Numerical modeling of Atlantic hurricanes moving into the middle latitudes

Hurricanes that form over the Atlantic Ocean very frequently migrate into the middle latitudes where they encounter much different oceanic and atmospheric conditions than in the tropics. Cool sea surface temperatures (SSTs) cause these storms to weaken and become thermodynamically decoupled from the ocean, while baroclinic atmospheric environments often cause them to reintensify into extratropical storms „Ÿ a process known as extratropical transition (ET). The changing structure of these storms in the middle latitudes presents many unique forecasting challenges related to the increasing asymmetry in moisture and wind fields.

An examination of three such events over Eastern Canada „Ÿ using a combination of observations and a numerical model „Ÿ forms the foundation of this work, with an emphasis on applying the research to weather forecast applications. The case studies include Hurricane Michael (2000), Hurricane Karen (2001) and Hurricane Juan (2003). Hurricane Michael intensified in a strongly baroclinc environment and evolved into an intense extratropical storm over Newfoundland. Karen also underwent ET, but weakened quickly while approaching Nova Scotia, while Hurricane Juan struck Nova Scotia as a category-two hurricane, experiencing only marginal weakening over anomalously warm SSTs. In essence, these cases represent a cross section of the behavior of many tropical cyclones in this part of the world.

Hindcast simulations are conducted for each event using the Canadian Mesoscale Compressible Community (MC2) model with a synthetic, observationally-consistent hurricane vortex used in the model's initial conditions. Sensitivity experiments are run for each case by modifying initial specifications of the vortex, model physics parameterizations, and surface boundary conditions like SST. For Hurricane Juan, it was determined that the anomalously-warm SSTs played a significant role in the landfall intensity. The model is also run in "forecast mode" to test its suitability as a prognostic tool for meteorologists. Significant improvements in the representation of the moisture and mass fields is observed in all three cases compared with numerical forecast models that did not employ vortex insertion in the initial atmospheric fields.