Using simulated tornado surface marks to help decipher near-ground wind fields

Numerical simulations of tornadoes show a rich complexity near the surface. The highest radial, azimuthal and sometimes vertical wind speeds generally occur within the lowest tens of meters above the ground. The tornado's structure and intensity proves to be strongly sensitive to properties of the near-surface inflow and can change rapidly in time for some conditions. Studying these effects in the field is problematic because mobile Doppler radar, which provides the best tornado wind field measurements, cannot probe below about ~20 m above ground. There is, however, a direct signature of the low-level wind field that is often available: the "surface marks" or "debris tracks" left behind by the tornado. A handful of studies 30-40 years ago attempted to measure tornado structure and intensity through deciphering tornado ground marks, but little scientific attention has been paid to them since. In the present work we revisit this field using high-resolution numerical large-eddy simulations of tornadoes with interacting debris, including an accounting on the surface of where soil is removed or deposited, to produce simulated tornado tracks together with the complete wind fields and debris clouds responsible for generating them. We will present results from a large series of tornado simulations with both single and multiple debris species, different tornado types, translation velocities, soil and surface properties, and both quasi-steady and evolving time histories. We generate tracks of simulated debris removal and deposition as well as the state of orientable debris (such as idealized "corn stubble" attached to the surface). For comparison we also collect simulated ground traces of wind field variables such as peak wind speeds or pressure drops encountered at each surface point.

In both observed and simulated tornado tracks a wealth of widely varying ground marks are encountered. Their interpretation is complicated because they are produced by a convolution of events in time and space. Some features prove robust while others appear sensitive to details that are likely unavailable in the field. In this work we address the physical origins of different types of surface marks, with the aim of deciphering what observed marks can tell us about the tornado structure, wind fields and time development that generated them, particularly if Doppler observations of conditions aloft are available as well. The simulations also allow us to reassess the successes and failures of interpretations of tornado surface marks that appear in the older literature.