Analog-Based Analysis of Non-Convective High Wind Events in Mid Mississippi River Valley

High-impact winds that occur outside of thunderstorm activity, non-convective wind events, are often overlooked, understudied, and dangerous. More fatalities occur from non-convective wind events than straight-line winds, and hurricanes. In addition to the many fatalities, a great number of power outages and property damage result from non-convective wind events. This study seeks to understand the mechanisms that produce non-convective wind events, and to provide forecasters with the tools to more accurately forecast the events.

To do so, data was gathered from the National Center for Environmental Information (NCEI) Automated Surface Observation System (ASOS) archive for the period of 1979-2016. Five observation stations in the Mid Mississippi River Valley were used to identify non-convective wind events. The event selection was based on the National Weather Service criteria for High Wind Warnings and Wind Advisories.

The event data was parsed for diurnal, monthly, event period, and wind direction patterns. In order to gain an understanding of the synoptic environment of non-convective wind events, composites of meteorological fields at each station were generated using the North American Regional Reanalysis (NARR) dataset. These composites included 500 hPa and 700 hPa geopotential heights and vorticity, 850 hPa and 925 hPa geopotential heights and isotachs, 850 hPa and 925 hPa geopotential heights, temperature, and temperature advection, mean sea level pressure, potential vorticity at various pressure levels, mid-level and low-level lapse rates, and cross sections of potential vorticity at each station.

In order to improve the composites results and provide tools to aid forecasters in creating probabilistic forecasts of non-convective wind events, the composites were subjected to a field score analysis of the events at each station. Using Saint Louis University’s Cooperative Institute for Precipitation Systems (CIPS) analog based system, individual cases that make up the composites were compared to previous analogs. If the cases met a certain field score for similarity, they remained in the study. New composites were made from the remaining cases.

The field score found from the above analysis was applied to the entire NARR dataset to determine if there were days that met this criteria but did not produce high winds, non-events or misses. Cases that met the field score criteria and produced non-convective wind events were considered events or hits. Misses were cases that met the field score criteria and did not produce a non-convective wind event. The hits and misses were further investigated by producing a contingency table to determine how reliable this field score criteria is for anticipating non-convective wind events. Misses were also analyzed by generating composites of the same fields as non-convective wind events. This comparison of misses to hits resulted in the identification of essential mechanisms for producing non-convective wind events.