An investigation of the development of supercells in the Indiana and Ohio tornado outbreak of 24 August 2016 using a WRF model simulation

On 24 August 2016, a tornado outbreak that produced 24 confirmed tornadoes impacted portions of Indiana and Ohio. Initially elevated multicellular convection transitioned to surface-based supercells between 1700 and 1830 UTC, after which time tornadoes began to occur. Such a transition is of particular interest owing to its relatively rare occurrence and because most convection-allowing models failed to predict it in this case, instead depicting a line of storms moving across the affected area. Three surface boundaries were present prior to and during the event. The first was an outflow boundary traced back to convection in Nebraska and Iowa the previous evening, along which an elevated storm cluster formed that later developed into three discrete supercells. The second was a differential heating boundary owing to cloud shading from a leading cluster of storms. This boundary moved northeastward across Indiana and its position coincided with the formation of each of the supercells and the production of their first tornadoes. The third was a westerly wind shift behind the outflow boundary that triggered a second round of supercells, some of which also produced tornadoes. A mesoscale convective vortex, also owing to storms from the previous evening, moved across northern Illinois and Indiana during the event augmenting the low-level and deep-layer vertical wind shear. A Weather Research and Forecasting (WRF) model simulation accurately captures the environment during this event, depicting a cluster of elevated thunderstorms developing in the vicinity of an outflow boundary during the early morning hours of 24 August. Trajectory analyses indicate that parcels entering the multicellular updrafts before 1530 UTC originate above 1 km, meaning that the convection is elevated. As diurnal heating progresses, more near-surface parcels enter the multicellular updrafts, each of which move atop convective outflow and weaken. Since nearly all of the vertical wind shear is below 1 km, mesocyclones do not form via tilting of horizontal vorticity until the storms ingest air from this layer. The storm on the southern end of this cluster becomes a supercell in the simulation around 1700 UTC, after it becomes surface based. Inflow trajectories pass through an area of high 0-1 km storm-relative helicity, likely owing to anvil shading southeast of the simulated storm. When these helicity-rich parcels enter the supercell updraft, strong near-surface rotation develops. A novel analysis of the perturbation pressure field from the WRF model output indicates that the development of relatively large vertical perturbation pressure gradients coincide with when near-surface air begins to enter the updrafts, resulting in upward accelerations in the lowest 2 km, below the level of maximum rotation, and downward accelerations above this level. In strengthening updrafts, upward-directed perturbation pressure accelerations due to buoyancy may offset the downward-directed non-linear perturbation pressure accelerations above the level of maximum rotation, allowing the updrafts to deepen and intensify. This simulation suggests that the transition from disorganized elevated convection to surface-based supercells begins as more near-surface air gradually enters the multicellular updrafts, leading to an increase in updraft rotation in new cells and in the magnitude of the upward-directed non-linear perturbation pressure gradients, which likely aid in the development of a single dominant supercellular updraft