Diagnostic analysis of short-term hurricane vortex evolution in high-resolution ensemble Kalman Filter data assimilation

The Hurricane Ensemble Data Assimilation System (HEDAS) , based on the Ensemble Square Root Filter (ESRF) approach, was developed at the Hurricane Research Division (HRD) of NOAA using an experimental version of the Hurricane Weather and Research Forecast model (HWRFx). The main focus of research with HEDAS is to study improvements to the initial representation of the hurricane vortex by assimilation of airborne Doppler radar, dropwindsonde, and Stepped Frequency Microwave Radiometer surface wind observations, collected during NOAA's Hurricane Field Program. HEDAS makes use of global analysis data from a 30-member ensemble obtained from NOAA/ESRL's ensemble Kalman filter (EnKF) global system and performs the assimilation on a 3-km nest centered on the hurricane vortex. As part of NOAA's Hurricane Forecast Improvement Program (HFIP), HEDAS will be run in a semi-operational mode during the 2010 Atlantic hurricane season. The results from the semi-operational implementation will be presented by Sellwood et al., at the 15-th Conference on Integrated Observing and Assimilation Systems for Atmosphere, Oceans, and Land Surface (IOAS-AOLS, AMS Annual Meeting, 2011).

The results of case studies to date indicate considerable skill of the high-resolution analysis from HEDAS in terms of intensity and spatial patterns of circulation at vortex scales when compared to observation-based wind and surface pressure analysis. In the case of Hurricane Bill (2009), however, the deterministic forecasts from the high-resolution HEDAS analyses exhibited significant vortex spin-down during first 12 hours, unlike the observed storm. The track forecasts were consistent with the best-track estimates. A similar tendency for the vortex spin-down was also identified in 1-hour long, forecasts during cycling in HEDAS. This tendency was, as expected, systematically corrected by the observations in the analysis. To better understand the evolution of the modeled vortex and the associated errors in the analysis and in relation to the subsequent forecast errors, in the current study relationships between the primary and secondary circulation and thermodymamical fields were evaluated using the ensemble data from the analysis as well as the results of the subsequent deterministic and ensemble forecasts. The diagnostics include the ensemble-based multi-variate error correlations and wind-mass balance and principal wave analysis in a vortex-relative reference frame.