Knowledge-inspired fusion strategies for the inference of PM<sub>2.5</sub> values with a neural network
<p>Ground-level concentrations of particulate matter (more precisely PM<span class="inline-formula"><sub>2.5</sub></span>) are a strong indicator of air quality, which is now widely recognised to impact human health. Accurately inferring or predicting PM<sp...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2025-06-01
|
| Series: | Geoscientific Model Development |
| Online Access: | https://gmd.copernicus.org/articles/18/3707/2025/gmd-18-3707-2025.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | <p>Ground-level concentrations of particulate matter (more precisely PM<span class="inline-formula"><sub>2.5</sub></span>) are a strong indicator of air quality, which is now widely recognised to impact human health. Accurately inferring or predicting PM<span class="inline-formula"><sub>2.5</sub></span> concentrations is therefore an important step for health hazard monitoring and the implementation of air-quality-related policies. Various methods have been used to achieve this objective, and neural networks are one of the most recent and popular solutions. In this study, a limited set of quantities that are known to impact the relation between column aerosol optical depth (AOD) and surface PM<span class="inline-formula"><sub>2.5</sub></span> concentrations are used as input of several network architectures to investigate how different fusion strategies can impact and help explain predicted PM<span class="inline-formula"><sub>2.5</sub></span> concentrations. Different models are trained on two different sets of simulated data, namely, global-scale atmospheric composition reanalysis provided by the Copernicus Atmosphere Monitoring Service (CAMS) and higher-resolution data simulated over Europe with the Centre National de Recherches Météorologiques ALADIN model. Based on an extensive set of experiments, this work proposes several models of knowledge-inspired neural networks, achieving interesting results from both the performance and interpretability points of view. Specifically, novel architectures based on boundary condition generative adversarial networks (BC-GANs, which are able to leverage information from sparse ground observation networks) and on more traditional UNets, employing various information fusion methods, are designed and evaluated against each other. Our results can serve as a baseline benchmark for other studies and be used to develop further optimised models for the inference of PM<span class="inline-formula"><sub>2.5</sub></span> concentrations from AOD at either the global or regional scale.</p> |
|---|---|
| ISSN: | 1991-959X 1991-9603 |