High-resolution urban air pollution mapping

Science, Journal Year: 2024, Volume and Issue: 385(6707), P. 380 - 385

Published: July 25, 2024

Variation in urban air pollution arises because of complex spatial, temporal, and chemical processes, which profoundly affect population exposure, human health, environmental justice. This Review highlights insights from two popular situ measurement methods—mobile monitoring dense sensor networks—that have distinct but complementary strengths characterizing the dynamics impacts multidimensional quality system. Mobile can measure many pollutants at fine spatial scales, thereby informing about processes control strategies. Sensor networks excel providing temporal resolution locations. Increasingly sophisticated studies leveraging both methods vividly identify patterns that exposures disparities offer mechanistic insight toward effective interventions. summarizes limitations these discusses their implications for understanding fine-scale impacts.

Language: Английский

---

Machine learning applications for electrospun nanofibers: a review

Journal of Materials Science, Journal Year: 2024, Volume and Issue: 59(31), P. 14095 - 14140

Published: July 30, 2024

Abstract Electrospun nanofibers have gained prominence as a versatile material, with applications spanning tissue engineering, drug delivery, energy storage, filtration, sensors, and textiles. Their unique properties, including high surface area, permeability, tunable porosity, low basic weight, mechanical flexibility, alongside adjustable fiber diameter distribution modifiable wettability, make them highly desirable across diverse fields. However, optimizing the properties of electrospun to meet specific requirements has proven be challenging endeavor. The electrospinning process is inherently complex influenced by numerous variables, applied voltage, polymer concentration, solution flow rate, molecular weight polymer, needle-to-collector distance. This complexity often results in variations nanofibers, making it difficult achieve desired characteristics consistently. Traditional trial-and-error approaches parameter optimization been time-consuming costly, they lack precision necessary address these challenges effectively. In recent years, convergence materials science machine learning (ML) offered transformative approach electrospinning. By harnessing power ML algorithms, scientists researchers can navigate intricate space more efficiently, bypassing need for extensive experimentation. holds potential significantly reduce time resources invested producing wide range applications. Herein, we provide an in-depth analysis current work that leverages obtain target nanofibers. examining work, explore intersection ML, shedding light on advancements, challenges, future directions. comprehensive not only highlights processes but also provides valuable insights into evolving landscape, paving way innovative precisely engineered various Graphical abstract

Language: Английский

---

A Novel Evolutionary Deep Learning Approach for PM2.5 Prediction Using Remote Sensing and Spatial–Temporal Data: A Case Study of Tehran

ISPRS International Journal of Geo-Information, Journal Year: 2025, Volume and Issue: 14(2), P. 42 - 42

Published: Jan. 23, 2025

Forecasting particulate matter with a diameter of 2.5 μm (PM2.5) is critical due to its significant effects on both human health and the environment. While ground-based pollution measurement stations provide highly accurate PM2.5 data, their limited number geographic coverage present challenges. Recently, use aerosol optical depth (AOD) has emerged as viable alternative for estimating levels, offering broader spatial higher resolution. Concurrently, long short-term memory (LSTM) models have shown considerable promise in enhancing air quality predictions, often outperforming other prediction techniques. To address these challenges, this study leverages information systems (GIS), remote sensing (RS), hybrid LSTM architecture predict concentrations. Training models, however, an NP-hard problem, gradient-based methods facing limitations such getting trapped local minima, high computational costs, need continuous objective functions. overcome issues, we propose integrating novel orchard algorithm (OA) optimize forecasting. This paper utilizes meteorological topographical features, satellite imagery from city Tehran. Data preparation processes include noise reduction, interpolation, addressing missing data. The performance proposed OA-LSTM model compared five advanced machine learning (ML) algorithms. achieved lowest root mean square error (RMSE) value 3.01 µg/m3 highest coefficient determination (R2) 0.88, underscoring effectiveness models. employs binary OA method sensitivity analysis, optimizing feature selection by minimizing while retaining predictors through penalty-based function. generated maps reveal concentrations autumn winter spring summer, northern central areas showing levels.

Language: Английский

---

Agent-based modelling: A stochastic approach to assessing personal exposure to environmental pollutants – Insights from the URBANOME project

The Science of The Total Environment, Journal Year: 2025, Volume and Issue: 967, P. 178804 - 178804

Published: Feb. 13, 2025

In the context of URBANOME project, aiming to assess European citizens' exposure air pollutants (PM10, PM2.5, NO2) and noise, an extensive data collection process was undertaken. This involved distribution stationary home sensors, portable smartphone applications, alongside participants logging their activities while using these devices. By leveraging socioeconomic socio-demographic statistical for residents Thessaloniki, we developed agent-based model estimate levels based on movement patterns, locations, collected from campaign. The highlights that individual's is closely linked type they perform, location, age, gender. Whether occurs indoors, or outdoors important determining intake levels. Activity selections were found be strongly influenced by income, social connections, indicating socio-economic factors significantly shape patterns. analysis also revealed considerable differences between PM measurements taken fixed monitoring stations sensors used in Notably, even agents residing same household displayed distinct levels, underscoring variability within localized environments. Preliminary results campaign compared with ABM outputs, showing median values up 20 % both noise inhalation intakes. research emphasizes importance such models developing future scenarios large cities aimed at fostering green transitions enhancing quality life. These provide valuable insights designing strategies reduce improve urban living conditions.

Language: Английский

---

Machine learning-based hybrid regularization techniques for predicting unconfined compressive strength of soil reinforced with multiple additives

Multiscale and Multidisciplinary Modeling Experiments and Design, Journal Year: 2025, Volume and Issue: 8(5)

Published: March 20, 2025

Language: Английский

---

Population exposure to outdoor NO₂, black carbon, and ultrafine and fine particles over Paris with multi-scale modelling down to the street scale

Atmospheric chemistry and physics, Journal Year: 2025, Volume and Issue: 25(6), P. 3363 - 3387

Published: March 20, 2025

Abstract. This study focuses on mapping the concentrations of pollutants interest to health (NO2, black carbon (BC), PM2.5, and particle number concentration (PNC)) down street scale represent population exposure outdoor at residences. Simulations are performed over area Greater Paris with WRF-CHIMERE/MUNICH/SSH-aerosol chain, using either top-down inventory EMEP or bottom-up Airparif, correction traffic flow. The higher in streets than regional-scale urban background, due strong influence road emissions locally. Model-to-observation comparisons were background stations evaluated two performance criteria from literature. For BC, harmonized equivalent BC (eBC) estimated concomitant measurements eBC elemental carbon. Using corrected flow, strictest met for NO2, eBC, PNC. inventory, also but errors tend be larger lower along those simulated especially NO2 concentrations, resulting fewer heterogeneities. impact size distribution non-exhaust was analysed both regional local scales, it is heavy-traffic streets. To assess exposure, a French database detailing inhabitants each building used. population-weighted (PWC) calculated by weighting populations which they exposed precise location their home. An scaling factor (ESF) determined pollutant estimate ratio needed correct order exposure. average ESF ring 1 PNC because modelled scale. It indicates that Parisian underestimated concentrations. Although this underestimation low an 1.04, very high (1.26), (between 1.22 1.24), (1.12). shows heterogeneities important considered less so PM2.5.

Language: Английский

---

Spatial-temporal analysis of urban air pollution related exposure and health impacts: Driving human-centered regulation and control

Urban Climate, Journal Year: 2024, Volume and Issue: 58, P. 102161 - 102161

Published: Oct. 12, 2024

Language: Английский

---

Mapping environmental noise of Guangzhou based on land use regression models

Journal of Environmental Management, Journal Year: 2025, Volume and Issue: 373, P. 123931 - 123931

Published: Jan. 1, 2025

Language: Английский

---

Assessing the role of spatial aggregation schemes with varying campaign durations of mobile measurements on land use regression models for estimating nitrogen dioxide

Environmental Pollution, Journal Year: 2025, Volume and Issue: 368, P. 125689 - 125689

Published: Jan. 13, 2025

Mobile air pollution measurements are typically aggregated by varying road segment lengths, grid cell sizes, and time intervals. How these spatiotemporal aggregation schemas affect the modeling performance of land use regression models has seldom been assessed. We used 5.7 million mobile nitrogen dioxide (NO2) collected over 160 days in Amsterdam (The Netherlands) subsampled them into five campaign durations (10-70 days). from each duration onto segments cells with spatial scales (25-200 m). A stepwise linear (SLRs) random forests (RFs) were trained for dataset to predict NO2 concentrations. The model accuracies validated using a 30% hold-out sample external Palmes long-term stationary (n = 105). At increased scales, prediction accuracy decreased RFs but SLRs when against measurements. Using measurements, varied across without any clear pattern. Regardless or segments, performed similarly at small (i.e., 25 m 50 Models based on less sensitive than those validations. Longer concentrations, though gain diminished after days. In conclusion, our results suggest that preferred scale gets larger as this approach likely reduces scale-dependent influences. plays more important role scales.

Language: Английский

---

Use of a Land Use Regression Model Methodology for the Estimation of Individual Long-Term PM_2.5 Exposure Profiles of Urban Residents in Jiujiang City, China

Journal of Geoscience and Environment Protection, Journal Year: 2025, Volume and Issue: 13(01), P. 233 - 243

Published: Jan. 1, 2025

Language: Английский

---