Date of Award




Document Type

Master's Thesis

Degree Name

Master of Science (MS)


Department of Atmospheric and Environmental Sciences

Content Description

1 online resource (xi, 99 pages) : illustrations (chiefly color), color maps.

Dissertation/Thesis Chair

Brian Tang


Convection (Meteorology), Vertical wind shear, Severe storms, Weather

Subject Categories

Atmospheric Sciences


Upstate New York has a variety of complex terrain that can interact with the background flow to create mesoscale heterogeneities in the lower troposphere. The major valleys of Upstate New York, the Hudson and Mohawk Valleys, often have increased moisture content and stronger surface winds than the higher terrain surrounding them. These features can have a profound effect on the evolution of convective storms, especially in cases characterized by low-to-moderate shear, which tends to favor pulse-like or multicellular convection. Analysis of composite radar imagery has indicated that convective storms often change mode while descending from the Catskills Mountains into the Hudson Valley, coinciding with an increase severe weather reports. Some storms exhibited back building once reaching the Hudson Valley. Back building is when a convective line has new cells initiating adjacent to the mature cells such that the line propagates upstream with regards to the low-level flow. Back-building mesoscale convective systems (MCS) have been connected to an increased threat of heavy rainfall and flash flooding, especially in the Northeast. A back-building mesoscale convective system (MCS) from 21 August 2019 was simulated using WRF-ARW to study the mesoscale interactions between the background flow, complex terrain features in and around the Hudson Valley, and the MCS’s convective cold pool. During the three hours proceeding the MCS, southerly terrain-channeled flow created a favorable pre-convective environment in the Hudson Valley through a low-level maximum in water vapor flux. Discrete convection from the Catskill Mountains intensified once reaching the Hudson Valley, creating a cold pool with an outflow boundary oriented across the valley. The channeled flow increased the low-level convergence along the southern portion of the outflow boundary causing high equivalent potential temperature air from the lower Hudson Valley to be lifted, initiating new convective cells. The MCS propagated down the valley until the channeled flow was cut off by another convective line entering the lower Hudson Valley. Decision trees were created to identify characteristics of the pre-convective environment that are conducive to back building in the Hudson Valley. Composite radar imagery was analyzed to identify cases with (n=15) and without (n=55) back building from June, July and August 2015-2020. HRRR 0-hour analyses, valid at 1800 UTC, were used to calculate area averaged variables in the Hudson Valley for each case and analyzed by the decision tree classifier. Variables related to surface-based instability, such as surface-based CAPE and lifted index, and low-level moisture content, such as 2-m AGL dew point depression, were chosen most often by the decision tree classifier. A high value of surface-based instability makes it more likely that a new updraft along the outflow boundary will grow into deep convection. A low 2-m dew point depression makes it more likely that a small vertical displacement of the surface parcel by the outflow boundary will result in saturation and positive buoyancy. Comparison of composite wind profiles from cases with and without back building revealed a difference in the wind speeds from 900 hPa to the tropopause. A mean tropospheric wind of 30 kt made it more likely that cells would move out of the Hudson Valley before forming a strong cold pool. A mean tropospheric wind of 20 kt resulted in slower cell motion, allowing stronger cold pools to form in the Hudson Valley and increasing the likelihood of back building.