SW6221006_Design Calculations_20230926

FILTERRA® : ANALYSIS OF
LONG-TERM PERFORMANCE
Prepared by
Contech Engineered Solutions LLC
9100 Centre Pointe Drive
West Chester, OH 45069
800-338-1122
Authors
Mindy Hills, Vaikko Allen, John Pedrick, Alex MacLeod, P.E.
C=:: NTECH°
ENGINEERED SOLUTIONS
Ito r ra®
January 1, 2023

FILTERRA°: ANALYSIS OF LONG-TERM PERFORMANCE
CONTENTS
Abstract 5
Background 5
Methods 10
Sampling 10
Hydraulic Evaluation 12
Biofiltration Media Composition Sampling and Analysis Techniques 13
Results 14
Sediment 16
Nutrients 18
Metals 32
Oil& Grease 41
Temperature and pH 43
Hydraulic Capacity 44
Biofiltration Media Composition 44
Plant Growth Progression 45
Discussion and Conclusion 46
Organic Matter 46
Vegetation 47
Root-zone Macro-and Microorganisms 47
Media QA/QC 48
Routine Maintenance 48
Performance Comparison to High Rate Biofiltration 49
Hydraulic Performance 50
Biofiltration Media Composition Correlation to Performance 51
Conclusion 54
References 54
Appendix A: Photo Log 56
Filterra Plant Progression Photos 56
Filterra Maintenance Photos 73
Hydraulic Evaluation Photos 76
Appendix B:Activation and Maintenance Records 80

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
TABLES
Table 1. Bioretention and high rate biofiltration performance for TSS and total phosphorus from
the 2020 Summary Statistics Report by the International Stormwater BMP Database 7
Table 2. Bioretention and high rate biofiltration performance for zinc and copper from the 2020
Summary Statistics Report by the International Stormwater BMP Database 7
Table 3. Filterra study test sites 8
Table 4. Water quality parameter reporting limits and analytical methods. 11
Table 5. Water quality parameter sample handling requirements 11
Table 6. Water quality performance summary for all Filterra study sites 15
Table 7. Study site A hydraulic evaluation summary 44
Table 8.Aggregate classification profile comparison 45
Table 9. Organic content profile comparison 45
Table 10. Plant growth progression 46
Table 11. Filterra long-term performance versus high rate biofiltration performance as reported
in the 2020 Summary Statistics Report by the International Stormwater BMP Database 49
Table 12. Study site A activation and maintenance record 80
Table 13. Study site B activation and maintenance record 81
Table 14. Study site C activation and maintenance record 82
FIGURES
Figure 1(a)(b). Study site A 2014 9
Figure 2(a)(b). Study site B 2009 9
Figure 3(a)(b). Study site C 2014 10
Figure 4. Field equipment flow diagram 12
Figure 5.TSS descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time 17
Figure 6.TSS line plot analysis comparing influent and effluent concentrations (a)Study site A
(b)Study site B (c) Study site C 17
Figure 7.TP descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time 19
Figure 8.TP line plot analysis comparing influent and effluent concentrations (a)Study site A (b)
Study site B (c) Study site C 20
Figure 9.TDP descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time 21
Figure 10.TDP line plot analysis comparing influent and effluent concentrations (a) Study site A
(b) Study site B (c) Study site C 22
Figure 11. OP descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time 23
Figure 12. OP line plot analysis comparing influent and effluent concentrations (a) Study site A
(b)Study site B (c) Study site C 24
January 1, 2023 Con tech Engineered Solutions LLC 1

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Figure 13.TN descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time 25
Figure 14.TN line plot analysis comparing influent and effluent concentrations (a) Study site A
(b) Study site B (c) Study site C 26
Figure 15. NO2, NO3-N descriptive statistical analysis (a) box and whisker plots of influent and
effluent concentrations (b) regression scatter plot of effluent concentrations and time 27
Figure 16. NO2, NO3-N line plot analysis comparing influent and effluent concentrations (a)
Study site A(b) Study site B (c) Study site C 28
Figure 17.TKN descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time 29
Figure 18.TKN line plot analysis comparing influent and effluent concentrations (a)Study site A
(b) Study site B (c) Study site C 30
Figure 19. NH4 descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time 31
Figure 20. NH4 line plot analysis comparing influent and effluent concentrations (a) Study site A
(b) Study site C 31
Figure 21.Tot. Zn descriptive statistical analysis (a) box and whisker plots of influent and
effluent concentrations (b) regression scatter plot of effluent concentrations and time 33
Figure 22.Tot. Zn line plot analysis comparing influent and effluent concentrations (a) Study site
B (b) Study site C 33
Figure 23.Tot. Cu descriptive statistical analysis (a) box and whisker plots of influent and
effluent concentrations (b) regression scatter plot of effluent concentrations and time 34
Figure 24.Tot. Cu line plot analysis comparing influent and effluent concentrations (a) Study site
B(b) Study site C 34
Figure 25.Tot. Cd descriptive statistical analysis (a) box and whisker plots of influent and
effluent concentrations (b) regression scatter plot of effluent concentrations and time 35
Figure 26.Tot. Cd line plot analysis comparing influent and effluent concentrations at Study site
B 36
Figure 27.Tot. Cr descriptive statistical analysis (a) box and whisker plots of influent and
effluent concentrations (b) regression scatter plot of effluent concentrations and time 37
Figure 28.Tot. Cr line plot analysis comparing influent and effluent concentrations at Study site
B 37
Figure 29.Tot. Pb descriptive statistical analysis (a) box and whisker plots of influent and
effluent concentrations (b) regression scatter plot of effluent concentrations and time 38
Figure 30.Tot. Pb line plot analysis comparing influent and effluent concentrations (a) Study site
B (b) Study site C 39
Figure 31.Tot. Ni descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time 40
Figure 32.Tot. Ni line plot analysis comparing influent and effluent concentrations at Study site
B 40
January 1, 2023 Con tech Engineered Solutions LLC 2

FILTERRA°: ANALYSIS OF LONG-TERM PERFORMANCE
Figure 33. O&G descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time 41
Figure 34. O&G line plot analysis comparing influent and effluent concentrations (a) Study site B
(b) Study site C 42
Figure 35.TPH descriptive statistical analysis box and whisker plots of influent and effluent
concentrations 43
Figure 36.TPH line plot analysis comparing influent and effluent concentrations at Study site B
43
Figure 37. Study site A Filterra media profile composition analysis of organic content after 14
years in operation 52
Figure 38. Study site A Filterra media profile composition analysis of silt and clay content after
14 years in operation 53
Figure 39. Study site A plant progression 2008 56
Figure 40. Study site A plant progression 2010 56
Figure 41. Study site A plant progression 2012 57
Figure 42. Study site A plant progression 2014 57
Figure 43. Study site A plant progression 2015 58
Figure 44. Study site A plant progression 2018 58
Figure 45. Study site A plant progression 2019 59
Figure 46. Study site A plant progression 2019, cut to single stem 59
Figure 47. Study site A plant progression 2020 60
Figure 48. Study site A plant progression 2021 60
Figure 49. Study site B activation 2005 61
Figure 50. Study site B motor oil from oil service station caked on mulch surface in 2005 61
Figure 51. Study site B plant replacement with Northern Bayberry from heavy motor oil
contamination in 2005 62
Figure 52. Study site B motor oil residue evident in drainage area feeding Filterra system 200662
Figure 53. Study site B plant progression 2006 63
Figure 54. Study site B plant progression 2008 63
Figure 55. Study site B plant progression 2009 64
Figure 56. Study site B plant progression 2009 64
Figure 57. Study site B plant progression 2009 65
Figure 58. Study site B continual motor oil contamination to Filterra system; monitoring
discontinued 65
Figure 59. Study site C plant progression 2005 66
Figure 60. Study site C plant progression 2007 66
Figure 61. Study site C plant progression 2008 67
Figure 62. Study site C plant progression 2008 67
Figure 63. Study site C plant progression 2009 68
Figure 64. Study site C plant progression 2010 68
Figure 65. Study site C plant progression 2011 69
January 1, 2023 Contech Engineered Solutions LLC 3

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Figure 66. Study site C plant replacement 2012 with Foster Holly; photo taken 2013 69
Figure 67. Study site C plant progression 2014 70
Figure 68. Study site C plant progression 2015 70
Figure 69. Study site C plant progression 2016 71
Figure 70. Study site C plant progression 2017 71
Figure 71. Study site C plant progression 2018 72
Figure 72. Study site C decommissioning 72
Figure 73(a)(b). Study site A pre-maintenance (a) and healthy roots evident post-maintenance
prior to mulch replacement (b) 2017 73
Figure 74. Study site A pre-maintenance 2019 74
Figure 75. Study site A healthy roots evident post maintenance prior to mulch replacement 2019
74
Figure 76(a)(b). Study site C pre-maintenance (a)and post-maintenance (b) 2016 75
Figure 77(a)(b). Study site C pre-maintenance (a)and healthy roots evident post-maintenance
prior to mulch replacement (b) 2017 75
Figure 78. Study site A Filterra system hydraulically tested 2021 76
Figure 79. Hydrant and meter set up 76
Figure 80. Flow monitoring equipment set up 77
Figure 81. Curb influent flow 77
Figure 82. Initial flow during unsaturated test phase 78
Figure 83. Head development during unsaturated test phase 78
Figure 84. Worms present during hydraulic evaluation 79
January 1, 2023 Con tech Engineered Solutions LLC 4

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
ABSTRACT
In situ performance of conventional, slow-flow bioretention and high rate biofiltration systems is
typically characterized via load and concentration reduction performance over a typical water
year. However, long-term performance monitoring is also beneficial,since bioretention soil media
is expected to have a useful life measured in decades without the need for replacement.
This report aims to advance our knowledge of biofiltration longevity through research that
assesses the ability of high rate biofiltration systems (designed at infiltration rates of 100 inches
per hour or greater)to provide consistent water quality and hydraulic performance at the decadal
time scale with standard maintenance.
Since 2007, Contech Engineered Solutions LLC has continuously evaluated water quality
performance at three different high rate biofiltration sites in the Mid-Atlantic via third-party field
sampling and laboratory analysis. All sites utilize Filterra®, a proprietary biofilter that shares
primary pollutant removal processes with conventional bioretention but has a much higher design
infiltration rate and a footprint that is typically less than 10%the size of conventional bioretention
systems. The study sites represent different sizes, ages, and land uses. All sites received regular
maintenance. Monitoring duration varied by site, ranging upwards to 13 years.
Hydraulic testing was conducted in a saturated and unsaturated condition to assess changes in
hydraulic capacity over time, for which suggested research methods and results are presented.
Media profile composition and plant growth progression are studied to understand changes in
system dynamics.
Based on available data, Filterra continues to meet or exceed performance expectations,
suggesting that routine maintenance is adequate to maintain design functionality long-term
without replacement of the bioretention soil media. Filterra performance continues to be similar
or better than conventional bioretention performance for TSS, nutrients and metals. Sampling
methods and results are presented for parameters of interest including TSS, nutrients, metals, oil
and grease, pH and temperature. Recommendations are made for ensuring long-term water
quality compliance, with a focus on providing proper and timely maintenance.
BACKGROUND
Conventional bioretention systems have been used to reduce stormwater runoff volume and rate,
and to improve the quality of stormwater discharge for over two decades. Design characteristics
vary regionally, but typically include a vegetated media bed of at least 18 inches in depth
comprised of a blend of sand and compost or topsoil.Typical design infiltration rates range up to
12 inches per hour with resultant ratios of bioretention area to contributing effective impervious
area of less than 10%. Where the long term reliable native soil infiltration rate is high enough to
infiltrate the entire design storm, bioretention systems are often designed without an underdrain
and effectively have no downstream discharge in routine storm events. Where native soil
infiltration rates are insufficient to eliminate runoff during the design storm, an underdrain is
typically installed within a gravel drainage layer below the bioretention soil. In this configuration,
January 1, 2023 Con tech Engineered Solutions LLC 5

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
a portion of the design storm is treated and released downstream, and the effluent quality of the
system must be considered.Although terminology varies regionally,with terms like bioretention,
rain garden and biofiltration often used interchangeably,for the purposes of this report,the term
bioretention refers to those systems that retain the design storm via infiltration after filtration
through vegetated soil media. The term biofiltration refers to those systems that release all or a
portion of the design storm through an underdrain after filtration through vegetated soil media.
The Filterra system was developed as a compact, high-performance alternative to conventional
bioretention and biofiltration and can be designed to retain and infiltrate all or a portion of the
design storm. Like conventional systems, it includes vegetation, a mulch layer for pretreatment
and moisture retention, engineered soil media and an underdrain that discharges treated
stormwater. Unlike conventional systems, the design infiltration rate of the filtration media
ranges from 50 to 175 inches per hour depending on local approvals. This significantly higher
infiltration rate is made possible by stringent quality control practices that ensure media
consistency,and by standardized design,construction,activation and maintenance practices. Due
to the smaller size of the Filterra system, the media experiences a higher pollutant load per area
and volume. However,the mulch layer over the top of the media is designed to protect the media
from degradation over time.
Filterra systems have been tested extensively following peer reviewed testing protocols including
the TAPE protocol (Ecology 2018).These studies have all been conducted on systems less than 5
years old and over a duration of 1 to 3 years. These tests have demonstrated that the pollutant
removal capabilities of the Filterra system meet or exceed the performance goals set by the
Washington State Department of Ecology TAPE program which is the premier field-based
stormwater control measure performance verification program in the United States. This study
was initiated to assess the ability of Filterra systems to remove common pollutants over time as
the system ages.
The International Stormwater BMP (Best Management Practice) Database is a clearinghouse for
test results from stormwater BMP field performance studies. In 2020, a performance summary
was released summarizing water quality performance of 14 types of stormwater control measures
for common pollutants (Clary et al. 2020). Included in the summary report is performance data
for Bioretention (BR) and High Rate Biofiltration (HRBF) which are defined as follows:
"Bioretention - Shallow, vegetated basins with a variety of planting/filtration media and
often including underdrains. Also called rain gardens and biofiltration"
"High Rate Biofiltration - Manufactured devices with high-rate filtration media that
support plants."
Although there are several types of proprietary biofiltration systems commercially available, the
data for HRBF in the summary report is entirely comprised of results from 6 Filterra sites, since
that was the only proprietary biofilter data in the database at the time of the report.The BR data
includes influent data from 43 studies and effluent from 41 studies.
January 1, 2023 Con tech Engineered Solutions LLC 6

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
The 2020 summary report concludes that both BR and HRBF provide significant removal of TSS.
HRBF also provides significant reduction in total phosphorus concentration, but runoff treated by
BR shows a significant increase in total phosphorus (Table 1).
Table 1. Bioretention and high rate biofiltration performance for TSS and total phosphorus from
the 2020 Summary Statistics Report by the International Stormwater BMP Database
Parameter TSS Total Phosphorus
Units (mg/L) (mg/L)
Stormwater Control BR HRBF BR HRBF
Measure
Median Influent 44 30.8 0.19 0.099
Value Effluent 10 3.8 0.24 0.05
Significant Median Value
Reduction (Mann Whitney Yes Yes Significant Yes
P-value 0.05) export
HRBF provided a significant reduction in total and dissolved copper and zinc. BR provided
significant reduction of total and dissolved zinc and total copper but reduction in dissolved copper
was insignificant (Table 2).
Table 2. Bioretention and high rate biofiltration performance for zinc and copper from the 2020
Summary Statistics Report by the International Stormwater BMP Database
Parameter Total Copper Total Zinc Dissolved Dissolved Zinc
Copper
Units (µg/L) (µg/L) (µg/L) (µg/L)
Stormwater BR HRBF BR HRBF BR HRBF BR HRBF
Control Measure
Median Influent 13.1 7.95 62 178 6.85 4.5 20.8 189
Value Effluent 7.13 3.75 12.8 60.6 7.54 3.4 12.5 79
Significant Median
Value Reduction Yes Yes Yes Yes No Yes Yes Yes
(Mann Whitney P-
value 0.05)
This report will compare the performance of Filterra systems over time to Filterra results from
short-term (1-3 years)testing under the TAPE protocol and other similar protocols from the 2020
Stormwater BMP Database Report. Changes over time in factors contributing to water quality
performance, including the saturated and unsaturated infiltration rate of the Filterra media,
media composition and plant growth will be investigated as well.
Long-term performance monitoring is helpful since bioretention soil media is expected to have a
useful life measured in decades without the need for replacement. This report aims to advance
January 1, 2023 Con tech Engineered Solutions LLC 7

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
our knowledge of biofiltration longevity through research that assesses the ability of high rate
biofiltration to provide consistent water quality and hydraulic performance at the decadal time
scale with standard maintenance.
Three Filterra study sites ranging in system age, size and land use were monitored in Maryland
and Virginia for various pollutants of concern to verify consistency in long-term performance
(Figure 1(a)(b), Figure 2(a)(b), Figure 3(a)(b)). Monitoring duration varied by study site ranging
from 3 to 13 years and covered system ages from 1 to 13 years since activation. Activation is
defined as when Filterra begins treating runoff after removal of flow barriers to protect the system
from construction-phase runoff and installation of vegetation, mulch, and dissipation stone.
Pollutant removal efficiency was evaluated via third-party field sampling and laboratory analysis
procedures. Pollutants monitored varied by site, including total suspended solids (TSS),
phosphorus, nitrogen, heavy metals, and oil and grease. Study test site descriptions are provided
in Table 3 below. Photos taken throughout the study period are provided for each study site in
Appendix A: Photo Log
Table 3. Filterra study test sites
Study Site ID A B C
Restaurant Oil Service Station Gas Station Retail
Land Use Commercial Parking Commercial Parking
Lot Lot Area
Location Virginia Beach,VA Baltimore, MD Hampton, VA
System Size (ft.) 6x4 6x6 6x8
Redtwig Dogwood,
Plant Type Nellie Stevens Holly Northern Bayberry Foster Holly
Activation Date 4/13/2007 6/1/2005 5/27/2005
Age at Time of
1- 11 3 -6 0- 13
Monitoring (yrs.)
Time Monitored (yrs.) 10 3 13
January 1, 2023 Con tech Engineered Solutions LLC 8

FILTERRA°: ANALYSIS OF LONG-TERM PERFORMANCE
.tea --_—,-.>w.es
i '''
k---
77
4 .: •� r• -�
L
-i'
.fi. ,4 , 414,
lc
.T `, 4 ,
:°! !
(b)
Figure 1(a)(b). Study site A 2014
j
^'j
itlft-
E
:e•r •R.,�•• _
/
(a) (b)
Figure 2(a)(b). Study site B 2009
January 1, 2023 Con tech Engineered Solutions LLC 9

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
:h ,N H '.. u' ,. •
•
•
q
(a) (b)
Figure 3(a)(b). Study site C 2014
The three test systems were sized according to Filterra regulatory approvals at the time of plan approval.
Systems were sized to treat design storms of% inch and 1 inch in Virginia and Maryland, respectively, at
the then current state-approved design infiltration rate of 100 inches/hr("j 1 gpm/sqft) based on test data
available at the time of the approvals. Sizing criteria was historically established following the rainfall
distribution and frequency data from the mid-Atlantic region to ensure 90%treatment of the total annual
rainfall volume.
METHODS
Sampling
Grab samples were collected following EPA sampling guidance (1992) over a 30-minute duration at 10-
minute intervals. Influent and effluent sampling were paced 5 minutes apart to allow for proper detention
time. Samples were collected near the beginning of the storm to capture the pollutant first flush.
Acceptable storm event criteria included (1) antecedent conditions of at least 6 hours of no greater than
trace precipitation and (2) more than 0.1 inch of total rainfall depth.
Sample information at collection included sample name, lab sample ID, date, sample times, sample
collection type, sampler name, parameters collected, sample container type, preservation, and any
unusual circumstances that may impact sample results. Samples were collected in preserved bottles as
appropriate, placed on ice and transferred to the appropriate analytical laboratory with a complete chain
of custody.
Influent sample collection occurred at the Filterra throat opening just before runoff entered the system
and effluent sample collection occurred at the Filterra outlet pipe invert into the downstream catch basin.
Sample collection and analysis was performed by Universal Laboratories at study sites A and C and
Microbac Laboratories at study site B. Laboratory contact information is found below.
January 1, 2023 Contech Engineered Solutions LLC 10

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Universal Laboratories
Project Manager:Dan Thornton, Project Supervisor
20 Research Drive
Hampton, VA 23666
[phone]
Microbac Laboratories, Inc.
Project Manager:Michael Arbaugh Sr., Division Manager
2101 Van Deman St.
Baltimore, MD 21224
[phone]
Reporting limits and analytical methods for water quality parameters are shown in Table 4.
Table 4. Water quality parameter reporting limits and analytical methods.
Reporting Limit(mg/L) Reporting Limit(mg/L)
Parameter Method (Universal) (Microbac)
Total Suspended Solids
(TSS) SM 2540 D 1.0 <10.0, 1.0
Total Phosphorus EPA 365.1 0.02 0.01
SM 4500
Dissolved Phosphorus P/B/E 0.02 0.01
Total Copper EPA 200.7 0.001, 0.005 0.005
Total Zinc EPA 200.7 0.005 0.005
Oil & Grease EPA 1664A 5.0 5.0
The required analytical container, sample handling, preservative and maximum allowable holding time
limits for each parameter are shown in Table 5.
Table 5. Water quality parameter sample handling requirements
Parameter Container Preservative Max. Allowable
Hold Time
Total Suspended Solids 1000 mL HDPE <6°C 7 days
(TSS)
Total Phosphorus 250 mL glass H2SO4 to pH<2/<6°C 28 days
Dissolved Phosphorus 250 mL glass filtration/H2SO4 to pH<2/<6°C 28 days
Total Copper 250/500 mL plastic HNO3 to pH<2/<6°C 6 months
Total Zinc 250/500 mL plastic HNO3 to pH<2/<6°C 6 months
Oil & Grease 1000 mL glass HCI to pH<2/<6°C 28 days
January 1, 2023 Contech Engineered Solutions LLC 11

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Hydraulic Evaluation
Field hydraulic evaluation was performed in July 2021 on the 14-year-old Filterra system at study site A to
correlate flow and quality performance. Hydraulic evaluation was not performed at the other two study
sites due to an oil spill at the oil service station at study site B that required full remediation and lack of
accessibility due to decommissioning of the Filterra system for site reconstruction at study site C.
Source Water and Flow Control
A fire hydrant, in combination with a rented city meter, was used as the influent supply for the field test
system (Figure 4).Source water from the hydrant was controlled manually with a ball valve on the hydrant
meter and directed to a flow meter via a combination of 2-inch flexible fire hose and 2-inch PVC pipe. Flow
was measured by a factory-calibrated Seametrics EX810 electromagnetic flowmeter and logged at a
minimum of 1 minute intervals.The logged flow data was used to verify that testing was conducted at the
target flow rates. Influent water was conveyed into 2-inch PVC piping connected to an upright-positioned
factory-calibrated rotameter (by King Instrument Company, manufacturer number 7205026163W). The
influent flow rate was regulated via a 2-inch globe valve on the discharge of the rotameter and adjusted
for the appropriate flow rate following the protocol. Water from the valve was directed into a section of
2-inch fire hose that discharged water into the gutter, mimicking real-world conditions for runoff. The
flow rate was held steady during the test at ±10% of the target value. The flow meters were calibrated
together at the test's start via time-bucket method,where water was introduced into a gradated tank and
timed. A ruler was secured to the inside of the tree grate frame to monitor fluctuations in the water
surface level in the head space of the Filterra system (Hills 2021).
Flow control valve
Rota meter
Water meter&valve assembly
WM
Hydrant Flexible hose Supports Magmeter To Test Unit
=100"straight pipe= c 20" straight pipe
Figure 4. Field equipment flow diagram
January 1, 2023 Con tech Engineered Solutions LLC 12

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Protocol
The field hydraulic evaluation was based on a prior protocol developed in consultation with Geosyntec
Consultants for the Filterra system. Protocol revisions were made to allow for automatic data collection
and data verification (Hills 2021).The general protocol included initiating hydraulic flow rate based on the
Filterra media surface area. Monitored media surface area, media depth and ponding depth allow for the
conversion between media flow rate (infiltration rate) and saturated hydraulic conductivity (Ksat) via
Darcy's Law.The infiltration rate is the rate at which water passes through the Filterra media,and may be
a function of media depth, hydraulic head, and moisture conditions among other factors (Geosyntec
2008).
Testing was performed in two phases:an unsaturated phase,followed by a saturation phase.The first test
allowed for observation of the system flow rate under (typical) dry conditions with less available water
content in the media,while the second test allowed for observation of the flow rate under(atypical,worst
case) saturated conditions. It is assumed the Filterra media is unsaturated at the beginning of the initial
test phase.The purpose of the unsaturated test was to bring the system to bypass and saturate the media,
followed by a draindown period to bring the media to field capacity. Flow was initially introduced near
design flow rate and increased incrementally by 20%until bypass was achieved. Flow rates were recorded
with each adjustment along with the accumulated water depth or water surface level (WSL) and elapsed
time per Thomas Scientific 1235C26 traceable stopwatches from the prior flow rate adjustment. Once
bypass was achieved, constant head conditions were maintained for five minutes to ensure saturation.
Bypass is defined as water exiting the system and traveling downstream of the system in an offline
application. Inflow was terminated after five minutes of bypass conditions and the rate of drawdown or
falling head was measured by recording the WSL with time over the media surface.
After a twenty-minute rest period passed since ceasing the initial unsaturated test, the saturation test
began with flow introduced near design flow rate and increased incrementally. Flow rates were again
recorded with each flow adjustment along with the WSL and elapsed time from the prior flow rate
adjustment until the system reached and maintained a steady WSL just below the bypass depth. Once an
approximate steady-state flow rate had been found,the flow rate and the WSL were noted, and the flow
rate was applied for five minutes. After these conditions were met, inflow was terminated, and the WSL
was noted. The rate of drawdown was then measured by recording the WSL with time over the media
surface (Geosyntec 2008; Hills 2009).
Biofiltration Media Composition Sampling and Analysis Techniques
To evaluate media characteristics, the mulch was scraped back in several locations within the system
following hydraulic testing. Media samples were collected at various depths throughout the media profile
and placed in properly labeled sample bags. Samples were analyzed for particle size analysis via following
ASTM F1632-03(ASTM 2018)and organic content via loss on ignition following ASTM F-1647(ASTM 2011),
respectively.
January 1, 2023 Con tech Engineered Solutions LLC 13

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
RESULTS
Summary statistics for sediment,nutrients,metals,oil&grease,temperature and pH are provided in Table
6 below.Water quality performance data analysis is presented by pollutant and represents combined data
sets across all three study sites over the three-to-thirteen-year monitoring duration. Reporting of
descriptive statistics generally follows the format used in the International Stormwater BMP Database
Summary Statistics report (2020). Data include sample counts, interquartile ranges of 25th and 75th
percentiles, and median influent, effluent, and removal efficiencies with 95% confidence intervals. Data
was analyzed in this manner to avoid outliers or concentrations below detection limits that may impact
results. Where non-detect concentrations were observed, concentration values equal to half the
detection limit were used for statistical analysis (Croghan and Egeghy 2003). Statistically significant
differences between influent and effluent median concentrations for each parameter were identified
using the non-parametric Mann-Whitney rank sum hypothesis test comparing the P-value to a significance
value of 0.05.
Influent and effluent concentrations were used to generate side-by-side box and whisker plots for each
pollutant. Data are presented on a log-scale where necessary to provide visual resolution on the y-axis.
The 25th and 75th percentiles for each pollutant data set are represented by the bottom and top of the
box, respectively, and listed in Table 6 below. The middle line of the box represents the 50th percentile
median for each pollutant.The notches at the median represent the 95%confidence internal around the
median and are also displayed in Table 6.The whiskers represent the furthest observation within 1.5 times
the interquartile range (IQR)from the quartiles. Near outliers displayed as orange plus symbols represent
observations further than 1.5 x IQR from the quartiles, and far outliers displayed as red asterisk symbols
represent observations further than 3.0 x IQR from the quartiles.
Regression analyses comparing effluent concentration versus time were conducted to understand
pollutant removal performance longevity. Confidence intervals were applied to the regression line to
show the 95% probability of the true regression line of the data set. R-squared values are also displayed
on each scatter plot figure as a measure of how well the regression model describes the data. Line plots
comparing influent versus effluent concentration by pollutant were also created to visually track the
impact of time on fluctuations in performance. Line plots were completed by study site for clarity and
comparison purposes.
January 1, 2023 Con tech Engineered Solutions LLC 14

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Table 6. Water quality performance summary for all Filterra study sites
IN vs.EFF
Parameter Sample count IQR 25th-75th percentiles Median (95%confidence interval) significant
diference P<0.05
Influent(IN) I Effluent(EFF) IN Conc.(mg/I) I EFF Conc.(mg/I) IN Conc. (mg/I) I EFF Conc.(mg/I) I Removal (%) (Y/N)
Sediment
TSS 88 88 13-57.25 1.6-7.45 31.2(21,37) 3.6(2.5, 5) 88.7(84, 90.5) Y
TSS-TAPEa 56 56 32.8-89.5 2.73-8.13 47.7(36,59.3) 5(3.7,6) 90.1 (88.9, 92.4) Y
Nutrients
TP 218 218 0.06-0.208 0.04-0.09 0.1 (0.09, 0.12) 0.05(0.05,0.06) 48.6(40,50) Y
TP-TAPEb 111 111 0.13-0.395 0.05-0.16 0.2(0.18, 0.26) 0.09(0.07,0.11) 60(41.7,63.3) Y
DP 190 190 0.03-0.12 0.02-0.07 0.06(0.05,0.07) 0.04(0.03,0.04) 35(30.6,45.5) Y
OP 172 172 0.03-0.103 0.01 -0.07 0.053(0.05,0.07) 0.03(0.03,0.04) 50(40, 50) Y
TN 89 89 0.7-3.2 0.5-2 1.4(1, 2.03) 1 (0.8, 1.25) 31.9(22.2,38.5) Y
NO2,3-N 57 57 0.1 -0.53 0.15-0.6 0.22(0.13,0.34) 0.32(0.2,0.44) -9.52(-78.6,0) N
TKN 74 74 0.655-3 0.405-1.45 1.4(0.84,2.01) 0.655(0.5, 1) 42(33.3, 55.6) Y
NH4 16 16 0.288-0.56 0.1 -0.26 0.34(0.29,0.52) 0.1 (0.1,0.26) 64.9(58.3,67.5) Y
Metals
Total Zn 58 58 0.083-0.305 0.033-0.08 0.14(0.109,0.188) 0.0455(0.038,0.07) 63(53.1,71.4) Y
Total Cu 56 56 0.011 -0.071 0.006-0.019 0.029(0.017,0.43) 0.01 (0.008,0.014) 57.3(40, 72.9) Y
Total Cd 13 13 0.001 -0.001 0-0.001 0.0084(0.001,0.001) 0.00059(0,0.001) 27.3(-172,68.8) N
Total Cr 14 14 0.002-0.006 0.001 -0.001 0.0042(0.002,0.007) 0.0008(0.001, 0.002) 76.5(52.2,87.5) Y
Total Pb 15 15 0.008-0.028 0.003-0.005 0.0147(0.008,0.028) 0.0025(0.003, 0.005) 68.8(0,83) Y
Total Ni 9 9 0.016-0.025 0.005-0.012 0.018(0.01,0.034) 0.005(0.005, 0.063) 64.3(-152,72.2) Y
Oil&Grease
O&G 25 25 5.6-11 2.5-5.7 7.2(6, 9.2) 2.5(2.5, 5.6) 58.3(48.2,66.1) Y
TPH 4 4 8.975-11 2.5-2.5 10.2(8, 11) 2.5(2.5,2.5) 75.2(68.8,77.3) Y
Other
pH 57 57 6.6-7.34 6.3-6.9 7(6.8,7.3) 6.6(6.5,6.7) N/A N/A
Temperature 35 35 14.3-25.9 14.3-25.4 16.5(14.5,25.3) 16.7(14.8,25) N/A N/A
a TAPE influent range of interest limited to >20 mg/L. Influent concentrations capped at 200 mg/L for removal efficiency calculation per TAPE guidelines.
b TAPE influent range of interest set at 0.1 to 0.5 mg/L. Influent concentrations capped at 0.5 mg/L for removal efficiency calculation per TAPE guidelines.
January 1, 2023 Con tech Engineered Solutions LLC 15

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Sediment
A statistical evaluation of 88 TSS sampling events demonstrated 88.7% median removal efficiency for
median influent and effluent concentrations of 31.2 mg/L and 3.6 mg/L, respectively. Influent
concentration range criteria set forth in the Technology Assessment Protocol - Ecology (TAPE) technical
guidance manual were applied to the TSS influent concentrations to exclude runoff events with very low
concentrations, resulting in 56 qualified events (Ecology 2018) that were above the lower 20 mg/L
threshold set by TAPE for TSS performance evaluation. For those events greater than 200 mg/L, the
influent concentration was capped at 200 mg/L in accordance with TAPE procedures. Applying TAPE
criteria to TSS influent concentrations increased median removal efficiency to 90% for median influent
and effluent concentrations of 47.7 mg/L and 5 mg/L, respectively. For both TSS and TSS-TAPE, median
influent and effluent concentrations were statistically different with very low effluent concentrations and
a P-value<0.001.
TSS box and regression plots (Figure 5) show the entire data distribution, excluding events with TSS
influent concentrations less than 20 mg/L per TAPE guidance. The box and regression plots show
significant TSS removal and no correlation between effluent concentration and time with the TAPE criteria
applied for the concentration range evaluated. Effluent concentrations fall consistently below the TAPE
treatment goal of 20 mg/L indicated by the red threshold line within the regression plot. This correlation
is further supported via line plot analysis in Figure 6 showing consistently low effluent concentrations over
time given variation in influent concentrations at all study sites.
1000 : 100 -
m J
4 .....
100 : 0
to - E7 47. 60 -
E ra
i — ----—_
0 1 + u 40 —— •———————
}2 ■ ■
f 10 : 1
L 0 •
it V
c ,' ( 20 - • - -
cu
H 1 : I W
a
I CC
I-
. -20 -col
H
H
0.1 -40
TSS_IN TAPE TSS_EFFTAPE 2006 2009 2012 2014 2017
Date(year)
Regression(R2=0.012)
Near outliers
x Far outliers 95%Confidence interval lines
———95%Prediction interval lines
(a) (b)
January 1, 2023 Contech Engineered Solutions LLC 16

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Figure 5. TSS descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
500 - 250
' —■—TSS_TAPE_IN —■—TSS TAPE
450 -
J J
E 400 - ---0---TSS-TAPE_EFF E 200 ■` ---0---TSS_TAPE_EFF
•` ■
c 350 - c
i 300 - ( E 150 ■
4, ++
a 0)i 250 -a
d
U U
C
200o100
a
< 150 \H / ■
50 ■H ■ �� ■ ■50 - i•. • /\iAc
•/ ■� .
o o 0 0 0 0 0 0 0 0 L�O°) ti�yO LO�ti LO�� ti§y� le LO\� LO�(O LO�1
N IN N IN IN N IN N NJ N
Date(year) Date(year)
(a) (b)
250
—■—TSS TAPE IN
J
to
200 •\ ---0---TSS_TAPE_EFF
0 \
i 150
c
a,
U
C
3 100
W ■
0.
er
\/
H ■ •
50 • • �•
•
'Le '6
y0 tisyy ti9y"' '5)
,N3 "16"‘ ti§yh le 1§C\
Date(year)
(c)
Figure 6.TSS line plot analysis comparing influent and effluent concentrations (a)Study site A (b)Study
site B(c)Study site C
January 1, 2023 Con tech Engineered Solutions LLC 17

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Nutrients
Box and regression plots for all nutrient forms monitored, with the exception of nitrite/nitrate nitrogen,
show statistically significant differences between influent and effluent median concentrations using the
non-parametric Mann-Whitney rank sum hypothesis test comparing the P-value to a significance value of
0.05.
Regression plots for all nutrient forms show no correlation between effluent concentration and time.
Higher nutrient effluent concentrations observed from 2008 through 2010 represent atypically high
influent concentrations at study site B. Line plot analysis for total phosphorus demonstrates low effluent
concentrations over time given variation in influent concentrations at all study sites. Other nutrient forms
including total dissolved phosphorus, orthophosphate, total nitrogen, total Kjeldahl nitrogen, and
ammonium, show greater correlation between influent and effluent concentrations in comparison to
other contaminants,with higher influent concentrations producing higher effluent concentrations.
Total Phosphorus
A statistical evaluation of 218 total phosphorus (TP) sampling events demonstrated 48.6% median
removal efficiency for median influent and effluent concentrations of 0.1 mg/L and 0.05 mg/L,
respectively. Influent concentration range criteria set forth in the Technology Assessment Protocol -
Ecology(TAPE)technical guidance manual were applied to the TP influent concentrations to exclude very
low concentration runoff events, resulting in 111 qualified events (Ecology September 2018) that were
above the 0.1 mg/L TAPE influent TP concentration threshold. For those events greater than 0.5 mg/L,the
influent concentration is capped at 0.5 mg/L according to TAPE procedures. Applying TAPE criteria to TP
influent concentrations increased median removal efficiency to 60% for median influent and effluent
concentrations of 0.2 mg/L and 0.09 mg/L, respectively. For both TP and TP-TAPE, median influent and
effluent concentrations were statistically different with very low effluent concentrations and a P-value <
0.001.
TP box and regression plots(Figure 7) show the entire data distribution excluding events with TP influent
concentrations less than 0.1 mg/L per TAPE guidance. The box and regression plots show significant TP
removal and no correlation between effluent concentration and time with the TAPE criteria applied for
the concentration range evaluated. This correlation is further supported via line plot analysis in Figure 8
showing consistently low effluent concentrations over time given variation in influent concentrations at
all study sites.
January 1, 2023 Contech Engineered Solutions LLC 18

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
100 : 4.50 ,
r
J
E 3.50 -
-
- 10 : X 0
E L 2.50 - ■
O ■
O . U
v _
v 1 #+ p 1.50 ---- _
u ■ -----
U
0 _ l�,J I 3 0.50 - -- —
0
f l W - ■-
I IJ w __
H 0.1 0. ■ram - -
, l a -0.50 - g.
--- I -----_
0.01 I -1.50
TP IN TAPE TP EFF TAPE 2006 2009 2012 2014 2017
Date(year)
+Near outliers Regression(R==0.074)
X Far outliers 95%Confidence interval lines
— — -95%Prediction interval lines
(a) (b)
Figure 7.TP descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
1.2 14 —■—TPTAPE IN
—■—TPTAPE IN _
---0---TP TAPE EFF --
12 - ■ -0---TP TAPE_EFF
1to
E ■ E
10 -
0 0.8 r
11)
I- i
8 -
++ C
g 0.6 U
v
• ■ ■ 0 6 -
o r■
■ c°
v 0.4 ■ ! a ■
a ■ Q 4 -
H ■ /I ~II ■ ■-■ ■ ■ d 10.2 9i�� � A ■ ■ ~ 27\ ■\
4 S' .,n ■ ..
0 I �°-� •
o gS/S-0- n� `'9 , o
r
m o N N cn a L.r) l0 N CO 2008 2009 2010
o o 0 0 0 0 0 0 0 0
N N N N N N N N N N Date(year)
Date(year)
(a) (b)
January 1, 2023 Con tech Engineered Solutions LLC 19

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
1.2 —■—TP TAPE IN
"Zr ---D-- TP TAPE_EFF
1 ■
0.8 ■
m ■
u 0.6
o • •
-• •
a 0.4 ■ •
• I7 ■ l
•
•
I- 0.2 Q;IF"'" i •
i
o
Ol O N m Ln lD I� CO
O O O O O O O O O O
N N N N N N N N N N
Date(year)
(c)
Figure 8.TP line plot analysis comparing influent and effluent concentrations (a)Study site A(b)Study
site B(c)Study site C
Total Dissolved Phosphorus
A statistical evaluation of 190 total dissolved phosphorus (TDP) sampling events demonstrated 35%
median removal efficiency for median influent and effluent concentrations of 0.06 mg/L and 0.04 mg/L,
respectively. Median TDP influent and effluent concentrations were statistically different with very low
effluent concentrations and a P-value < 0.001. TDP box and regression plots (Figure 9) show significant
TDP removal and no correlation between effluent concentration and time. Line plot analyses in Figure 10
show correlation between influent and effluent concentrations, with higher influent concentrations
producing higher effluent concentrations.
January 1, 2023 Con tech Engineered Solutions LLC 20

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
10 = X- X 3.5 -
X •�„ Joao 3 -
a 1 = E 2.5 -
Ec
X X o
c - 2 -
o c
L
it
ar_+ �.�� U 1.5 -
c �'-X c
a - j + 0 1
c I u - — _
c0�0.1 - +-. I ---------
a - n I 3 0.5 - ■ a
rziII L� I a -0 - �� - - ---------
f 1 - ---
I I -0.5 -------------
0.01 I I 1 -1 .
DP IN DP_EFF 2006 2009 2012 2014 2017
Date(year)
Regression(R2=0.046)
Near outliers
:I(Far outliers 95%Confidence interval lines
— — -95%PredIction interval lines
(a) (b)
Figure 9.TDP descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
—■—TDP IN 10 —■—TDP_IN
0.3 - ---0---TDP --_EFF 9 - -0---TDP_EFF
I
8
ao
0.25 - ■ E 7 -
E1 c
c 0.2 - 6
■ ` 5 - ■
2 I c
� 0.15 - �■ ■ U V
U • c 4 O
■ u P
u0.1 � ; •■ ■ a 3 -
0. ■ ■i • ■ G
I- 0.05 .Pi i �I ti —r\; ■ 1 ;8,I o f" MI - --ts {-9-- I\
i 11 / '4�a.. -c
0
0 0 0� a-Ia-Ia-Ic-I COc-Ic-Ic-I 2008 2009 2010
0 0 0 0 0 0 0 0 0 0 0
N N N N N N N N N N N
Date(year) Date(year)
(a) (b)
January 1, 2023 Contech Engineered Solutions LLC 21

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
0.45 —■—TDP IN
■ TDP_
IN
■ • TDP_EFF
0.35 I ■ ■
■
134 J I
0.3
•
• 0.25 ■
E▪ 0.2
■ i
■
chi 0.15
1•I •
O '
Pei
0 r r
01 O —1 N M if1 lO n c0
O O O O O O O O O O
N N N N N N N N N N
Date(year)
(c)
Figure 10.TDP line plot analysis comparing influent and effluent concentrations (a) Study site A(b)
Study site B (c) Study site C
Orthophosphate
A statistical evaluation of 172 orthophosphate (OP) sampling events demonstrated 50% median removal
efficiency for median influent and effluent concentrations of 0.053 mg/L and 0.030 mg/L, respectively.
Median OP influent and effluent concentrations were statistically different with very low effluent
concentrations and a P-value<0.001.
OP box and regression plots (Figure 11)show significant OP removal and no correlation between effluent
concentration and time. Line plot analyses in Figure 12 show correlation between influent and effluent
concentrations, with higher influent concentrations producing higher effluent concentrations.
January 1, 2023 Con tech Engineered Solutions LLC 22

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
10 = -
1.6 -
J
J X E 1.2 -
C
aa 1 = X o
E -
as
:� _ a,
co v
+`+ II+ 4 - ■
Co — __
ai ..........
____
• 0.1 _ 0.4 = - _16___--
N.
Li
a pi�I �YT ■ _� -�
E7 o
7
-0.4 — —
0.01 I
OP_IN OP_EFF 2006 2009 2012 2014 2017
Date(year)
+Near outliers Regression(R2=0.088)
XFar outliers 95%Confidence interval lines
— — -95%Prediction interval lines
(a) (b)
Figure 11. OP descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
U.L5 - 1.8 -
-■—OP_IN P —■—OP_IN
---0---OP_EFF 1.6 - ---0---OP_EFF
0.2 - 1.4 -
_ ■
Iwo ■ E 1.2 - 1 Q
E I I
0.15 - ; ;■ ■■ I
o o 1 _ I
✓ � ' 4
,_ ++ 0.8 - = Q ,
a • 1
u • ■ ', ■l
o
o
I • si ■ u ;
a 0.05 - ■ ■m,,�■�i I• ■ J
a0 0.4 • ��`/'■; ; ■
ri%rt It
.4,0
• i'dim mod b---a-Ci �'10 �■ O
O 0 0 ci ri N. °° 2008 2009 2010
o o 0 0 0 0 0 0 0 0 0
N N N N N N N N N N N
Date(year) Date(year)
(a) (b)
January 1, 2023 Contech Engineered Solutions LLC 23

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
0.7
—■—OP_IN
0.6 i -- - OP_EFF
■
•
II
0.5
! 0.4
o ■
■ 1
•
E 0.3 ■
L
it I
■!■ ■
n
, ; _ •
g Cr
0 Q -ciN§43.
Ql O N M V1 l0 I, 00
o
O O O O O O O O O O
N N N N N N N N N N
Date
(c)
Figure 12. OP line plot analysis comparing influent and effluent concentrations (a) Study site A(b) Study
site B(c)Study site C
Total Nitrogen
A statistical evaluation of 89 total nitrogen (TN) sampling events demonstrated 31.9% median removal
efficiency for median influent and effluent concentrations of 1.4 mg/L and 1.0 mg/L, respectively. Median
TN influent and effluent concentrations were statistically different with very low effluent concentrations
and a P-value<0.001.
TN box and regression plots (Figure 13) show significant TN removal and no correlation between effluent
concentration and time. Line plot analyses in Figure 14 show correlation between influent and effluent
concentrations, with higher influent concentrations producing higher effluent concentrations.
January 1, 2023 Con tech Engineered Solutions LLC 24

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
100 7 -
- :1- 8
to
10 - + 1=
on I
f- a 6
E - o a a ■•
o E c [ - •
■ ■
•
E - -
�� 8 2 -• ■
V I ■�
o I I :-r�:- --r■■-:�-■- ars__ _;■
Z I _3 0 ■ ■ ■
1- 0.1 : I E
W
- I 1.- -2 - -----
0.01 — . --, -4
TN IN TN_EFF 2006 2009 2012 2014 2017
Date(year)
Near outliers Regression(R2=0.040)
Far outliers -95%Confidence interval lines
— — -95%Prediction interval lines
(a) (b)
Figure 13.TN descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
18 —■—TN_IN —■—TN_IN
16 - ---0---TN_EFF •
• ---a---TN_EFF
• 10 -
14 -
J
to to E 12 - E 8 4
a• 10 - c • ■
■ = 6 - • ; ■
8 - 4... `,i
c c
ai ar �.
c 6 - c 4 -
O 0 • i\
u -b
z 4 - 4 z �� ,
1— :,, ///' . .• •
2 - C „/ 1.,,i\ \ A ..., :
2009 2010 2011 2012 2013 2014 2009 2010 2011 2012
Date(year) Date(year)
(a) (b)
January 1, 2023 Contech Engineered Solutions LLC 25

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
12
—•—TN_IN
• ---0---TN EFF
10 — 4
J 0
ba E 8 • •
1
1
o ,
o ,
.41 6 ,
� 1
L � 1
' 1
u 4
11 ■\ • ■ 1
O 1 ` I •
V i �� • 1
. , 1\ •\ :I2 1 ■ i
I bT(o-- _ - - - � - _-_�_
oT 8 1
2009 2010 2011 2012
Date(year)
(c)
Figure 14.TN line plot analysis comparing influent and effluent concentrations (a) Study site A(b) Study
site B(c)Study site C
Nitrite/Nitrate-Nitrogen
A statistical evaluation of 57 Nitrite/Nitrate-Nitrogen (NO2, NO3-N) sampling events demonstrated -
9.52% median removal efficiency for median influent and effluent concentrations of 0.22 mg/L and 0.33
mg/L, respectively. Median NO2, NO3-N influent and effluent concentrations were not statistically
different with a P-value of 0.202. NO2, NO3-N is highly soluble in soils, and export is common in
bioretention media due to nitrification of ammonium to nitrite and nitrate.
NO2, NO3-N box and regression plots (Figure 15) show no significant NO2, NO3-N removal and no
correlation between effluent concentration and time. Line plot analyses in Figure 16 show correlation
between influent and effluent concentrations, with higher influent concentrations producing higher
effluent concentrations.
January 1, 2023 Con tech Engineered Solutions LLC 26

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
10 = 1.75
Joao 1.5
J - W E
to T -----
+ 0
+ —
o 1.25 -------
O• 1 = I + I 1 1.
:r +' ■
-
L
c - uc 0.75 ■ ■ ■
�� -
u - ; �� V --------■ •' -------_.
C
O \may ;' 0.5
c� • ■ ow
■
0.1 = al
------ -
• -' -
----------- ----��_ ■
■■ ■ __ -,�6 NJ ■ ■ ■ ■ ■ ' II `-`^_
•
O - O
Z - Z
N-0.25
- 0
Z ———————
0.01 1 1 -0.5 ——
NO2 NO3 IN NO2 NO3 EFF 2008 2010 2013 2016 2019
Date(year)
Near outliers iRegression(R2=0.018)
x Far outliers 95%Confidence interval lines
— —-95%Prediction interval lines
(a) (b)
Figure 15. NO2, NO3-N descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
1. - _L.„
■ ■ —■—NO2_NO3-IN —■—NO2_NO3_IN
3 1 ---0---NO2_NO3_EFF \ 1.4 ---0---NO2_NO3_EFF
4
E S E 1.2 I
c 0.8 - I� n c i
O i i •' ii Q a+ 1 h
raIII
1 IIIlI 1
y+ 1 \ 11 11 11
0.6 P 1 U 0.8
1 1i / %\ 1 N C 11
(o 1 \ i i� \ /11
U0.4 1 `�Z \ 1 I \ 1 Z
• 1en \\ ,, \\ I '• I p 0.4
Z 0.2 •CI ■ 0 /
o 0 \6 \ / \ � ■■_ 0 0.2 •1 ■ ■ ', ■ Z \\\ ' ■Z 0 I I afl , i■ 0 I■_��■ 0 l CI
y000 ti0,O tiON' ti0ti� tiON' tiON' tiON' tiON' 'VON
2008 2009 2010
Date(year) Date(year)
(a) (b)
January 1, 2023 Contech Engineered Solutions LLC 27

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
3
—■—NO2 NO3
■ - -
IN
2 5 —CI—NO2_NO3_EFF
E
c
0 2
+D
s-
L
5 1.5 �� Q
•
z '•�■;
0 , 1 4
z l ; ;Q o
0.5 '
Date(year)
(c)
Figure 16. NO2, NO3-N line plot analysis comparing influent and effluent concentrations (a)Study site A
(b) Study site B (c) Study site C
Total Kjeldahl Nitrogen
A statistical evaluation of 74 total Kjeldahl nitrogen (TKN) sampling events demonstrated 42% median
removal efficiency for median influent and effluent concentrations of 1.4 mg/L and 0.655 mg/L,
respectively. Median TKN influent and effluent concentrations were statistically different with very low
effluent concentrations and a P-value< 0.001.
TKN box and regression plots (Figure 17) show significant TKN removal and no correlation between
effluent concentration and time. Line plot analyses in Figure 18 show correlation between influent and
effluent concentrations, with higher influent concentrations producing higher effluent concentrations.
January 1, 2023 Con tech Engineered Solutions LLC 28

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
100 z 10 -
J 8 - ■
to
10 = ki, E
•
E - 4+ c 6 ---_ __
o _
L
4 �
^ Ii aa) _ ■
1 - c '�■_ ■
c = �� � o r- ■
a - u 2 """-----f------
---------------
o I I c ■� ■ i--r-------■ ----
I
o _
0.1 = I •
w -
Y -2 - ---__
0.01 I -4 —. .
TKN IN TKN_EFF 2008 2010 20.13 2016 2019
Date(year)
Near outliers Regression(R2=0.099)
x Far outliers 95%Confidenc e interval lines
——-95%Prediction interval lines
(a) (b)
Figure 17.TKN descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
1,, 112
• •■—TKN_IN ■ —■—TKN_IN
14 1 ---0---TKN_EFF
10 ---0---TKN_EFF
J
\ 12 0a
to
E - 8
10 c
c o ■ ■
a+c y \■ll `�aJU
U 6 - c , P `�c 3 4 it
; : •,
V f ■Z 4 Y % N0 ■ �`.H ■ ■ ~ 2 ■ , u �� ``
2 q ■ / ■ a
\ r a '��D 8
•
�/00ki
2009 2010 2011 2012 2013 2014 2015 2016 2017 2008 2009 2010
Date(year) Date(year)
(a) (b)
January 1, 2023 Contech Engineered Solutions LLC 29

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
10 -
9 - —■—TKN_IN i
g _ ■ ---0-- TKNEFF
J
dp 7 - ■
E
6 -
0
fa 5 -
L
■
c 4 ■
V ■
3 Q1 ■
u 1 Q
Z Y 2 i i\ i ■
a6'8
0 fl• d"'�8
2009 2010 2011 2012 2013 2014 2015 2016 2017
Date(year)
(c)
Figure 18.TKN line plot analysis comparing influent and effluent concentrations (a)Study site A (b)
Study site B (c) Study site C
Ammonium
A statistical evaluation of 16 ammonium (NH4) sampling events demonstrated 64.9% median removal
efficiency for median influent and effluent concentrations of 0.34 mg/L and 0.1 mg/L, respectively.
Median NH4 influent and effluent concentrations were statistically different with very low effluent
concentrations and a P-value<0.001. NH4 was not monitored at study site B.
NH4 box and regression plots (Figure 19) show significant NH4 removal and no correlation between
effluent concentration and time. Line plot analyses in Figure 20 show correlation between influent and
effluent concentrations, with higher influent concentrations producing higher effluent concentrations.
January 1, 2023 Con tech Engineered Solutions LLC 30

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
10 - 1 -
_ �
�\
- 0.75 -
oa
J
X E_
E 0 0.5 - -
fa
o 4-. _
c ' --t
ra 1 : X c0.25 - • `■ ■ ■
c _
ai I u° -------------------
C - �+
co., c N 0 -
a ' I
I ' ` _
z I "'_ ' ^ R
I \
_ -0.25 -
z
' I �_�
0.1 -0.5
NH4_IN NH4_EFF 2013 2014 2016 2017
Date(year)
+Near outliers —Regression(R2=0.249)
x Far outliers •95%Confidence interval lines
— —-95%Prediction interval Ines
(a) (b)
Figure 19. NH4 descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
0.8 3 5—■—NH4_IN —■—NH4_IN
•
0.7 - • -0-- NH4_EFF 3 ---0---NH4_EFF
3 0.6
� JoIDA ao 2.5
E 0.5 - E
c o 2
o
7. 0.4 ■ +,
■ 1-
/ c 1.5
aci 0.3 • ,., •
C 0•,....7m
c
o ■ ` 0 1
u 0.2 / 1 u
Z ■
. , _
2
Z 0.1 -d 0-D- -0 Z 0.5 -'Q. / ■` ■
■p-q ■—■�ci
2014 2015 2016 2017 2014 2015 2016 2017 2018
Date(year) Date(year)
(a) (b)
Figure 20. NH4 line plot analysis comparing influent and effluent concentrations (a) Study site A(b)
Study site C
January 1, 2023 Con tech Engineered Solutions LLC 31

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Metals
Box and regression plots for all metals monitored, with the exception of cadmium which was likely
influenced by influent concentrations near the detection limit, show statistically significant differences
between influent and effluent median concentrations using the non-parametric Mann-Whitney rank sum
hypothesis test comparing the P-value to a significance value of 0.05.
Regression plots for all metals show no correlation between effluent concentration and time. Line plot
analysis for all metals generally demonstrate low effluent concentrations over time given variation in
influent concentrations at all study sites.
Total Zinc
A statistical evaluation of 58 total zinc (Tot. Zn) sampling events demonstrated 63% median removal
efficiency for median influent and effluent concentrations of 0.14 mg/L and 0.0455 mg/L, respectively.
Median Tot. Zn influent and effluent concentrations were statistically different with very low effluent
concentrations and a P-value<0.001.
Tot. Zn box and regression plots (Figure 21) show significant Tot. Zn removal and no correlation between
effluent concentration and time.This correlation is further supported via line plot analysis in Figure 22 for
study sites where Tot. Zn was monitored, showing consistently low effluent concentrations over time
given variation in influent concentrations.
10 = 3 -
2.5 -
- ao
'aE
1 = X c 2 -
E 2
- I 4-
o• - ° 1.5 -
c
c 0.1 = I 1 -
d _ f 1 1 �� v° -
c - I Y aci 0.5 -
N I W 0 - Sae ■— ■
-0.5 -
0.001 -1
Tot.Zn IN Tot.Zn EFF 2006 2009 2012 2014 2017
Date(year)
+Near outliers Regression(R2=0.004)
x Far outliers 95%Confidence interval lines
——-95%Prediction interval lines
(a) (b)
January 1, 2023 Con tech Engineered Solutions LLC 32

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Figure 21.Tot. Zn descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
0.8 - 3 -
• —■—Tot.Zn_IN —■—Tot.Zn_IN
0.7 - —0—Tot.Zn EFF 4 ---0---Tot.Zn EFF
1 2.5 - # —
•E 0.6 - • E
2 -
c 0.5 r
o
i- 0.4 • v 1.5 -
aci • c
o 0.3 - i ■ / u 1 - ■
■ ■\■ N
N 0.2 � I ++ N.
, •!._-: I ■ o 0.5 ■ •H 0.1 , �■ ,.8 Iil /\ • IN li _ p
0 -'0- 0 w+-�"��' "
2008 2009 2010 0 0 0 0 ci ci ci ci
0 0 0 0 0 0 0 0 0 0 0
N N N N N N N N N N N
Date(year) Date(year)
(a) (b)
Figure 22.Tot. Zn line plot analysis comparing influent and effluent concentrations (a) Study site B (b)
Study site C
Total Copper
A statistical evaluation of 56 total copper(Tot. Cu)sampling events demonstrated 57.3% median removal
efficiency for median influent and effluent concentrations of 0.029 mg/L and 0.01 mg/L, respectively.
Median Tot. Cu influent and effluent concentrations were statistically different with very low effluent
concentrations and a P-value<0.001.
Tot. Cu box and regression plots (Figure 23)show significant Tot. Cu removal and no correlation between
effluent concentration and time.This correlation is further supported via line plot analysis in Figure 24 for
study sites where Tot. Cu was monitored, showing consistently low effluent concentrations over time
given variation in influent concentrations.
January 1, 2023 Con tech Engineered Solutions LLC 33

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
10 - -
J 0.175 -
to
1 - E
to -E
_ c 0.125 -
.a
0.1 - I X x)K •a�i _ ■
I- v 0.075 ■
�. _ �`� +
aJ - ,
I _
++ -■--
0 0.01 = �--� f�1 •0.025 - ■ - ■ Y • i
3 . I I --/Wr ■-aJL ■
U _ I W _ um% ■ ■ �- �i__
4 3 - _
I- 0.001 - I v -0.025 -
- I 1- _ ------_
0.0001 -0.075
Tot. Cu IN Tot.Cu EFF f 2006 2009 2012 2014 2017
Date(year)
1-Near outliers Regression(RZ=0.048)
xFar outliers 95%Confidence interval lines
——-95%Prediction interval lines
(a) (b)
Figure 23.Tot. Cu descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
0.3 - 0.3 1
—•—Tot.Cu_IN ■ —0—Tot.Cu_IN
---0---Tot.Cu EFF ---0---Tot.Cu_EFF
—0.25 - ■ - 0.25
\ J
to
E
C 0.2 - - 0.2 -
0 C
'4, 0
II:
p a+
� 0.15 ■ A i 0.15 -
1■— ■ I c
chi ■ I ■ i ' a.
Q U
p • , ` C
U 0.1 - i o 0.1 ■
3
0 • U ¢■
�' I •I �+ ■
H 0.05 -I I /I/■ 1! p 0.05 ■ ■ ■
■ ■■ •\d r 6 d-22.---a.-------a ,§ 1—__0 I\/\ ■
0 0 �mil IN'
2008 2008 2008 2009 2009 2009 2010 2010 2007 2009 2011 2013 2015 2017
Date(year) Date(year)
(a) (b)
Figure 24.Tot. Cu line plot analysis comparing influent and effluent concentrations (a) Study site B (b)
Study site C
January 1, 2023 Con tech Engineered Solutions LLC 34

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Total Cadmium
A statistical evaluation of 13 total cadmium (Tot. Cd) sampling events demonstrated 27.3% median
removal efficiency for median influent and effluent concentrations of 0.0.0084 mg/L and 0.00059 mg/L,
respectively. Median Tot. Cd influent and effluent concentrations were not statistically different with a P-
value of 0.110. The lack of statistical difference between median influent and effluent concentrations is
likely influenced by influent concentrations near the detection limit of 0.005 mg/L.
Tot. Cd box and regression plots (Figure 25) show no significant Tot. Cd removal and no correlation
between effluent concentration and time. Low influent and effluent concentrations for most storm events
make any correlation between influent and effluent concentration difficult to evaluate in the line plot
analyses in Figure 26 for study site B where Tot. Cd was monitored.
10 = 0.25 -
0.2 -
1 b\.o
... E tu) 0.15 -
E = c
o
0.1 - a 0.1 -
ra
a X 20.05 -
a o- X
u V
0 0.01 = a 0 - ■■ ■ ■ ■
-a - 3 —
+� - w-0.05 - 'o
0.001 - r - -- I 13 ,
-0.1 - __---------
0.15 -
0.0001 '
Tot.Cd IN Tot.Cd EFF 2008 2008 2009 2010 2010
Date (year)
Near outliers Regression(R2=0.114)
} Far outliers 95%Confidence interval lines
——-95%Prediction interval lines
(a) (b)
Figure 25.Tot. Cd descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
January 1, 2023 Con tech Engineered Solutions LLC 35

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
0.25 -
—■—Tot.Cd_IN
---0---Tot.Cd_EFF
0.2 - q
J
to
E ,
,
0.15 - i
0
, D
y0.1 - 1 \\s
a., I \
C 1 `
•
•u \\
0.05 - , \\
yi
0 0-0-00—oTd —b
2008 2009 2010
Date(year)
Figure 26.Tot. Cd line plot analysis comparing influent and effluent concentrations at Study site B
Total Chromium
A statistical evaluation of 14 total chromium (Tot. Cr) sampling events demonstrated 76.5% median
removal efficiency for median influent and effluent concentrations of 0.0042 mg/L and 0.0008 mg/L,
respectively. Median Tot. Cr influent and effluent concentrations were statistically different with very low
effluent concentrations and a P-value< 0.001.
Tot. Cr box and regression plots (Figure 27) show significant Tot. Cr removal and no correlation between
effluent concentration and time.This correlation is further supported via line plot analysis in Figure 28 for
study site B where Tot. Cr was monitored, showing consistently low effluent concentrations over time
given variation in influent concentrations.
January 1, 2023 Con tech Engineered Solutions LLC 36

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
10 = 0.004 - _
_
1 E 0.003 - •
E - ■ ---------"---'---
7. 0.002 `
0.1 = c ■
f0 G1
_ 0 0.001 - ■ ■
cu u U
• 0.01 - c ■ ---■____________ ■ ■
O - '
U I 41 0
1� 0 -
u - I I + i
L -1 LU
I
1-• 0.001 - I 6
+J-0.001 - ----
J � __
___
0.0001 I -0.002 . I 1
Tot.Cr IN Tot.Cr_EFF 2008 2008 2009 2010 2010
Date(year)
Near outliers Regression(R2=0.032)
xFar outliers •95%Confidence interval lines
——-95%Prediction interval Ines
(a) (b)
Figure 27.Tot. Cr descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
0.045 -
—■—Tot.Cr_IN ■
0.04 -
---0---Tot.Cr_EFF
J
0.035 -
on
c 0.03 -
0
0 0.025 -
L
M
C
Q.) 0.02 -
C
o
u 0.015 •
L
V
a+ 0.01
0
0.005 ■ ■0 ■
pp-.I ■■-
2008 2009 2010
Date(year)
Figure 28.Tot. Cr line plot analysis comparing influent and effluent concentrations at Study site B
January 1, 2023 Con tech Engineered Solutions LLC 37

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Total Lead
A statistical evaluation of 15 total lead (Tot. Pb) sampling events demonstrated 68.8% median removal
efficiency for median influent and effluent concentrations of 0.0147 mg/L and 0.0025 mg/L, respectively.
Median Tot. Pb influent and effluent concentrations were statistically different with very low effluent
concentrations and a P-value of 0.022.
Tot. Pb box and regression plots (Figure 29) show significant Tot. Pb removal and no correlation between
effluent concentration and time. This correlation is further supported via line plot analysis in Figure 30
where Tot. Pb was monitored, generally showing low effluent concentrations over time given variation in
influent concentrations. Export instances occurred in 3 of the 15 sampling events, which may be due to
heavier lead loading at the gas and oil service station study sites in a prior storm event resulting in residual
concentration influencing these sampling events.
10 = 0.6 -
- J 0.5
- --------------
•
E 0.4 -
-Zr
p- 1 : C
E X ° .3
c - X 2
o - c 0.2 - -------------
+r - a,
12 V
c 0.1 : 0 0.1 -
V X
c - c 0 -•• • ■ • •
0 Icu
-
V - r 3 _______________
a '� : W-0.1
c 0.01
i a-0.2 - -
-
I _ ~-0.3 -__________
0.001 , -0.4
Tot. Pb IN Tot. Pb EFF 2007 2008 2010 2011 2012 2014
Date(year)
Near outliers Regression(R==0.007)
x Far outliers
95%Confidence interval lines
— — -95%Prediction interval lines
(a) (b)
Figure 29.Tot. Pb descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
January 1, 2023 Con tech Engineered Solutions LLC 38

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
0.6 - 0.07 -
-■—Tot.Pb_IN —■—Tot.Pb_IN
P
OS - --o -Tot.Pb_EFF 0.06 - Di ---0---Tot.Pb_EFF
ao ; CIA I
E E 0.05 - '
0.4 -
c
+r
�` 4' 0.0 4 - '
�
2 : �`` L
ra
c •
;
0.3 - '
cci % `. c 0.03 J "ir1
c '
u 0.2 - u ■ \\ � ' \
o
H 0.1 -
• `.. I- 0.01 ■ i•\
2008 2009 2010 2007 2008 2009 2010 2011 2012 2013
Date Date
(a) (b)
Figure 30.Tot. Pb line plot analysis comparing influent and effluent concentrations (a) Study site B (b)
Study site C
Total Nickel
A statistical evaluation of 9 total nickel (Tot. Ni) sampling events demonstrated 64.3% median removal
efficiency for median influent and effluent concentrations of 0.018 mg/L and 0.005 mg/L, respectively.
Median Tot. Ni influent and effluent concentrations were statistically different with very low effluent
concentrations and a P-value of 0.048.
Tot. Ni box and regression plots (Figure 31) show significant Tot. Ni removal and no correlation between
effluent concentration and time.This correlation is further supported via line plot analysis in Figure 32 for
study site B where Tot. Ni was monitored, generally showing low effluent concentrations over time given
variation in influent concentrations.
January 1, 2023 Con tech Engineered Solutions LLC 39

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
10 - .
- —
0.12 - ---
1 - E
E 0.0 8 -c
o
oc�a ---------------"
s 80.04 -
48 0.1 = u
a, c -
c = I v° ■ L ■
o - ---, --�--- c 0.00 -
2 - - 3 ---------------------- `
H0.01 - w�� z-0.04 _ ------•
Y
o - ——
- I- ——
—
0.001 1 I -0.08 . ,
Tot. Ni_IN Tot. Ni_EFF 2008 2008 2009 2010 2010
Date(year)
+Near outliers Regression(R2=0.204)
x Far outliers 95%Confidence interval lines
——-95%Prediction interval Ines
(a) (b)
Figure 31.Tot. Ni descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
0.1 -
—■—Tot.Ni_IN
0.09 - ---D---Tot.Ni_EFF Q
0.08 - ,
,
0.07 - 1
-
, et
c 0.06 -
o `.
0 0.05 - I
4- I \
0.04 - i \\ ■
c i •
o0.03 -
Z /.1
.,c,; 0.02 ..! \\
m—m— �\
0.01 1(9
El--a � \b
0
2008 2009 2010
Date(year)
Figure 32.Tot. Ni line plot analysis comparing influent and effluent concentrations at Study site B
January 1, 2023 Con tech Engineered Solutions LLC 40

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Oil& Grease
A statistical evaluation of 25 oil and grease (O&G)sampling events demonstrated 58.3% median removal
efficiency for median influent and effluent concentrations of 7.2 mg/L and 2.5 mg/L, respectively. Median
O&G influent and effluent concentrations were statistically different with very low effluent concentrations
and a P-value<0.001.
O&G box and regression plots (Figure 33) show significant O&G removal and no correlation between
effluent concentration and time. Line plot analyses in Figure 34 show no correlation between influent and
effluent concentration over time.
1000 - 14
- X - 12
- 10 —_______——
ao C
E 100 - o
o - 4- O
c - a�i ■
c
6 --- -- ---
---------
■
V
o c 4
• 10 = lam, v
• - ^�` rJ ' W 2 __-- _ _
O _ I ,
i
06
I 21_1_
0 0
1 -2
O&G IN O&G EFF 2008 2010 2013 2016
Date(year)
Near outliers Regression(R2=0.026)
x Far outliers 95%Confidence interval lines
——-95%Prediction interval lines
(a) (b)
Figure 33. O&G descriptive statistical analysis (a) box and whisker plots of influent and effluent
concentrations (b) regression scatter plot of effluent concentrations and time
January 1, 2023 Con tech Engineered Solutions LLC 41

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
400 - 14 -
—■—O&G_IN
■ —■—O&G_IN 4
350 - 12 - 7, ---0---0&G_EFF
---0---O&G_EFF
300 - , '
o�a10 -E
; 1
E '
■
m 200 - ' / �,
c 150 -
o U 4 i '
CCD 100 0 ! l i \i
ea 0 0 ■ a■—■ ti ■
0 so - 2
2008 2009 2010 00� 00o ti ' 'O,O 6,7 le Otis
Date(year) ' 'O�h l
Date(year)
(a) (b)
Figure 34. O&G line plot analysis comparing influent and effluent concentrations (a) Study site B (b)
Study site C
Total Petroleum Hydrocarbons
A statistical evaluation of 4 total petroleum hydrocarbon (TPH) sampling events demonstrated 75.2%
median removal efficiency for median influent and effluent concentrations of 10.2 mg/L and 2.5 mg/L,
respectively. Median TPH influent and effluent concentrations were statistically different with very low
effluent concentrations and a P-value of 0.029.
The TPH box plot (Figure 35) shows significant TPH removal. Sample size was not adequate for proper
regression analysis to demonstrate whether correlation exists between effluent concentration and time.
Line plot analysis in Figure 36 for study site B where TPH was monitored shows low effluent concentrations
over time given variation in influent concentrations
January 1, 2023 Con tech Engineered Solutions LLC 42

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
100 -
J
tu)
E
C
0
.470
10 - I
. I
a,
u
C .
0
u
2 .
a
1—
1 _,
TPH_IN TPH_EFF
Figure 35.TPH descriptive statistical analysis box and whisker plots of influent and effluent
concentrations
12 1 —■—TPH_IN
■ ---0---TPH_EFF •
10
J ■
dp
E 8 ■
C
0
f° 6
L
it
a,
u
C
0 4
u
I
H
G -0
2
0
2008 2009
Date(year)
Figure 36.TPH line plot analysis comparing influent and effluent concentrations at Study site B
Temperature and pH
An evaluation of 35 temperature sampling events demonstrated no statistical difference between median
influent and effluent temperatures of 16.5 mg/L and 16.7 degrees C, respectively.An evaluation of 57 pH
sampling events demonstrated no statistical difference between median influent and effluent pH of 7 and
6.6, respectively.
January 1, 2023 Con tech Engineered Solutions LLC 43

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Hydraulic Capacity
Separate analyses were completed in 2021 at study site A, after the water quality sampling phase, to
evaluate Filterra hydraulic conductivity in unsaturated and saturation conditions. Following procedures
performed in the Filterra Field Flow Rate Evaluation Report (Geosyntec 2008), hydraulic evaluation was
conducted in three distinct phases described below with results displayed in Table 7. The Filterra media
hydraulic evaluation demonstrated a typical condition median infiltration rate of 177 in/hr, in line with
Filterra's Washington Department of Ecology General Use Level Designation approved infiltration rate of
175 in/hr.The Filterra media saturated, atypical operating conditions displayed a median infiltration rate
of 107 in/hr.
Steady Flow Rate, Rising Head —The initial unsaturated and saturation tests included a period
when steady inflow was maintained above the hydraulic capacity of the Filterra system and water
accumulated in the system causing the water surface level (WSL)to rise.The duration of the rising
water level, inflow rate, change in accumulated head and average WSL during the period or rising
head was used to calculate an infiltration rate.
Modulated Flow Rate, Constant Head— During the saturation test, an inflow rate that allowed
constant accumulated head for five minutes was determined. The constant head data was
calculated similarly to the Steady Flow Rate, Rising Head data, however changes to WSL was set
to zero.
No Inflow, Falling Head — The ponded water receded after inflow ceased in both the initial
unsaturated and saturation tests and rate of fall was recorded. Analysis was calculated similar to
those calculations in the Steady Flow Rate, Rising Head test described above.
Table 7. Study site A hydraulic evaluation summary
Modulated
Test ID Steady Flow, Flow, Falling Head Method Method
Rising Head Constant Average Median
Head
Unsaturated
Infiltration 191 N/A 163 177 177
(in/hr)
Saturated
Infiltration 107 109 75 97 107
(in/hr)
N/A: constant head is not held during unsaturated test.
Biofiltration Media Composition
Particle size analysis demonstrates average gravel, sand, and silt and clay percentages vary by less than
3.4%among all profile depths when compared to a media depth of 12 inches (Table 8).The profile depth
of 12 inches was selected as the reference for comparison because historic depth profile analysis
demonstrates Filterra media generally meets specification at this depth and is not altered over time.
January 1, 2023 Con tech Engineered Solutions LLC 44

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Table 8.Aggregate classification profile comparison
Depth Gravel (>2mm) Sand (>53µm,<2mm) Silt and Clay(<53µm)
Comparison Difference(%) Difference(%) Difference(%)
3"vs. 12" -3.37 0.21 3.20
6" vs. 12" -1.02 -0.42 1.50
9"vs. 12" 1.76 -2.33 0.60
Average: -0.88 -0.85 1.77
Median: -1.02 -0.42 1.50
Organic content analysis demonstrates the average organic content varies by 0.44% among all profile
depths, and by 0.03%when the 3-inch media bed depth is excluded (Table 9).The data was analyzed with
the 3-inch bed depth organic difference excluded since the surface layer contained significantly higher
organic content in comparison to the other depths sampled.
Table 9. Organic content profile comparison
Depth Comparison Organic Difference(%)
3" vs. 12" 1.27
6"vs. 12" 0.02
9" vs. 12" 0.04
Average (all): 0.44
Median (all): 0.04
Avg. (excluding 3"): 0.03
Med. (excluding 3"): 0.03
Plant Growth Progression
Activation and maintenance records are provided in Tables 11-13 in Appendix B for each study site. Plant
height and width measurements were recorded during most maintenance events.Stem diameter was also
recorded in some instances. Results indicate plant growth progression over time at all study sites.
Reduction in plant height or width from the prior measurement at maintenance is due to plant pruning.
Table 10 shows plant height and width for study sites A through C increased 1.3 to 2.4 fold and 1.7 to 3.2
fold, respectively, over the different monitoring periods. The amount of waste removed, defined as
sediment,trash, debris, etc., was also recorded,with the largest waste retrieval summing 49 cubic feet at
study site A per Table 12. Waste removed is a conservative value since the mulch volume placed at the
prior maintenance was subtracted from the waste volume removed at the following maintenance, which
would have degraded some since the prior maintenance.
January 1, 2023 Con tech Engineered Solutions LLC 45

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Table 10. Plant growth progression
Fold Change from Initial Measurement
Study Site 'Plant Height(ft.) 'Plant Width (ft.) Duration (yrs.)
A 2.4 2.1 9.0
B 1.5 3.2 4.5
bC-1 2.6 3.0 5.0
bC-2 1.3 1.7 3.0
'Pruning occurred at maintenance throughout the plant measurement period resulting in
occasional reduction in plant height and width.
bC-1 represents Redtwig Dogwood at activation and C-2 represents later replacement with Foster
Holly due to plant injury.
DISCUSSION AND CONCLUSION
Biofiltration treatment mechanisms rely on a synergistic community of living organisms such as plants,
microorganisms and organic media to ensure long-term sustainable quality and hydraulic performance.
Plants and organics facilitate a sustainable biological cycle through regeneration of hydraulic function and
pollutant removal capacity through decomposition, degradation and uptake of captured pollutants.
Organic matter within the engineered soil media is sustained overtime to replenish adsorption capacity
through influx of organic material in stormwater runoff, plant root die off, and mulch degradation and
replacement at maintenance. Microbes support nutrient cycling through organic matter and compound
decomposition and biodegrade pollutants into less toxic forms. These biofiltration components support
the hydraulic and water quality performance longevity evidenced in the long-term data collection at the
different Filterra study sites up to 13 years with routine maintenance.
Organic Matter
As a biofiltration surface layer, mulch is the first line of defense for treatment via physical filtration and
chemical complexing, but also protects the underlying treatment media from scour and occlusion. The
media stays protected and infiltration rates are maintained while most of the sedimentation occurs on
the surface of the mulch within a biofiltration practice. The consistent water quality and hydraulic
performance longevity demonstrated by Filterra is likely due in part to regular maintenance which
includes removal and replacement of the mulch layer.In addition to trapping particulate matter that could
migrate into the media bed, it also provides many benefits thought to improve long-term functionality.
For example, supporting the biological community by providing organic replenishment to the media,
pollutant treatment, adsorption site regeneration, and moderating temperature and moisture within the
media bed. Ongoing maintenance will reduce the likelihood of needing to replace the media over the long
term.
Organic, wood-sourced mulch should be used due to its metals adsorption capacity and intrinsic
properties including humic compounds consisting of carboxyl and hydroxyl functional groups, cation
exchange capacity, surface area and pH. Wood mulch is known to capture oil and grease among other
organic compounds.Wood mulch is also a host for microbial and macro-organism activity which supports
January 1, 2023 Con tech Engineered Solutions LLC 46

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
plant health and pollutant degradation. Organic matter within the soil media retains moisture, provides
carbon to the microbial community, supports vegetative growth, and enhances pollutant removal.These
components are critical to long-term success by supporting inert and reactive filtration during storm
events as well as the biological transformations and sequestration that occur between storm events.
Vegetation
Vegetation is key to sustainability by supporting microbiological activity and maintaining an assimilative
capacity. Captured pollutants are biodegraded by microorganisms into forms available for plant uptake
via phytoremediation. As vegetation biomass increases, as observed in the growth progression data in
Table 10, so does Filterra's ability to capture and process more pollutants. Plants regenerate the Filterra
media pollutant removal capacity by making the media adsorption sites available for the next storm event.
Vegetation improves sustainability of the Filterra system by enhancing pollutant removal and uptake as
well as maintaining design hydraulic flow rates through root expansion, penetration, exudate production
and die-off. Roots shrink and swell during wetting and drying cycles keeping preferential pathways within
the filtration media open. Plant roots and associated microbiological growth provide exudates which build
and maintain soil structure.This increases macropore development for maintaining infiltration rates.
When healthy vegetation is part of the living ecosystem that makes up a biofilter, media porosity is
increased,soil structure is improved,and compaction is reduced. Plant roots continuously penetrate filter
media as the plant grows and the roots themselves die and regrow forming micro channels.This prevents
media compaction and increases porosity, maintaining aeration and hydraulic rates. Infiltration rates
observed during hydraulic field evaluation at study site A under typical operating conditions demonstrate
the Filterra system sustained hydraulic capacity after 14 years with routine maintenance and supporting
vegetation. At the surface, plant movement by wind or activity of birds, rodents and insects which
associate with the plants can increase hydraulic rates by breaking apart the sedimentation crust that
occludes the surface. Plants also enhance volume reduction through evapotranspiration.
Root-zone Macro-and Microorganisms
Biofilters with plants and organic media have more microbial density and diversity than non-vegetated,
non-organic media filters and therefore have more ability to transform and uptake pollutants (Hills et al.
2017). Microorganisms degrade and transfer pollutants into less toxic forms through nutrient cycling.
Nutrient cycling can include chelation for plant uptake, and sequestration of pollutants through carbon
and nutrient assimilation (Coyne 1999). Microorganisms alter the soil chemistry in the rhizosphere that
enhances pollutant removal efficiency. Plants increase organic matter in the soil through decomposition
of biomass, including the roots themselves, known as cell sloughing, which provides a carbon source to
the microorganisms in the media (Tugel 2000).
Additionally, mycorrhizae fungi create a symbiotic relationship with plant roots whereby plant roots
excrete sugars for the fungi while the fungi provide "pollutants"to the plants in the form of nutrients for
further biomass production. Mycorrhizae fungi increase the surface area of plant roots, which enhances
absorption of phosphorus, nitrogen, and metals, which are all macro and micro plant nutrients vital for
plant growth and reproduction (Lewis and Lowenfels 2010).
January 1, 2023 Con tech Engineered Solutions LLC 47

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Macroorganisms, like earthworms, live symbiotically with microorganisms. Earthworms increase organic
content in media by burying and consuming organic material deposited on the surface. Earthworms also
support continued hydraulic and pollutant biodegradation by increasing microbial activity, recycling
nutrients and altering soil structure through cast production. Earthworms also increase infiltration by
improving porosity and drainage with burrow creation. Earthworms are an indicator of healthy soils, as
evidenced during the hydraulic field evaluation at study site A(Figure 84) (Tugel 2000).
Media QA/QC
Filterra's history of performance success is predicated on a robust media QA/QC program. There is
oversight beginning with the raw materials through the commercially produced Filterra media blend.
Standards of practice have been developed utilizing rigorous verification testing for qualifying, sourcing,
verifying, producing, storing, and handling Filterra media. Media certification is based on a controlled
manufacturing process with post-production media validation required to ensure that the blend meets
specification. QA/QC procedures are critical to ensuring media consistency and function per design
specification (Hills et al. 2016).
Routine Maintenance
Contech recommends annual to semi-annual maintenance depending on the site location. As
demonstrated in the maintenance records in Table 12 through Table 14, routine maintenance is
recommended for quality and hydraulic design sustainability. Not following a regular maintenance
schedule may result in enduring later expensive restorative costs. Prior hydraulic evaluation on older
Filterra systems with large gaps in maintenance history demonstrated slower hydraulic capacity,
supporting the importance of keeping a regular maintenance schedule (Hills 2009).
Filterra maintenance requires removing degraded mulch along with sediment, trash and debris and
replacing with new mulch. Proper vegetative pruning should also occur as necessary not only for aesthetic
value but to preserve access for sediment, debris and spent mulch removal during future maintenance
visits.The mulch is typically the only component of the system that needs to be replaced regularly due to
decomposition. Mulch replacement will extend the service life of the soil media indefinitely. Organic
mulch replacement is necessary to support the chemical and biological processes with the ecosystem,
microbial activity, media regeneration and preservation, and water holding capacity. Additionally,
maintenance permanently removes contaminants associated with mulch and accumulated sediment
(Herrera and Geosyntec 2010).
For bioretention practices,degraded mulch looks much like a layer of soil overtime, and these smaller soil
particles and captured pollutants begin to migrate into the media bed, causing flow restriction. Longer
than recommended maintenance intervals may require removal of the first few inches of media to restore
hydraulic capacity. Therefore, new mulch cannot simply be placed overtop of old mulch; spent mulch
removal and replacement is required. A technical memorandum on bioretention maintenance produced
by the EPA suggests surface layer media replacement may also rejuvenate water quality performance
January 1, 2023 Con tech Engineered Solutions LLC 48

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
based on research by University of Maryland demonstrating sedimentation and heavy metal accumulation
in the top 2 to 4 inches of media (EPA 2016).
Performance Comparison to High Rate Biofiltration
Water quality results for the three Filterra long-term performance study sites were very similar to the
2020 International Stormwater BMP Database Summary Report (Table 11)(Clary et al. 2020) results for
High Rate Biofiltration (HRBF).The median influent and effluent TSS and total phosphorus concentrations
were nearly identical for both data sets. Median total zinc influent concentration was lower at the Filterra
long-term performance study sites as compared to the HRBF results from the BMP Database at 140 p.g/L
and 178 µg/L, respectively. Median effluent zinc concentration was also lower at the Filterra long-term
performance study sites as compared to the HRBF BMP Database results at 46 µg/L vs 60.6 µg/L,
respectively. Median influent total copper concentration was higher at the Filterra long-term performance
study sites at 29 µg/L vs 8 µg/L. Median effluent total copper concentration was also higher at the Filterra
long-term performance study sites at 10 µg/L vs 4 µg/L. Removal efficiency was significant for all
parameters.These are the only parameters with data from both sources for both types of BMPs.
Table 11. Filterra long-term performance versus high rate biofiltration performance as reported in the
2020 Summary Statistics Report by the International Stormwater BMP Database
Parameter TSS Total Phosphorus Total Copper Total Zinc
Units (mg/L) (mg/L) (µg/L) (µg/L)
Stormwater High Rate Filterra High Rate Filterra High Rate Filterra High Rate Filterra
Biofiltration Long-term BiofiltrationLong-term BiofiltrationLong-term BiofiltrationLong-term
Control Measure Performance Performance Performance Performance
Median Influent 30.8 31.2 0.099 0.1 7.95 29 178 140
Value Effluent 3.8 3.6 0.05 0.05 3.75 10 60.6 46
Significant Median
Value Reduction I Yes Yes Yes Yes Yes Yes Yes Yes
(Mann Whitney P-
value 0.05)
As noted previously,the HRBF category in the 2020 International Stormwater BMP database summary
report is comprised entirely of six Filterra field studies, conducted over a period of one to three years
and all initiated within a few years of installation. None of the long-term performance study sites are
included in the International Stormwater BMP Database.Therefore, the nearly identical data sets from
both sources indicate that Filterra long-term performance is similar to initial Filterra performance for
these parameters.
The synthesis of similar long-term performance data for conventional bioretention systems was beyond
the scope of this study. Some evidence suggests that conventional bioretention performance may
improve over time, particularly as labile nutrients and dissolved metals originating from the media itself
are flushed and vegetation matures (Herrera 2016). Further research is recommended at multiple sites
to assess changes in long-term performance for conventional bioretention systems.
January 1, 2023 Con tech Engineered Solutions LLC 49

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Hydraulic Performance
Pollutant load reduction provided by biofilters is a function of concentration reduction, runoff reduction
(via infiltration and evapotranspiration)and capture efficiency(the proportion of average annual flow that
is treated). This research demonstrates that Filterra systems continue to provide consistently high
concentration reduction for typical stormwater pollutants over time. Filterra capture efficiency depends
on the ratio of the system to its contributing drainage area,which is dictated by local regulations, and any
changes in hydraulic capacity over time. It was infeasible to conduct hydraulic testing repeatedly over the
life of each system. Instead, the media flow rate was tested after completion of the pollutant removal
testing to establish a worst-case scenario for final flow rate.This proved to be difficult for Site B and Site
C, as impacts on both sites prevented the final flow test. Site B experienced an oil spill in 2011 resulting in
a full remediation of the Filterra system media. Replacing the media with new media rendered any future
flow data irrelevant when compared with the pollutant removal testing. Site C was razed in late 2018 and
fenced off preventing any additional testing or maintenance. Site A was therefore the only site remaining
at the time of the final flow test.
Given that all 3 systems were originally designed based on a media flow rate of 100"/hr,the unsaturated
median rate of 177"/hr proves that the system continues to outperform expectations even after 14 years.
The saturated median media flow rate of 107"/hr also meets the original design flow rate.
A decrease in infiltration rate was observed from the unsaturated test to the saturated test. The
unsaturated flow rate is a measurement of the flow through the system when water begins to enter the
system up to the point that the media is fully saturated. Preferential flow paths and moisture deficit can
draw the water into the media at a higher rate than under saturated processes via matric potential or
suction head that draws water into unsaturated soils.The unsaturated flow rate would represent the flow
rate that would be seen in a Filterra system in the field without an antecedent rain event (i.e. storms
within a few hours). On the other hand, the saturated flow rate is a conservative measurement of flow
through the media after the media has been fully saturated. Antecedent moisture conditions are more
uniform where saturated flow processes dominate. Saturation of the media can impact flow because
water droplets occupy media void space and organic particles within the media can absorb water and
change shape, modifying preferential flow paths.Therefore,the saturated media flow rate represents the
flow rate that would be seen in a Filterra system with a large antecedent rain event.
The time between maintenance events at Site A was longer than the recommended 6 to 12 months on
several occasions throughout the monitoring period, which may have contributed to increased fines in
the first several inches of media. Per the organic content and silt and clay analysis,the results are outside
of the media specification in the first 3 inches of media, as demonstrated in Table 8 and Table 9, which
likely reduced the potential 175 inch/hour infiltration rate recognized by many localities. Replacement of
the first several inches of media may be necessary after 10 to 15 years to restore the original system flow
rate.
January 1, 2023 Con tech Engineered Solutions LLC 50

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Biofiltration Media Composition Correlation to Performance
Particle size analysis demonstrated less than 3%differences in the average gravel, sand, and silt and clay
percentages among all media profile depths.The particle size analysis data collected remained within the
Filterra media specification after 14 years of operation,supporting the quality and hydraulic performance
observed over the monitoring life of the system. The particle size and organic content analysis data
showed the media is preserved, sustaining pollutant removal and hydraulic function.
Organic content analysis demonstrated the average organic content varied more in the media surface
layer in comparison to the other profile depths due to increased organic material (Figure 37). Moving
upwards from a 12-inch media bed depth to 9 and 6 inches showed no significant difference in organic
content (< 0.04%), however a higher percent difference of 1.27%was observed in the 0 to 3-inch media
surface layer(Table 9).Additionally,slightly elevated levels of silt and clay were present in the top 3 inches
of media (Figure 38),which is to be expected after 14 years of operation with several missed or extended
maintenance periods. Moving upward from a 12-inch media bed depth to 9-and 6-inch depths showed a
gradual increase in silt and clay,with the highest percent difference of 3.2%observed in the surface media
layer (Table 8). While the system still maintained expected quality and hydraulic performance, the data
suggests that replacement of the first several inches of media after 10 years may be beneficial. Media
surface layer replacement ensures hydraulic function is not compromised as the surface layer may
become richer in organic material over time due to natural degradation processes.
January 1, 2023 Con tech Engineered Solutions LLC 51

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
0
f+ •
"! /
r � �
v• •
1,
• • ,.i •
♦
z
z
0
_ . _
� z
3
- O
• W
• cc
'
Figure 37. Study site A Filterra media profile composition analysis of organic content after 14 years in
operation
January 1, 2023 Con tech Engineered Solutions LLC 52

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
I•
O v•• . - •
A.
•`_ ..
I
—
cn
k1 W
cc
Figure 38. Study site A Filterra media profile composition analysis of silt and clay content after 14 years
in operation
January 1, 2023 Con tech Engineered Solutions LLC 53

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
CONCLUSION
Filterra long-term performance is similar to short-term (1-3 year),third-party verified Filterra field
studies for TSS, phosphorus, copper and zinc per the 2020 International BMP Database. Filterra long-
term performance meets or exceeds conventional bioretention performance with each system type
providing significant reduction in TSS,total and dissolved zinc, and total copper. Filterra also
demonstrated significant total phosphorus and dissolved copper reduction while conventional
bioretention showed insignificant dissolved copper removal and a net export of phosphorus (Clary et al.
2020).
Biofiltration systems with plants and organic media support the hydraulic and water quality
performance longevity evidenced in the long-term data collection. Filterra performance should remain
consistent over time with routine maintenance based on long-term quality and hydraulic performance,
and media composition analysis. Annual to semi-annual maintenance depending on the site location is
recommended for quality and hydraulic design sustainability and avoiding restorative costs.
Maintenance requires removing degraded mulch along with sediment,trash and debris, replacing it with
new mulch, and vegetative pruning as needed.
Additional long-term emerging contaminant studies and a better understanding of vegetation's role in
contaminant reduction are needed. Long-term performance and comparison information provided in
this report, along with future research needs, will help design engineers and approval entities make
more informed decisions on selecting stormwater control measures as sustainable solutions.
REFERENCES
ASTM (American Society for Testing and Materials). 2011. Standard test methods for organic matter
content of athletic field rootzone mixes. F1647-11. West Conshohocken, PA: ASTM
International.
ASTM (American Society for Testing and Materials). 2018. Standard test method for particle size
analysis and sand shape grading of golf course putting green and sports field rootzone mixes. F1632
03.West Conshohocken, PA:ASTM International.
Coyne, M. S. 1999. Soil microbiology:an exploratory approach. Albany, NY: Delmar Publishers.
Croghan, C., and Egeghy, P. 2003. Methods of dealing with values below the limit of detection using SAS.
RTP, NC: United States Environmental Protection Agency Southern SAS User Group.
Ecology(Washington State Department of Ecology). 2018. Technial guidance manual for evaluating
emerging storm water treatment technologies. Technology Assessment Protocol-Ecology(TAPE).
Olympia, Washington: Ecology Water Quality Program.
Geosyntec (Geosyntec Consultants). 2008. Filterra Field Flow Rate Evaluation Report.Acton, MA:
Geosytnec.
January 1, 2023 Con tech Engineered Solutions LLC 54

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Herrera (Herrera Environmental Consultants, Inc.). 2016. Pacific Northwest Bioretention Performance
Study Synthesis Report. Prepared for City of Redmond, Washington by Herrera Environmental
Consultants, Inc. Seattle, WA: Herrera.
Herrera (Herrera Environmental Consultants, Inc.) and Geosyntec(Geosytnec Consultants). 2010.
Filterra Bioretention Systems Technical Basis for High Flow Rate Treatment and Evaluation of
Storm water Quality Performance. Seattle, WA: Herrera and Geosyntec.
Hills, M. 2021. Filterra Field Hydraulic Evaluation SOP. Ashland,VA: Contech Engineered Solutions LLC.
Hills, M.,Allen,V., and Lenhart,J. 2017. "Investigating the Microbiology of Bioretention." In Proc., World
Environmental and Water Resources Congress 2017,416-430. Sacramento, CA:ASCE
Hills, M., Kay, E., Newcomb, M. 2009. Filterra Field Flow Rate Evaluation Report Addendum.Ashland,VA:
Contech Engineered Solutions LLC
Hills, M., Macleod, A., and Lenhart,J. 2016. "Ensuring Bioretention System Performance Success:
Guidance for the Verification of Bioretention Media via Quality Assurance and Control Testing." In Proc.,
World Environmental and Water Resources Congress 2016. West Palm Beach, FL:ASCE
Lewis, W., and Lowenfels,J. 2010. Teaming with microbes: The organic gardener's guide to the soil food
Web. Portland, OR:Timber Press
Tugel,A., Lewandowski, A., Happe-vonArb, D. 2000. Soil Biology Primer. Ankeny, I: Soil and Water
Conservation Society
USEPA(United States Environmental Protection Agency). 2016. Operation and maintenance of green
infrastructure receiving runoff from roads and parking lots technical memorandum. Washington D.C.:
USEPA.
WRF (The Water Research Foundation). 2020. International storm water BMP database:2020 summary
statistics. Denver, CO: WRF.
January 1, 2023 Con tech Engineered Solutions LLC 55

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
APPENDIX A: PHOTO LOG
Filterra Plant Progression Photos
-
" ate.
? -.,`S
r•
vt
Figure 39. Study site A plant progression 2008
-'.1., •'• •
m
.. - V '' 1+4G, a•� 1_ r i .-•
+� _ I '� * 1
�. .' 2
1 .
1. t
t)' \. �j. i•-'.*b• ...
1FTC-1
y; '',�. ...,ram' •AO •i`, _ . .
-- i//_
Filielli
IL. -_ 1011—
.# i .vs
Figure 40. Study site A plant progression 2010
January 1, 2023 Con tech Engineered Solutions LLC 56

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
A
(-2e n •� ; .
ar'S�
S -a
�11�1
y,. —-- — v
/ ,_
/ I . Mil,1
Figure 41. Study site A plant progression 2012
' c\ tr;bN' r
Iti -... •
. +`,,.i ,. - ."7 0'*'•p o. !Ih..c,., ..>4i2
- ^ s tire.'
i.
.l` { Oi-1 v t dam. 'xf'
' 'r:4›...!L-- - t•I•or'.. /(4ke".- ''....,- 1: ..' . •-• -
, r: _
A
•
10
�- •
.0S
-4
Figure 42. Study site A plant progression 2014
January 1, 2023 Con tech Engineered Solutions LLC 57

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
4.'4
rI .�C ,Lr,k'.•-.
r
•
V
r
a!<< kkkk(l 2,
A: ,
I.
Figure 43. Study site A plant progression 2015
• ,' . M ' . 1 r ` '
„,.. -'-'-'••••'14. ii. .,.1.'..' 4•00:_.::.A::::::: .....:41:,•(41411:..... '.s,._ -
'ram- , Ap
` ��\;\\ `2VIA
-a: • 1
•
,.1, .• 1I,. 1
Figure 44. Study site A plant progression 2018
January 1, 2023 Con tech Engineered Solutions LLC 58

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
74, , .. , wino. — ezawsk-;---antmvifw-ipm=-7-_ -I=
iiili• - . --, 1. .. e' -.. r. . •..; ). 1 •;',.:
, . 1( .. ''. a. •
!• • • . -1 •
i • ' ., ,
•
..,. , .. .
-•- •4'- •-•''4
- •0 r
.4._.•P‘s •• :::1.1....1 -• --
• •-v-,i,
.:,•• - ' ''' . "'•'• ...."'" ...-.
b; .•••?. . -, ;.•-. ' - .-.' • . •.: s.-- •••
, . ,_, ... • . . ./....;;........4..
• - Il % r
. . ..... ,,.,
-- . ...
... a; • .
.... .
•
4.-, '.-- - • ,'-.- .
. _
- . ..,_ • -, -.,.: - .
. .
Figure 45. Study site A plant progression 2019
• • • is •% . .4 •- .
'' .•': • .' ;42.,‘. '4. • - '''' •4.
„.. ..... IL
s wOlir: 01
"' e -•
- '. '.t. ,•v 1 4'9P .
411%
-il v;,
a kv 41I7.: ' • . .. .mmillm. - °IP
.,
.
- • • 0 01 ii. , •
• . ,k 1 .. f. r I ' 4.I A• ' .*% •
' ' . It S'Ill• ' IP. . .# 1 4.
. • *. • • .i'off i,i,
1. . • -
4.*
•r -' . ...
. .4 -- %.;4 i' '4; .4
• S •- a -, AP.I'll,VC ., .
‘4t.. .."1.
•;., .' • •• •• !opt . ,(.. i
4-2• . "t. 'ti li ' . JA
tr.- ",- • . . • . .. •:.r, 11
..If I; : .4 ' P. . eA4
isitt, . •
' r". .......
: ....... .. ...r,... .
. .
.. . • • ,
. . •
. . ..
• ..: .
-• - -
.. ''' 0001r"flir .4.0--1•Peet-- .
;.•.-
•— - 1, Ale . ;
_..
.--- - -"---..- # • •
- i
„,„:_ _ . • ..,
..,,,„,;_,• •
•...4.- '411.1.1110.,,
.1 ••• i''•: 'AP _
Figure 46. Study site A plant progression 2019, cut to single stem
January 1, 2023 Con tech Engineered Solutions LLC 59

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
--.- - - . ' ;. . 7 g Jw...„4,1
, . . 7 ' A 1
sir c • %. ' ti
•r
.e'•i' 4, %.jA,
l �w,' . ,ram
.j
10- • V. 4,1*
46 . . .
• , .
•41. .. .
• - •
r •
.' 'jAt * . 4'. T .S •' ;*
. lit\ \
•. _ 'Pr -416. —
Figure 47. Study site A plant progression 2020
• . • -
r
•
, -Z ' *40/:lit:-4;,:e.., '
3.. i V iir, ter K °' . r
..... �.."'
•
Figure 48. Study site A plant progression 2021
January 1, 2023 Con tech Engineered Solutions LLC 60

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
i
Al .4 11... a ,
.a. ti- .}. . .
il
' '- ,' r
t
Figure 49. Study site B activation 2005
. -T
optcs
r i4
' .R
. 4 X '' - "..s. 1 1 cat:: .: _�
w.
L .c
.•` 1 .
r • •,•,
� ` ', ., ., , vA
f r! 4•' 6
•
Figure 50. Study site B motor oil from oil service station caked on mulch surface in 2005
January 1, 2023 Con tech Engineered Solutions LLC 61

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
11
yam • J-_ Vb.
. .
N - IIlIN1 1 f -_
" . - --4-::1.r_.: - ,i . ,00
.,..
_ f ..„ , _
. _
-i - .1-
_.
kki,t‘ • , _
,....., _
y ,
. •
__ , , . .
. .
Figure 51. Study site B plant replacement with Northern Bayberry from heavy motor oil contamination
in 2005
i
A
1116‘_ • - ,.
1 _
. .
.:
Figure 52. Study site B motor oil residue evident in drainage area feeding Filterra system 2006
January 1, 2023 Con tech Engineered Solutions LLC 62

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
_ _ -. •,. ..-1—.44 •,, f...
. • . ' •''‘. 1.1.„ - ;•.;O.•.
. '.' • r le t40, L -- ...A. dr-
-,•.;
, -
. .... .. •
......'
t
. '. .s'1., .0114.7 Ntli• ''
' --..-...PP.
dir
. , MIP -,.iZallarmi.
.
...
.. •
• —-
— - •
.-
---.
•
141
- a-- . - ..•
- - - - .
:• . ...t. ,......:.........,,,rom......
•
Figure 53. Study site B plant progression 2006
• •,,,Z. ,,7,, 117PFPIr 4--..; , ,,,.. ,
,..
•,i
,-
„AO
W' - - '',4'''... . 4.,/, ‘ • r; ri. Igi I
- ,CAP.. • T -1,./4-- '' ' ,, .0.- • '-', 1 I
_'1, ' 'V/' ..,50, **41•11/..:,_
-:. l
4-,-- ' ' • :ii• -,. .. ,
• ---If. -''.
, • —•
. .,
,-, \‘, ,,,, ,4, ". • - , — -... - '.7,--. ' ' .
^: fk.3 / •-• '. za. --*4 i •••, l'i .....1,N.1,,`.0.. • ..
. i s!.'''''-'' • • .•v 17,e -- -- ;A:• a" ••• X e• .0'54
z. . ,).•f -
.: —,...
: :1:4•.-i- - s .s ,', ...: *% V , '''' -,
'0...4.
I ,
...
......
, ilitto
-441‘ ka
. .
_ ---•*".
-
_ s
• . ..._
' .I 1 .
-• ' ,-s-
410 INItill*t ... _•.
• -" .
-• 4 .
Figure 54. Study site B plant progression 2008
January 1, 2023 Con tech Engineered Solutions LLC 63

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
L Y
p4,"'.. . -..aft
_ t.„...:_ .
/ ., .
,. .y iW.
•
. ..
..„ ,,.. „......
,,,,„tf . , iipo
, t_. ,
,.. . ..„ . .. ‘....„. .
,1.u. ,1.` `%
lei'_
.
•
/
•
�/
•r
irilliall*- ,
' 1 -
. '
Figure 55. Study site B plant progression 2009
I
gmij
•9r 4 •of n '� Y` , . '.. .. rr�.
Imo.- - a . Y•er
_ 1,: ..air a
-:7 I, �-
'''''tsas41104s446,_
. • 1,.%.4'el, 71441/1111.."- / : ...
Figure 56. Study site B plant progression 2009
January 1, 2023 Contech Engineered Solutions LLC 64

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
TOD
-71
, ri i
'; i 4`, '4 °' a fs'N ,_�- .. t t w n • , t
. re. /pp s
II,
j; iu
st . .
r_..
... . .".. .,k . ,...
_, ,a;iiffykib,"-1_,-,
Figure 57. Study site B plant progression 2009
* .: —
• ♦ r
•
to - �'i
Y
i i \ .
•rI{ - •mosio-1-M* (. /� , .4 • .. ,
1 1 ♦ air • _ �� �. `•- /
Y• .ji s_
it/ . I—
s _
4i',I
•Ak `
I, •
1 w
, . ti
pi
niAll .
_ . • .. -
Figure 58. Study site B continual motor oil contamination to Filterra system; monitoring discontinued
January 1, 2023 Con tech Engineered Solutions LLC 65

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
----7"--7--
"1-"1.'..1-11 ,11 ,..111111 III 4
t .. 1 — ----4 MT P
rl / -
- fr ,
1 —1: 4 —
_, r-Ar._,, ...4, 11 •
•
1 , ,
, i •. . 1 1,....-----, i , . _
, = -_-_ • , -41 s.1 All 1 :•
I 01 .I .. ,,,,,t.. - , .
. ___
, , - I
_
. -k...u.,.:.-,-Ei
11.2__T_al 7ip 1 r-
•. ,,,..,..- , 4, •
__
t.,- A 4. •
. '-‘1,- - .. ,e-:-/- , i = *•., v.
=,..71k• t.,,,..".% ., :5i..-- --
, ' -.. --;...-- -`...--Air- , ...., . _ „,...., .• 6 k,•.r-i .5....._...., ::,..f
' iy. 8.-•..::.__.
1116-...`"#.4..44,0,, -:: - -1'!,',..,..i.: 1-- -.4'•10,.. ..;,,t; .,--..-r. ...
•A: att.- -......,.....
„ -
- .,, " o''''f::-: 4',At• I -•."qt... ••- C- '''',"•"1,."...:,,„,.,,,' .. "•4.1•40v
••,-.1.+.4,4"-l''''-'•'• '
.. , •
''•• ...,- v • ... " , ,
•
•
s,
.. _ _ - _ - •...,'''"",064;...' '' --"*.'‘ - '. •a. --riddle
Figure 59. Study site C plant progression 2005
1
----_-,
— "
91:",`,1:',!'"
...•
• ._ 3;, .ast . , ....
T. c...- r-r-- • *7
;Fai"-
-a-f
, . . • . ..
. . _
i ...
- - - 4.4........ . . x,. - ..„. .
----
- ...' ". .
•
- lir 1 •
-0 ‘" • A
, • - .. 4
- I
Figure 60. Study site C plant progression 2007
January 1, 2023 Con tech Engineered Solutions LLC 66

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
k, 0«.1"
., `s r. ,,t-
•
k:. ; ,•. '•�1 \- Cp' }�• • ate►• Y• •+
M _,Ns ,e--+� •lam ( ! \' , 1•' ,,�• ••.� f
. 1
, : r` {:
....a ,..
,, • •
vim,
t;.
Figure 61. Study site C plant progression 2008
4-
. , 41!..,,,, if,,,,
. . . .
. 7-4. _ t " 1,-,4 ki '- 0i,604 .
'` se \,111! ` %� ,,yam
. .,,mil :r1. ..*i ,.
�:. r� - ti
r.-
fir+' 1tit
,...1k.
,
"` ''' -- ^. SSA.. 1�
Figure 62. Study site C plant progression 2008
January 1, 2023 Con tech Engineered Solutions LLC 67

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
.r ... �/ Sw R: , s
• Ne_. q '"�a• •.�•• t A
•
`-rc�.`, e 77�i may.
wt k I '�.
. :F. xtv,) . A . 4,, . ..,
t 11 . ,..,,,,k
•
Figure 63. Study site C plant progression 2009
t ....i ,
i r , �Y 1
.'...-it 44 .4' — •••O
• •, .r`[.,. e 4..t .r art'' ''. `;';'•-'
•
•
of /� ` - �.
. oikeetesks....i,.... .MIONF,7 F.-tili:•1/4‘ z.'w-
E •'=t'_ _ice.
• '�iwc- w ar
Figure 64. Study site C plant progression 2010
January 1, 2023 Con tech Engineered Solutions LLC 68

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
.,-7+ ,• . / f•
30
'* ,- .. Srµ , ,}` r
• .. •
. '.._
i
r ■■ 47.. , 3
' "�'`1 T
'al
_ --
�:
Figure 65. Study site C plant progression 2011
•' 4
5%i
] a s
J
:- - -
ti-
...0 • , . _
I. _ ' I - s
Figure 66. Study site C plant replacement 2012 with Foster Holly; photo taken 2013
January 1, 2023 Con tech Engineered Solutions LLC 69

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
• j
` 1.
'!' - • t -, -at �f
�aha. �.�py,,♦'.. _~ yea KI
Ai
Figure 67. Study site C plant progression 2014
Fr
i f4.
ei
I '4
` '
r
Figure 68. Study site C plant progression 2015
January 1, 2023 Con tech Engineered Solutions LLC 70

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
: . i'...:=I::;,...-:'.' - ' 1. .... 4 r..
x"r
Inakiis
44. f i�•
_ 3t
h'�LL/ +
v -
\..A
Figure 69. Study site C plant progression 2016
, '
�.
. 4; -. _`+;.,� '01. kN: �
Iv
�.- tti..t4:.."5
, . _
- .
•,?.. ,t,; . ,.s���,. , .-- •.
. .y + .I
_ .,
= r'
1- ✓
-. \\\\\\\\\\\�����/
/,
Figure 70. Study site C plant progression 2017
January 1, 2023 Con tech Engineered Solutions LLC 71

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
•
•
� !y
sit
40.
h ~
+.h
17FY
"tyk • ,
Figure 71. Study site C plant progression 2018
s
• i
w ..
0111
Figure 72. Study site C decommissioning
January 1, 2023 Con tech Engineered Solutions LLC 72

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Filterra Maintenance Photos
4 i, . : ,.., - , __A i.i-,..t..--,i. i`.7.-.\--s-F.,i--' ',',1(''. -.=,-...'' ."
4\
{ j y 1d•.,�,i �..K.F� cr7"F•
•
, q' •1}'A w. i .t t l�pF
•
•
v,i ,,:. . r. t �� •...1 -." , r, i�.ii C�Af �0,),„ l �Y` ,41i,,,,,.LTt' y •k .:;' .,
. ;' c "a _ " ' • le,:. �.i.a ,; yip ry n e: .'4
e t ) y:. i. . «: i t
t, I ra m . . : , .% ,4. .'1'F: .' :., y�}`A ,, j'
iSr .. '� `Ajar h' . 1
_ .•fit > r .ar k44
s
(a) (b)
Figure 73(a)(b). Study site A pre-maintenance (a) and healthy roots evident post-maintenance prior to
mulch replacement (b) 2017
January 1, 2023 Contech Engineered Solutions LLC 73

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
itedirii i p r , . - - ' -
a ;;;^y ;
_ V. : ' 4 ... -16 14
: .44,,41 .. ..,,-. .;.,
,.. _ , ,- ,-- r l''
. 41,... ,,,. •
�.i - r \
• al `, 1/4)
Figure 74. Study site A pre-maintenance 2019
_.,--....., _,.---,, Olt ,, z .
,. .-
, .
1,...
.-.4
..„.
--
Yi'
. ,•
.,....... - „ , .
.„, " ..,.,•.
.,. ,
..
. .
_ . _ ,
, . . .
. . .1 •
• w ' �.
Figure 75. Study site A healthy roots evident post maintenance prior to mulch replacement 2019
January 1, 2023 Con tech Engineered Solutions LLC 74

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
•`I , . •/'/ J�-. '��''7 ' \I -�{..• -TI .
. . t •�It` n ' / 44 a �1 ��/n WI,L •
:r a _� ��r»7t i , if
'r I!I �PYf ,41.r 4 1
A •Al R~ , /' r t \
//
t ` ` `'
Aitri
e . of
..:Eat r�'S,el ...f i. � 4 x "
r
I'� 1 / ; �. t .`
•♦ • w.
. `Y r 1 V 4i (L. s
(a) (b)
Figure 76(a)(b). Study site C pre-maintenance (a)and post-maintenance (b) 2016
�`• + , •
1 _ '_ •_ ._ {
`., ► .
1 w
\{. ,), ,-�,1 h tT,y ji .ti .1 it y '
•
ate/i�t`- '; J�°�. tt '4t !' t ; • t•`: * t-
a 4 a ,, a tiK 1 .. ' S ' rt,' %`t ...
.,ti •"� �y`\ f.:�yyx;��'t''t,�'S4 _ l•
y •
`#`w ft, t { r. • `t
• 3 \b�.'7 ` ttY 1 vY
T
(a) (b)
Figure 77(a)(b). Study site C pre-maintenance (a)and healthy roots evident post-maintenance prior to
mulch replacement (b) 2017

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Hydraulic Evaluation Photos
El l'i- ••
Z t k;
_ r�` '
1 _ ` Y
r�f'�
- �� 6
I �9fr91d' �'` L.' 't v., ! _
Figure 78. Study site A Filterra system hydraulically tested 2021
110040.1
\•. -
I ` ., II._,:� -r- .II _
___ .
_ _ ..
,Ar...
. ...., , ,,,,,,, ..,,,, f -
_ _____
, .
-1
41.
l .:i.
i
/ ' r`, -( '.., ' . fit
.r. ♦•^^\.:.lrs J �(,'„1, ri,�, .` T. rr4gr',.
4 'rf"r' ;vfir. 4 i,4' $ 4ay'it�G-
C 1 .3
r,. r ' 't K h h F•. , n P
� f
Figure 79. Hydrant and meter set up
January 1, 2023 Con tech Engineered Solutions LLC 76

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
M .
H a
i v� _ (�
3.
, ''. ' N rl 4 1 1q°..ki'll4 \''°. \ti il\
k;. •
0 , .a. , t.* la ".1% , N
/ i\. b ?"'
4 '
L I • it -.,otiti V./i. `u' 7.y
„,w 1ie„•, tom'
4�1 9 . . i, , �\
-• r0 i .1‘,
4 . ' , 1 - 4
% / / \
Figure 80. Flow monitoring equipment set up
`" ' I
Y
3-.-
t -d, t �' n iY Q,Y -;( •e
r.
fy.�#7",0f. r .d- q ,.tr- i 4.
r - ,, , ,
_„,.. :.,.,
,,:
..:..-.,
-
A d ry' /, �� :„, i 5 k.gf T i
'n � A1 io"�'�!,a4',.. A�� r.s —
Y x� AA n ..
j%4 ,4� 7
S r
'',"; .14A41 i",4r.;l'," iriforz:4„.., 7.-,,,,,,' .. ' :if •.4" " ,, y,
Figure 81. Curb influent flow
January 1, 2023 Con tech Engineered Solutions LLC 77

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
j. W
.
.1 T
y
•
,. i . "l t
Frifr . .r 1
Figure 82. Initial flow during unsaturated test phase
tea"k .
• k„ 0 1'
.. .4 ,.
Y - r
• "•v' �'w
t
'
\\ - -
; K, . ,.,A1,, ..7...... N. •._ • .,,,„.:
. .„ ,.._.,.4_,...__„. .... .,... , .w ., .,.,„, . ..07;.:._
w/ . _ . . , •,,, , ,w, ,. t
,... It..,: , :,..„..,..., , " ‘,,,,,,,.. 41, 1, ...,
..., • -
r �.
111111111111.111111111111.11111111
Figure 83. Head development during unsaturated test phase
January 1, 2023 Con tech Engineered Solutions LLC 78

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
• fk. tifeg.' . - , I ' ;AZ, . '0,‘
4 ,, •
,..,.. • 4.,.. b., 1.• '
f. •
y,�
H
4
Figure 84. Worms present during hydraulic evaluation
January 1, 2023 Con tech Engineered Solutions LLC 79

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
APPENDIX B: ACTIVATION AND MAINTENANCE RECORDS
Table 12. Study site A activation and maintenance record
Date Plant Height (ft.) Plant Width (ft.) Stem Diameter(in.) Waste (cuft.)
b4/13/2007 aN/A
12/11/2007 2.5 2.3 2N/A 0.0
5/11/2008 2.0 1.0 2N/A 0.0
2/25/2009 aN/A
1/26/2010 aN/A
8/30/2010 6.8 4.3 aN/A 4.7
7/25/2011 aN/A
1/23/2012 4.3 2.8 2.0 0.7
9/13/2012 4.5 3.0 4.0 1.4
3/28/2013 4.8 3.2 2.0 3.4
10/14/2013 4.3 3.8 2.0 2.7
4/23/2014 5.5 4.4 2.5 3.4
4/3/2015 5.3 4.3 2.5 0.7
10/26/2015 4-5 3-4 2-3 0.0
4/6/2016 5 -6 4-6 2-3 2.7
10/11/2016 3 -4 2-3 1-2 0.0
4/28/2017 5.5 3.0 5.0 0.0
10/9/2017 4-5 3-4 2-3 8.0
4/19/2018 5 -7 3 -4 3-4 1.4
10/11/2018 5 -7 3-4 3-4 1.4
4/8/2019 4-5 3-4 2-3 4.7
10/24/2019 4-5 2-3 3-4 2.0
4/8/2020 4-5 1-2 3 -4 3.4
11/5/2020 4-5 3-4 3-4 3.4
4/14/2021 5 -7 3-4 3-4 5.4
Sum 49.1
aN/A
bActivation, planted Nellie Stevens Holly
January 1, 2023 Con tech Engineered Solutions LLC 80

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Table 13. Study site B activation and maintenance record
Date Plant Height (ft.) Plant Width (ft.) Waste (cuft.)
b12/1/2004 aN/A
`6/17/2005 4.7 2.7 2.4
11/1/2005 6.6 5.3 1.4
5/3/2006 aN/A
d4/17/2007 5.2 5.6
1/14/2008 5.5 6.2 3.4
6/22/2008 aN/A
4/30/2009 7.1 5.0 1.4
11/4/2009 aN/A
e9/13/2010 6.3 8.6 3.4
f2/17/2011
Sum 11.9
a N/A
bactivation, planted Foster Holly
`Dead plant, heavy silt/oil on system surface, replaced w/ Northern Bayberry
dMotor oil on system surface
'Dead plant, motor oil on system surface
fOil spill, maintenance and monitoring terminated
January 1, 2023 Con tech Engineered Solutions LLC 81

FILTERRA®: ANALYSIS OF LONG-TERM PERFORMANCE
Table 14. Study site C activation and maintenance record
Date Plant Height (ft.) Plant Width (ft.) Stem Diameter(in.) Waste Removed (cuft.)
b5/27/2005 aN/A
11/3/2005 4.0 3.0 5.4
4/11/2006 aN/A
5/1/2007 5.8 8.0 0.0
1/29/2008 7.8 9.0 1.4
4/29/2008 aN/A
12/17/2008 aN/A
9/11/2009 aN/A
8/30/2010 10.3 7.3 3.0 4.0
7/25/2011 8.0 9.0 0.0
1/25/2012 7.2 6.6 2.0 4.0
`9/13/2012 4.5 2.3 2.0 2.7
6/17/2013 5.3 2.8 1.5 2.0
11/11/2013 4.5 3.3 2.0 0.0
6/25/2014 4.8 2.7 1.0 0.7
12/16/2014 4.5 3.0 1.0 0.0
6/23/2015 5.4 3.9 1.0 3.4
12/29/2015 5 -7 3-4 1 -2 N/A'
6/13/2016 4-5 0- 1 1 -2 2.7
12/7/2016 4-5 2-3 1 -2 6.0
6/6/2017 5 -7 2-3 1 -2 6.0
12/13/2017 4-5 2 -3 1-2 2.0
6/6/2018 5 -7 3-4 1 -2 2.0
Sum 42.5
aN/A: Maintenance occurred but record not available
bActivation, planted Redtwig Dogwood
`Replaced Dogwood with Foster Holly
January 1, 2023 Con tech Engineered Solutions LLC 82