Wind Characteristics of Landfalling Hurricanes Measured by Texas Tech University StickNets

Prior controversy among the meteorological and engineering communities concerning differences in the interpretation of wind speed measurements within land-falling hurricanes, most notably Hurricane Andrew (1992), has led to the development of techniques enabling comparison between diverse wind measurements as well as establishing a framework for future measurement deployments. Three decades later, Hurricane Ian (2022) highlighted the need for high-quality, high-resolution measurements due to a void in reliable surface observations in the landfall region. These voids directly impact immediate post-storm operations such as the coordination of relief supplies, proper alignment of first responders and utility workers, and the readiness of response operations such as the Red Cross. Other post-storm functions are also impacted by a lack of wind measurements, such as the assessment of wind loss by the insurance and reinsurance industry. Long term impacts of unresolved wind fields also include the connection of wind damage to local building codes, inhibiting societies growth towards more tropical cyclone resistant communities. There remains a tremendous need to fill these measurement gaps while providing real-time measurements to the impacted communities to further our understanding of the risk that life and property may endure in landfalling tropical cyclones.

This study aims to bridge the gap between the meteorological and engineering communities by using the largest database of surface-based tropical cyclone measurements at landfall to date leveraging 10 historical Texas Tech University Hurricane Research Team (TTUHRT) deployments since the 2008 Atlantic hurricane season. The presented wind measurements were made using a fleet of up to 48 rapidly deployable measurement stations, called StickNet, which records high-resolution (1-10 Hz) wind speed, direction, and pressure measurements documenting the surface wind field of record. Previous studies have shown that the comparison of nearby wind speed measurements may present large errors when disregarding fundamental measurement differences. To combat these differences, a standardization procedure was used to remove the influences of averaging time, measurement height, and surrounding surface exposure. The raw and standardized wind speeds were inserted into a database along with several other derived parameters important for both meteorological and engineering studies. Among the most important parameters influencing the wind speed measurements is the upstream surface roughness. Stratification of measurements by their upstream roughness reveals secondary relationships within the measurements. Further, the incorporation of corresponding WSR-88D radar measurements provides more information about each tropical cyclone, its location, and boundary layer structure. Coupling of the two measurement systems allows for comparison between the aloft and surface wind fields. Understanding these comparisons may allow for future investigations linking real-time meteorological observations into modeling the structural risk a landfalling tropical cyclone wind field may pose on various coastal regions.

In this current study, each measurement underwent quality control and is standardized to a previously defined framework. Given the sensitivity of the roughness length on the calculated standardized wind speed, two techniques were used for evaluation. The first method calculates the raw 10-minute roughness length directly from the time segment of interest. This method allows for the inclusion of extreme roughness length values. A second, new, approach was created to leverage the power of averaging and preserve a directionality-based approach. This “sector” standardization method stratified and then averaged the previous surface roughness measurements in 30 directional bins. Although this method eliminates the extreme values through averaging, it gives a more representative roughness length based on the wind direction for each platform. The wind speeds are standardized with both surface roughness evaluation methods. Cyclone center location was found by linearly interpolating between NHC Best Track data points, resulting in a 10-minute center location estimate. This center location and resulting center motion provided a storm relative position for each platform. Relevant data from National Weather Service WSR-88Ds were gridded to constant horizontal and vertical planes. An adaptation of the Fuzzy Logic Method was used to identify regions of convection or stratiform precipitation for each time segment in the database. The start of a new procedure to find the inner and outer location of the eyewall based on radar reflectivity was also created with the future goal to automate the cyclone center location, eyewall location, and structure characteristics in real time. With the inclusion of StickNet measurements, Best Track data, radar data, and the calculated parameters the largest database to date results with over 55,000 unique ten-minute average segments and over 190 parameters per segment.