The Dynamical Influences of Cloud Shading on Supercell Thunderstorms

Open Access
Frame, Jeffrey William
Graduate Program:
Doctor of Philosophy
Document Type:
Date of Defense:
October 06, 2008
Committee Members:
  • Paul Markowski, Dissertation Advisor
  • Paul Markowski, Committee Chair
  • Yvette Pamela Richardson, Committee Member
  • Jerry Y Harrington, Committee Member
  • William Menaul Frank, Committee Member
  • Andrew Mark Carleton, Committee Member
  • numerical modeling
  • convection
  • radiative transfer
  • supercells
Numerical simulations of supercell thunderstorms which include parameterized radiative transfer and surface fluxes are performed using the Advanced Regional Prediction System (ARPS) model. The tilted independent column approximation (TICA) is adopted for use in the ARPS model because the existing method of parameterized radiative transfer, the independent column approximation (ICA), permits only the vertical transfer of shortwave radiation. The computed radiative fluxes from both the TICA and ICA are compared to output from a three-dimensional Monte Carlo radiative transfer solver and it is determined that the TICA fluxes more closely match those from the Monte Carlo model than do those from the ICA. Additionally, the TICA is able to capture the extensions of shadows that occur when the solar zenith angle deviates significantly from zero, which cannot be captured by the ICA. The maximum low-level air temperature deficits within the modeled cloud shadows is 1.5 to 2.0 K, which is only about half that previously observed. The loss of strong solar heating of the model surface within the shaded regions cools the surface temperatures, and changes the sign of the sensible heat flux near the edge of the shadow. This stabilizes the model surface layer and suppresses vertical mixing at low levels within the shaded area. This reduction in vertical mixing means that higher momentum air from aloft is prevented from mixing with air near the surface that has lost momentum to surface friction. The net result of this is a shallower, but more intense vertically-sheared layer near the surface. As the supercell's rear-flank gust front propagates into this modified shear layer, the layer of cold outflow air becomes shallower and it accelerates eastward. In the case of a stationary storm, the cold outflow undercuts the updraft and mesocyclone, depriving them of warm and moist inflow, and ultimately weakening the storm. These results are not likely applicable to all simulations of supercells with radiation because varying surface characteristics alter the amount of frictional drag experienced by the low-level flow. Additionally, the propagation of the rear-flank gust front is heavily modulated by both the strength and the location of the outflow, which are influenced by the choice of the storm-relative wind profile and the microphysics package. If shortwave radiation is excluded from the model, a shallow stable layer forms over the entire domain and the storm becomes elevated and weakens. The direct absorption and emission of radiation by clouds does not significantly affect the simulated supercells. The base-state environment is changed to see under which conditions cloud shading and friction combine to force the undercutting of the updraft. Neither a morning model initialization nor a cold season model initialization prevent this from occurring in any of the simulations which produce an anvil shadow. The ground-relative wind is also varied because the surface fluxes of both heat and momentum are not Galilean invariant. A storm in which both the rear-flank gust front and updraft slowly move along the major axis of the anvil shadow becomes undercut, much like the stationary storm. A fast moving storm, however, does not become undercut because less time exists to cool the model surface and to decouple the surface layer if the storm moves faster. If the gust front moves into the anvil shadow and the updraft moves normal to the shadow (i.e., the northward movement of the updraft for an eastward-extending anvil), cyclic behavior can result, although this is highly dependent on storm motion. If the gust front propagates into the full sun (i.e., southward movement), the storm is relatively unaffected by the presence of radiation because the dynamics that govern gust front propagation remain relatively unchanged.