I have extensive research expertise in investigating how terrain (the vertical variation of landforms) and land cover (such as forest, grassland, urban areas, water, etc.) impact tornadoes. More specifically, I am interested in how the wind field and intensity of tornadoes as well as the path deviate as a direct response to variations in terrain or land cover. My research team and I have used single- and dual-Doppler radar observations from a rapid-scanning (360 degree sampling in 2 seconds) mobile radar, coupled with Digital Elevation Model and the National Landcover Database within a GIS to investigate the effects of terrain and land cover on the strength, structure, genesis, and decay of observed atmospheric tornadoes (Butler 2017; Houser 2020; Muncy 2021; Price 2023). I use statistical analyses to test for relationships between these various factors. This work was funded under National Science Foundation grants AGS-1749504 and AGS-2244884.
The primary conclusions from these studies, when taken in aggregate are: 1) Tornadoes tend to have weaker vertical motions vorticity and rotational velocity at high elevations relative to low elevations – In other words, tornadoes are generally stronger in localized valleys than on top of nearby hills (Figures 1, 2). 2) Tornado winds tend to diverge at higher local elevations relative to the nearby mean elevation, and they tend to converge at lower elevation (Figure 3); 3) Tornadoes tend to curve to the left going uphill and to the right going downhill, which also agrees with other studies (Figure 4); 4) Larger tornadoes are affected by elevation changes more than smaller tornadoes (Figure 5); 5) Certain locations with prominent terrain features experience more – or less – tornadoes than other locations; 6) Vorticity, vertical velocity and convergence are stronger when tornadoes pass over land cover with higher surface roughness values; 7) Vortex structure often changes as tornadoes pass over gradients in surface roughness; and 8) smaller tornadoes are more prone to circulation structure changes when encountering surface roughness gradients than large tornadoes.
Simply stated, both terrain and land cover affect tornado properties.
However, despite these generalized conclusions, there are many individual cases where the behavior of a specific tornado is different than, or even opposite to the results stated above. These differences are likely due to the notable limitations of the studies, which include: i) an inability to sample near the surface (z < 100 m) wind field; ii) variable heights of the lowest available radar observations; iii) an inability to parse the contributions of storm-scale features and environmental heterogeneities; and iv) factors associated with the terrain or surface roughness that were not accounted for such as approach angle, variability of terrain and ground cover within the area of intersection of the vortex. Because most of these limitations cannot be overcome using observations alone, numerical simulations are required to gain high-fidelity representations of tornadoes interacting with terrain and land cover, and to isolate the effects that these physical characteristics have on the tornado from other non-surface related factors.
My next project will be coupling observations with a high resolution numerical model where we (the research team) will systematically evaluate changes in the terrain and/or land cover with the wind speeds (vertical and horizontal), minimum central pressure, and path of both idealized (simple numerical representations of swirling winds) and observed tornadoes.
-
- Figure 1: Relationship between terrain height (meters, right y axis, color shade) and tornado intensity (velocity differential [meters/second], left y axis, red line) over time for the 24 May, 2011 Lookeba, OK tornado based upon RaXPol single Doppler data. (From Houser et al 2020: Fig. 7a.)
-
- Figure 2: Histogram displaying the frequency of occurrence (y axis) and range of 1000 bootstrapped replicate sample means for tornado rotation (vorticity in [s-1], x axis) for higher than average terrain elevation (red) vs lower elevation (blue) for 7 tornadoes analyzed in Price 2023. The trend for the blue on the right hand side implies stronger vorticity (rotation) at lower elevations. (From Price 2023: Figure 5.24.)
-
- Figure 3: Histogram displaying the frequency of occurrence (y axis) and range of 1000 bootstrapped replicate sample means for convergence (x axis) for higher than average terrain elevation (red) vs lower than average terrain elevation (blue) for the same 7 tornadoes as shown in Figure 2. Based off of dual-Doppler analyses generated from radar observations. (From Price 2023: Figure 5.31.)
-
- Figure 4: Change in direction (bearing, in compass degrees) as tornadoes move uphill (blue) vs downhill (red) for the cumulative results of 10 tornadoes. The clear separation between the red and blue indicates that tornadoes nearly always turn to the left (- bearing change) when moving uphill and turn to the right (+ bearing change) when moving downhill.
-
- Figure 5: Comparison of the effects that changing elevation has on the rotation of large tornadoes (left) vs small tornadoes (right). Large = radius > 250 m; small =radius < 250 m. a: As large tornadoes moved uphill (red), they STRENGTHENED. Similarly, as large tornadoes moved downhill, they WEAKENED. This trend is more clearly distinguishable versus the trend for small tornadoes, which shows the opposite relationship. There is more overlap between histograms for small tornadoes, indicating that this relationship is less pronounced.
-
- Figure 6: a) Terrain (colors) and tornadogenesis points (green dots) for all tornadoes in (TEXT INCOMPLETE)
-
- Figure 7: Histogram displaying the frequency of occurrence (y axis) and range of 1000 bootstrapped replicate sample means for tornado rotation (vorticity in [s-1], x axis) for higher than average surface roughness values (purple) vs lower than average surface roughness values (green) for 7 tornadoes analyzed in Price 2023. The trend for the purple on the right hand side means that tornado rotation is stronger, in general, when surface roughness values are higher (i.e., more friction) than when they are lower.
-
- Figure 8: Various results from Houser’s research team’s work. Left (a): Time series between vorticity (yellow) vs surface roughness length (white) for the 25 May 2012 Russel, KS Tornado illustrating the positive correlation between vorticity and surface roughness. Center (b): 9-case aggregation of bootstrapped vorticity deviations (from case medians) for upper (purple) vs lower (green) quartiles of surface roughness acquired from dual-Doppler analyses. Right (c) – Aggregated bootstrapped path direction changes going uphill (red) and downhill (blue) for the same 9 tornadoes as in the center plot.
References:
Butler, K. M. (2017). The Effects of Land Cover Type on Tornado Intensity in the Southeastern US (M.S. thesis, Ohio University).
Houser, J. B., N. McGinnis, K. M. Butler, H. B. Bluestein, J. C. Snyder, and M. M. French, 2020: Statistical and empirical relationships between tornado intensity and both topography and land cover using rapid-scan radar observations and a GIS. Mon. Wea. Rev., 148, 4313–4338, https://doi.org/10.1175/MWR-D-19-0407.1
Muncy, T. J. (2021). Topographic and Surface Roughness Influences on Tornadogenesis and Decay (M.S. thesis, Ohio University).
Price, B. (2023). Using Dual-Doppler Analyses to Investigate the Impacts of Terrain and Surface Roughness on the Three-Dimensional Wind Flow Field. (M.S. Thesis, Ohio University).