IntroductionUrban water quality is affected by multiple stress factors including impervious surface area; rainfall intensity and duration; and land slope, soil type, and human land-use activity (Park et al. 2009). These stresses can lead to widescale erosion, excessive nutrient loading, eutrophication, dead zones, and a surge in toxic chemical concentrations (USEPA 2000c; Davis and McCuen 2005; LeFevre et al. 2015). To mitigate the increased volume of water and reduce nutrients, metals, and other pollutants in stormwater runoff, many different stormwater control measures (SCMs) have been designed and implemented in the urban landscape. Bioretention, generally consisting of layers of engineered media (sand, gravel, soil, and organic matter) topped with vegetation, is one of the most widely used SCMs for this purpose and is highly effective in hydrologic management and water quality improvement (Davis and McCuen 2005; Davis et al. 2009; Hunt et al. 2012). Many studies have shown bioretention to be extremely effective in metal uptake and suspended solids removal. Removals of nitrogen (N) and phosphorus (P) are more variable and are highly dependent on the system design and media characteristics, respectively (Davis et al. 2001, 2003; Glass and Bissouma 2005; Hunt et al. 2006; Li and Davis 2008; Liu et al. 2014; Hurley et al. 2017; Shrestha et al. 2018; Skorobogatov et al. 2020).Interest is growing in the incorporation of reused, recycled, and reclaimed materials in various construction projects (Kirchoff et al. 2003; Herrera Environmental Consultants 2012). One such recycled material, compost, is of particular interest and can consist of numerous different source constituents including biosolids, manure, food waste, and yard waste. In the US, nearly 30% of municipal solid waste produced nationwide is compostable (USEPA 2019). Converting this waste stream to compost and compost-based projects could significantly reduce waste volume and production in the US.Compost is an attractive SCM additive due to its potential to increase organic matter and nutrients content, water-holding capacity, cation exchange capacity, soil porosity, soil stabilization and aggregation, and buffering capacity of the SCM media (He et al. 1992; Mitchell 1997; Weng et al. 2002; Kirchoff et al. 2003; Paus et al. 2014). These properties can be beneficial for vegetation establishment and growth, water infiltration and storage, and heavy metal and organic contaminant retention; however, compost is often high in N and P, which may be counterproductive in SCMs (Kirchoff et al. 2003). Previous research has noted that bioretention soil media (BSM) amended with various types of compost had the potential for pollutant removal, but has also led to excessive nutrient leaching [both 80% yard/20% food (Iqbal et al. 2015; Mullane et al. 2015); several types: yard/food, manure/food, manure/bedding, and vermicompost (Hurley et al. 2017); and manure/food/wood (Shrestha et al. 2018)].Compost characteristics can impact nutrient leaching from compost-amended bioretention media. Hurley et al. (2017) noted that composts that excluded manures, biosolids, and food waste can lead to lower P leaching compared with composts that contain these materials. Additional amendments or pretreatment of the composted material before incorporation into the BSM media may mitigate some leaching. Aluminum-based water treatment residual (WTR) has shown promise in previous studies as a P mitigation strategy (Lucas and Greenway 2011; Palmer et al. 2013; Poor et al. 2019; Ament et al. 2021, 2022). WTR is a byproduct of the drinking water treatment process that offers high surface area and large concentrations of aluminum and/or iron which, when used in bioretention, has shown improvement in the adsorption of dissolved P. The aluminum and iron (hydr)oxides provide strong adsorption sites for dissolved inorganic and organic P species (O’Neill and Davis 2012; Liu and Davis 2014; Ament et al. 2021).Many jurisdictions have recommended compost as a bioretention media component; however, these prior studies that indicated N and P leaching suggested that these recommendations should be reevaluated. This work uses a set of mesocosm column studies to examine the mechanisms of nutrient release and transport from the addition of compost to BSM. Leaching characteristics from synthetic stormwater–applied mesocosms were measured and used to analytically compare unamended BSM with two different compost source material types (biosolids and green waste–derived compost) at two compost mixing ratios (15% and 30% compost/BSM). The dynamics of effluent N and P concentrations were evaluated, both on a per-storm basis and over the time frame of several simulated storm events. Additionally, two different nutrient mitigation steps were evaluated as possible beneficial treatments for compost incorporation into the bioretention media (WTR mixed with green-waste compost and mechanical rinsing of the biosolids compost).Methods and MaterialsMaterialsBSM meeting Maryland Department of Transportation State Highway Administration (MDOT SHA) Specification 920.01.05, containing washed silica sand, furnished topsoil, and aged hardwood mulch (MDOT-SHA 2022) were supplied by Stancill’s (Waldorf, Maryland). Biosolids-derived compost, marketed as Orgro High Organic Compost (Baltimore City Composting Facility, Baltimore, Maryland) originated from the Back River Wastewater Treatment Plant in Baltimore, Maryland. Green waste–derived compost, marketed as Leafgro, was sourced from the Montgomery County Yard Trim Composting Facility (Dickerson, Maryland). Aluminum-based WTR was provided by the Dalecarlia drinking water treatment facility (US Army Corps of Engineers, Washington, DC). Both composted materials were produced aerobically and are compliant with MDOT SHA Specification 920.02.05 (MDOT-SHA 2022).To remove fines from the sand and gravel, both were placed in buckets and washed with tap water until the water ran clear. All compost, sand, WTR, and BSM were sieved with a 1-cm sieve prior to column setup. Due to the claylike structure of WTR, the material was air-dried for 2 days after sieving. Table 1 presents compost and BSM characteristics as measured in the University of Maryland Environmental Engineering Laboratory: bulk density and KCl extractable nitrogen (Castle 2009), CaCl2 extractable phosphorus (Moore and Joern 2009), and Mehlich-3 extractable phosphorus (Mehlich 1984).Table 1. Bulk density, KCl nitrogen, CaCl2 phosphorus, and Mehlich-3 phosphorus extraction results for bioretention soil media, both composts (green waste and biosolids), and washed biosolids compost (average ± standard deviation)Table 1. Bulk density, KCl nitrogen, CaCl2 phosphorus, and Mehlich-3 phosphorus extraction results for bioretention soil media, both composts (green waste and biosolids), and washed biosolids compost (average ± standard deviation)MixtureDry bulk density (kg/m3)KCl nitrogen (mg-N/kg)CaCl2 phosphorus (mg-P/kg)Mehlich-3 phosphorus (mg-P/kg)BSM1,900±4704.3±0.640.05±0.000.93±0.08Biosolids340±1116,000±27074±2.6820±200Green waste290±22740±7710±0.3673±35Washed biosolids430±255,900±7101.9±3.6250±50Green waste with WTR337—4.6—Mesocosm Experimental DesignClear acrylic column mesocosms (Fig. S1) were used to conduct storm simulation studies in a controlled greenhouse environment under a constant air temperature of approximately 20°C and natural lighting. The columns had 19.1-cm inner diameters, were 122 cm tall, and had a small acrylic wedge with a 14.7° slope at the bottom to prevent pooling. Each column was packed with a 7-cm layer of gravel (13–19 mm, 2 L), 7-cm layer of sand (2 L), and 77 cm (22 L) of BSM/modified media. Layers were loosely packed by the weight of the next layer and light shaking. The sand and gravel layers were separated from the test media by a 1-mm2 mesh screen to prevent media loss from the system. After the first experimental run, additional media were added to return the media height to 77 cm. Columns were wrapped in foil to limit sunlight exposure to the media profile. Tall fescue grass (Festuca arundinacea) sod was obtained from Behnke Nursery in Beltsville, Maryland; circles were cut in the sod to fit into the columns, loose soil was shaken from the sod, and sod was placed on top of the media.The media layer in each of the columns was composed of mixed materials as described in Table 2. The control column was made of 100% BSM and the compost-amended columns were mixed and filled based on volume. For the 30% washed biosolids (30WB) column, the compost media was soaked in tap water by the research team at a 1∶1 ratio by volume for 1 h, drained, and air-dried overnight before being mixed with the BSM. The drainage water was analyzed as described subsequently. The media used in the 15% green waste with WTR (15G+WTR) was prepared by evenly mixing 3.75% WTR by air-dried volume with 81.25% BSM and 15% green-waste compost.Table 2. Media compositions in mesocosm columnsTable 2. Media compositions in mesocosm columnsColumn IDBSM30G15G15G + WTR30B15B30WBBSM (%)a100708581.3708570Compost (%)a0301515301530Compost sourceN/AGreen wasteGreen wasteGreen wasteBiosolidsBiosolidsBiosolidsWTR (%)a0003.8000For all column trials, synthetic stormwater solutions were made using tap water with the added constituents presented in Table S1. Sodium bisulfite at 2.2 mg/L was added to neutralize the chlorine in the tap water. Applied synthetic stormwater had TN concentrations of 4.0±1.0 mg-N/L. Only dissolved pollutants were used in the stormwater input. Concentrations of total N and total P vary widely in urban stormwater. A median value for TN and TP of 1.85 and 0.20 mg/L, respectively, has been calculated based on large data sets (Pamuru et al. 2022).Synthetic stormwater events were applied to each column weekly for 6-h intervals over a period of 8 weeks via drip tubing. This application rate was based on the median rainfall depth and duration in Maryland (Kreeb 2003). To analyze the experimental system under inundated conditions, a total rainfall depth of 4.6 cm was used for the simulations. At a 20:1 drainage area-to-SCM ratio, 92 cm (27 L) of total water was applied at a flow rate of 15 cm/h (72 mL/min) per stormwater simulation. For two stormwater events, the flow rate was altered to investigate effects of changing influent flow and volume on effluent nutrient concentrations with the same 6-h duration. The flow rate was halved to 7.5 cm/h (36 mL/min) for Storm 5 (resulting in 45 cm of water applied) and doubled to 30 cm/h (144 mL/min) for Storm 6 (180 cm of water applied). However, with only one event for each, inadequate data are available to discern any impacts and they are not discussed further.To determine long-term water quality effects, Mesocosm 30B (total of 20 weeks) and Mesocosm 30G (total of 37 weeks) were continued with tap water–only application and a flow rate 15 cm/h for 6 h. At 10 and 20 total weeks for 30B and 18 and 37 total weeks for 30G, a synthetic stormwater event (using the solution described in Table S1) was applied to these columns and effluent was measured for TN, TP, and their respective species.During applied stormwater events, samples were collected every 30 min for the first 2 h and every hour for the remaining 4 h. Eight to 10 samples were collected during each event. Additionally, samples of the influent and tap water used to prepare the influent were obtained once per storm. Effluent volume was not measured. All samples were immediately frozen until analysis. Before use, all glassware and sample bottles were washed with Alconox detergent, rinsed with tap water, then deionized water, soaked overnight in a 5 N hydrochloric acid (HCl) bath, rinsed with deionized water three final times, and left to air dry.Water QualityAll grab samples were tested for total N and P, and Samples 1, 4, 7, and the final grab sample were additionally filtered (0.22-μm filter) and tested for all N and P species. General comparisons between treatments were done using the mean effluent concentrations obtained per applied stormwater event, and error bars denote standard deviation unless otherwise stated.Total and dissolved P were analyzed by digesting the samples (APHA 2005) and then evaluated using the Murphy and Riley (1962) colorimetric method. Orthophosphate [soluble reactive phosphorus (SRP)] was analyzed with the Murphy and Riley (1962) colorimetric method without digestion. Particulate phosphorus (PP) was calculated by subtracting dissolved P from total P, and dissolved organic phosphorus (DOP) was calculated by subtracting orthophosphate from dissolved P.Total N was analyzed using a Shimadzu SSM-5000A (Columbia, Maryland) with Total Nitrogen Measuring Unit. Nitrite was measured with the 4500-NO2-B Colorimetric Method (APHA 2005); nitrate was measured using a Dionex ICS-1100 (Sunnyvale, California) with ASRS 4-mm suppressor and Dionex IonPac AS22 column. Ammonium was measured with the 4500-NH3-F Phenate Method (APHA 2005). Organic nitrogen (organic N) was calculated by subtracting all measured species (nitrate, nitrite, and ammonium) from the total N measurement.Any sample that measured above the range of the standard curve was diluted until the measurement was within the range of the standard curve. Samples that measured below the standard curve were considered below detection limit and presented as half of the lowest standard (USEPA 2000b). To ensure consistently accurate measurements deionized water and standards checks were done throughout the tests. Some samples below the detection limit were spiked with a known concentration to prevent reporting a false minimum value.Statistical AnalysisData were determined to be statistically different using the nonparametric Wilcoxon rank-sum test with the null hypothesis being that the two sets are from the same population (Hollander and Wolfe 1999). General increases or decreases in data trends were determined using the nonparametric Mann-Kendal tau test with the null hypothesis being that there is no trend in the data set along the x-axis (Gibbons 1985). All null hypotheses were rejected when the calculated critical value was greater than the 95% confidence interval (p<0.05).Results and DiscussionCompost Addition and TypeEffluent TN and TP mean concentrations are presented for all tested media and influent synthetic stormwater in Figs. 1 and 2, respectively (both log scale), as a function of applied stormwater for eight storm events. Discontinuities appeared when a new storm started. The nutrient concentrations in the column discharges varied widely based on the media mix used.The BSM-only demonstrated some removal for TN from the very first event and sustained TN removal throughout the study. TP concentrations in the effluent exceeded that of the influent for the first few samples of the first storm event, originating from the topsoil or mulch present in the BSM. After this, the TP concentrations in the bioretention column discharge were also less than the influent. On average, BSM had effluent concentrations of 0.12±0.09 mg-P/L for TP and 1.9±0.7 mg-N/L. These values are in the range found for field bioretention systems using media with low P contents (McNett et al. 2011). The BSM showed effective removal for both nutrients and was the only media mixture to consistently do this. These data indicate that use of the hardwood mulch as the organic amendment allows for N and P removal and does not produce nutrient leaching. This agrees with the work of Franklin et al. (2015), who found that mineralization of N from organic materials was inversely related to the carbon (C) to N ratio and that organic materials with high C:P ratios leached little P or immobilized P.According to the USEPA, reference conditions for ambient total N concentrations should not exceed 0.71 mg/L and total P concentrations should not exceed 31.25 μg/L for rivers and streams in the Eastern Coastal Plain (which includes Maryland) (USEPA 2000a). Other geographic areas have similar reference conditions.Initial TN concentrations exceeded 1,000 mg/L for both 15% and 30% biosolids compost addition; values ranged from 10 to 100 mg/L TN for the green-waste compost. Both 30% green-waste and biosolids composts discharged greater than 1 mg/L TP for the duration of the study. From influent concentrations, effluent concentration reduction was only found for the 15G+WTR for TP (32% reduction, p=0.02) and BSM (only) for both TP (70%, p<0.001) and TN (56%, p<0.001).With the addition of 15% compost to the BSM mix, effluent TN and TP concentrations (compared with traditional BSM) increased by 54-fold and 8.8-fold for biosolids, respectively, and 2.1-fold and 6.1-fold for green waste, respectively (all of which were statistically significant, p<0.001). These concentrations were 27-fold and 4.8-fold greater TN and TP, respectively, for biosolids and 1.1-fold and 3.0-fold greater, respectively, for green waste than influent concentrations (only TN concentrations for 15G were not found to be significantly different from influent concentrations, p=0.5).At 30% compost addition, biosolids effluent concentrations compared with BSM significantly increased (p<0.001) to 67-fold and 19-fold greater for TN and TP, respectively. Similarly, when compared with BSM-only, green-waste compost at 30% had 5.7-fold and 19-fold increase (p<0.001) in effluent concentrations for TN and TP, respectively. This equated to 25-fold and 6.8-fold increases compared with influent concentrations (TN and TP, respectively) for biosolids and 7.7-fold and 2.9-fold for green waste, respectively (all of which were statistically significant, p=0.001). From 15% compost to 30% compost (twofold increase), TN and TP effluent concentrations increased by a factor of 1.2 and 2.2 for biosolids and 2.7 and 3.0 for green waste, respectively. Leaching of N and P from compost-amended media under bioretention conditions has been noted by others (Iqbal et al. 2015; Mullane et al. 2015; Hurley et al. 2017). Hurley et al. (2017) observed that compost consisting of yard waste materials leached much less P than composts with manures, biosolids, and food waste.An initial flush phenomenon, most evident for TN discharge (Fig. 1), was found from all compost bioretention mesocosms. The initial flush occurred during each daily synthetic storm event, and also the mean concentrations of the first events were higher, falling with subsequent storms, both as noted previously by Mullane et al. (2015). Following the initial flush was a more consistent condition in which concentrations appeared to stabilize and become nearly constant with respect to time. For each compost amendment, constant leaching was achieved after a different depth of applied water; this corresponded to an initial flush of approximately 2 m for green waste and 4 m for biosolids compost. The difference in time to constant leaching was likely due to the difference in initial availability of nutrients; biosolids had 22 times KCl extractable N, 7.4 times CaCl2 extractable P, and 1.2 times greater Mehlich-3 P compared with green waste (Table 1).Although the stabilized concentrations found in the effluent from compost treatments were far lower than the initial flush values (11%–84% reduction from the first 2 m to the last 6 m of applied stormwater for green waste and 24%–94% reduction from the first 4 m to the last 4 m for biosolids), BSM stormwater effluent was still found to be statistically lower than all tested compost treatments (various compost amendments had increases of 2.0 to 24 times TN concentrations and 7.3 to 22 times TP concentrations). Sustained leaching of N and P from compost-amended (80% yard/20% food) bioretention media has also been reported by Mullane et al. (2015).Further evidence of both the green waste and biosolids compost reaching conditions of constant nutrient release can be found in Fig. 3, showing up to 45 m of applied water for both 30B and 30G. Both compost-amended columns showed evidence of a storm-related initial flush (except the ∼45 m storm for 30% green waste); however, concentrations on average reached very near influent values. Compared with influent concentrations (3.9±0.5 mg-N/L and 0.34±0.06 mg-P/L for TN and TP, respectively), average constant-leaching concentrations of TN for 30G and 30%B were 5.8±1.3 and 4.5±1.4 mg-N/L, and for TP were 0.65±0.16 and 0.84±0.12 mg-P/L, respectively. Although closer to influent concentrations, these values still represent leaching of nutrients and not stormwater treatment.The initially high nutrient concentrations found in the effluent from compost-amended mesocosms, compared with influent concentrations and BSM, were largely due to the greater initial availability of N and P in the composts, as evidenced by their respective extractions (Table 1). Compared with BSM, biosolids compost had 882 and 3,721 times higher extractable N and P; green-waste compost had 172 and 723 times greater extractable N and P compared with BSM. Compost mineralization will release N and P initially at high concentrations, but decreasing with time (Claassen and Carey 2007; Franklin et al. 2015; Al-Bataina et al. 2016). Franklin et al. (2015) noted that the masses of P leached from composts were directly proportional to the total amount of P present (via acid digestion).Additionally, high initial nutrient values may be linked to fine particulates escaping from the system, evidenced in the N and P speciation data in Fig. 4. The initial flush N species (α columns in Fig. 4) for green waste (first 2 m of applied stormwater) and biosolids (4 m of applied water) consisted largely of ammonium and organic N, compared with the steady state (β and γ columns in Fig. 4), which were largely nitrate. For phosphorus, PP dominated in the initial flush and SRP in the steady state. Compost fines were initially washed from the mesocosm, leading to the dominance of the ammonium, particulate N (organic N), and PP, as also noted by Iqbal et al. (2015) and Mullane et al. (2015) in their study of yard/food waste compost mixes. These prior studies noted the largest fraction of particles in the 1–11-μm range.As the study progressed, nitrate and SRP increased in fraction of total nutrient effluent, whereas particulate associated species concentrations were reduced. Organic matter mineralization in compost leads to the formation of ammonium (via ammonification) and nitrate (through nitrification) (Al-Bataina et al. 2016). The nutrient export noted in the long-term studies was likely due to the slow release of N and P through compost organic matter mineralization over time. Although not a bioretention study, Mangum et al. (2020) also found substantial reductions in organic N and PP with increased synthetic stormwater application during investigation of submerged gravel wetlands amended with compost.Media ModificationsTo examine possible ways to curb compost nutrient leaching, two media modifications were analyzed: WTR addition to 15G (15G+WTR) and initial compost washing of the 30B (30WB). Washing the biosolids compost was a mechanical measure to reduce readily leachable concentrations found in the initial flush of nutrients; the addition of WTR was intended to offer additional complexing sites for orthophosphate removal from infiltrating stormwater, thus reducing P concentration in the effluent.Both compost modifications resulted in reductions in mean discharge TP concentrations compared with their unmodified counterpart, but only 30WB showed reductions in mean TN concentrations. Washing biosolids compost, on average, reduced concentrations by 19% (from 2.4±1.4 mg-P/L with 30B to 1.9±0.9 mg-P/L for 30WB; p=0.002) for TP and 47% (p=0.002) for TN (from 176±395 to 76±80 mg-N/L, respectively). Green waste TP concentrations were reduced by 61% by WTR addition (from 0.69±0.26 to 0.27±0.35 mg-P/L for 15G and 15G+WTR, respectively). Although 15G+WTR did not show initial reduction in TN concentration compared with 15G (average in first 2 m of 9.8±12 mg-N/L compared with 5.3±2.5 mg-N/L, respectively; p=0.71), the constant-leaching phase (last 6 m) showed a statistically significant concentration reduction of 34% (from 3.5±0.6 mg-N/L for 15G to 2.3±3.3 mg-N/L for 15G+WTR; p<0.001). Other studies have similarly noted that the incorporation of Al-based WTR or other iron- and aluminum-based sorptive media decreased P leaching from bioretention media amended with from 10% to 40% compost (Palmer et al. 2013; Shrestha et al. 2018; Poor et al. 2019; Ament et al. 2021, 2022).Although compost modifications, washing for biosolids, and WTR addition for green waste resulted in reduced nutrient export compared with their nonmodified counterpart, the BSM-only treatment consistently had the lowest effluent nutrient concentrations. With mean BSM effluent concentrations of 0.12±0.09 mg/L TP and 1.9±0.7 mg/L TN, the BSM (which has organic matter from the topsoil and wood mulch, but no added compost) (Table 1) dischargers were 58% and 55% lower than 15G+WTR and 94% and 98% lower than 30WB, respectively.Changes in N and P species corresponding to the nutrients reductions for the two media modifications (30B to 30WB and 15G to 15G+WTR) can be seen in Fig. 4. Although both modifications showed some form of reduction in nutrient effluent concentrations for their respective compost type, the dominant phase (initial flush vis-à-vis constant leaching) was different. Comparing 30B with 30WB, reductions in nutrients concentrations was noted largely in the initial flush (α in Fig. 4), as expected. Here, nitrate and ammonium concentrations were found to be 66% (p=0.11) and 70% (p=0.13) lower on average than from the unwashed media, respectively. SRP and PP were 25% (p=0.004) and 40% (0.07) lower, respectively.Comparatively, within the initial flush, 15G to 15G+WTR did not show reductions for TN and smaller reductions for TP (compared with the constant-leaching phase). Within the constant-leaching phase, however, the WTR modification to the green-waste compost resulted in reduced effluent concentrations for all measured N species (except for nitrite, which was often found below the detection limit) (73%, p=0.03 for ammonium; 54%, p<0.001 for nitrate; and 34%, p=0.07 for organic N) and P species (84%, p<0.001 for SRP; 79%, p<0.001 for DOP; and 81%, p=0.001 for PP). The release of PP during the initial flush was not affected by the WTR addition because this release occurred through washout of fine compost particles causing the high initial flush nutrients release even with the WTR addition. The compost and WTR were mixed throughout the entire depth of the media. Layering of media to capture particulate and dissolved P released from the compost may be more effective in overall P removal (Poor et al. 2019; Ament et al. 2021).The difference in when (initial flush via-a-vis constant leaching) reductions in nutrient effluent concentrations occurred was largely due to the controlling media modification mechanism. Washing of the biosolids compost was designed to focus on initial nutrient reduction by removing easily leachable nutrients and resulted in initial drops in nutrient availability (Table 1) of 63% KCl extractable N, 97% CaCl2 extractable P, and 70% Mehlich-3 extractable P. This pretreatment resulted in the larger reductions noted within the initial flush. However, the addition of WTR to green-waste compost was designed to affect the long-term release of P and resulted in the greater reduction during the constant-leaching phase.Similar reductions in ortho- and dissolved organic P were found by Liu and Davis (2014) with the addition of Al-based WTR to a field bioretention cell, most likely via enhanced P adsorption to the charged WTR surfaces. Additionally, WTR reductions of TN may have been a result of increased cation exchange capacity due to WTR amendment, which has been shown to increase NH4-N and glycine (organic N) retention (Gallimore et al. 1999; O’Neill and Davis 2012).Despite the nutrient reductions measured with the media modifications compared with compost amendment alone, unamended BSM (with topsoil and wood mulch, but no compost) continued to discharge nutrient concentrations well below the amended materials. Compared with BSM, 15G+WTR concentrations were twofold and threefold higher, and for 30WB, concentrations were 40-fold and 16-fold greater for TN and TP, respectively. The better performance seen in the BSM system compared with the compost-amended media was largely due to the initial flush differences. Compost-amended bioretention systems continued to have nutrient leaching within the first few applied stormwater events. However, based on longer-term constant leaching averages, green waste with WTR effluent was comparable to BSM with effluent TN concentrations 1.25 times greater and TP concentrations 1.33-fold greater.From a practical perspective, the compost washing process is not a feasible treatment. It was labor-intensive and produced a wash water that was very high in nutrients that needed subsequent treatment/disposal (total N = 3,640±1,600 mg/L, predominantly ammonium; total P = 27±5.8 mg/L, mostly organic P).Compost use as a bioretention amendment may depend on media layering and configuration. Several studies of compost/BSM mixtures showed poor performance with nutrient leaching (Mullane et al. 2015; Hurley et al. 2017; Shrestha et al. 2018), as was seen in this study. However, previous studies have used layering techniques or additional media amendments (e.g., biochar and WTR) with composted material and have demonstrated improved effluent results compared with the traditional BSM mixture (Palmer et al. 2013; Poor et al. 2019; Shrestha et al. 2018).Nutrient Mass ExportA major benefit to compost incorporation into the BSM design is improved water-holding capacity (Kirchoff et al. 2003). Although not measured in this study, with increased water-holding capacity, compost has the potential to reduce the discharge volume from the incoming stormwater. Using mass as the metric may make the export of nutrients between compost-amended media and BSM more comparable. Conceptually, with characteristics of a sandy loam soil, BSM has an expected field capacity of 15%–30% depending on silt content; however, biosolids compost and green-waste compost have field capacity values of 80% and 84%, respectively (Duzgun et al. 2021). With a media volume of 22 L, BSM is expected to reduce influent volumes by 3.3–6.6 L per storm event (mean of 4.5 L). Media with 15% compost, based on water-holding capacity, would reduce effluent water by 5.4–8.4 L (6.9 L mean) per storm event, and 30% compost would reduce effluent volume by 7.6–10.2 L per storm event (8.9 L mean). However, shorter antecedent drying periods and/or poorly draining surrounding native/site soils may reduce the maximum capacity of compost for water retention.Based on the preceding mean water-holding capacity values, calculated effluent mass export of TN and TP are presented in Fig. 5. Through the duration of this study, only BSM and green waste showed similar (30G, at 45 m) or reduced (15G, 24% at 8 m; 15G+WTR, 22% at 8 m; and BSM, 55% at 8 m) export of total N mass compared with untreated influent synthetic stormwater, and only the BSM (71% at 8 m) and 15G+WTR (37% at 8 m) had reduced TP export. However, all compost treatments were found to have a consistent decrease in nutrient leaching with time, followed by a relatively constant nutrient release.Linearly projecting the constant nutrient leaching phase measured in this study indicates that 30G and 15B would begin to reduce influent TP mass export at 170 and 175 m of applied water, respectively. With an average annual rainfall in Maryland of 1.11 m and a drainage area of 1:20, these values are equivalent to 7.6 and 7.8 years for 30G and 15B, respectively. However, it is not clear that TP reductions would occur if no sorption sites would be available on the media. Extrapolations indication that 15G, 30B, and 30WB would never reduce influent TP mass export.Linear projections also showed 30G and 30B begin to reduce influent TN mass export at 46 m (2.1 years) and 1,035 m (46 years), respectively, whereas 15B and 30WB never reduce TN mass export. Nitrogen removal could occur via biological and/or plant-induced mechanisms. All of these projections represent a significant leaching time before reaching a state of nutrients removal.Nutrients and Plant GrowthThe tall fescue grew well in all media. The grasses increased in aboveground height (Fig. S1) and the roots grew deep into the media. The growth was not quantified, but it could be seen through the acrylic columns.Recommendations for bioretention system surface areas have been found to range from 3% to 43% of the drainage area (Stander et al. 2010), but around 5% to 10% appears most common (Li et al. 2009; Brown et al. 2015). This leads to ratios of rainfall-runoff of approximately 1∶10 to 1∶20. Even though stormwater N and P concentrations are moderately low, as used in this study, because of the greater stormwater volume that is directed to the bioretention systems, N and P mass loads are generally high enough to support vegetation growth without any addition of nutrient sources such as compost or chemical fertilizers. Tall fescue annual nutrient requirements are approximately 98 kg N/(ha-year) (2 lb N/1,000 sq ft) to 147 kg N/(ha-year) (3 lb N/1,000 sq ft) and 11 kg P/(ha-year) (0.5 lb P2O5/1,000 sq ft) to 21 kg P/(ha-year) (1 lbP2O5/1,000 sq ft) (Turner 2013). In our column study, the TN and TP loads (all dissolved) were approximately 790 and 39 hg/(ha-year), respectively.ConclusionsThis research was designed to determine the implications of compost addition as a bioretention media modification by investigating nutrient dynamics and identifying critical elements related to nutrient retention and release in bioretention systems. Additionally, this study sought to explore media modifications supplemental to compost application for nutrient reductions and to shed light on nutrient release mitigation potential. The addition of compost to BSM had significant negative effects on the release of N and P from the bioretention media. Although green waste treatments generally performed better than biosolids in effluent TN (1–2 orders of magnitude lower) and TP (2%–31% lower) concentrations, BSM without compost (with wood mulch as an organic amendment) consistently outperformed all media amended with compost (53% to 99% lower TN concentrations and 83% to 95% lower TP concentrations, not including the modified media).Major factors affecting the increased release of nutrients from composted amendments vis-à-vis BSM could be linked to initial extractable nutrients and organic matter content. Composted material had 172-fold to 3,721-fold higher extractable N and P and up to 39% greater organic matter content than BSM. The greater nutrient availability was evidenced in an initial flush of nutrients in which the concentrations were 10% to 91% greater than the subsequent constant-effluent-leaching concentrations, with the highest concentration flushes occurring for nitrogen from 30B in early events. This flush was a large contributor to the increased concentrations found in all compost treatments compared with BSM.After the initial flush of nutrients seen in the treatments, a continued steady-state release of nutrients dominated by nitrate and orthophosphate was measured. In this constant-leaching phase, it appears that continued break down of organic material through aerobic mineralization led to the increased nutrient concentrations found in compost-amended media compared with BSM. A mass-based examination of nutrient export through total volume capture may yield more comparable effluent results based on potentially 30%–140% reduction in effluent volume due to improved water-holding capacity from composted material versus the traditional sandy loam BSM.The mechanical washing of biosolids resulted in reduced extractable N (KCl) and P (CaCl2 and Mehlich-3), leading to reduced concentrations of both labile N and P; however, the benefits of these reductions were largely seen in the initial flush of TN, producing comparable TN export to 15B, whereas TP export was virtually unchanged. The addition of WTR significantly improved TP retention (72% reduction with WTR) and slightly improved TN retention in the constant-leaching phase (34% reduction) for the tested green-waste compost.Based on results of this research, amending bioretention with biosolids or green-waste compost resulted in leaching of nutrients and cannot be recommended if N and P leaching into stormwater is a concern. Additional amendments to the bioretention media (washing or WTR addition) reduced leaching of nutrients but did not eliminate it nor did it improve performance over that of BSM alone.References Ament, M., S. Hurley, M. Voorhees, E. Perkins, Y. Yuan, J. Faulkner, and E. Roy. 2021. “Balancing hydraulic control and phosphorus removal in bioretention media amended with drinking water treatment residuals.” ACS ES&T Water 1 (3): 688–697. https://doi.org/10.1021/acsestwater.0c00178. Ament, M. R., E. D. Roy, Y. Yuan, and S. H. Hurley. 2022. “Phosphorus removal, metals dynamics, and hydraulics in stormwater bioretention systems amended with drinking water treatment residuals.” J. Sustainable Water Built Environ. 8 (3): 04022003. https://doi.org/10.1061/JSWBAY.0000980. APHA (American Public Health Association). 2005. Standard methods for the examination of water and wastewater. 22nd ed. Washington, DC: APHA. Brown, R. A., T. P. O’Connor, and M. Borst. 2015. “Divergent vegetation growth patterns relative to bioinfiltration unit size and plant placement.” J. Sustainable Water Built Environ. 1 (3): 04015001. https://doi.org/10.1061/JSWBAY.0000796. Castle, S. 2009. Soil inorganic nitrogen: KCl extraction. Boulder, CO: Univ. of Colorado Boulder Aridlands Ecology Lab Protocol. Claassen, V. P., and J. L. Carey. 2007. “Comparison of slow-release nitrogen yield from organic soil amendments and chemical fertilizers and implications for regeneration of disturbed sites.” Land Degrad. Dev. 18 (2): 119–132. https://doi.org/10.1002/ldr.770. Davis, A. P., and R. H. McCuen. 2005. Stormwater management for smart growth. New York: Springer. Davis, A. P., M. Shokouhian, H. Sharma, and C. Minami. 2001. “Laboratory study of biological retention for urban stormwater management.” Water Environ. Res. 73 (1): 5–14. https://doi.org/10.2175/106143001X138624. Davis, A. P., M. Shokouhian, H. Sharma, C. Minami, and D. Winogradoff. 2003. “Water quality improvement through B: Lead, copper, and zinc removal.” Water Environ. Res. 75 (1): 73–82. https://doi.org/10.2175/106143003X140854. Franklin, D., D. Bender-Özenç, N. Özenç, and M. Cabrera. 2015. “Nitrogen mineralization and phosphorus release from composts and soil conditioners found in the Southeastern United States.” Soil Sci. Soc. Am. J. 79 (5): 1386–1395. https://doi.org/10.2136/sssaj2015.02.0077. Gallimore, L. E., N. T. Basta, D. E. Storm, M. E. Payton, R. H. Huhnke, and M. D. Smolen. 1999. “Water treatment residual to reduce nutrients in surface runoff from agricultural land.” J. Environ. Qual. 28 (5): 1474–1478. https://doi.org/10.2134/jeq1999.00472425002800050012x. Gibbons, J. D. 1985. Nonparametric statistical inference. 2nd ed. New York: Marcel Dekker. Glass, C., and S. Bissouma. 2005. “Evaluation of a parking lot bioretention cell for removal of stormwater pollutants.” WIT Trans. Ecol. Environ. 81: 10. https://dooi.org/ 10.2495/ECO050691. Herrera Environmental Consultants. 2012. Pollutant export from bioretention soil mix; 185th Avenue NE. Redmond, WA: Herrera Environmental Consultants. Hollander, M., and D. A. Wolfe. 1999. Nonparametric statistical methods. 2nd ed. New York: Wiley-Interscience. Hurley, S., P. Shrestha, and A. Cording. 2017. “Nutrient leaching from compost: Implications for bioretention and other green infrastructure.” J. Sustainable Water Built Environ. 3 (3): 04017006. https://doi.org/10.1061/JSWBAY.0000821. Iqbal, H., M. Garcia-Perez, and M. Flury. 2015. “Effect of biochar on leaching of organic carbon, nitrogen, and phosphorus from compost in bioretention systems.” Sci. Total Environ. 521–522 (Jul): 37–45. https://doi.org/10.1016/j.scitotenv.2015.03.060. Kirchoff, C. J., J. F. Malina, and M. E. Barrett. 2003. Characteristics of composts: Moisture holding and water quality improvement. CRWR Online Rep. No. 03-09. Austin, TX: Center for Research in Water Resources, Univ. of Texas at Austin. Kreeb, L. B. 2003. “Hydrologic efficiency and design sensitivity of bioretention facilities. College Park, MD: Honor’s Research, Univ. of Maryland. LeFevre, G. H., K. H. Paus, P. Natarajan, J. S. Gulliver, P. J. Novak, and R. M. Hozalski. 2015. “Review of dissolved pollutants in urban storm water and their removal and fate in bioretention cells.” J. Environ. Eng. 141 (1): 04014050. https://doi.org/10.1061/(ASCE)EE.1943-7870.0000876. Li, H., and A. P. Davis. 2008. “Heavy metal capture and accumulation in bioretention media.” Environ. Sci. Technol. 42 (14): 5247–5253. https://doi.org/10.1021/es702681j. Liu, J., and A. P. Davis. 2014. “Phosphorus speciation and treatment using enhanced phosphorus removal bioretention.” Environ. Sci. Technol. 48 (1): 607–614. https://doi.org/10.1021/es404022b. Liu, J., D. Sample, C. Bell, and Y. Guan. 2014. “Review and research needs of bioretention used for the treatment of urban stormwater.” Water 6 (4): 1069–1099. https://doi.org/10.3390/w6041069. Lucas, W., and M. Greenway. 2011. “Phosphorus retention by bioretention mesocosms using media formulated for phosphorus sorption: Response to accelerated loads.” J. Irrig. Drain. Eng. 137 (3): 144–153. https://doi.org/10.1061/(ASCE)IR.1943-4774.0000243. Mangum, K. R., Q. Yan, T. K. Ostrom, and A. P. Davis. 2020. “Nutrient leaching from green waste compost addition to stormwater submerged gravel wetland mesocosms.” J. Environ. Eng. 146 (3): 04019128. https://doi.org/10.1061/(ASCE)EE.1943-7870.0001652. McNett, J. K., W. F. Hunt, and A. P. Davis. 2011. “Influent pollutant concentrations as predictors of effluent pollutant concentrations for mid-Atlantic bioretention.” J. Environ. Eng. 137 (9): 790–799. https://doi.org/10.1061/(ASCE)EE.1943-7870.0000373. MDOT-SHA. 2022. Maryland department of transportation state highway administration standard specifications for construction and materials. Baltimore, MD: MDOT. Mitchell, D. 1997. “State highway departments find it pays to use compost, Part II.” Bio Cycle 38 (8): 67–72. Moore, P. A., Jr., and B. C. Joern. 2009. “Water- or dilute salt-extractable phosphorus in soil.” In Methods of phosphorus analysis for soils, sediments, residuals, and waters, edited by M. L. Self-Davis. 2nd ed. Blacksburg, VA: Southern Cooperative. Mullane, J., M. Flury, H. Iqbal, P. Freeze, C. Hinman, C. Cogger, and Z. Shi. 2015. “Intermittent rainstorms cause pulses of nitrogen, phosphorus, and copper in leachate from compost in bioretention systems.” Sci. Total Environ. 537 (Dec): 294–303. https://doi.org/10.1016/j.scitotenv.2015.07.157. Pamuru, S. T., E. Forgione, K. Croft, B. V. Kjellerup, and A. P. Davis. 2022. “Chemical characterization of urban stormwater: Traditional and emerging contaminants.” Sci. Total Environ. 813 (Mar): 151887. https://doi.org/10.1016/j.scitotenv.2021.151887. Park, M.-H., X. Swamikannu, and M. K. Stenstrom. 2009. “Accuracy and precision of the volume-concentration method for urban stormwater modeling.” Water Res. 43 (11): 2773–2786. https://doi.org/10.1016/j.watres.2009.03.045. Paus, K., J. Morgan, J. Gulliver, and R. Hozalski. 2014. “Effects of bioretention media compost volume fraction on toxic metals removal, hydraulic conductivity, and phosphorus release.” J. Environ. Eng. 140 (10): 04014033. https://doi.org/10.1061/(ASCE)EE.1943-7870.0000846. Poor, C., K. Conkle, A. MacDonald, and K. Duncan. 2019. “Water treatment residuals in bioretention planters to reduce phosphorus levels in stormwater.” Environ. Eng. Sci. 36 (3): 265–272. https://doi.org/10.1089/ees.2018.0254. Shrestha, P., S. Hurley, and B. Wemple. 2018. “Effects of different soil media, vegetation, and hydrologic treatments on nutrient and sediment removal in roadside bioretention systems.” Ecol. Eng. 112 (Mar): 116–131. https://doi.org/10.1016/j.ecoleng.2017.12.004. Skorobogatov, A., J. He, A. Chu, C. Valeo, and B. van Duin. 2020. “The impact of media, plants, and their interactions on bioretention performance: A review.” Sci. Total Environ. 715: 136918. https://doi.org/10.1016/j.scitotenv.2020.136918. Stander, E. K., M. Borst, T. P. O’Connor, and A. A. Rowe. 2010. “The effects of rain garden size on hydrologic performance.” In Proc., Environmental and Water Resources Institute (EWRI) World Environmental and Water Resources Congress. Reston, VA: ASCE. USEPA. 2000a. Ambient water quality criteria recommendations information supporting the develop of state and tribal nutrient criteria for rivers and streams in Nutrient Ecoregion XIV Eastern Coastal Plain. EPA-822-B-00-022. Washington, DC: United States Environmental Protection Agency. USEPA. 2000b. Assigning values to non-detected /non-quantified pesticide residues in human health food exposure assessments. Washington, DC: Office of Pesticide Programs, US Environmental Protection Agency. USEPA. 2000c. Low impact development (LID), a literature review. EPA-841-B-00-005. Washington, DC: United States Environmental Protection Agency. Weng, L., E. J. Temminghoff, S. Lofts, E. Tipping, and W. H. Van Riemsdijk. 2002. “Complexation with dissolved organic matter and solubility control of heavy metals in a sandy soil.” Environ. Sci. Technol. 36 (22): 4804–4810. https://doi.org/10.1021/es0200084.