which it released depends on the type of residual, soil, and other factors involved (Latif et al., 2018; Chen et al., 2017; Behera and Balasubramanian, 2014). Previous studies by Zainal et al. (2021) and Lee et al. (2018) showed that about 93% of atmospheric SO₄^2– was derived from non-fire events and that it was present in higher concentrations during pre-smoke events than during smoke events and post-smoke events that occurred in year 2019 in Malaysia. In Terengganu, the higher correlation of SO₄^2– with high AOD levels and PM_2.5 concentrations during LP1 can be attributed to the large-scale emissions from the petroleum sector (main economic sector) which affect the air quality in the ECPM region (Ragothaman and Anderson, 2017). During HP1 and HP2, the BC and OC levels were directly proportional to the level of AOD and ambient PM_2.5 concentrations in the ECPM region, in large part due to BB activities. Meanwhile, due to the complicated chemical reactions of SO₄^2–, it is reasonable to assume that the SO₄^2– derived from both local and non-local sources, which originated primarily from fossil fuel combustion.

In summary, according to the HYSPLIT trajectory analyses, the high ambient PM_2.5 concentrations and AOD levels during the HP1 period resulted from intense BB activities in the northeast of the ECPM region (Indochina) while the HP2 period was influenced by BB activities located in the southwest (Indonesia). Inter-monsoon and SWM weather conditions influenced the transport route of transboundary smoke in Indochina and Indonesia during HP1 and HP2. The air quality was influenced by mixed emissions from local and non-local sources during LP1. Meanwhile, LP2 was found to have the cleanest environment out of all studied periods as it experienced the NEM which helped wash out the pollutants and deposit them away from the area. Therefore, BC, OC, and SO₄^2– had a strong positive influence on AOD levels and ambient PM_2.5 concentrations during HP1 and HP2. However, SO₄^2– levels were more significant during LP1 due to the complex chemical reactions involving SO₄^2– in local emission sources. Understanding the complex interactions between aerosol properties, the microphysical state of the atmosphere can help gain crucial information on the behavior of the aerosols themselves and should be considered in future studies to gain a better insight into the long-range transport of air pollutants. Moreover, aerosol species of nitrate also should be included in future studies to see how significant it is in influencing the PM_2.5 concentration and AOD level, especially during BB in ECPM region.

 
4 CONCLUSIONS

In conclusion, the hourly trend of PM_2.5 concentrations and AOD levels in the ECPM regions (Terengganu, Kelantan, and Pahang) in 2019 revealed two significant peaks of high pollution events. Hence, four different pollution periods with high and low PM_2.5 concentration trends and AOD levels in March–April (HP1), August–September (HP2), May (LP1), and December (LP2) were identified to determine the sources of high PM_2.5 concentrations and AOD levels in the studied regions. Both high and low concentrations of PM_2.5 and AOD levels were observed through HYSPLIT backward trajectory and CALIOP data analyses. High concentrations of PM_2.5 and AOD levels potentially originated from and were influenced by BB in Indochina during HP1, burning activities in Kalimantan and Sumatra during HP2, local (natural and anthropogenic activities) and non-local emissions during LP1, and natural sources during LP2. During HP1, the tropospheric aerosols were scattered below 5 km altitudes in the daytime meanwhile at nighttime, the tropospheric aerosols were scattered at two different heights which were at 5 km and between 10 and 15 km. The tropospheric aerosols in HP2 were scattered below 5 km (day- and nighttime). During LP1, the tropospheric aerosols were observed at two different heights which were below 2 km and around 12–13 km during daytime, and at nighttime it scattered at three different atmosphere altitudes which were 2 km, 3–4 km, and 9–10 km. The levels of tropospheric aerosols in LP2 were relatively low and did not become the main concern in this study because it’s likely they would be deposited in the atmosphere.

The Spearman bivariate correlation analysis results showed that the AOD levels during the HP1 and HP2 periods were most influenced by BC, followed by OC, SO₄^2–, sea salt, and dust. During LP1, SO₄^2– exerted the strongest influence on AOD levels, followed by BC, OC, dust, and sea salt. The BC, OC, and SO₄^2– levels, especially during HP1, HP2, and LP1, showed a strong positive correlation with AOD levels and PM_2.5 concentrations. Relative to HP1 and HP2, the SO₄^2– levels had a significant effect on the AOD levels during LP1. This study obtained important data on the sources of pollutant emissions during various polluted episodes, as well as the types of aerosol species that influence the air quality in ECPM regions. These results can help guide policymakers and develop monitoring measures to protect the air quality of the ECPM. Further research into the behavior of aerosols and their interactions at different atmospheric altitudes during different monsoon seasons in Malaysia is needed to better understand how aerosols are transported over long-range routes.

ACKNOWLEDGMENTS

We gratefully acknowledge the Ministry of Higher Education (MoHE) Malaysia (FRGS/1/2020/WAB02/UKM/02/1) and Shanghai International Science and Technology Cooperation (NO. 19230742500) for supporting and funding this work. All authors would like to express gratitude for the provision of MODIS, CALIOP, and MERRA-2 data from NASA, the weather data from Malaysian Meteorological Department (MET), and the air quality data from the Department of Environment (DOE) Malaysia.