Evaluation of high-resolution surface analyses and forecasts with ensemble data assimilation in regions of complex terrain

Accurate, high-resolution atmospheric surface analyses and short-range forecasts are important from many reasons. Our recent studies suggest that the errors in atmospheric surface analysis and forecast are flow-dependent. Because of its flow-dependent background error term, ensemble Kalman filter can overcome limitations that were found in 3-dimensional variational assimilation of near-surface observations over complex terrain. In this study, we performed data assimilation and numerical prediction experiments over the Intermountain West region for a one-month period with an advanced research version of Weather Research and Forecasting (WRF) model and its 3-dimensional variational data assimilation (3DVAR) systems, as well as an ensemble Kalman filtering system developed by NCAR Data Assimilation Research Testbed (DART/WRF). The hourly surface mesonet observations, along with available conventional data are assimilated into WRF model with horizontal resolution at 10 km, 3.3km and 1.1km. Results show surface data assimilation improves the predictability of near surface atmospheric conditions, specifically in short-range forecasts. However, for weak- and strong-synoptic forcing scenarios, errors in surface analyses and forecasts behave differently. Analyses and forecast results are also sensitive to the model resolution and various configurations of data assimilation. For instance, simulations at high-resolution perform better than those at coarser resolution. Specifications of terrain heights and land use parameters in the WRF model also have a significant impact on the predictability of near surface atmospheric conditions. The performance of ensemble data assimilation at high-resolution is evaluated and compared with 3DVAR. Results will be presented during the conference.