High-spatiotemporal-resolution inverse estimation of CO and NO_x emission reductions during emission control periods with a modified ensemble Kalman filter

摘要

Emission control strategies are among the most effective ways to improve air quality, but the emissions reduced were usually estimated with high uncertainty. Here we present a modified ensemble Kalman filter (EnKF) to reduce the uncertainties of carbon monoxide (CO) and nitrogen oxides (NOx) emissions at one week and 5 km resolution by assimilating surface CO and nitrogen dioxide (NO2) observations. By decoupling analysis steps from ensemble simulations, the modified EnKF avoids filter divergence and enables reuse of costly ensemble simulations, making high-spatiotemporal-resolution inversion affordable. This method is tested by a set of observing system simulation experiments. By assimilating synthetic observations from 400 sites, the errors of CO and NOx emissions over Beijing-Tianjin-Hebei (BTH) in the a priori emission inventory are reduced by 78% and 76%, respectively. Further application of this method estimates the emission reductions during China's Victory day parade using real surface observations. The changes of the emissions in each city are identified by this method, which suggests that the CO and NOx emissions over BTH region during the parade are reduced by 36% and 44%, respectively. Using the inversed emission inventory, the biases of CO and NO2 simulations during the parade are reduced by 95% and 91%, respectively. This highlights the potentials of this method for improving high-spatiotemporal-resolution emission estimation.