A diagnostic study of non-convective high winds in the 12-13 November 2003 Great Lakes cyclone

On 12-13 November 2003 an intensifying mid-latitude cyclone tracked across the Midwest and Great Lakes regions. Non-convective high winds up to 76 knots caused $36 million in damage, eight deaths, and 23 injuries from Iowa to Pennsylvania as well as a major seiche on Lake Erie. The most outstanding destruction from this storm occurred in Lower Michigan where a utility company called it the worst storm since the Edmund Fitzgerald storm in 1975.

Despite the extent of the wind damage, there was little consensus among forecasters at the regional National Weather Service offices as to the origin of the high winds or the mechanisms responsible for them. Storm event summaries from forecast offices across the Midwest and Great Lakes cited the following possibilities: a fast-moving cold front, strong cold-air advection, isallobaric winds, and a deep tropopause fold. To help mitigate the hazards associated with non-convective windstorms, it is necessary to develop a comprehensive, multi-scale conceptual model of these events.

This project provides an account of the synoptic and subsynoptic dynamical features that likely contributed to the non-convective high winds across the Midwest and Great Lakes regions during the 12-13 November 2003 cyclone. Geographically, we focus on conditions over Lower Michigan, Lake Erie, and northern Ohio where more than one-half of the storm-total damage was reported and wave-height displacement on Lake Erie reached 14 feet. The pattern in the surface wind field over the region was found to differ from that observed over the Midwest and other parts of the Great Lakes. Diagnostic analyses suggest that surface gusts over the region of greatest damage were connected with high-momentum air descending from the tropopause. The ability of air at the tropopause to descend through the boundary layer is likely maximized where the mechanical turbulence at the base of the tropopause fold aligns with the convective turbulence at the top of the boundary layer. This is shown in the present storm to occur over the maximum in regional boundary layer heights, which is a main new result of this research.