SW5231001_Stormwater Report_20240119

Quantifying the Water Quality and Hydrologic Performance of a Submerged Gravel
Wetland Treating Industrial Runoff in Greensboro, North Carolina
Caleb Mitchell, Dr. William Hunt, and Sarah Waickowski
North Carolina State University
3110 Faucette Drive, Raleigh,North Carolina
Final Report
North Carolina Land and Water Fund
Research Project 2017-1005
12/6/2021

Contents
List of Figures 3
List of Tables 6
Acknowledgements 7
Executive Summary 7
Literature Review 8
Introduction 8
Stormwater Control Measures 8
Free Water Surface-Flow Constructed Wetlands 9
Horizontal Subsurface-Flow Constructed Wetlands 11
Vertical Subsurface flow Constructed Wetlands 13
Hybrid Subsurface flow Constructed Wetlands 13
Subsurface flow Gravel Wetlands for Stormwater Treatment 13
Objectives 16
Materials and Methods 16
Site Description 16
Pre-Construction 16
Post-Construction 17
Monitoring Design 19
Pre-construction 19
Post-construction 20
Hydrologic Data Analysis 21
Water Quality Data Analysis 31
Results 32
Post-construction 32
Hydrology 32
Water Quality 37
Discussion 44
Conclusion and Recommendations 45
Minimum Design Criteria 45
References 48
1

Appendix 53
2

List of Figures
Figure 1. Typical secondary wastewater concentrations that are treated by constructed wetlands
(USEPA, 2000b) and traditionally constructed residential stormwater runoff(Bedan and Clausen,
2009) (BOD=Biochemical Oxygen Demand, TSS=Total Suspended Solids,
NH3/NH4=Ammonia/Ammonium,NO3=Nitrate, TN=Total Nitrogen, TP=Total Phosphorus) 10
Figure 2. Schematic of free water surface constructed stormwater wetland. Taken from Hazen
and Sawyer. 11
Figure 3. Schematic of horizontal subsurface-flow wetland. 12
Figure 4. Removal efficiencies for planted HFs treating stormwater and milk condensate runoff
from Warrnambool Cheese and Butter Company Victoria, Australia (Ad=Arundo donax,
Pa=Phragmites australis). Hydraulic loading rates of 3.75 cm/d(1.5 in/d) and 7.5 cm/d(3 in/d)
are shown Data from Idris et al., (2012). 12
Figure 5. Schematic of subsurface-flow gravel wetland treating stormwater. Taken from Hazen
and Sawyer 14
Figure 6. City of Greensboro vehicle service center `STORMWATER OUTFALL C.' Looking
upstream: a) southeast (SE) and southwest (SW) inlet, b)northwest(NW) inlet, c) swale, d) spill
containment baffle and rip rap cascade, and e) outlet dry basin. The red dots indicate the location
of monitoring wells as pictured in Figure 7B and C. 17
Figure 7. Watershed map of the Greensboro SSGW. 18
Figure 8. Pre-construction seasonal high-water table (SHWT) monitoring wells and atmospheric
pressure transducers. A)perforations spaced every 2 to 3-inches, b)pipe sock and bentonite clay
mound, c) sealed monitoring well, and d) atmospheric pressure transducer zip-tied to
`STORMWATER OUTFALL C' sign. 20
Figure 9. ISCO 6712 Portable Sampler with 730 Bubbler Module in job box 20
Figure 10. Modified Phillip Dunne Infiltrometer measuring surface saturated hydraulic
conductivity (KsAT). 21
Figure 11. Diagram of sharp crested contracted v-notch weir. Taken from(Shen, 1981). 23
Figure 12. Underdrain 60-degree sharp crested v-notch weir.Notice the free and fully ventilated
flow from the notch 24
Figure 13. Southwest(SW) inlet looking downstream at the water level logger PVC well casing
followed by the sharp crested 90-degree v-notch weir. 25
Figure 14. Schematic of contracted rectangular(rectangular notch) sharp crested weir. Taken
from Kindsvater and Carter(1959) 26
Figure 15. Northwest(NW) inlet looking downstream at the water sample and bubbler tubing
followed by the compound sharp crested 90-degree v-notch and contracted rectangular weir.
Notice the hydraulic cement yet to be placed at the invert. 27
Figure 16. Southeast(SE) inlet looking downstream at the water level logger PVC well casing
followed by a compound 90-degree v-notch and constricted rectangular sharp crested weir. 27
Figure 17. Broad crested overflow weir and original underdrain with upturned PVC T. 28
3

Figure 18. Manual and tipping bucket rain gauge supported by a 10-cm(4-in)by 10-cm(4-in)
post. Notice the field bag, sampling bottles, and job box with solar panel and auto sampler. The
NW inlet is to the left and SSGW in the background to the right. 30
Figure 19. Adjusted rainfall totals per storm across the monitoring period. Storms above 1-inch
are highlighted in dark red, storms between 1 and 0.5-inch are highlighted in yellow, and storms
less than 0.5-inches are highlighted in green. 33
Figure 20. Water levels within the subsurface-flow gravel wetland during the monitoring period
with media layers, water level control structure, and maintenance activities. Note the dimensions
of the water level control structure and diameter of the underdrain are not to scale (NTS). 33
Figure 21. Estimated seepage between the subsurface gravel wetland and the surrounding
watershed groundwater throughout the monitoring period as a function of saturated subbase in-
situ hydraulic conductivity, KSAT 34
Figure 22. Cumulative volume inflow, outflow, and flow reduction for storm events. Note
erroneous storms have been removed. 35
Figure 23. Comparison of peak storm flow distributions using box and violin plots between
inflow and outflow of the subsurface gravel wetland. Note that all 226 storms are included. 36
Figure 24. Comparison of storm flow volume distributions using box and violin plots between
inflow and outflow of the subsurface gravel wetland. Note that all 226 storms are included. 36
Figure 25. Sediment and nutrient comparison of inflow versus outflow concentration boxplot
distributions 37
Figure 26. Sediment and nutrient percent reduction in concentration from inflow versus
underdrain flow as a function of time. 38
Figure 27. Sediment and nutrient percentile concentrations comparing inflow versus underdrain
flow. 38
Figure 28. TSS concentration comparison of inflow versus underdrain flow as a function of
percentile. Note the horizontal lines indicate poor(red)to good(green)macroinvertebrate water
quality in the Piedmont of North Carolina(McNett, 2010) 39
Figure 29. TN concentration comparison of inflow versus underdrain flow as a function of
percentile. Note the horizontal lines indicate fair(yellow)to good (green) macroinvertebrate
water quality in the Piedmont of North Carolina(McNett, 2010). 39
Figure 30. TP concentration comparison of inflow versus underdrain flow as a function of
percentile. Note the horizontal lines indicate fair(yellow)to good(green) macroinvertebrate
water quality in the Piedmont of North Carolina(McNett, 2010). 40
Figure 31. TSS load comparison of inflow versus outflow boxplot distributions as a function of
season. 41
Figure 32. TN load comparison of inflow versus outflow boxplot distributions as a function of
season. 41
Figure 33. TP load comparison of inflow versus outflow boxplot distributions as a function of
season. 42
Figure 34. TSS load percent reduction boxplot distributions as a function of season. 42
4

Figure 35. TN load percent reduction boxplot distributions as a function of season. 43
Figure 36. TP load percent reductions boxplot distributions as a function of season. 43
Figure 37. Coefficient of discharge, Ce, as a function of h/P and P/B for 90-degree v-notch sharp
crested weirs. Notice that for weirs with P ten times smaller than B the value of Ce is 0.578
regardless of increasing head, h. Taken from (Shen, 1981) 53
Figure 38. Coefficient of discharge, Ce, as a function of v-notch angle, O. Taken from (Shen,
1981). 53
Figure 39. Adjustment factor for viscosity and surface tension, kb, as a function of v-notch angle.
Taken from(Shen, 1981). 54
Figure 40. Crest length correction factor, kb, for various weir contraction ratios, b/B. Taken from
(Kindsvater and Carter, 1959) 54
Figure 41. Coefficient of discharge, Ce, as a function of h/P for sharp crested rectangular weirs.
Taken from(Kindsvater and Carter, 1959). 55
Figure 42. Comparison of peak storm flow distributions using box and violin plots between
inflow and outflow of the subsurface gravel wetland. Note that only the 159 non erroneous
storms are included. 56
Figure 43. Peak storm flow percent reduction boxplot and violin distributions as a function of
season. Note that all 226 storms are included. 56
Figure 44. Comparison of volume of storm flow distributions using box and violin plots between
inflow and outflow of the subsurface gravel wetland. Note that only the 159 non erroneous
storms are included. 57
Figure 45. Volume percent reduction boxplot and violin distributions as a function of season.
Note that all 226 storms are included. 57
5

List of Tables
Table 1. Comparison of SSGWs pollutant removal efficiencies treating stormwater. UNH and
Farrell Park are case studies of SSGWs. NJDEP and Knox County standards are based on the
UNH case study. (Knox, 2008; NJDEP, 2014; Vanasse Hangen Brustlin and Dipietro, 2012) 14
Table 2. SSGWs design features including minimum and maximum soil layer depths and
seasonal high water table separation. 15
Table 3. Triangular(v-notch)weir discharge coefficients, Ce, and head adjustments, kwh, as a
function of notch angle from Shen, (1981). 23
Table 4. Rectangular notch weir crest length correction factor,kb, as a function of contraction
ratio, b/B from Kindsvater and Carter, (1959) 26
Table 5. Sharp crested rectangular weir coefficient of discharge, Ce, as a function of h/P for
various contraction ratios,b/B. Interpreted and converted to nondimensional form from Figure
41 (Kindsvater and Carter, 1959)using M = 322gCe 26
Table 6. Water quality laboratory analysis methods and calculations. 31
Table 7. Median sediment and nutrient load percent reductions as a function of season. 44
6

Acknowledgements
North Carolina State University would like to thank the City of Greensboro, the North
Carolina Land and Water Fund(formerly the North Carolina Clean Water Management Trust
Fund), Dragonfly Pond Works, and Mellow Marsh Farms for making this project a success.
North Carolina State University would also like to thank Sarah Waickowski, Shawn Kennedy,
and Erin Carter for their work designing the wetland, installing monitoring equipment, and
collecting data, respectively. Honorable mentions include Amber Ellis, Duwane Thomas,
Rebecca Rahn, Carmen Tormey, Karen Sloan, and Andrew Yount for their help planting wetland
plants and David Stapleton for raking the surface with the bunker raker.
Executive Summary
Urban development creates unsustainable flashy hydrology which consequently pulses
intense hydraulic, sediment, and pollutant loads into receiving fluvial systems. Natural
(undisturbed) landscapes yield sustainable hydrology with distinct subsurface pathways,
predominately shallow interflow and groundwater surge. To create sustainable development in
terms of optimizing innovative post-construction or retrofitted stormwater control measures, this
research assesses the water quality and hydrologic performance of an undersized subsurface
gravel wetland treating and controlling stormwater.
There are multiple hydraulic configurations for gravel wetlands including: (1)horizontal
flow systems which contain continually saturated gravel media, (2)vertical flow systems that can
either have saturated or unsaturated media, and (3) hybrid flow systems which utilize a
combination of hydraulic fluxes and saturation conditions.
In Greensboro,North Carolina the undersized subsurface gravel wetland was able to
detain and fully treat the 13-mm (0.51-in) storm and utilized a hybrid flow setup. Flow initially
entered and filled a wet forebay where it overtopped a berm to enter the wetland basin. Then
vertical flow, through unsaturated surface wetland media(sand) and perforated risers, directed
water down into the subsurface submerged gravel media where horizontal flow dominated. The
formation of a schmutzdecke clogged the surface with up to 85%reduction in surface saturated
hydraulic conductivity.
Raking of the surface was performed twice during the first year after construction to
maintain the intended hydrology. Raking increased the surface saturated hydraulic conductivity
by up to 214%. Thereafter, macrophyte establishment maintained intended subsurface wetland
hydrology. After 2.3-years of operation, the wetland surface saturated hydraulic conductivity
reduced to 40 mm/hr(1.59 in/hr), within the current design guidance for filtration-based
practices in North Carolina 25 to 50 mm/hr(1 to 2 in/hr).
The median concentration percent removals (n=28 storms)were 92% for TSS, 43% for
TN, 47% for organic-N, 43% for TKN, 54% for NOx, 40% for NH3, 67% for TP, 53% for
Ortho-P, and 75% for Particle-P. The median percent reductions were 77, 23, and 50% for TSS,
TN, and TP loads, respectively. The areal annual loading rates were 388, 7.58, and 1.92 kg/ha/yr
(346, 6.76, and 1.71 lbs/ac/yr) for the inflow.
7

Literature Review
Introduction
Urban stormwater runoff and its historic control by humans has undoubtedly devastated
communities and ecosystems that depend on streams and rivers. It is understood that peak
stormwater flows generated from connected urban and suburban impervious surfaces can
hydraulically blow out culverts and bridges, destroy in-stream ecosystems, and put downstream
communities more at risk to flash floods. Historically connected impervious surfaces convey
peak flows offsite as quickly as possible, reducing local flooding and keeping transportation
infrastructure clear of water. However, once this quickly-conveyed, sediment-hungry water
reaches the local stream it can erode and quickly degrade habitat by pulsing unnatural hydraulic,
sediment, and pollutant loads through the downstream reach (Walsh et al., 2005). With respect to
climate change and the increasing intensity, depth, and duration of storm events, this urban
stream syndrome problem seems hopeless.
However, there have been strategies to combat this issue of degraded and dangerous
fluvial systems such as in-stream restoration, floodplain connection,riparian buffers, and
watershed management(Palmer et al., 2014). All these restoration techniques share one thing in
common: sustainably managing fluvial systems. As in-stream geomorphic restorations based
upon reference reach design continue to fail at providing their intended benefits (Miller and
Kochel, 2013, 2009; Palmer et al., 2014), innovative watershed-scale restoration techniques are
necessary to combat the complex issue of the urban stream syndrome. Such techniques include
implementing and testing novel stormwater management tools that mimic natural hydrology.
Stormwater Control Measures
When developing land under the rules of the National Pollution Discharge Elimination
System(NPDES), two of the major goals are to ensure that the post-construction runoff
hydrograph peak is less than or equal to the pre-construction runoff hydrograph peak and that the
water quality is not harmfully altered. Even when performing well at peak flow reduction,
traditional stormwater control measures (SCMs) can unintendedly induce longer durations of
bank forming flow in streams and worsen larger watershed flooding (McCuen, 1979; National
Research Council, 2009). To help alleviate this issue, there has been a primary shift in the goals
of stormwater control to instead mimic predevelopment conditions through the implementation
of low impact development(LID) and retrofitting of traditional SCMs. The goal of maintaining
predevelopment hydrology is usually meet through the specifications of LID by disconnecting
the directly connected impervious surfaces (Bedan and Clausen, 2009; Davis, 2007) and
allowing surface runoff to flow across soil and vegetation(usually a lawn) in hopes of promoting
both shallow interflow and ultimately groundwater surge (Cizek and Hunt, 2013). This can be
done by daylighting existing grey infrastructure (e.g.,pipes, road gutters, and impervious
conveyance channels) and allowing the water to instead flow through a natural or ecologically
engineered system like a dry detention pond,wet retention pond (Blecken et al., 2017; DeBusk et
al., 2010),permeable roadway (Bean et al., 2007; Collins et al., 2010; Page et al., 2014, 2015),
bioretention cell (Davis, 2007; Davis et al., 2009, 2012; Passeport et al., 2009), sand filter
8

(DeBusk et al., 2010), suspended pavement tree cell (Page et al., 2014, 2015), or stormwater
wetland(Merriman et al., 2017; Moore and Hunt, 2012).
However, some of these systems have public stigma like dry ponds, wet ponds, and
stormwater wetlands due to the idea that all ponded surface water, whether temporary or
permanent, harbors the growth and survival of mosquitos and thus vector diseases (Greenway et
al., 2003; Hunt et al., 2005; USEPA, 2000a). This is a misinformed stigma because mosquitos
will not survive with proper wetland vegetation, permanent water storage, and predators as
discussed by Knight et al. (2003). This stigma can be eliminated with the implementation of
stormwater control measures (SCMs) that utilize subsurface flows like those flows that are found
in a subsurface gravel wetland(Saeed and Sun, 2012; Suliman et al., 2006). Also,because
airports need to have clear airways free of avian life as well as clean controlled stormwater
runoff, a subsurface flow system like a subsurface gravel wetland (SSGW) will eliminate/retrofit
currently used airport stormwater wetlands that are highly suitable habitat for large avian life
(US EPA, 1999; USEPA, 2000a). However, the merits of using a subsurface gravel wetland to
help control this ever increasing stormwater runoff have not been fully verified or even
quantified in North Carolina(Ballestero and Houle, 2015).
A subsurface gravel wetland is a variation of and potential retrofit of existing SCMs. This
hybrid SCM, a subsurface gravel wetland, needs to be fully qualified and quantified on the basis
of its runoff control, water quality benefits, construction constraints, and economic viability, so
that land developers and managers can credit (Waickowski et al., 2017), utilize, and install the
practice properly. To understand the water treatment mechanisms that will be present in a
subsurface gravel wetland and pollutant concentration differences between stormwater and
wastewater three types of wetlands will be described: (1) free water surface flow wetlands, (2)
subsurface gravel wetlands treating wastewater, and(3) subsurface gravel wetlands treating and
controlling stormwater.
Free Water Surface-Flow Constructed Wetlands
Free water surface (FWS)-flow constructed wetlands are built with the intention of
mimicking natural wetlands (Moore and Hunt, 2012; Rooney et al., 2015). Topographic
adjustments to the landscape such as excavating deep pools and shallow water conveyance are
used to ensure that during droughts the wetland has a permanent pool of water, a habitat for
aquatic macroinvertebrates, amphibians, and fish. This reduces the frequency of mosquito larvae
occurrences by providing habitat for predators (Greenway et al., 2003; Hunt et al., 2005; Knight
et al., 2003; USEPA, 2000a). Berms are constructed to provide microtopography which create a
longer flow path thus encouraging soil water interactions for longer durations at varying
elevations of the oxygen and temperature stratified water column (Wright et al., 2006).
Organic carbon amendments to the anaerobic wetland environment are placed to
encourage the growth of denitrifying bacteria (Burchell et al., 2007; Lenhart et al., 2012).
However, the amount of carbon pool amendments in these wetlands needed to promote
denitrification is dependent on the nitrogen concentrations of the incoming water(Burchell et al.,
2007; Lenhart et al., 2012). USEPA (2000b) characterized minimum and maximum pollutant
9

concentrations for secondary municipal wastewater typically received by constructed wetlands,
whereas Bedan and Clausen (2009) characterized residential stormwater pollutant concentrations
for both LID and traditional developments. The minimum secondary wastewater and
traditionally developed residential stormwater runoff concentrations are listed in Figure 1.
70
60
60 — —
50
• 40
• 30 24
O 20 — 20
10 9
4 3.4 1.45
0.15 0.3 0.185
Wastewater [mg/L] Residential Stormwater [mg/L]
®BOD TSS ®NH3/NH4 DNO3 ®TN ©TP
Figure 1. Typical secondary wastewater concentrations that are treated by constructed wetlands
(USEPA, 2000b) and traditionally constructed residential stormwater runoff(Bedan and
Clausen, 2009) (BOD=Biochemical Oxygen Demand, TSS= Total Suspended Solids,
NH3/NH4=Ammonia/Ammonium, NO3=Nitrate, TN=Total Nitrogen, TP=Total Phosphorus)
For stormwater FWS wetlands there are five zones delineated by internal topography: (1)
deep pools, (2)transitions between deep and shallow water, (3) shallow water, (4)temporary
inundation areas, and(5)the upper banks. The deep pools (and forebays), which occupy 20 to
25% of the wetland surface area, are for energy dissipation,particulate settling, and soluble
organic compound removal in the anaerobic zones near the water-detritus-soil interface (Dotro et
al., 2017; Hunt et al., 2007; Vymazal, 2014). The smallest amount of surface area is occupied by
the transition zone where a gentle slope, no steeper than 1.5:1 (H:V), connects the deep pools to
the shallow water and promotes both aerobic and anaerobic conditions with large gradients in
redox potential (Dotro et al., 2017; Hunt et al., 2007). The shallow water zone, depths of 5 to 10-
cm(2 to 4-in), allows for a wide variety of plant life to grow and contains high levels of
dissolved oxygen due to atmospheric diffusion and photosynthetic activity of periphyton and
algae (Vymazal, 2014; Hunt et al., 2007). It promotes nutrient transformations by facultative
heterotrophic bacteria(Dotro et al., 2017) . During storm events (Figure 2)the temporary
inundation zone provides a floodplain where biodiversity and hyporheic exchange can occur
between the catchment and wetland (Adyel et al., 2017; Hunt et al., 2007). The upland area,
upper bank with slopes no steeper than 3:1 (H:V), connects the wetland to the surrounding land
with a variety of vegetation ranging from obligate upland to obligate wetland species depending
on the water table depth (Hunt et al., 2007).
10

Between Storms During Storms
Low Flow DrawoaNn
4 Overflow Structure
• I i r i I IA , ,; —outnowto
' k k sue,
•
Wetland sod '° �� ,
Clay Leer +mw
Easting Sell
Figure 2. Schematic of free water surface constructed stormwater wetland. Taken from Hazen
and Sawyer.
Typically, FWS wetlands for stormwater are planted with a variety of native emergent,
submerged, and free floating macrophytes within these various zones. The plants are selected
based on their tolerance to both inundation during the wet season and drought during the dry
season. The U.S. Fish and Wildlife Service has classified more than 6,700 species that are
obligate and facultative to wetlands (USEPA, 1999). Obligate wetland species are found growing
exclusively in wetlands, whereas facultative wetland species grow in both upland and wetland
areas (USEPA, 1999). However, FWS wetlands that treat tertiary wastewater are most often
planted with a monoculture of emergent plant genus like Typha, Phragmites, Scirpus, and
sometimes submerged plants like Potamogeon and Elodea, and floating plants like Eichornia
and Lemna (Dotro et al., 2017; Vymazal, 2014). The floating macrophytes, like filamentous
algae, and duckweed(Lemna spp.) typically enter the wetland through influent water and
intermingle with the dominating emergent and submerged macrophytes (USEPA, 1999).
Horizontal Subsurface-Flow Constructed Wetlands
Horizontal subsurface-flow constructed wetlands (HFs) operate as anaerobic treatment
systems due to the usage of water level control structures that promote internal water storage
(IWS) and perforated inlet distribution pipes that loads the wetland subsurface (Figure 3). HFs
are used most efficiently to polish tertiary wastewater of specific species of nutrients specifically
nitrified nitrogen (Dotro et al., 2017). A HF's capacity and ability to treat wastewater is
dependent upon the hydraulic loading rate (HLR) and hydraulic retention time (HRT) (Mayo and
Mutamba, 2004). The HF bed(usually granular material like gravel) can clog and cause wetland
short-circuiting through surface flow. Clogging can also back up water into the upstream
drainage system/sewer when the inlet and outlet are not properly placed to decrease preferential
flow paths and HLR exceeds 30 cm/d (Suliman et al., 2006). Macrophytes are planted to improve
media hydraulic conductivity near the surface, yet they also reduce media porosity due to
rhizome growth(Dotro et al., 2017; Nivala et al., 2012; Silveira et al., 2015; Vymazal, 2013).
11

However, the reduction in media porosity (i.e., clogging) is accompanied with the added benefits
of increased aerobic and carbonic microsites in the rhizosphere to promote transformation and
uptake of nutrients (Dotro et al., 2017;Nivala et al., 2012; USEPA, 2000a).
0
------___ n .,•
Figure 3. Schematic of horizontal subsurface flow wetland.
Eight 1.6-m(5.2-ft)by 0.8-m (2.6-ft)by 0.4-m (1.3-ft) length by width by depth HFs
were constructed and monitored by Idris et al. (2012) in Victoria Australia to treat 10,000,000
liters (2,641,720 gallons) of annual stormwater runoff(pavement, pans, and roof) and milk
condensate (skimmed milk powder production) from a milk processing plant. The HFs were all
proceeded by a series of bulk and wetland specific water storage tanks. The tanks provided
primary treatment through particulate sedimentation and aerobic nutrient transformations
removing 42% and 26% of the BOD and TSS, respectively, before the stormwater entered the
HFs. The average (± SD) untreated influent stormwater concentrations were 21 ± 9, 28 ± 13, 5.2
± 1.2, and 6.2± 1.3 mg/L for BOD, TSS, TN and TP respectively. The removal percentages of
the planted HFs are listed in Figure 4.
100 - 96 94
87 91 ."'.
90 84 , 87
80 78 75 75
71 73 71 •
0
70 NN61 60
60 " 55
U
50 \\ 41 % .
W
20 •
EEEEEEEE
•
Ad 3.75 cm/d Pa 3.75 cm/d Ad 7.5 cm/d Pa 7.5 cm/d
El BOD CI TSS El TN OTP
Figure 4. Removal efficiencies for planted HFs treating stormwater and milk condensate runoff
from Warrnambool Cheese and Butter Company Victoria, Australia (Ad=Arundo donax,
Pa=Phragmites australis). Hydraulic loading rates of 3.75 cm/d(1.5 in/d) and 7.5 cm/d(3 in/d)
are shown Data from Idris et al., (2012).
12

Vertical Subsurface flow Constructed Wetlands
Vertical subsurface-flow constructed wetlands (VFs) are designed and operated as
aerobic treatment systems, through intermittent surface loading and resting periods typically 2 to
3 d and 4 to 5 d, respectively (Vymazal, 2013, 2014). This pulse dosing allows for the void
spaces to be occupied by air during the interval between feeding batches when the water drains
away through the drainage network at the bottom of the filter(Dotro et al., 2017). Clogging of
the media bed is typically prolonged by intermittent loading to promote aerobic removal of
organic carbon, earthworms to cause bioturbation of the media, and planting of macrophytes
(Dotro et al., 2017; Nivala et al., 2012). The latter can temporarily improve surface infiltration
rates by forming preferential flow paths along the stalks and roots, but as biofilm and solids
accumulate in the upper portions of the media the system can further clog (Nivala et al., 2012).
The rhizosphere can also release organic compounds that aid in denitrification when the
subsurface is saturated(Dotro et al., 2017). Silveira et al. (2015) evaluated the nitrogen removal
performance of two pilot-scale VFs with a subsurface saturated layer receiving raw wastewater.
This system is called the French system since it is directly loaded with untreated wastewater.
These systems are not allowed in the United States or cold climates due to possible human
contact with the wastewater or freezing as it percolates down through the media from the surface
(Dotro et al., 2017; Nivala et al., 2012). During the winter Silveira et al. (2015) noted long
durations of clogging causing slow drainage and low peak drainage flow. During the winter
drainage peak flows were 0.02 and 0.05 m/h (0.8 to 2 in/h) for two VFs monitored by Silveira et
al. (2015), and between 0.1 and 0.45 m/h (3.9 to 18 in/h) for both VFs for the rest of the
experiment.
Hybrid Subsurface flow Constructed Wetlands
Typical hybrid flow constructed wetlands are utilized to treat raw sewage in two stages.
Stage one consists of VFs in parallel and stage two consists of HFs in parallel or vice versa.
Typically, each VF in parallel receives the wastewater for 3 to 4 d then rests (drains) for one
week. VF systems promote full nitrification but no denitrification without an anaerobic zone
within the media bed. Thus, horizontal flow constructed wetlands (HF) were added to the
treatment process to denitrify the nitrified VF effluent in a hybrid system called VF +HF. Many
other variations of hybrid flow constructed wetlands include HF +VF, FWS +HF, VF +FWS,
and various combinations and arrangements of these wetland types. However, the VF+HF are
most efficient at ammonia removal, the FWS are most efficient at total nitrogen removal, and all
hybrid constructed wetlands (CWs)remove NH4-N at a similar rate to a single VF. Although, all
hybrid CWs remove total nitrogen more efficiently than a single VF or HF (Vymazal, 2013).
Molle et al. (2008) investigated the nitrogen removal capabilities of two subsurface flow
constructed wetlands in series, vertical then horizontal (i.e., VF +HF), and determined that
during warmer temperatures (>15°C) total nitrogen removal was 83%.
Subsurface flow Gravel Wetlands for Stormwater Treatment
Subsurface gravel wetlands treating stormwater (Figure 5) are innovative and much of the
design guidance (CRWA, 2009; DNREC, 2019;NJDEP, 2014; Vanasse Hangen Brustlin and
13

Dipietro, 2012) is based upon the few studies conducted by the University of New Hampshire
Stormwater Center(UNHSC) (R. Roseen et al., 2010) and the subsequent guidance they have
produced(Ballestero et al., 2016; Ballestero and Houle, 2015; R. M. Roseen et al., 2009). For the
UNHSC SSGW, nitrate removal was significantly less during the winter months due to the lack
of an anaerobic zone triggered by higher dissolved oxygen concentrations and slower microbial
activity caused by the lower temperatures (Ballestero and Houle, 2015). Conversely,phosphorus
removal was higher during the winter due to increased dissolved oxygen concentrations and
acidic conditions which promotes precipitation of phosphate with iron and aluminum(R. Roseen
et al., 2010). The performance of SSGWs treating stormwater are presented in Table 1.
Between Storms During a Storm Outlet
Control
1 1
Perforated / 1 , <Milo,to
v
Riser I! 1 I ` 'i P t i I t stream
Vegetaton
- aWetland Soil itr
@gC
Jr-.
Choker St nii
Aggregate Storage
MINN
aaaoowauva
(Nay Liner
Figure 5. Schematic of subsurface flow gravel wetland treating stormwater. Taken from Hazen
and Sawyer
Table 1. Comparison of SSGWs pollutant removal efficiencies treating stormwater. UNH and
Farrell Park are case studies of SSGWs. NJDEP and Knox County standards are based on the
UNH case study. (Knox, 2008;NJDEP, 2014; Vanasse Hangen Brustlin and Dipietro, 2012).
Particulates Nutrients Heavy Metals
TSSt0I Pathogens[b] TN[e] TKN[d] NO3[e] BOW TPtgl TDPIh1 Zinc Copper Iron
SSGW
Removal Efficiency[%]
UNH 99 80 55 95
Farrell
98 91 63 80 68 88 30 79 42 79
Park
NJDEP
90 90
Standards
Knox Co. 80 70 20 60 50
[a]TSS=Total Suspended Solids [e]NO3=Nitrate Nitrogen
[b]Pathogens=E.Coli,Enterococci,or Fecal Coliform [f]BOD=Biochemical Oxygen Demand
[c]TN=Total Nitrogen [g]TP=Total Phosphorus
[d]TKN=Total Kjeldahl Nitrogen [h]TDP=Total Dissolved Phosphorus
14

The design and maintenance guidance of these systems is limited. However, there are
gravel depth recommendations and siting guidance with respect to the seasonal high groundwater
table (SHWT). Media depth and SHWT separation guidance are listed in Table 2. Roseen et al.
(2010) determined that SSGW forebay maintenance would need to occur every three years to
ensure aerobic conditions not impeded by oxygen demanding detritus and thick plant litter,but
this was for New Hampshire where the summers are short, and the winters are long and bring
intense snowfall. The frequency of forebay maintenance and SSGW clogging (i.e., reduction in
surface media hydraulic conductivity) may be different in a warmer climate due to a quicker
development of the schmutzdecke (Hunt et al., 2015). Further, mosquito larvae frequency,
biodiversity, ecosystem services, carbon and GHG sequestration, treatment capacity as either an
infiltration or filtration practice (i.e., loose or tight underlying in-situ soil), public perception,
improvement of downstream aquatic habitat, and human safety in the right of way have yet to be
quantified for SSGWs. Potentially SSGWs, if highly maintained, could retrofit golf course
bunkers/sand traps into SCMs (Lemons, 2013). However, design features of these systems need
to be optimized based upon watershed characteristics (e.g., stable vs. unstable), wetland media
depth and composition, plant types, influent pollutant loads, surface to subsurface storage ratio,
subsurface water storage volume to water quality volume (R. Roseen et al., 2010), and watershed
to wetland area loading ratio. Additionally, denitrification potential, biomass production, and
pathogenic removal still need to be quantified for these systems.
Table 2. SSGWs design features including minimum and maximum soil layer depths and
seasonal high water table separation.
Minimum and Maximum Depths of Soil Minimum Seasonal High
Layers(in) Groundwater Table Separation(in)
or Qualitative
Citation Wetland Choker Gravel
Surface Stone Media
UNHSC: 70% #78 Stone, #57 Stone,
Sand,>15% Pea Gravel Driveway
Organic Matter, (cpnominal= Gravel,
<15% Clays 0.25 to 0.5 River Rock
(Here-in: 100% inch) (( nominal—O.5
Sand) to 0.75 inch)
DNREC,2019 8 4 24 to 48 No separation necessary
Dipietro, 2012 - - 30 No guidance
CRWA,2009 8 - 24 May be appropriate for use in
areas with high water level
NJDEP, 2014 8 3 24 12
R. M. Roseen et 8 3 24 Groundwater interception
al.,2009 allowed/encouraged for continual
subsurface submergence
Knox,2008 - - - 24 if receiving hotspot runoff
Here-in: 6 3.5 12 12 if groundwater drinking source
Mitchell,2021
15

Objectives
However, the research proposed here-in investigates three specific aspects of a hybrid
flow SSGW: (1)hydraulic properties and control of stormwater, (2)hydrologic mitigation of
urbanization, and (3) TSS, nutrient(N and P), and metals removal from municipal vehicular
storage yards. It is hypothesized that due to the sporadic hydroperiod of a SSGW subjected to
flashy urban runoff, as opposed to a steady-state hydroperiod of a wastewater SSGW, clogging
will occur less frequently due to the development of aerobic conditions between storm events.
Materials and Methods
Site Description
Pre-Construction
A riparian buffer of 132 trees with median diameter at breast height(DBH) of 23-cm(9-
in) surrounded the City of Greensboro Vehicle Service Center Stormwater Outfall C (Figure 6).
The drainage pipe infrastructure for the service center daylighted into the forested wet swale
(Figure 6A-C). The swale was traversed with tin sheets attached to a horizontal telephone pole
buried into berms (Figure 6D). The swale water level was set right at the bottom of the tin sheets
by a downstream clay berm covered with non-woven geotextile and rip rap (Figure 6D). This
acted as a flow baffle for spill containment of floatable material (e.g., trash, motor oil, and gear
grease). After overtopping the berm, the water would plunge down the rip rap cascade into the
dry basin and exit through the outlet pipe, see Figure 6E.
16

A B C CIE
, r
�I /'
,,
`r. �'� ithia:k..' -tom
7
7B
,N O�\ /Nr�
`\ rNN
, N
0 rrNIN,
\\ —
s n
O as
'�
-'- \\ 7C
\A
zr \\
7N.4 247 WI W. 0' `\.
d
Figure 6. City of Greensboro vehicle service center 'STORMWATER OUTFALL C. 'Looking
upstream: a) southeast(SE) and southwest (SW) inlet, b) northwest(NW) inlet, c) swale, d) spill
containment baffle and rip rap cascade, and e) outlet dry basin. The red dots indicate the
location of monitoring wells as pictured in Figure 8B and C.
The 2018 conditions of Stormwater Outfall C needed retrofitting. The NW inlet created a
scour hole that was undermining the soil support around the haunch of the pipe (Figure 6B). The
swale was eroding the berm along the Gillespie Golf Course (Figure 6C). The rip rap on the
cascade had slipped down into the dry basin which left the non-woven geotextile exposed
(Figure 6D). The outlet pipe had accumulated logs which created a higher potential for blockage
(Figure 6E).
Post-Construction
In July 2019, a subsurface-flow gravel wetland was constructed between the City Vehicle
Service Center and Gillespie Golf Course in Greensboro,North Carolina. The 1,335-m2 (14,365-
ft2) submerged gravel wetland was built on silty sand in-situ soil described as "sand-rock." It was
built with a 265-m2 (2,850-ft2)wet forebay followed by a 970-m2 (10,450-ft2) saturated planted
sand and gravel media bed. The 150-cm(5-ft) deep forebay collected the Service Center's 5.7-ha
(14.0-ac) of ultra-urban (89% impervious)watershed. The dominate soil was Enon-Urban, a
hydrologic soil group C. The composite Curve Number(CN)was estimated as 97 with 98 and 86
17

for the impervious and pervious, respectively. There were three pipe inlets, northeast (NW),
southeast (SE), and southwest (SW), and the direct run-on hillslope contributing 66.0, 30.1, 2.1,
and 1.8% of the drainage area, respectively (Figure 7).
• Mdnhoirs Q Runon'VYeterl,e1 0 2550 100 150 200 N
■ Drop Inlets Southwest Watershed .'.'� m� US Feet
Storm sewer r1 Southeast WatershedF
-1Fdebay Berm rI Northwest Watershed
1 Gravel Wetland I Pervious .' 7, r
,Greeroi•o Vehicle Service Center
ii401 Patton Avenue.
Greensb'oi o, North Carolina; 27406 ,r
Map 0 eator: Caleb Mitchell - "
Date Printed: 11/4/2021
(timemnimi"
■ n ■
■
311' ■ ■ v
■
o
h►
n ■ o u
•
• • •
u • a -
r
1 ■ I
■
II illta
• e
■
■ •■ • 1
■
u
■■ ■
■
0 l]17
o
II o
0 0 - s..,...
I
_ !�91w 1 'la lc rde
'27"N ••• [.
Figure 7. Watershed map of the Greensboro SSGW.
18

The hydraulic loading ratio (HLR, see Equation 1) for the entire wetland(i.e., forebay,
berm, and media)was 42:1. The runoff which collected in the forebay discharged into the
wetland after overtopping the 98-m2 (1,059-ft2) forebay berm. The forebay berm was vegetated
and armored with rip rap along its slopes. The top of berm was 13.7-m(45-ft) long by 3.0-m (10-
ft)wide and acted as a broad crested weir—setting the normal pool elevation of the forebay. The
HLR for the surface of the media was 60:1.
A
HLR = Ad (1)
W
Where: HLR is hydraulic loading ratio, unitless
Ad is the watershed drainage area, ha (ac)
AW is the wetland surface area, ha(ac)
The water quality storm of 25-mm (1-in)was estimated to generate 990-m3 (35,000-ft3)
of runoff which was not entirely captured in the 30-cm(12-in) deep bowl surface storage of 310-
m3 (10,995-ft3) above the gravel bed. Taking the water quality volume to bowl temporary storage
volume (WQV:BSV)the wetland is 31% sized. The largest storm to be captured was 13-mm
(0.52-in), however this is due to dry antecedent conditions and a low intensity storm.
Monitoring Design
Pre-construction
The existing forested swale, spill containment baffle, rip rap cascade, and dry basin was
monitored for seasonal high-water table (SHWT) at two locations (see Figure 6) from October
2018 to May 2019. Two monitoring wells were constructed out of 5-cm(2-in) diameter schedule
40 polyvinyl chloride (PVC)pipes that were perforated every 7.5-cm(3-in), wrapped in pipe
sock, and hydraulically sealed at the surface by bentonite clay-soil mounds. Three HOBO U20L-
04 Water Level Loggers (Onset Computer Corporation) recorded absolute pressure and
temperature every 2-mins. Two were in the bottom of the wells and one in the atmosphere zip-
tied to a sign labeled"STORMWATER OUTFALL C" (see Figure 8).
19

A► B C 1 �i a4t.'� D
K -7,
mod . =' _ : � . f 4' r� > , �� � T�RM� -
:� �=� '� OUTFAL
r , { :, R
Figure 8. Pre-construction seasonal high-water table (SHWT) monitoring wells and atmospheric
pressure transducers. A)perforations spaced every 2 to 3-inches, b)pipe sock and bentonite clay
mound, c) sealed monitoring well, and d) atmospheric pressure transducer zip-tied to
'STORMWATER OUTFALL C'sign.
Post-construction
The subsurface gravel wetland was monitored for hydrology and water quality at the
inlets, outlets, and within the media bed. Rainfall was monitored at the NW inlet using a Rain
Collector 7852 0.25-mm (0.01-in)tipping bucket rain gauge (Davis Instruments, Hayward, CA)
and calibrated with a manual Stratus Rain Gauge (Productive Alternatives, Fergus Falls, MN).
Discharge was monitored using compound 60 and 90-degree v-notch and contracted rectangular
sharp crested weirs at the NW, SE, and SW inlets, underdrain, and a 15-cm (6-in)breadth broad
crested weir at the overflow. At the NW inlet and underdrain, ISCO 6712 Portable Samplers with
730 Bubbler Modules (Teledyne Technologies, Thousand Oaks, CA)were used to measure water
level above weir inverts every 2-min to calculate discharge and collect flow weighted water
samples (Figure 9). Water levels above weir inverts, within the gravel media, and in a reference
well were measured every 2-mins using HOBO U20L-04 Water Level Loggers (Onset Computer
Corporation, Bourne, MA) for the SE and SW inlets, overflow, and reference SHWT,
respectively.
't041..11‘`,A.,_,ik,.V!W If1 'F 03F-a► =1._ _`_
my
Ili
I er..
A`
V
y ';
:::2 41(12(' S .,
1
A
Figure 9. ISCO 6712 Portable Sampler with 730 Bubbler Module in job box.
20

Surface saturated hydraulic conductivity (Ksat)was measured during and post-
construction using the Modified Philip-Dunne (MPD) Infiltrometer(Upstream Technologies,
New Brighton, MN; ASTM, 2018) at six locations within the gravel media basin: two upstream
(W 1 and El), two midstream(W2 and E2), and two downstream(W3 and E3). During
construction, subbase in-situ soil Ksat was also measured at two locations within the forebay:
upstream and downstream. The Ksat of the surface of the gravel media was measured post-
construction,pre-maintenance,post-maintenance, and at the end of the monitoring period.
Mtr
F( it , �'a� ¢ifs' a 0
fit
4 1 "
+rd"
. .. _ _, __ ( __
o —
Figure 10. Modified Phillip Dunne Infiltrometer measuring surface saturated hydraulic
conductivity (KsAT).
Event mean concentrations (EMCs) for nitrate+nitrite (NO3+2), ammonia(NH3), total
Kjeldahl nitrogen (TKN), total suspended solids (TSS), total phosphorus (TP), and ortho-
phosphate (Ortho-P; P043-to H2PO4)were monitored at the NW inlet and underdrain for 28
storm and four baseflow events. Also, five-day biological oxygen demand(BOD5), anionic
surfactants (MBAS), and dissolved and total cadmium (Cd), copper(Cu), lead (Pb), and zinc
(Zn) EMCs were monitored for 11 storm and four baseflow events. EMCs for oils and grease
(O&G)were monitored for the four baseflow events. Redox potential (Eh) and pH were recorded
at 2-min intervals within the gravel bed from December 2019 to March 2020 using a Multi-
Parameter Water Quality TROLL 9500 (In-Situ Incorporated, Fort Collins, CO). Dissolved
oxygen(DO)was recorded at 2-min intervals at the bottom and top of the gravel layer and in the
underdrain from November 2020 to November 2021 using HOBO U26 Dissolved Oxygen Data
Logger(Onset Computer Corporation, Bourne, MA).
Hydrologic Data Analysis
Monitoring well and atmospheric pressure data were used to calculate the depth of water
above the Water Level Loggers using Equation 2. At each monitoring well, the elevation and
location were surveyed to determine the absolute location.
21

D = YH2o(Dwell — Patm) (2)
Where: D is depth of water in monitoring well, m(ft)
7H2o is unit weight of water, 1000 kg/m3 (62.4 lbs/ft3)
Pwell is the absolute pressure in the well, Pa(psi)
patm is the atmospheric pressure, Pa(psi)
Depth of water above the Water Level Loggers was used to calculate depth of water
above the weir inverts using the difference in elevation. Discharge was calculated for each weir
based upon geometric configuration (Horton, 1906; Kindsvater and Carter, 1959; Shen, 1981).
The general modified discharge equation for v-notch (triangular-notch) sharp crested weirs is
given by Shen (1981) as described in Equation 3 and visualized in Figure 11.
8 0 5
Q _ Ce 15 2g tan 2 he (3)
Where: Q is discharge, m3/s (cfs)
Ce is the nondimensional effective (adjusted) coefficient of discharge defined in
Figure 37 and Figure 38
g is the acceleration due to gravity, m/s2 (ft/s2)
0 is the angle of the notch, degrees
he is the effective(adjusted)height(head)of the upstream liquid surface above
the vertex of the notch, he = h + kVh where kvh is defined in Figure 39 and
Table 3, m(ft)
22

•_t/1c IN.
Crest surface
7450
Upstream face
of weir plate
Detail of sides
B
of notch
\
i
\
A
g •N;1 t
P
Figure 11. Diagram of sharp crested contracted v-notch weir. Taken from (Shen, 1981).
Table 3. Triangular (v-notch) weir discharge coefficients, Ce, and head adjustments, kvh, as a
function of notch angle from Shen, (1981).
Triangular Notch Angle Coefficient of Discharge, Ce, Viscosity/Surface Tension
(degree) (unitless) Head Adjustment, kvh, m(ft)
22.5° 0.592 0.0027 (0.009)
45° 0.581 0.0015 (0.005)
60° 0.577 0.0012 (0.004)
90° 0.578 0.0009 (0.003)
120° 0.589 0.0009 (0.003)
23

A 60-degree v-notch sharp crested weir was installed at the underdrain(Figure 12) to
control the internal water storage (IWS) level within the gravel media and calculate discharge
using Equation 4.
Q = 0.251f(D + kvh)2.5 (4)
Where: Q is discharge, m3/s (cfs)
D is water depth above the invert of the weir, m(ft)
kvh is head adjustment for viscosity/surface tension as defined in Figure 39 and
Table 3, m(ft)
rI/
1110 A, 11,°M , t
v
fir
i
i 7 ,,
k "
11
• /
I
fil 1
III 4
Figure 12. Underdrain 60-degree sharp crested v-notch weir. Notice the free and fully ventilated
flow from the notch.
A 90-degree v-notch sharp crested weir was installed at the SW inlet (Figure 13) to
calculate discharge using Equation 5.
Q = 0.436 J(D + k„h)2.5 (5)
24

-.1 i...-�.. } � it �
4.4410
• I •
•,ate ` Jam_ "_ - \ ■ Pt
1 •
Figure 13. Southwest (SW) inlet looking downstream at the water level logger PVC well casing
followed by the sharp crested 90-degree v-notch weir.
The general modified discharge equation for suppressed and contracted rectangular sharp
crested weirs is given by Horton (1906) and Kindsvater and Carter(1959) in Equation 6.
Q = CeJbehe'5 (6)
Where:Ce is the effective coefficient of discharge for suppressed and contracted weirs
and is defined in Figure 41 and
Table 5, (unitless)
g is acceleration due to gravity, m/s2 (ft/s2)
be is the effective crest length, be = b + kb where kb is defined in Figure 40 and
Table 4, m(ft)
kb is the crest length correction, m(ft)
he is the effective head, he = h + krh, where krh is 0.91-mm(0.003-ft), m(ft)
25

8 \\\
A.
Figure 14. Schematic of contracted rectangular(rectangular notch) sharp crested weir. Taken
from Kindsvater and Carter (1959).
Table 4. Rectangular notch weir crest length correction factor, kb, as a function of contraction
ratio, b/B from Kindsvater and Carter, (1959).
Contraction Ratios,b/B, (unitless) Crest Length Correction Factor, kb, (unitless)
0 0.0065
0.20 0.0065
0.30 0.0070
0.40 0.0075
0.50 0.0100
0.60 0.0113
0.70 0.0132
0.80 0.0138
0.90 0.0125
1.00 -0.0035
Table 5. Sharp crested rectangular weir coefficient of discharge, Ce, as a function of h/P for
various contraction ratios, b/B. Interpreted and converted to nondimensional form from Figure
41 (Kindsvater and Carter, 1959) using M = a�9 Ce.
Contraction Ratios,b/B, (unitless) Equation for the Coefficient of Discharge, Ce,
as a Function of h/P
1.0 (Fully Suppressed) Ce = 3.42 + 0.4243 P
0.75 Ce = 3.36 + 0.1989 P
Compound 90-degree v-notch contracted rectangular sharp crested weirs were installed at
the NW(Figure 15) and SE (Figure 16) inlets to calculate discharge using Equations 7.
26

0.436J(D„ + kvh)2.5 D < H„
Q 0.436,5(H„ + kvh)2.5 + Cej(L + kb)(Dr + krh)1.5 D > II, (7)
Where: D,is the water depth above the invert of the v-notch weir, m (ft)
H,is the height of the v-notch weir before compounding, m (ft)
L is the crest length of the rectangular weir, m(ft)
Dr is the water depth above the compounding of the rectangular weir, m (ft)
....
a• •
Figure 15. Northwest (NW) inlet looking downstream at the water sample and bubbler tubing
followed by the compound sharp crested 90-degree v-notch and contracted rectangular weir.
Notice the hydraulic cement yet to be placed at the invert.
'.:, . wiP" .7P, ,...",„:10:44114,frrie-2011, t,
t --
i7c E, • .f.`- �,r yd 4�
gg
ye : ,. b'a{�; �.
, . ., •,; , '',, --4t.. .—..„1 . 1\
yl. F i i �i'lt� !r 4\ 7
Figure 16. Southeast(SE) inlet looking downstream at the water level logger PVC well casing
followed by a compound 90-degree v-notch and constricted rectangular sharp crested weir.
27

A 15-cm(6-in)breadth rectangular broad crested weir was used to set the temporary
inundation depth before overflow and calculate the discharge that bypassed the media filtration,
see Figure 17, using Equation 8 (Horton, 1906).
Q = 0.463.LD1.5 (8)
A----;------- -----'.--- , ,.. --,41- i , . ----,„st
7 .
r-.
A /
2,
/\ ,,,*;:i
, . ----"
1
. 7
i 4g1 ,... ,..
Figure 17. Broad crested overflow weir and original underdrain with upturned PVC T.
Volume was calculated for the discharge at each weir during the period between water
level recordings by using a Riemann Sum mid-point approximation, as shown in Equation 9. The
duration between recording intervals was 120-s. Storm volumes were calculated for the smaller
of 48 hours after rainfall or until the start of the next storm.
n
1.1 Q,At (9)
1=1
Where: V is volume of water, m3 (ft3)
i is the index of measured values from start to end of flow
n is the total number of time steps during the storm
At is the duration between depth records, s
Weirs installed at the outlet of ponds and SCMs provide ideal dimensions and flow
conditions that match those in the weir experimentation literature (Horton, 1906; Kindsvater and
Carter, 1959; Shen, 1981). Often weirs installed in the field, specifically at pond inlets, lack
calibration and are not the required ideal dimensions to create quiescent backwater at normal
depths upstream of the weir. Instead, the upstream channel produces initial velocities not
28

accounted for in empirical weir discharge equations. Therefore, volume was also estimated using
the Curve Number Method as described in Equation 1010.
((P — 0.2S)2 (10)
— uv A
P + 0.8S
Where: uv is the volume unit conversion from ha-mm to m3 or ac-in to ft3,
uv = 10 or uv = 3630 ft3
ha*mm ac*in
P is precipitation depth for storm of interest, mm(in)
S is the storage of the watershed before runoff, S = ((- °) — 10)uL, mm(in)
uL,is the length unit conversion from in to mm, uL = 25.4 m-m. or uL = 1 in
in in
A is the watershed area, ha(ac)
Continuous tipping bucket rain gauge rainfall was adjusted to the manual rain gauge
rainfall (Figure 18)with Equation 11. The adjusted rainfall was calculated for every 2-min
interval for each period between manual readings.
Rma
_ n
Rad] Rap (11)
Rtip
Where: Radj is adjusted depth of rainfall, mm(in)
Rman is the depth of rainfall recorded with the manual rain gauge for each site
visit, mm(in)
Rt;p is the depth of rainfall recorded with the tipping bucket rain gauge, mm (in)
29

S •
I. ti- 'a Iii' —0 tom t �
of '
t.,jv
1
Figure 18. Manual and tipping bucket rain gauge supported by a 10-cm (4-in) by 10-cm (4-in)
post. Notice the field bag, sampling bottles, and job box with solar panel and auto sampler. The
NW inlet is to the left and SSGW in the background to the right.
Continuous rainfall data was separated into storms based on antecedent rainfall. When
there were six hours without rain a new storm was defined. Storms less than 0.1-inch typically
produce minimal runoff are large error in flow measurement.
Subsurface hydraulic gradients were determined between the subsurface gravel wetland
water levels and the surrounding groundwater table of the watershed using Darcey's Law as
described in Equation 12.
Q = KSAT (dH I)A (12)
Where: Q is the seepage exchanged between the subsurface gravel wetland and
surrounding watershed, m3/hr(ft3/hr)
KSAT is the saturated hydraulic conductivity, m/hr(in/hr)
dH is the incremental change in head between the water levels within the
subsurface gravel wetland and surrounding groundwater levels, m (ft)
dl is the flow path of interest length between the subsurface gravel wetland and
groundwater levels, m(ft)
A is the surface area available for seepage perpendicular to the flow path, m3 (ft3)
30

Water Quality Data Analysis
To determine the EMCs of non-measured pollutants including total nitrogen(TN),
organic nitrogen(Org-N), particulate phosphorus (Part-P), and particulate Cd, Cu, Pb, and Zn
utilize the equations in Table 6.
Table 6. Water quality laboratory analysis methods and calculations.
Pollutant Method/Calculation
NO3+2 EPA Method 353.2 Rev 2.0 (1993)
SM 4500 NO3 F-2000
NH3 EPA Method 350.1 Rev 2.0 (1993)
SM 4500 NH3 G-1997
TKN EPA Method 351.2 Rev 2.0 (1993)
SM 4500 N Org D-1997
Org-N Org-N=TKN—NH3
TN TN=TKN+NO3+NO2
TSS SM 2540 D-1997
TP EPA Method 365.1 Rev 2.0 (1993)
SM 4500 P F-1999
Ortho-P EPA Method 365.1 Rev 2.0 (1993)
SM 4500 P F-1999
Part-P Part-P= TP—Ortho-P
BOD5 SM 5210 B-2001
MBAS SM 5540 C 2000
O&G EPA 1664 B
Dis-Cd EPA 200.8 Rev 5.4 (1994)
Part-Cd Part-Cd=Total Cd—Dis-Cd
Total Cd EPA 200.8 Rev 5.4 (1994)
Dis-Cu EPA 200.8 Rev 5.4 (1994)
Part-Cu Part-Cu=Total Cu—Dis-Cu
Total Cu EPA 200.8 Rev 5.4 (1994)
Dis-Pb EPA 200.8 Rev 5.4 (1994)
Part-Pb Part-Pb= Total Pb—Dis-Pb
Total Pb EPA 200.8 Rev 5.4 (1994)
Dis-Zn EPA 200.8 Rev 5.4 (1994)
Part-Zn Part-Zn=Total Zn—Dis-Zn
Total Zn EPA 200.8 Rev 5.4 (1994)
To determine the inflow and outflow pollutant loads utilize Equation 13 and assume that
any volume that overflows, therefore bypassing the media, does not get treated. In other words,
assume the inflow event mean concentration (EMC) is the same as the overflow EMC since
overflow water quality was not monitored.
31

M = Vc (13)
Where: M is the load or mass of substance passing a specified point, kg (lb)
V is the volume passing the same specified point, m3/s (ft3/s)
c is the event mean concentration of pollutant, mg/L (lb/ft3)
Wilcoxon Signed Rank Tests were used to determine the statistical difference between
paired inlet and underdrain median EMCs and loads for the 28 water quality sampled storms
(Helsel and Hirsch, 1992).
Area annual loading rates were calculated using the average annual precipitation for the
entire 21-month monitoring period as described in Equation 14.
Pmeasured M
W = (14)
APsampled
Where: W is the cumulative area annual loading rate of substance, kg/ha/yr(lb/ac/yr)
Pmeasured is the average annual precipitation recorded for the entire 21-month
monitoring period, mm/yr(in/yr)
A is the watershed area, ha(ac)
Psampled is the total precipitation recorded for all sampled storms, mm(in)
Results
Post-construction
Hydrology
From October 13th, 2019 to August 23rd, 2021 there were 226 storms ranging from 0.002
to 8.9-inches. There were 133 storms greater than 0.1-inch, 57 greater than 0.5-inch, and 24
greater than 1-inch (Figure 19). Thus, based on rainfall depth alone 76 storms or 57% of storms
were fully captured and treated (i.e, the green zone of Figure 19). Whereas, for a fully sized
system, 89% of storms would have been captured.
Water levels within the subsurface-flow gravel wetland remained below the surface for
more than 64% of the 759 days which the monitoring well was installed and recording (Figure
20). Water overflowed the outlet structure for 0.85% of the time.
The water levels within the subsurface gravel wetland were at a higher elevation than the
surrounding groundwater levels for 727 days or 96% of the monitoring period. Predominately,
the wetland was exfiltrating (losing)water into the ground as opposed to infiltrating (gaining)
groundwater, indicated by negative and positive values, respectively, in Figure 21.
32

10A0 .
•
•
•
•• • • • • •
L • •• •
' . Season
o . •
. • •
. • •• .� Fall
•
o • • • • a •• ••.• • • • • • • • ••• • •• • Spring
• .
n 0,10- • • • • • • • • •
$ •
•
• . • • Summer
o • ' • •'' •• • ' • • • Winter
F- • • . • •
e . •
c ••• •••S.
CC • • • • • • •
001- • • • • •• •• •OD •• •• ••••• • • • •
•• •
•
Oct Dec Feb Apr Jun Aug Oct Dec Feb Apr Jun Aug Od
Date 2019-2021
Figure 19. Adjusted rainfall totals per storm across the monitoring period. Storms above 1-inch
are highlighted in dark red, storms between 1 and 0.5-inch are highlighted in yellow, and storms
less than 0.5-inches are highlighted in green.
74 _ I I 1 Top of TE mporary ondi r
I
-73E- IL 11 t 111,1 1 I Ili.i Iiii _ ■ IV - I
J
Underdrain avert 6-in of Sand(ASTM C33) l iiiiiii
1'ii: "1�i in' ' - 4 1 111 11 41it a 1,11 111JiRlllullli1 1 141111111( IL IT
c IN �"t NI 1 ur131amplatra a p8 Stu1e)�r11 AIr U""1 '�Nr 1
\ y .�I 1 �J
Q 744- II •
o Perforated Underdrain(NTS)__ I.
• s
ra 12-in of Gravel(#57 Stone)
743-
AllirliWilliallig
742-
Aug Od Dec Fe^ -' Jun Aug Oct Dec Feb Apr Jun Aug Od Le
Date 2019-2021
Figure 20. Water levels within the subsurface flow gravel wetland during the monitoring period
with media layers, water level control structure, and maintenance activities. Note the dimensions
of the water level control structure and diameter of the underdrain are not to scale (NTS).
33

0 0000 I i 1
I !
-0.0025- -
t
Zz
— Ksat=9 9e-05 mlhr
, 1\1
00-0 0050 1 — Ksat=3.0e-02 INhr
a '', Ksat=6.9 e-01 i n h r
l 11I(1
-0.0075-
f
-0 0100-
Aug Oct Dec Feb Apr Jun Aug Oct Dec Feb Apr Jun Aug Oct Dec
Date 2019-2021
Figure 21. Estimated seepage between the subsurface gravel wetland and the surrounding
watershed groundwater throughout the monitoring period as a function of saturated subbase in-
situ hydraulic conductivity, KSA7:
Cumulative volumes from storm events show that consistently the inflow was larger than
the outflow(Error! Reference source not found.). However, 72 storms were removed due to
the large errors that occur during monitoring low head discharges less than 0.06-m(0.1-ft).
These storms, typically less than 2.5-mm (0.1-in), produced erroneous relatively large outflow
volumes not physically possible due to the lack of groundwater infiltration (Figure 21) and
relatively small inflow volumes. Of the 71,483-m3 (2,524,381-ft3) of cumulative inflow, for the
remaining 159 storms, 15%was reduced due to exfiltration to the surrounding groundwater table
or evapotranspiration (ET) to the atmosphere. Up to 2,100-m3 (75,000-ft3) of ET is estimated to
have occurred based on the product of wetland surface area and the 62-in of recorded ET at
North Carolina A&T Research Farm throughout the monitoring period. However, since
vegetation was not substantially established until a year after planting, the ET was likely only
863-m3 (30,465-ft3) or 1.2%reduction of inflow volume. Depending on the compacted subbase
hydraulic conductivity, an additional 0.24 to 1,600-m3 (8.4 to 58,000-ft3) of exfiltration(0.0003
to 2.3%volume reduction) is estimated to have been loss due to seepage into the groundwater
(Figure 21). In other words, up to 3.5% or the volume reduction can be attributed to exfiltration
and ET. Therefore, the additional 11.5%volume reduction was attributed to the leak in the
underdrain weir(Figure 12)that started a year after installation.
34

2500000-
2000000-
U
1500000-
0
— Cumulative Inflow
— Cumulative Outflow
P
1000000- — — Cumulative Flow Reduction
•
■
500000-
0-
Nov Jan Mar May Jul Sep Nov Jan Mar May
Date 2019-2021
Figure 22. Cumulative volume inflow, outflow, and flow reduction for storm events. Note
erroneous storms have been removed.
The distribution of outflow storm peak flows was consistently lower and less variable
than the inflow when considering all storms (Figure 23) and even lower and less variable when
removing the 72 erroneous storms (Figure 42). Median peak flow reductions were 71% for all
storms. By season, median peak flow reductions were 54, 25, 26, 88% for spring, summer, fall,
and winter, respectively(Figure 43).
The distribution of inflow volumes during storms were consistently more variable than
the outflow when considering all storms (Figure 24) and when removing the 67 erroneous storms
the inflow distribution increased and became relatively less variable (Figure 44). Median percent
volume reduction for all storms was -42%. By season, median percent volume reduction was -68,
-308, -11, and-9% for spring, summer, fall, and winter, respectively(Figure 45).
The best-fit site average surface saturated hydraulic conductivity at construction was
2,474 mm/hr(97.4 in/hr) and was reduced by up to 85% in some portions of the wetland during
the first four months of operation with an average reduction of 67% (±18%). After 2.3-years of
operation, the reduction was 99% (±4.2%) resulting in a best-fit site average surface saturated
hydraulic conductivity of 40 mm/hr(1.59 in/hr).
35

1 e+02-
i
1e.00-
Site
LL Inflow
N
_c 1 e-02
Outflow
a •
1 e-04-
Inflow Outflow
Sampling Location
Figure 23. Comparison of peak storm flow distributions using box and violin plots between
inflow and outflow of the subsurface gravel wetland. Note that all 226 storms are included.
1e- -
II __
�1e.03- Site
o II
E — 1=3 Inflow
; Outflow
� 1\ /1
_ ,
)..1(
1e- -
Inflow Outflow
Sampling Location
Figure 24. Comparison of storm flow volume distributions using box and violin plots between
inflow and outflow of the subsurface gravel wetland. Note that all 226 storms are included.
36

Water Quality
Underdrain EMCs were consistently lower and less variable for all pollutants monitored,
see Figure 25. The Wilcoxon Signed Rank Test showed that the inflow EMCs were 13.4, 1.7,
and 3.1 times higher than the underdrain flow EMCs for TSS, TN, and TP, respectively.
Additionally, for all pollutants the inflow EMCs were statistically different than the underdrain
flow EMCs (i.e.,p<<0.05), except for NH3 (p=0.3557).
TSS TN TKN
• 4. •
100- • 3 •
-
2-
2-
50 1 I 1 -I 1
0- Pollutant
SSGW-IN SSGW-UND SSGW-IN SSGW-UND ---.1'-IN SSGW-UND
TSS
i OMON NOX • NH3 El TIJ
E • 06- • 06-
El TKIJ
c 2- •
o • 0-4- 0 4- • El OrgN
•
i- • 02 I I • 02 El
• NOX
`� r I �_ NH3
0 0.0- 0-0- El TP
U SSGW-IN SSGW-UIJD SSGW-UNr SSGW-IN SSGW-UND
TP PartP OrthoP
El PartP
0.4- • • al 0rthoP
• 0-3. 0.20-
0.3
02 0.15 •
0.2- 0.10-
0.1 I t I 0.1 l 0.05•
0.0— 0.00- —
SSGW-IN SSGW-UND 3SGPr-IN SSGW-UND SSGW-IN SSGW-UND
Sampling Location
Figure 25. Sediment and nutrient comparison of inflow versus outflow concentration boxplot
distributions.
Treatment efficiencies improved after initial leaching and were maintained as the gravel
wetland aged, see Figure 26. Median EMC percent reductions were 92, 43, and 67% for TSS,
TN, and TP, respectively.
Percentiles (i.e., 1 - exceedance probability) showed than, for all pollutants, over 75% of
the sampled storms had higher inflow EMCs compared to underdrain flow EMCs (Figure 27).
When excluding NH3, there were over 92% of the sampled storms with higher inflow versus
underdrain flow EMCs.
For TSS, see Figure 28, nearly all the inflow EMCs were above the levels deemed poor
for aquatic macroinvertebrates in the Piedmont of North Carolina (i.e., > 7 mg/L; McNett, 2010).
Yet, 67% of the time the underdrain flow EMCs were lower than levels deemed good for aquatic
macroinvertebrates (i.e., <4 mg/L).
For TN, see Figure 29, only 25% of the inflow EMCs were above the levels deemed fair
for aquatic macroinvertebrates in the Piedmont of North Carolina (i.e., > 1.17 mg/L; McNett,
37

2010). However, more than 82% of the underdrain flow EMCs were below the levels deemed
good for aquatic macroinvertebrates (i.e., < 0.69 mg/L).
For TP, see Figure 30, more than 67% of the inflow EMCs were above the levels deemed
fair for aquatic macroinvertebrates in the Piedmont of North Carolina(i.e., >0.2 mg/L; McNett,
2010). Yet only 28% of the underdrain flow EMCs were below the levels deemed good for
aquatic macroinvertebrates (i.e., < 0.06 mg/L).
TSS TN TIN
0. 50- • : • • • so �• •• : • •
•. .
N
0_ • '• • • 0- •• • •
-200- • •
-50- • • -50- ' •
-400- -100- -100-
-500-, • -150-• -150- • Species
Jan May Sep Jan May Jan May Sep Jan May Jan May Sep Jan 1.lay
• TSS
OrgN NOX NH3 • TN
> •�• .. 100' 1 • . . • 0- •0. .
m 50- •. 4 • 3 .. 50- • :r . ' TKN
•• • -500
g0- •• 0- •• • • • OrgN
W • -1000
c -50- • • -50- • NOX
1500-100-
• NH3
V-100- . •
-150- •- -2000- • • •
01 Jan May Sep Jan May Jan May Sep Jan May Jan May Sep Jan May TP
'I:
• PartP
TP PartP OMOP
j• r• • 100 t • OMoP
M • 75- .•••
•. . 4 ••• • •Me • •r •.
50- • • • 0- •
• • • 50- • •
0 25 . -100- ••
-50- _ 0- • 200 •
Jan May Sep Jan May Jan May Sep Jan May Jan May Sep Jan May
Date 2019-2021
Figure 26. Sediment and nutrient percent reduction in concentration from inflow versus
underdrain flow as a function of time.
100- N
50- ✓. to
4-
• • Z
3-
1- '
2- O
E 1 z
c 0.6- IZ Site
0.4- •
b 02- X SSGW-IN
ac� 0.0 - • • • - - -
0.6- SSGW-l1ND
c 0.4- � r _
U 0.2- u
0.0
0.4r
0.3-
0 1- . • _ -•--.--• -•—• -
0.3 v
0.2- -
0.0
t_Arlo
0.00 0.25 0.50 0.75 1.00
Percentile Concentration[0,1J
Figure 27. Sediment and nutrient percentile concentrations comparing inflow versus underdrain
flow.
38

•
100 7
-r-• •-----��
Of i
Site
•
c to- -- SSGW-IN
• SSGW-UND
o -e
U
V7
1-
H
•
•
t- •
•
0.0'0 u 1 0.75 1.00
Percentile Concentration[0, 1]
Figure 28. TSS concentration comparison of inflow versus underdrain flow as a function of
percentile. Note the horizontal lines indicate poor(red) to good(green) macroinvertebrate water
quality in the Piedmont of North Carolina (McNett, 2010).
4-
3-
5 Site
+ SSGW-IN
2-
• SSGW-UND
0
U _.
F
-
0.00 0 25 0.50 0.75 1.00
Percentile Concentration[0,1]
Figure 29. TN concentration comparison of inflow versus underdrain flow as a function of
percentile. Note the horizontal lines indicate fair(yellow) to good(green) macroinvertebrate
water quality in the Piedmont of North Carolina (McNett, 2010).
39

0.4-
•
0.3-
Site
b • SSGW-IN
00.2 • SSGW-UND
U
0.1-
0.00 0.2° i E-' 0.75 1.00
Percentile Concentration[0. 1]
Figure 30. TP concentration comparison of inflow versus underdrain flow as a function of
percentile. Note the horizontal lines indicate fair(yellow) to good(green) macroinvertebrate
water quality in the Piedmont of North Carolina (McNett, 2010).
Pollutant loads for TSS, TN, and TP were consistently lower and less variable for the
inflow versus the outflow, see Figure 31, Figure 32, and Figure 33. However, for the one
summer storm that was sampled, the TN and TP outflow loads were higher than the inflow
(Figure 32 and Figure 33). The Wilcoxon Signed Rank Test showed that the inflow loads were
5.8, 1.2, and 1.9 times higher than the outflow loads for TSS, TN, and TP, respectively. For TSS
and TP the inflow loads were statistically different than the outflow loads (i.e.,p<<0.05), but not
for TN (p = 0.289).
Percent removal of loads for TSS, TN, and TP were consistent for spring and summer,
yet for summer there was only one sampled storm (Figure 34, Figure 35, Figure 36). However,
for fall and winter,percent removal of loads for TN and TP were less consistent and not always
positive.
40

Spring Summer
300
5.5
10D-
5.0-
30- 4.5-
10- 4.0-
rn Season
Spring
0 0 Summer
Fall tinter u 0 Fall
100 .".'inter
130 •
10- 1J-
•
-
1-
Inflow Outflow Inflow Outflow
Sampling Location
Figure 31. TSS load comparison of inflow versus outflow boxplot distributions as a function of
season.
Spring Summer
2.0- 0.4-
0 3-
1 0-
0.5
Season
Spring
Outflow nOC w Outflow
m 11 Summer
3 Fall Winter
Z 0 Fall
3.0- 0 Winter
3.0-
1D 1 0-
0.3- 0-3-
i1-
0.1-
Inflow Outflow Inflow Outflow
Sampling Location
Figure 32. TN load comparison of inflow versus outflow boxplot distributions as a function of
season.
41

Spring Summer
0.10-
0.05•
•
Season
0 immti
n 0 03- ra Spring
-0 - r Inflor, Outflow Summer
Fall rte
Fall
1
0 Winter
0.30-
L 0.10-
1
0-01- 0 03-
Inflow Outflow Inflow Outflow
Sampling Location
Figure 33. TP load comparison of inflow versus outflow boxplot distributions as a function of
season.
Spring Summer
36 94-
98-
36.92-
95- 3690-
36.88-
o ns-
0 36.86-
Season
°' 0 Spring
Spring Sum-.,.
c - il Summer
o
Fal I Winter Fall
a• 100- �- I
-0 IIMILF- 0- winter
m
0
J
Cl) -250• •
I- 0
-500-
-100-
-750•
Fall Winter
Season
Figure 34. TSS load percent reduction boxplot distributions as a function of season.
42

Spring 7 Summer
80-
iiE
-202_625
60-
-202.650
0
c 40- -202 675
0
Season
m -202.700 Spnng
CL Spring Summer
c e Summer
d Fall Winter
m 100� dFall
d ter-' Winter
m 0- 1-4!.11-1
0
0
J
F-100-
-200-
-200-
•-300- -400- •
-400-
Fall Winter
Season
Figure 35. TN load percent reduction boxplot distributions as a function of season.
SPIN Summer
86.0- -404 82-
85.5 40484-
-404.86-
_85.0-
-404.88-
m
0 84,5- 404 90
0 Season
I Spnng Summer * Spnng
eSummer
0 100 Fall Winter Fall
100-
d I 0 Winter
v
to 0-
J 0- I
-100-
-200- •
-100-
• -300-
-200- • -400- •
Fall Winter
Season
Figure 36. TP load percent reductions boxplot distributions as a function of season.
Median percent reductions for TSS, TN, and TP loads as a function of season are listed in
Table 7. Overall, the median percent reductions were 77, 23, and 50% for TSS, TN, and TP
loads, respectively. The areal annual loading rates were 388, 7.58, and 1.92 kg/ha/yr(346, 6.76,
and 1.71 lbs/ac/yr) for the inflow.
43

Table 7. Median sediment and nutrient load percent reductions as a function of season.
Percent Reduction(%)
Season TSS TN TP
Spring 95.6 54.3 85.1
Summer 36.9 -202.7 -404.9
Fall 75.5 17.8 43.0
Winter 92.4 40.0 60.4
Discussion
Despite being undersized, the subsurface-flow gravel wetland(SSGW) reduced
cumulative storm volumes, peak flows, EMCs, and pollutant loads. Loss of water via exfiltration
and ET from the SSGW was minimal. Thus, during the second year of monitoring, the water
balance was closed with the added losses attributed to a leaky wooden underdrain weir box
(Figure 12). Additionally, the ET was also likely to not have started until the vegetation was
established more than a year after construction(September 2020), see Figure 22. Even with
underestimated inflow discharge due to the lack of quiescent backwater conditions upstream of
the inlet weirs, the peak flows were still substantially reduced. The largest reduction in peak flow
was for winter, and this is common in cold climate gravel wetlands (Ballestero and Houle, 2015)
as lower temperatures promote plant senescence and less bioturbation.
The surface saturated hydraulic conductivity reduced by up to 85% in some places during
the fall (November 2019), four months after construction. This clogging was due to the
development of a schmutzdecke and it contributed to the reduced peak flows. Almost four
months later(March 2020), after the surface was raked (January 2020), the surface saturated
hydraulic conductivity had increased by up to 214% in some places to a best-fit site average of
1,611 mm/hr(63.4 in/hr). After being in operation for 2.3-years, even with a two logio reduction
in surface saturated hydraulic conductivity, the best-fit site average of 40 mm/hr(1.59 in/hr)was
still in the range of filtration-based practices within North Carolina(NCDEQ, 2020). Once the
plants established in the first growing season, they shaded and bioturbated the schmutzdecke
providing preferential flow paths along shoots, stalks, and roots. However, the surface organic
layer continued to accumulate at an average rate of 17.5 ± 6.25 mm/yr(0.69±0.25 in/yr).
Event mean concentrations (EMCs)were reduced to non-toxic levels for stream benthic
macroinvertebrates in the Piedmont of North Carolina. However, during the first few storms,
nutrients were leached from the gravel wetland due to newly planted plugs likely full of fertilizer
from the nursery, high infiltration rates reducing treatment contact time, lack of mature microbial
communities, and emergent copepod crustaceans around the outlet structure that excrete NH3.
During the second winter and spring (January to April 2021), the underdrain EMCs spiked up for
TN, TKN, OrgN,NH3, and OrthoP. The release of TKN and NH3 are likely due to the geese
eating plants,burrowing up to 15-cm(6-in) into the surface, and defecating in the wetland. The
lack of maintenance (i.e., raking of the surface) and additional plant senescence during the
second winter also could have contributed to the leaching of OrgN. The additional nitrogenous
44

oxygen demand during the winter combined with lower reaeration rates due to the progression of
clogging could have promoted the chemical reduction and leaching of OrthoP.
Sediment and nutrient loads were still reduced even without reliable volume reductions.
This trend was apparent for all seasons with adequate data(i.e., more than one storm sampled).
During the summer, only the TSS load was reduced while the TN and TP loads were exported
double and quadruple, respectively. However, spring provided the largest and most consistent
load reduction for all pollutants likely due to the increased plant and microbial activity fueled by
available carbon sources that had decayed over winter.
Conclusion and Recommendations
Minimum Design Criteria
Gravel Wetland MDC 1. Maximum Ponding Depth for Design Volume.
The maximum depth of ponding above the wetland surface shall not exceed 3-feet
provided the surface water quality volume to subsurface internal water storage volume
(WQV:IWS) does not exceed 5:1. Therefore, the deeper the temporary inundation is the
deeper the total soil and gravel layer needed using Equation 15.
WQV
DSOIL&GRAVEL � IWS cPDINUNDATION (15)
Where: DsOIL&GRAVEL is the total depth of the soil and gravel, m (ft)
WQV is the water quality volume generated from the contributing
watershed during the 25 to 38-mm(1 to 1.5-in) storm, mm(in)
IWS is the internal water storage volume that is retained in the subsurface
at normal pool between storms, m3 (ft3)
cp is the porosity or the ratio of pore/void volume and total volume,
untiless
DINUNDATION is the depth of the temporary inundation at which the WQV is
stored before overflow and treatment bypass, m(ft)
Gravel Wetland MDC 2. Required Pretreatment Device/Forebay.
An aerobic pretreatment device (i.e., swale, hydrodynamic separator, concrete catch
basin, or forebay)must be provided that is at least 10% of the WQV or wetland surface
area. Dry aerobic forebays are preferred over wet anaerobic forebays. Water level control
structures that drawdown the forebay between storms and can be adjusted to change the
normal forebay pool are recommended. Wet postbays (analogous to outlet deep pools in
wetlands) can reduce the amount of gravel and wetland soil needed.
Gravel Wetland MDC 3. Seasonal High Water Table Requirements.
45

Separation is preferred but intersecting the SHWT may provide saturation without the
need for impermeable liners in coarser soils (i.e., HSG A and B). Separation is required
when the groundwater is the drinking water(e.g., Blue Baby Syndrome).
Gravel Wetland MDC 4. Maximum Peak Attenuation Depth and Minimum Freeboard.
The maximum depth of peak storm flow shall not exceed 5-feet above the wetland
surface. There shall be at least 1-foot of freeboard. The invert of the overflow weir shall
not exceed 4-feet above the wetland surface.
Gravel Wetland MDC 5. Inlet and Underdrain Configuration.
Inlet subsurface distribution(e.g., perforated pipe, underground detention chambers) shall
be provided in coarse gravel and cobble aggregate. Secondary surface loading is
encouraged when surface to subsurface perforated distribution risers are also included.
The underdrain should set the IWS invert at 4 to 8-inches below the wetland soil surface.
The underdrain not be connected to the other distribution pipes, but a clean out should
still be provided. The underdrain should have an adjustable drawdown orifice (i.e.,
upturned PVC T with cap) to drain the IWS.
Gravel Wetland MDC 6. Minimum Subsurface Flow Length and Length to Width Ratio.
The water shall flow horizontally through the subsurface gravel layer for at least 30-feet
total. Vertical perforated risers shall promote surface to subsurface hydraulic connections
provided they are not connected to the underdrain. The length to width (L:W)ratio shall
be a minimum of 2:1. Baffles, berms, or cells in-series shall be provided when linear
systems are not feasible.
Gravel Wetland MDC 7. Media Depths and Mix Specification.
The media shall consist of, at a minimum, ¶-foot base layer of driveway gravel (i.e., #57
stone), 0.3-foot filter layer of chocker pea gravel (i.e., #78 stone), and 0.5-foot surface
wetland soil layer(i.e., 100% sand or bioretention mix, see Table 2). Size the layers
according to Equations 16 and 17.
D15,COARSESUBLAYER G 5D85,SETTINGBED (16)
D50,COARSESUBLAYER 25D50,SETTINGBED (17)
Where: D15,COARSE SUBLAYER is the 15th percentile particle size of the larger
material which the setting bed is placed on
D85,SETTING BED is the 85th percentile particle size of the smaller material
which is placed on the coarse sublayer
D50,COARSE SUBLAYER is the 50th percentile particle size of the larger
material which the setting bed is placed on
46

D5o,SETTING BED is the 50th percentile particle size of the smaller material
which is placed on the coarse sublayer
Gravel Wetland MDC 8. Media P-Index.
The phosphorus (p-index) shall not exceed 30 in nutrient sensitive watersheds (NSWs)
and shall not exceed 50 elsewhere.
Gravel Wetland MDC 9. Planting Plan.
Utilize wetland temporary inundation zone plantings. Herbaceous species planted on 2-
foot centers or less, shrub species on 5-foot centers, and trees on 10-foot centers.
47

References
Adyel, T. M., Hipsey, M. R., Oldham, C. E. (2017). Temporal dynamics of stormwater nutrient
attenuation of an urban constructed wetland experiencing summer low flows and
macrophyte senescence. Ecological Engineering, 102, 641-661.
https://doi.org/10.1016/j.ecoleng.2016.12.026
ASTM. (2018). Standard Practice for Measuring Field Infiltration Rate and Calculating Field
Hydraulic Conductivity Using the Modified Philip Dunne Infiltrometer Test.ASTM
International,D8152-18, 1-13. https://doi.org/10.1520/D8152-18.1.6
Ballestero, T. P., Houle, J. (2015). Design and Maintenance of Subsurface Gravel Wetlands.
Retrieved from
https://www.nh.gov/dot/org/projectdevelopment/materials/re se arch/proj ects/documents/156
80W-FINALREPORT.pdf
Ballestero, T. P., Houle, J. J., Puls, T. A. (2016). Univeristy of New Hampshire Stormwater
Center Subsurface Gravel Wetland Design Specifications, 1-11.
Bean, E. Z., Hunt, W. F., Bidelspach, D. A. (2007). Evaluation of Four Permeable Pavement
Sites in Eastern North Carolina for Runoff Reduction and Water Quality Impacts. Journal
of Irrigation and Drainage Engineering, 133(6), 583-592.
Bedan, E. S., Clausen, J. C. (2009). Stormwater Runoff Quality and Quantity From Traditional
and Low Impact Development Watersheds(1).Journal of the American Water Resources
Association, 45(4), 998-1008. https://doi.org/Doi 10.1111/J.1752-1688.2009.00342.X
Blecken, G.-T., Hunt, W. F., Al-Rubaei, A. M., Viklander, M., Lord, W. G. (2017). Stormwater
control measure (SCM) maintenance considerations to ensure designed functionality. Urban
Water Journal, 14(3), 278-290. https://doi.org/10.1080/1573062X.2015.1111913
Burchell, M. R., Skaggs, R. W., Lee, C. R., Broome, S., Chescheir, G. M., Osborne, J. (2007).
Substrate Organic Matter to Improve Nitrate Removal in Surface-Flow Constructed
Wetlands.Journal of Environment Quality. https://doi.org/10.2134/jeg2006.0022
Cizek, A. R., Hunt, W. F. (2013). Defining predevelopment hydrology to mimic predevelopment
water quality in stormwater control measures (SCMs). Ecological Engineering, 57, 40-45.
https://doi.org/10.1016/j.ecoleng.2013.04.016
Collins, K. A., Hunt, W. F., Hathaway, J. M. (2010). Side-by-Side Comparison of Nitrogen
Species Removal for Four Types of Permeable Pavement and Standard Asphalt in Eastern
North Carolina.Journal of Hydrologic Engineering, 15(6), 512-521.
https://doi.org/10.1061/ASCEHE.1943-5584.0000139
CRWA. (2009). Constructed Stormwater Subsurface Gravel Wetland.Low Impact Best
Managment Practice (BMP)Information Sheet.
Davis, A. P. (2007). Field Performance of Bioretention: Water Quality. Environmental
Engineering Science, 24(8). https://doi.org/10.1089/ees.2006.0190
Davis, A. P., Hunt, W. F., Traver, R. G., Clar, M. (2009). Bioretention Technology: Overview of
Current Practice and Future Needs.Journal of Environmental Engineering, 135(3), 109-
48

117. https://doi.org/10.1061/ASCE0733-93722009135:3109
Davis, A. P., Traver, R. G., Hunt, W. F., Lee, R., Brown, R. A., Olszewski, J. M. (2012).
Hydrologic Performance of Bioretention Storm-Water Control Measures.Journal of
Hydrologic Engineering, 17(5), 604-614. https://doi.org/10.1061/(ASCE)HE
DeBusk, K. M., Hunt, W. F., Hatch, U., Sydorovych, O. (2010). Watershed Retrofit and
Management Evaluation for Urban Stormwater Management Systems in North Carolina.
Journal of Contemporary Water Research &Education, 146(1), 64-74.
https://doi.org/10.1111/j.1936-704X.2010.00392.x
DNREC. (2019). Delaware Post Construction Stormwater BMP Standards & Specifications.
Dover, DE. Retrieved from
http://www.dnrec.delaware.gov/swc/Drainage/Documents/Sediment and Stormwater
Program/2019 February/Regulatory Guidance Documents 2019/BMP Stds and Specs - EFF
FEB 2019.pdf
Dotro, G., Langergraber, G., Molle, P.,Nivala, J., Puigagut, J., Stein, O., Von Sperling, M.
(2017). Treatment Wetlands. (G. Dotro, Ed.),Biological Wastewater Treatment Series.
London, UK: IWA Publishing. https://doi.org/10.1201/9781420012514
Greenway, M., Dale, P., Chapman, H. (2003). An assessment of mosquito breeding and control
in four surface flow wetlands in tropical-subtropical Australia. Water Science and
Technology, 48(5), 249-256. Retrieved from https://iwaponline.com/wst/article-
pdf/48/5/249/423483/249.pdf
Helsel, D. R., Hirsch, R. M. (1992). Statistical methods in water resources. Statistical Methods in
Water Resources. https://doi.org/10.2307/1269385
Horton, R. E. (1906). Weir Experiments, Coefficients, and Formulas. Washington D.C. Retrieved
from https://pubs.usgs.gov/wsp/0150/report.pdf
Hunt, W. F., Burchell, M. R., Wright, J. D., Bass, K. L. (2007). Stormwater wetland design
update:zones, vegetation, soil, and outlet guidance. Urban Waterways AGW-588-12.
Raleigh,NC.
https://doi.org/http://www.bae.ncsu.edu/stormwater/PublicationFiles/WetlandDesignUpdate
2007.pdf
Hunt, William F., Apperson, C. S., Lord, W. G. (2005). Mosquito Control for Stormwater
Facilities.North Carolina Cooperative Extension Service- Urban Waterways. Retrieved
from http://www.bae.ncsu.edu/stormwater/PublicationFiles/Mosquitoes2005.pdf
Hunt, William F, Lord, B., Loh, B., Sia, A. (2015). Plant Selection for Bioretention Systems and
Stormwater Treatment Practices (SpringerBr). SIngapore: Springer US.
https://doi.org/10.1007/978-981-287-245-6
Idris, S. M., Jones, P. L., Salzman, S. A., Allinson, G. (2012). Performance of the Giant Reed
(Arundo donax) in Experimental Wetlands Receiving Variable Loads of Industrial
Stormwater. Water, Air, and Soil Pollution, 223, 549-557. https://doi.org/10.1007/sl l 270-
011-0881-y
Kindsvater, C. E., Carter, R. W. (1959). Discharge Characteristics of Rectangular Thin-Plated
49

Weirs. Transactions of the American Society of Civil Engineers, 124(1), 772-801.
Knight, R. L., Walton, W. E., O'meara, G. F., Reisen, W. K., Wass, R. (2003). Strategies for
effective mosquito control in constructed treatment wetlands. Ecological Engineering, 21,
211-232. https://doi.org/10.1016/j.ecoleng.2003.11.001
Knox. (2008). Knox County Tennessee Stormwater Management Manual Submerged Gravel
Wetland, 2, 229-238.
Lemons, J. (2013). Methods and Materials for Golf Course Sand Bunker Construction.
Lenhart, H. A., Hunt, W. F., Burchell, M. R. (2012). Harvestable Nitrogen Accumulation for
Five Storm Water Wetland Plant Species: Trigger for Storm Water Control Measure
Maintenance?, 138(9), 972-978. https://doi.org/10.1061/(ASCE)EE.1943-7870.0000550
Mayo, A. W., Mutamba, J. (2004). Effect of HRT on nitrogen removal in a coupled HRP and
unplanted subsurface flow gravel bed constructed wetland. Physics and Chemistry of the
Earth, 29, 1253-1257. https://doi.org/10.1016/j.pce.2004.09.005
McCuen, R. H. (1979). Downstream Effects of Stormwater Management Basins. Journal of the
Hydraulics Division, 105, 1343-1356. https://doi.org/10.1002/9781118669709.ch15
McNett, J. (2010). Proposing a New Method of Stormwater BMP Assessment and Evaluating the
Toxicity of Forebay Sediments. North Carolina State University, Raleigh.
Merriman, L. S., Moore, T. L. C., Wang, J. W., Osmond, D. L., Al-Rubaei, A. M., Smolek, A.
P., ... Hunt, W. F. (2017). Evaluation of factors affecting soil carbon sequestration services
of stormwater wet retention ponds in varying climate zones. Science of The Total
Environment, 583, 133-141. https://doi.org/10.1016/j.scitotenv.2017.01.040
Miller, J. R., Kochel, C. R. (2013). Use and performance of in-stream structures for river
restoration: A case study from North Carolina. Environmental Earth Sciences, 68(6), 1563-
1574. https://doi.org/10.1007/s12665-012-1850-5
Miller, J. R., Kochel, R. C. (2009). Assessment of channel dynamics, in-stream structures and
post-project channel adjustments in North Carolina and its implications to effective stream
restoration. Environmental Earth Sciences, 59(8), 1681-1692.
https://doi.org/10.1007/s12665-009-0150-1
Molle, P., Prost-Boucle, S., Lienard, A. (2008). Potential for total nitrogen removal by
combining vertical flow and horizontal flow constructed wetlands: A full-scale experiment
study. Ecological Engineering, 34(1), 23-29. https://doi.org/10.1016/j.ecoleng.2008.05.016
Moore, T. L. C., Hunt, W. F. (2012). Ecosystem service provision by stormwater wetlands and
ponds: A means for evaluation? Water Research, 46, 6811-6823.
https://doi.org/10.1016/j.watres.2011.11.026
National Research Council. (2009). Urban Stormwater Management in the United States. The
National Academies Press. Washington, DC: The National Academies Press.
https://doi.org/10.17226/12465
NCDEQ. (2020).NCDEQ Stormwater Design Manual C-2. Bioretention Cell Design Objective.
50

Raleigh,NC.
Nivala, J., Knowles, P., Dotro, G., Garcia, J., Wallace, S. (2012). Clogging in subsurface-flow
treatment wetlands: Measurement, modeling and management. Water Research, 46(6),
1625-1640. https://doi.org/10.1016/J.WATRES.2011.12.051
NJDEP. (2014). New Jersey Stormwater Best Management Practice Manual Subsurface Gravel
Wetlands, 1-19.
Page, J. L., Winston, R. J., Mayes, D. B., Perrin, C. A., Hunt, W. F. (2014). Retrofitting
Residential Streets with Stormwater Control Measures over Sandy Soils for Water Quality
Improvement at the Catchment Scale.Journal of Environmental Engineering, 141(5).
https://doi.org/10.1061/(ASCE)EE
Page, J. L., Winston, R. J., Mayes, D. B., Perrin, C., Hunt, W. F. (2015). Retrofitting with
innovative stormwater control measures: Hydrologic mitigation of impervious cover in the
municipal right-of-way. Journal of Hydrology, 527, 923-932.
https://doi.org/10.1016/j.jhydro1.2015.04.046
Palmer, M. A., Hondula, K. L., Koch, B. J. (2014). Ecological Restoration of Streams and
Rivers: Shifting Strategies and Shifting Goals.Annual Review of Ecology, Evolution, and
Systematics, 45(1), 247-269. https://doi.org/10.1146/annurev-ecolsys-120213-091935
Passeport, E., Hunt, W. F., Line, D. E., Smith, R. A., Brown, R. A. (2009). Field Study of the
Ability of Two Grassed Bioretention Cells to Reduce Storm-Water Runoff Pollution.
Journal of Irrigation and Drainage Engineering, 135(4), 505-510.
https://doi.org/10.1061/ASCEIR.1943-4774.0000006
Rooney, R. C., Foote, L., Krogman,N., Pattison, J. K., Wilson, M. J., Bayley, S. E. (2015).
Replacing natural wetlands with stormwater management facilities: Biophysical and
perceived social values. Water Research, 73, 17-28.
https://doi.org/10.1016/j.watres.2014.12.035
Roseen, R. M., Ballestero, T. P., Houle, J. J. (2009). UNHSC Subsurface Gravel Wetland Design
Specifications. Durham,NH. Retrieved from
http://www.unh.edu/erg/csteyhttp://www.unh.edu/erg/cstev
Roseen, R., Watts, A., Houle, J., Farah, K., Ballestero, T. (2010). Investigation of Nutrient
Removal Mechanisms of a Constructed Gravel Wetland Used for Stormwater Control in a
Northern Climate. Durham,NH. Retrieved from
http://neiwpcc.org/neiwpcc_docs/Gravel WetlandNutrientCyclingFinalReport3-31-10.pdf
Saeed, T., Sun, G. (2012). A review on nitrogen and organics removal mechanisms in subsurface
flow constructed wetlands: Dependency on environmental parameters, operating conditions
and supporting media, 112, 429-448. https://doi.org/10.1016/j.jenvman.2012.08.011
Shen, J. (1981). Discharge Characteristics of Triangular-Notch Thin-Plate Weirs. US Geological
Survey Water Supply Paper, 1-45. https://doi.org/10.3133/wsp1617B
Silveira, D. D., Belli Filho, P., Philippi, L. S., Kim, B., Molle, P. (2015). Influence of partial
saturation on total nitrogen removal in a single-stage French constructed wetland treating
raw domestic wastewater. Ecological Engineering, 77, 257-264.
51

https://doi.org/10.1016/j.ecoleng.2015.01.040
Suliman, F., Futsaether, C., Oxaal, U., Haugen, L. E., Jenssen, P. (2006). Effect of the inlet—
outlet positions on the hydraulic performance of horizontal subsurface-flow wetlands
constructed with heterogeneous porous media.Journal of Contaminant Hydrology, 87(1-2),
22-36. https://doi.org/10.1016/J.JCONHYD.2006.04.009
US EPA. (1999). Treatment Wetland Habitat and Wildlife Use Assessment. Retrieved from
http s://nepis.epa.gov/Exe/ZyNET.exe/20004911.TXT?ZyActionD=ZyDocument&Client=E
PA&Index=1995+Thru+1999&Docs=&Query=&Time=&EndTime=&SearchMethod=l&T
ocRestrict=n&Toc=&TocEntry=&QField=&QFieldYear=&QFieldMonth=&QFieldDay=&I
ntQFieldOp=0&ExtQFieldOp=0&XmlQuery=
USEPA. (1999). Treatment Wetland Habitat and Wildlife Use Assessment. Gainesville, Florida:
U.S. Environmental Protection Agency Office of Wastewater Management.
USEPA. (2000a).Manual Constructed Wetlands Treatment of Municipal Wastewaters.
Cincinnati, Ohio, USA. Retrieved from
http://nepis.epa.gov/Exe/ZyPDF.cgi/3 0004TBD.PDF?Dockey=30004TBD.PDF
USEPA. (2000b). Wastewater Technology Fact Sheet Wetlands: Subsurface Flow. Washington,
DC. Retrieved from https://www3.epa.gov/npdes/pubs/wetlands-subsurface_flow.pdf
Vanasse Hangen Brustlin, I., Dipietro, T. (2012). City of South Burlington Stormwater Utility
Case Study-Farrell Park Gravel Wetland. South Burlington, VT. Retrieved from
http://www.sburlstormwater.com/wp-content/uploads/2012/12/FarrellParkGravel-
Wetland Final-12-3-12—FINAL4theWeb.pdf
Vymazal, J. (2013). The use of hybrid constructed wetlands for wastewater treatment with
special attention to nitrogen removal: A review of a recent development. Water Research,
47, 4795-4811. https://doi.org/10.1016/j.watres.2013.05.029
Vymazal, J. (2014). Constructed wetlands for treatment of industrial wastewaters: A review.
Ecological Engineering, 73, 724-751. https://doi.org/10.1016/j.ecoleng.2014.09.034
Waickowski, S., Hunt, W., UNRBA. (2017).North Carolina Stormwater Control Measure
Credit Document. Retrieved from https://files.nc.gov/ncdeq/Energy Mineral and Land
Resources/Stormwater/BMP Manual/SSW-SCM-Credit-Doc-20170807.pdf
Walsh, C. J., Roy, A. H., Feminella, J. W., Cottingham, P. D., Groffman, P. M., Morgan, R. P.
(2005). The urban stream syndrome: current knowledge and the search for a cure. Journal
of the North American Benthological Society, 24(3), 706-723. https://doi.org/10.1899/04-
028.1
Wright, J. D., Burchell, M. R., Evans, R. O., Shelby, J. D. (2006). Hydrologic Effect of Three
Restoration Techniques on a prior converted wetland in eastern nc. In Hydrology and
Management of Forested Wetlands, Proceedings of the ASABE International Conference
(pp. 138-149). ASAE Pub.
52

Appendix
0.63 --
0.62
0.8 0.5
U? 0.61 1.0 0.9 0.8-0.7 0.4
0.60 P/ /1.5 i __ I 0.3/
/ I I 0.2
tai
7 f +0.59 /- r - l-
r _ / (-I T
0.58 is +1 Pa. 0.1
1 0.578(avg)
-- -1 Pct
0.57 P limits Tt t
WC'/P of Figs. fr f f
0.56 1-
t. , -i , I .
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
VALUE OF h/P
Figure 37. Coefficient of discharge, Ce, as a function of h/P and P/B for 90-degree v-notch sharp
crested weirs. Notice that for weirs with P ten times smaller than B the value of Ce is 0.578
regardless of increasing head, h. Taken from (Shen, 1981).
0.61 - . . 1 1 , 1 1 1 1
Note: Average values of Ce in the range
0.60 of values of h/P from 0.03 to -
C� � 0.35 and P/B from 0.16 to 1.04
0.59 , - °‘>..„„.....Q.................ol ....-•11 I
D 0.58
Q
>
0.57 ,
0.56 - - -
0 20 40 60 80 100 120 140
VALUE OF 0, IN DEGREES
Figure 38. Coefficient of discharge, Ce, as a function of v-notch angle, 0. Taken from (Shen,
1981).
53

0.024
EXPLANATION
—❑ 4 Symbol Name
0.020 V Barr
X Cone
{ 0 Yarnell
0 Greve
w 0.016 �` } —� A Lenz
w
LL % + Numachi,and others
Z —\ L i
\
s
°` 0.012 \� —�
o \
D x
J
j 0.008
0 0 ❑
0.004 l x
x
o .
10 20 30 40 50 60 70 80 80 100 110 120
VALUE OF 0,IN DEGREES
Figure 39. Adjustment factor for viscosity and surface tension, kh, as a function of v-notch angle.
Taken from (Shen, 1981).
+0.015 - - - - - -
+0.010
c 0
.73
4it +0.005 , + : TTT: T \
w
0
a)
A 0
>
—0.005 - - - l
0 0.20 0.40 0.60 0.80 1.00
Value of b/B
Figure 40. Crest length correction factor, kb,for various weir contraction ratios, b/B. Taken
from (Kindsvater and Carter, 1959).
54

4.2 • T I
Legend
From experimental data
— — — Interpolated
4.0
1•
09
3.8
Cea3.22+0.40 h/P
v
U
$
° 0 3.6
d aC —
j i "� 4
° /
01 o r
i
ice ''
3.4 > -
i
i
�'' 0.5— ---'-1
�'' ---.— r 0.4
3.2 r ~" r
y 0.2
y
4,--
0
3.0
0 0.4 0.8 1.2 1.6 2.0 2.8
Value of h/P
Figure 41. Coefficient of discharge, Ce, as a function of h/P for sharp crested rectangular weirs.
Taken from (Kindsvater and Carter, 1959).
55

1e*00-
1 Site
5 I'
inflcn
__ Out
1e
Inflow Outflow
Sampling Location
Figure 42. Comparison of peak storm flow distributions using box and violin plots between
inflow and outflow of the subsurface gravel wetland. Note that only the 159 non erroneous
storms are included.
100
o- V
100-
-200-
-300-
-400 O
a -500-
A
v -600- ((Season
1 -700- IT Spnng
o -800- Summer
C•I_ -900- • Fall
g-1000- 1.1 Winter
-1100-
re
d-1200-
-1300-
-1400-
-1500-
-1600-
-1700-
Spring Summer Fall Winter
Season
Figure 43. Peak storm flow percent reduction boxplot and violin distributions as a function of
season. Note that all 226 storms are included.
56

Site
Inflow
• Outflow
1e+
y
Inflow Outflow
Sampling Location
Figure 44. Comparison of volume of storm flow distributions using box and violin plots between
inflow and outflow of the subsurface gravel wetland. Note that only the 159 non erroneous
storms are included.
0-
-1000-
-2000- •
•
a -3000-
c
0
• Season
-4000
Spnng
-• -5000- • Summer
o -6000 • Fall
a • Winter
0
0 -7000-
-8000-
-9000-
-10000-
Spnng Summer Fall Winter
Season
Figure 45. Volume percent reduction boxplot and violin distributions as a function of season.
Note that all 226 storms are included.
57