An application of potential vorticity inversion to improving the numerical prediction of the March 1993 superstorm

Abstract

In this study, a methodology is proposed to improve the model initial conditions, based on available surface temperature observations from ships, buoys, and drifters. It is tested with the numerical prediction of the 12-14 March 1993 superstorm that is initialized at its incipient stage over the Gulf of Mexico. In this methodology, the authors make use of the piecewise potential vorticity (PV) inversion technique and treat the temperature errors at the lowest level as a surrogate PV anomaly. After inverting the wind and mass perturbations from the surface thermal anomaly and its pertinent interior PV anomaly, a three-dimensional, dynamically consistent set of "errors" are obtained and added to the model initial conditions to improve the representation of the lower troposphere over the data-sparse ocean. It is found that the numerical model prediction, initialized with the modified initial conditions, exhibits significant improvements in the early rapid deepening and the track of the superstorm over ocean, the development of a prefrontal squall line, and the central sea level pressure traces during the life cycle of the cyclone, as verified against observations. These results show that the methodology proposed is promising in improving the representation of lower-tropospheric meteorological variables in the model initial conditions, based on available surface observations over data-sparse regions.