Mesoscale precursors to the Hurricane Gaston flooding event as diagnosed from observations and numerical simulations

Freshwater flooding has been shown to be a leading cause of loss of life in tropical cyclones (TCs) accounting for over half of the TC related deaths in the United States between 1970 and 1999 (Rappaport 2000). Recently, a greater understanding of the extratropical transition process (ET) has lead to better forecasts of TC flooding events. Comprehensive research on ET by Evans and Hart (2001, 2003; Hart 2003), Klein et al. (2000, 2002), Harr and Elsberry (2000), Atallah and Bosart (2003, 2007) and numerous others has allowed forecasters to apply their understanding of quasi-geostrophic (QG) theory to more accurately predict where heavy rainfall will occur. However, rainfall forecasts have remained problematic for storms that do not undergo ET or remain too small for QG theory to apply. This study will use Hurricane Gaston (2004) to investigate the dynamics that drive heavy precipitation in these “non-traditional” storms.

Hurricane Gaston produced 10-12 inches of rain and flash flooding across Richmond, Virginia on 30 August, 2004 while remaining warm core and highly compact. Numerical models and human forecasters struggled with the mesoscale nature of the storm and heavily precipitating convective feature. Early study of observational data using the Rapid Update Cycle (RUC) analysis revealed an area of extremely high convective available potential energy (CAPE) and dew points just north of Gaston in southeast Virginia on the morning of 30 August. It was determined that a clearing in the cloud field allowed solar radiation to warm the surface in the morning hours, causing a rapid increase in low level theta-e. Short-lived supercells quickly formed along a convergent band rotating into southeast Virginia.

While the observational study provided evidence of low-level ingredients that supported the initial convection, the upper-level maintenance mechanisms were more difficult to diagnose. Intense mesoscale ridging with areas of inertial instability was indicated in the RUC upper level analysis, but results were inconsistent and inconclusive. To explore the dynamics below the resolvable scale of the observations, the Nonhydrostatic Mesoscale Atmospheric Simulation System (NHMASS) model was utilized to simulate key mesoscale features. Nested domain simulations of 18, 6, and 2 km all verified well with the observed low-level features once an acceptable initialization dataset was found.

A persistent area of inertial instability in a sharply curved mesoscale ridge was found at the convective outflow level in the 6 and 2 km simulations. The anticyclonic shear on the equatorward side of the polar jet-streak also contributed to lowering the inertial stability north of the convection, providing a preferred channel for outflow mass evacuation. This fits the conceptual framework for where inertial instability can be found in the mid-latitudes and helps set up a deep tropospheric mesoscale circulation not unlike those found in mesoscale convective complexes (MCCs). This type of system, where both conditional and inertial instability are present, is referred to as convective symmetric instability and is responsible for the long duration of the precipitation over Richmond, Virginia. It is theorized that convective symmetric instability could be common in TC landfall events on the East Coast, helping to produce self-sustaining ageostrophic secondary circulations that extend the life-span of heavily precipitation convective systems.