Assimilation of surface-based profiler observations using an observation system simulation experiment: Part 2: forecast impact

Part 1 of this study (Otkin et al. 2010) described the impact of a potential suite of surface-based remote-sensing profiling systems on the analysis accuracy during an Observation System Simulation Experiment (OSSE) over the continental U.S. The assimilation experiments evaluated the impact of four profiling systems: Doppler wind lidar (DWL), Raman lidar (RL), microwave radiometer (MWR), and Atmospheric Emitted Radiance Interferometer (AERI). In this study, model forecasts created using the final analyses from 7 assimilation experiments in which different combinations of profiler observations were assimilated are compared to data from a high-resolution “truth” simulation to determine the impact of these observations on the accumulated precipitation and low-level moisture flux convergence. This study uses the Method for Object-Based Diagnostic Evaluation (MODE) package in the Model Evaluation Toolbox (MET) from the Developmental Testbed Center to examine matched pairs of accumulated 6-hr precipitation objects between the “truth” simulation and assimilation forecasts. Overall, assimilating temperature and water vapor profiles from the RL, MWR, or AERI in combination with conventional observations (CONV) tends to improve the location of the accumulated precipitation field. Wind profiles from the DWL in combination with CONV observations, however, produce a better intensity forecast but not necessarily an improvement in the location and timing of the precipitation. When assimilating both mass (RL, MWR, AERI) and velocity (DWL) observations, the results indicate that a substantially better forecast for both the location and intensity of the 6-hr accumulated precipitation field is achieved. Similar behavior occurs in the 850 hPa moisture flux convergence field, which suggests that assimilating a combination of mass and velocity profiler observations may improve precipitation forecasts.