The phosphorus saturation index (PSI) and P saturation ratio (PSR) were tested across a wide range of bioretention soil mixtures (BSMs) for their relationship to total and dissolved P in column leachate. The BSMs tested were made using different feedstocks including sand alone, food and yard waste compost, biosolids and yard compost, and high Fe biosolids. The PSI of the mixtures ranged from 0.23 ± 0.03 (biosolids and yard waste compost 15%, oyster shells 5%) to 1.26 ± 0.02 (biosolids and yard waste compost 80%). The PSR of the mixtures ranged from 0.05 ± 0 (100% sand) to 3.12 ± 0.12 (biosolids and yard waste compost 80%). A total of 12 storm events were staged using both synthetic rainwater (total P = 1.71 ± 0.3 mg L) and actual stormwater (total P = 0.08 ± 0.03 mg L). Excluding the sand-only mixture, all treatments were a source of P for the duration of the study. However both total and dissolved P concentrations decreased over time. Effluent concentrations of total P in the first event ranged from 0.59 to 75.55 mg L but decreased by the final event to between 0.15 and 10.73 mg L. The PSR was found to be a good predictor of P leaching from all BSMs with an of 0.73 for total mass of P leached across all leaching events. The PSI was a poor predictor ( < 0.3). The PSR also predicted total P in leachate for individual events more effectively than the PSI. Total P in the BSMs had no relationship to P in the leachate.

A column study was conducted to test the ability of bioretention soil mixtures (BSMs) to remove nutrients, metals, and polyaromatic hydrocarbons (PAHs) from stormwater collected from an urban highway. Infiltration rate, plant growth response, and turbidity of the effluent were also measured. The BSMs were made from a range of types and rates of composts and additional materials such as water treatment residuals, sawdust, and oyster shells. Sand was used as a control. Total N and P in stormwater measured 1.8 ± 1 and 0.08 ± 0.03 mg L −1 . All treatments were a source of these nutrients. Metal concentrations in the stormwater were low, with mean Cu and Zn concentrations of 39.8 ± 19.1 and 173 ± 113 mg L −1 , and Cd and Pb close to detection limits. All treatments absorbed Cu and Zn from stormwater with varying levels of removal efficiency. The three treatments tested removed 84 to 100% of the PAHs from the stormwater. In general, contaminant removal (N, P, and Zn) efficiency was not related to infiltration rate, with a slight decrease in Cu removal efficiency observed with increased infiltration rate (R 2 = 0.32). These results indicate that the BSMs tested were a source of nutrients but were generally effective at removing metals and PAHs from stormwater.

Bioretention systems are increasingly used as alternatives to engineered systems for stormwater management in urban areas. These systems typically consist of sand and compost and are designed to allow for rapid infiltration and filtering of stormwater. In some cases, these systems drain directly into streams. In other cases, the water percolates through the subsoil into groundwater. The systems are also designed for a wide range of settings including curbside systems, systems in parking lots and systems alongside buildings. The contaminants of concern in the stormwater include nutrients (N and P), metals (primarily Cu, Pb and Zn), organics (primarily PAHs) sediment and pathogens. Recent work has shown that untreated stormwater is toxic to fish and that filtering stormwater through a sand, compost and bark mixture was sufficient to eliminate any negative impacts (McIntyre et al., 2016). The current regulations on bioretention systems for green stormwater infrastructure in WA State specify that the compost used not include municipal biosolids or animal manures. The rational for this guidance is that biosolids and manures are considered to have higher nutrient and metal concentrations than yard waste or food scraps. The current guidance does not take into account that composts can have widely different characteristics based on feedstock blends and different ratios of each component of the mixture. Previous work demonstrated that feedstocks were not a significant predictor of performance for bioretention systems . In contrast, the phosphorus saturation index (PSI), a test to estimate the potential for phosphorus leaching, was shown to be highly predictive of efficacy of the different mixtures for both nutrients and metals. MethodsWe conducted a replicated field trial to test a range of different bioretention soil mixtures (BSM) for their ability to remove metals and nutrients from stormwater. The study included 40% food yard waste compost and 60% sand by volume, the current recommended mixture in WA State and a 100% sand as control treatments. Other treatments included biosolids: yard waste compost produced by King County added at different rates with sand, and +/-Fe based water treatment residuals, and high Fe biosolids from DC Water, used both directly and following composting. A list of the treatments used in the study is shown below. The control treatments and the treatments that include biosolids from DC Water are shown in bold.