1 Introduction

Since the World Health Organization (WHO) formally announced the SARS-CoV-2 as a global pandemic on March 11, 2020, there have been 110,224,709 confirmed cases of COVID-19, including 2,441,901 deaths on February 20, 2021. Without any doubts, COVID-19 pandemic has dramatic impact on numerous dimensions such as economic shock, logistics, supply chain, tourism industry, education, mental health, agriculture, food security, and animal sectors (Basilaia and Kvavadze 2020; Gagliano et al. 2020; Grida et al. 2020; McKibbin and Fernando 2020; Proaño 2020; Sahu 2020; Seleiman et al. 2020; Williams 2021). Apart from the above-mentioned impacts, a work from home policy has been widely adopted as a part of government enforced quarantine measures to reduce the outbreak of pandemic (Purwanto et al. 2020; Setyawan and Lestari 2020). As a consequence of a transition to remote working amid COVID-19 pandemic, numerous studies have been focused on the improvement of air quality level particularly in urban area around the world (Lian et al. 2020; Mahato et al. 2020; Stratoulias and Nuthammachot 2020).

Although there have been many studies focus on the chemical characterization of air pollutants in the ambient air of Thailand, little is known about the impact of lockdown policy on the improvement of air quality (ChooChuay et al. 2020a, b, c; Pongpiachan and Iijima 2016; Pongpiachan et al. 2013, 2017). Several investigations have been conducted to elucidate the relationship between the “COVID-19 lockdown” and “air pollutant levels” in the ambient air of Thailand (Oo et al. 2021; Wetchayont 2021). A significant reduction of NO2 level up to 20.1% was observed in the Bangkok Metropolitan Area (BMA) during the government enforced quarantine in 2020 (Oo et al. 2021). The monthly average of NO2 in BMA observed in 2020 was also 9.8% lower than 2019 for the same time duration (Oo et al. 2021). Another in situ detections found a significant reducing trend of PM2.5, PM10, O3, and CO contents in the ambient air of Bangkok during the COVID-19 outbreak year in comparison with the same periods in 2019 (Wetchayont 2021). In spite of numerous studies focused on the influences of pandemic on reducing air pollutants, there is a rising concern that the psychological and social impact of COVID-19 might have played an important role in various aspects such as societies’ mental health, nursing services of elderly patients, anxiety/depression and suicidal ideation (Meng et al. 2020; Serafini et al. 2020; Sharma et al. 2020; Tee et al. 2020; Yom-Tov et al. 2021). On the contrary, there is a limited number of studies shed light on the intimate relationship between “COVID-19 infected number” and “PM2.5 level” in the ambient air of urban cities. Overall, the main principles of this study are to (i) elucidate the impacts of COVID-19 daily case reports on the variation of PM2.5 concentration and (ii) investigate the influences of lockdown policy and hot-spot numbers on the fluctuation of PM2.5 content in the ambient air of Bangkok in January 2021. It is crucial to underline that all hot-spot numbers were provided by FIRMS (The Fire Information for Resource Management System) which integrates remote sensing and GIS technologies to deliver MODIS (Moderate Resolution Imaging Spectroradiometer) hot-spot positions to GISTDA (The Geo-Informatics and Space Technology Development Agency (Public Organization)). MODIS is a major instrument aboard the Terra (Terra passes from north to south across the equator in the morning) and Aqua (Aqua passes from south to north over the equator in the afternoon) satellites. It is also important to highlight that the MODIS instrument is using both the Terra and Aqua spacecraft. It has a viewing swath width of 2330 km and views the entire surface of the Earth every one to two days. Its detectors measure 36 spectral bands between 0.405 and 14.385 µm, and it acquires data at three spatial resolutions—250 m, 500 m, and 1000 m. In this study, the definition of “hotspot” indicates any fire locations detected by MODIS in the administrative provinces of central Thailand (i.e. Bangkok, Ang Thong, Chai Nat, Kamphaeng Phet, Lopburi, Nakhon Nayok, Nakhon Pathom, Nakhon Sawan, Nonthaburi, Pathum Thani, Phetchabun, Phichit, Phitsanulok, Phra Nakhon Si Ayutthaya, Samut Prakan, Samut Sakhon, Samut Songkhram, Saraburi, Sing Buri, Sukhothai, Suphan Buri, and Uthai Thani).

2 Materials and Method

The Bangkok Metropolitan Administration (BMA) has PM10 and PM2.5 air quality monitoring networks over Bangkok area use two US-EPA FEM BAM models BAM-1020 Beta Attenuation Mass Monitor and BAM-1022 Portable Continuous Beta Attenuation Mass Monitor. “BAMs” are generally regarded as the standard for continuous ambient PM2.5 measurement in air quality monitoring. BAMs is using Beta ray attenuation, which is one of the most widely used methods worldwide for regulatory monitoring of automatically measures and records airborne particulate matter concentration levels in micrograms per cubic meter (μg m−3). They are generally simple, reliable, comparatively inexpensive, and relatively easy to operate. BAM-1022 received US-EPA Federal Equivalent Method (FEM) designated PM2.5 in 2013 and latest modifications in 2019. In addition, all PM2.5 data were collected by using BAMs located at 49 BMA districts. Some more descriptions on locations and characteristics of the site chosen are described as follows. In this study, there are ten BMA districts (i.e. Bang Bon, Bang Kapi, Bang Khae, Bang Khen, Bang Kho Laem, Bang Khun Thian, Bang Na, Bang Phiat, Bang Rak, and Bang Sue) connected with the Thai word “Bang” which means place and/or location usually near a river or canal. Although there are some unconscious biases related to “Bang” districts as suburban areas and/or outer districts, as they grew into a larger town, these “Bang” districts reaches a population of somewhere between 20,000 and 30,000 people then they will begin to be informally regarded as a city. For instance, Bang Kapi, Bang Rak, and Bang Sue have population of 148,964 persons, 46,472 persons, and 128,995 persons, respectively (Bangkok Metropolitan Administration 2014). It is interesting to note that Nong Chok, Lat Krabang, Bang Khun Thian, and Khlong Sam Wa have comparatively large areas with the values of 236,261 km2, 123,859 km2, 120,687 km2, and 110,686 km2, respectively (Bangkok Metropolitan Administration 2014). It is also crucial to underline that Khlong San, Ratchathewi, and Bangkok Yai have relatively high population densities with the values of 12,432 person km−2, 10,355 person km−2, and 11,327 person km−2, respectively (Bangkok Metropolitan Administration, 2014). All statistical analysis was conducted using IBM SPSS Statistics 23. The bivariate Pearson Correlation generates a sample correlation coefficient, R, which detects the strength and direction of linear relationships between hot-spot numbers and PM2.5 contents. Furthermore, the Pearson Correlation assesses whether there is statistical evidence for a linear relationship among the same pairs of hot-spot numbers and PM2.5 contents in the population. It is also crucial to note that the data must meet the following requirements which are (i) the values for all parameters (i.e. hot-spot numbers and PM2.5 contents) across cases are unrelated, (ii) the value for any parameter cannot affect the value of any parameter for other cases, (iii) the cases must have non-missing values on both parameters.

In this study, the independent samples t test was used to compare the averages of 2 independent groups, which are PM2.5 concentrations detected at 49 districts of BMA as categorized by COVID-19-infected number (n) (i.e. n < 300 vs. n ≧ 300, n < 500 vs. n ≧ 500, n < 700 vs. n ≧ 700) to determine whether there is statistical evidence that the related population averages are significantly different. The one-way analysis of variance (ANOVA) is generally employed to evaluate whether there are any statistically significant differences between the averages of PM2.5 contents measured in three different periods (i.e. 01/01/21–10/01/21, 11/01/21–20/01/21, and 21/01/21–31/01/21). It is also important to highlight that the one-way ANOVA is an omnibus test statistic and cannot explain which specific groups were statistically significantly different from each other; it only describes that at least two groups were different. To obtain a valid result, there are several assumptions that need to be passed prior to the statistical analysis. For instance, the dependent parameter (i.e. PM2.5 content) should theoretically be detected at the interval or ratio level. The independent parameter should have independence of observations as well as consist of two or more categorical, independent groups. In addition, there should be no significant outliers and the PM2.5 content should be approximately normally distributed for each monitoring site.

3 Results and Discussion

As illustrated in Fig. 1A–C, three main peaks of PM2.5 contents were detected in the first week (i.e. 04/01/21–06/01/21), the second week (i.e. 14/01/21–16/01/21), and the third week (i.e. 21/01/21–23/01/21) of January 2021. The associations of daily peak concentration of PM2.5 with daily COVID-19-infected number and hot-spot number (see Fig. 2) were carefully investigated using t test, Analysis of Variance (ANOVA), and Pearson Correlation Analysis (PCA). A t test was employed to correlate the arithmetic mean of PM2.5 and COVID-19-infected numbers (n) in three different categories which were Group-I (i.e. n < 300 vs. n ≧ 300), Group-II (i.e. n < 500 vs. n ≧ 500), and Group-III (i.e. n < 700 vs. n ≧ 700) (see Table 1). No significant differences (p > 0.05) were observed in Group-I indicating that the COVID-19-infected number of 300 can be considered of minor importance for making decisions to work from home (WFH). On the contrary, some significant differences (p < 0.05) were detected in air quality observatory sites located in Taling Chan, Phasi Charoen, Nong Khaem, Phar Nakhon, Bang Bon, Bang Khae, and Bang Khun Thian when using the COVID-19-infected numbers of 500 and 700 as illustrated in Table 1. Although these seven districts have comparatively high population with the average value of 130,842 ± 46,383 (i.e. Taling Chan (105,857), Phasi Charoen (129,238), Nong Khaem (153,175), Phar Nakhon (55,373), Bang Bon (107,140), Bang Khae (191,966), Bang Khun Thian (173,144)), the density population is highly deviated with the value in the range of 1435 Person km−2 (Bang Khun Thian) to 10,002 Person km−2 (Phar Nakhon) (Bangkok Metropolitan Administration 2014). It appears rationale to conclude that both the population and the density population cannot be used to explain some significant differences between arithmetic mean of PM2.5 and COVID-19-infected numbers observed in these seven districts. As clearly illustrated in Fig. 3, Taling Chan, Phasi Charoen, Nong Khaem, Phar Nakhon, Bang Bon, Bang Khae, and Bang Khun Thian are located in west and southwest of Bangkok. These seven districts are relatively close to Samut Sakhon province, the epidemic hot-spot of the second outbreak of COVID-19 in Thailand. On December 18th 2020, a 67-year-old merchant at a seafood market in Samut Sakhon was tested positive and a few days later up to 1300 persons were also detected with COVID-19. The second wave of outbreak were observed in up to 33 provinces with 28 provinces throughout the nation were declared “red zone” and subjected to lockdown measures (Tan and Lim 2021). The impact of second wave was more pronounced with the volume of cumulative nationwide cases amounted to 24,751-almost tripled the total cases reported in 2020 (Suphanchaimat et al. 2021). As a part of COVID-19 prevention measures, the Royal Thai Government has approved the extension of the Emergency Decree nationwide until 28th February, 2021 to contain the local transmission. The Emergency Decree dramatically decrease the number of passengers travelling to the infected areas adjacent to Bangkok due to some active preventive measures such as closure of areas at risk of infection, a ban on gatherings or illegal assembly in crowded areas, and strict screening of the movement of migrant workers. These active preventive measures can significantly reduce PM2.5 contents particularly during the first week of January, 2021. The fact that all PM2.5 concentrations detected in the above-mentioned BMA districts were significantly lower when the COVID-19-infected numbers were greater than 500 and 700 indicates that the COVID-19-infected situation reports play a major role in governing public decision to WFH.

Fig. 1
figure 1

A Daily variations of PM2.5 contents collected at Dusit, Pom Prap, Samphanthawong, Phaya Thai, Wang Thonglang, Pathum Wan, Bang Rak, Sathon, Yan Nawa, Wathana, Suan Luang, Bang Na, Chatuchak, Don Mueang, Sai Mai, Bang Kapi, Khan Na Yao, and Laat Krabang from 1 to 31st January, 2021. B Daily variations of PM2.5 contents collected at Min Buri, Nong Chok, Prawet, Khlong San, Bangkok Yai, Bangkok Noi, Taling Chan, Thawi Watthana, Phasi Charoen, Nong Khaem, Bang Bon, Thung Khru, Phar Nakhon, Huai Khwang, and Khlong Toei from 1 to 31st January, 2021. C Daily variations of PM2.5 contents collected at Bang Sue, Laat Phrao, Lak Si, Bang Khen, Saphan Sung, Bueng Kum, Khlong Sam Wa, Chom Thong, Bang Phlat, Bang Khae, Bang Khun Thian, Bang Kho Laem, Din Daeng, Phra Khanong, Rat Burana, and Ratchathewi from 1 to 31st January, 2021

Full size image
Fig. 2
figure 2

Covid-19-infected number and hot-spot number in Thailand from 1 to 31st January 2021

Full size image
Table 1 Statistical descriptions of PM2.5 concentrations detected at 49 districts of BMA as categorized by COVID-19-infected number
Full size table
Fig. 3
figure 3

A map of Bangkok as classified by the significant levels of ANOVA test by using PM2.5 concentrations collected at 49 districts at 3 different monitoring periods

Full size image

Further assessment connected with the impacts of WFH on air quality improvement were conducted by categorizing monitoring periods of PM2.5 into three groups namely Group-I (i.e. 01/01/21–10/01/21), Group-II (i.e. 11/01/21–20/01/21), and Group-III (i.e. 21/01/21–31/01/21). As clearly displayed in Table 2, the levels of PM2.5 were significantly lower in Group-I than other two groups as detected in Chatuchak, Khlong Sam Wa, Ratchathewi, Don Mueang, Sai Mai, Laat Phrao, Lak Si, Bang Khen, Nong Chok, Wang Thonglang, Bang Sue, Din Daeng, Bang Kapi, Saphan Sung, Pathum Wan, Huai Khwang, Phaya Thai, Khlong Toei, Bueng Kum, Bangkok Noi, Khan Na Yao, Wathana, Bang Na, Prawet, Laat Krabang, Dusit, Min Buri, Suan Luang, Phra Khanong, Bang Rak, Bang Phlat, Pom Prap, and Phar Nakhon. These significant differences detected at the 33 districts can be ascribed to several reasons, including geographical locations and pollution sources. Previous studies highlight traffic emissions as one of the major sources of PM2.5 in ambient air of Bangkok (ChooChuay et al. 2020a; Pongpiachan and Iijima 2016; Pongpiachan et al. 2015, 2017). It is well known that these 33 districts have comparatively higher traffic volume in comparison with those of other 16 districts (i.e. Yan Nawa, Samphanthawong, Sathon, Thung Khru, Taling Chan, Khlong San, Bangkok Yai, Phasi Charoen, Rat Burana, Bang Kho Laem, Chom Thong, Thawi Watthana, Bang Bon, Nong Khaem, Bang Khae, and Bang Khun Thian). For instance, the number of vehicles at intersections of Kaset (Bang Khen), Democracy Monument (Ratchathewi), Prasert Manukitch Lat Pla Kha (Laat Phrao), Klong Prapa (Bang Sue), and Thiam Ruam Mit (Huai Khwang) have number of vehicles per day with the values of 156,442, 132,011, 109,572, 108,063, and 105,804, respectively (Bangkok Metropolitan Administration 2019). The average traffic volume of these five districts (i.e. 122,378 ± 21,776) was almost 4.9, 6.0, and 7.3 times higher than those of Sathon (i.e. Witthayu; 25,096), Thung Kru (i.e. Phacha U Tit & Kru Nai; 20,528), and Yan Nawa (i.e. Naret; 16,848), respectively (Bangkok Metropolitan Administration 2019).

Table 2 ANOVA test of PM2.5 concentrations detected at 49 districts of BMA as categorized by 3 different monitoring periods
Full size table

It is also interesting to note that some significant reductions of PM2.5 were observed in the eastern parts of Bangkok as illustrated in Fig. 3. This can be explained by the fact that the majority of business districts and commercial zone are more likely to be concentrated in the eastern part of Bangkok than the western region. Thus, it appears rationale to assume that the implemented quarantine measures tend to play a major role in reducing PM2.5 levels particularly in the business areas of eastern Bangkok. In this study, the hot-spot numbers in Thailand from 1 to 31st January 2021 were also carefully analyzed using Pearson Correlation Analysis (PCA) to correlate with PM2.5 contents collected at 49 districts of BMA in January 2021. As shown in Fig. 4, the correlation coefficients of hot spots were comparatively low with the values of R < 0.4. On the contrary, the relatively high correlation coefficients of R > 0.9 were observed with PM2.5 collected within different BMA air quality observatory sites. These findings underline the importance of traffic emissions as one of the main contributors of PM2.5 in BMA districts during the observatory period.

Fig. 4
figure 4

Pearson correlation coefficients between hot-spot numbers and PM2.5 contents collected at 49 districts of BMA in January 2021

Full size image

4 Conclusions

Numerous empirical investigations underline the importance of lockdown policy as one of key factors which dramatically reduce air pollutants in ambient air around the world particularly in the middle of pandemic. Earlier, there was no clarity whether the COVID-19-infected number can psychologically affect decision making on self-quarantine during the coronavirus crisis. PM2.5 contents observed in some BMA districts were significantly lower when the COVID-19-infected numbers were larger than 500 and 700 underlines the importance of WFH for improving air quality. The second outbreak can significantly reduce PM2.5 levels detected in eastern part of Bangkok particularly during the first week of January, 2021. In addition, the relatively low Pearson’s correlation coefficients between hot-spot numbers and PM2.5 concentrations indicates that agricultural waste burnings can be considered as the second contributor after vehicle exhausts.