Performance evaluation of multiple particulate matter monitoring instruments under higher temperatures and relative humidity in Southeast Asia and design of an affordable monitoring instrument (ManPMS)

Abstract The rapid industrialization of nations in Southeast Asia (SEA) has led to a decline in these countries’ air quality, including high levels of particulate matter (PM). Monitoring these air pollutants is crucial to understanding the pollution status of the area and developing management plans for improvement. The metrological conditions in the region present challenges as high temperature and high humidity have been known to cause errors in the measurements. This study investigated the performance of five PM monitoring instruments with different working principles. The air temperature was mostly over 25 °C with relative humidity usually remaining above 80%, which is typical of SEA weather. Measurements from all instruments had good correlations with each other as their linear regressions yielded slopes of 1 ± 0.15 and R2 > 0.65. Moreover, this study found that depending on the chosen reference instrument, not all factors affect the devices equally. In particular, using Partisol as a reference, the PM2.5 concentration, air temperature, and relative humidity had less impact upon the relative bias level compared to using Leckel as a reference. In addition, the high cost of monitoring instruments also poses financial constraints on how many monitoring stations can be deployed. To tackle this issue, this study presents ManPMS whose design is based on that of the USEPA Title 40 Part 50 with slight modifications. The cost to manufacture and assemble the instrument was only 2/3 the price of a typical instrument with similar performance.


Introduction
Southeast Asia (SEA) consists of 10 nations that have been rapidly developing during the 21 st century. However, the air quality in these nations has been negatively affected in both urban and rural settings. To meet the higher energy demand from industrialization, many countries have chosen coal-fired power plants as a cost-effective solution. [1] However, these power plants are heavy polluters, [2] including particulate matter (PM), the interest of this study. PM is classified according to size including (a) TSP (total suspended particulate matter less than 100 mm in size), (b) PM 10 (particulate matter less than 10 mm in size), and (c) PM 2.5 (particulate matter less than 2.5 lm in size). [3] Technological advances and environmental awareness are slowly shifting the trend to oil-fired power plants, [4] which release significantly less PM. [5] Despite a positive change taking place in the region, due to cost constraints and political reasons, new coal-fired power plants are still being planned and commissioned. [6][7][8] Besides air pollution from industrial sources, those from traffic sources also contribute to the decline in the air quality of SEA because nations in the region do not have well-defined guidelines to reduce traffic congestion. [9] Currently, only Malaysia and Singapore have up-todate air quality standards for PM 10 and PM 2.5 , [10,11] which follow the World Health Organization (WHO) guidelines of 2021. [12] Vietnam also had its own air quality standard QCVN 05:2013/BTNMT, [13] but it has not been updated since 2013. For the daily average of PM 2.5 concentration, the standard is set at 25, 35, and 50 lg/m 3 for Singapore, Malaysia, and Vietnam, respectively. In comparison, the WHO guidelines recommend the concentration to be set at 15 lg/m 3 . In Hanoi (Vietnam), the PM 10 and PM 2.5 concentrations can be high as 117.1 lg/m 3 and 65.2 lg/m 3 , respectively, during hour-long traffic jams. [14] Thus, in the short term, SEA faces many challenges in improving its air quality with the increasing urbanization and industrialization.
The air quality in rural areas in the region has been negatively impacted as well. The open burning of agricultural waste and deforestation has been increasing to keep up with the demand for food and export. [15] The smoke from the activity has been recorded to travel across boundaries and oceans to affect adjacent countries. Haze is an atmospheric phenomenon in which visibility is degraded by aerosols. Air quality monitoring stations in Hat Yai (Thailand) detected open burning smoke originating from Indonesia during haze events. [16] Because of the constant burning, combined with the metrological conditions in the regions, the intensity and frequency of haze events would only worsen in the near future. [17] In particular, in Hanoi (Vietnam), the combination of high humidity and lower temperature caused haze events with the PM 2.5 concentrations to reach 130 lg/m 3 in 2019 [18] and PM 10 concentrations to reach 200 lg/m 3 in 2015. [19] As a result, the danger of open burning affects a large region of SEA as the smoke can travel over 500 km. Previous work has demonstrated that the increase in mortality is connected with the increase in PM 10 and PM 2.5 exposure as these particles can reach the alveoli of the lung to cause respiratory problems as well as cardiovascular and neurological concerns. [20,21] Most typical aerosol monitoring instruments are bulky, heavy, and expensive (approximately $5,000-$30,000 per instrument, and up to $100,000 per station. [22] These instruments are usually imported from abroad, which adds complexity to installation and maintenance. In addition, the location of the monitoring station often has strict requirements and must be accessible for calibration and maintenance processes. Such conditions present financial obstacles for developing nations in SEA. In recent years, some rural towns in Vietnam have increased investment in building more monitoring stations to meet national requirements such as TCVN 5067:1995 [23] and TCVN 9469:2012. [24] However, the total number of stations remained sparse and air quality assessment can only be used for the general forecast as warnings for extreme weather events such as haze are currently limited. In particular, only two, out of the nine monitoring stations that had been planned, were deployed in Ho Chi Minh in 2022. [25] In addition to the financial difficulties during the construction and operation of monitoring stations, the weather and climate of the SEA also affect the accuracy of measurements. Overall, the climate in SEA can be characterized by high average temperature (> 25 C), high average humidity (> 75%), and high rainfall (> 100 mm) annually. The high temperature can cause a loss in semi-volatile material due to evaporation and decomposition, [26] which would distort the final chemical composition of the sample. In addition, when the relative humidity is larger than 70%, PM can absorb water from the air. [27] Heavy rain can generate a large quantity of aerosol water which can artificially increase the sample weight. [28] Thus, the climate of the region presents challenges for the accuracy of the monitoring data as conditions in SEA can affect the composition and the weight of the sample.
Understanding the hardships that SEA faces and the importance of highquality air quality monitoring data, this study aims to evaluate the performance of different types of PM monitoring instruments under typical SEA weather conditions by investigating the effects of concentration, temperature, and relative humidity against reference instruments. PM 2.5 would be the focused parameter in this study because of its effect on human health. In addition, this study also presents ManPMS, a reference gravimetric method based on the United States Environmental Protection Agency (EPA) design as a proposal for a cost-effective monitoring instrument. Most of the component of this instrument was locally sourced. The PM 10 and PM 2.5 separators were manufactured by the Vietnam Academy of Science and Technology (VAST) as they were not available in Vietnam during the study period.
The Leckel was designed following the European EN 12341:2014 standards [29] according to the manufacturer. The flow rate of the instrument was set at 2.3 m 3 /h and used PTFE coated glass fiber 47 mm x 0.45 lm filter (Measurement Technology Laboratories, USA). On the other hand, the Partisol was designed following Title 40 -Part 50 of the United States Code of Federal Regulation (CFR) standards. [30] Its flow rate was set at 1.0 m 3 /h and used the same filter as the Leckel. Both instruments are certified to be used as ambient air quality monitoring systems by their respective legislation.
The Airborne is a portable automated PM monitoring instrument that detects and measures PM concentrations using laser diodes. It features channel diameters of 0.3, 0.5, 1.0, 3.0, 5.0, and 10.0 lm, as well as sampling times ranging from 1 s to 24 h. All channel diameters were used in this study, the sampling frequency was set to 1 h, and the flow rate was set to 1.7 m 3 /h. The TEOM is a stationary automated PM monitoring instrument. It measures PM concentration by detecting the oscillating frequency of the filter when PM is deposited. The operating temperature was set at 50 C, and the flow rate was 1.0 m 3 /h.

Sample collection
All monitoring activities for all instruments were conducted at the rooftop of the Center for Research and Technology Center (CRETECH) building (18 Hoang Quoc Viet Street, Cau Giay District, Hanoi, Vietnam), which is a 6 story-building, from 11 February 2022 to 26 April 2022. The building is 200 m away from the main traffic road of the district and 160 m from the smaller one ( Figure 1). All samples collected were processed at the CRETECH laboratory (ISO 17025 [31] certified). The gravimetric analysis of PM 2.5 filters was carried out in a temperature and humidity-controlled laboratory complying with EN 12341:2014 standards. [29] The temperature was kept between 20-23 C and humidity at 30-35%. The filters were weighed before and after sampling after being conditioned for at least 24 h. The electrostatic charge of the filters was removed with an ionized air blower (CSD-0911, MEI-SEI, Japan) before the filters were weighed by a microbalance (CP2P-F, Sartorius, UK) with its readability rating at 0.001 mg. Each filter was weighed five times with a reading deviation of less than 2.0 lg. The experiment was conducted for 10 weeks during the late spring and early summer months in Hanoi, which represented high temperature and high relative humidity conditions, typical of SEA weather ( Figure 2). These measurements were taken from the auxiliary system on ManPMS. The daily average temperature ranged from 15-34 C and the daily average relative humidity was measured from 52-97%.

Data analysis
Because Leckel and Partisol are certified reference gravimetric instruments by their corresponding governmental bodies, they were used as references for this study to compare the performance of other instruments. An intercomparison between the reference instruments and the others was performed using linear regression. The slopes of the regression line and the R 2 values were used to determine the relationship between instruments. The root mean square error (RMSE) between the pair of instruments was calculated to evaluate the average error. The RMSE values were calculated based on Equation (1). Next, this study examined the effects of PM 2.5 concentrations, temperature, and relative humidity on the relative bias levels. These parameters were divided into smaller groups and one-way ANOVA was conducted to determine whether there was a significant difference between these groups. The one-way ANOVA tests were conducted to compare the effect of concentration on the relative bias, which is calculated based upon Equation (2). A positive relative bias indicates that the sampling instrument was recording a higher concentration while a negative relative bias shows that the reference instrument was recording this concentration.

RMSE
where reference is the measurement from the reference instruments, sample is the measurement from the other instruments, and

Design of ManPMS and validation of ManPMS
The design of ManPMS is based on that of Appendix J and Appendix L of Title 40 -Part 50 of the CFR [30] with a few modifications. The instrument consists of (a) PM 10 and PM 2.5 separators, (b) the sample filter, (c) the pump, and (d) the auxiliary system. The illustration instrument is depicted in Figure  3 and the cost analysis of parts and testing is presented in Table 2. The detailed description of each system is described in detail below.

PM 10 and PM 2.5 separators
The PM 10 and PM 2.5 separators were manufactured and assembled by VAST in this study and their dimensions were based on Appendix L of Title 40 -Part 50 of the CFR. [30] Before reaching the sampling filter, the incoming PM needs to be separated from dust, water, and other contaminants. In the EPA original design, the PM 10 and PM 2.5 separators are connected via a 1 m tube. In ManPMS, the separators are directly connected to reduce the overall dimension. The air passes through the PM 10 separator to remove particles larger than 10 lm and passes through the PM 2.5 separator to remove particles larger than 2.5 lm. The remaining airflow should only contain PM 2.5 before being collected by the sampling filter. Both of the separators used in this instrument are inertial impactors, which categorizes particles based upon their aerodynamic size. The theory of how they operate was well studied. [32] The incoming PM and unwanted particles strike a target plate. Depending on the speed and diameter of the target plate, as well as the particle size and density, only particles with certain aerodynamic diameters continue with sufficient momentum to be carried by the airflow to the next stage. This study focused on the properties of the PM 2.5 separator as PM 2.5 is the main parameter. The characteristics of a separator are determined by its efficiency curve and the sample geometric standard deviation (GSD) for accuracy and precision, respectively. The A1 Test Dust (Powder Technology, USA) was used to examine the separator properties as the A1 Test Dust provides a variety of particle sizes from 0.1 to 15 lm. The aerosol was dispersed using RBG 1000 (Palas, Germany) at 10 mg/h for different amount of time to achieve 0.5, 4, and 6 mg loading. The target plate was immersed under 1 mL of DOWSIL 702 Pump Fluid (Dow, USA). The efficiency curve in Equation (3) and the GSD in Equation (4) derived in this study were simplified from previous studies. [33,34] where E is the efficiency (%), a is the curve coefficient, D 50 is the cut point (lm) where E ¼ 0:5, and d is the aerodynamic diameter (lm) of the particle. The GSD was calculated by: where D 84 and D 16 are the cut points (lm) where E ¼ 0:84 and E ¼ 0:16, respectively. The efficiency curve indicates how well the separator collects particles of different sizes. Ideally, D 50 should equal the diameter of the target PM, which is 2.5 lm for the PM 2.5 separator. The curve parameter a ¼ 3:3683 was calculated by the EPA. The requirements and parameters of the EPA curve can be found at Title 40 -Part 50 and Part 53 of the CFR. [30,35] To calculate the parameters of the curve, the PM 2.5 separator was tested under different sample loadings at 0.5, 4, and 6 mg. The efficiency E was measured at various diameters d from 1.4 to 3.2 lm. The best-fit curves at different loadings were compared to the EPA curve ( Figure 4 and Table 3).
Overall, the best-fit curves from all of the loading closely matched the EPA curve. The curve coefficient a under all loading was underestimated compared to the EPA curve. However, the difference was not significant as the value from the EPA was still within the standard error. According to Title 40 -Part 50 of the CFR, [30] D 50 ¼ 2:560:2lm, under 0.5 and 4 mg loadings, the PM 2.5 separator performed within the required standards. At 6 mg, however, the D 50 value was slightly outside the acceptable range; thus, the maximum loading allowed in this study was 6 mg. The higher loading changes the property of the surface of the impact plate by having incoming particles impact the deposited ones. Previous studies have shown that higher loading decreases the D 50 : [36,37] The GSD represents how sharply the separator removes larger particles. Smaller values of GSD are preferred because it indicates that the separator is more precise. Based on previous studies, GSD values less than 1.20 are considered small, and values higher than 1.30 are considered high. [33,36,37] The PM 2.5 separator under all loading is less sharp than that of the EPA testing. Despite so, the GSD values were within the normal ranges as other studies had recorded values from 1.2 to 1.4. [34,38] Overall, the characteristics of the PM 2.5 separator manufactured by VAST met the requirements of the EPA and were within the normal range of other studies. Therefore, the separator was deemed suitable for installation.

Sample filter, pump, and auxiliary system
The sample filters were PTFE-coated glass fiber 47 mm in diameter and 0.45 lm in pore size (Measurement Technology Laboratories, USA). The Rotary Vane Pump (VTE 6, Thomas, Germany) was used in assembling the instrument. The pump performance was evaluated at 0, 75, and 120 lg/m 3 concentrations. Overall, the average flow rate was 0.99 ± 0.01 (0.96%) m 3 /h, which met the EPA standards ( Figure S1).
The auxiliary system monitors the ambient and filter conditions. To monitor the ambient conditions, a temperature and relative humidity sensor 597 A (Met One Instruments, US) was installed. Moreover, for monitoring the filter conditions, a temperature sensor and a cooling fan were used. In the original EPA design, the filter temperature sensor is there only to verify that there are no large temperature differences between the filter and the ambient condition. However, in this instrument, the filter temperature is kept at 25 ± 0.5 C by the cooling fan to protect the sample from high temperatures. The cool air was taken from the CRETECH building air conditioning system. There is a cutout on the back of the case, below the filter cassette, where the cooling fan was inserted. Hence, the fan cools down the filter indirectly by cooling the entire case. The fan turns on when the filter reaches 25.5 C and turns off when the temperature is reduced to 24.5 C. Because the filter is indirectly cooled, it usually took 3-5 min for the temperature to drop and 20-25 min to remain within the range. It was noted that on 13, 14, 26, and 27 April, the fan was on for most of the day because of how sunny these days were.

Field results
All measurements from the instruments are presented in Figure 5 and are summarized in Table 4. Overall, all instruments were able to display the natural increase and decrease in concentration at the monitoring site as peaks and troughs were well captured. Measurements from different instruments showed strong positive correlations (R > 0.8) to each other (Table  5). However, viewing in detail, there was a slight discrepancy between the gravimetric reference and automated instruments. The gravimetric reference method instruments recorded the average PM 2.5 mass concentration at 21.52 ± 6.13 lg/m 3 while the automated devices only measured the average concentration at 15.78 ± 4.79 lg/m 3 . The average discrepancy was 26.24%, which is within the 3-50% range previously recorded. [39] In general, the performance of ManPMS is closest to that of the reference instruments chosen for this study, especially with Partisol as they are the most similar in design. Next, for TEOM, its 50 C operating temperature is its most significant source of error. The high temperature vaporizes semi-volatile chemicals in the PM, thus, TEOM underestimates the concentration. Lastly, for Airborne, since it was calibrated based on a specific type of aerosol, any difference in the PM physical dimension contributes to the error of the final measurement. These are the general factor that affects the relative bias of these instruments when compared to the reference devices. The difference in performance and the effects of metrological conditions is discussed below to see if these conditions add to the well-known erroneous issues. The difference in performance and the effects of metrological conditions are further discussed in the following sections (Table 5).

Performance and effect of concentration
As ManPMS has the same working principle as the reference instruments, its performance closely matched theirs (Figure 6(a-d)). The linear regressions between these instruments had their slopes within 1 ± 0.05 and R 2 > 0.8. A slightly higher R 2 value was found when comparing ManPMS and Partisol (R 2 ¼ 0.965) than that of Leckel (R 2 ¼ 0.847) because ManPMS and Partisol had the same flow rate at 1 m 3 /h while Leckel flow rate was at 2.3 m 3 /h. On the other hand, TEOM and Airborne had a lower performance compared to ManPMS. In particular, the linear regressions between TOEM and Airborne against Partisol had their slopes within 1 ± 0.1 and 0.8 > R 2 > 0.7 (Figures 6(e-f)), while the linear regressions against Leckel had their slopes within 1 ± 0.15 and 0.75 > R 2 > 0.65 (Figure 6(b,c)). Overall, measurements from ManPMS, TEOM, and Airborne better fitted the result from Partisol than from Leckel.
When comparing ManPMS against the reference instruments, Leckel showed that the effect of PM concentration was significant on the relative bias while Partisol did not show similar results (Figure 7(a-d)). This   suggests that the difference in flow rate and temperature control affected the sample degradation rate. At lower concentrations, positive relative bias was found, indicating that ManPMS recorded higher concentrations. These results indicate that the temperature control from ManPMS degraded the sample slower than for Leckel. When the concentrations increased, the relative bias reduced and became negative, meaning Leckel recorded higher concentrations. The sample loading from Leckel is larger than that of ManPMS at the same PM concentration due to the higher flow rate of Leckel. At higher concentrations, the higher filter loading was sufficiently large that the interactions between the PM with themselves and the filter counteract the degradation effect of the higher flow rate. The forces that higher filter loading induced were also shown in the literature to influence the quality control of gravimetric analysis [40] and filtration efficiency. [41] Next, the influence of PM concentration was significant when TEOM was compared against both reference instruments (Figure 7(b-e)). The negative relative bias indicated that TEOM recorded lower concentrations compared to the reference instruments. This was expected as the higher internal temperature of TEOM evaporated semi-volatile chemicals in the PM. However, there was no specific trend observed. There was less relative bias only at the 15-20 and 30-35 lg/m 3 ranges. We noted that on days when these concentrations were recorded, relative humidity stayed high both day and night. During the periods from 16-19 February and 4-6 March, the relative humidity remained over 95% all day, and for the periods from 16-26 March and 15-23 April, the relative humidity remained over 90% all day. So, the reduction in relative bias may be due to the metrological condition stabilizing TEOM operation. Lastly, the influence of PM concentration was not significant when Airborne was compared against both reference instruments (Figure 7(c-f)). The error from Airborne was attributed to coincidence error, which is common among optical particle counters.

Effect of ambient temperature
Ambient temperature is an important aspect that affects the accuracy of an instrument. Its impact on the instruments was analyzed by dividing the temperature range into 4 equal intervals and calculating the relative bias of each interval. When comparing the ManPMS against the reference instruments, Leckel showed that the effect of temperature was significant on the relative bias while Partisol did not show similar results (Figure 8(a-d)). Because ManPMS internal temperature was kept at 25 C, cooler PM entering slightly evaporated its semi-volatile chemicals. This would explain the negative relative bias at the 15-20 C range. Seemingly, warmer PM entering ManPMS hygroscopically grew from the condensed water, leading to the positive relative bias at the 30-35 C range. The ranges in the middle were observed with less relative bias. However, the influence of semi-volatile evaporation and hygroscopic growth was only seen in Leckel but not Partisol for comparison. As ManPMS and Leckel are similar in design and specification, the discrepancy in the concentration recorded by these instruments was too small for the relative bias to be significantly different at the variety of temperature ranges.
Next, comparing against the reference instruments, TEOM showed that the effect of temperature was significantly different for Partisol (Figure 8(b-e)). As the ambient temperature increased closer to the TEOM operating temperature, less relative bias was found. As the common issues with TEOM are usually concerns about semi-volatile chemicals evaporating, the higher ambient temperature would have already evaporated these chemicals. For Airborne, the effect of temperature on relative bias was observed when compared against Leckel. Because Airborne is calibrated with a certain shape and size of the particle, the physical dimensions of the PM may be different at higher temperatures which would lead to an increase in relative bias.

Influence of relative humidity
Besides ambient temperature, relative humidity also affects the accuracy of an instrument. Its impact on the instruments was analyzed by dividing the humidity range into 4 equal intervals and calculating the relative bias of each interval. Only TEOM showed that relative humidity had a significant effect on the relative bias when compared against both reference instruments. In particular, as the humidity increased, the relative bias level decreased, rather than increased as shown in previously. [42,43] Even though higher humidity would cause the PM to hygroscopically grow, the relative humidity remained constantly high for the duration of the experiment and usually fluctuated 10% between noon and midnight. Thus, it may be the continuous high level of humidity that contribute to the stability of TEOM. The same trend was also observed in Airborne despite relative humidity not having a significant effect on the relative bias level. Moreover, Airborne should, theoretically, be more prone to hygroscopic growth. Recent studies have shown that at high relative humidity, its effect on hygroscopic growth was either a plateau or insignificant ( Figure 9). [44,45]

Discussion and conclusion
This study has also shown that ManPMS performance was comparable to the two reference instruments while being 2/3 of the cost. The authors of this study were able to assemble the instrument with only local suppliers, greatly reducing assembly and maintenance efforts. The PM 2.5 manufactured by VAST was tested to determine that its efficiency curve was suitable for operation. The cooling fan was able to maintain the filter temperature at 25 ± 0.5 C to protect the sample from high temperatures. With the increasing need for air monitoring and the budget that SEA nations, it is a cost-effective way of providing a sufficient number of monitoring instruments and lower operational costs.
Moreover, this study also presented the intercomparison between the performance of two reference instruments (Leckel and Partisol) against three other instruments (ManPMS, TEOM, and Airborne) with different working principles. Under the high temperature and high humidity weather of Hanoi, typical to that of SEA, all instruments were able to display the fluctuation in concentration. However, the automatic PM monitoring instruments tended to underestimate the values compared to the gravimetric reference devices. Meanwhile, measurements from all instruments still showed strong correlations to each other as their coefficients of correlation were larger than 0.8. The linear regressions obtained good linear correlations, having slopes within 1 ± 0.15 and good to medium consistency with R 2 > 0.65.
In addition to evaluating the measurements from the instruments, this study also investigated the effect of varying PM 2.5 concentrations, temperature, and relative humidity upon the relative bias level. There were mixed results as no single factor was concluded to affect all instruments equally. Depending on the reference being used, the level of relative bias remained the same throughout the ranges. The summary of the p-values from all of the ANOVA tests is presented in Table 6 when examining the impact of the ambient conditions on the relative bias.
For concentration, its influence on ManPMS and TEOM was significant at the 1% level while on Airborne, the effect was only significant at the 10% level. Because ManPMS and TEOM require a filter in their design, the loading on the filter influences the degree of relative bias. The temperature control of ManPMS and higher sample loading of Leckel were determined to be the main factors in explaining the relative bias level.
For temperature, its effect on all instruments was significant at both the 5% and 1% levels. As volatile nitrate was recorded to be more than 20% of the total PM weight in Hanoi, [46] the constant high temperature caused volatile substances to decompose or evaporate and impact the relative bias level. In addition, the condensation from the cooler temperature of ManPMS was observed and explained the discrepancy.
Lastly, for relative humidity, only TEOM was significantly affected at the 5% level. Because TEOM is the only instrument that heated the sample, the process may have removed the water in the PM. The continuous high relative humidity during the sampling period may have stabilized the physical and chemical properties of the PM. [44,45] This would explain why there were no significant variations in the relative bias level when the relative humidity changes. Thus, mixed results depended on the reference instrument being used to calculate the relative bias, and not every instrument was similarly affected by meteorological conditions. For future monitoring activities in the region, these three factors in this study should be taken into account.
Comparing against the Vietnam Air Quality Standard QCVN 05:2013/BTNMT, [13] the air quality during the sampling period did not exceed the requirement of 50 lg/m 3 . However, the sampling location was over 150 m away from the main roads and was on a 6-story-high building. On the other hand, comparing the results against the WHO standards [12] of 15 lg/m 3 , there is a discrepancy in the air quality depending on which type of instruments was selected. Recorded on the reference gravimetric instruments (ManPMS, Leckel, and Partisol), there were 61 days when the air quality exceeded 15 lg/m 3 . Looking at the automatic instrument, however, there were 41 days when the air quality exceeded that same concentration. The reference instruments were able to accurately depict the air quality of the area which was underperforming for 80% of the sampling period. Data from the automatic instruments only showed that the air quality was exceeding WHO standards for 55% of the sampling time. Thus, the importance of high-quality measurement needed to be emphasized as it plays an important role in understanding the current air quality of the area.