Robust Spatiotemporal Estimation of PM Concentrations Using Boosting-Based Ensemble Models

Particulate matter (PM) as an air pollutant is harmful to the human body as well as to the ecosystem. It is crucial to understand the spatiotemporal PM distribution in order to effectively implement reduction methods. However, ground-based air quality monitoring sites are limited in providing reliable concentration values owing to their patchy distribution. Here, we aimed to predict daily PM10 concentrations using boosting algorithms such as gradient boosting machine (GBM), extreme gradient boost (XGB), and light gradient boosting machine (LightGBM). The three models performed well in estimating the spatial contrasts and temporal variability in daily PM10 concentrations. In particular, the LightGBM model outperformed the GBM and XGM models, with an adjusted R2 of 0.84, a root mean squared error of 12.108 μg/m2, a mean absolute error of 8.543 μg/m2, and a mean absolute percentage error of 16%. Despite having high performance, the LightGBM model showed low spatial prediction accuracy near the southwest part of the study area. Additionally, temporal differences were found between the observed and predicted values at high concentrations. These outcomes indicate that such methods can provide intuitive and reliable PM10 concentration values for the management, prevention, and mitigation of air pollution. In the future, performance accuracy could be improved through consideration of different variables related to spatial and seasonal characteristics.

Comparison of PM10 Sources at Traffic and Urban Background Sites Based on Elemental, Chemical and Isotopic Composition: Case Study from Krakow, Southern Poland

In large urban agglomerations, car traffic is one of the main sources of particulate matter. It consists of particulate matter directly generated in the process of incomplete liquid fuel burning in vehicle engine, secondary aerosols formed from exhaust gaseous pollutants (NOx, SO2) as well as products of tires, brake pads and pavement abrasion. Krakow is one of the cities in Europe with the highest concentrations of particulate matter. The article presents the results of combined elemental, chemical and isotopic analyses of particulate matter PM10 at two contrasting urban environments during winter and summer seasons. Daily PM10 samples were collected during the summer and winter seasons of 2018/2019 at two stations belonging to the network monitoring air quality in the city. Mean PM10 concentrations at traffic-dominated stations were equal to 35 ± 7 µg/m3 and 76 ± 28 µg/m3 in summer and winter, respectively, to be compared with 25.6 ± 5.7 µg/m3 and 51 ± 25 µg/m3 in summer and winter, respectively, recorded at the urban background station. The source attribution of analyzed PM10 samples was carried out using two modeling approaches: (i) The Positive Matrix Factorization (PMF) method for elemental and chemical composition (concentrations of elements, ions, as well as organic and elemental carbon in daily PM10 samples), and (ii) Isotope Mass Balance (IMB) for 13C and 14C carbon isotope composition of carbonaceous fraction of PM10. For PMF application, five sources of particulate matter were identified for each station: fossil fuel combustion, secondary inorganic aerosols, traffic exhaust, soil, and the fifth source which included road dust, industry, construction work. The IMB method allowed the partitioning of the total carbon reservoir of PM10 into carbon originating from coal combustion, from biogenic sources (natural emissions and biomass burning) and from traffic. Both apportionment methods were applied together for the first time in the Krakow agglomeration and they gave consistent results.

PM2.5 Pollution Strongly Predicted COVID-19 Incidence in Four High-Polluted Urbanized Italian Cities during the Pre-Lockdown and Lockdown Periods

Background: The coronavirus disease in 2019 (COVID-19) heavily hit Italy, one of Europe’s most polluted countries. The extent to which PM pollution contributed to COVID-19 diffusion is needing further clarification. We aimed to investigate the particular matter (PM) pollution and its correlation with COVID-19 incidence across four Italian cities: Milan, Rome, Naples, and Salerno, during the pre-lockdown and lockdown periods. Methods: We performed a comparative analysis followed by correlation and regression analyses of the daily average PM10, PM2.5 concentrations, and COVID-19 incidence across four cities from 1 January 2020 to 8 April 2020, adjusting for several factors, taking a two-week time lag into account. Results: Milan had significantly higher average daily PM10 and PM2.5 levels than Rome, Naples, and Salerno. Rome, Naples, and Salerno maintained safe PM10 levels. The daily PM2.5 levels exceeded the legislative standards in all cities during the entire period. PM2.5 pollution was related to COVID-19 incidence. The PM2.5 levels and sampling rate were strong predictors of COVID-19 incidence during the pre-lockdown period. The PM2.5 levels, population’s age, and density strongly predicted COVID-19 incidence during lockdown. Conclusions: Italy serves as a noteworthy paradigm illustrating that PM2.5 pollution impacts COVID-19 spread. Even in lockdown, PM2.5 levels negatively impacted COVID-19 incidence.

Comparing Methods to Impute Missing Daily Ground-Level PM10 Concentrations between 2010–2017 in South Africa

Good quality and completeness of ambient air quality monitoring data is central in supporting actions towards mitigating the impact of ambient air pollution. In South Africa, however, availability of continuous ground-level air pollution monitoring data is scarce and incomplete. To address this issue, we developed and compared different modeling approaches to impute missing daily average particulate matter (PM10) data between 2010 and 2017 using spatiotemporal predictor variables. The random forest (RF) machine learning method was used to explore the relationship between average daily PM10 concentrations and spatiotemporal predictors like meteorological, land use and source-related variables. National (8 models), provincial (32) and site-specific (44) RF models were developed to impute missing daily PM10 data. The annual national, provincial and site-specific RF cross-validation (CV) models explained on average 78%, 70% and 55% of ground-level PM10 concentrations, respectively. The spatial components of the national and provincial CV RF models explained on average 22% and 48%, while the temporal components of the national, provincial and site-specific CV RF models explained on average 78%, 68% and 57% of ground-level PM10 concentrations, respectively. This study demonstrates a feasible approach based on RF to impute missing measurement data in areas where data collection is sparse and incomplete.

Public Health Considerations for PM10 in a High-Pollution Megacity: Influences of Atmospheric Condition and Land Coverage

This paper analyzes the PM10 concentrations and influences of atmospheric condition (AC) and land coverage (LC) on a high-pollution megacity (Bogota, Colombia) from a public health viewpoint. Information of monitoring stations equipped with measuring devices for PM10/temperature/solar-radiation/wind-speed were used. The research period lasted eight years (2007–2014). AC and LC were determined after comparing daily PM10 concentrations (DPM10) to reference limits published by the World Health Organization (WHO). ARIMA models for DPM10 were also developed. The results indicated that urban sectors with lower atmospheric instability (AI) had a 2.85% increase in daily mortality (DM) in relation to sectors with greater AI. In these sectors of lower AI, impervious LC predominated, instead of vegetated LC. An ARIMA analysis revealed that a greater extent of impervious LC around a station led to a greater effect on previous days’ DPM10 concentrations. Extreme PM10 episodes persisted for up to two days. Extreme pollution episodes were probably also preceded by low mixing-layer heights (between 722–1085 m). The findings showed a 13.0% increase in WHO standard excesses (PE) for each 10 µg/m3 increase in DPM10, and a 0.313% increase in DM for each 10% increase in PE. The observed average reduction of 14.8% in DPM10 (−0.79% in DM) was probably due to 40% restriction of the traffic at peak hours.

Evaluation of Boosted Regression Tree for the Prediction of the Maximum 24-Hour Concentration of Particulate Matter

Air pollution is a considerable health danger to the environment. The objective of this study was to assess the characteristics of air quality and predict PM10 concentrations using boosted regression trees (BRTs). The maximum daily PM10 concentration data from 2002 to 2016 were obtained from the air quality monitoring station in Kuching, Sarawak. Eighty percent of the monitoring records were used for the training and twenty percent for the validation of the models. The best iteration of the BRT model was performed by optimizing the prediction performance, while the BRT algorithm model was constructed from multiple regression models. The two main parameters that were used were the learning rate (lr) and tree complexity (tc), which were fixed at 0.01 and 5, respectively. Meanwhile, the number of trees (nt) was determined by using an independent test set (test), a 5-fold cross validation (CV) and out-of-bag (OOB) estimation. The algorithm model for the BRT produced by using the CV was the best guide to be used compared with the OOB to test the predicted PM10 concentration. The performance indicators showed that the model was adequate for the next day’s prediction (PA=0.638, R2=0.427, IA=0.749, NAE=0.267, and RMSE=28.455).

Spatiotemporal Evaluation of PM10 Concentrations within the Greater Athens Area, Greece. Trends, Variability and Analysis of a 19 Years Data Series

This work investigates the spatiotemporal variation of suspended particles with aerodynamic diameter less than or equal to 10 μm (PM10) during a nineteen years period. Mean daily PM10 concentrations between 2001 and 2018, from five monitoring stations within the greater Athens area (GAA) are used. The aim is to investigate the impact of the economic crisis and the actions taken by the Greek state over the past decade on the distribution of PM10 within the GAA. Seasonality, intraweek, intraday and spatial variations of the PM10 concentrations as well as trends of data, are statistically studied. The work may assist the formation of PM10 forecasting models of hourly, daily, weekly, monthly and annual horizon. Innovations are alternative ways of statistical treatment and the extended period of data, which, importantly, includes major economic and social events for the GAA. Significant decreasing trend in PM10 series concentrations at all examined stations were found. This may be due to economic and social reasons but also due to measures taken by the state so as to be harmonised with the European Directives concerning the protection of public health and the atmospheric environment of the European Union (EU) members.