Date of Award




Document Type


Degree Name

Doctor of Philosophy (PhD)


Department of Atmospheric and Environmental Sciences

Content Description

1 online resource (xviii, 178 pages) : illustrations (chiefly color), color maps.

Dissertation/Thesis Chair

Robert G Fovell

Committee Members

Lance Bosart, Daniel Keyser, Brian Tang


Cyclones, Cyclone forecasting, Boundary layer (Meteorology), Atmospheric turbulence, Atmospheric models

Subject Categories



Subgrid-scale turbulence in numerical weather prediction models is typically handled by a PBL parameterization. These schemes attempt to represent turbulent mixing processes occurring below the resolvable scale of the model grid in the vertical direction, and they act upon temperature, moisture, and momentum within the boundary layer. This dissertation utilizes idealized and full-physics numerical model simulations to understand how variations in turbulent mixing parameterizations may influence sensible weather forecasts of baroclinic cyclones across a variety of spatial and temporal scales. Furthermore, a primary pathway through which PBL turbulence projects upscale during baroclinic cyclone events is identified using a combination of Eulerian and Lagrangian analysis techniques. A series of sensitivity experiments were performed to determine how influential the choice of PBL scheme was to the forecasted surface cyclone track and precipitation footprint. Turbulent mixing was then modified using a single PBL scheme to isolate the source of the demonstrated variability in cyclone evolution. The primary physical process through which turbulent mixing projects on the larger-scale baroclinic environment was identified with the assistance of trajectory plume analysis. Results suggested that varying the critical bulk Richardson number (BRN) in a K-profile PBL scheme was similar to selecting a different PBL parameterization. Differences in boundary layer moisture availability, arising from reduced entrainment of dry, free tropospheric air, led to variations in the magnitude of latent heat release above the warm frontal region, producing stronger upper-tropospheric downstream ridging in simulations with less PBL mixing. The more amplified flow pattern impeded the northeastward propagation of the surface cyclone and resulted in a westward shift of precipitation. In addition, the local influence of parameterized PBL turbulence was examined, complementary to the aforementioned large-scale projection. A focus was placed on the Yonsei University (YSU) and the 2016 version of the Global Forecast System (GFS) PBL parameterization schemes due to their widespread usage in the community and similar assumptions owing to their shared lineage to the older Medium Range Forecast PBL scheme. Single-column-model experiments were conducted to establish the source of the differences in turbulent mixing imposed by each scheme across two contrasting environments. The GFS scheme was found to mix more vigorously through a greater depth than the YSU over all of the environments tested. The differences were largely a result of variations in the critical BRN and the stable velocity scale between the two schemes. Lastly, the influence of PBL mixing variations between the YSU and GFS on cool-season precipitation were illustrated using simulations of a mixed-precipitation event over the northeastern U.S. Further sensitivity tests were performed to establish the reasoning behind the varying precipitation forecasts between each PBL scheme. The stratocumulus (SC) mixing scheme within the GFS was found to drastically reduce frozen precipitation accumulation relative to the rest of the scheme. Further inspection of the SC scheme suggested the additional turbulent mixing forced by the scheme was unnecessary for the simulated conditions. A modification to the GFS SC scheme was proposed and the behavior of PBL schemes within precipitating boundary layers was discussed.

Included in

Meteorology Commons