Evaluating the Viability of Surface Marks to Estimate Tornado Winds in Rural Areas

Tornadoes cause significant damage, injuries and fatalities worldwide. Tornado-based design for structures was introduced in ASCE 7-22, and design for tornado loading will continue to become more common in tornado prone regions and for certain structures in the United States. One parameter used for tornado loading design is wind speed for a particular return period. Currently, tornado wind speeds are estimated from the enhanced Fujita Scale which is a subjective scale that measures the wind speed that is needed to cause a certain degree of damage to a structure. However, when a tornado does not impact a structure, when a structure’s building quality is poor, or during strong tornadoes the estimated wind speed is hard to obtain. For this reason, methods like the treefall analysis have been developed and used. However, in strong and violent tornadoes and in cases where trees are scarce (e.g. cropland), a wind speed estimate can not be accurately obtained. In some tornado cases, surface marks are left behind due to debris deposit or suction (Fig. 1). Using a method introduced by Van Tassel and Prosser, Fujita used the surface marks to estimate the wind speed of a 1965 Palm Sunday tornado.

The method assumes that the marks are related to the rotational and translational movement of the tornado, and thus wind speed estimates can be calculated through measurements of the cycloidal surface mark. Using aerial imagery, the width (w) and the height (2R) of the cycloidal surface mark is measured and a Gmax (n) is calculated by solving for n in equation 1. GMax is a ratio of tangential velocity and translational velocity. Once the GMax is calculated, the tangential speed V can be calculated as the product of GMax and the translational speed of the tornado. The maximum wind speed estimate based on a single surface mark is the sum of the translational and tangential speed.

Equation 1:

w/(2R)=n^-1((n²-1)^1/2-π/2+sin^-1(n^-1))

This method was used to estimate the wind speeds of the 2013 Washington, IL tornado which has a peak wind speed of 190 mph as determined by the enhanced Fujita scale in the damage survey. Areas of the track where cycloidal surface marks were split up into 10 sections. In each section, the loop dimensions were measured using ArcGIS, and a windspeed was calculated. An average of all the wind speeds for each individual loop was calculated to obtain the estimated wind speed for that section of the track. From the cycloidal surface marks, the Washington tornado had an average wind speed estimate from as low as 184 mph to as high as 219 mph across multiple areas of its path. The high wind area had individual surface mark wind speeds from 205 mph to 231 mph in addition to limitations with the accuracy of loop width and height measurements. Estimates of wind speed were then compared with wind speed estimates using the EF scale from nearby structures. The peak wind of 219 mph occurred immediately after Washington, IL and a residential neighborhood. The damage survey of nearby structures for this section resulted in a wind speed estimate of 190 mph. However, the estimates from the damage survey further down the track were significantly less than the estimates from the surface mark method. This can possibly be due to many reasons including the limited number of structures that were impacted in the rural area.

The results from the Washington, IL tornado show realistic wind speed estimates given the constraints of the enhanced Fujita scale, and a more thorough study should be done. Many questions remain regarding the source (suction or debris deposit) and frequency of these cycloidal surface marks and how the season of occurrence and soil properties influence the shape and appearance of the surface marks. As time allows more of these questions will be addressed and the viability of this method will be further explored by comparing wind speed estimates from the treefall method to the surface mark method in the 2021 Western Kentucky tornadoes.