Kai Fan, Ranil Dhammapala, Kyle Harrington, Brian K. Lamb, Yunha Lee 

A machine learning (ML) based modeling framework has been successfully used to provide operational forecasts of O3 at Kennewick, WA. This paper shows its performance when applied to other observation locations to predict O3 and PM2.5 concentrations. The 10-time, 10-fold cross-validation method was used to evaluate the model performance in the Pacific Northwest (PNW). Similar to Kennewick, ML1 captures more high-O3 events, but also generates more false alarms, and the accuracy of ML2 is better (R2 = 0.79), especially for low-O3 events. Compared to AIRPACT, the combined modeling framework reduces the normalized mean bias (NMB) from 7.6% to 2.6%. In terms of Air Quality Index (AQI) forecasts, improvements occur for each AQI level which reflects more accurate O3 predictions and better capture of more high-O3 events. For PM2.5, ML1 and ML2 demonstrate similar capabilities to predict high-PM2.5 events and ML2 keeps its accuracy for low-PM2.5 predictions, so there is no need to combine the two methods. During the evaluation period, AIRPACT under-predicts the wildfire season PM2.5 concentrations in the PNW (NMB = -27%) and over-predicts at some sites in the cold season up to 200%, while ML2 has a lower NMB in both seasons (NMB = 7.9% in the wildfire season and 2.2% in the cold season) and correctly captures more high-PM2.5 events. The ML modeling framework is now operational for forecasts of O3 and PM2.5 at over 100 observation sites in the PNW.