List All Projects List Graduate Projects List Senior Design Projects List REU Projects Other Undergraduate Projects
Project # 43
Community-Engaged Research to Enhance Local Air Quality Predictions: Integrating Satellite-Based Particulate Matter Estimation into Multimodal Data-Fusion Technique
| Team Lead: | Justin Kim |
| Other Team Members: | Alice Tran |
| Faculty Advisor(s): | Michelle Hummel |
| Department: | CE |
| Project Type: | REU |
| Title: | Community-Engaged Research to Enhance Local Air Quality Predictions: Integrating Satellite-Based Particulate Matter Estimation into Multimodal Data-Fusion Technique |
| Abstract: | Severe air pollution can cause or exacerbate various health issues in the population. Recently, a multimodal data-fusion method that uses the stochastic advection and diffusion (SAD) model was developed to predict fine particulate matter (PM2.5) concentrations at a regional scale by processing data from multimodal air quality monitors. This model has demonstrated its ability to improve prediction accuracy compared to relying solely on state- or local-agency-operated regulatory-grade air monitors or low-cost sensors. However, the SAD model has so far only incorporated data from ground-level air monitors. Satellite imagery can also serve as a valuable data source to expand the multimodal data fusion framework. Recently, NASA published a new collection of high-resolution satellite image products (TEMPO dataset), including one for PM2.5. In this project, we integrate this new satellite-based data source to the SAD model to add details on surface PM2.5 and improve prediction quality. To achieve this, we compare the PM2.5 estimates from TEMPO with corresponding ground-level observations then evaluate the accuracy of the TEMPO dataset to consider estimation uncertainty and data credibility. Based on these results, TEMPO data will be integrated into the data-fusion framework to predict PM2.5 concentrations at higher spatial resolution and with broader spatial coverage compared to ground-based monitoring systems. Insights from this research can support improved prediction and planning in communities facing impaired air quality. |
| Poster: | View Poster #43 (PDF) |
| Judges Scoring: | Submit Evaluation (judges only) |