CICEET Progress Report for the period 8/1/02 through 1/31/03

Project Title: Engineering Bioretention for Treatment of Storm Water Runoff
Principal Investigator(s): Allen P. Davis

Accomplishments
Scheduled Tasks:
The work for the fifth six-month period of the project focused on Task 1-- Column/Mechanistic Experiments and Task 3 -- Data Analysis.

Progress on Tasks
Three 6-hr bioretention column experiments composed of the same types and amounts of media, but with different media configurations, were completed. In addition, four column experiments using the same media have been finished to evaluate bioretention performance during repetitive storm events. The data, including infiltration rates and pollutant removal efficiencies of infiltrating runoff collected from nineteen column experiments using different media, were employed to calculate the total input, output, and removed mass of pollutants.

The media compositions for the three 6-hr columns are summarized in Table 1. The objective for this part of the study was to investigate the effect of media configuration on the infiltration rate of runoff, as well as pollutant removals.

For the repetitive experiments, the interval for each experiment is four to seven days. Generally, mulch and soil contain abundant organic matter and could serve as the media for plant growth. Also, organic matter can serve as the carbon source for microorganism growth. Both of these are helpful to the operation of bioretention. In addition, because of the bigger media particle size, Sand (II) can treat a larger volume of runoff before clogging. Therefore, mixing mulch and soil with Sand (II) seems to be a good mixture for the upper media. Below the top layer, Sand (I) can efficiently remove pollutants and is a better choice among other media with respect to pollutant removal. The top two layers can serve as the bioretention media for rapid infiltration of first flush runoff with efficient pollutant removal. Finally, soil with high pollutant removal capacity in the bottom can enhance the bioretention performance. Therefore, in this study, a three-layered media was used based on the hypotheses above. The top layer was formed by mixing 3.06 kg mulch, 3.06 kg Soil (IV), and 6.13 kg Sand (II) homogeneously (30 cm). The middle layer was 23.2 kg Sand (I) (55 cm), and the bottom was 5.9 kg Soil (IV) (10 cm) (See Figure 1).

A paper entitled "Evaluation of Bioretention for Treatment of Urban Storm Water Runoff" has been selected as a Podium Presentation and will be presented at the World Water and Environmental Resource Congress 2003, Philadelphia, PA, June 22, 2003. This conference is sponsored by the American Society of Civil Engineers. Also, an abstract "Multiple-Event Study of Bioretention for Treatment of Urban Storm Water Runoff" was submitted to the 7th International Conference- Diffuse Pollution and Basin Management, which will be held during August 17 to 23, 2003 in Dublin, Ireland. The International Water Association sponsors this latter conference. Finally, a manuscript entitled "Evaluation and Modification of Bioretention Media for Treatment of Urban Storm Water Runoff" is in preparation for submission to a referred journal.

Difficulties Encountered
No major difficulties have been encountered.

Preliminary Data
For the three new media mixes, infiltration rates of runoff ranged from 1.47 to 5.87 cm/min. The infiltration rate in the column with homogeneous media was the highest (5.06-5.87 cm/min). The column with Soil (IV) in the bottom resulted in slightly higher infiltration rate (1.67 cm/min) than the one with Soil (IV) in the top (1.47 cm/min). This is probably because the higher water head caused by the infiltrated runoff in the upper Sand (II) layer helps runoff to flow through the bottom soil layer.

The pollutant removal efficiencies for these three column experiments are summarized in Table 2. All experiments demonstrated excellent removal efficiencies for oil/grease (>99%). Lead was similarly removed in both layered columns (>99%) but some lead (17~99% removal) leached out along with the SS in the column with homogeneous media. TSS was leached out from the homogeneous media during the testing period (-509) to 93% removal), whereas both layered media removed over 90%. With respect to TP, the level in the effluent increased when large amounts of SS leached out. For nitrate-N, the removal efficiency ranged from 6% to 8%. For ammonium-N, the removal efficiency for these three experiments ranged from 14% to 18%. Different configurations of media produced similar low removal for these two pollutants during the 6-hr period. Therefore, different configurations of bioretention media can result in much different infiltration rates. For pollutant removal, some suspended solids could leach out because of the fast infiltration of runoff and lead to poor performance.

For the second set of repetitive experiments, the infiltration rate of runoff during the first 6-hr was stable (5.1 cm/min). During the subsequent three runs, runoff infiltrated into the media slowly in the beginning and later became steady-state flow (5.1 cm/min). The pollutant removal results are summarized in Table 3. More data collection is ongoing.

Data Analysis for 6-hr Column Experiments
Pollutant removal efficiencies for previous 6- hr column experiments are summarized in Progress Reports 2 and 3. Based on all data, effects of media components, silt/clay content in the media, and media configurations on bioretention performance are now discussed.

In this study, infiltration rate and pollutant removal efficiency are two primary factors used to evaluate the effectiveness of a bioretention column during a 6-hr run. Based on the data, although several media types can remove >99% O/G and lead, the total removed mass of these two pollutants from runoff varied widely because of the different infiltration rates. In addition, Sand (II) only removed 10% TP, which is much poorer than Soil (I). However, the total removed mass of TP by Sand (II) is higher than by Soil (I). Therefore, the total mass of input, output, and removed pollutant, which combines the effects of both infiltration rate and pollutant removal efficiency, are considered in this study.

Different sands and soils result in different infiltration rates because of their different particle sizes and textures. Input, output, and removed mass of pollutants for different native media are summarized in Figure 4. Silt and clay in the medium slow the runoff infiltration rate. In order to investigate the effect of these factors on runoff treatment, the results for columns employing different homogeneous media mixtures (Sand (I)/Soil (I)) are compared and summarized in Figure 5.

Also, different media configurations (layered and homogeneous) result in different runoff infiltration rates and pollutant removal efficiencies. Comparisons of column results employing different media configurations (composed of Sand (I) and Soil (I)) are shown in Figure 6. Summarizing all the analyses, all media employed in this study can remove O/G, TSS, and lead efficiently, especially Sand (II). However, Sand (II) has the worse TP removal capacity among all native media. Therefore, Sand (II) is a good option serving as a single medium if the runoff contains high concentrations of O/G, TSS, and lead. By mixing Sand (II) with other media such as soil, TP removal efficiency can also be improved.

Anticipated Success in Meeting Project Objectives in Scheduled Project Period
The final two field studies (See Figure 2 and Figure 3) were conducted during actual rainfall events. These studies were just completed on February 4, 2003 at two bioretention cells recently constructed on the University of Maryland campus. These samples are now being analyzed. It is expected that all project objectives will be met by May 2003.

Tasks and activities for next reporting period

Tasks for the next reporting period
The tasks for the next reporting period are to finish all proposed project work. The objectives for these tasks are to finalize conclusions based on all results and provide recommendations for future bioretention design.

Work plan to accomplish tasks
One Ph.D. student is working full time on the project. An undergraduate student has been assisting 1/2-time over the fall and will continue in the spring. All laboratory and field results will be evaluated and concluded.

Concerns or difficulties
At this point, we have no concerns or difficulties.

Expenditures
Expenditures are in the range expected for the work accomplished to date.