Modeling benefits and tradeoffs of green infrastructure: Evaluating and extending parsimonious models for neighborhood stormwater planning

Green infrastructure is often proposed to complement conventional urban stormwater management systems that are stressed by extreme storms and expanding impervious surfaces. Established hydrological and hydraulic models inform stormwater engineering but are time- and data-intensive or aspatial, rendering them inadequate for rapid exploration of solutions. Simple spreadsheet models support quick site plan assessments but cannot adequately represent spatial interactions beyond a site. The present study builds on the Landscape Green Infrastructure Design (L-GrID) Model, a process-based spatial model that enables rapid development and exploration of green infrastructure scenarios to mitigate neighborhood flooding. We first explored how well L-GrID could replicate flooding reports in a neighborhood in Chicago, Illinois, USA, to evaluate its potential for green infrastructure planning. Although not meant for prediction, L-GrID was able to replicate the flooding reported and helped identify strategies for flood control. Once evaluated for this neighborhood, we extended the model to include water quality through the representation of dispersion and settling mechanisms for two pollutant surrogates-total nitrogen and total suspended solids. With the extended model, Landscape Green Infrastructure Design Model-Water Quality (L-GrID-WQ), we examined benefits, costs, and tradeoffs for different green infrastructure strategies. Bioswales were slightly more effective than other green infrastructure types in reducing flooding extent and downstream runoff and pollution, through increased infiltration and settling capacity. Permeable pavers followed in effectiveness and are suggested where spatial constraints may limit the installation of bioswales. Although green infrastructure supports both flooding and pollution control, small tradeoffs between these functions emerged across spatial layouts: strategies based on only curb-cuts better controlled pollution, while layouts that followed the path of water flow better controlled flooding. By illuminating such tradeoffs, L-GrID-WQ can support green infrastructure planning that prioritizes unique concerns in different areas of a landscape.

Interactions of particulate- and dissolved-phase heavy metals in a mature stormwater bioretention cell

Bioretention is an increasingly common stormwater control measure (SCM) for mitigation of stormwater quantity and quality. Studies from lab to field scale have shown successful removal of total metals from stormwater, especially Cu and Zn which are ubiquitous in the urban environment yet detrimental to aquatic ecosystems. While bioretention effectively removes particulate matter and particulate bound (PB) contaminants, removal performance of dissolved metals has been neglected in field studies. After approximately two decades of these systems being implemented, with a typical design-life of 20 years, performance of mature systems is unknown. This study examined the performance of a 16- to 18-year-old bioretention cell by characterizing Cu and Zn partitioning and removal. Flow-weighted composite samples of stormwater and bioretention effluent were collected and analyzed for total and dissolved metals. Size-fractioned road-deposited sediments (RDS) were collected and analyzed for metals and particle size distribution. The comparison of RDS and PB metals showed that PB-Zn was enriched in stormwater, indicating higher mobility of PB-Zn compared to PB-Cu. The mature bioretention system effectively removed particulates and PB-metals with average load reductions of 82% and 83%, respectively. While concentrations for dissolved metals were low (<40 μg/L), no significant difference between influent and effluent was observed. Effluent concentrations of total and dissolved Cu, total organic carbon, and particulates were not significantly different from those measured over 10 years ago at the site, while total Zn effluent concentration slightly increased. MINTEQ speciation modeling showed that Cu was approximately 100% bound with dissolved organic matter (DOM) in all bioretention effluent. While Zn was also mostly bound with DOM in effluent, some events showed free ionic Zn reaching concentrations in the same order of magnitude. Media amendments, maintenance, and monitoring of SCMs should be considered where further removal of dissolved metals is necessary for the protection of aquatic environments.

Sources of variation in nutrient loads collected through street sweeping in the Minneapolis-St. Paul Metropolitan Area, Minnesota, USA

Excess non-point nutrient loading continues to impair urban surface waters. Because of the potential contribution of tree litterfall to nutrient pollution in stormwater, street sweeping is a promising management tool for reducing eutrophication in urban and suburban regions. However, nutrient concentrations and loads of material removed through street sweeping have not been well characterized, impeding the development of pollution reduction credits and improvement of models for stormwater management. We evaluated the role of canopy cover over streets, street sweeper type, season, and sweeping frequency in contributing to variation in concentrations and loads of nitrogen (N), phosphorus (P), and solids recovered in street sweepings, using analyses of samples collected during regular street sweeping operations in five cities in the Minneapolis-St. Paul Metropolitan Area, Minnesota, USA. We expected that nutrient concentrations and loads would be highest in seasons and places of higher tree litterfall. We also expected that regenerative-air sweepers would recover higher loads compared to mechanical broom sweepers. Total N and P concentrations in sweepings increased most strongly with canopy cover in June, October, and November. Total N and P recovered in street sweepings similarly increased with canopy cover in June, October, and November, and peaked in early summer and autumn, times of high litterfall. In contrast, total dry mass in sweepings was greatest in early spring, following winter snowmelt. However, nutrient loads and concentrations did not differ between sweeper types. Our results add to growing evidence of the importance of street trees in contributing nutrient pollution to urban surface waters. Street sweeping focused on high-canopy streets during early summer and autumn is likely an effective management tool for stormwater nutrient pollution.

Land use and rainfall influences on bacterial levels and sources in stormwater ponds

Urban stormwater runoff is a known source of microbial contamination of stormwater ponds. However, less is known about the influences of land use and rainfall on microbial quality over time in these receiving waters. In this study, two fecal indicator bacteria (FIB), namely Escherichia coli and thermotolerant coliforms, were monitored in three stormwater ponds in Calgary, Alberta, Canada. The stormwater ponds were selected due to their potential as water sources for beneficial uses such as irrigation, which requires lower water quality than drinking water, thereby alleviating the pressure on the city's potable water demands. The selected stormwater ponds vary in size and shape, contribution catchment size, and percentages of several primary land use types. Microbial source tracking for human, dog, seagull, Canada goose, ruminant, and muskrat was also conducted to determine sources of bacterial contamination in the stormwater ponds. Sampling was conducted near the pond surface and adjacent to the shoreline, specifically near the outfalls that discharge stormwater runoff into the ponds and the inlets that convey water out of the ponds. Overall, the FIB concentrations in the vicinity of pond outfalls were significantly or relatively higher than those near pond inlets. The contamination in the McCall Lake and the Country Hills stormwater ponds showed higher amounts of human markers (40 to 60%) compared to the Inverness stormwater pond (< 20%), which coincided with their higher FIB concentration medians. The results revealed that stormwater drained from catchments with a higher percentage of commercial land use was more contaminated than those with primary residential land use, while the impacts of residential development on the FIB levels in the Inverness stormwater pond were not obvious. Furthermore, FIB concentrations in the ponds increased in response to both rain events and inter-event dry periods, with human-specific markers being predominant despite the high levels of animal markers during inter-event dry periods. Human-origin sources might be among the main microbial loading contributors in the pond catchments in general. All these findings can inform the development or improvement of measures for mitigating microbial pollution, strategies for reusing stormwater, and maintenance programs.

How effective is the retention of microplastics in horizontal flow sand filters treating stormwater?

Microplastics accumulate in stormwater and can ultimately enter freshwater recipients, and pose a serious risk to aquatic life. This study investigated the effectiveness of lab-scale horizontal flow sand filters of differing lengths (25, 50 and 100 cm) in retaining four types of thermoplastic microplastics commonly occurring in stormwater runoff (polyamide, polyethylene, polypropylene, and polyethylene terephthalate). Despite the differences in particle shape, size and density, the study revealed that more than 98% of the spiked microplastics were retained in all filters, with a slightly increased removal with increased filter length. At a flow rate of 1 mL/min and after one week of operation, 62-84% of the added microplastics agglomerated in the first 2 cm of the filters. The agglomerated microplastics included 96% of high-density fibers. Larger-sized particles were retained in the sand media, while microplastics smaller than 50 μm were more often detected in the effluent. Microplastics were quantified and identified using imaging based micro Fourier Transform Infrared Spectroscopy. The efficient retention of microplastics in low-flow horizontal sand filters, demonstrated by the results, highlights their potential importance for stormwater management. This retention is facilitated by various factors, including microplastic agglomeration, particle sedimentation of heavy fibers and favorable particle-to-media size ratios.

Performance monitoring of constructed floating wetlands: Treating stormwater runoff during the construction phase of an urban residential development

In the context of climate change and global trend towards greenfield urbanisation, stormwater and transported pollutants are expected to increase, impairing receiving environments. Constructed floating wetlands (CFWs) can improve stormwater retention pond performance. However, performance data are currently largely restricted to mesocosm experiments, limiting design enhancement fit for field implementation. The present 12-month field study aims to fill part of these gaps by identifying limitations and necessary design improvements for CFWs on a large retention pond/lake. Water in a 2.6-ha lake receiving stormwater from a 45-ha urban area under development in subtropical Queensland, Australia, was recirculated during dry weather periods to minimise algal growth and the risks of blooms. Pollutant removal efficiencies of two full-scale CFWs were evaluated during storm events and dry weather periods as a function of inlet and outlet pollutant concentrations, flow and rainfall. Inlet TSS and TN concentrations in runoff during the construction phase of the development exceeded required water quality limits while TP inflow concentrations were low and often below the detection limit. Median pollutant load reduction efficiencies during storm-events were - 20 % TSS, -2 % TN and 22 % TP at CFW1 and 51 % TSS, 3 % TN and 17 % TP at CFW2, respectively. TSS and TN concentration removal efficiencies at CFW1 were low and highly variable, partly due to low inlet concentrations, high flow velocities and short hydraulic retention times (<1 day). However, CFW1 significantly reduced TSS concentrations during dry weather periods. In contrast, CFW2 significantly reduced TSS concentrations during both storm events and during inter-event periods. This study highlights treatment limitations associated to the operational conditions of CFWs at field-scale not identifiable in a mesocosm-scale study. Further research is necessary to investigate treatment performance of CFWs during the operational phase of the development with higher nutrient levels.

Transport and sources of nitrogen in stormwater runoff at the urban catchment scale

Revealing the transport and sources of nitrate in urban stormwater runoff can effectively manage nitrate pollution in urban areas. This study used the chemical properties of stormwater along with δ¹⁵N-NO₃^- and δ¹⁸O-NO₃^- isotopes to identify the transport and sources of nitrate within an urban catchment. The results showed that the NO₃-N concentration and total dissolved nitrogen (TDN) composition differed among roof runoff, road runoff, and drainage runoff. The highest NO₃^-N concentration was found in roof runoff and NH₃-N dominated the TDN composition. However, the erosion of pervious surfaces and litter may have led to higher DON/TDN values in road runoff. The TDN composition of drainage runoff was consistent with that of roof runoff. Furthermore, among the various rainfall characteristics, the depth and intensity of rainfall were significantly correlated with the NO₃-N concentrations in roof runoff and road runoff, while antecedent dry days had little effect. According to a Bayesian mixing model, the average contributions of the nitrate load in drainage runoff were ranked as road runoff (51.6%) > rainwater (29.2%) > and roof runoff (15%), which is consistent with the results of previous studies. Rainwater nitrate may have ranked second due to the confluence time, pollution level, and other factors that made rainwater reduce the pollution characteristics of roof runoff. The dominant contribution of road runoff to the NO₃^-N concentration of drainage runoff could be attributed to the large runoff volume. Hence, effective measures should be taken to minimize the NO₃-N concentration in roof runoff, while runoff volume reduction should be the primary concern for controlling road runoff pollution. This work is helpful for obtaining a better understanding of the transport and sources of nitrate that vary dynamically within different hydrological flow pathways, and the outcomes are expected to enhance targeted measures to mitigate nitrate pollution in urban water systems.

Low-cost biomonitoring and high-resolution, scalable models of urban metal pollution

As the global toll on human lives and ecosystems exacted by urban pollution grows, planning tools still lack the resolution to identify priority sites where toxic pollution can be most efficiently averted at a spatial scale that matches funding and management. Here we tackle this gap by demonstrating novel scalable methods to monitor and predict urban metal pollution at high resolution (<5 m) across large areas (10,000-100,000 km²) to guide pollution reduction and stormwater management. We showcase and calibrate predictive models of Zn, Cu, and a synthetic index of pollution for the Puget Sound region of Washington State, U.S., a densely urbanized yet important ecosystem of conservation interest, and exemplify their transferability across the entire United States. We leveraged widely and freely available datasets of car traffic characteristics and land use as predictor variables and trained the models with biological monitoring data of metal concentrations in epiphytic moss from >100 trees based on new rapid and low-cost protocols introduced in this study. Our model predictions, showing that 50% of the total Cu and Zn pollution across the Puget Sound watershed is deposited over only 3.3% of the land area, will allow cities to effectively and efficiently target toxic hotspots.

Rapid plant responses following relocation of a constructed floating wetland from a construction site into an urban stormwater retention pond

This study compared plant growth, nutrient partitioning and total nutrient uptake by tall sedge (Carex appressa) plants in large-scale Constructed Floating Wetlands (CFWs). Two CFWs with a total area of 2088 m² were installed in a 2.6 ha man-made urban lake to treat stormwater runoff during the construction phase of a 45-ha residential development. After 12 months of operation, parts of the CFWs, with a total area of 147 m², were removed from the urban lake and relocated into a well-established 0.127-ha stormwater retention pond at another site. Biomass and nutrient concentrations of C. appressa shoots above the floating mat and roots below the mat were analysed at both sites 12, 16 and 25 months after initial planting. Plants at the urban lake maintained an extensive root network but there was no increase in total plant biomass at 16 and 25 months after planting. In contrast, the relocated plants in the stormwater pond showed extensive shoot growth but a significant decline in root biomass. C. appressa at the urban lake removed and sequestered 1.00 ± 1.04 g m^-2 N, 0.11 ± 0.07 g m^-2 P and 1.03 ± 0.81 g m^-2 K while plants at the pond removed 11.20 ± 2.29 g m^-2 N, 1.37 ± 0.26 g m^-2 P and 16.13 ± 2.88 g m^-2 K during 12 and 25 months after planting. This study demonstrated that C. appressa adapted rapidly to changes in nutrient availability. The implications are interesting as nutrient levels can be low in constructed lakes during the initial phase of urban developments but can increase rapidly as the development progresses. The study demonstrated multiple benefits of CFWs for stormwater treatment during the early construction stages of an urban development and the potential benefits of relocating and establishing CFWs in existing stormwater retention ponds and lakes.

Nutrient uptake by constructed floating wetland plants during the construction phase of an urban residential development

This study investigated plant growth, nutrient partitioning and total nutrient uptake by tall sedge (Carex appressa) plants in two large-scale Constructed Floating Wetlands (CFW1 and CFW2). These CFWs were installed to treat stormwater runoff discharging into a newly-constructed 2.6-ha lake during the construction phase of a 45-ha residential development. Nutrient concentrations of C. appressa shoot above the mat, biomass within the mat, and roots below the mat were analysed 0, 12 and 16 months after planting. Extensive root growth was evident after 12 and 16 months. Some nutrients (nitrogen, phosphorus, sulphur) were distributed almost evenly among the above-, within-, and below-mat components, while others (aluminium, copper, iron, manganese) were concentrated in or on the roots. Given the low concentrations of nutrients within the water column, large amounts of nutrients were removed from stormwater by the plants. Total nitrogen uptake was 20.20 ± 2.88 kg in CFW1 and 15.00 ± 2.07 kg in CFW2 over the 16-month study period. Total potassium uptake was 12.59 ± 1.64 kg in CFW1 and 7.20 ± 1.56 kg in CFW2. Phosphorus uptake was low as a consequence of low phosphorus availability in the water. High aluminium, iron and manganese concentrations were found in the roots, demonstrating that C. appressa removed and sequestered large quantities of these water pollutants from urban stormwater runoff. For example, total aluminium uptake was 7.82 ± 1.73 kg in CFW1 and 5.62 ± 0.75 kg in CFW2. This study demonstrated multiple benefits of CFWs for stormwater treatment in the early stages of an urban development.

The influence of heavy metals in road dust on the surface runoff quality: Kinetic, isotherm, and sequential extraction investigations

This study examines the adsorption and desorption characteristics of heavy metals in road dust (RD) for the aspect of integrated stormwater management. The chemical fractionations of Cu, Zn, Ni, and Cd were determined by a three-step sequential extraction protocol. Pseudo-first-order and Pseudo-second-order kinetic models, along with Langmuir, Freundlich, and Temkin isotherms were adopted to simulate the batch experimental data. The proportional shift of metals' chemical fractionations in original RD, adsorption equilibrium, and desorption equilibrium were determined. Results show that RD has a remarkable affinity to adsorb heavy metal within a short time. The adsorption processes were well described by the Pseudo-second-order kinetic model (R² = 0.98-0.99) and Freundlich isotherm (R² = 0.89-0.98) for most of the given metals indicating that the chemical adsorption was probably the rate-controlling step and the binding energy for each site was not identical. The maximum adsorption capacities for Cu, Cd, Zn, and Ni were 6300 mg kg^-1, 5800 mg kg^-1, 4000 mg kg^-1, and 3200 mg kg^-1, respectively. A linear fit to the equilibrium pH and the total amounts of the adsorbed metals indicates a strong pH-dependent adsorption. According to the proportional shift of metals' chemical fractionations during the adsorption and desorption processes, the exchangeable fractions of heavy metals in RD were irreversible. It suggests that a portion of the surface sites of RD would be not exchangeable once it was occupied.

Assessment of stormwater pollutant loads and source area contributions with storm water management model (SWMM)

Decentralized urban runoff management requires detailed information about pollutant sources and pathways. However, scarce data of local water quality compel simplified approaches in water quality modelling. This study investigated the use of constant source concentrations in modelling pollutant loads. The source area contributions of total suspended solids, total phosphorus, total nitrogen, lead, copper and zinc were modelled with SWMM based on literature event mean concentrations (EMCs) for different land cover types and on-site rainfall and discharge data for a residential area in southern Finland. The simulated pollutant loads were compared with loads measured at the catchment outlet. Large differences were evident in the modelled catchment-scale and land cover specific loads, depending on the EMC data source. The simulated loads exceeded the measured loads especially during wet conditions, which was explained by the dilution effect of large stormwater volumes on measured EMCs. In addition, the mismatch was explained by the lack of local data for the source area EMCs and by the unaccountability of the mechanisms affecting loads along the pollutant pathways from source areas to sewer outlet. The spatial simulation of stormwater pollutant loads enabled the assessment of source area contributions at the catchment scale, as well as the pollutant pathways and the total diffuse pollution load. For a single pollutant, one or two important pollutant sources contributed the majority of the catchment load, which provides useful information for stormwater management. However, for a group of pollutants, no single land cover type dominated the pollutant loads, reflecting the challenges in decentralized water quality management in the scale of a residential area. Overall, the results emphasize that the widely used stormwater quality modelling with constant EMCs is uncertain even when on-site water quality and rainfall-runoff data from a catchment outlet are available.

Assessing the polycyclic aromatic hydrocarbon (PAH) pollution of urban stormwater runoff: a dynamic modeling approach

Urban stormwater runoff delivers a significant amount of polycyclic aromatic hydrocarbons (PAHs), mostly of atmospheric origin, to receiving water bodies. The PAH pollution of urban stormwater runoff poses serious risk to aquatic life and human health, but has been overlooked by environmental modeling and management. This study proposed a dynamic modeling approach for assessing the PAH pollution and its associated environmental risk. A variable time-step model was developed to simulate the continuous cycles of pollutant buildup and washoff. To reflect the complex interaction among different environmental media (i.e. atmosphere, dust and stormwater), the dependence of the pollution level on antecedent weather conditions was investigated and embodied in the model. Long-term simulations of the model can be efficiently performed, and probabilistic features of the pollution level and its risk can be easily determined. The applicability of this approach and its value to environmental management was demonstrated by a case study in Beijing, China. The results showed that Beijing's PAH pollution of road runoff is relatively severe, and its associated risk exhibits notable seasonal variation. The current sweeping practice is effective in mitigating the pollution, but the effectiveness is both weather-dependent and compound-dependent. The proposed modeling approach can help identify critical timing and major pollutants for monitoring, assessing and controlling efforts to be focused on. The approach is extendable to other urban areas, as well as to other contaminants with similar fate and transport as PAHs.