Ensemble Kalman Filter Assimilation of Multisensor Observations from Convection for Storm-Scale Analysis

点击翻译按钮获取中文摘要

Observational capabilities on the storm scale have increased dramatically during the past decade. The national network of operational Doppler radars provides observations of convective storms and their environments every five minutes. In the near future, rapidly scanning phased array radars, dual polarization data, and regional high-density networks of radars could also be available. Surface data from local surface mesonets have also become increasingly available. The availability of these data will spur the development of new retrieval and data assimilation methods for storm scale phenomena.Observations of Doppler velocity and reflectivity are not often available near the ground. At the same time the convective cold pool and storm wind field plays a critical role in determining significant surface weather, convective initiation, and storm rotation. Therefore retrieval of the wind, temperature, and moisture fields near the surface is both difficult and crucial toward increasing understanding of storm dynamics and improving forecasts and warnings.Recently the ensemble Kalman filter (EnKF) has been shown to be a viable method for obtaining convective-scale atmospheric state estimates from Doppler radar observations. Assimilation experiments using synthetic data sampled from convective storm models indicate that this method can be used to retrieve the unobserved fields within storms. Initial experiments using observed Doppler winds indicate that EnKF has potential for real data as well. An attractive feature of the EnKF is that it is a relatively simple methodology and can be implemented into a forecast model with little effort. The major goal of this research is to assess the potential of EnKF in assimilating real observations of convective storms through the use of a series of observing system simulation experiments. Unique aspects of this work include the use of an ultra-high-resolution numerical cloud model to produce synthetic observations as well as considering the role of model error associated with the microphysical parameterization. Both radar and in situ surface observations will be generated synthetically. The methodology will be to use multiple Doppler radars and surface-based observations to retrieve wind, temperature, and hydrometeor fields at low levels. Retrieval of low-level fields are particularly important toward understanding storm dynamics, since these features play an important role in determining storm evolution and severity.The research increases intellectual knowledge in both the theoretical and applied meteorological communities. The development of robust methods for estimating near-surface wind, temperature, and moisture fields within storms from various observational platforms will greatly enhance theoretical understanding of storm dynamics within more realistic environments. Research results can then be applied in the operational community to enable regional forecasters to focus on smaller regions where the conditions are most conducive for severe storm development. Increased knowledge of storm scale dynamics will also help local forecasters improve the warning process by increasing lead times and reducing false alarms.