Real-Time PM2.5 Monitoring Issues and their Impacts on Air Quality Forecasting

Forecasting air quality has grown substantially in the last ten years. For the majority of agencies forecasting, ozone was the only pollutant predicted until 2003. In recent years, health studies and new EPA standards are creating increased focus of PM2.5 impacts on public health. As a result, more air quality agencies are now forecasting PM2.5 in addition to ozone. The monitoring methods for ozone are accepted, which results in generally a high level of data quality. The methods for continuous PM2.5 monitoring are more complex with varied techniques; this results in more data quality issues. As PM2.5 forecasting increases, so do the methods and tools used to make predictions. This paper focuses on how uncertainties in PM2.5 measurements impact different air quality forecast methods (human forecasts, statistical models, and numerical modeling). Specifically, we examine the impact of uncertainties in PM2.5 data on forecast verification results of these different measurement methods. In addition, problems with comparing real-time modeled PM2.5 with real-time observations will be presented.