Tornadoes and debris

The debris trajectory animations of Figures 6 to 11 were provided by Professor Mark Sterling, whose ability to use advanced EXCEL functions seems to be significantly greater than mine. His contribution is much appreciated.

Previous work

In 2017 Mark Sterling and I published the paper “Modelling wind field and debris flight in tornadoes”, which described the integration of a tornado wind field model and the debris flight equations to look at the pattern of compact debris movement in tornadoes of different types. Typical results for falling and flying debris are shown in figure 1 below and give an indication of the complexity of the debris trajectories that were predicted.

Now whilst the tornado wind model that was used in the analysis was a considerable improvement over those that existed at the time, in that it gave a consistent three dimensional velocity formulation, it did however have one major drawback. This was the fact that the vertical velocity component was unbound and increased with height, albeit quite slowly. In a more recent paper in 2020 “The lodging of crops by tornadoes”, we developed an improved model, in which the vertical velocity peaked at a certain height and then decreased at greater heights. In this blog post I will briefly explore  the use of this wind model to predict compact debris flight paths using the same methodology as in the first paper, and in doing so will illustrate the importance of the tornado model on debris trajectory prediction.

The tornado wind model

The expressions for the radial, circumferential and vertical velocities in the 2020 model are given in figure 2. Here the velocities are normalized by the maximum circumferential velocity and the radial and vertical distances by the radius at which the maximum velocity occurs. Note that this is different from the 2017 paper where the maximum radial velocity was used for normalization. The parameter K is related to what will be termed the swirl ratio S (the ratio of the maximum circumferential to maximum radial velocity) by a function of the parameter gamma, which is a shape parameter that affects the shape of the radial and vertical profiles. (Unfortunately this web template doesn’t support Greek letters, so I have to spell them out). Figure 3 shows typical velocity profiles for different values of this parameter.  It can be seen that for gamma = 2, the peak of the vertical velocity is at the vortex centre, as in a typical single cell vortex, whilst for higher values it moves away from the centre becoming more like a two cell vortex (but note there is no downflow at the vortex centre in this case.

• 

• 

Debris flight equations

The equations for compact debris flight are given in figure 4. These are the same as in the 2017 paper, although expressed a little differently. The debris velocities (lowercase) in the three directions are again normalized by the maximum tangential tornado velocity. Two dimensionless parameter are identified – the Tachikawa number Ta that relates the flow force on the debris particle to its weight, and a tornado Froude number Fr. Different dimensionless parameters were used in the 2017 paper, because of the different reference velocity that was used

Solutions

Putting together the velocity equations in figure 2 and the particle flight equations in figure 4, it can be seen that there are four parameter that define debris trajectories – the tornado parameters S, gamma and Fr, and the debris Tachikawa number Ta. In addition any one flight trajectory will be defined, at least in its early stages by the dimensionless values of the radius and height at its release point. If these six parameters are specified then the equations of debris flight can be solved in a straightforward manner.  In what follows we define a base case situation as in figure 5, and then vary each of the parameters around this base case value. We present the results in the animations of figures 6 to 11.Each animation shows four plots – the trajectories projected onto a vertical plane through the tornado centre; the trajectories projected onto a horizontal plane; the trajectories in a rotating plane in the radial and vertical directions, and a plot of the variation of particle kinetic energy with time. The latter acts as a damage indicator of debris flight, but also clearly shows whether or not the solution converges or diverges with time. Note that the dimensionless time shown in the kinetic energy plots is proportional to the time of revolution of the vortex – a time of 2 pi corresponds to one vortex revolution. 

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

Figure 6. Effect of variations in Tachikawa number

First consider the effect of changing Tachikawa number, Ta – see Figure 6. This represents changes in the nature of the debris. A low value of Ta represents heavy debris and vice versa. It can be seen that at low values of Ta, the debris tracks can reach significant heights and the debris undergoes a diverging motion when viewed in the radius / height plane, with a diverging kinetic energy oscillation. At some point in the trajectory the debris hits the ground and the energy falls to zero. The base case situation at Ta = 100 is still mildly diverging but the trajectory does not intersect the ground plane for the length of the calculation. As Ta increases further, the debris takes up a stable path in the radius / height plane travels around a small circular trajectory, with the kinetic energy converging to a stable value. This suggest that light debris can reach an equilibrium where it is held aloft by the tornado. The position around which the circular motion takes place is around a normalized radius of 1.3 and a normalized height of 0.9. The value of height is much less than calculated in the 2017 paper, reflecting the fact that the vertical velocity does not decrease indefinitely with height for the new model as it did in the old.

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

Figure 7. Effect of variations in Froude number

The effect of variations in Froude number is shown in Figure 7. The primary effect that increase in Fr has is to increase the centrifugal force on the debris. At low values, the trajectories are stable and similar to that of the base case. As the values increase above 1.0 the oscillations become larger due to the increased centrifugal forces and eventually become unstable, with the trajectories meeting the ground at high values.

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

Figure 8. Effect of variation in Swirl Ratio

The effects of variations in the Swirl ratio shown in Figure 8 are complex, with diverging trajectories (and ground impact) at both low and high values, and a region of stable trajectories between values of around 1.0 to 1.9. At low values the trajectories are destabilized by the high values of radial velocity, and at high values are destabilized by high values of the circumferential velocity.

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

Figure 9 Effect of variations in gamma

The change in values of gamma from the one cell form of gamma = 2 to the quasi-two cell form of gamma = 4 shown in Figure 9 results in little change to the debris trajectories from the base case, although the oscillations in the kinetic energy fall as gamma increases.

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

Figure 10. Effect of variations in radial starting position

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

Figure 11. Effect of variations in vertical starting position

The debris trajectories remain stable as the normalized radius varies between 1 and 1.9 but outside those limits the trajectories diverge and intersect with the ground (Figure 10). Similarly the trajectories are only stable for normalized values for height between 0.8 and 1.2 (Figure 11). Thus the starting point window for the trajectories to ultimately attain a stable form is quite small.

Concluding remarks

A number of points arise from the results presented above.

• Even for the simple wind and debris flight formulation adopted, debris trajectories can be quite complex.
• A comparison of the results obtained with the old and the new wind field model show very considerable differences, due to the different vertical velocity formulation. analysis reveals that the debris trajectories can be specified by a small number of debris and tornado parameters, with the Tachikawa number and the Swirl Ratio being the most significant.
• There are regions within parameter space for which the debris trajectories become stable – i.e. the debris flies indefinitely.