The unappreciated threat from inertial instability

On 16 February 2007, the Front Range of the Rockies and the mountain areas experienced strong winds (45 m/s gusts), heavy snow (6-10 inches in the mountains and 3+ inches in Fort Collins), hail, graupel, and lightning. Many of the mountain roads were closed for two days, which would have been one of the busiest ski days of the year. Given that the ski industry loses $800,000 an hour when I-70 is closed on a weekend, this storm had a significant economic/societal impact. Some of the high-impact weather associated with this event occurred from convective bands formed under the northwesterly flow aloft. These bands occurred on the anticyclonic shear side of the jet stream in a region with deep upper-tropospheric negative absolute vorticity, indicating inertial instability. Both human and operational NWP model forecasts failed to produce the convective bands in this case.

Such an event is reminiscent of three other events that have been documented in the literature over the last twenty years. Despite this growing body of knowledge, the threat faced from precipitation bands and clear-air turbulence associated with the release of inertial instability remains largely unaddressed observationally, numerically, or theoretically in the atmospheric sciences. As such, the predictability of these bands is unknown.

There are many remaining challenges to be answered. How often such precipitation bands occur in nature remains unknown. Although diagnosing the existence of these bands after their formation is relatively straightforward using an ingredients-based approach, the dynamics of these bands remains a well-veiled mystery, revealing a substantial challenge for idealized numerical experimentation. Despite this dour scenario for predicting and understanding precipitation bands associated with inertial instability, high-resolution NWP models offer some hope that such events could be adequately forecast in the future. Indeed, high-resolution simulations of three of the four cases indicate that such models have the ability to predict such events.