Christine Pham / Earth & Environmental Sciences / Faculty Mentor: Yunyao Li
Due to the growing impacts of climate change, wildfires have become increasingly frequent and intense, with significant consequences for air quality and, by extension, human health. A major area of concern related to wildfire emissions is particulate matter. Because of its links to adverse health outcomes, various organizations have developed forecasting models to predict surface-level particulate matter 2.5µm (PM2.5) concentrations. However, predicting PM2.5 concentrations is challenging, especially due to the unpredictable nature of wildfires. As a result, the accuracy of these models can vary significantly.
This case study focuses on the 2025 Los Angeles wildfire. To evaluate the accuracy and effectiveness of different forecasting models, we compare six widely used models—HRRR, GEOS-CF, GEFS, NAAPS, NAQFC, ECCC—and their ensemble mean against EPA ground-level observations to determine their predictive reliability and spatial resolution at varying scales.
By interpolating and analyzing model outputs at the census tract level, we can better understand the impacts of wildfire-related PM2.5 on human health. This study also examines how predictive air quality models can inform policymaking and decision-making at both local and federal levels. Additionally, identifying each model’s strengths and limitations provides valuable insights for improving forecasting accuracy and enhancing future air quality modeling efforts.
Leave a Reply