Wildfire smoke forecasts using HYSPLIT-based emission inverse modeling system and satellite observations

Extreme wildfire events became more frequent in recent years. During such events, the satellite measurements provided valuable information to monitor and predict the impacts of the extreme fires. In this study, an emission inverse modeling system to estimate wildfire smoke source strength, vertical distribution, and temporal variations by assimilating satellite observations with the HYSPLIT model will be presented. In the HYSPLIT-based emission inverse modeling system, a cost function is defined to mainly quantify the differences between model predictions and satellite measurements of column integrated air concentrations, weighted by the model and observation uncertainties. Smoke sources that minimize this cost function provide the optimal smoke emission estimates. The system will be tested using several recent wildfire events. Hindcasts using the emission estimates by the inverse system will be performed. Comparison between this new emission estimation system and the US Forest Service BlueSky emission prediction that is currently used by the US National Weather Service's operational forecasts will be conducted.