Oklahoma bioretention technology demonstration and the effects filter media heterogeneity on performance

Abstract

Ten bioretention cells were constructed in Oklahoma as part of a full-scale technology project to demonstrate phosphorus reduction efficiency. The design used a sand/fly ash blend as the filter media, for enhanced phosphorus attenuation, and incorporated surface sand plugs for improved infiltration in poor draining soils. Engineering considerations, general design procedures, site parameters, construction and planting details and costs are documented.

A three-dimensional finite element model was developed to simulate flow through a bioretention cell and address some of the questions that arose during design and construction regarding the effects of soil amendment implementation, and sand plug size and placement on cell performance. Three general configurations were modeled for three different scenarios. A filter-only configuration was evaluated to assess the effect of filter media hydraulic conductivity heterogeneity on flow and transport. The second configuration added a top soil and sand plug layer with 6 sand plugs measuring 1.5 m by 1.5 m, which was similar to the constructed cells. The final configuration evaluated a top soil and sand plug layer with 14 smaller sand plugs measuring 1 m by 1 m. Three different scenarios were evaluated for each configuration that varied by size and distribution of the filter media heterogeneity. The first scenario used the measured scale and range variability, the second used the same scale with double the variation, and the third used the same variability, but increased the scale volume by a factor of 27.

Model results indicated that variability in fly ash content created complex flow through the filter medium, but did not result in significant preferential flow. Sand plugs created flow concentration but did not dominate flow within the cell, and the number of sand plugs was not significant provided that their total area was sufficient to maintain the desired drainage rate. Mean effluent concentration did not exceed the Oklahoma criterion for scenic rivers until after 22 years and 33 years for the filter only and sand plug configurations, respectively. Modeling predicted more than 144 years of P removal. All distributions show similar removal efficiencies indicating that reasonable mixing effort will enable proper cell performance.