Application of machine learning models for PM2.5 prediction in bengaluru using precursor air pollutants and meteorological data

Theoretical and Applied Climatology, Journal Year: 2025, Volume and Issue: 156(3)

Published: March 1, 2025

Language: Английский

---

Relationships between fixed-site ambient measurements of nitrogen dioxide, ozone, and particulate matter and personal exposures in Grand Paris, France: the MobiliSense study

International Journal of Health Geographics, Journal Year: 2025, Volume and Issue: 24(1)

Published: March 27, 2025

Past epidemiological studies, using fixed-site outdoor air pollution measurements as a proxy for participants' exposure, might have suffered from exposure misclassification. In the MobiliSense study, personal exposures to ozone (O3), nitrogen dioxide (NO2), and particles with aerodynamic diameters below 2.5 μm (PM2.5) were monitored quality monitor. All spatial location points collected GPS receiver mobility survey used retrieve background hourly concentrations of pollutants nearest Airparif monitoring station. We modeled 851,343 min-level observations 246 participants. Visited places including residence contributed majority minute-level observations, 93.0%, followed by active transport (3.4%), rest on-road rail transport, 2.4% 1.1%, respectively. Comparison station-measured each individual indicated low Spearman correlations NO2 (median across participants: 0.23), O3 (median: 0.21), PM2.5 0.27), varying levels correlation microenvironments (ranging 0.06 0.35 according microenvironment). Results mixed-effect models that was very weakly explained (R2 < 0.07) all pollutants. The R2 only few higher than 0.15, namely in microenvironment (R2: 0.25) 0.16) separated 0.20). Model fit slightly increased decreasing distance between Our results demonstrated relatively pollutants, confirming proxies can lead However, type are shown affect extent

Language: Английский

---

Predicting air pollution changes due to temperature increases in two Brazilian capitals using machine learning – a necessary perspective for a climate resilient health future

International Journal of Environmental Health Research, Journal Year: 2025, Volume and Issue: unknown, P. 1 - 15

Published: April 2, 2025

Given that climate change can exacerbate the health impacts of air pollutants, we evaluated impact temperature increase scenarios on pollutant levels (O3, PM2.5, and PM10) in Porto Alegre Recife, Brazil. Air pollutants meteorological data were collected, simulations performed using a Support Vector Machine model with radial basis function kernel, applying increases 0.5°C, 1.0°C, 1.5°C, 2.0°C to predict future concentrations. The analyzed seasonally annually. Pearson correlation principal component analyses (PCA) explored relation conditions. Simulations revealed rising temperatures do not uniformly lead increased concentrations; instead, effects are highly dependent local climatic In Alegre, O3 throughout year, peak 14.14% during summer + scenario, while PM2.5 PM10 also showed marked seasonal increases. Conversely, decreased some seasons but autumn, particulate matter summer. findings underscore need for systems consider these dynamics their management strategies through location-specific investigations emphasize importance policy-driven adaptive measures build climate-resilient systems.

Language: Английский

---

Mitigating Air Pollution Risks with Deep Learning: A Quantum-Optimized Approach for Nitrogen Dioxide Prediction in Los Angeles

Journal of Machine and Computing, Journal Year: 2025, Volume and Issue: unknown, P. 709 - 719

Published: April 5, 2025

Air pollution causes about seven million pre mature deaths globally every year, making it a critical issue that requires urgent attention. The key to mitigating its devastating effects lies in understanding nature, identifying sources and trends, predicting its. Accurate Real-time air forecasting is challenging task due spatiotemporal dynamics, requiring sophisticated modeling approaches. In our study, employed the Sequential Array-based Convolutional LSTM (SACLSTM) framework, which captures spatial temporal correlations by integrating deep CNNs for analysis with models prediction. To further enhance model's accuracy, optimized SACLSTM parameters using Quantum-based Draft Mongoose Optimization Algorithm (QDMOA). Using ten days of nitrogen dioxide (NO₂) data from Los Angeles County, developed sequential encoder-decoder network capable levels into future. By reformatting satellite quality images 5D tensor, achieved precise predictions concentrations across various locations time periods Angeles. Our results are thoroughly documented metrics visualizations, clearly demonstrating factors behind improved accuracy. comparison highlights effectiveness approach providing reliable forecasts.

Language: Английский

---

Analysis of land surface temperature drivers in Beijing’s central urban area across multiple spatial scales: An explainable ensemble learning approach

Energy and Buildings, Journal Year: 2025, Volume and Issue: unknown, P. 115704 - 115704

Published: April 1, 2025

Language: Английский

---

Optimization of Highway Performance and Safety by Integrated Multi-Criteria Decision-Making Techniques

International Journal of Pavement Research and Technology, Journal Year: 2024, Volume and Issue: unknown

Published: June 22, 2024

Language: Английский

---

SolarGAN for Meso-Level Solar Radiation Prediction at the Urban Scale: A Case Study in Boston

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(23), P. 4524 - 4524

Published: Dec. 2, 2024

Evaluating solar radiation distribution at the urban scale is crucial for optimizing placement and size of installations managing heat. This study introduces a method predicting using 2D mapping data, applying Generative Adversarial Network (GAN) model to city Boston. Traditional simulation methods, such as 3D modeling satellite imagery, require complex resource-intensive data inputs. In contrast, this research allows open-source geographic information—such building footprints, heights, terrain—to predict various spatial scales (150 m, 300 500 m). The GAN model, detailed results, trained paired datasets information heatmaps. It achieved high accuracy resolution, with m demonstrating best performance (R2 = 0.864). model’s capability generate high-resolution (2 m) maps from simplified inputs demonstrates potential GANs climate prediction, offering rapid efficient alternative traditional methods. approach holds significant planning, particularly in photovoltaic (PV) system layouts UHI effect.

Language: Английский

---

Exploring Summer Variations of Driving Factors Affecting Land Use Zoning Based on the Surface Urban Heat Island in Chiang Mai, Thailand

ISPRS International Journal of Geo-Information, Journal Year: 2024, Volume and Issue: 13(7), P. 228 - 228

Published: June 30, 2024

Numerous studies have examined land surface temperature (LST) changes in Thailand using remote sensing, but there has been little research on LST variations within urban use zones. This study addressed this gap by analyzing summer zoning (LUZ) blocks the 2012 Chiang Mai Comprehensive Plan and their relationship with biophysical parameters (NDVI, NDBI, MNDWI). The approach integrated detailed data sensing for granular analysis. Correlation stepwise regression analyses (SRA) revealed that NDBI significantly impacted most block types, while NDVI MNDWI also influenced LST, particularly 2023. findings demonstrated complexity of dynamics across various LUZs Mai, SRA results explaining 45.7% to 53.2% over three years. To enhance environment, adaptive planning strategies different categories were developed will be considered upcoming revision Plan. offers a new method monitor heat island phenomenon at level, providing valuable insights planning.

Language: Английский

---

SolarGAN For Meso-Level Solar Radiation Prediction at the Urban-Scale: A Case Study in Boston

Published: Jan. 1, 2024

Language: Английский

---

Applicability of Machine Learning to Predict the Flexural Stresses in Jointed Plain Concrete Pavement

Journal of structural design and construction practice., Journal Year: 2024, Volume and Issue: 30(1)

Published: Oct. 12, 2024

Language: Английский

---