Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Civil and Environmental Engineering

First Advisor

Marty E. Tittlebaum


In Part One, three bench-scale rock filters containing 2 feet of gravel were investigated. Two of the filters were planted with Sagittaria lancifolia and Scirpus validus, while the third filter was an unvegetated control filter. During a preliminary portion of this study, the BOD$\sb5$ surface loading was held constant at 4.96 g/day/m$\sp2$. The BOD$\sb5$ mass removal percentages at this time averaged 75%, 60%, and 44% for the Scirpus, Sagittaria, and control systems, respectively. Following this constant loading rate, an 80-day experiment was run on the filters using eight combinations of two flow rates and four influent BOD$\sb5$ concentrations, each combination remaining constant for ten days. These combinations resulted in BOD$\sb5$ surface loadings from 4.63 to 30.96 g/day/m$\sp2$. Overall average BOD$\sb5$ removal percentages during this latter portion of the study were 69%, 57%, and 47% for the Scirpus, Sagittaria, and control systems, respectively. ORP and DO measurements within these systems indicated no free oxygen available at any depth. TKN removal was higher in the plant systems relative to the control, with the Scirpus system achieving a higher overall removal than the Sagittaria system. This increased nitrogen removal may be due to nitrification occurring in the thin aerobic zone surrounding portions of the plant roots. The data gathered from the Part One bench-scale study was used in Part Two to develop a computer model for predicting effluent BOD$\sb5$ concentrations. This computer model was then applied to a full-scale municipal system. As determine by the sum of squared residuals parameter, consistently more accurate predictions were obtained with the computer model than with two currently used equations for all three bench-scale systems and the full-scale system. An analysis of the computer model BOD$\sb5$ mass balance suggests that microbial degradation occurring on the rock surfaces is responsible for the greatest decrease in BOD$\sb5$; furthermore, this degradation rate increases in the presence of aquatic plants. By making several simplifying assumptions, the computer model is reduced to a manageable equation using easily obtained parameters.