Degree Type


Date of Award


Degree Name

Master of Science


Geological and Atmospheric Sciences



First Advisor

William Gallus


During the spring and summer months, convective precipitation is a common occurrence across the central United States. This frequent precipitation often comes in the form of mesoscale convective systems (MCSs), which provide a large amount of the necessary water to the agricultural industry of the region. MCSs, which require large amounts of moisture and wind shear to form, are often fueled in this region by the Great Plains Low-Level Jet (LLJ). These phenomena are responsible for a large majority of the moisture transport into the region and are a main ingredient in summer MCSs. While numerical weather prediction has improved greatly in recent years, accurate forecasting of the LLJ still remains a challenge. Therefore, a further understanding of LLJ simulation is needed to better predict MCSs. To understand how models can simulate the LLJ, atmospheric conditions with the planetary boundary layer (PBL) must be analyzed and understood within numerical models. Parameterization is important in representing small-scale processes and closing systems of dynamic equations, but model sensitivity to these parameterization schemes is something that has not been greatly explored. This work focuses on investigating how three different PBL schemes, the Mellor-Yamada-Janjic (MYJ) scheme, the Mellor–Yamada–Nakanishi–Niino (MYNN) scheme, and the Yonsei University (YSU) scheme, impact the simulation of the PBL and the impact those conditions have on LLJs.

First, the Weather Research and Forecasting model was used to simulate 30 cases previously studied by Squitieri and Gallus (2016a,b), including 15 strongly-forced cases (Type C) and 15 weaker-forced cases (Type A). The behavior of the PBL and LLJ were analyzed for each case to determine the impacts different PBL scheme had on the model output. It was found that MYJ and MYNN had similar tendencies when simulating the afternoon PBL, but when simulating the LLJ, MYNN tended to be more of an outlier, especially when forecasting the location of the LLJ peak. YSU, being a nonlocal closure scheme compared to two local schemes, behaved differently from the other two schemes, more often simulating a dry and warm boundary layer compared to the other schemes as well as observations, which had impacts on the LLJ.

Second, a modified version of the MYNN scheme, referred to hereafter as the mod-MYNN scheme, was compared to the original as well as observational data to see if modified physics were able to improve the forecast of the LLJ and subsequent MCSs. It was found that both schemes perform similarly to one another, but the modified scheme performed slightly more like a nonlocal scheme, displaying similar biases to those of YSU in the first part of the experiments, although both schemes fail to completely capture the MCSs and resulting nocturnal precipitation accurately.


Copyright Owner

Michael Garberoglio



File Format


File Size

104 pages