The influence of meteorology and emissions on the spatio-temporal variability of PM10 in Malaysia
Introduction
Outdoor air pollution from particulate matter (PM) is a worldwide issue, with 91% of the world's population living in areas where air pollution exceeds WHO guidelines [WHO, 2018]. Understanding the scale and severity of outdoor particulate air pollution, as well as its interaction with local meteorology and the complex regional climate, is critical to improving our understanding of pollution impacts on visibility, economy, and public health [Latif et al., 2018; Lee et al., 2017]. Equatorial Asia is particularly impacted by degraded and deteriorating air quality (AQ), particularly in densely-populated areas in and around its major cities. PM10 and PM2.5 (i.e., PM with an aerodynamic diameter of smaller than 10 μm and 2.5 μm, respectively) are common pollutants in the area originating from both natural and anthropogenic sources. Since approximately 1970, Equatorial Asia has experienced significant urbanization [Muntean et al., 2018; United Nations, 2018] which has led to increased PM emissions at the local and regional scales. Emissions from wildfires are also key in dictating unhealthy AQ conditions [Andreae, 1990; Crippa et al., 2016], thus urgent actions are needed to control deforestation rates [Miettinen et al., 2011; van der Werf et al., 2008] and reduce associated social and climatic impacts [United Nations, 2013].
This work focuses on Malaysia, both Peninsular bordering Thailand in the north and Singapore in the south, and the northern portion of the island of Borneo which is shared with Indonesia and Brunei. The country's population was 32.68 million in 2019, with its most populated state, Selangor, on the central-west coast of the peninsula and where the capital city Kuala Lumpur is, with 6.55 million people [Department of Statistics, 2020]. High levels of surface PM10 affect not only public health [Latif et al., 2018; Ostro et al., 2004], but also the local economy through increased health costs, production losses, and lost tourism revenue, among others [Awang et al., 2000; Tacconi, 2016]. In urban areas, the combination of high PM emissions originating from both mobile and stationary sources with specific local meteorological conditions [Afroz et al., 2003] may result in hazardous background PM10 levels even during non-fire periods [Awang et al., 2000; Lee et al., 2018]. Wildfires are also a key regional source of PM10, occurring partly due to natural ignition and partly due to agricultural practices such as land clearing for palm oil or timber plantations [Carlson et al., 2012; Marlier et al., 2015; Purnomo et al., 2018; Wicke et al., 2011]. During the fire season, PM10 emissions can be transported from Sumatra, Indonesia to Malaysian territory within 48 h [Aouizerats et al., 2015; Azmi et al., 2010] often causing severe haze events. For example, in 1997 PM10 values reached up to 20 times background levels, reducing visibility to as little as 50 m and causing closures of local airports [Davies and Unam, 1999], with haze-related PM2.5 concentrations significantly associated with an increase in respiratory tract conditions [Emmanuel, 2000] and causing an estimated 10,800 excess deaths that year [Marlier et al., 2013]. Similarly, during the 2015 fire season background PM concentrations increased by a factor of 30–100 in the region, with an estimated 185 million people persistently exposed to PM10 concentrations exceeding WHO guidelines, and contributed to 11,880 excess deaths [Crippa et al., 2016]. The frequency of wildfire activity and the range of transport of fire emissions are closely related to the regional influence of the wet and dry monsoons [Latif et al., 2018; van der Werf et al., 2008]. During the wet Chinese monsoon in the winter, winds blow from the north-east, generally increasing precipitation and wet deposition processes and hence decreasing fire ignition frequency and pollution transport. Conversely, during the dry Australian monsoon in the late summer, dry winds blow from the south-east, increasing pollution levels by both enhancing dry conditions for fuel ignition and enabling the long-range transport of PM from fires on the islands of Borneo and Sumatra [Hansen et al., 2019; Juneng et al., 2011; Reid et al., 2012]. Additionally, El Niño Southern Oscillation (ENSO) phases influence the severity of the local fire season [Chen et al., 2017; Field et al., 2009], increasing fire counts during dry El Niño years and decreasing them during wet La Niña years [Chen et al., 2016; Marlier et al., 2013; Reid et al., 2012]. The extreme El Niño years of 1997 and 2015 saw record peaks in burned area in the region, resulting in extreme pollution events with dire social and economic effects throughout the Maritime Continent.
Other regional meteorological phenomena at a variety of spatio-temporal scales are also expected to influence mixing conditions in the lower atmosphere and thus impact near-surface PM10 concentrations on a given day at a given location [Reid et al., 2012]. Past studies have sought to identify correlations of PM10 with other pollutants and with meteorology [Azmi et al., 2010; Dominick et al., 2012; Juneng et al., 2011], and found a significant influence of temperature and humidity on PM10 concentrations. Higher temperatures are mostly positively correlated with PM, both because they generally occur within the burning season and thus are associated with high primary emissions from wildfires [Azmi et al., 2010; Dominick et al., 2012], and also because they favor chemical reactions involved in secondary aerosol formation [Seinfeld and Pandis, 2016; Sheehan and Bowman, 2001]. Conversely, high humidity was found to be negatively correlated with PM10 as a result of rainy conditions which increase wet deposition rates [Azmi et al., 2010; Dominick et al., 2012]. Researchers have linked seasonal fluctuations of PM10 to the differing impacts of synoptic circulation over different geographical regions in Malaysia and identified two different intra-seasonal time-scales of PM10 fluctuations, 10–20 days and 30–60 days, which were hypothesized to be indirectly induced by the intra-seasonal variations of the climate of the Maritime Continent [Juneng et al., 2009].
Over the past two decades, these complex pollution dynamics have garnered the attention of the scientific community in regard to Equatorial Asia in general and Malaysia in particular. However, a scarcity of publicly accessible observational datasets in the region has precluded thorough investigations of the spatio-temporal variability of AQ in the region over climatologically relevant time scales both in observational studies and modeling approaches, where ground-based observations are critical to assess models' fidelity. Several observational studies have made use of statistical methods to assess the status of AQ and speculate on pollution sources in Malaysia, with the Greater Klang Valley receiving the most attention due to its position as the biggest urban and industrial hub and often most polluted area in the country [Azmi et al., 2010; Juneng et al., 2011; Rahman et al., 2015]. However, studies making use of observations have so far been limited in their spatial and/or temporal extent, with analyses conducted either on particular regions within Malaysia [Azmi et al., 2010; Latif et al., 2014], over short periods of time [Dominick et al., 2012; Juneng et al., 2009], or both [Juneng et al., 2011; Rahman et al., 2015]. Other studies have made use of regional chemical transport models to attribute pollution contributions from both fire and non-fire sources [Hansen et al., 2019; Lee et al., 2018], with several also focusing on quantifying health impacts associated with increased PM due to wildfire occurrence [Crippa et al., 2016; Marlier et al., 2013; Mead et al., 2018; Reddington et al., 2014]. However, the restricted availability of observational data has limited the temporal extent and/or certainty in regional-scale pollution exposure estimates [Mead et al., 2018].
This study makes use of ground-based long-term PM10 observations from a dense monitoring network established by the Department of Environment (DOE) in Malaysia, which enables assessment of spatio-temporal patterns in AQ experienced by the country. Our study is based on the entire extent of PM10 observations, measured at 52 monitoring sites across Malaysia from 1996 to 2015. These observations are used to investigate the following research questions: (i) to what extent, if any, has Malaysia's AQ deteriorated over time as a consequence of increased urbanization (and/or) fire activity? and (ii) what are the main emission and meteorological variables that are key to explaining spatio-temporal variability in pollution levels across Malaysia? This work aims to answer these questions through statistical data analysis and a statistical model that should provide both a baseline and a reference for future chemical transport modeling studies in the region.
Section snippets
PM10 ground observations data
Fig. 1 shows the location of the 52 DOE stations monitoring PM10 on both the Peninsula and Borneo, along with the year each station began recording data. The sites have been collecting hourly observations of PM10 concentrations through beta attenuation or tapered element oscillating microbalance instruments since 1996 as part of the Continuous Air Quality Monitoring (CAQM) program. Observations have been generally continuous over time, with a percentage of hourly data between each station's
Exploratory analysis of PM10 observations
The time series of PM10 concentrations are analyzed to investigate inter- and intra-annual variability as well as spatio-temporal patterns of AQ conditions (including extreme events) at each station. Specifically, PM10 intra-annual variability is investigated based on monthly averages computed at each location by averaging daily mean PM10 when more than 75% of days in the month are available. Subsequently, all years on record are aggregated to derive monthly profiles of PM10 concentrations
Seasonality of PM10 by region
Regional differences in the median monthly profiles of daily PM10 concentrations suggest the influence of both fire and urban emissions in the particulate pollution of Malaysia (Fig. 3). The profiles show a clear seasonality among most regions, with PM10 concentrations being generally higher during the dry-monsoon period (May–October) than the wet monsoon period (December–March). This seasonal PM10 peak is consistent with the region's increased fire activity in the summer and fall months as
Discussion and concluding remarks
This study makes use of long-term observations of PM10 concentrations recorded at 52 monitoring sites over Malaysia to investigate the influence of urban and fire emissions as well as of key meteorological variables on the observed spatio-temporal variability of PM10. Our findings indicate that:
- •
Meteorological variables are key to explain the variability in observed PM10 concentrations. Many vary by region and season, and are strongly related to the two monsoon climatic patterns influencing the
Author contributions
MA, PC and DB jointly designed the analysis methods. MA performed the analyses. MIM and MTL provided the observational data. MA and PC jointly wrote the paper. All authors provided feedback on the paper and interpretation of the results.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The authors gratefully acknowledge the National University of Malaysia (Universiti Kebangsaan Malaysia, UKM), METMalaysia (Malaysian Meteorological Department) and the Malaysian Department of Environment (DOE) for providing access to the observational data used in this study.
References (56)
- et al.
Review of air pollution and health impacts in Malaysia
Environ. Res.
(2003) - et al.
Impact of urbanization growth on Malaysia CO2 emissions: evidence from the dynamic relationship
J. Clean. Prod.
(2017) - et al.
Smoke-haze from the 1997 Indonesian forest fires: effects on pollution levels, local climate, atmospheric CO2 concentrations, and tree photosynthesis
For. Ecol. Manag.
(1999) - et al.
Spatial assessment of air quality patterns in Malaysia using multivariate analysis
Atmos. Environ.
(2012) - et al.
Spatio-temporal characteristics of PM10 concentration across Malaysia
Atmos. Environ.
(2009) - et al.
Factors influencing the variations of PM10 aerosol dust in Klang Valley, Malaysia during the summer
Atmos. Environ.
(2011) - et al.
Long term assessment of air quality from a background station on the Malaysian Peninsula
Sci. Total Environ.
(2014) Impact of regional haze towards air quality in Malaysia: a review
Atmos. Environ.
(2018)- et al.
Reducing forest and land fires through good palm oil value chain governance
Forest Policy Econ.
(2018) - et al.
Correlations between fine particulate matter (PM2. 5) and meteorological variables in the United States: Implications for the sensitivity of PM2. 5 to climate change
Atmos. Environ.
(2010)
Exploring land use changes and the role of palm oil production in Indonesia and Malaysia
Land Use Policy
Biomass burning in the tropics: impact on environmental quality and global climate
Popul. Dev. Rev.
Importance of transboundary transport of biomass burning emissions to regional air quality in Southeast Asia during a high fire event
Atmos. Chem. Phys.
Air quality in Malaysia: impacts, management issues and future challenges
Respirology
Trend and status of air quality at three different monitoring stations in the Klang Valley, Malaysia, Air Quality
Atmos. Health
Committed carbon emissions, deforestation, and community land conversion from oil palm plantation expansion in West Kalimantan, Indonesia
Proc. Natl. Acad. Sci.
The 2015 Borneo fires: what have we learned from the 1997 and 2006 El Niños?
Environ. Res. Lett.
A pan-tropical cascade of fire driven by El Niño/Southern Oscillation
Nat. Clim. Chang.
Population exposure to hazardous air quality due to the 2015 fires in Equatorial Asia
Sci. Rep.
Gridded emissions of air pollutants for the period 1970–2012 within EDGAR v4.3.2
Earth Syst. Sci. Data
Fossil CO2 and GHG Emissions of all World Countries
High resolution temporal profiles in the Emissions Database for Global Atmospheric Research
Sci. Data.
Effects of aging on organic aerosol from open biomass burning smoke in aircraft and laboratory studies
Atmos. Chem. Phys.
New Malaysia Ambient Air Quality Standard, Edited
Demographic Statistics Fourth Quarter 2019, Malaysia
East Asia’s Changing Urban Landscape: Measuring a Decade of Spatial Growth
Impact to lung health of haze from forest fires: the Singapore experience
Respirology
Human amplification of drought-induced biomass burning in Indonesia since 1960
Nat. Geosci.
Cited by (9)
Spatio-temporal variations and socio-economic drivers of air pollution: Evidence from 332 Chinese prefecture-level cities
2023, Atmospheric Pollution ResearchSpatial assessment of PM10 hotspots using Random Forest, K-Nearest Neighbour and Naïve Bayes
2021, Atmospheric Pollution ResearchCitation Excerpt :This can be a good reason why the best model performance classified the map's central area as regions with high PM10 hotspot. It is not surprising that several studies have been carried out around this region as well (Rahman et al., 2015; Khan et al., 2016; Alifa et al., 2020). These indicate that the RF model's classification is more scientific, realistic and acceptable.
Air quality impacts of the 2019–2020 Black Summer wildfires on Australian schools
2021, Atmospheric EnvironmentCitation Excerpt :Thus, high temperatures and low humidity can be related to PM2.5 emissions via increased fire activity and intensity (McClure and Jaffe, 2018; Val Martin et al., 2015). This study also characterised the non-linearity inherent in local meteorology-PM2.5 relationships during Black Summer, whereas previous studies have identified significant linear influences of meteorological conditions on particulate concentrations during fires (e.g. Alifa et al., 2020; Balachandran et al., 2017). The non-linear covariate relationships noted here may partially indicate the time- and location-dependent passage and evolution of weather systems across the south-eastern Australian landmass and surrounding oceans, their interactions with the wildfires, and also interactions with features such as complex topography and the urban area (Jiang et al., 2012, Fig. 1).