Association between ambient air particulate matter and human health impacts in northern Thailand

Air pollution in Thailand is regarded as a serious health threat, especially in the northern region. High levels of particulate matter (PM2.5 and PM10) are strongly linked to severe health consequences and mortality. This study analyzed the relationship between exposure to ambient concentrations of PM2.5 and PM10 by using data from the Pollution Control Department of Thailand and the burden of disease due to an increase in the ambient particulate matter concentrations in northern Thailand. This study was conducted using the Life Cycle Assessment methodology considering the human health damage impact category in the ReCiPe 2016 method. The results revealed that the annual average years of life lived with disability from ambient PM2.5 in northern Thailand is about 41,372 years, while from PM10 it is about 59,064 years per 100,000 population. The number of deaths from lung cancer and cardiopulmonary diseases caused by PM2.5 were approximately 0.04% and 0.06% of the population of northern Thailand, respectively. Deaths due to lung cancer and cardiopulmonary diseases caused by PM10, on the other hand, were approximately 0.06% and 0.08%, respectively. The findings expressed the actual severity of the impact of air pollution on human health. It can provide valuable insights for organizations in setting strategies to address air pollution. Organizations can build well-informed strategies and turn them into legal plans by exploiting the study’s findings. This ensures that their efforts to tackle air pollution are successful, in accordance with regulations, and contribute to a healthier, more sustainable future guidelines on appropriate practices of air pollution act/policy linkage with climate change mitigation.

charge fees in China, which resulted in three benefits: reduced congestion and emissions, an improved health impact index, and a decrease in the number of illegal travels.In Bangkok, Thailand, Chavanaves 17 calculated the health and economic gains achieved compared to a business-as-usual scenario.The associated health burden was evaluated using impact characterization factors (CFs) developed for different spatial situations in Thailand.
Ruchirawat et al. 18 examined the potential health impacts of carcinogenic air pollution exposure in urban areas compared to rural areas, using a toxicological model of exposure to early biological effects.They assessed several biomarkers to evaluate the potential health risks associated with this exposure.Van Zelm et al. 19 updated CFs for human health damage caused by PM 10 and ozone in Europe in 2000, considering slight increases in NH3, NO x , SO 2 , PM 10 , and NMVOC emissions.Cohen et al. 20 utilized accelerator-based ion beam analysis (IBA) techniques to quantify and characterize PM 2.5 pollution for a range of elements from hydrogen to lead in Hanoi, Vietnam.Apte et al. 21examined global trends of intra-urban intake factors (IFs) for dispersed ground-level primary pollutant emissions across countries, regions, and cities of different sizes.www.nature.com/scientificreports/Kassomenos et al. 22 investigated the relationship between ozone (O 3 ) and PM 10 exposure and public health, revealing a quantification of the disease burden from PM 10 and O 3 -related mortality and morbidity using a Life Cycle Impact Assessment focused on Greece, specifically Athens.Gronlund et al. 23 analyzed and evaluated CFs (DALY/kgPM 2.5emitted ) in US metropolitan regions, along with the results of dose-response factors, severity factors, and intake fractions.The studies discovered that the average annual health burden in the US due to PM emissions and CFs was 2.2 times higher.
Tang et al. 24 assessed human health damage factors (DFs) using the Special Report on Emission Scenarios (SRESs) of the Intergovernmental Panel on Climate Change (IPCC).DFs included malaria, diarrhea, cardiovascular disease, starvation, coastal floods, and inland flooding.Van Zelm et al. 25 determined regionalized CFs for human health damage from PM 2.5 and ozone, as well as vegetation damage from ozone, using a chemical transport model.Khaniabadi et al. 26 studied the effects of PM 10 , NO 2 , and O 3 on health in Kermanshah City, Iran, finding that PM 10 accounted for 62% of premature deaths, while NO 2 and O 3 accounted for 11% and 27% of deaths, respectively.They also found that a 10 μg/m 3 increase in PM 10 , NO 2 , and O 3 resulted in a relative risk (RR) of 1.066, 1.012, and 1.020, respectively.
Tang et al. 27 estimated ozone DFs by region and investigated the impacts of long-distance migration on the DFs using a global chemical transport model (CTM).Additionally, Tang et al. 28 employed a chemical transport model to evaluate the DFs of human health damage caused by PM 2.5 in 10 different locations worldwide.
Considering the scarcity of health-related data, determining the specific health effects of PM 2.5 for each area in Thailand is challenging.Additionally, some data such as RR, mortality rate (MR), years of life lost (YLL), and breathing rate (BR) for each province are only available as average data at the national level and not specific to local areas.Many studies rely on WHO reports or use regional and global values.It is crucial to have more accurate and reliable health data for Thailand to understand how severe air pollution affects human health in different contexts and air pollution sources in each province of Northern Thailand.Therefore, this study employs the Life Cycle Assessment (LCA) approach, following International Organization for Standardization (ISO) 29 standards (ISO 14040 30 and ISO 14,044), to investigate the health impacts of PM 2.5 and PM 10 .The study aims to analyze the relationship between exposure to ambient concentrations of PM 2.5 and PM 10 using data from the Pollution Control Department (PCD), providing a quantification of the burden of disease from PM 2.5 and PM 10 -related mortality and morbidity due to exposure in terms of Disability Adjusted Life Years or DALYs 31 .
The study is limited to assessing the impact of outdoor air pollution.This is because the concentration data used in the study only measure outdoor air pollution levels.It is worth noting that a significant portion of the   www.nature.com/scientificreports/population spends the majority of their time in non-air conditioned environments, such as homes, workplaces, and public spaces.Therefore, indoor pollution sources, such as cooking fumes and other pollutants, were not taken into consideration in this study.
Indoor pollution tends to disperse rapidly and does not accumulate to the same extent as outdoor pollution.As a result, the study focused specifically on the effects of outdoor air pollution on human health.The limitations related to indoor pollution sources and their potential health impacts were not addressed in this particular study.

Methodology
Goal and scope of this study.The goal of this study is to assess the human health damage caused by PM 2.5 and PM 10 in northern Thailand.The study employed the Life Cycle Assessment (LCA) methodology, specifically focusing on the human health damage impact category within the ReCiPe 2016 method 6 .The impact of particulate matter on human health is commonly measured in Disability-Adjusted Life Years (DALY).The study examines the relationship between exposure to ambient concentrations of PM 2.5 and PM 10 , utilizing the annual mean PM concentration data from Thailand's Pollution Control Department (PCD) 1 during the period of 2014 to 2018.Additionally, the study aims to quantify the burden of disease associated with PM 2.5 and PM 10, including mortality and morbidity resulting from exposure to these pollutants.
Focus area.This study focuses on the northern region of Thailand, which comprises ten provinces: Chiang Rai, Mae-Hong-Son, Chiang Mai, Phayao, Lamphun, Lampang, Phrae, Nan, Tak, and Nakhon-Sawan.Most areas in northern Thailand are characterized by hilly terrain and serve as the source of several significant rivers.These hill ridges run in a north-south orientation, stretching parallel from west to east, intersected by various major valleys, particularly those near Chiang Mai, Chiang Rai, Lampang, and Nan provinces.The highest mountain in the region is Doi Inthanon, rising approximately 2595 m above mean sea level, located in Chiang Mai.Along the eastern border, adjacent to the northeastern part, lies a mountainous area known as the central highlands.Between the western mountains and the central highlands, there is a central valley in the southern portion of the region 32 .Agricultural land use in the northern region spans approximately 6,368,630 hectares, estimated to account for 40 percent of the total land use.This includes approximately 41 percent allocated to paddy fields and 32 percent dedicated to field crops 33 .
The inventory data for this study included PM 2.5 and PM 10 concentrations, represented as annual mean concentrations in μg/m 3 from the Pollution Control Department (PCD) of Thailand in ten provinces for the period of 2014 to 2018.The locations of monitoring stations are indicated by cross symbols on Landsat-9 images as shown in Fig. 4 and Table 1.
The input parameters for calculating the CF or human health damage are presented in Table 2.The calculation of human health damage was taken into consideration.
Regarding PM 2.5 , the Relative Risk (RR) value for cardiopulmonary disease is 1.013 per μg/m 3 which includes, ischemic heart disease, stroke, and respiratory diseases.The RR for lung cancer is 1.014 per μg/m 335 .For PM 10 , the RRs for cardiopulmonary and lung cancer are 1.0005 36 .and 1.27 per μg/m 337 , respectively.This study obtained YLL for four age groups (30-49, 50-59, 60-69, and 70 and older), and the MR per health effect were taken from the World Health Organization 34 .The calculation used the number of adults in northern Thailand aged ≥ 30 years, which amounted to 14,252 persons 38 .

Human health damage.
In this section, the CF for human health damage caused by PM 2.5 and PM 10 in 10 provinces of northern Thailand is evaluated.CFs are addressed based on the marginal change in DALYs of humans resulting from a marginal change in the annual concentrations of PM 2.5 and PM 10 .The applied fate and exposure analysis model is presented in Fig. 5.
CFs for human health damage.The data of PM 2.5 and PM 10 concentrations, collected from PCD, ranged from the period from 2014 to 2018.This annual dataset allows for the quantification of the burden of disease from PM 2.5 and PM 10 related mortality and morbidity due to exposure to these pollutants through Life Cycle Impact Assessment.The calculations and assumptions for CF for human health damage consist of three factors: (a) intake factor (IF), (b) human effect factor (EF) and (c) damage factor (DF).Overall, CF (DALY/kg emitted ) for human health damage caused by PM 2.5 and PM 10 are provided in Eq. (1) below 22,25 : where IF is the intake factor, indicating the amount of pollutant inhaled by the population per unit of pollutant emitted (kg inhaled /kg emitted ), EF is the human effect factor, which reflects the change in disease incidence resulting from a change in exposure to the pollutant (DALYs/kg inhaled ), DF is the damage factor, which quantifies the harm caused by the pollutant.CF is the characterization factor, indicating the annual marginal change in DALYs per unit increase in the ambient concentrations (DALY/substance emitted ) Intake factor (IF).Instead of using the more commonly used intake fraction, this study employs an intake factor, IF pop,x having dimensionless units to expresses the population intake of pollutant k (I pop,k in kg/year) per unit emission of substance x 39 .The reason for this choice is that the intake fraction would imply the intake of a fraction of the emission itself, whereas in the case of secondary aerosols, the primary emission is a different substance.
The calculation of IF pop,x can be performed using the formula described by Kassomenos et al. 22 (1) where N is the number of inhabitants in the area, C m is the annual average concentration of the pollutant, m (kg/m 3 ), M x is the annual emissions of substance x (kg/year), IH is the average human health intake rate (13 m 3 / day = 4745 m 3 /year, US EPA 1997).
For this study, it is assumed that the CF values were calculated per person (i.e., N=1).Additionally, concentration increments of 1% were used, i.e., dC/dM (the change in concentration per unit change in emissions, M x ) was considered as a 1% increase of the average concentration C m .
Effect factor (EF).An effect factor (EF) represents the human effect caused by the pollutant in the receptor and quantifies the change in disease incidence resulting from a change in exposure.It is determined by dividing the concentration-response function (CRF) in m 3 /year/kg by the BR (m 3 /year) (2) (3) EF e,k,j = dINC k,j dEXP k,j = CRF e,k,j BR www.nature.com/scientificreports/ The region-specific CRF can be calculated using the following Equation 25: where CRF is the concentration-response function (m 3 /year/kg), RR e,k is the relative risk associated with obtaining a specific health effect, e, due to exposure to the pollutant, k (per μg/m 3 ), MR e,j is the mortality rate for the specific health effect, e, in region j in terms of deaths/person/year, C k,j is the yearly average background concentration of pollutant k in in region (μg/m 3 ), BR is the breathing rate (m 3 /year).By applying these calculations, the specific CRF values can be obtained, which quantify the relationship between pollutant exposure and health effects in a given region.

Damage factor 25 (DF).
A damage factor (DF e,k,j ) is a measure of the years of life lost (YLL) associated with a specific health effect (e) per incidence case.The YLL value associated with the health effect (e) per incidence case in region j for a specific pollutant k is estimated using data from the from the World Health Organization 34 .The DF can be calculated using the following equation: This equation allows for the quantification of the impact of a specific health effect caused by a pollutant in a given region in terms of years of life lost.

Results and discussion
Human health damage.On the other hand, the highest IF value for inhaled PM 10 was observed at Nakhon-Sawan, NS (41T) station.Nakhon-Sawan is a large city that encompasses both urban and agricultural areas.The station at Nakhon-Sawan only measured PM 10 and did not have PM 2.5 measurements during the study period.The region is predominantly dedicated to agriculture, with rice and maize crops being the most common.Nakhon-Sawan also has a high population density.Based on the available PM 10 monitoring data, the results indicated that Nakhon-Sawan had the highest IF value for PM 10 , followed by Chiang Mai, CNX (36T), and Lampang, LPG (37T), stations, respectively.
These findings suggest that the concentration and distribution of PM 2.5 and PM 10 pollutants, as well as population density and specific local factors such as traffic and agricultural activities, contribute to variations in the IF values across different stations and provinces in northern Thailand.

Results of EF.
The results of the effect factor (EF) analysis, expressed as EF factors, are presented in Table 4.The focus of this study was on EF for mortality associated with dominant diseases, namely lung cancer disease and cardiopulmonary diseases, which included ischemic heart disease, stroke, and respiratory diseases 34 , caused by PM 2.5 and PM 10 pollutants.
The variables used in the calculation of EF include the concentration-response function (CRF), relative risk (RR), mortality rate (MR), annual average background concentration of the pollutant, and breathing rate (BR).With the exception of CRF, which was related to the annual average background concentration of the pollutant at each station, baseline parameters were mostly used for the calculations.These variables can influence the EF value of each specific area.
The highest EF value for PM 2.5 -related mortality from lung cancer and cardiopulmonary diseases was found at the Tak, TK (76 T), station, with values of approximately 7.86E-10 DALYs/kg inhaled and 5.03E−09 DALYs/ kg inhaled , respectively.It is worth noting that Tak, where the TK (76 T) station is located, is in close proximity to neighboring countries like Myanmar, and the increased concentration observed in this area from 2016 to 2018 could be influenced by transboundary pollution.
For PM 10 exposure, the highest EF values for lung cancer and cardiopulmonary diseases were observed at the Lampang, LPG (39 T), station, which is situated in the coal power plant area in the Mae-Moh district of Lampang.The calculated EF values for lung cancer and cardiopulmonary diseases at this station were approximately 1.75E−09 DALYs/kg inhaled and 2.17E−10 DALYs/kg inhaled , respectively.
These results indicate the varying EF values for different diseases and pollutants across the monitoring stations in northern Thailand.Specific local factors, such as transboundary pollution and proximity to coal power plant areas, can contribute to higher EF values in certain regions.

Results of CFs.
The results of the characterization factor (CF) calculation for human health damage, which represents the marginal change in DALYs (Disability-Adjusted Life Years) due to a marginal change in the annual concentrations of PM 2.5 and PM 10 , are shown in Figs. 6 and 7.The highest CF value for PM 2.5 was found at the www.nature.com/scientificreports/Chiang Mai, CNX (36T), located in the traffic area of Chiang Mai city at the Yupparaj Wittayalai School, with a value of approximately 1.23E−04 year/kg.The next highest CF value was observed at the Lampang, LPG (40T), station, situated in the power plant area of Lampang, at approximately 1.15E−04 year/kg.These results correspond to the intake factor (IF) values of PM 2.5 , indicating that the most affected areas in terms of health damage from PM 2.5 are Chiang Mai and Lampang provinces.
Regarding the CF of PM 10 , the results showed that the highest CF values were observed at Nakhon-Sawan, NS (41T), and Chiang Mai, CNX (36T), stations with values of 8.00E−05 year/kg and 7.85E−05 year/kg, respectively.These CF results align with the IF values of PM 10 , which are highest in the urban areas of Nakhon-Sawan and Chiang Mai.It is worth noting that the CF of PM 2.5 appears to be higher than that ofPM 10 , indicating that particulates with an aerodynamic diameter less than or equal to 2.5 micrometers have a greater impact on human health.
In addition, this study calculated the average years of life lived with disability (YLD) associated with ambient PM 2.5 in northern Thailand, which amounted to approximately 48,372 years per 100,000 people.This calculation was based on the PM 2.5 emissions data for Thailand in 2018 from the Emissions Database for Global Atmospheric Research (EDGAR) 40 .Furthermore, it was found that the number of deaths from lung cancer and cardiopulmonary diseases caused by PM 2.5 was approximately 0.04% and 0.06% of the population of northern Thailand, respectively.Similarly, for PM 10 emissions in Thailand in 2018, the YLD for the entire population of northern Thailand was estimated to be around 59,064 years.The percentage of deaths from lung cancer and cardiopulmonary diseases caused by PM 10 was approximately 0.06% and 0.08%, respectively.
Comparison with other studies.The results of this study were compared to those obtained from published Life Cycle Assessments (LCAs) that focused on characterization factors (CFs) for human health damage 19,22,23,25 .The intake factor (IF) values for PM 2.5 in this study ranged from 5.1.E−07 to 1.2.E−06, whereas Van Zelm et al. 25 reported a value of 1.94E−06 for primary PM 2.5 for the entire country of Thailand.
It is worth noting that the IF values obtained in this study for the five PM 2.5 monitoring stations in the northern region were lower than the IF value reported by Van Zelm et al. for the entire country.This difference can be attributed to the fact that the sources of PM 2.5 in northern Thailand, such as agricultural waste burning or forest fires, may be different from other regions in Thailand.Therefore, the IF values for PM 2.5 in the northern region may not be representative of the entire country.This comparison in illustrated in Fig. 8, which presents the IF values for PM 2.5 from this study along with values from other studies.
When comparing the characterization factor (CF) for PM 2.5 in this study with the regionalized CF by Van Zelm et al. for the entire country of Thailand, it was found that the CF in this study is lower.Van Zelm et al. study aimed to quantify the overall human health damage caused by a specific pollutant per unit of emission of a primary PM 2.5 precursor across Thailand.Therefore, the human health damage for the five PM 2.5 stations in northern Thailand is lower than that for the entire country.
Furthermore, the average results of this study were compared to CF values from 63 densely populated urban areas in the United States 23 .It was found that the CF values in this study are lower than those for the US  sources and concentrations between countries.It suggests that the PM 10 emissions released in the northern region of Thailand may be lower compared to other countries in Europe and Greece.However, when comparing the IF values for each station in northern Thailand to the literature values, it is observed that some stations, such as Nakhon-Sawan, NS (41 T), exhibit higher IF values.This indicates that these stations have higher concentrations of PM 10 , leading to higher population intake per unit of emission.This can be attributed to specific local factors and sources in those areas.
In summary, the IF and CF values for PM 10 in northern Thailand, when compared to other studies, show lower values on a broader regional scale (Europe and Athens), but there are variations within the region itself.
Summary.The differences observed in the CF values for both PM 2.5 and PM 10 between this study and other studies can be attributed to various factors related to emissions sources, meteorology, climate, geography, and other contextual variables.Each country has its own unique mix of air pollutants emitted, which can be influenced by industrial activities, transportation, energy production, and other sources.
Additionally, meteorological conditions such as wind patterns, temperature, and precipitation can affect the dispersion and transport of air pollutants, leading to variations in concentration levels.Climate factors such as temperature inversions or stable atmospheric conditions can also contribute to higher pollutant concentrations in specific areas.
Geographical factors play a role as well.Countries and regions with different topography, land use patterns, and proximity to pollution sources may experience varying levels of pollution and associated health impacts.Urban areas, with their higher population densities and increased pollution sources, may have different CF values compared to rural or remote areas.
Furthermore, CF calculations are influenced by other factors such as mortality rates, years of life lost, and incident cases, which can vary from country to country.These factors reflect the specific health burden and vulnerability of the population in each location, and they contribute to the estimation of CF and the overall health damage associated with air pollution.
To ensure accurate and precise estimation of health damages specific to each country, it is crucial to use appropriate and context-specific input parameters for CF calculations.This includes considering the unique emissions sources, meteorological conditions, climate patterns, geography, and health-related factors of the particular country or region under study.
With regards to the results of this study, we found that there is some limitation of these input parameters for calculating the CF for human health damage that were used.Due to the lack of observed data, values from scientific reports, existing research, and publications were referred to.Only the results of IF and CF in Europe and Thailand are compared.Considering the IF and CF estimated of pollutant in northern Thailand for this study, the results of the fate and exposure analysis were significant.It is important to observe and analyze how modeled concentrations and temporal variations of emission substances in different regions are represented.This study was limited due to a lack of observed data for calculating the EF of human caused by the pollutant, which represents the change in disease incidence due to a change in exposure as a result of the estimated EF.More specifically, the value of the region-specific CRF is necessary for accurate calculation.Moreover, the DF calculation used statistical data of YLL associated with the health effect, which was estimated per region.Therefore, more observed data is required to estimate the result of the diseases on which we have focused.In summary, the combined intake, effect, and DF presented in CFs can reveal differences in area depending on the context that may influence PM characterization for different areas.In the long term, more specific data for CF calculation is required for related variables such as RR value, MR, or YLL in each area in order to evaluate the most accurate

Figure 2 .
Figure 2. Number of hospital admissions affected by air pollution in northern Thailand in 2020.

Figure 3 .
Figure 3.The cause-and-effect relationship of fine particulate matter emissions to human health damage 6 .

Figure 4 .
Figure 4. Location coordinates of 16 monitoring stations operating in northern Thailand.

Figure 5 .
Figure 5.The procedure of determining CF on human health damage.

10 Figure 10 .
Figure 10.Comparison of IF for PM 10 with other study.

Figure 11 .
Figure 11.Comparison of CF for PM 10 with other studies.

Table 1 .
The monitoring data set on annual mean of PM 2.5 and PM 10 concentrations in northern Thailand.

Table 3 .
Results of IF.high population denstity.As a result, Chiang Mai had the highest IF value for PM 2.5 , followed by Lampang, LPG (40T), another station in Chiang Mai, CNX (35T) and Nan, NN (75T), stations, respectively.

Table 4 .
Results of EF for mortality due to PM 2.5 , PM 10 exposure in Northern Thailand.

06 2.00.E-06 2.08.E-06 2.15.E-06 2.38.E-06 1.82.E-06 1.73.E-06 2.06.E-06 1.82.E-06 2.11.E-06 1.86.E-06 1.57.E-06 2.05.E-06 1.74.E-06 2.08E-06 2.02.E-06 CR CR
22w.nature.com/scientificreports/metropolitanareas.The differences can be attributed to variations in emissions sources and the characteristics of a large country like the United States, which includes urban, rural, and remote locations.Additionally, the input data used for CF calculations, such as concentration and dose-response factors, annual mortality rate, and specific diseases in each metropolitan area, influenced the CF values.This comparison is depicted in Fig.9, which presents the CF values for PM 2.5 from this study along with values from other studies.The comparison results of the intake factor (IF) and characterization factor (CF) for PM 10 with other studies are presented in Figs.10 and 11.When comparing the IF values for PM 10 from this study with studies conducted in Europe19and greater area of Athens, Greece22, it was found that the IF values for all 14 stations in this study are lower.This indicates that the population intake of PM 10 per unit of emission of substance is lower in northern Thailand compared to Europe and Athens.The differences in IF values can be attributed to variations in emission Vol.:(0123456789) Scientific Reports | (2023) 13:12753 | https://doi.org/10.1038/s41598-023-39930-9CF(yr/kg) CF of PM 2.5 (yr/kg) Figure 9.Comparison of CF for PM 2.5 with other studies.1.82.E-