Spatiotemporal analysis and forecasting of air quality in the greater Dhaka region and assessment of a novel particulate matter filtration unit

Abstract

Bangladesh is one of the most polluted nations in the world, with an average Air Quality Index (AQI) of 161 in 2021; its capital, Dhaka, has the worst air quality of any major city in the world. The present study aims to analyze the spatiotemporal distribution of air quality indicators in the greater Dhaka region, forecast weekly AQI, and assess the performance of a novel particulate matter filtration unit in removing particulate matter. Air quality indicators remained highest during the dry season with an average of 128.5 μm/m 3, while the lowest concentration was found in the monsoon season with an average of 19.096 μm/m 3 . Analysis revealed a statistically significant annual increasing trend of CO, which was associated with the growing number of brick kilns and usage of high-sulfur diesel. Except for the pre-monsoon AQI, concentrations of both seasonal and yearly AQI and PM 2.5 showed decreasing trend, though predominantly insignificant, demonstrating the improvement in air quality. Prevailing winds influenced the seasonal distribution of tropospheric CO & NO 2 . The study also employed a seasonal autoregressive integrated moving average (ARIMA) model to forecast weekly AQI values. ARIMA (3,0,4) (3,1,3) at the 7-periodicity level performed best forecasting the AQI values among all developed models with low root mean square error (RMSE)-29.42 and mean absolute percentage error (MAPE)-13.11 values. The predicted AQI values suggested that the air quality would remain unhealthy for most weeks. The experimental simulation of the particulate matter filtration unit, designed in the shape of a road divider, generated substantial cyclonic motion while maintaining a very minimal pressure drop. In the real-world scenario, using only cyclonic separation and dry deposition, the suggested air filtration system removed 40%, 44%, and 42% of PM 2.5 , PM 10 , and TSP, respectively. Without employing filters, the device removed significant amounts of particulate matter, implying enormous potential to be used in the study area. The study could be useful for policy makers to improve urban air quality and public health in Bangladesh and in other developing countries. Supplementary Information The online version contains supplementary material available at 10.1007/s10661-023-11370-y.

A Prediction Hybrid Framework for Air Quality Integrated with W-BiLSTM(PSO)-GRU and XGBoost Methods

“…Numerous researchers have enhanced classical models to yield improved prediction outcomes. For instance, the seasonal autoregressive integrated moving average (SARIMA) model [14,15] was adeptly employed to establish a prediction framework for the air quality index by fully considering time factors. In pursuit of long memory air quality data monitoring, Pan and Chen [16] proposed an autocorrelation data control chart based on the autoregressive fractional integral moving average (ARFIMA) model, which efficiently maintains the data quality by taking into account the complex time dependence.…”

Predicting Particulate Matter (PM10) Levels in Morocco: A 5-Day Forecast Using the Analog Ensemble Method.

“…Mo et al (2019) have highlighted that providing accurate 𝑃𝑀 10 forecasts for decision-makers is crucial to protecting public health and preventing potential health risks to the population. Although many studies have attempted to predict particulate matter using various tools (Agarwal & Sahu, 2023;Folifack Signing et al, 2024;Gul et al, 2022;Khan et al, 2022;Masood & Ahmad, 2023;Rahman & Kabir, 2023;Verma et al, 2023), research specifically predicting future levels of 𝑃𝑀 10 in Morocco remains limited. Some studies limit their use to a single station where the models were developed, producing only one forecast for the next hour or the next day (Adnane et al, 2022;Bouakline et al, 2022), or they were only validated at a single station, which limits the generalization of their models to other areas Ajdour, Leghrib, Chaoufi, Fetmaoui, et al, 2020;Saidi et al, 2023).…”