Effectiveness of Highway Bioswales in Treating Stormwater Runoff in Maryland

This long-term field study assessed the performance of highway bioswales in controlling stormwater and stormwater pollution associated with Maryland’s highway system; bioswale performance was assessed relative to grassed swales with which the bioswales were paired at three sites across the state (Washington County, Frederick County, and Howard County). For one pair of swales in Howard County with the most extensive and highest quality data, analysis of 155 runoff events over an eight-year period showed no statistically significant differences in mean stormflow volume, although data (84 events) from a second pair with a shorter record showed about a 30% reduction. Peak runoff reduction fell in the range of 14-19%, but this “apparent” difference is likely overstated because of systematic differences (~10%) in event rainfall and peak rainfall intensity between the swales (grassed swale > bioswale). Underdrainage dominated bioswale runoff at all three sites, contributing 93 – 97% of total runoff from the three bioswales with the most extensive data. The high efficiency by which rainfall was “captured” and discharged was attributed to extremely high measured infiltration capacity of the engineered soil media and the unregulated underdrains that are characteristic of the bioswale design.

With respect to stormwater quality, we examined differences in paired event mean concentrations for more than 25 different measured water pollutants/constituents. We found no significant differences in chloride between any of the three paired swales. Since we expected that road-salting on these highways would produce comparable levels of chloride in runoff, the lack of significant differences provided confirmation that our hydrologic and water quality monitoring were being done with acceptable accuracy and precision. Using paired comparisons, we found significantly lower orthophosphate-P concentrations for the same three bioswales compared to the grassed swale controls which we attributed to greater soil P chemi-sorption; overall, the bioswales reduced event mean orthophosphate-P concentrations by 37 – 86% relative to the grassed swale controls. Total concentrations were also significantly reduced (19 – 72%). We found significantly higher concentrations of both nitrate-N (210 – 590%), total N (24 – 125%), and several trace metals (Cu: 40 – 590%; Cd: 47 – 120%; and Zn: 200 – 960%) in combined runoff from all three bioswales, however. We attributed these responses primarily to excessive leaching of the biosoil materials by percolating water. The latter results strongly suggest that regulatory authorities should consider some major design modifications to improve overall bioswale water quality performance with respect to these pollutants.