Date of Award


Document Type


Degree Name

Master of Science


Department of Engineering Physics

First Advisor

Michael K. Walters, PhD


Diffusion-Transport (D-T) modeling is a branch of numerical weather prediction concerned with eddy diffusion of particulate pollutant plumes and their transport by the wind. When conducting D-T modeling, establishing the height of the planetary boundary layer (PBL) is crucial to defining the vertical bounds within which a plume can become thoroughly mixed. The PBL can be deduced from observations or model simulation. Three sounding analysis PBL algorithms were considered - the Potential Instability Mixing Depth (PIMIX), Potential Temperature (POTEMP), and Gradient Richardson Number (RICH) algorithms. A turbulent kinetic energy (TKE) based PBL algorithm was also evaluated. The purpose of this research was threefold. First, observed atmospheric soundings were input, and algorithm output was compared to human analyses of the observed soundings (Analysis). Second, Regional Atmospheric Modeling System (RAMS) generated forecast soundings were input, and algorithm output was compared to human analyses of the forecast soundings (Simulation). Finally, algorithm output from simulation was compared with the human analyses from analysis (Verification). These PBL comparisons were put into one of four categories: (1) hit, (2) indication of deep convection, (3) miss, or (4) algorithm failure. Algorithm performance was ranked based on the number of hits, then on indications of deep convection. PIMIX was the best analysis tool, while both POTEMP and TKE were the best simulation methods. All algorithms had a similar number of hits for verification, but PIMIX had more estimates indicative of deep convection, so was ranked best.

AFIT Designator


DTIC Accession Number



The author's Vita page is omitted.