Ensemble-Based Hurricane State Estimation, Intensity Prediction, and Targeting

点击翻译按钮获取中文摘要

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).While much has been learned about hurricane structure and intensity change from observational analyses and mesoscale model simulations, much less is known about the predictability of these features. Arguably the most important contribution to the well-known lack of improvement in dynamical forecasts of intensity change is the need for dynamically consistent initial conditions. The challenge in creating initial conditions is the assimilation of observations in a hurricane environment, which has large spatial variability. To address this issue, the Principal Investigator (PI) will apply an ensemble approach to hurricane state estimation, predictability, and targeting, with an emphasis on structure and intensity change. The research builds upon recent work by the PI and collaborators using ensemble techniques for data assimilation, predictability, and targeting. Research activities include assimilating observations using an ensemble Kalman filter in a systematic manner, starting with updates of only the axisymmetric structure, and proceeding thereafter by including updates to increasing azimuthal wavenumbers. The predictability time scale of each component will be evaluated using standard error metrics and more recently developed methods based on information theory, including predictable component analysis. Storms to be investigated include: Rita, Katrina, and Ophelia from the Rainband and Intensity Experiment (RAINEX); Emily from the Tropical Cloud Systems and Processes (TCSP) experiment; and Bonnie (1998). Storms will be analyzed using two experiment types: (1) actual observations, and (2) simulated observations drawn from truth simulations constrained to approximately follow the track and intensity of the actual storms. Ensemble-based sensitivity analysis will be used to assess the impact of observations, including targeting, through observation denial in the real-data cases. Intellectual merit: The research plan builds a bridge between observational studies, which have established properties of hurricane structure and intensity change but lack dynamical continuity, and theoretical and modeling studies that have the spatial and temporal continuity to assess the dynamics of these features, but lack a close link to the observations of real storms. The assimilated datasets that result from the research offer the spatial and temporal resolution of numerical simulations, but should also remain faithful to available observations. These data will be used to systematically assess the predictability of storm structure and intensity, which is a crucial aspect of hurricane dynamics, for which little is known. Broader impact: The broader impact to society of improved forecasts of hurricane intensity is well known. Given a successful execution of the research, a transition to operations seems feasible. Moreover, demonstrated ability to target observations to routinely improve predictions of hurricane intensity would have a transformational impact on operational hurricane forecasting and thus also on emergency planning for these storms.