Abstract
Non-point source pollution, such as stormwater runoff, has become a leading threat to the quality of water resources and aquatic ecosystems near highly developed watersheds. Sudden discharges of stormwater from paved surfaces results in flooding, erosion, sewer overflows, and pollution into receiving waters. Improved stormwater management is needed to protect global water resources. Green infrastructure (GI) stormwater management practices mimic natural landscape hydrology by slowing, spreading, and infiltrating stormwater runoff before discharging it to receiving waters. GI is increasingly designed into urban landscapes to protect waterways from detrimental effects of urban stormwater, but it is still a young and developing technology with many performance knowledge gaps. This dissertation aims to explore the performance of a variety of modern GI practices over an annual range of weather conditions and storm events along Lorton Road in Northern Virginia. There are three primary objectives: (1) compare overall performance of four different GI designs, (2) determine the transport and attenuation of deicing salt in infiltration-based GI, and (3) track denitrification within GI using dual stable nitrate isotope analysis.
There are many different GI designs with sometimes greatly varying levels of stormwater management performance. Additionally, GI performance can be dependent on watershed, storm event, local climate, and maintenance characteristics. Studies have documented the performance of individual GI designs, but few have compared multiple GI designs side by side in the same location and climate. The evaluation of the performance of different operational GI designs receiving similar stormwater runoff conditions is needed to minimize climate and watershed variance and help guide watershed managers in GI selection. This study compares the performance of four different GI designs (bioretention, grass channel (GC), compost amended grass channel (CAGC), and bioswale) receiving the same weather conditions along Lorton Road in Northern Virginia. Stormwater runoff volumes and water quality parameter concentrations were measured at inlets and outlets of each GI during 27 storm events in all seasons over 14 months. The four different GI designs had a wide range of performances with respect to traditional stormwater quality criteria, some acting as pollutant sinks and others as pollutant sources. The bioretention and GC had significantly higher total surface load reduction averages of all water quality parameters than the CAGC and bioswale.
Winter deicing salt application has led to water quality impairment as stormwater carries salt ions (Cl- and Na+) through watersheds. GI is not yet designed to remove salt, but may have potential to mitigate its loading to surface waters. Two infiltration-based GI practices (bioretention and bioswale) were monitored year-round over 28 precipitation events to investigate the transport of salt through modern stormwater infrastructure. Both the bioretention and bioswale significantly reduced effluent surface loads of Cl- and Na+ (76% to 82%), displaying ability to temporarily retain and infiltrate salts and delay their release to surface waters. Changes in bioretention soil chemistry revealed a small percentage of Na+ was stored long-term by ion exchange, but no long-term Cl- storage was observed. Limited soil storage along with groundwater observations suggest the majority of salt removed from stormwater by the bioretention infiltrates into groundwater. Infiltration GI can seasonally buffer surface waters from salt, but are also an avenue for groundwater salt loading.
Strategies to mitigate watershed nitrogen export are critical in managing water resources. GI has shown ability to remove nitrogen from stormwater, but the removal mechanism is often unclear. Denitrification removes nitrate from water permanently, making it the most desirable removal mechanism. The year-round field performance of the bioretention was monitored to investigate the transport of nitrogen and the occurrence and contribution of denitrification. Stormwater runoff volumes, nitrogen concentrations, and nitrate isotope ratios (δ15N-NO3- and δ18O-NO3-) were measured at the inlet and outlet of the bioretention during 24 storm events over 14 months. Nitrate concentration reductions displayed seasonal trends, with higher reductions happening in warmer months and lower reductions or increases occurring in winter. Cumulative bioretention nitrate and total nitrogen load reductions were 73% and 70%, respectively, but only two out of 24 monitored events displayed denitrification isotope trends, indicating other nitrogen removal mechanisms (i.e. infiltration and plant uptake) are primarily responsible for nitrogen surface effluent reductions. Only approximately 1.4% of the total reduced nitrate surface effluent load over the monitoring period was attributable to denitrification. Conditions leading to monitored denitrification suggest future GI designs should consider increasing hydraulic retention time (HRT) to encourage the important ecosystem service denitrification provides.
