Dataset of air pollutants (PM_2.5, PM₁₀, CO) concentrations in the export processing area of Dhaka, Bangladesh

Abstract

This study presents a valuable dataset on air quality in the densely populated Dhaka Export Processing Zone (DEPZ) of Bangladesh. It included a dataset of Particulate Matter (PM_2.5, PM₁₀) and CO concentrations with Air Quality Index (AQI) values. PM data was collected 24h, and CO data was collected 8h monthly from 2019 to 2023 using respirable dust sampler APS-113NL for PM_2.5, APS-113BL for PM₁₀, and LUTRON AQ9901SD Air Quality Monitor Data Logger used to measure CO concentration data. Data sampling locations are selected based on population density, and employment data for DEPZ is also included, highlighting a potential rise in population density. This article also forecasted pollutant concentrations, AQI values, and health hazards associated with air pollutants using the Auto Regressive Moving Average (ARIMA) model. The performance of the ARIMA model was also measured using root mean squared error (RMSE) and mean absolute error (MAE). However, this can be used to raise awareness among the public about the health hazards associated with air pollution and encourage them to take measures to reduce their exposure to air pollutants. In addition, this data can be instrumental for researchers and policymakers to assess air pollution risks, develop control strategies, and improve air quality in the DEPZ.

Specifications Table

Subject	Atmospheric Science, Pollution
Specific subject area	Air pollutants data (PM2.5, PM10, CO) at Dhaka's export processing area.
Type of data	Chart Graph Figure
How data were acquired	Respirable dust sampler APS-113NL for PM2.5 and APS-113BL for PM10 was used to collect 24 h in a month and LUTRON AQ9901SD Air Quality Monitor Data Logger was used in day time 8 h, in a month to measure CO concentration data.
Data format	Raw, Analyzed
Description of data collection	Using a respirable dust sampler APS-113NL measured PM2.5, APS-113BL monitored PM10, and LUTRON AQ9901SD Air Quality Monitor Data Logger was used to measure CO concentration data month-wise from January 2019 to December 2023 and listed in an Excel sheet. Then, AQI was calculated using the Environmental Protection Agency's (EPA) breakpoint table and summarized.
Data source location	DEPZ, Bangladesh (Latitude: 23.94821, Longitude: 90.27727).
Data accessibility	Repository name: Mendeley Data DOI: 10.17632/cdn2vbvzgr.3 URL: https://data.mendeley.com/datasets/cdn2vbvzgr/3

1. Value of the Data

• •This dataset can be used to forecast air quality and assess the impact of industrial activities on air and public health. This information can be used to develop and implement air quality management plans.
• •PM_2.5, PM₁₀, and CO are all known to have adverse health effects. High levels of these pollutants can increase the risk of respiratory problems, heart disease, and other health problems. The dataset is used to assess the health impact of air pollution and develop interventions to reduce exposure.
• •PM_2.5, PM₁₀, and CO can also negatively impact the environment. The dataset can be used to assess the environmental impact of air pollution and develop measures to reduce pollution.
• •The dataset can be used to ensure industries inside the EPZ area comply with National Environmental policy and regulations.
• •The dataset can inform working people about air quality in the export processing area. This information can help people to make informed decisions about their health and well-being.

2. Data Description

Air is a fundamental prerequisite for the survival and development of all living things, affecting health and economic growth [1,2]. This dataset focuses on air quality monitoring within DEPZ in Bangladesh, aiming to assess potential air quality measures and potential health risks for its large workforce due to these pollutants. The data covers average PM_2.5, and PM₁₀ concentrations measured 24 h once a month, and CO measured (midday 8 h) once in a month from January 2019 to December 2023. Air quality data was collected at the BEPZA main gate within DEPZ and this will provide valuable information, at the same time further analysis should consider potential variations across the zone due to wind patterns and specific industrial activities. While weather data wasn't collected alongside air pollutants, incorporating general Dhaka city climatic conditions for the data collection period can be explored. In addition, Dhaka features a tropical wet and dry climate, which translates to hot, humid summers and warm, drier winters. The hallmark of this climate is the monsoon season, which brings the vast majority of rainfall. Research by Chakraborty (2019) indicates that nearly 80 % of Dhaka's annual average rainfall of 1854 mm (73 inches) falls between May and September, coinciding with the southwest monsoon [3].

Exposure to high levels of air pollution significantly raises the risks of breathing difficulties, cough, lower respiratory tract infections, depression, and other health conditions. Children under five years, the elderly, and people with comorbidities such as diabetes, heart, or respiratory conditions are most vulnerable [4]. Surveillance of air pollutants is essential for evaluating control measures for air pollution in a city affected by increasingly heavy vehicles and daily increasing air pollution [5]. The control system will be more effective if the measured data and forecast are reliable and confidential [6].

Recognizing the valuable city-wide air quality data collected by the Department of Environment (DoE) for Dhaka, our dataset offers a complementary perspective by zooming in on the DEPZ area. This hyperlocal focus allows for a targeted assessment of health risks for the large DEPZ workforce potentially exposed to localized pollutants that might not be reflected in broader DoE air quality monitoring data. DEPZ-specific data, when combined with the DoE's data, provides a more comprehensive picture of air quality variations across Dhaka. This combined view empowers researchers, policymakers, and residents to track trends, pinpoint potential pollution sources within and around DEPZ, and develop targeted air quality management strategies for the area. This article describes a PM dataset of linked Mendeley data that was collected from the DEPZ area in Bangladesh, and covered from Jan 2019 - Dec 2023 to assess potential health risks for a large workforce. DEPZ is a densely populated zone with a varying employee range of 67859 to 94527 (Table 1).

Table 1.

Investment and employment scenario of DEPZ.

Year	Investment (Million US $)	Export (Million US $)	Employment (No.)
2018–19	1436.95	26,981.97	94,527
2019–20	1525.45	28,796.53	90,985
2020–21	1605.71	30,456.35	67,859
2021–22	1676.77	32,579.22	78,310
2022–23	1737.34	34,381.16	79,045

• Source: Bangladesh Export Processing Zone Authority (BEPZA).

• website: https://www.bepza.gov.bd/epz-profile/dhaka-epz.

However, the focus was on PM_2.5, PM₁₀, and CO concentrations measured monthly using calibrated instruments APS-113NL for PM_2.5, APS-113BL for PM₁₀, and LUTRON AQ9901SD for CO at the BEPZA main gate. Data was stored date-wise in an excel sheet, and analysis was done using RStudio (version 2023.09.1 + 494).

3. Experimental Design, Materials, and Methods

The PM_2.5, PM_10, and CO concentration values were processed to determine AQI value following equations (i), (ii), and (iii), respectively, and the breakpoint Table (Table 2), all prescribed by environmental protection agencies (EPAs) [[7], [8]].

$$P_{PM2.5}=\left\{\left(P_{high}-P_{low}\right)/\left(C_{high}-C_{low}\right)\right\}\times \left(AQI-C_{low}\right)+P_{low}$$ (i)

$$P_{PM10}=\left\{\left(P_{high}-P_{low}\right)/\left(C_{high}-C_{low}\right)\right\}\times \left(AQI-C_{low}\right)+P_{low}$$ (ii)

$$P_{CO}=\left\{\left(P_{high}-P_{low}\right)/\left(C_{high}-C_{low}\right)\right\}\times \left(AQI-C_{low}\right)+P_{low}$$ (iii)

Where,

Table 2.

Air quality index (AQI) and breakpoint prescribed by environmental protection agencies.

PM2.5 (µg/m3)	PM10 (µg/m3)	CO (ppm)	AQI	Level of health concern
0–15.4	0–54	0–4.4	0–50	Good
15.5- 40.4	55–154	4.5–9.4	51–100	Moderate
40.5- 65.4	155–254	9.5–12.4	101–150	Unhealthy for sensitive group
65.5- 150.4	255–354	12.5–15.4	151–200	Unhealthy
150.5- 250.4	355–424	15.5–30.4	201–300	Very unhealthy
250.5–350.4	425–504	30.5–40.4	301–400	Hazardous
350.5–500.4	505–604	40.5–50.4	401–500	Very hazardous

P_high = Index breakpoint corresponding to C_high, P_low = Index breakpoint corresponding to C_low, C_high = The concentration breakpoint ≥ C, C_low = The concentration breakpoint ≤ C.

Forecasting of air pollutants PM_2.5, PM₁₀ and CO concentrations:

Box-Jenkins methods are used here for future prediction of the dataset. This method is classified as a linear model presenting both stationary and non-stationary time series [6,8]. Box-Jenkins methods, which include the Autoregressive (AR) models, the Integrated (I) models, and the Moving Average (MA) models, are of practical importance in forecasting. Four steps must be considered to obtain the model by the Box-Jenkins methodology: data preparation, model selection, estimation, and forecasting. Finally, the ARIMA model is used for prediction purposes using Rstudio (version 2023.09.1 + 494).

Air Quality monitoring summary:

Table 3.

Table 3.

Observed Air Quality summary from 2019 to 2023.

Variables	Year	Min	Max	Mean	AQI	Health Concern
PM2.5 (µg/m3) (24hr)	2019	10.30	121.30	52.40	101–150	Unhealthy for sensitive group
	2020	9.10	101.90	23.17	51–100	Moderate
	2021	10.00	131.00	53.99	101–150	Unhealthy for sensitive group
	2022	10.80	140.70	58.83	101–150	Unhealthy for sensitive group
	2023	14.00	115.50	57.07	101–150	Unhealthy for sensitive group
PM10 (µg/m3) (24hr)	2019	46.00	323.00	182.75	101–150	Unhealthy for sensitive group
	2020	41.00	303.00	91.00	51–100	Moderate
	2021	45.00	333.00	178.83	101–150	Unhealthy for sensitive group
	2022	48.00	343.00	187.75	101–150	Unhealthy for sensitive group
	2023	63.00	317.00	197.67	101–150	Unhealthy for sensitive group
CO (ppm) (8hr)	2019	3.70	14.50	9.74	101–150	Unhealthy for sensitive group
	2020	3.30	13.90	5.73	51–100	Moderate
	2021	3.60	14.80	9.50	101–150	Unhealthy for sensitive group
	2022	3.90	15.10	9.87	101–150	Unhealthy for sensitive group
	2023	4.90	14.30	10.25	101–150	Unhealthy for sensitive group

PM_2.5: Hazardous air contaminant. Particles of a diameter of 2.5 micrometers or less, small enough to pass past the throat and nose and into the lungs [[9], [10]]. Because air contaminants interact with one another, PM_2.5 levels, like other air pollutants, are rising at an alarming rate year after year. It varies according to the season (Fig. 1). Pollutant levels were relatively low during COVID-19 but dramatically surged in 2022 when the DEPZ reopened. In 2022, the highest PM_2.5 measurement was 140 micrograms, while the lowest was 101 micrograms.

Fig. 1.

Trend analysis of PM_2.5 from 2019 to 2023.

PM₁₀: Dangerous air pollutant. Particles with a diameter of 10 micrometers or less and these particles are small enough to pass through the throat and nose and enter the lungs [8] as air pollutants work combined. Like other air pollutants, PM₁₀ levels are also trending to increase year by year at an alarming rate. It shows seasonal variations. During the coronavirus period, the pollutant level was pretty low and suddenly increased in 2022 when DEPZ reopened. The highest PM₁₀ value was 343 micrograms in 2022, and the lowest was 303 micrograms in 2020 (Fig. 2).

Fig. 2.

Trend analysis of PM₁₀ from 2019 to 2023.

AQI: In the trend and time series graphs, 2021 (at random) indicates a rapid decrease or improvement in the air during COVID-19. Furthermore, air quality varies seasonally (Fig. 3). A chart depicting the hazardous status of air quality at the end of the year and the beginning of the year. The AQI was more significant prior to 2020, and after a brief decline, the AQI began to rise again, and it is continuing to rise on a daily basis. The AQI is rising, and the air is becoming increasingly contaminated as the industry resumes operations.

Fig. 3.

Trend analysis of AQI from 2019 to 2023.

CO: Carbon monoxide is another detrimental air pollutant. Due to carbon monoxide, the body starts to replace oxygen with carbon monoxide and creates a poisonous environment [11]. According to the trend and additive time series graph, the trend of CO is increasing significantly (Fig. 4). The season variation is visible and depicts a higher CO value at the beginning of the year. The highest volume is 14 micrograms. It influences the reason for the high demand for export at the end and starting of the year.

Fig. 4.

Trend analysis of CO from 2019 to 2023.

Forecasting of PM_2.5, PM_10, and CO concentrations in the study area

This data article included 2024 to 2027 forecasted concentrations of PM_2.5, PM_10, and CO in the study area. Forecasting of these pollutants is done using the ARIMA model in RStudio software, and represents the forecasted value graphically. Forecasted full dataset available in data respiratory (https://data.mendeley.com/datasets/cdn2vbvzgr/3).

However, using the ARIMA model, Fig. 5 shows the forecasted PM_2.5, PM₁₀, and CO concentration levels from 2024 to 2027. The black line indicates observed data until around 2023, after which forecasts are made, represented by the blue line surrounded by a confidence interval shaded in grey. The PM_2.5 and PM₁₀ concentration value shows significant fluctuations over the years, but the CO value shows no significant fluctuations over the years, and all are expected to remain within a certain range according to the forecast.

Fig. 5.

Graphical representation of the forecasted value of three pollutants.

The x-axis represents years from 2019 to 2027, while the y-axis represents PM_2.5, PM_10, and CO levels. The historical data of PM_2.5, PM_10, and CO levels is shown by a black jagged line with significant fluctuations. Post-2024, a blue line with a grey shaded area around it represents forecasted PM_2.5, PM_10, and CO levels and their confidence intervals, respectively.

According to Table 4, it can be said for PM_2.5, the model has moderate prediction errors, with RMSE slightly higher than MAE, suggesting potential for improvement; for PM₁₀ the model has larger prediction errors, especially in terms of RMSE, indicating a need for further refinement, and for CO the model shows relatively good prediction accuracy, with both RMSE and MAE values being low.

Table 4.

Performance measure of ARIMA forecasting model applied on pollutants.

Pollutants	RMSE	MAE
PM2.5	20.30706	13.64468
PM10	57.70847	40.93531
CO	3.184394	2.661667

Limitations

This is a significant limitation as factors like wind speed, direction, and precipitation can all impact how pollutants disperse in the atmosphere. Air quality can vary significantly across a zone due to factors like industrial activity or traffic patterns. Focusing on just one location doesn't capture this spatial variability. The dataset only covers PM_2.5, PM_10, and CO, while other pollutants like SO₂, NO_2, and O₃ can also be harmful.

Ethics Statement

This dataset does not involve human subjects, animal experiments, or data collected from social media platforms.

CRediT authorship contribution statement

Mustafizur Rahman: Conceptualization, Methodology, Software, Formal analysis, Writing – original draft, Visualization, Supervision. Faijunnesa Rashid: Data curation, Writing – original draft, Formal analysis. Diwakar Kumar: Data curation, Writing – review & editing. Md. Ahosan Habib: Data curation, Writing – review & editing. Ahmad Ullah: Writing – review & editing.

Data Availability

• Dhaka EPZ air pollution: PM2.5, PM10, and CO measurements and predictions (Original data) (Mendeley Data).

• The datasets are available at https://data.mendeley.com/datasets/cdn2vbvzgr/3.