Improve Ensemble-Based State Estimation and Forecasting with Simultaneous Parameter Estimation

Incorrect parameter values used in PBL schemes are a likely cause for their systematic errors. Model errors reduce the accuracy of state estimation and subsequent forecasting. The ensemble Kalman filter (EnKF) could account for model errors by estimating the optimal flow dependent parameters when it estimate regular model state, thus could improve the state estimation and the subsequent forecasting. In this study EnKF is used to estimate the flow dependent optimal values of two fundamental parameters in the ACM2 PBL scheme in the Weather Research and Forecast (WRF) model simultaneously as it estimates the regular model state through the state vector augmentation method. The simulation is conducted for the period of 1200 UTC August 29, 2006 to 0000 UTC September 3, 2006 over Texas. The parameter estimation EnKF is compared with WRF deterministic forecasting and regular EnKF (i.e., no parameter estimation EnKF). Cold drift appears in the WRF deterministic forecasting, which is suspected to be partially caused by simulated too strong northerly wind. Parameter-estimation EnKF is shown to help eliminate the cold bias from WRF forecasting and provides better performance than no parameter-estimation EnKF by providing the best wind fields and more optimized flow dependent parameters.