Abstract This study evaluates the performance of a low-cost air quality monitoring device, the Oizom Polludrone, for sustainable air quality assessment in Kuala Terengganu, Malaysia. The objectives were to analyse temporal variations of major air pollutants and meteorological parameters, examine relationships between pollutants and meteorological factors, and assess the reliability of the low-cost device through comparison with data from the Department of Environment (DOE) station. Hourly data on particulate matter (PM₁, PM_2.5, and PM₁₀), gaseous pollutants (CO, NO₂, SO₂, and O₃), and meteorological variables (temperature, relative humidity, wind speed, ultraviolet radiation, and atmospheric pressure) were collected over a three-month monitoring period. Results showed that particulate matter concentrations were consistently higher than gaseous pollutants, identifying particulate pollution as a dominant air quality concern in the study area. Strong positive correlations were observed between PM_2.5 and PM₁₀ (r = 0.997) as well as PM₁  (r = 0.995), indicating common emission sources and similar atmospheric behaviour. Meteorological parameters, particularly wind speed and relative humidity, were found to influence pollutant dispersion and accumulation. Comparison with DOE reference data yielded a moderate level of agreement for PM_2.5 measurements (R² = 0.3549), indicating that the low-cost device captured general pollution trends despite some variability. Overall, the findings demonstrate that the Oizom Polludrone is a reliable and cost-effective tool for supplementary and indicative air quality monitoring in tropical urban environments such as Kuala Terengganu. While not intended to replace regulatory-grade monitoring stations, its application can enhance spatial and temporal air quality coverage, supporting sustainable urban air quality management.

Keywords: Low-cost air quality sensor, PM_2.5, Sensor validation, Tropical climate, Urban air pollution

Citation:  Rifin, M.S.N.M., Ahmad, A.N., Mansor, A.A., Ramly, Z.T.A., and Abdullah, S. 2026. Performance evaluation of a low-cost air quality monitoring device under tropical urban conditions in Kuala Terengganu, Malaysia. Natural and Life Sciences Communications. 25(3): e2026061.

Graphical Abstract 

INTRODUCTION

Maintaining air quality has become increasingly important due to rapid industrialisation, urbanisation, population growth, and intensified exploitation of natural resources (Guo et al., 2024; Qiao et al., 2025). These processes have contributed to a steady rise in air pollution, which remains a serious environmental issue worldwide. Air pollution is composed of a complex mixture of particulate matter (PM) and gaseous pollutants, including carbon monoxide (CO), nitrogen dioxide (NO₂), sulfur dioxide (SO₂), and ozone (O₃), primarily originating from industrial activities and vehicular emissions (Guo et al., 2023; Radarit et al., 2024; Ramly et al., 2025). Numerous studies have demonstrated strong associations between air pollution exposure and adverse human health outcomes, such as lung cancer, cardiovascular disease, chronic bronchitis, asthma, respiratory disorders, reduced life expectancy, and premature mortality (Leelasittikul et al., 2020; Spyropoulos et al., 2021). According to estimates by the World Health Organization, increasing urban air pollution in developing countries has resulted in more than two million premature deaths annually, in addition to widespread respiratory illnesses (Guo et al., 2021). In Malaysia, growing concerns over ambient air quality are largely driven by rapid industrialisation, urban expansion, and transboundary air pollution events (Abdullah et al., 2018). Industrial activities and fossil fuel combustion represent major sources of air pollution with exposure to ozone and airborne particulate matter linked to increased hospital admissions and mortality rates (Usmani et al., 2020).

Conventional air quality monitoring in Malaysia is conducted through Continuous Air Quality Monitoring (CAQM) stations operated by the Department of Environment (DOE) (Abdullah et al., 2019). These stations employ reference-grade instruments that provide accurate and reliable measurements but require substantial financial investment for installation, operation, and maintenance. Consequently, their spatial coverage is limited, particularly in smaller cities and semi-urban regions. In response, low-cost air quality monitoring sensors have increasingly been adopted to supplement existing monitoring networks (Shabbir et al., 2025). These sensors enable near–real-time air quality monitoring, are relatively easy to operate, and require minimal maintenance (Shahid et al., 2025). An additional advantage lies in their scalability, which allows dense monitoring networks to be established at lower cost, thereby enhancing spatial resolution of air quality data (Chojer et al., 2020). Low-cost air pollution monitoring devices typically integrate multiple sensors with auxiliary components to detect, quantify, and transmit data on targeted pollutants and associated meteorological parameters (US EPA, 2024). These devices facilitate rapid responses to changes in air quality conditions and support data-driven environmental management strategies (Oizom, 2024). One such device is the Oizom Polludrone, which has been developed to provide real-time measurements of ambient air pollutants. The Polludrone is equipped with sensors capable of measuring particulate matter (PM₁, PM_2.5, and PM₁₀), gaseous pollutants including CO, NO₂, SO₂, and O₃, as well as meteorological parameters such as temperature, relative humidity, wind speed, atmospheric pressure, and ultraviolet (UV) radiation (Oizom, 2024).

Although low-cost sensors are increasingly used for air quality assessment, their measurement accuracy can be influenced by environmental factors such as humidity, temperature, and wind conditions, particularly in tropical climates. As a result, performance evaluation against reference-grade monitoring stations is essential to determine their reliability and suitability for scientific and regulatory applications (Suriano and Prato, 2023). Previous studies conducted in Malaysia have highlighted the influence of meteorological conditions on pollutant concentrations and sensor performance, particularly in urban environments (Asy’aari et al., 2020). However, research focusing on low-cost air quality sensor performance under the humid tropical conditions of Malaysia’s east coast remains limited. Through this integrated assessment, the study provides insights into air quality characteristics, meteorological influences, and the reliability of low-cost monitoring technologies in a coastal tropical environment, contributing to improved air quality management strategies in Malaysia.

MATERIALS AND METHODS

Site selection

Air quality monitoring was conducted at two urban locations (Figure 1) along the east coast of Peninsular Malaysia, both situated in Kuala Terengganu, Terengganu State. The first site was Sekolah Kebangsaan Chabang Tiga (latitude 05°18′29.13″ N, longitude 103°07′13.41″ E), which hosts a Department of Environment (DOE) Continuous Air Quality Monitoring (CAQM) station. The second site was Universiti Malaysia Terengganu (UMT) (latitude 5°24′27″ N, longitude 103°05′17″ E), an urban campus environment influenced by traffic, academic activities, and nearby residential areas. The Oizom Polludrone device was deployed at the UMT site, while reference air quality data were obtained from the DOE CAQM station at Sekolah Kebangsaan Chabang Tiga. The straight-line distance between the Universiti Malaysia Terengganu (UMT) site and the Department of Environment (DOE) CAQM station at Sekolah Kebangsaan Chabang Tiga is approximately 11.79 km. Both sites are located within the same urban airshed of Kuala Terengganu and are influenced primarily by traffic emissions, small-scale commercial activities and residential combustion sources. No major heavy industrial facilities are located within a 3 km radius of either monitoring site. The UMT site is situated near campus access roads and moderate daily traffic flows, while the DOE station is positioned within a mixed residential and commercial zone exposed to urban traffic emissions. The Oizom Polludrone and the DOE CAQM station were not co-located but were situated within the same urban zone of Kuala Terengganu and influenced by similar emission sources. The objective of this comparison was to evaluate indicative sensor performance under real-world conditions rather than regulatory compliance. Differences in micro-environmental conditions, including surrounding land use, airflow patterns, and installation height, were therefore expected and considered during data interpretation. The Polludrone sensor was installed outdoors at approximately 3 m above ground level in an open area, away from direct obstructions, to ensure adequate air circulation.

Figure 1. Location of the air quality monitoring sites in Kuala Terengganu, Terengganu, Malaysia, showing the Universiti Malaysia Terengganu (UMT) site and the Department of Environment (DOE) Continuous Air Quality Monitoring (CAQM) station at Sekolah Kebangsaan Chabang Tiga.

Data collection

Air quality monitoring was conducted over a continuous three-month period at each site. Pollutant concentrations and meteorological parameters were recorded at 1-hour intervals, resulting in 4,320 data points per monitoring location. The measured air pollutants included particulate matter (PM₁, PM_2.5, and PM₁₀) and gaseous pollutants (carbon monoxide, CO; nitrogen dioxide, NO₂; ozone, O₃; sulfur dioxide, SO₂). Meteorological parameters included air temperature, relative humidity, wind speed, atmospheric pressure, and ultraviolet (UV) radiation. All air pollutant and meteorological measurements at the UMT site were obtained using the Oizom Polludrone, a low-cost, smart air quality monitoring device designed for continuous outdoor monitoring (Figure 2). The Polludrone employs a combination of laser-based optical sensors for particulate matter, electrochemical sensors for gaseous pollutants, and non-dispersive infrared (NDIR) technology. Meteorological sensors are integrated into the system to capture ambient environmental conditions. The device was factory-calibrated prior to deployment, following manufacturer specifications, and operated continuously throughout the monitoring period. Reference air quality data were obtained from the DOE CAQM station located at Sekolah Kebangsaan Chabang Tiga (Figure 3). The CAQM station employs reference-grade instruments that comply with national and international air quality monitoring standards. (Ash'aari et al., 2020) These include technologies such as Beta Attenuation Monitors (BAM) or equivalent Federal Equivalent Methods (FEM) for particulate matter, and certified gas analysers for gaseous pollutants.  Hourly averaged data from the DOE station were obtained for comparative analysis with the Polludrone measurements.

Figure 2. Oizom Polludrone low-cost air quality monitoring device used in this study, equipped with particulate matter, gaseous pollutant, and meteorological sensors.

Figure 3. Department of Environment (DOE) Continuous Air Quality Monitoring (CAQM) station at Sekolah Kebangsaan Chabang Tiga, Kuala Terengganu, serving as the reference monitoring site (DOE, 2022).

Data analysis

Prior to analysis, all datasets underwent quality control and preprocessing procedures. Data records affected by sensor downtime, communication failure, or obvious instrument malfunction were excluded from the analysis. Missing data were not interpolated, only valid hourly observations recorded simultaneously by the respective instruments were retained. Descriptive statistical analysis was performed to characterise the distribution and temporal variability of air pollutants and meteorological parameters. Statistical indicators included mean, median, variance, standard deviation, skewness, and kurtosis (Kawichai et al., 2021). Spearman rank correlation analysis was used to assess relationships between air pollutants and meteorological parameters. This non-parametric approach was selected because several variables exhibited non-normal distributions, as indicated by high skewness and kurtosis values. Correlation coefficients were interpreted based on their magnitude, with values greater than 0.5 considered strong and values below 0.5 considered weak (Azman et al., 2024). Statistical significance was assessed at the 0.01 and 0.05 levels. All correlation analyses were conducted using IBM SPSS Statistics. Performance evaluation of the Oizom Polludrone was carried out through comparison with DOE CAQM reference data. PM_2.5 concentrations were selected for performance assessment due to their relevance to public health and data availability at both monitoring sites. Performance evaluation was limited to PM_2.5 due to data availability constraints from the DOE real time monitoring system. While other pollutants are displayed through the Air Pollutant Index (API), raw hourly concentration data are not publicly accessible for quantitative comparison. Therefore, statistical performance indicators including R², Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Mean Bias Error (MBE) were calculated for PM_2.5 only. Hourly PM_2.5 data from the Polludrone and DOE station were temporally aligned based on corresponding timestamps. Linear regression analysis was applied to examine the relationship between Polludrone and DOE PM_2.5 measurements. The regression model quantified the degree of agreement between the two datasets using the regression equation and coefficient of determination (R²). No calibration or correction model was applied in order to assess the raw, out-of-the-box performance of the low-cost sensor under real-world tropical conditions.

RESULTS

Trend of air pollutants in urban areas.

All descriptive and correlation analyses were based on Polludrone measurements collected at the UMT site. The descriptive statistical summary of air pollutants measured in Kuala Terengganu is presented in Table 1. Carbon monoxide (CO) recorded a mean concentration of 0.40 ± 0.16 mg/m³, with a median value of 0.37 mg/m³ and low variance (0.02), indicating relatively stable concentrations during the monitoring period. The CO distribution was highly positively skewed (skewness = 8.14) with very high kurtosis (113.07), reflecting predominantly low concentrations with occasional extreme values. Nitrogen dioxide (NO₂) showed a low mean concentration of 3.58 ± 8.48 µg/m³. The large variance (71.90), combined with high skewness (12.45) and kurtosis (308.56), indicates substantial episodic variability. Ozone (O₃) concentrations had a mean of 17.55 ± 45.17 µg/m³ and a median of 12.01 µg/m³, with very high variance (2,040.74), skewness (11.35), and kurtosis (141.54), demonstrating considerable fluctuation throughout the study period. Sulfur dioxide (SO₂) exhibited a mean value of 29.78 ± 340.90 ppb with extremely high variance (116,214.98), skewness (13.79), and kurtosis (209.09), indicating infrequent but intense concentration spikes.

Particulate matter concentrations were generally higher and more consistently distributed than gaseous pollutants. PM_2.5 recorded a mean concentration of 22.24 ± 14.50 µg/m³ and a median of 19.26 µg/m³, with moderate variance (210.20), skewness (1.05), and kurtosis (0.84). PM₁₀ showed a higher mean concentration of 43.23 ± 24.94 µg/m³ and a median of 38.27 µg/m³, with variance of 621.82, skewness of 1.03, and kurtosis of 0.88. Similarly, PM₁ had a mean concentration of 19.51 ± 13.23 µg/m³ and a median of 16.81 µg/m³, with moderate variance (175.03), skewness (1.00), and kurtosis (0.67). These results indicate that particulate matter concentrations followed approximately normal distributions with slight positive skewness.

Meteorological parameters exhibited characteristic tropical coastal conditions. Relative humidity (RH) recorded a high mean of 87.40 ± 9.05%, with a median of 89.73%, moderate variance (81.90), and slight negative skewness (-0.55). Atmospheric pressure showed a mean of 1,008.56 ± 9.64 hPa and a median of 1,009.51 hPa, with high kurtosis (183.35) and strong negative skewness (-12.76), indicating occasional low-pressure events. Temperature averaged 29.20 ± 3.37°C with limited variability (variance = 11.38), while the UV index exhibited high variability (1.98 ± 2.85) and positive skewness (1.29). Wind speed averaged 1.32 ± 3.03 m/s, with high skewness (9.58) and kurtosis (101.32), reflecting predominantly calm conditions interrupted by sporadic strong wind events.

Table 1. Summary statistics of air pollutant concentrations and meteorological parameters measured in Kuala Terengganu using the Oizom Polludrone during the study period.

Correlation between air pollutants and meteorological factors in Kuala Terengganu

Spearman correlation analysis between air pollutants and meteorological parameters is summarised in Figure 4. Carbon monoxide (CO) showed strong positive correlations with PM_2.5 (r = 0.571), PM₁₀ (r = 0.559), and PM₁ (r = 0.580), while weak positive correlations were observed with NO₂ (r = 0.262) and SO₂ (r = 0.186). CO exhibited weak negative correlations with atmospheric pressure (r = -0.176), relative humidity (r = -0.051), and wind speed (r = -0.003), and weak positive correlations with temperature (r = 0.237) and UV index (r = 0.113). Nitrogen dioxide (NO₂) displayed weak positive correlations with relative humidity (r = 0.279) and weak negative correlations with temperature (r = -0.198), UV index (r = -0.347), and wind speed (r = -0.196). Correlations between NO₂ and particulate matter were very weak. Ozone (O₃) showed moderate positive correlations with temperature (r = 0.491) and UV index (r = 0.383), while moderate negative correlation was observed with relative humidity (r = -0.530). Weak negative correlations were also recorded between O₃ and pressure (r = -0.277). Sulfur dioxide (SO₂) exhibited very weak positive correlations with particulate matter fractions (PM_2.5, PM₁₀, and PM₁) and weak negative correlations with pressure (r = -0.125) and wind speed (r = -0.064). Particulate matter fractions were very strongly intercorrelated, particularly PM_2.5 with PM₁₀ (r = 0.997) and PM₁ (r = 0.995), and PM₁₀ with PM₁ (r = 0.986). Weak negative correlations were generally observed between particulate matter and pressure, UV index, and wind speed, while weak positive correlations were observed with temperature. Among meteorological parameters, relative humidity showed strong negative correlations with temperature (r = -0.911) and UV index (r = -0.773). Temperature exhibited a strong positive correlation with UV index (r = 0.788), while wind speed showed weak positive correlations with temperature (r = 0.233) and UV index (r = 0.381).

Figure 4. Spearman correlation matrix illustrating relationships between air pollutants and meteorological parameters measured in Kuala Terengganu during the monitoring period.

The performance of the low-cost air monitoring device of Oizom Polludrone with the reference data in Kuala Terengganu.

The comparative statistical summary between Oizom Polludrone and DOE measurements is presented in Table 2. The DOE data recorded a mean concentration of 23.40 ± 12.07 µg/m³, while the Oizom Polludrone recorded a comparable mean of 22.32 ± 14.74 µg/m³. Median concentrations were 21.15 µg/m³ for DOE and 19.26 µg/m³ for Oizom. Both datasets exhibited positive skewness (DOE: 1.28; Oizom: 1.08) and moderate kurtosis (DOE: 1.50; Oizom: 0.92). The Oizom Polludrone exhibited a wider concentration range (82.46 µg/m³) compared to DOE (59.64 µg/m³), with lower minimum and higher maximum recorded values. The root mean squared error (RMSE) was 21.33 µg/m³, while the mean absolute error (MAE) was 15.52 µg/m³. The mean bias error (MBE) was 10.98 µg/m³, indicating a tendency of the Oizom Polludrone to overestimate PM_2.5 concentrations relative to the DOE reference station. The regression analysis was conducted using temporally aligned hourly PM2.5 data from the Polludrone and DOE CAQM station. Linear regression analysis between DOE and Oizom Polludrone data yielded the regression equation y = 0.7271x + 5.3 with a coefficient of determination of R² = 0.3549 (Figure 5). This indicates a moderate level of agreement between the two datasets, with approximately 35.49% of the variance in Polludrone measurements explained by the DOE reference values.

Table 2. Comparative descriptive statistics of hourly PM_2.5 concentrations measured by the Oizom Polludrone and the DOE Continuous Air Quality Monitoring (CAQM) station.

Figure 5. Linear regression relationship between hourly PM_2.5 concentrations measured by the Oizom Polludrone and the DOE reference station in Kuala Terengganu.

DISCUSSION

Overall, the DOE reference data recorded a slightly higher mean pollutant concentration compared to the Oizom Polludrone, although the difference between the two instruments was relatively small. This close agreement indicates that the Oizom sensor is capable of capturing general air pollution trends with reasonable accuracy, despite being a low-cost monitoring device. Similar findings have been reported in recent studies, which demonstrate that modern low-cost air quality sensors can provide reliable indicative measurements when deployed under stable ambient conditions (Liu et al., 2020; Singh et al., 2021; Rahman et al., 2022). However, the larger standard deviation observed in the Oizom dataset indicates greater variability in its measurements, reflecting the inherent sensitivity of low-cost sensors to micro-environmental fluctuations such as airflow changes, humidity variation, and sensor calibration drift (Kang et al., 2022). The slightly lower median values recorded by the Oizom Polludrone suggest a tendency to underestimate pollutant concentrations during low-exposure periods. This behaviour is consistent with previous research indicating that optical-based sensors may underestimate concentrations when particle levels are low or influenced by moisture, due to limitations in light-scattering algorithms (Bulot et al., 2020). Nevertheless, both instruments exhibited positively skewed distributions, indicating the occurrence of episodic high-pollution events. In tropical environments such as Malaysia, these peaks are often linked to traffic surges, biomass burning, and rapid meteorological transitions during monsoon periods (Fong et al., 2018; Yin, 2020). The similarity in skewness values demonstrates that the Oizom Polludrone is capable of detecting pollution episodes concurrently with the reference instrument, even if discrepancies exist in magnitude.

Differences in kurtosis between the two datasets further highlight the contrasting measurement characteristics of regulatory-grade and low-cost sensors. The DOE reference monitor displayed a more stable distribution, while the Oizom device captured greater short-term variability. This observation aligns with earlier studies showing that low-cost sensors are more responsive to rapid atmospheric changes, whereas reference-grade instruments are designed for long-term stability (Ramadani et al., 2025). These differences reflect fundamental variations in sensor design rather than measurement errors (Diez et al., 2024). The wider data range observed in the Oizom measurements represents one of the most notable distinctions between the two monitoring systems. Low-cost sensors are known to respond more strongly to sudden environmental changes, often producing more extreme readings during short-lived pollution events (Macías-Hernández et al., 2023; Liu et al., 2024). The lower minimum values recorded by the Oizom Polludrone may be attributed to sensor zero-drift or transient atmospheric conditions that momentarily reduce detected particle counts, a phenomenon commonly documented in low-cost sensor evaluations (Atfeh et al., 2025). Conversely, the higher maximum values observed may indicate overestimation during high-pollution episodes, particularly under conditions of elevated humidity or rapid changes in aerosol composition (Bran et al., 2022). While such behaviour suggests the need for periodic calibration or correction factors (Mei et al., 2025), it also confirms the sensor’s capacity to detect pollution surges, which is essential for real-time monitoring applications.

The moderate coefficient of determination (R² = 0.3549) obtained from the regression analysis reflects partial agreement between the Oizom Polludrone and the DOE reference data. This level of correlation is comparable to previous studies evaluating low-cost sensors against regulatory instruments (Chatoutsidou et al., 2025). The remaining unexplained variability likely arises from differences in measurement principles, sensor sensitivity to environmental conditions, and spatial separation between monitoring sites. Reference-grade DOE stations employ technologies such as Beta Attenuation Monitors and chemiluminescence analysers, which provide high accuracy and stability. In contrast, the Oizom Polludrone relies on optical, electrochemical, and NDIR sensors that are more susceptible to environmental interference (Bucek et al., 2021).

Non-co-location of the instruments may have further contributed to the observed discrepancies. Small differences in sensor height, airflow patterns, nearby emission sources, and surrounding structures can substantially influence measured pollutant concentrations, particularly in urban and semi-urban environments (Manu and Rysanek, 2024; Mai et al., 2025). Additionally, environmental factors such as humidity, temperature, and aerosol composition play a critical role in low-cost sensor accuracy. Previous studies emphasize the importance of applying humidity compensation and calibration models to improve agreement with reference measurements (Patra et al., 2021). As no correction model was applied in this study, environmental variability likely contributed to the scatter observed in the regression results.

Overall, the findings indicate that while the Oizom Polludrone may not provide measurements suitable for regulatory or compliance-based assessments, it performs effectively as an indicative monitoring tool. The device successfully captures temporal variability and general pollution trends, making it valuable for supplementary monitoring, community-scale exposure assessment, and studies requiring high spatial coverage (Chieh et al., 2025; Marto et al., 2025). These results support existing evidence that low-cost sensors, when appropriately calibrated and carefully deployed, can complement conventional air quality monitoring networks by enhancing data availability and spatial resolution (Dangare et al., 2025). From a policy perspective, the integration of low-cost sensors such as the Polludrone into Malaysia’s existing air quality monitoring framework could significantly enhance spatial coverage, particularly in semi-urban and coastal regions where regulatory stations are limited.

CONCLUSION

PM_2.5 and PM₁₀ recorded higher concentrations than gaseous pollutants, indicating that particulate pollution remains a key air quality concern in urban Kuala Terengganu. Elevated particulate levels were associated with traffic emissions, construction activity, and other urban processes, compounded by limited atmospheric dispersion under generally low wind speeds. In contrast, gaseous pollutants such as CO, NO₂, and O₃ exhibited comparatively lower mean concentrations, reflecting the lower intensity of industrial activity on the east coast relative to more urbanised regions of Peninsular Malaysia. Meteorological conditions during the monitoring period were characterised by high relative humidity, warm temperatures, and predominantly calm winds, which are typical of Terengganu’s tropical coastal climate. Performance evaluation results demonstrate that the Oizom Polludrone exhibits reasonable agreement with DOE reference measurements, particularly in capturing temporal variability and general pollution trends. Although discrepancies were observed, attributable to sensor sensitivity to environmental factors and inherent limitations of low-cost instruments, the regression analysis confirmed a moderate level of agreement between the two datasets. These findings indicate that the Oizom Polludrone is suitable as a supplementary monitoring tool rather than a replacement for regulatory-grade stations. In conclusion, this study confirms that low-cost air quality monitoring devices such as the Oizom Polludrone offer significant potential to enhance the spatial and temporal coverage of air quality data in Malaysia. When appropriately calibrated and integrated with existing monitoring networks, such devices can provide valuable real-time information to support air quality assessment, early detection of pollution episodes, and informed environmental management.

ACKNOWLEDGEMENTS

We acknowledge Universiti Malaysia Terengganu by providing a Matching Grant 1+2 (Ref: UMT/PPP/2- 2/2/15 Jld.4 (71)) (VOT: 53598) for funding this study. Additionally, we would like to express our gratitude to the Air Quality Division of the Malaysian Department of Environment for the air quality data.

AUTHOR CONTRIBUTIONS

Muhammad Syakirul Naim Mat Rifin: Conceptualization (Lead), Data Curation (Lead), Formal Analysis (Lead), Investigation (Lead), Writing-Original Draft (Lead); Aimi Nursyahirah Ahmad: Investigation (Equal), Data Curation (Equal), Writing-Review & Editing (Equal); Amalina Abu Mansor: Methodology (Lead), Validation (Lead); Zamzam Tuah Ahmad Ramly: Formal Analysis (Supporting), Visualization (Lead); Samsuri Abdullah: Conceptualization (Supporting), Supervision (Lead), Project Administration (Lead), Writing-Review & Editing (Lead), Funding Acquisition (Lead).

CONFLICT OF INTEREST

The authors declare that they have no conflicts of interest.

REFERENCES

Abdullah, S., Ismail, M., and Ahmed, A.N. 2018. Identification of air pollution potential sources through principal component analysis (PCA). International Journal of Civil Engineering and Technology. 9: 1435-1442.

Abdullah, S., Ismail, M., and Ahmed, A.N. 2019. Multi-layer perceptron model for air quality prediction. Malaysian Journal of Mathematical Sciences. 13: 85-95.

Ash'aari, Z.H., Aris, A.Z., Ezani, E., Kamal, N.I.A., Jaafar, N., Jahaya, J.N., Manan, S.A., and Saifuddin, M.F.U. 2020. Spatiotemporal variations and contributing factors of air pollutant concentrations in Malaysia during movement control order due to pandemic COVID-19. Aerosol and Air Quality Research. 20: 2047-2061. https://doi.org/10.4209/aaqr.2020.06.0334

Atfeh, B., Barcza, Z., Groma, V., Tordai, Á.V., and Mészáros, R. 2025. Performance assessment of low-and medium-cost PM_2.5 sensors in real-world conditions in central Europe. Atmosphere. 16: 796. https://doi.org/10.3390/atmos16070796

Azman, M.S.M., Mansor, A.A., Ahmad, A.N., Ismail, M., Jarkoni, M.N.K., and Abdullah, S. 2025. Physio-chemical indoor air quality analysis and CO₂ ventilation forecasting using artificial neural networks in boat manufacturing. Natural and Life Sciences Communications. 24: e2025014. https://doi.org/10.12982/NLSC.2025.014

Bran, S.H., Macatangay, R., Surapipith, V., Chotamonsak, C., Chantara, S., Han, Z., and Li, J. 2022. Surface PM_2.5 mass concentrations during the dry season over northern Thailand: Sensitivity to model aerosol chemical schemes and the effects on regional meteorology. Atmospheric Research. 277: 106303. https://doi.org/10.1016/j.atmosres.2022.106303

Buček, P., Maršolek, P., and Bílek, J. 2021. Low-cost sensors for air quality monitoring-the current state of the technology and a use overview. Chemistry-Didactics-Ecology-Metrology. 26: 41-54. https://doi.org/10.2478/cdem-2021-0003 

Bulot, F., Russell, H., Rezaei, M., Johnson, M., Ossont, S., Morris, A., Basford, P., Easton, N., Foster, G., Loxham, M., et al. 2020. Laboratory comparison of low-cost particulate matter sensors to measure transient events of pollution. Sensors. 23: 7657. https://doi.org/10.3390/s23177657

Chatoutsidou, S.E., Chalvatzaki, E., Mihalopoulos, N., Kassandros, T., Bagkis, E., Karatzas, K., Melas, D., and Lazaridis, M. 2025. Evaluating low-cost sensors for particle mass concentrations, personal exposure and internal dose characterization at Eastern Mediterranean sites: Can they stand as efficient alternatives? Journal of Aerosol Science. 191: 106711. https://doi.org/10.1016/j.jaerosci.2025.106711

Chieh, T.C., Lung, S.C.C., Chang, L.T., Liu, C.H., Tsou, M.C.M., and Wen, T.Y.J. 2025. Long-term monitoring of particulate matter in an Asian community using research-grade low-cost sensors. Environmental Monitoring and Assessment. 197: 653. https://doi.org/10.1007/s10661-025-14098-z

Chojer, H., Branco, P.T.B.S., Martins, F.G., Alvim-Ferraz, M.C.M., and Sousa, S.I.V. 2020. Development of low-cost indoor air quality monitoring devices: Recent advancements. Science of The Total Environment. 727: 138385. https://doi.org/10.1016/j.scitotenv.2020.138385

Dangare, T., Matandirotya, N.R., Handizvihwe, G., Mathe, P., Mushore, T.D., Matandirotya, E., and Mashonjowa, E. 2025. Leveraging low-cost sensors and machine learning for air quality insights in an urban location of Zimbabwe: A case study. Scientific African. 30: e02992. https://doi.org/10.1016/j.sciaf.2025.e02992

Diez, S., Lacy, S., Coe, H., Urquiza, J., Priestman, M., Flynn, M., Marsden, N., Martin, N., Gillott, S., Bannan, T., et al. 2024. Long-term evaluation of commercial air quality sensors: An overview from the QUANT (Quantification of utility of atmospheric network technologies) study. Atmospheric Measurement Techniques. 17: 3809-3827. https://doi.org/10.5194/amt-17-3809-2024

Fong, S.Y., Abdullah, S., and Ismail, M. 2018. Forecasting of particulate matter (PM₁₀) concentration based on gaseous pollutants and meteorological factors for different monsoons of urban coastal area in Terengganu. Journal of Sustainability Science and Management. 13: 3-17.

Guo, Q., He, Z., and Wang, Z. 2023. Change in air quality during 2014-2021 in Jinan City in China and its influencing factors. Toxics. 11: 210. https://doi.org/10.3390/toxics11030210

Guo, Q., He, Z., and Wang, Z. 2024. The characteristics of air quality changes in Hohhot City in China and their relationship with meteorological and socio-economic factors. Aerosol and Air Quality Research. 24: 230274. https://doi.org/10.4209/aaqr.230274

Guo, Q., Wang, Z., He, Z., Li, X., Meng, J., Hou, Z., and Yang, J. 2021. Changes in air quality from the COVID to the post-COVID era in the Beijing-Tianjin-Tangshan region in China. Aerosol and Air Quality Research. 21: 210270. https://doi.org/10.4209/aaqr.210270 

Kang, Y., Aye, L., Ngo, T.D., and Zhou, J. 2022. Performance evaluation of low-cost air quality sensors: A review. Science of The Total Environment. 818: 151769. https://doi.org/10.1016/j.scitotenv.2021.151769

Kawichai, S., Prapamontol, T., Cao, F., Liu, X.Y., Song, W.H., Kiatwattanacharoen, S., and Zhang, Y.L. 2021. Significant contribution of C₃-type forest plants' burning to airborne PM_2.5 pollutions in Chiang Mai Province, Northern Thailand. Chiang Mai University Journal of Natural Sciences. 20: e2021088. https://doi.org/10.12982/CMUJNS.2021.088

Leelasittikul, K., Koonkumchoo, P., Buranapuntalug, S., Pongpanit, K., and Yuenyongchaiwat, K. 2020. Effect of air quality on cardio-respiratory systems in Northern Thailand (Chiang Mai, Chiang Rai and Nan Province). Chiang Mai University Journal of Natural Sciences. 19(2): 702-717. https://doi.org/10.12982/CMUJNS.2020.0045

Liu, X., Jayaratne, R., Thai, P., Kuhn, T., Zing, I., Christensen, B., Lamont, R., Dunbabin, M., Zhu, S., Gao, J., et al. 2020. Low-cost sensors as an alternative for long-term air quality monitoring. Environmental Research. 185: 109438. https://doi.org/10.1016/j.envres.2020.109438

Liu, X., Pushpawela, B., Jayaratne, R., Thai, P., Christensen, B., Zing, I., Lamont, R., Dunbabin, M., Wang, H., Gao, J., et al. 2024. Response of low-cost sensors to high PM_2.5 concentrations during bushfire and haze events. Aerosol Science and Technology. 58(9): 978 - 990. https://doi.org/10.1080/02786826.2024.2368733

Macías-Hernández, B.A., Tello-Leal, E., Barrios, O., Leiva-Guzmán, M.A., and Toro, R. 2023. Effect of environmental conditions on the performance of a low-cost atmospheric particulate matter sensor. Urban Climate. 52: 101753. https://doi.org/10.1016/j.uclim.2023.101753

Mai, C., Wang, Z., Chen, L., Huang, Y., Li, M., Shirazi, A., Altaee, A., and Zhou, J.L. 2025. Field-based calibration and operation of low-cost sensors for particulate matter by linear and nonlinear methods. Atmospheric Pollution Research. 16(12): 102676. https://doi.org/10.1016/j.apr.2025.102676

Manu, S. and Rysanek, A. 2024. A co-location study of 87 low-cost environmental monitors: Assessing outliers, variability, and uncertainty. Buildings. 14(9): 2965. https://doi.org/10.3390/buildings14092965

Marto, J.P., Lemley, G.M., and Moronta, D.H. 2025. Mapping street-level air pollution: One year of mobile monitoring in disadvantaged communities across New York State. Journal of the Air & Waste Management Association. 75(10): 808-822. https://doi.org/10.1080/10962247.2025.2547638

Mei, H., Wei, P., Ghadikolaei, M., Gali, N., Wang, Y., and Ning, Z. 2025. Performance validation and calibration conditions for novel dynamic baseline tracking air sensors in long-term field monitoring. Atmospheric Measurement Techniques. 18(8): 1771-1785. https://doi.org/10.5194/amt-18-1771-2025

Oizom Instruments Pvt. Ltd. 2024. Environmental monitoring solutions for air quality. https://oizom.com/

Patra, S.S., Ramsisaria, R., Du, R., Wu, T., and Boor, B.E. 2021. A machine learning field calibration method for improving the performance of low-cost particle sensors. Building and Environment. 190: 107457. https://doi.org/10.1016/j.buildenv.2020.107457

Qiao, S., Guo, Q., He, Z., Feng, G., Wang, Z., and Li, X. 2025. Spatiotemporal trends and drivers of PM_2.5 concentrations in Shandong Province from 2014 to 2023 under socioeconomic transition. Toxics. 13(11): 978. https://doi.org/10.3390/toxics13110978

Radarit, K., Wiriya, W., Chai-Adisaksopha, C., Chantara, S., and Prapamontol, T. 2024. Significantly increased accumulations of PAHs in scalp hair during smoke-haze period among female adolescents in Chiang Mai, Thailand. Natural and Life Sciences Communications. 23: e2024013. https://doi.org/10.12982/NLSC.2024.013

Rahman, M.M., Shuo, W., Zhao, W., Xu, X., Zhang, W., and Arshad, A. 2022. Investigating the relationship between air pollutants and meteorological parameters using satellite data over Bangladesh. Remote Sensing. 14(12): 2757. https://doi.org/10.3390/rs14122757

Ramadani, U., Radojević, S., Lazović, I.M., Radivojević, D.S., Obradović, J., Živković, M. and Tasić, V. 2025. Application of low-cost air quality monitoring system in educational facilities in Belgrade, Serbia. Atmosphere. 16(9): 1103. https://doi.org/10.3390/atmos16091103

Ramly, Z.T., Ishak, M.Y., Abdullah, A.M., Ismail, M., Abdullah, S., and Mansor, A.A. 2025. Investigating the effectiveness of coal-fired power plant operations: Management, technical and air pollution aspects. Emerging Science Journal. 9(1): 86-113. https://doi.org/10.28991/ESJ-2025-09-01-06

Shabbir, M., Saeed, T., Saleem, A., Bhave, P., Bergin, M., and Khokhar, M.F. 2025. A paradigm shift: Low-cost sensors for effective air quality monitoring and management in developing countries. Environment International. 200: 109521. https://doi.org/10.1016/j.envint.2025.109521

Shahid, S., Brown, D.J., Wright, P., Khasawneh, A.M., Taylor, B., and Kaiwartya, O. 2025. Innovations in air quality monitoring: Sensors, IoT and future research. Sensors. 25(7): 2070. https://doi.org/10.3390/s25072070

Singh, D., Dahiya, M., Kumar, R., and Nanda, C. 2021. Sensors and systems for air quality assessment monitoring and management: A review. Journal of Environmental Management. 289: 112510. https://doi.org/10.1016/j.jenvman.2021.112510

Spyropoulos, G.C., Nastos, P.T., and Moustris, K.P. 2021. Performance of aether low-cost sensor device for air pollution measurements in urban environments. accuracy evaluation applying the air quality index (AQI). Atmosphere. 12(10): 1246. https://doi.org/10.3390/atmos12101246

Suriano, D. and Prato, M. 2023. An investigation on the possible application areas of low-cost PM sensors for air quality monitoring. Sensors. 23(8): 3976. https://doi.org/10.3390/s23083976

United States Environmental Protection Agency. (2024, December 20). Low-cost air pollution monitors and indoor air quality. https://www.epa.gov/indoor-air-quality-iaq/low-cost-air-pollution-monitors-and-indoor-air-quality

Usmani, R.S.A., Saeed, A., Abdullahi, A.M., Pillai, T.R., Jhanjhi, N.Z., and Hashem, I.A.T. 2020. Air pollution and its health impacts in Malaysia: A review. Air Quality, Atmosphere & Health. 13(4): 1093-1118. https://doi.org/10.1007/s11869-020-00867-x

Yin, S. 2020. Biomass burning spatiotemporal variations over South and Southeast Asia. Environment International. 145: 106153. https://doi.org/10.1016/j.envint.2020.106153

OPEN access freely available online

Natural and Life Sciences Communications

Chiang Mai University, Thailand. https://cmuj.cmu.ac.th