Citation
Sustainable design of urban porous landscape detention

Material Information

Title:
Sustainable design of urban porous landscape detention
Creator:
Kocman, Shauna Marie
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
xviii, 121 leaves : illustrations ; 28 cm

Subjects

Subjects / Keywords:
Rain gardens -- Design and construction ( lcsh )
Storm water retention basins -- Design and construction ( lcsh )
Sustainable design ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 114-121).
Thesis:
Civil engineering
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Shauna Marie Kocman.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
656803233 ( OCLC )
ocn656803233
Classification:
LD1193.E53 2010d K62 ( lcc )

Downloads

This item has the following downloads:


Full Text
SUSTAINABLE DESIGN OF URBAN POROUS LANDSCAPE DETENTION
BASINS
by
Shauna Marie Kocman
B.S., Environmental Engineering, Colorado State University, 2001
M.S., Civil Engineering, University of Colorado at Denver, 2006
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Civil Engineering, Sustainable Urban Infrastructure Project
2010


2010 by Shauna Marie Kocman
All rights reserved.


This thesis for the Doctor of Philosophy
degree by
Shauna Marie Kocman
has been approved
by
Anu Ramaswami
Rajagopalan Balagi
April 16, 2010


Kocman, Shauna Marie (Ph.D., Civil Engineering)
Sustainable Design of Urban Porous Landscape Detention Basins
Thesis directed by Professors James C.Y. Guo and Anu Ramaswami
ABSTRACT
Porous Landscape Detention basins (PLD) capture and filter stormwater from micro-
rainfall events while taking advantage of the intrinsic quality of plants to act as water
treatment systems. The current design recommendations leave opportunity for the
incorporation of waste symbiosis and holistic design concepts. This thesis
investigates the physical, chemical and biological performance of the currently
recommended PLD design and two waste-incorporated designs.
The beneficial reuse of urban waste stream materials into sustainable filtration mixes
is assessed and paired with environmental life cycle analysis. Results of waste
screening tests indicate that the currently recommended 15% peat and 85% sand
(control) may be replaced by two viable waste-incorporated mixes 1) Compost-Paper-
Sand (CPS) mix at 7.5%, 7.5% and 85%, respectively, and, 2) Compost-Paper-Sand-
Tires (CPST) mix at 7.5%, 7.5%, 76% and 8% respectively. Bench scale tests
compared the water flow and water filtration capabilities of the CPS and CPST with
the control first in un-vegetated and then in vegetated systems.
Un-vegetated test results showed no significant difference between CPS and control
in terms of infiltration rates, flow attenuation over time (with sediment loading of 6
kg/m2), and water quality tested for nutrients (TKN, NO2+NO3 and TP), pathogens
(total coliforms) and total metals (copper, lead and zinc). Although without
vegetation, the CPST did show less flow attenuation over time indicating a longer life
span, this advantage disappeared in vegetated systems. With vegetation, all three
systems performed similarly in terms of flow attenuation over time (sediment loading
of 13 kg/m2 equal to 5 years in an example PLD). Vegetation is an added benefit as
it increases nutrient removal and the lifespan of the PLD.
An example PLD is used to evaluate the lifespan, cost savings and environmental
impact of installing one of the waste-incorporated mixes instead of the control. With
vegetation, all three systems require replacement due to infiltration rates below the


minimum 2.5 cm/hr in 14 to 16 years. The economic first cost analysis and
environmental LCA of the example site shows that installation of the waste-
incorporated CPS mix saves $2,800 and 6.6 MTC02E.
Vegetated PLD is successful at stormwater treatment including the removal of
sediment (93%-100% removal), nutrients (57%-98% removal), metals (83%-99%
removal) and pathogens (87%-99% removal). Replacing the currently used peat and
sand mixture with the waste-incorporated CPS mix is recommended based on
environmental benefit and performance criteria. In locations which are highly
sensitive to phosphorous, the peat and sand mixture should still be installed due to
higher removal rates (98%). Because tires are buoyant and may overflow the PLD
contaminating downstream waterways, more research and risk assessment is
necessary before using shredded tires.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
Signed _____
James C.Y. Guo
Signed
Anu Ramswami


DEDICATION PAGE
I dedicate this thesis to my parents and my grandparents, who gave me an
appreciation of learning and taught me the value of perseverance and resolve. I also
dedicate this to my husband and my 3 children for their support and understanding
while I was completing this thesis.


ACKNOWLEDGMENT
Thank you to my advisors, James C. Y. Guo and Anu Ramaswami, for their
contribution and support to my research. Their insights and continual guidance was
essential in completing this monumental task. I also wish to thank all the members
of my committee for their valuable participation and insights.
Gratitude goes to the babysitters that allowed me to study, work in the lab and write
my dissertation my parents, my brother Daniel, Lauren, my cousins Jeremy and
Cedar, Viola and others too numerous to mention.
This research has been supported by funding from the Department of Education
GAANN grant, ARCS Scholarship, CCHE Waste-to-Value grant, the Urban Drainage
and Flood Control District and the Urban Water Research Institute. Thank you to
Ken Mackenzie from UDFCD and Ben Urbonas from UWRI who have given their
time and expertise to this research. Additionally, materials and time have been
donated by the AcuGreen, Denver Botanical Gardens and the City and County of
Denver Health Department.


TABLE OF CONTENTS
FIGURES............................ ................xvi
TABLES............................. ................xvii
1 SUSTAINABLE URBAN INFRASTRUCTURE..............l
1.1 SUSTAINABILITY................................1
1.2 URBAN INFRASTRUCTURE..........................1
1.3 STORMWATER MANAGEMENT.........................2
1.3.1 LANDSCAPE AND STORMWATER RUNOFF...............2
1.3.2 BMP FOR COLORADO UDFCD........................3
1.3.3 PREVIOUS STUDIES..............................5
1.4 OBJECTIVES....................................7
2 WASTE SCREENING AND ENVIRONMENTAL BENEFIT.....8
2.1 INTRODUCTION AND OBJECTIVE....................8
2.2 LITERATURE REVIEW ON MATERIALS SUBSTITUTION...9
2.3 METHODOLOGY..................................11
2.3.1 LEACH TEST...................................12
2.3.2 FLOW RATE TEST...............................13
2.3.3 GERMINATION TEST.............................14
viii


2.3.4 ENVIRONMENTAL LCA...........................15
2.4 RESULTS.....................................17
2.4.1 LOCAL AVAILABILITY AND COST.................17
2.4.2 LEACHING CRITERIA...........................19
2.4.3 FLOW RATES..................................22
2.4.4 CONFIRMATION TEST- GERMINATION..............24
2.5 SCREENED BEDDING MATERIAL MIXTURE...........25
2.6 GREENHOUSE GAS BENEFIT......................27
2.7 CONCLUSION..................................30
3 INFILTROMETER DESIGN AND FLOW MODEL.........32
3.1 INTRODUCTION................................32
3.2 SURFACE STORAGE BASIN.......................33
3.3 SUBSURFACE FILTERING SYSTEM.................36
3.4 OPERATION OF POROUS LANDSCAPING BASIN.......39
3.5 LABORATORY TESTS............................40
3.5.1 PREVIOUS DESIGN STUDIES.....................40
3.5.2 INFILTROMETER DESIGN AND TESTING............41
3.6 DESIGN EXAMPLE AND SCHEMATICS...............48
3.7 CONCLUSION..................................52
4 WASTE-INCORPORATED BENCH SCALE TEST- BARE SOIL... 53
IX


4.1 INTRODUCTION AND BACKGROUND...............53
4.2 BENCH SCALE TEST METHODOLOGY..............54
4.3 WATER QUALITY IMPACTS.....................60
4.4 CLOGGING EFFECTS..........................68
4.5 DESIGN EXAMPLE............................73
4.6 CONCLUSION................................76
5 BENCH SCALE TEST WITH VEGETATION..........78
5.1 INTRODUCTION AND BACKGROUND...............78
5.2 BENCH SCALE TEST METHODOLOGY..............79
5.3 VEGETATION................................84
5.4 WATER QUALITY IMPACTS.....................89
5.4.1 NUTRIENT REMOVAL..........................92
5.4.2 METALS REMOVAL............................96
5.4.3 PATHOGEN REMOVAL..........................97
5.4.4 SUSPENDED PARTICLES.......................99
5.5 CLOGGING EFFECTS.........................103
5.6 CONCLUSIONS..............................108
6 CONCLUSIONS AND RECOMMENDATIONS..........110
APPENDIX A.......................................112
REFERENCES ......................................113
x


FIGURES
Figure
1 -1 Picture of PLD Aurora, CO..............................................4
1- 2 PLD Cross Section......................................................4
2- 1 Batch Test Setup for Leaching and Row Rate............................13
2-2 Batch Test Setup and Outflow for Collection of Leachate and Row Rate..13
2-3 Picture of Batch Test Setup for Germination Rate. Grass Seeds Were Planted in
Various Media Mixtures, Watered and Set by a Sunny Window..............15
2-4 Reduction of Copper Leaching From Various Mixes After Multiple Rushes of
Water. The Gray Area Indicated the Permissible Waste Mixture Passing the
Metals Leaching Test...................................................22
2-5 Effect of Varying Amounts of Compost and Paper on Row Rate Shows
Compost Increases Row rate and Paper Slows the Row Rate................23
2-6 Increase in Row Rate is Related to the Amount of Shredded Tires.......24
2-7 Germination Tests Confirmed no Difference in the Ability of the Four Mixes to
Support Vegetation.....................................................25
2-8 Picture of Control and Permissible Waste-Incorporated Media Mixtures for PLD
.......................................................................27
2-9 Net GHG Benefit of Installing the CPST Waste-Incorporated Mix Instead of the
Business as Usual (Control) in an Example PLD..........................29
2- 10 Cost Savings to Installing the CPST Waste-Incorporated Mix Instead of the
Business as Usual (Control) in an Example PLD..........................30
3- 1 Layout of Porous Landscaping Basin....................................33
3-2 Stormwater Quality Control Volume for Porous Landscaping Basin Design...35
xi


3-3 Illustration of Infiltrometer Operation under Saturated Conditions.......37
3-4 Soil Column Design with 2-Layered System and Lowered Outflow for Field
Conditions................................................................42
3-5 Soil Column Design with Elevated Outflow for Saturated Conditions........43
3-6 InfiUrometers Built in Laboratory........................................44
3-7 Measurements for Soil and Water Levels Inside the Soil Column............44
3-8 Geotextile between Large Aggregate and Filter Layer Inside the Column...45
3-9 Manometer Setup to be Placed inside the Column...........................45
3-10 Variation of Infiltration Rates for Sample Columns (Control: Peat and Sand )46
3- 11 Comparison between Observed and Calculated Filtering Thickness.........51
4- 1 Setup of Triplicate Soil Columns and Barrels of Stormwater in the Lab....55
4-2 Variation of Infiltration Rates as the Filtration Mixtures Approach Saturation
after 72 hours of Continuous Flow.........................................56
4-3 Outfall N-43 IE in the South Platte River Where Stormwater for the Experiment
was Collected.............................................................58
4-4 Plot of TKN Concentration in Water for Various Accumulated Sediment Loads
Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66 kg/m2.
The Results Showed No Consistent Statistical Difference Between the Waste-
Incorporated Media Mixes (CPS, CPST) Versus the Control for Leaching or
Filtering of TKN............................................................60
4-5 Plot of Total Phosphorous Concentration in Water for Various Accumulated
Sediment Loads Including Tap Water (No Sediment) and Stormwater with 0.33
and 2.66 kg/m2. The Results Showed No Consistent Statistical Difference
Between the Waste-Incorporated Media Mixes (CPS, CPST) Versus the Control
for Leaching or Filtering of TP.............................................61
Xll


4-6 Plot of Total Copper Concentration in Water for Various Accumulated Sediment
Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66
kg/m2. The Results Showed No Consistent Statistical Difference Between the
Waste-Incorporated Media Mixes (CPS, CPST) Versus the Control for Leaching
or Filtering of Copper...............................................................62
4-7 Plot of Total Coliform Forming Units Water for Various Accumulated Sediment
Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66
kg/m2. The Results Showed that the Control (Peat and Sand) Leached the Most
Coliforms. All Mixtures Showed Consistently High Pathogen Removal Rates 63
4-8 Picture of Floating Particles from the Filtration Mix. The Paper-Tire Deposit
Shows How Density Stratification Occurs When Floating Particles Settle...........65
4-9 Plot of Density of Soil Samples in Depths 0-10 cm of the Soil Column. Results
Show Lighter Material in the Top Layer of the Soil Shows Density Stratification
of Control.......................................................................66
4-10 Plot of Density of Soil Samples in Depths 0-10 cm of the Soil Column. Results
Show Lighter Material in the Top Layer of the Soil Shows Density Stratification
of CPS...........................................................................66
4-11 Plot of Density of Soil Samples in Depths 0-10 cm of the Soil Column. Results
Show Lighter Material in the Top Layer of the Soil Shows Density Stratification
of CPST..........................................................................67
4-12 Plot of Sieve Analysis of Samples at Various Depth (0-10 cm) of Control (Peat
and Sand Mixture). Results Show Small Particles in the Top Soil Layer
Indicating Sediment Filtered out of the Stormwater Accumulates on the Top 1
cm.....................................................................69
4-13 Plot of Sieve Analysis of Samples at Various Depth (0-10 cm) of CPS Results
Show Small Particles in the Top Soil Layer Indicating Sediment Filtered out of
the Stormwater Accumulates on the Top 1 cm.............................70
4-14 Plot of Sieve Analysis of Samples at Various Depth (0-10 cm) of CPS Results
Show Small Particles in the Top Soil Layer Indicating Sediment Filtered out of
the StormWater Accumulates on the Top 1 cm.............................71
xm


4-15 Plot of Reduced Infiltration Rate, fs, Normalized by fc (Hortons Constant
Infiltration Rate) Versus the Accumulative Sediment Load for Various Media
Mixtures. Results Show Reduction of Infiltration Rate with Accumulative
Sediment Load.............................................................73
4- 16 Plot of Reduced Infiltration Rate, fs, Normalized by fc (Hortons Constant
Infiltration Rate) Versus Time for Various Media Mixtures. Results Show
Reduction of Infiltration Rate Over Time for Example PLD..................76
5- 1 Steps for Bench Scale Testing..........................................................80
5-2 Picture of Grow Lights above Columns for Growing Grass in the Lab...........83
5-3 Picture of Un-Germinated Seeds Floating in Water above Soil Surface.........84
5-4 Plot of Plant Counts 25 Days After Planting the seeds in the Cake Layer. Results
Indicate no Difference between the Treatment Groups............................85
5-5 Picture of Healthy Grass Growing in the Column 25 Days after 1st Planting...86
5-6 Picture of Soil Surface after the 1st Planting of Grass was Choked and Died.86
5-7 Picture of Grass in the Control (Peat and Sand) Column after 2 Months without
Water. The Picture Shows the Stunted Growth of Vegetation in Control (Peat
and Sand) After a Dry Period...........................................87
5-8 Picture of Grass in the CPS Column after 2 Months without Water. Healthy
Vegetation in CPS after a Dry Period...................................87
5-9 Picture of Grass in the CPS Column after 2 Months without Water. Healthy
Vegetation in CPS after a Dry Period...................................87
5-10 Average Number of Plants in the Columns with 2 Watering Schemes: Watering
Every Other Day and without Water for 2 Months.........................88
5-11 Plant Height after 2 Months without Water Shows CPS Supports Plant Growth
through a Dry Period...................................................88
5-12 Plot of Removal Rate for Nutrients and Metal with and without Vegetation.
Results Show Increased Percent Removal of Contaminants with Vegetation ....90
xiv


5-13 Plot of TKN Concentration in Water for Various Accumulated Sediment Loads
(0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed
Vegetation does not Effect TKN Removal.......................................93
5-14 Plot of NO3+NO2 Concentration in Water for Various Accumulated Sediment
Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed
Consistently Higher Concentrations in Outflow from the CPS than the Control
Mix..........................................................................94
5-15 Plot of Total Phosphorous Concentration in Water for Various Accumulated
Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results
Showed Vegetation Increased the Removal Rate of TP from the Control Mix ..95
5-16 Plot of Percent Removal of Total Phosphorous for Various Accumulated
Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results
Showed Increased Removal Capacity for TP with Vegetative Conditions.95
5-17 Plot of Total Copper Concentration in Water for Various Accumulated
Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results
Showed Vegetation does not Effect Copper Removal....................96
5-18 Plot of Total Copper Concentration in Water for Various Accumulated
Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results
Showed Breakthrough of Zinc in CPST Mixture.........................97
5-19 Plot of Total Coliform Forming Units in Water for Various Accumulated
Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results
Showed All Mixes Were Continually Successful at Filtering Pathogens from
Stormwater throughout the Experiment................................98
5-20 Picture of Boating Particles from the Filtration Mixture Plugging the Overflow
in the Lab..............................................................100
5-21 Picture of Light Particles which Overflowed a PLD.......................100
5-22 Plot of the Amount of Particles (TSS) from the Various Filtration Mixes Which
Were Found to Boat in Clean Water. Results Indicate as Much as 2,000 mg/1
TSS May Overflow the PLD................................................101
5-23 Boating Particles in Water Samples of Overflow from the Soil Columns................102
xv


5-25 Particles on the Paper of Filtered Samples from Overflow after Grass was
Growing. The Samples Show that Peat, Compost and Tires Particles Float and
May Overflow the PLD....................................................................103
5-26 Plot of Infiltration Capacity of Various Media Mixtures Before and After
Vegetation is Growing in the Cake Layer. Regeneration of Infiltration Capacity
from Growth of Vegetation Shows the Control and CPS Benefited more than
CPST..............................................................................104
5-27 Plot of Saturated Infiltration Capacity of Various Media Mixtures Before and
After Vegetation is Growing in the Cake Layer. Results Show Increase in
Saturated Infiltration Capacity after Grass was Germinated into the Cake Layer
........................................................................................105
5-28 Plot of Reduced Infiltration Rate, fs, Normalized by fc (Hortons Constant
Infiltration Rate) Versus Accumulative Sediment Load for Various Media
Mixtures, with and without Vegetation. Results Show Regeneration of Flow
Rate in Control and CPS Mixes after Grass is Germinated in the Cake Layer .106
5-29 Plot of Reduced Infiltration Rate, fs, Normalized by fc (Hortons Constant
Infiltration Rate) Versus Time for Various Media Mixtures, with and without
Vegetation. Results Show Vegetation's Increases the Time to Clogging in the
Control and the CPS Mixes...............................................................107
xvi


TABLES
Table
1 -1 Existing PLD research.....................................................6
2-1 Summary of literature on PLD waste stream replacements...................11
2-2 Possible local waste streams to replace virgin materials for incorporation in PLD
..........................................................................12
2-3 PLD vegetation seed mixture prescribed in UDFCD criteria manual..........14
2-4 Number of seeds planted in each batch test...............................15
2-5 Data sources to calculate GHG emissions impact...........................17
2-6 Cost comparison of virgin and waste materials............................19
2-7 Leachate pH from control and waste incorporated media mixtures after 48 hours
contact time..............................................................19
2-8 Leachate pH from crushed concrete........................................20
2-9 Leachate pH from media mixtures on top of crushed concrete...............20
2-10 Leaching results of various media mixtures indicate that three mixtures pass the
leaching test.............................................................21
2-11 Permissible ranges of waste-incorporated filtration mixes.................26
2- 12 GHG benefit in MTC02E of replacing business-as-usual (control) with a
waste incorporated CPST mix...............................................28
3- 1 Large column studies of systems similar to PLD...........................41
3-2 Compaction and density of soil-mix sample................................46
3-3 Variation of hydraulic heads measured at 5 stations......................47
xvii


3-4 Hydraulic conductivity in filtration layer after 72 hours of water flow....47
3- 5 Observed case Studies for bio-retention medium (Li and Davis 2008a and
2008b)......................................................................51
4- 1 Average water quality results of stormwater outfall N-43 IE................58
4-2 Average event mean concentrations for Denver area..........................59
4-3 Sediment loading to the soil column surface during the experiment..........64
4-4 Table of density of soil sample samples at various depth (0-10 cm). Results
show lighter material settled on the top soil layer resulting in density
stratification..............................................................67
4- 5 Table of percent of particles passing 75 um sieve for samples of the filtration
layer at various depths (0-10 cm)...........................................68
5- 1 PLD vegetation seed mixture as prescribed in UDFCD criteria manual.........82
5-2 Amount and type of grass seeds planted in each column......................82
5-3 Germination slows after 20 days as Sshown by the average number of plants 20
and 25 days after planting..................................................84
5-4 Table of contaminant removal rate comparing un-vegetated and vegetative
conditions. Results show vegetation increases the nutrient removal rate.....90
5-5 Contaminant concentrations in inflow and outflow water through various media
mixtures with and without vegetation compared to the EPA freshwater criteria
............................................................................92
5-6 Percent removal of total coliform forming units without and with vegetation.
The results show high percent removal of pathogens from stormwater..........98
5-7 Accumulative sediment loading to the soil surface during the experiments with
and without vegetation......................................................99
xviii


1 SUSTAINABLE URBAN INFRASTRUCTURE
1.1 SUSTAINABILITY
In a 1987 report to the United Nations, Gro Brundland defined sustainable
development as development that meets the needs of the present without
compromising the ability of the future generations to meet their own needs (WCED
1987). Sustainable urban infrastructure systems therefore are those that provide for
the needs of urban residents in both the short and long term while also maintaining
the environmental, social and economic resources. The combination of these
resource parameters to represent sustainability was coined the Triple Bottom Line for
businesses by John Elkington (Elkington 1998). At the 2002 World Summit on
Sustainable Development in Johannesburg, South Africa, these same parameters were
designated as People, Prosperity and the Planet (Gadepalle et al. 2007; UNEP 2002).
Sustainable development is achieved by the use of green engineering principles.
These principles help to enhance long-term sustainable infrastructure through design
and use of products that minimize adverse impact on human health and the
environment. The Green Engineering Principles (Anastas and Zimmerman 2003,
EPA 2010) include holistic design, conservation of natural resources, life-cycle
thinking, use of inherently safe and benign materials, minimizing depletion and
preventing waste, and implementation of innovative and culturally sensitive solutions.
Waste symbiosis from local industrial byproducts, for example, is an innovative reuse
and waste minimization technique that provides for sustainable development
1.2 URBAN INFRASTRUCTURE
In 2009, urban areas served as home and employment centers to more than half of the
worlds population (UN 2007). By 2030 that number is expected to rise to 60%
(UNEP 2007). In the U.S. urban areas as defined by the Census Bureau include areas
with a population over 50,000 and urban clusters include 2,500 to 49,999 people (US
Census Bureau 2000). In the US, over 79% of people live in urban areas (US Census
1


Bureau 2000). Locally, of the 4.6 million people in Colorado, 3.7 million live in
urban areas (USDA 2007). Population projections indicate a population increase of
65% in Colorado from 2000 to 2030. Of the expected 6.2 million people in Colorado
in 2030, 5.3 million (85%) will live along the urbanized front range.
Urbanized areas are characterized by population density and intensity of development
and infrastructure. Urban infrastructure refers to the engineered systems that provide
services of water, wastewater, energy and transport (of human goods and
information) within an urban area (Ramaswami 2005). Sustainable development can
be measured with performance, environmental, social equity and economic metrics.
The focus of this thesis is to explore holistic and sustainable design of stormwater
detention basins in Denver, Colorado and the arid Western region, addressing
performance and environmental aspects of sustainable design.
1.3 STORMWATER MANAGEMENT
1.3.1 LANDSCAPE AND STORMWATER RUNOFF
The built environment will expand to support increasing urban populations. To
highlight the accelerated rate of land use, it is estimated that in 2030 more than half of
all the built environment in the U.S. will have been constructed since 2000 (Nelson
2004). New development changes the natural hydrologic character of the landscape.
Large impervious areas, such as parking lots and roadways, increase the stormwater
runoff volumes, reduce groundwater recharge and increase pollutant loads.
Additionally, in an urban environment, the disturbance of land by construction can
trigger the leaching of soil-bound nutrients (Smith 2005).
In fact, the EPA indicates that runoff from urban areas is the leading source of
impairments to surveyed estuaries and the third largest source of water quality
impairments to surveyed lakes (EPA 2004b). Urban drainage in turn affects the
water supply sustainability. For example, degraded water quality has forced five
communities in Colorado to utilize aggressive reverse osmosis (RO) systems to treat
drinking water (American Membrane Technology Association 2006). On the other
hand, the city of New York was able to save millions of dollars through focusing their
2


efforts on source water protection to avoid increased treatment capacity. (Appleton
2002).
Quantity of Runoff:
Dealing with stormwater is an essential function of urban infrastructure. Historically,
the focus of stormwater management has been on public safety, acting as a flood
control measure for a given hydrological area. Traditional stormwater conveyance
and collection systems (concrete channels) have been a common approach to dealing
with large volumes of runoff water from impervious areas. However, purely
structural catchments areas such as constructed basins and detention tanks are
successful to reduce the peak flow and runoff volume at the design or historic level,
but provide little to no pollutant removal for frequent events (Federal Highway
Administration 2006). Urban runoff solutions must include both quantity and quality
controls, i.e. improving the on-site recharge to the local groundwater and reducing the
pollutant loads at the source.
Quality of Runoff:
Since 1990, the focus of stormwater management has shifted to include water quality
concerns. By adapting and maximizing the design of detention basins to enhance the
treatment and infiltration of runoff, stormwater quality can be enhanced. A study by
CH2M Hill found centralized conventional drainage design was unable to match the
performance of the natural watershed (CH2M Hill 2001). Low impact development
(LID) techniques take advantage of micro-scale approaches to mold the development
of land to function similar to natural drainage systems, thus replicating ecosystem
services which the open space would have performed (Sample and Heaney 2006).
1.3.2 BMP FOR COLORADO UDFCD
Monitoring urban rainfall runoff and protecting water quality can be addressed
through site-sensitive Best Management Practices (BMPs) which take advantage of
structural basins and low impact development techniques. BMPs are site sensatie
solutions based on location. The Urban Drainage Flood Control District (UDFCD)
has been charged with assisting local governments in the Denver metropolitan area
with multi-jurisdictional drainage and flood control issues. The UDFCD has
developed design criteria for stormwater management for the semi-arid climate which
3


includes comprehensive list of installation and maintenance instructions for BMPs
specific to Denver metro area.
Figure 1-1 Picture of PLD Aurora, CO
In particular the porous landscape detention (PLD) is a design which the UDFCD
promotes for small installations reducing the impact on developable land. The PLD is
a constructed sedimentation facility intended to capture and filter stormwater for
micro rainfall events taking advantage of the intrinsic quality of plants to act as water
treatment systems (Guo 2007). The current design consists of a vegetated zone on
top of a filtration mix underlain with large aggregate and drains. This design shown
in Figure 1 is specified in Drainage Criteria Manual, Volume 3, Chapter 5.6
(USWDCM 2001).
Figure 1-2 PLD Cross Section
4


The UDFCD criteria manual notes one specific disadvantage of the PLD is the
potential for clogging. Additionally, the specific water quality benefits of the PLD in
the Denver area are yet to be defined (USWDCM 2001).
BMP design is still evolving and new research leads to upgrades in current
recommendations. The current design recommendations leave opportunity for the
incorporation of waste symbiosis and holistic design concepts. The current media
mix design consists of peat, sand and gravel. Peat, which is imported, is very
expensive ($130 per cubic yard at Paulino Gardens in 2009) and has associated
environmental impacts of transportation and of peat mining. Local waste streams
offer an opportunity for replacement of portions of the media. For example options
for peat replacement include, compost, shredded paper and other organic waste
stream materials (Tucker 2007). In addition the sub-layer may utilize waste stream
materials such as recycled aggregate (McCambridge et al. 2004) and shredded tires
(JaiTire 2008). However, as shown in Table 1-1 no previous study has evaluated the
impact of waste materials on the performance and sustainability aspects of a PLD.
1.3.3 PREVIOUS STUDIES
The PLD promoted by the UDFCD for the Denver Metro area is a vegetated detention
basin which acts to treat stormwater. The PLD is also referred to as bioretention,
biofiltration, rain garden and vegetated infiltration basins in other parts of the United
States. PLD, or bioretention, was originally developed in Prince Georges County,
Maryland and has been used as a stormwater BMP since 1992 (Prince George's
County Department of Environmental Resources (PGDER) 1993; USEPA 1999).
PLDs have been proven to enhance stormwater by reducing peak runoff volumes, and
by providing filtration of sediments and sorption of pollutants. (Davis et al. 2001;
Davis et al. 2006; Davis et al. 2003; Hunt 2003; Hunt et al. 2006). The UDFCD
adopted the recommendations as presented by the Prince Georges County, Maryland
with slight adjustments based on the hydrologic and geologic conditions common to
Colorado (Guo and Urbonas 1996a; 1999). Table 1-1 presents studies completed to
date of low impact development techniques similar to PLDs including bioretention
areas and vegetated infiltration basins.
5


Table 1-1 Existing PLD research
Location Author (year) Scale Technique Performance Sustainability
Physical- Infiltration Chemical * Biological Waste Reuse Life Cycle
CA (Caltrans 2004) field Various BMPs to retrofit sites Visual observation N, M TSS, fecal coliform X X Life Cycle Cost
PA (Hunt 2003) lab+ field bioretention mulch top layer and anoxic zone water balance N (lab)+ M TSS, BOD (field) X- Planted but no data X Sorption of metals
NC (Sharkey 2006) lab + field bioretention with anaerobic zone X N X X X
MD (Kim et al. 2003a) lab bioretention- anoxic zone and various soil mixes X N, turbidity samples X Newspaper, woodchips, sawdust, alfalfa X
(Sun, 2004) lab bioretention uptake of metals by grasses water balance M grasses evaluated for metals uptake X X
(Davis 2007) field bioretention inflow and outflow N, M, TSS X planted but no data X X
(Hsieh, 2004; Hsieh and Davis 2005) lab+ field bioretention with media mixes outflow N, M, TSS, O/G, X mulch X
(Davis et al. 2001) lab bioretention flow rate M X planted but no data mulch Sorption of metals
(Davis et al. 2006) lab+ field bioretention flow rate N X planted but no data X X
(Prince Georges County 2000) field LID techniques yes N X X X
WA (CH2M Hill 2001) field LID techniques flow rate X X X Life Cycle Cost
(Ames et al. 2001) lab retention basins with media mixes inflow M.TSS, TPH X X X
AU (Siriwardene 2007) lab sediment transport in gravel filter outflow TSS X X lifetime for clogging
NY (Appleton 2002) field various LID X X X X System- wide Economic savings
CO (Guo and Urbonas 1996b; 2002) field Various LID WQCV X X X X
(Guo and Hughes 2001) field Infiltration bed WQCV X X X X
Note: X not included in the study; M- metals, N- nutrients, TSS- total suspended solids, BOD-
biological oxygen demand, O/G oil and gas, TPH- total petroleum hydrocarbons
WQVC water quality control volume
6


Since the 1990s the majority of the research on PLDs has been completed at the
University of Maryland under Dr. Allen Davis (Davis 2007; Davis et al. 2001; Davis
et al. 2006; Davis et al. 2003; 2004; Hsieh and Davis 2005a; 2005b; Hsieh et al. 2007;
Kim et al. 2003; Sun and Davis 2007). No previous study has taken a combined look
at performance criteria and environmental benefits of a waste incorporated enhanced
PLD. Additionally, past studies have been focused on coastal areas with different
weather patterns and hydrologic conditions, rather than that of the arid west.
1.4 OBJECTIVES
The objectives of this thesis are:
1) to select the best waste material reuse for sustainable PLD sub-base system
design (Chapters 1 and 2)
2) to quantify the life-cycle environmental benefits of waste reuse for stormwater
PLD. (Chapter 2)
3) to investigate the operation of a 2-layered PLD (Chapter 1 and 3)
4) to investigate the impacts of waste materials and vegetation on performance of
the PLD addressing (Chapter 4 and 5)
2a) infiltration capacity for on-site stormwater volume disposal, and
2b) effectiveness of contaminant removal for stormwater quality
enhancement.
The objectives were completed through the following phases of work:
Literature review and method development (Chapters 1-5)
Waste material screening (Chapter 2)
Environmental life cycle analysis (Chapter 2)
Develop and test soil column design at the bench scale (Chapter 3)
Model a 2-layered PLD flow from Bench scale test (Chapter 3)
Bench scale test with waste materials and bare soil conditions (Chapter 4)
Bench scale test with vegetation (Chapter 5)
7


2 WASTE MATERIALS SCREENING AND ENVIRONMENTAL BENEFIT
This chapter expands on the paper entitled Waste Incorporated Sustainable Design
of Stormwater Detention Basins 1) Waste Screening Tests by Shauna M Kocman,
Anu Ramaswami, James C. Y. Guo. The paper was submitted to the ASCE Journal
of Environmental Engineering in April 2010.
2.1 INTRODUCTION AND OBJECTIVE
Stormwater management is a site-sensitive solution requiring optimal design
techniques for the geographic region and locally available construction materials.
Consequently the associated economic and environmental benefits are closely tied to
location also. The stormwater infrastructure chosen for this study, the porous
landscape detention basin (PLD), is currently recommended for use in the Denver
area (USWDCM 2001). The current design recommendations for materials leave
opportunity for the incorporation of waste symbiosis and sustainable design concepts.
The upper filtering layer is composed of a peat and sand mixture. Peat, which is
imported, is very expensive and has associated environmental impacts of
transportation and of peat mining (Cleary et al. 2005). Both large and small (sand)
aggregates are virgin materials and carry associated impacts (Reiner 2007). To date
there has been no comprehensive waste screening study to evaluate the use of various
waste stream materials into PLDs and the resulting hydrologic performance and life
cycle environmental impact. Local environmental benefits include reduction in peak
runoff flows, improvements in water quality and reduction in pathogen counts. Life
cycle environmental benefits include waste reuse and greenhouse gas benefits over
the life cycle of a PLD, from materials extraction to installation.
To the best of our knowledge only one study has investigated the use of waste
materials as replacements in PLD. In 2003 Kim, Seagren and Davis investigated the
use of various organic materials (paper, alfalfa, etc) for use as electron donor
substrate for denitrification in the lower layers of a modified bioretention system but
did not evaluate the impact on infiltration rates or environmental benefit (Kim et al.
2003). Other studies which incorporate leaf or mulch compost are not specifically
focused on waste reuse and do not include compost created from a mix of organic
8


waste materials. These studies have focused on the optimization of mixtures of
native soil, sand, mulch and/or leaf compost mixtures based on measurements of
infiltration rates and/or pollutant reduction (Ames et al 2001, Thompson 2008, Hsieh
and Davis 2005, Bratieres 2008). The performance of field installations of
bioretention areas with native soil, sand, and compost has been monitored and is
reported in the literature,(Ames et al. 2001; Davis 2007; Hsieh 2004; Hunt et al.
2006). None of the above studies have evaluated the life cycle environmental benefit
of offsetting virgin materials by incorporating waste reuse.
The objective of this chapter is to find suitable waste materials and evaluate the
environmental benefit (local and life cycle) of incorporating those materials. Possible
waste materials through literature review are inventoried. Those materials were
screened through performance tests (local availability, cost, leaching and infiltration
rates) and confirmed by germination rate testing. Permissible waste-incorporated
mixtures are defined and the environmental benefit is evaluated using greenhouse gas
savings.
2.2 LITERATURE REVIEW ON MATERIALS SUBSTITUTION
The Denver region was the focus of this study so the PLD design, materials, testing
procedures and environmental benefit are based on this location. A list of possible
materials was created based on literature review and exploration of local materials.
The procedure for screening required that materials pass three criteria: 1) local
availability and cost, 2) leaching and 3) infiltration rate. A last confirmation test on
germination rates ensured the passing mixture(s) would support vegetative growth.
The first step was to create a list of possible waste materials. While few studies have
focused on incorporation of more than leaf/ mulch compost in PLDs, guidance can be
gained from focusing on landscaping studies that have previously documented use of
waste materials. An initial literature review shows aggregate replacement options in
landscaping, include recycled aggregate, crushed glass, and shredded rubber tires
(Moller and Leger 1998; Tang et al 2007). Peat replacement options include organics
used as soil amendments such as compost, recycled paper, sawdust, and spent hops
(Burgos et al. 2006; Castaldi et al. 2005; Davis and Wilson 2007Glenn et al. 2002;
Grimes and Cooper 1999; Herwijnen et al. 2007; Kim et al. 2003; Motavalli and
Discekici 2000; Molphy et al. 2001, Tucker 2007,). Further investigation into
organics revealed a local source of compost utilizes various waste streams including
those considered here, spent hops, saw dust, wood chips and biosolids (Yost 2008).
Therefore this compost was considered as an avenue to capture more than one waste
9


stream. Because sources of materials for compost production vary by time and
location, Class A certified compost was chosen to ensure quality (Yost 2008).
Some studies have shown incorporating waste stream materials may have unintended
negative water quality impacts. For example, the leaching nutrients from compost are
a concern (Burgos et al. 2006; Castaldi et al. 2005; Grimes and Cooper 1999;
Herwijnen et al. 2007). Paper has shown to initially deplete garden soils of nitrogen
so may possible offset the nutrients in compost (Glenn et al. 2002). Table 2-1
presents the potential waste stream materials for replacement based on literature
review of landscaping applications. Additional information as whether data are
available for leaching, life cycle analysis have been conducted and if the material is
locally available in the Denver area is also included in Table 2-1.
10


Table 2-1 Summary of literature on PLD waste stream replacements
Waste Stream Author and Year Other Applications Data Available for Leaching? LCA of GHG? Local Availability
Aggregate
Recycled crushed glass (Clean Washington Center 1996) Landscaping as groundcover or mulch no no no
(Moller and Leger 1998) Topsoil mix yes no
Shredded Rubber (EPA 2009, Tang et al. 2007) Mulch, water filtration no no yes
Recycled aggregate (McCambridge et al. 2004) Landscape material and road base no no yes
Substrate Media
Compost (Burgos et al. 2006; Castaldi et al. 2005; Grimes and Cooper 1999; Herwijnen et al. 2007) Peat replacement yes no yes
Office paper (Motavalli and Discekici 2000) Soil topdressing yes no yes
Recycled Newspaper (Glenn et al. 2002; Kim et al. 2003; Molphy et al. 2001) Peat replacement yes no yes
Spent hops (Tucker 2007) Soil amendment and compost no no no
Pine needles (Pote and Daniel 2008) Groundcover no no no
Sawdust (Davis and Wilson 2007) Soil amendment And compost no no no
2.3 METHODOLOGY
Local landscape suppliers were contacted for availability and cost of possible
materials. Landscape supply companies in Denver do not carry pine needles and do
not recommend use of pine needles as a soil amendment due to their acidic nature.
Availability of crushed glass in many locations is unreliable and transporting it is cost
prohibitive in places where sand is readily available. The compost incorporated many
waste stream organics, such as spent hops, saw dust, and wastewater residuals. Spent
11


hops and sawdust were not readily available and were incorporated in the compost, so
were not investigated separately. As a result, local waste materials choices were
crushed recycled concrete, shredded paper, compost and shredded tires presented in
Table 2-2.
Table 2-2 Possible local waste streams to replace virgin materials for incorporation
in PLD
Waste Stream Replacement Local Source
Crushed Rubber Aggregate JaiTire Industries, AcuGreen
Recycled concrete Aggregate Oxford Recycling Center
Compost Peat A1 Organics
Office paper Peat Waste Management
Recycled Newspaper Peat Denver Post newsprint and white waste Waste Management
2.3.1 LEACH TEST
The narrowed list of possible materials was subjected to leach testing. The testing
began with replacement of one virgin material with one waste material e.g., replacing
peat with compost only. Subsequent experiments included combining various waste
materials e.g., compost and paper mixture for peat replacement. Water quality
samples were collected from 4 liter batches of the mixtures by adding deionized water
to the top of each batch and collecting the outflow. Samples were analyzed for pH,
nutrients and metals. A sension 1 pH meter was used to read the pH of the
samples. Samples were sent to Metro Wastewater Reclamation District for analysis
of total keldjal nitrogen (TKN), nitrate plus nitrate (NO2+NO3), total phosphorous
(TP), and total metals. Figures 2-1 and 2-2 are photographs of the experimental set-
up for batch testing.
12


Figure 2-1 Batch Test Setup for Leaching and Flow Rate
Figure 2-2 Batch Test Setup and Outflow for Collection of Leachate and Flow Rate
The minimum acceptable infiltration rate for a PLD is 2.5 centimeters per hour
(cm/hr) (1 inch per hour (in/hr)) and the minimum filtration depth is 45.7 cm (18
inches). Therefore the maximum contact time for stormwater filtering in a PLD is
expected to be 18 hours just before failure. An exceedingly conservative 48 hour
contact time was chosen the first leach test. A still conservative contact time of 18
hours was chosen for subsequent tests. Reduction in leaching was investigated
through samples taken at the first and tenth flushes. Samples of leachate were
analyzed at Metro Wastewater Reclamation Districts certified laboratory for
nutrients and metals. The results of material mixtures were compared to the EPA
freshwater criteria and control, peat and sand leachate results (EPA 2006).
2.3.2 FLOW RATE TEST
The same 4-liter batches of materials were then tested for flow rates. Water was
added to the top and the outflow was measured with known volume and stopwatch.
Flow rates were measured and compared to the control (peat and sand). Mixtures
13


within a range of 80% to 120% of the control flow rate qualified for confirmation
germination tests.
2.3.3 GERMINATION TEST
After the completion of the flow rate and leaching tests, the new batches of the
passing mixtures were created for germination tests. The germination tests were
conducted with a mixture of native grass seeds in small batches of approximately 200
square centimeters of bedding material. Germination and growth rates were
compared to the control (peat and sand). Grass seeds were counted in relative
quantities recommended in Volume 3 Criteria Manual for the Denver Metropolitan
area and mixed into the top 1/8 inch of media mix (USWDCM 2001). The
recommended seed mixture is shown in Table 2-3. The surface area of the batch test
would not allow for all 14 species of plants therefore only the available grasses were
used. Table 2-4 presents the relative number of grass seeds were spread on the
filtration media mixes and watered.
Table 2-3 PLD vegetation seed mixture prescribed in UDFCD criteria manual
COMMON NAME SCIENTIFIC NAME VARIETY PLSLbs per Acre Ounces per Acre
Sand bluestem Andropogon hallii Garden 3.5
Sideoats grama Bouteloua curtipendula Butte 3
Prairie sandreed Calamovilfa longifolia Goshen 3
Indian ricegrass Oryzopsis hymenoides Patoma 3
Switchgrass Panicum virgatum Blackwell 4
Western Wheatgrass Pascopyrum smithii Ariba 3
Little bluestem Schizachyrium scoparium Patura 3
Alkali sacaton Sporobolus airoides 3
Sand dropseed Sporobolus cryptandrus 3
* Pasture Sage Artemisia frigida 2
* Blue aster Aster laevis 4
* Blanket flower Gaillardia aristata 8
* Prairie coneflower Ratibida column ifera 4
' Purple prairiedover Dalea (Petalostemum) purpurea 4
Sub-Totals: 27.5 22
Total lbs per acre: 28.9
Source (USWDCM 2001)
14


Table 2-4 Number of seeds planted in each batch test
Seed Name Amount (lbs/acre) seeds/lb Batch size (sq ft) seeds/ Batch
sideoats gramma 3 191,000 0.5 7
prarie sandreed 3 274,000 0.5 9
swtichgrass 4 270,000 0.5 12
western wheatgrass 3 110,000 0.5 4
little bluestem 3 260,000 0.5 9
sand dropseed 3 825,000 0.5 28
The containers in Figure 2-3, were set near a sunny window, watered and monitored
daily for germination and growth. The number of plants germinated was tracked and
reported on day 8, 10, 12 and 17. Germination rates did not increase between day 12
and 17 as shown. Therefore the day 17 plant counts were used to compare the
mixtures ability so support seed germination and plant growth.
Figure 2-3 Picture of Batch Test Setup for Germination Rate. Grass Seeds Were
Planted in Various Media Mixtures, Watered and Set by a Sunny
Window
2.3.4 ENVIRONMENTAL LCA
The final result testing lead to a multi-criteria permissible range of mixtures for
incorporation in PLDs. Two sample mixtures in the mid-range of the permissible
amounts were chosen to evaluate the environmental impact based on a design
example site. The environmental benefit of incorporating the waste materials can be
measured in greenhouse gas (GHG) emissions. The materials included in this study
replaced virgin (sometimes foreign) materials with use of local waste materials. The
impact of the virgin materials and the equivalent replacement are evaluated by offset
of virgin material and GHG emissions.
15


The EPA Waste Reduction Model (WaRM) and published GHG emissions specific to
peat mining and transportation (EPA 2008, Cleary et al. 2005) and aggregate (Reiner,
2007) were combined to calculate the GHG savings. The boundaries considered for
impact from materials were extraction to installation and included transportation. A
design example was created to calculate the net benefit of a waste-incorporated PLD
installed at a field test site.
Peat, which is imported, has associated impacts of land use, fossil fuel combustion,
and peat decomposition (ADAS UK Ltd and Enviros Consulting Ltd 2005). Peat
bogs are drained before mining which increases decomposition and release of carbon
dioxide and methane gas. The peat is then extracted, processed and transported long
distances. Once the peat is used in soil applications it decomposes further releasing
more GHGs. One life cycle analysis completed on Canadian Peat in 2005 estimated
an emission factor of approximately 0.05 tons of carbon dioxide equivalents per ton
of peat extracted in 2000 (Cleary et al. 2005).
Transportation for aggregates, which are locally mined, is shorter than peat but
scarcity in the Denver Metro area is a concern. The primary source of alluvial coarse
aggregates in Denver is the South Platte River corridor. As sites near Denver are
depleted, it is expected that the distance traveled for coarse aggregate will soon
increase (Reiner 2007). As the travel distance to obtain virgin aggregate materials is
expected to increase, energy use and cost will increase as well. Additionally, reuse of
local materials minimizes the impact of disposal in a landfill.
The waste materials and the replaced virgin materials (peat and sand) are presented in
Table 2-5. Use of waste materials results in GHG savings (negative number) while
use of virgin materials results in emissions of GHG (positive number). Offsetting one
virgin material results in both the savings from the use of the waste material plus the
savings of offsetting the virgin material. For example a ton of peat (+.84 MTC02E)
replaced by a ton of compost (-.2 MTC02E) results in a net benefit of -1.04
MTC02E.
16


Table 2-5 Data sources to calculate GHG emissions impact
Material MTC02E impact per ton material Model Includes Literature Source
Compost from mixed organics -0.2 20 miles transport, sequestered carbon and avoided landfill (EPA 2008)
Paper -0.05 20 miles transport, avoided landfill (EPA 2008)
Rubber -0.04 20 miles transport, avoided landfill (EPA 2008)
Aggregate +.006 40 miles transport, processing and consumption (Reiner, 2007)
Peat Moss +.84 2000 miles transport, processing, carbon release from mining (Cleary et al. 2005; EPA 2002; EPA 2004a)
Note: MTC02E metric ton carbon dioxide equivalents
See Appendix A for more detail about the GHG emissions from peat
The total peat produced in 2000 was 1.3 million tons with a total GHG emission,
excluding transport, of 0.612 MTC02E (Cleary et al. 2005). Transportation was
added based on EPA fuel economy standards of 5 miles per gallon for trucks. The
EPA reported GHG emission factor for diesel fuel is 22.23 pounds per gallon (EPA
2002; EPA 2004). The average truckload of peat is 17.4 tons (Cleary et al. 2005). A
total of .612 MTC02E per ton of peat plus transport from Canada to Denver (2,000
miles) creates .24 MTC02E per ton of peat is the total GHG cost of peat (EPA 2002;
EPA 2004).
2.4 RESULTS
2.4.1 LOCAL AVAILABILITY AND COST
The waste material screening criteria of availability/cost, leaching and flow rate, lead
to a list of unsuitable materials and a range of optimal mixtures for the suitable
materials. This study was designed for a location-sensitive solution for the Denver
metro area, representing an urban area in the arid west. Following literature review,
local suppliers were contracted for availability and cost. The possible materials were
17


narrowed to compost from mixed stream of organics, paper, shredded tires and
crushed concrete. The cost comparison for the chosen materials, compost, shredded
paper, shredded tires, and crushed concrete, is presented in Table 2-3. All the waste
materials are less than or equal to the cost of the virgin materials and therefore pass
the cost screening criteria.
Sand and aggregate are currently recommended in the design mixture and were
obtained from Pioneer Sand (Golden, Colorado) and Sante Fe Sand and Gravel
(Englewood, Colorado).The replacements considered for aggregates included
recycled glass, shredded tires and crushed concrete. Availability of crushed glass in
many locations is unreliable and transporting it is cost prohibitive in locations where
sand is readily available. Locally, Coors Bottling collects the majority of available
glass and melts it into new beer bottles. Some fines were screened off and donated
from Coors Bottling. A source of surplus crushed glass is not readily available in the
Denver Metro area. The City of Durango collects and sells glass cullet and the
grinder was broken at the time of this study. Additionally the transportation to the
Metro area is approximately 240 miles and would be expensive. Therefore, the glass
cullet currently fails the availability and cost category for use in the Metro area.
Shredded tires and crushed concrete are readily available. Acugreen (Denver,
Colorado) donated shredded tires and provided a tour of the facility. Crushed concrete
is recycled from many demolition projects and was donated from Oxford Recycling
(Englewood, Colorado).
The compost was obtained from A1 Organics and incorporated many waste stream
organics, such as spent hops, saw dust and wastewater residuals. Spent hops and
sawdust were not reliably available and were incorporated in the compost, so were
not investigated separately. Shredded waste paper is available from Waste
Management. For the batch test waste paper was shredded from the newspaper and
some office paper. Landscape supply locations do not carry pine needles and do not
recommend use of pine needles as a soil amendment due to their acidic nature. The
cost comparison for the remaining materials, compost, shredded paper, shredded tires,
crushed concrete, is presented in Table 2-6. All the waste materials are less than or
equal to the cost of the virgin materials and therefore pass the cost screening criteria.
18


Table 2-6 Cost comparison of virgin and waste materials
Virgin Material (cost in $/cy) Replacement Material Density (lbs/cy material) Cost ($/cy material) Comparison ($/cy replacement/ $/cy virgin)
Peat ($170/cy) A1 Compost 1,030 lbs/cy $35 .27
Peat ($170/cy) Shredded paper 39 lbs/cy $10 .08
Sand ($ 17/cy) Shredded Tires 2,000 lbs/cy $17 1
Va aggregate ($ 17/cy) Crushed concrete 2,900 lbs/cy $11 .44
Note: lbs/cy = pounds per cubic yard = 0.59 kg/m3
2.4.2 LEACHING CRITERIA
After screening for availability and cost, the waste materials were subjected to leach
tests in order to ensure the protection of water quality. Leaching of nutrients and
metals was found to be dependent on both the amount of the waste material in the
mixture and contact time with water. Batch Tests for pH indicated that the organic
material slightly acidifies the water (pH 6.09 -6.23 while crushed concrete drastically
increases the pH (11-12). The results of the pH testing for are presented in Table 2-7.
Table 2-7 Leachate pH from control and waste incorporated media mixtures after 48
hours contact time
Material pH after 48 hours
Peat and sand 6.23
Compost and sand 6.09
Paper and sand 6.12
Tires 6.78
Concrete 12.17
Concrete aggregate was quickly eliminated as the pH is very alkaline (11- 12 standard
units) even after many flushes of water. Table 2-8 presents the results of the pH
batch test with 100% crushed concrete.
19


Table 2-8 Leachate pH from crushed concrete
Flush number Contact Time pH
1 48 hour 12.61
2 48 hour 11.03
4 48 hour 12.29
7 48 hour 12.43
8 continuous flow 11.13
10 continuous flow 11.53
Additional batch tests were constructed with media layers representing the filtration
layer in the field installation to test the effect of recycled concrete below various soil
mixes. The recommendation for layers in Volume 3 Criteria Manual is 30 cm (12
inches) allowance for ponding water, 46 cm (18 inches) of soil media and 20 cm (8
inches) of aggregate under drain. In other words the water allowance is
approximately 30% of the depth, soil media is 50% and the aggregate is 20%.
Therefore the batch tests were constructed with 9 cm (4 in) of water, 15 cm (5 Vi in)
of soil media and 6 cm (2 Vi in) of aggregate for a total of 30 cm (12 in) depth. Three
flushes (1, 2, 3) with a 24 contact time were measured. Deionized water was added
the top of the soil media, left for 24 hours, drained and pH was measured. Three
additional flushes (4, 5, 6) were continuous flow. The results of the testing are
presented in Table 2-6.
Table 2-9 Leachate pH from media mixtures on top of crushed concrete
Soil Media on top of Crushed Concrete PH Flush Number (24 hr contact) PH Flush Number (continuous flow)
1 2 3 4 5 6
15% Compost and 85%Sand 11.44 11.68 11.59 11.83 10.57 10.74
15% Peat and 85% Sand 12.06 11.90 11.94 11.55 11.61 10.70
16% Compost, 42% Tires, 85% Sand 11.53 11.73 11.47 10.83 10.50 10.47
Pure shredded tires leach high concentrations of copper (68 ug/L) and zinc (13,500
ug/L) during 48 hours of contact time. Less contact time and a mixture of 50% crumb
rubber and 50% sand reduces the concentrations significantly (23 ug/L copper and
less than 20 ug/L zinc). After 10 subsequent flushes of water, both copper and zinc
are reduced to non-detect levels (<20 ug/L). Similarly, higher concentrations of
20


compost and longer contact times increase the amount of nutrients and metals in the
leachate. With a mixture of 15% compost and 85% sand (complete replacement of
peat) the leachate contained 54 mg/L TKN, 31 mg/L TP and high values of copper
(307 ug/L), lead (66 ug/L) and zinc (489 ug/L) after 48 hours of contact. Whereas
after 18 hours of contact time and only 5% compost the nutrients ( 6.5 mg/L TKN and
.41 mg/L TP) and metals (20 ug/L copper and lead and zinc both non-detect) were
considerably lower. After ten flushes of water the nutrients and metals in all cases
had been reduced by more than 30%. Table 2-10 presents the results of the leaching
tests and whether the mixture passed this screening test after 10 flushes.
Table 2-10 Leaching results of various media mixtures indicate that three mixtures
pass the leaching test
MEDIA Pass/ Fail Flush Number Contact Time (hours) Nutrients (mg/L) Metals (ug/L)
TKN Nitrate +Nitrite Total P Cu Pb Zn
EPA Freshwater Criteria (EPA 2006) NR 10 0.055 13 65 120
15% peat 85% sand Control 1st 48 11 0.15 0.27 14.6 ND(5) ND(20)
Control 1st 18 4.4 0.59 0.42 4.5 ND(5)
Control 10th 18 2.5 0.09 0.23 2.5 ND(5) ND (20)
15% compost 85% sand fail 1st 48 53.9 196 31.4 307 66.4 489
15% paper 85%sand PASS 1st 48 ND(,3) 0.03 0.23 ND(2) ND(5) ND (20)
10% paper, 5% compost, 85% sand fail 1st 18 6.5 45.5 0.41 19.7 ND(5) ND (20)
PASS 10th 18 2.5 0.32 0.05 3.7 ND(5) ND (20)
5% paper, 10% compost, 85% sand fail 1st 18 12.5 35.4 1.94 34.3 ND(5) 24.3
fail 10th 18 4.1 0.06 2.01 13.2 ND(5) ND (20)
100% tires fail 1st 48 43.9 0.72 0.23 68.3 ND(5) 13500
50% tires, 50% sand fail 1st 18 7.4 0.41 0.09 22.4 ND(5) ND (20)
PASS 10th 18 2.8 0.05 0.09 4.1 ND(5) ND (20)
10% tires, 90% sand PASS 1st 18 3.9 0.48 0.05 10.4 ND(5) ND (20)
PASS 10th 18 2.4 0.16 1.02 6.1 ND(5) ND (20)
Note: Failed results are in italics.
As presented in Table 2-10, compost and tires both leach metals, such as copper, and
paper does not. Therefore, replacing only a portion of peat with compost and a
portion of the sand with tires was considered to reduce the impact of leaching.
Because the results indicate that metals and nutrients leaching decreases with flushes,
mixtures which pass the tenth flush were considered for further evaluation. Figure 2-
21


4 presents the results of copper leaching after the first and tenth flush of various
media mixtures. The shaded area represents the EPA fresh water quality standard of
13 ug/L copper (EPA 2006). Figure 2-4 indicates that a mixture of less than 10%
compost and less than half replacement of tires with sand must be used to attain the
water quality standard within 10 flushes. Therefore, based on the water quality
criteria, any mixture with less than 42% tires and less than 10% compost is
permissible.
Figure 2-4 Reduction of Copper Leaching From Various Mixes After Multiple
Flushes of Water. The Gray Area Indicated the Permissible Waste
Mixture Passing the Metals Leaching Test
2.4.3 FLOW RATES
After leaching tests, paper, compost and tires remained as viable materials.
Permissible mixtures of the materials based on leaching results were then compared
to the flow rate. The results of the flow rates indicate that paper slows the infiltration
while tires increase infiltration rates. An optimal range of paper and compost may be
reached based on the comparison to the flow rate of the control. Figure 2-5 shows the
permissible amount of compost in the mixture is about 6.5-11% and 4-8.5% paper
based on flow rate alone.
22


Figure 2-5 Effect of Varying Amounts of Compost and Paper on Flow Rate Shows
Compost Increases How rate and Paper Slows the How Rate
The shape, size and buoyancy of materials also affect the infiltration rate. Shredded
tires, for example float in water. When water is added to tires only (100% tires), the
tires float and the flow rate is basically free flow, as fast as the outlet allowed. This
case was not considered. Figure 2-6 shows the increase in flow rate when replacing
tires with sand. The increase in flow rate should be considered when mixing tires and
paper, where tires can offset the reduction in flow rate from paper. The shaded area
represents the permissible range of shredded tires in the mixture.
23


Effect of Replacing Sand with
Shredded Tires
Figure 2-6 Increase in Row Rate is Related to the Amount of Shredded Tires
2.4.4 CONFIRMATION TEST- GERMINATION
The results of the previous three tests lead to permissible mixtures to be confirmed
through germination test. The germination test was conducted with the following
mixes:
1. Control: 15% peat and 85% sand
2. 10% compost, 5% shredded paper, and 85% sand
3. 5% compost, 10% shredded paper and 85% sand
4. 6% paper, 10% compost, 42% tires and 42% sand.
As presented in Figure 2-7, the testing confirmed the success of germination of the
four mixtures. The conclusion of the batch test germination rates was a mixture of
between 5% to 10% compost and less than 42 % tires will meet the germination test
requirement.
24


Germination Rate
mix number
mix #1 -15% peat, 85% sand
o mix #2 -10% compost, 5% paper, 85% sand
x mix #3 5% compost, 10% paper, 85% sand
o mix #4 -10% compost, 6% paper, 41% tires, 42% sand
Figure 2-7 Germination Tests Confirmed no Difference in the Ability of the Four
Mixes to Support Vegetation
2.5 SCREENED BEDDING MATERIAL MIXTURE
The summary of the permissible bedding mixture is presented in Table 2-11 based on
the three screening criteria, 1) cost/availability, 2) leaching, 3) flow rate, and
confirmation germination tests. Results indicate that mixtures of compost, shredded
paper, and shredded tires pass the screening criteria and confirmation tests. The three
materials were available and cost effective. Leaching and flow rates determined the
amount of compost, paper and tires in the mix. The permissible amount of compost is
between 6.5-10%, shredded paper between 5-8.5% and shredded tires less than 10%
of the complete mixture.
25


Table 2-11 Permissible ranges of waste-incorporated filtration mixes
Virgin Material Replacement Material Permissible amount of material in the mixture based on the following criteria: Permissible Amount of Material
Cost/ Availability Leaching Flow Rate 17 day germination rate
Peat Compost Pass 10% or less* 6.5-11% 5-10% 6.5-10% (5-8% if mixed with tires)
Shredded paper Pass 15% or less 4-8.5% 5-10% 5-8.5%
Sand Tires Pass 42% or less <8% 42% or less 0-8%
Complete replacement of the organic portion (peat) is possible with a mixture of
paper and certified compost. If the mixture contains only sand (not tires), a mixture
of 6.5-10% compost and 5-8.5% shredded paper may be used. Due to leaching from
both tires, the amount of compost must fall between 5-8% of the total mix if tires are
also incorporated. If tires are incorporated in the mix to offset some sand, equal
amounts of paper (7.5%) and compost (7.5%) would fall within the permissible
amounts for the compost, paper, sand and tires scenario. A mixture of equal amounts
of paper (7.5%) and compost (7.5%) falls within the permissible amount while
obtaining benefit of each mixture. A mixture between the permissible amounts of
each material creates a buffer for error in measurement while reaping the benefits of
waste material replacement. The possible mixtures are shown in Figure 2-8.
26


15% Peat
85% Sand
7.5% Compost
7.5% Paper
85% Sand
7.5% Compost
7.5% Paper
77% Sand
8% Tires
Figure 2-8 Picture of Control and Permissible Waste-Incorporated Media Mixtures
for PLD
2.6 GREENHOUSE GAS BENEFIT
The environmental benefit of incorporating the waste materials was evaluated with
greenhouse gas emissions. The benefit of waste materials (compost, paper, and
shredded tires) and the offset virgin materials (peat and sand) are both accounted for
in the net GHG benefit, (eg use of compost and paper must account for both the use
of the material and the avoided peat). Therefore, if 1 ton of compost were to replace
1 ton of peat the benefit must include .2 MTC02E per ton of compost plus .84
MTC02E per ton of sand.
A design example based on the GHG savings from replacing the virgin materials with
waste materials is presented in Table 6. A field site on Auraria campus in Denver,
Colorado was chosen to retrofit an older sand filter with a PLD. The footprint of the
PLD is 213.7 square meters (2,300 square feet). The depth of the bedding material is
45.7 cm (18 inches, 1.5 feet) so the volume of material required is 97.7 cubic meters
(3,450 cubic feet, 127.7 cubic yards). Two options for bedding material are
considered: the business as usual (BAU) currently recommended mixture of 15% peat
and 85% sand by volume and the waste-incorporated mixture of 7.5% compost, 7.5%
paper, 76% sand and 8% tires by volume. Bedding materials are generally sold by
weight and the recommendation is based on volume therefore the density was
measured. The density of the materials was evaluated in the lab and the total weight
27


of each material in the example PLD of 97.7 cubic meters (127.7 cubic yards) is
presented in Table 2-9.
Table 2-12 GHG benefit in MTC02E of replacing Control with a_Waste
Incorporated CPST mix
Mix Number Material % by Volume Material Weight in Example PLD (tons) [avoided tons of virgin material] GHG MTC02E /ton Total MTC02E from the installation of the Example PLD (+ released avoided)
Business-As Usual (control) Peat 15 6.70 +0.84 +6.51
Sand 85 146.54 +0.006
waste- incorporated PLD (CPST) Compost 7.5 4.93 [3.35 tons peat] -0.2 -0.62
Paper 7.5 0.19 [3.35 tons peat] -0.05
Tires 8 10.22 [15.52 tons sand] -0.04
Sand 76 131.02 +0.006
As presented in Table 2-9, a total of 6.5 MTC02E is released from the use of peat
and sand in the example PLD. The total GHG savings from installing the waste-
incorporated mixture is 0.6 MTC02E. The net benefit includes both the savings from
the use of waste materials plus the GHG offset by avoiding the virgin materials.
Additionally the benefit is location specific as 0.18 MTC02E of the 0.84 MTC02E
are from transportation of Peat from Canada to Colorado. Additional information
about the impacts of peat use, processing, transportation and degradation in soil are
included in Appendix A. For example, the net benefit shown in Figure 2-9 is a total
of 7 MTC02E.
28


GHG Comparison of Two Media Mixes
Figure 2-9 Net GHG Benefit of Installing the CPST Waste-Incorporated Mix
Instead of the Business as Usual (Control) in an Example PLD
The total greenhouse gas benefit of offsetting all the peat with compost and paper and
a portion of the sand with tires is 7.1 MTC02E for 98 cubic meters (128 cubic yards
of material or 3,450 square feet of PLD surface area). Since the environmental
impact of peat is high the total cost of all virgin materials (6.5 MTC02E) can be
offset by replacing the peat with compost and paper (net benefit of 6.6 MTC02E).
An additional benefit of 0.4 MTC02E is realized by offsetting a portion of the sand
with crumb rubber.
Similarly the cost benefit of offsetting the BAU with WI mix is attributed to the
compost and paper. The cost of tires is the same as sand and does not contribute to
the cost savings. Based on the example PLD, the cost savings to install the WI mix
compared to the BAU is $2,800 as presented in figure 2-10.
29


Cost Comparison for Two Media Mixes
\
cn
O
O
_i
<
i
o
Cost
Savings
$2800
Figure 2-10 Cost Savings to Installing the CPST Waste-Incorporated Mix Instead of
the Business as Usual (Control) in an Example PLD
2.7 CONCLUSION
Engineers and contractors may be inclined to change bedding material specifications
when materials are unavailable or expensive. Randomly substituting materials into
specifications may result in unknown impacts to the operation of the PLD in terms of
water quality, infiltration rates and plant growth. The relative amount of material in
the mixture, such as shredded tires, compost and paper, was found to effect the
performance. For example, complete replacement of peat with compost initially
would result in excess metals and nutrients leaching into the stormwater, thereby
degrading water quality. Tire particles were found to leach high amounts of copper
and zinc after long contact times. Therefore, tires and compost should be used correct
percentages. The current recommendation of peat and sand may be used as well as the
waste-incorporated mixture of compost and paper with less than 8% replacement with
tires and a mixture of compost and paper.
A variety of waste materials were selected and screened based on availability, cost,
leaching, flow rate and germination rates. The screening process led to a list of
unsuitable materials and range of permissible mixtures for the suitable materials. Two
final mixtures were recommended for use in a waste-incorporated media for a PLD.
The currently used 15% peat and 85% sand may be replaced by either of two
mixtures. One option is to replace the peat with equal amounts of paper (7.5%) and
30


compost (7.5%). Another option is a mixture replacing both the peat and sand
portions which would consist of 7.5% paper, 7.5% compost, 77% sand and 8% tires.
The greenhouse gas benefit of offsetting both peat and sand resulted in a total benefit
of 7.13 MTC02E for an example PLD of 3,450 square feet of PLD surface area. Of
the 7.13 MTC02E, the greatest portion of the savings (6.63 MTC02E) is realized
from the replacement of the peat with compost and paper.
Additionally, lighter materials such as organics (peat, compost, paper) and tires were
found to be buoyant in water. Although this was not a screening criterion, the impact
of buoyancy effects will be studied in further bench scale testing. Floating particles
in stormwater plus added particles from the bedding material may cause obstructions
in outflow designs and possible impacts to downstream water quality.
31


3
INFILTROMETER DESIGN AND FLOW MODEL
3.1 INTRODUCTION
The porous landscaping detention basin (PLD) in Figure 3-1 reduces on-site
stormwater runoff volume and peak discharge using subsurface infiltration. Beneath
the vegetated basin bottom is an aggregate sub-base that is typically divided into an
upper filtering layer comprised of fine aggregate, and a lower reservoir layer
comprised of larger aggregate. The geotextile fabric provides the separation between
these two layers. Stormwater that is intercepted by the surface basin will be infiltrated
into the subsurface reservoir where the seepage flow is filtered, stored, and gradually
released into the perforated pipes that are tied into the downstream sewer manhole.
The water detention process is usually assumed to begin with dry subsurface layers.
During an event, all the aggregate voids are filled up with water before the seepage
flow can be fully developed through the saturated medium.
The infiltration rate on the basin bottom represents the inflow to the PLD system
while the seepage rate through the subsurface medium represents the outflow. The
operation of a PLD is controlled by either the infiltrating rate or the seepage rate,
whichever is smaller (Guo 1998). If the subsurface seepage flow cannot sustain the
infiltrating flow, the water mounding will be built up to balance the inflow and
outflow rates. This phenomenon is manifested by standing water often observed in
many water quality enhancement basins (Guo 2001). The similar phenomena were
also observed in the subsurface drip irrigation system that has become a common
method for the irrigation of field crops, trees, and landscaping. When the pre-
determined discharge of the emitter is larger than the soil infiltration capacity, water
pressure at the dripper outlet increases and can become built up. This pressure
buildup in the soil decreases the pressure difference across the dripper and,
subsequently, decreases the trickle flow (Shani et al. 1996). Therefore, it is advisable
that the subsurface geometry beneath the PLD sustains the continuity of flow.
32


Figure 3-1 Layout of Porous Landscaping Basin
The design goal of a PLD system is to seek the balance between the surface and sub-
surface flows under the available hydraulic head and required water quality control
storage volume. This chapter presents an attempt to integrate the surface and
subsurface hydrology and hydraulic processes together to design a PLD (Guo et al
2009).
3.2 SURFACE STORAGE BASIN
A PLD is designed as an on-site storm water disposal facility. The storage volume of
a PLD is often sized for the water quality control volume (WQCV) or equivalent to
the flush volume. From previous studies, WQCV is approximately equal to 3- to 4-
month event (Guo and Urbonas 1996). For convenience, WQCV is directly related to
the local rainfall distribution. There are many recommended probabilistic
distributions derived for complete rainfall data series, such as exponential
distribution (Bedient and Huber, 1992), one-parameter Poisson distribution
(Wanielista and Yousef, 1993), and two-parameter model of Gamma distribution
(Woolhiser and Pegram, 1979). In this study, the one-parameter exponential
distribution is adopted to fit the frequency distribution of rainfall event depths (Guo
2002). The exponential distribution is described as:
33


Equation 3-1
in which f(p) is the frequency distribution for local rainfall depth, p. Integrating
equation 3-1 yields the cumulative probability distribution as:
Equation 3-2
F(0< p in which F(P) is the non-exceedance probability for rainfall depth to be less than or
equal to the design rainfall depth, P. Considering surface depression, a runoff-
producing rainfall depth can be converted into its runoff volume as:
Equation 3-3
P0 = C(P Pj)
in which Pa is WQCV in mm per watershed, C is the runoff coefficient, and P, is the
incipient runoff depth [mm]. As recommended, an incipient runoff depth of 2.5 mm
is introduced to filter out extremely small rainfall events (Guo and Urbonas in 1996,
Driscoll et al. in 1989). Normalizing equation 3-3 yields
Equation 3-4
in which Pm is the local average rainfall event-depth that can be found elsewhere
(EPA 1986). Substituting equation 3-4 into equation 3-2 yields
Equation 3-5
F(0< p< P0) = \-e P CP"
in which F(0 captured by the design water control volume, i.e. P0. In this study, equation 3-5 is
termed the synthetic runoff capture curve that is normalized by local average rainfall
event-depth, runoff coefficient, and runoff incipient depth. Re-arranging equation 3-5
yields:
34


Equation 3-6
z3l
Cv -\-aeCPm
Equation 3-7
p
a = e m
In which Cv is the runoff volume capture rate between zero and unity, and a =
constant determined by incipient runoff depth. The value of a represents the
watershed natural depression capacity. Figure 3-2 presents a set of normalized runoff
capture curves produced using equation 3-6 with runoff coefficients of 0.4, 0.6, 0.8,
0.9 and 1.0. It is noticed that the curvature of runoff capture curve increases when the
runoff coefficient decreases. This tendency reflects the fact that the higher the
imperviousness in a watershed, the less the surface depression and detention. As a
result, the response of a watershed to rainfall is quick and direct.
-C=0.40 -C=0.60 -C=0.80 C=0.90 -*-c=i.O I Po/Pm (Basin Slze/Average Rainfall Depth)
Figure 3-2 Stormwater Quality Control Volume for Porous Landscaping Basin
Design
35


For a selected runoff volume capture rate, Figure 3-2 provides the required WQCV
for a PLD basin as:
Equation 3-8
y0 = p0a
where V0 is the WQCV [m3] and A is the catchment area tributary to basin [m2].
Safety is always a concern when designing a PLD. Often the water depth in a PLD is
set to be 15 to 30 cm (6 to 12 inches). With a selected basin depth, the basin cross
sectional area is determined as:
Equation 3-9
where A0 is the average cross section area, and Y is the basin depth. To enhance the
infiltrating process, the basin bottom shall be on a flat to mild slope.
3.3 SUBSURFACE FILTERING SYSTEM
Drain time is critically important to the operation of a PLD because it controls the
sediment removal rate. Based on the urban pollutant characteristics, a drain time for
the PLD is usually set to be between from 12 to 24 hours (USWDCM 2001). Figure
3-3 illustrates the flow through the two filtering layers including sand-mix and then
gravel.
The model is limited to saturated conditions as the infiltration rates through the
filtration layers reached steady state through 2 filtering layers. This is not applied to
the entire wetting cycle in soil saturation. Additionally, the ponding water is assumed
to be constant depth. Under a constant head, the steady flow condition is derived as:
Equation 3-10
/=v,=v2
In which/is the infiltrating rate, and V is the seepage flow velocity through each
layer, The subscriptions 1 and 2 represent the variables associated with the sand-
mix and gravel layers, respectively. A saturated seepage flow through a medium is
proportional to the energy gradient as:
36


Equation 3-11

= K,
dHx
Equation 3-12
dH2
K,
H,
Where K is the hydraulic conductivity, H is the energy head, and dH is the energy
loss.
Figure 3-3 Illustration of Infiltrometer Operation under Saturated Conditions
In practice, the design infiltrating rate depends on the drainage nature of the selected
soil-mix. With a pre-selected design infiltrating rate, the total filtering thickness for
the two filtering layers is calculated as:
Equation 3-13
D = fTd
Where D is the total thickness for two filtering layers,/is the infiltration rate, and Tj
is the PLD drain time. The fundamental challenge in PLD design is how to divide the
37


total thickness between the two filtering layers because the layer thickness is directly
related to the hydraulic gradients for seepage flow through the system.
Equation 3-14
D = H] +H2
where Hi is the sand-mix thickness and H2 is the gravel layer thickness. As illustrated
in Figure 3-3, the available hydraulic head for the PLD system is
Equation 3-15
H = Y + D
where Y is the water loading depth in PLD. In this study, the optimal performance of
a PLD is defined by the infiltration flow and the subsurface thickness that allow the
seepage flow to consume the hydraulic head available as:
Equation 3-16
H=dHl +dH2
Aided by equations 3-10, 3-11, and 3-12, the head losses through the two filtering
layers are:
Equation 3-17
dH. = H.
K,
Equation 3-18
dH2
2
Aided by equations 3-15, 3-16, 3-17 and 3-18, the optimal performance of a PLD is
described as:
Equation 3-19
L+Jm_______
D f f
K, K
-)
in which f/Ki >1, and K2>Ki
38


Equation 3-20
H2=D-Hx
Equation 3-19 is valid when f/Ki >1 and K2>Kj. In other words, the infiltration rate is
greater than the seepage rate and the sand-mix layer is above the gravel layer.
Equations 3-19 and 3-20 are derived to be the guidance to divide the total required
filtering thickness into two layers.
3.4 OPERATION OF POROUS LANDSCAPING BASIN
The operation of a PLD is subject to clogging due to sedimentation in the basin. The
pollutant deposit is often accumulated on top of the basin bottom and then diffused
into the top layer of the sand-mix medium. When the infiltration rate decays on the
basin bottom, the friction losses are reduced accordingly. Namely,
Equation 3-21
Equation 3-22
dH2
Mi
k2
In which fs is the reduced infiltrating rate due to clogging effect. Clogging to a PLD
system generally occurs as a thin, hard cake layer (0.2 to 0.5 cm) of sediment on
the basin bottom. As reported, migration of solids diffuses into the top 5 to 10 cm of
the sand-mix layer while the hydraulic conductivity in the lower sand-mix layer
remains unchanged (Li and Davis 2008a and 2008b; Mays 2005). As illustrated in
Figure 3-3, the reduced infiltration flow under a clogging condition may not
completely consume the hydraulic head available. As a result, the residual pressure in
the system, is calculated as:
Equation 3-23
H, Y + D-dH\ -dH2
39


where H, is the residual pressure head. If the cake layer presents an additional friction
loss, the value of H, will decrease; otherwise it represents the pressure built up in the
filtering layers. The reduced infiltrating rate implies a prolonged drain time, or a
period of standing water in the PLD, as:
Equation 3-24
in which Tw is the increased drain time or period of standing water. As the basin
bottom is gradually clogged, the infiltration flow rate continues to decrease while the
residual pressure continues to increase. When the residual pressure is becoming to
equal the available head, the system will about to cease functioning because the basin
is completely plugged (Guo et al 2009).
3.5 LABORATORY TESTS
3.5.1 PREVIOUS DESIGN STUDIES
Lab studies focusing on stormwater BMPs similar to the PLD have been limited.
Additionally large column studies are uncommon due to many variables affecting
infiltration in surface soils. The complexities of systems as they exist in the field are
challenging to replicate and measure in lab conditions.
Generally, in order to isolate one or several factors, column studies are completed
with small diameter columns (< 25 cm or <10 inches) in controlled lab conditions.
Soil column studies focusing on physical characteristics for measurement of saturated
hydraulic conductivity (Davis et al. 2001; Mays and Hunt 2005; Yang et al. 2004) are
by up flow. Costs for this type of systems are cost prohibitive and researchers
generally use one column. Additionally field conditions are best replicated by down
flow.
Studies focusing on unsaturated conditions generally use smaller diameter columns
and add water from the top for down flow (Hsieh and Davis 2003; Siriwardene et al.
2007). Research completed by the team at the University of Maryland utilize small
40


columns and growing boxes and apply synthetic stormwater to the top (Davis 2007;
Davis et al. 2001; Davis et al. 2006; Davis et al. 2003).
A large diameter column has the advantage of replicating field conditions and
reducing the effects of sidewall flow (Corwin 2000; Hunt 2003). A few studies of
soil characteristics have utilized larger columns (>30 cm or >12inch diameter)
(Corwin and LeMert 1994; Sun 2004). Even fewer studies related to BMPs have
utilized large diameter columns. The studies are presented in Table 1-3 below.
Table 3-1 Large column studies of systems similar to PLD
Author, Year Study Column Size Measurement equipment
(Hunt 2003) column anoxic zone D= 30.5cm (12") H=122 cm (48") Outflow with graduated cylinder and timer
(Ames, Inkpen et al. 2001) Column for infiltration rates of selected media D= 61 cm (24) H= 91 cm (36") Pressure transducer (flow rate) Water content reflectometer Temperature probe Data logger
Sun 2004 uptake of metals by grasses D=31cm (12") H=31cm (12") Soil cores, vegetation and water samples collected
Generally methods for these include adding water from the top, measuring inflow
and outflow, and collecting water and or soil samples. Some studies also collect data
about the moisture content and water suction in unsaturated conditions (Ames et al.
2001). None of the studies have collected comprehensive data about infiltration
capacity, clogging and water quality. Ames et al had difficulty reproducing
infiltration rates between tests (Ames et al. 2001)
3.5.2 INFILTROMETER DESIGN AND TESTING
The design for this study included a large diameter PVC column, pressure ports and a
variable height outflow shown in Figure 1-3 and Figure 1-4. The large diameter
cylinder (38 cm) which was >600 times the particle size was used to reduce the
influence of short circuiting. Pressure ports and manometers were installed to
measure changes in pressure through the filtration soil-mix. Each column was
equipped with an overflow to maintain constant head of 30 cm (12 in) when
necessary. The outflow at the bottom was designed to change depths from below the
41


course aggregate to above the aggregate as in Figure 1-5. The elevated outflow
served to saturate the filtration layers.
Figure 3-4 Soil Column Design with 2-Layered System and Lowered Outflow for
Field Conditions
42


Figure 3-5 Soil Column Design with Elevated Outflow for Saturated Conditions
Previous reports recommended that a soil-mix thickness of 30 to 60 cm (or 12-24
inches) can effectively remove pollutants in stormwater (Hunt 2003; NCDENR 2007;
Sun 2004; Toronto and Region Conservation Authority 2007). As reported, metals
are removed in the top 20 to 45 cm (8 to 18 inches) of soil-mix layer (Davis et al.
2003; Sun 2004; Winogradoff 2001). In practice, the soil-mix layer thickness is
recommended to be 45 cm (18 inches) to allow for both adequate pollutant removal
and root zone for vegetation. In the laboratory, a 38-cm (15-inch) circular
infiltrometer in Figures 3-4 and 3-5 was utilized to represent a section of the PLD A
view of the inside of the column is presented in Figure 3-6. Soil sample columns
were prepared to mimic the field conditions as closely as possible. As illustrated in
Figure 3-5, the sample column in the infiltrometer is built with an upper filtration
layer of soil-mix, a lower layer of gravel, and a perforated bottom drain. The 1.9 cm
(3A -inch) crushed granite was spread in the bottom. ASTM C33 washed sand and
Canadian peat were combined at the ratio of 85% sand and 15% peat by volume for
the soil-mix layer. To evaluate the effect of the geotextile on the infiltration rates,
Sample Column I had a geotextile layer separating the soil mix layer from the
aggregate layer; and sample Columns J and F were constructed and tested without the
geotextile. Figure 3-7 shows the geotextile set inside the soil column.
43


Figure 3-6 Infiltrometers Built in Laboratory.
Figure 3-7 Measurements for Soil and Water Levels Inside the Soil Column
44


Figure 3-8 Geotextile between Large Aggregate and Filter Layer Inside the Column
Manometers were installed at the ports seen in Figure 3-6 and 3-7. Inside the
columns the manometers consisted of a hose barb, flexible tubing and a cover, as in
Figure 3-9. The outside of the column flexible tubing allowed for a static pressure
measurement.
Figure 3-9 Manometer Setup to be Placed inside the Column
Table 3-2 presents the compaction and density of the soil-mix samples prepared for
infiltrometer tests. For this study, the filtering layers were structured with an upper
45-cm (18-inch) sand-mix layer and a bottom layer of 20-cm (8-inch) gravel. As
illustrated in Figure 3-6, the total thickness of a sample column is set to be 65 cm (26
inches). With a constant head of 30 cm (12 inch), the total head applied to the
infiltrating flow is 95 cm (38 inches) as shown in Table 3-2.
45


Table 3-2 Compaction and density of soil-mix sample
Material Soil-mix Sample Density grams per cubic centimeter (pounds per cubic foot) Compaction Ratio
Loose Soil-mix Compacted Soil-mix
Sand 1.39 (86.79) 1.76(109.65) 1.26
Peat 0.42 (26.05) 0.84 (52.43) 2.01
85% Sand and 15% Peat 1.16(72.31) 1.75 (109.22) 1.51
0:00:00 24:00:00 48:00:00 72:00:00
hours:min:sec
Figure 3-10 Variation of Infiltration Rates for Sample Columns (Control: Peat and
Sand)
Manometers were installed on the infiltrometer wall at 4 stations to measure the
variation of static heads. The locations of manometers are expressed by the vertical
distances above the bottom of the sample column as shown in Table 3-2.
46


Table 3-3 Variation of hydraulic heads measured at 5 stations
Column Infiltrating Locations of manometers above ground in cm
Sample ID Rate 66.0 61.0 27.9 22.9
cm /hr entrance sand layer upper sand layer lower sand layer geotextile layer
Reading in manometers above ground in cm
Col F without geotex 22.9 97.8 95.3 31.1 22.9
Col I with geotex 24.3 96.5 94.0 29.2 22.9
Col J without geotex 25.9 96.5 94.6 48.6 22.9

Col F without geotex 8.8 97.8 96.2 27.9 22.9
Col I with geotex 12.4 96.5 94.6 27.9 22.9
Col J without geotex 13.5 96.5 95.3 30.5 22.9
Note: Col F without geotex is the Soil Column F without geotextile measured after 72 hr
operation.
Table 3-4 Hydraulic conductivity in filtration layer after 72 hours of water flow
Column Sample ID Infiltration Rate cm/hr Hydraulic Conductivity for Soil-mix cm/hr
Col F without geotex 8.8 4.9
Col I with geotex 12.4 6.1
Col J without geotex 13.5 7.3
The sample columns were tested with the wet initial condition by filling each
column to 30.5 cm (12 inches) of water and soaking the sample column overnight
with the drain valves closed. The next morning the valve was opened and outflow
measurements were recorded. The outflow rates from sample columns were recorded
for 72 hours continuously. Figure 3-9 is the plot of the decayed infiltration rate
through the sand mix layer over 72 hours. The best fitted Hortons infiltration
equation for the soil mix tested is (Horton 1933):
Equation 3-25
/, =9.8 + 37.2*r-144'
Where/t is the infiltration rate in cm/hr at time t in hours, and k = decay constant in
1/hour. Equation 3-25 has a correlation coefficient, r2 = 0.85. The differences in the
47


infiltration decay among sample columns are attributed to sample preparations,
compaction condition, and test operations. Soil compaction alone can have a
significant impact on infiltration rates in sandy soils (Pitt et al. 1999). Although the
impact of the geotextile fabric on the flow rate was not measurable through the
sample columns, when the columns were deconstructed the soil-mix particles were
found to migrate into the gravel layer. This visual observation has was also found by
Haliburton and Wood. (Haliburton and Wood, 1982). Based on three soil column
tests, the average final infiltration rate is found to be 11.6 cm/hr (or 4.6 inch/hr). The
hydraulic conductivity coefficients were found to be 6.1 cm/hr (or 2.3 inch/hr) for soil
mix and 64.0 cm/hr (or 25.2 inch/hr) for gravel. Results listed in Table 3-3, these
values are within the range reported before (Schwartz and Zhang, 2003). Literature
reports of hydraulic conductivities in bioretention systems are 1.3 to 15 centimeters
per hour (cm/hr) (0.5 to 6 inches per hour (in/hr)) (Davis et al. 2001; Hunt 2003; Hunt
and White 2001). During field tests of installed PLDs Hunt found a range of 1.3 to
3.3 cm/hr (0.5 to 1.3 in/hr). The minimum design criteria set by the UDFCD is 2.5
cm/hr (1 in/hr) (UDFCD 1999).
3.6 DESIGN EXAMPLE AND SCHEMATICS
The PLD basin located in the City of Denver, Colorado is employed as an example to
illustrate the design procedure. The catchment draining into this basin has a tributary
area of 1.0 hectare (2.5 acre) and runoff coefficient of 0.60. As reported, the average
event depth for the Denver metropolitan area is 1.0 cm (or 0.4 inch) (EPA 1986).
With a runoff volume capture rate of 80%, the WQCV for this basin is calculated as:
Equation 3-26
~pi -0.25
a = ePm = e 10 =0.79
Equation 3-27
z3l -p0
Cv=\-aecp" = 1 -0.79 e0-60*10 =0.8 OrPo = 0.86cm
Equation 3-28
V0 = P0A = 0.86 cm x 1.0 ha = 86 m3
48


From the laboratory test,/is 11.6 cm/hr (or 4.6 inch/hr), Kj is 6.1 cm/hr (or 2.4
inch/hr) for soil mix and K2 is 64.0 cm/hr (or 25.2 inch/hr) for gravel. Consider a
drain time of 6 hours. The required dimension for the subsurface filtering system is
calculated as:
Equation 3-29
D = fTd =11.6x6 = 69.6 cm
Equation 3-30
H = Y + D = 30.5 + 69.6 = 100.1 cm
Equation 3-31
H_
D
'-=1
f Y
K,
D
f f
(-----)
K, K
1-
(^-1)------L-
6.0 69.6
11.6 11.6
6.1 64.0
0.72
or Hi = 50.1 cm and H2 = 19.5 cm
Substituting the above dimension into equations 3-17 and 3-18 yields
Equation 3-32
dH. = H. =^x50.1 = 95.3 cm
1 Kx 1 6.1
Equation 3-33
dH2 = H2 =^xl9.5 = 3.5 cm
2 K2 64.0
As expected, the total friction losses for the seepage flow through the two layers
satisfy equation 3-15. This is the optimal performance for this PLD basin, according
to equation 3-19. In practice, the infiltration rate at the site is estimated with
uncertainties. Secondly, the infiltration rate is subject to the clogging effect. For
comparison, the performance of this PLD is further assessed for the condition that the
49


infiltration rate is reduced to 7.5 cm/hr. The corresponding head losses are calculated
as:
Equation 3-34
K, 6.1
= 61.6 cm
Equation 3-35
K7 64.0
= 2.3 cm
Equation 3-36
H, = H -dHx-dH2 =100.1-61.6-2.3 = 36.2 cm
Equation 3-37
69.6
7.5
- 6.0 = 3.3 hr an extended period of standing water.
Under the clogging condition, this PLD drains slowly. After the design drain time of
6 hours, standing water is developed in the basin. The residual pressure head, Ht, is
either absorbed in the clogging (cake) layer or built up in the soil medium.
The above analysis was repeated for the bioretention medium studies (Li and Davis
2008a and 2008b). In the laboratory, the clean sand column was used as the
subsurface medium to filter the solids in storm water. As listed in Table 3-4, a 0.1 to
0.6-cm cake layer was observed to form on the basin bottom. The top layer of the soil
column is gradually clogged by the migrated solids. Table 3-4 summarizes the
observed thickness of the top clogged layer. The clean sand column started with a
hydraulic conductivity of 45 cm/hr. After the top layer is clogged, the equivalent
hydraulic conductivity coefficients for the soil column were measured as listed in
Table 3-4. Consider that the clogged top layer is equivalent to a sand-mix layer and
the bottom clean layer is equivalent to a gravel layer. Equations 3-19 were applied to
3-7 case studies to divide the total medium thickness into the top and bottom layers.
Figure 3-10 presents good agreement between the calculated and observed thickness
of the top layer.
50


Table 3-5 Observed case studies for bio-retention medium (Li and Davis 2008a and
2008b)
Infiltration Flow Rate f cm/h Total Filtering Depth D cm Conductivity Coefficients Thickness
Top Layer K, cm/h Bottom Layer k2 cm/h Equivalent Ke cm Top Layer cm Bottom Layer cm Cake Layer cm
4.80 5.50 2.00 45.00 3.00 3.50 2.00 0-0.3
4.90 10.50 2.00 45.00 3.00 7.00 3.50 0-0.4
9.50 5.50 2.00 45.00 4.00 4.10 1.40 0-0.3
9.50 10.50 2.00 45.00 3.00 7.40 3.10 0-0.1
19.80 5.50 2.00 45.00 5.00 3.70 1.80 0-0.3
19.70 10.50 2.00 45.00 6.00 6.70 3.80 0-0.6
20.10 5.50 2.00 45.00 7.00 4.30 1.20 0-0.2
PLD SubBase Dimension
Observed Optimal Dimension H1 / H (Top Sand Layer)
Figure 3-11 Comparison between Observed and Calculated Filtering Thickness
51


3.7 CONCLUSION
A PLD should be designed using the concept of hydrologic system to take both the
surface and sub-surface flows into consideration. In this study, it is recommended that
the PLD be sized for the on-site storm water quality control volume according to the
selected drain time and target pollutant removal rate. The total filtering thickness
underneath the PLD shall be determined by the selected drain time and infiltration
rate.
The filtering layers beneath a PLD shall be structured to completely consume the
hydraulic head available in the system. The optimal dimension of the sub-base
medium is found to be closely related to the design infiltrating and seepage rates. In
this study, equation 19 was derived to provide the optimal thickness ratio between the
sand-mix and gravel layers. The optimal sub-base dimension delivers the highest
seepage flow rate for the unclogged condition.
Equation 3-38 is numerically sensitive to f/Ki, but not to f/K2 because the hydraulic
conductivity coefficient of gravel is usually much higher than the infiltrating rate or
the ratio, f/K2, which is numerically close to zero. For simplicity, the thickness for
the sand-mix layer is approximated as:
Equation 3-38
h, = (i--^-)D + -^-r
' / /
In practice, it is critically important that the ratio, f/Kj, is properly selected to avoid
undesirable prolonged standing water in the PLD.
In this study, the infiltration rate for the soil-mix layer varies from 50 to 7.5 cm/hr (20
to 3 inch/hr). The final infiltration rate is approximately 7.5 to 12.5 cm/hr (3 to 5
inch/hr) after an operation of 72 hours. The hydraulic conductivity coefficient was
varied within a small range through the soil-mix column. All these uncertainties are
attributed to the residual pressure in the PLD system. As a common practice,
perforated pipes are installed in the subsurface system. A sub-drain pipe creates an
accelerated hydraulic gradient to collect the excessive water and to alleviate the build-
up pressure.
52


4 WASTE-INCORPORATED BENCH SCALE TEST- BARE SOIL
This chapter expands on the paper Waste Incorporated Sustainable Design of
Stormwater Detention Basins 2. Bench Scale Tests by Shauna M Kocman, James C.
Y. Guo, Anu Ramaswami submitted to ASCE Journal of Environmental Engineering
April 2010.
4.1 INTRODUCTION AND BACKGROUND
Sustainability as it relates to urban water must take into account both quality and
quantity of water resources, as well as beneficial reuses of urban waste streams now
and into the future. Low impact development (LID) techniques take advantage of
micro-scale approaches to mold the development of land to function similarly to
natural drainage systems, thus replicating ecosystem services which the open space
would have performed (Sample and Heaney 2006). PLDs are designed based on
water quality capture volumes from the contributing watershed to reduce peak flows
and optimize sediment removal rates (Guo and Urbonas 1996b; Guo and Urbonas
2002). Studies have shown PLDs enhance stormwater by reducing peak runoff
volumes, and by providing filtration of sediments and absorption of pollutants. (Davis
et al. 2001; Davis et al. 2006; Davis et al. 2003; Hunt 2003; Hunt et al. 2006). The
top layer of the PLD consists of a vegetated porous detention basin underlain by a
drain.(USWDCM 2001). This chapter investigates the impact of varying the soil-mix
filter layer based on the performance and behavior of the system.
While many studies have been done on PLDs (Davis 2007; Davis et al. 2006; Davis et
al. 2003; Guo et al. 2009; Hsieh and Davis 2005b; Hsieh and Davis 2005c; Kim et al.
2003; Sun and Davis 2007), traditional studies have focused on measuring infiltration
rates and pollutant removal, but few have combined other performance characteristics
such as sediment accumulation and pathogen removal rates. Although waste reuse is
not a new concept, no previous study has taken a combined look at performance
criteria and environmental benefits of a waste-incorporated enhanced PLD.
53


4.2 BENCH SCALE TEST METHODOLOGY
This research focuses on performance and sustainability of vegetated, waste
incorporated PLD. The physical performance of the systems focuses on both
hydraulic conductivity of soil layers and infiltration capacity of the system, both of
interest for field design and maintenance recommendations. Additionally mass
removal is an important measurement of water quality performance.
The water quality portion focuses on contaminate removal. Fate and transport models
include advection, dispersion, and diffusion equations which are complicated and not
practical for field conditions. Field conditions include complexities such as root
growth, metals sorption, and sedimentation which complicate the system. Therefore
mass accounting will be used for effectiveness of contaminate removal.
Infiltration capacity is described by Hortons infiltration equation (Horton 1933). As
water is applied to the soil surface the infiltration rate reduces infiltration capacity by
an exponential decay function as presented in Equation 4-1.
Equation 4-1
f,=fc+(fo~fc)e-k'
Where/, is the infiltration rate at time t [L/t],/o is the initial infiltration rate [L/t],
fc is the infiltration capacity or equilibrium infiltration rate after the soil has been
saturated [L/t]. The variable k is the decay constant specific to the soil [l/t].
Over time the infiltration capacity depends on soil properties and assumes a
maximum capacity. The infiltration capacity documented by Pit et al to depended
heavily on compaction (Pit et al. 1999): The effect of compaction on sandy soils is
very large, reducing the infiltration rates by between 5 and 10 times.
Through previous waste screening and batch testing permissible mixtures of waste
materials to be incorporated into the PLD were defined. In this study the control and
two permissible waste-incorporated mixtures were further evaluated through bench
scale testing. The peat and sand mixture, currently recommended by the Urban
Drainage Flood Control District for the Denver Metro area, served as a control
(USWDCM 2001). The large (15 inch, 38 cm) diameter infiltrometer designed for
54


this study was utilized in the laboratory to represent a section of the recommended
field installation of the PLD. The columns were packed to replicate the current
recommendation of 45.7 cm (18 in) of filtration media and 20.3 cm (8 in) of
aggregate under drain (Guo et al 2009). The under layers were consistent between all
the columns including the geotextile, gravel layer, and PVC under drain. To avoid
pressure back-up, the under drains were overdesigned to drain above the infiltration
rate. The performance of the selected material mixtures in terms of infiltration and
contaminate removal were measured and assessed. Stormwater was applied to the
packed columns and monitored for performance criteria (flow rates, clogging, water
quality parameters and sediment buildup). Triplicate columns in Figure 4-1 were
created with filtration layers consisting of the following three mixtures:
1. Control (15% peat and 85% sand)
2. CPS (7.5%compost, 7.5% shredded paper and 85% sand)
3. CPST (7.5% compost, 7.5% shredded paper, 8% tires and 77% sand)
Figure 4-1 Setup of Triplicate Soil Columns and Barrels of Stormwater in the Lab
Duplicate bare soil bench tests were completed. The first consisting of a leaching and
a filtering phase and the second a filtering phase only. Clean tap water was added to
the system initially to measure for any leaching from the materials. Then synthetic
stormwater was added to measure filtering of contaminates. The flow rate and water
quality parameters were monitored throughout both the leaching and the filtering
phases. During step 1 an initial saturated flow rate was measured by elevating the
outflow of the soil columns and adding tap water for 72 hours.. The initial condition
acts to saturate the column and measure any leaching of sediment, chemicals, or
pathogens from the media mixture that may occur. As presented in Figure 4-2, after
the columns were saturated the flow rates of triplicate samples of all three mixtures,
except one outlier, approach 8.3 to 13 centimeters per hour (cm/hr) (3.3 to 5.2 inches
55


per hour (in/hr)). The current design minimum requirement is 2.5 cm/hr (1 in/hr),
which all three mixtures obtain. The current design minimum requirement is 2.5
cm/hr (1 in/hr), which all three mixtures obtain.
An initial saturated flow rate for the first bare soil test with new media-mixture was
measured during 72 hours of elevated outflow. The infiltration rates over the 72 hour
duration are plotted in Figure 4-2 and follow Hortons infiltration model stated in
equation 4-1 (Horton 1933).
Clean Water Infiltration Rate
Elapsed Time
hours
a Control (Peat.Sand) Compost, Paper, Sand
X Compost, Paper, Sand, Tires -----Horton Curve for Peat.Sand
- Horton Curve for Compost,Paper, Sand - - Horton Curve for Compost, Paper, Sand, Tires
Figure 4-2 Variation of Infiltration Rates as the Filtration Mixtures Approach
Saturation after 72 hours of Continuous Flow
Hortons estimate for infiltration rate over time is based on the equation:
Equation 4-2
/(0 = fc +(/ -fc)e(~k)'
Where/(7) is the infiltration rate at time t,f0 is the initial infiltration rate [L3/t] and fc
is the infiltration rate at field capacity [L3/t]. The A: is a constant with units 1/time.
56


The best fit equations for the infiltration rate of the three mixtures shown in Figure 4-
2 are:
Equation 4-3
Peat and Sand : fconlrol (t) = 9.9 +(50 -9.9) Equation 4-4
Compost, Paper,Sand :fcps(t) = 14.5 + (56 -14.5)e<_009)'
Equation 4-5
Compost, Paper, Sand,Tires: fcpsl (t) = 9.1 +{35 -9.1)
Where/control (t) is the infiltration rate [cm/hr] of the control (peat and sand) mixture
at time t. /cps (t) is the infiltration rate [cm/hr] at time t of the compost, paper and
sand mixture and the/cpst (t) is the infiltration rate [cm/hr] at time t of the compost,
paper, sand and tires mixture. In equations 4-3, 4-4 and 4-5 fa is expressed in cm/hr,
fc is the infiltration rate at field capacity expressed in cm/hr, and t is the elapsed time
in hours. The A: is a constant with units 1/hr. Equations 4-3, 4-4 and 4-5 have
correlation coefficients of r2= 0.90, 0.83, 0.66.
After the conclusion of the initial clean water leaching phase, synthetic stormwater
was applied to the column in individual applications of 30 cm (12-inch) depths to
represent runoff events which fill the PLD. The stormwater used throughout the study
was collected from an storm sewer outfall N-431E on the South Platte draining from
Denver. Urban runoff was collected from this outfall in 30 gallon plastic drums and
transported to the lab.
This outfall carries approximately 2 cubic feet per second (cfs) dry weather flow and
has a 475 cfs capacity for wet weather flows. The outfall drains 80% of the Lower
Platte Valley Basin (4.47 square miles) consisting of mixed industrial, commercial
and residential land uses. A picture of the outfall is presented in Figure 4-3.
57


Figure 4-3 Outfall N-431E in the South Platte River Where Stormwater for the
Experiment was Collected
The summary of a 4 year record of dry weather sampling for outfall N-431E is
presented in Table 4-1. The total suspended solids (TSS) results during the four years
of sampling data have a mean 61 and a standard deviation of 190 mg/L. The average
concentrations for the Denver area are presented in Table 1-9 for comparison
(USWDCM 2001). After additional sediment is added and stormwater is applied to
the columns, an influent and effluent sample was collected and analyzed for TSS.
Table 4-1 Average water quality results of stormwater outfall N-43 IE
Sample Min Max Average
PH 7.7 8.4 8.1
TSS (mg/L) <1 990 61
TKN (mg/L) >1.0 6.59 1.5
Nitrate +Nitrite (mg/L) >.5 5.66 3.0
Total P (mg/L) .08 1.77 .3
E-coli per 100ml 80 20,000 2,745
58


Table 4-2 Average event mean concentrations for Denver area
Sample Industrial Commercial Residential Undeveloped
TSS (mg/L) 399 225 240 400
TDS (mg/L) 58 129 119 678
TKN (mg/L) 2.7 3.3 3.4 3.4
Nitrate +Nitrite (mg/L) 0.91 0.96 0.65 .5
Total P (mg/L) 0.42 0.43 0.65 0.4
Total Copper (ug/L) 84 43 29 40
Total Lead (ug/L) 130 59 53 100
Total Zinc (ug/L) 520 240 180 100
Stormwater was applied with the same characteristics and equivalent volumes as
urban runoff defined by Guo and Urbonas (2002). The water was spiked with
additional sediment to accelerate the clogging process in order to relate surface
loading to field conditions where the clogging process occurs over years. Water
quality samples were collected for the duration of the stormwater applications.
Water quality samples were collected and analyzed for pH, total suspended solids
(TSS), total dissolved solids (TDS), pathogens, total keldjal nitrogen (TKN), nitrate
plus nitrate (N02+N03), total phosphorous (TP), and total metals. The nutrient and
metals were analyzed at Metro Wastewater Reclamation Districts laboratory. The
total metals sweep (Beryllium, Chromium, Manganese, Nickel, Copper, Zinc,
Arsenic, Selenium, Molybdenum, Silver, Cadmium, Antimony) were analyzed by
EPA 200.8 ICPMS methods. Total coliforms were measured as an indicator of
concentration of pathpogens and were both analyzed in-house and at Industrial Labs
by the membrane filtration method (Clesceri et al. 1998). The pH, TS, TSS, TDS
and pathogens were analyzed at the Auraria Campus laboratory in Technology
building following the Standard Methodology for Examination of Water and
Wastewater guidelines (Clesceri et al. 1998).
Two separate tests were conducted with stormwater. The first consisting of seven
applications and the second with twelve applications of stormwater. After 7
stormwater applications and a total of 1.1 kg/m2 of sediment was accumulated, soil
cores were collected. Then the top 10 cm of filtration layer was removed and the
second test was started.
The soil core samples were analyzed for density stratification and particle size
distribution. Samples of clean soil after compaction and dirty soil after stormwater
59


applications were collected from each column at depths of 0-1 cm, 4-6 cm and 8-10
cm. Soil cores were collected with a 1.45 cm diameter tube. The samples were dried
and weighed to calculate the density. Sieve analysis was conducted for particle size
distribution by ASTM D-421 Standard Practice for Dry Preparation of Soil Samples
for Particle-Size Analysis and Determination of Soil Constants (2009).
4.3 WATER QUALITY IMPACTS
During the initial clean water test and subsequent stormwater applications, the
systems were monitored for leaching and filtering of nutrients, metals, pathogens and
suspended solids. All treatments initially leached total keldjal nitrogen (TKN) and
total phosphorous (TP) during the clean water application and ultimately filtered
TKN and TP from the system. The average removal rate of TKN was between 32%
and 44% with an inflow of 1.7 mg/L and between 71% and 77% removal with 2.6
mg/1. Figure 4-4 presents the leaching of TKN during the application of clean water
and then removal of TKN during the filtration of stormwater.
Leaching and Filtering of TKN
2.5
2
1.5 -
1
0.5
0
Leaching
Filtering
Tap Water Tap Water Final Stormwater
0 0 .33
Accumulative Sediment Load (kg/m2)
Inflow
Control
CPS
BCPST
Figure 4-4 Plot of TKN Concentration in Water for Various Accumulated Sediment
Loads Including Tap Water (No Sediment) and Stormwater with 0.33
and 2.66 kg/m2. The Results Showed No Consistent Statistical
Difference Between the Waste-Incorporated Media Mixes (CPS, CPST)
Versus the Control for Leaching or Filtering of TKN
60


Initially all three mixtures leached TP and then filtered TP. Between 48% and 83%
of TP was removed from an inflow of 0.62 mg/L and between 80% and 86% was
removed from an inflow of 1.3 mg/L. The inflow and outflow concentrations are
presented in Figure 4-5.
Leaching and Filtering of Phosphorous
E,
Q.
I
h
1.4
1.2
1
0.8
0.6
0.4
0.2
0
Filtering
Leaching
Inflow
Control
HOPS
CPST
Tap Water Tap Water Final Stormwater Stormwater
0 0 .33 2.66
Accumulative Sediment Load (kg/m2)
Figure 4-5 Plot of Total Phosphorous Concentration in Water for Various
Accumulated Sediment Loads Including Tap Water (No Sediment) and
Stormwater with 0.33 and 2.66 kg/m2. The Results Showed No
Consistent Statistical Difference Between the Waste-Incorporated Media
Mixes (CPS, CPST) Versus the Control for Leaching or Filtering of TP.
All the treatments filtered copper, even from the clean tap water. Copper may
leach into tap water from copper pipes in a building (Denver Water 2009). The
removal rates varied between 45% and 81% during an inflow of 21 ug/L and between
93% and 95% with the application of 113 ug/L. The difference in removal rate
between the control and the two treatments (CPS and CPST) is not statistically
significant. The inflow and outflow concentrations of copper in ug/L are presented in
Figure 4-6.
61


w
Q.
a
o
o
3
o
50
45
40
35
30
25
20
15
10
Filtering of Copper
270 mg/L Cu
113mg/L Cu
i -
.fee. k
Inflow
Control
CPS
BCPST
Tap Water
0
Tap Water Final Stormwater Stormwater
0 .33 2.66
Accumulative Sediment Load (kg/m2)
Figure 4-6 Plot of Total Copper Concentration in Water for Various Accumulated
Sediment Loads Including Tap Water (No Sediment) and Stormwater
with 0.33 and 2.66 kg/m2. The Results Showed No Consistent
Statistical Difference Between the Waste-Incorporated Media Mixes
(CPS, CPST) Versus the Control for Leaching or Filtering of Copper.
During the clean water application total coliforms were initially leached from the
systems. The control (peat and sand) leached the most total coliforms (950 cfu) while
the compost, paper and sand mixture added the least (90 cfu) to the outflow. Visual
observation in the lab confirmed the compost did have high microbial activity but
testing showed few coliforms. During the composting process heat is generated and
pathogens are killed. Certified compost must meet strict pathogen limits before sale
(US Composting Council 2009; Yost 2008). The runoff water was high in total
coliforms (5,700 to 60,0000 cfu). Additionally, one stormwater application was
spiked with bird droppings (26,000 cfu) and the filtering capacity of the systems was
measured with initial and final outflow samples. The results indicate a high filtering
capacity for total coliforms as presented in Figure 4-7. Removal rates varied between
88% and 92% with the application of 5,700 cfu and between 95% and 99% with
application of 60,000 cfu.
62


Pathogen Leaching and Filtering
60,000 cfu
inflow
Control
HOPS
0CPST
Figure 4-7 Plot of Total Coliform Forming Units Water for Various Accumulated
Sediment Loads Including Tap Water (No Sediment) and Stormwater
with 0.33 and 2.66 kg/m2. The Results Showed that the Control (Peat
and Sand) Leached the Most Coliforms. All Mixtures Showed
Consistently High Pathogen Removal Rates
Suspended sediment is a contaminant also and a goal of PLD is to filter total
suspended solids (TSS). The removal rate of solids from the systems varied between
93% and 100% during test 1 and test 2. The sediment loaded in the lab was increased
to accelerate the clogging process and then related to sediment loads in stormwater in
the field. As presented in Table 4-3, a total of 1.08 kg/m2 at an average of .15 kg/m2
per application were filtered through the columns during test 1 and a total of 6.06
kg/m2 at an average of .5 kg/m2 per application during test 2.
63


Table 4-3 Sediment loading to the soil column surface during the experiment
Number of Stormwater Applications Test 1 Test 2
Average Standard Deviation Accumulative Load Average Standard Deviation Accumulative Load
TSS (kg/m2) TSS (kg/m2)
1 0.10 0.01 0.10 0.16 0.01 0.16
2 0.08 0.01 0.19 0.33 0.02 0.49
3 0.07 0.00 0.26 0.18 0.01 0.66
4 0.07 0.01 0.33 0.33 0.02 0.99
5 0.05 0.02 0.38 0.50 0.02 1.49
6 0.26 0.05 0.64 0.61 0.02 2.10
7 0.44 0.02 1.08 0.55 0.02 2.66
8 0.58 0.01 3.24
9 0.56 0.01 3.80
10 0.57 0.05 4.36
11 0.86 0.08 5.23
12 0.83 0.05 6.06
During testing in the lab it became evident that an additional water quality impact
may come from the filtration mixture. As seen in the screening tests the lighter
particles, such as tires and organics, float. When mixed with sand, the particles are
trapped in the lower layers by the heavier particles and will not float. When the top
layer is subject to mixing from rain the lighter particles are dislodged and float to the
top, as shown in Figure 4-8. Floating material in stormwater, such as leaves, twigs
and cigarette butts combine with the light materials in the filtration mix and may be
carried downstream in the event of overflow or plug the outlet.
64


Figure 4-8 Picture of Floating Particles from the Filtration Mix. The Paper-Tire
Deposit Shows How Density Stratification Occurs When Floating
Particles Settle
With each filling of the basin, light materials float and become incorporated with the
sediment from the stormwater. Density stratification occurs and the upper layer
becomes less dense than the underlying layers. Density tests conducted on soil
samples at 3 depths indicate that the top 0-1 cm of soil become less dense than lower
layers after application of stormwater. Figures 4-9, 4-10 and 4-11 show the
difference in the density of the clean and dirty layers at different depths. The top
layers of the final dirty mixtures (0-1 cm) are less dense than mixture at lower depths.
65


Peat and Sand
Density Stratification
* Clean
Dirty
Depth of Soil (cm)
Figure 4-9 Plot of Density of Soil Samples in Depths 0-10 cm of the Soil Column.
Results Show Lighter Material in the Top Layer of the Soil Shows
Density Stratification of Control
Compost, Paper and Sand Density Stratification
1.60 S 1.40 a 1-2 1 1 00 2 o.0o 0.60 ? 0.40 S 0.20 0.00


X *
Clean Dirty




2 4 6 0 10 Depth of Soil (cm)
Figure 4-10 Plot of Density of Soil Samples in Depths 0-10 cm of the Soil Column.
Results Show Lighter Material in the Top Layer of the Soil Shows
Density Stratification of CPS
66


Compost, Paper, Sand and Tires
Density Stratification
Figure 4-11 Plot of Density of Soil Samples in Depths 0-10 cm of the Soil Column.
Results Show Lighter Material in the Top Layer of the Soil Shows
Density Stratification of CPST
As the water infiltrates, sediment is accumulated in the matrix of the media, creating a
cake layer. This accumulation on the top layers reduces the infiltration rates. It was
observed that this cake layer includes both stormwater sediment and light particles.
The cake layer formed by the sediment accumulation and lighter material is shown
from the density stratification and the sieve analysis. As presented in Table 4-4, the
top layer (0-1 cm) is consistently lighter than the lower layers, indicating lighter
particles floated and settled into the top layers.
Table 4-4 Table of density of soil sample samples at various depth (0-10 cm).
Results show lighter material settled on the top soil layer resulting in
density stratification
Control Compost, Paper, Sand Compost, Paper, Sand, Tires
Soil Sample Depth Density (g/cc)
0-1 cm 1.0 1.2 1.0
4-6 cm 1.3 1.5 1.5
8-10 cm 1.5 1.5 1.6
67


4.4 CLOGGING EFFECTS
The clogging process can be measured from reduction in infiltration rates and
observed from sediment accumulation in the filtration layers. Sediment accumulates
on top of the filtration mixture creating a cake layer. Collecting soil cores from the
columns and a sieve analysis shows evidence of sediment accumulation on the top 1
cm of filtration layer. Stormwater particles are small and the increase in particles
smaller than 75 um after application of stormwater indicates that the sediment
accumulated in the top layers. The sieve analysis as shown by the particle size
distribution plots, Figures 4-12, 4-13, and 4-14, show the change in smaller particles
in the top 0-lcm. The change is significant when viewing the percent of particles
passing the 75um sieve as presented in Table 4-5.
Table 4-5 Table of percent of particles passing 75 um sieve for samples of the
filtration layer at various depths (0-10 cm)
Control Compost, Paper, Sand Compost, Paper, Sand, Tires
Soil Sample Depth Percent Passing 75 um Sieve
0-1 cm 3.2 3.0 3.0
4-6 cm 0.7 1.0 1.6
8-10 cm 0.8 0.8 0.9
68


Control Sieve Analysis of Fine Aggregates
0.01 0.1 1 10
Sieve Size (mm)
|a 0-1 clean e 4-6 clean 8-10 clean 0-1 dirty 4-6 dirty emo dirty]
Figure 4-12 Plot of Sieve Analysis of Samples at Various Depth (0-10 cm) of
Control (Peat and Sand Mixture). Results Show Small Particles in the
Top Soil Layer Indicating Sediment Filtered out of the Stormwater
Accumulates on the Top 1 cm.
69


Control Sieve Analysis of Fine Aggregates
Figure 4-13 Plot of Sieve Analysis of Samples at Various Depth (0-10 cm) of CPS .
Results Show Small Particles in the Top Soil Layer Indicating Sediment
Filtered out of the Stormwater Accumulates on the Top 1 cm.
70


Compost, Paper, Sand and Tires
Sieve Analysis of Fine Aggregates
Figure 4-14 Plot of Sieve Analysis of Samples at Various Depth (0-10 cm) of CPS .
Results Show Small Particles in the Top Soil Layer Indicating Sediment
Filtered out of the StormWater Accumulates on the Top 1 cm.
Sediment accumulation on the top layer causes a reduction in infiltration rate. The
change of infiltration rate was measured throughout the continuous application of
stormwater. For mathematic convenience, the reduced infiltration rate,/* is
normalized by/c the Hortons constant infiltration rate defined in equation 4-3,4-4
and 4-5. The accumulative sediment load, Ls, is expressed as weight of sediment per
unit area in kg/m2.
71


The decay of infiltration is plotted in Figure 4-15 for various sub-base mixtures. The
relationship can be depicted by an exponential decay function between fffc and Ls
[kg/m2].
Equation 4-6
fs
Peat and sand : = -4.449e
f
J c control
-0.1361/,
Equation 4-7
Compost, paper and sand: = 2.921e~'m911
/
c cps
Equation 4-8
/
Compost, paper, sand and tires '
fc
= 2.8303e_00814,J
cpst
2
Where/i is the infiltration rate [cm/hr] after accumulative sediment load Li5[kg/m ]
and fc is Hortons constant infiltration rate [cm/hr]. The subscripts, control, cps and
cpst, indicate the media-mixture associated with each equation. Where^
V fc control IS the
constant infiltration rate [cm/hr] of the control (peat and sand) mixture. fc cps is the
infiltration rate [cm/hr] of the compost, paper and sand mixture and the/ccpst is the
infiltration rate [cm/hr] of the compost, paper, sand and tires mixture. Equations 4-6,
4-7 and 4-8 have correlation coefficients r = 0.89, 0.93, 0.92 respectively.
72


Clogging Rate
.. o
IS.
5
o
o
o
5
o
>
TJ
+->
(0
ro 5
DC (0
* Peat and Sand Compost, Paper, Sand x Compost, Paper, Sand, Tires
Figure 4-15 Plot of Reduced Infiltration Rate, fs, Normalized by fc (Hortons
Constant Infiltration Rate) Versus the Accumulative Sediment Load for
Various Media Mixtures. Results Show Reduction of Infiltration Rate
with Accumulative Sediment Load
The minimal infiltration rate according to the local stormwater criteria for PLD is 2.5
cm/hr (1 in/hr). Therefore the system is considered clogged when fs becomes less
than 2.5 cm/hr.
4.5 DESIGN EXAMPLE
As indicated in Figure 4-15, infiltration through the filtering layers is closely related
to the accumulative amount of sediment loaded onto the infiltrating bed. The design
example intends to illustrate how to interpret the accumulative sediment load into the
basins operation. The annual sediment yield generated from the tributary area can be
estimated by the annual mean event concentration of sediment and the annual runoff
volume as:
73


Equation 4-9
^ = CPAributary
Equation 4-10
LS=CSV
In which V is the annual runoff volume in [L3], C is the runoff coefficient, P is the
annual rainfall depth in [L], A tributary is the watershed tributary area [L2], C$ is the
mean sediment concentration [M/L3], and Ls is the annual sediment load [M/L3]. In
practice, the PLD is designed to intercept a portion of the tributary area. As a result,
the annual loading to the PLD is estimated as:
Equation 4-11
A
tributary
a ---------
A
nPLD
Equation 4-12
Bs = CsaCP
In which a is the ratio of the tributary area intercepted by PLD, Apld is the surface
area of the PLD [L2] and Bs = annual unit-area sediment load to the basin of the PLD
[M/L /year]. Figure 4-15 shows the decay of the infiltration rate with respect to the
accumulative sediment load, Ls, which can be converted into the PLDs service years
as:
Equation 4-13
Bs
Where N is the number of PLDs service years, Ls is the accumulative sediment load
into laboratory tested PLD [M/L ] and Bs is the annual accumulative sediment load
into the basin of interest [M/L2/year]. Equation 4-13 assists the engineer to convert
Figure 4-15 into any PLDs service years for an investigation of the life-cycle
operation.
For example, a PLD in Denver is designed based on the local requirements
(USWDCM 2001) using the average annual rainfall in Denver, and average sediment
74


concentration observed in Colorado (Doefer and Urbonas 1993). The example PLD
has a surface detention capacity up to a water depth of 0.305 m (12 inches). It will
capture and treat runoff from a parking lot. The ratio (a) of the parking lot area as
the contributing watershed to the PLD is 20 to 1. The event mean sediment
concentration (Cs), TSS, in runoff from commercial areas in Colorado is recorded as
240 mg/L. Annual precipitation (P) in Denver area is .4 meters (15.4 in). Aided by
equation 12 with Cs = 240 mg/L, a=20, C=0.9 for parking lot, and P=0.4 m, the
annual unit-area sediment load to the example PLD is calculated as:
Equation 4-14
Bs = (240mg //)(20)(.9)(.4m) = 1.728kg / m2
The accumulative sediment load (Ls) on the x-axis in Figure 4-15 can then be
converted into years of service for the example PLD. Figure 4-16 presents the
reduction in infiltration over time for the example PLD. The decay of infiltration rate
fs down to 2.5 cm/hr (1 in/hr) varies for each sub-base mixture based on the Hortons
constant infiltration rate fc. For this example, the PLD is considered clogged at
fjfc=2.5/fc.
As shown in Figure 4-2 and equations 4-2, 4-3 and 4-4, fc is 9.9 cm/hr for the control
(peat and sand), 14.5 cm/hr compost, paper and sand and 9.1 cm/hr compost, paper,
sand and tires. Therefore the PLD is considered clogged in Figure 4-16 where fjfc is
2.519.9 for the control, 2.5/14.5 for compost, paper and sand and 2.5/9.1 for compost,
paper, sand and tires.
75


Clogging Rate
o
= $>
s
O
o
o
5
o
5"
T3
O
+-i
<0
2
CC (0
5
4
3
2
1

A
A
A A A
Jr
t*. *A
+ +s -k V

X
!
*
Control ^=0.89
CPS r2 = 0.93
CPST r2 = 0.92
0 5 10 15
Time (years)
a Peat and Sand Compost, Paper, Sand + Compost, Paper, Sand, Tires
Figure 4-16 Plot of Reduced Infiltration Rate,/i? Normalized by/c (Hortons
Constant Infiltration Rate) Versus Time for Various Media Mixtures.
Results Show Reduction of Infiltration Rate Over Time for Example
PLD
4.6 CONCLUSION
The infiltration tests indicated that improvements in water quality result in sediment
accumulation on the PLDs bottom, and clogging through the sub-base filtering media
over time. All three mixtures filtered TKN, phosphorous and copper from the system
during the stormwater applications. Removal rates for TKN, TP, and Copper varied
from 32% to 77%, 48% to 86%, and 48% to 95% respectively. The pathogen removal
rates were between 88% and 99%. Sediment removal rates remained high throughout
the two tests, 93% to 100%. Accumulation of sediment creates a cake layer on top of
the filtration layer. Sieve analysis indicated the cake layer formed in all three
treatments was on the top surface up to 1 cm. Additionally, the cake layer consists of
both solids from storm water and floating particles from the PLDs media mixture.
76


Individual components of media perform differently and mixing may alter the
characteristics of the system. Shredded tires float and are hydrophobic. When mixed
with sand the layers become compacted and are held in place by the sand but tires
from the top layer still float. Dry organic materials such as peat and compost are
buoyant as well. Shredded paper floats initially and will sink as they become absorb
water. These materials may affect the cake layer, the outlet and possibly the
downstream waterways.
A model has been developed to approximate the clogging of a PLD in the field based
on sediment loading. The model can be used to estimate the life span and
maintenance of the PLD. A design example indicates clogging occurs when the
infiltration rates is reduced to the minimal 2.5 cm/hr (1 in/hr) which occurs in 11 to
16 years. Replacement of the top 10 cm of material is suggested to renew the
infiltration rate.
77


5 BENCH SCALE TEST WITH VEGETATION
5.1 INTRODUCTION AND BACKGROUND
The PLD is designed to treat stormwater through a vegetated basin and soil-mix
filtration layer (USWDCM 2001). The previous chapter defined the possible media-
mixes for the filtration layer to achieve stormwater treatment goals. This chapter
investigates the impact of vegetation on the performance of the system. The goals
of this chapter are to: 1. Ensure the ability of the filtration mixture to support plant
growth. 2. Evaluate the impact of vegetation on performance and 3. Evaluate the
possibility of the PLD as a source of water quality impact.
Plants have been shown to reduce contaminates in water and soil through
degradation, uptake and other biochemical mechanisms (Langergraber 2005; Munch
et al. 2005; Schoonover et al. 2005; Sirivedhin and Gray 2006; Vidon and Hill 2004).
Many engineered systems such as wetlands and phytoremediation function
successfully due to the intrinsic treatment capacity of plants (Elodie et al. 2009;
Gottschall et al. 2007; Huett et al. 2005; Kohler et al. 2004; Syversen 2005)
Additionally, the roots of plants create macropores in the soil, creating water
pathways. Both the treatment capacity and the infiltration rate are affected by the
addition of plants in the PLD.
Permissible filtration soil-mixtures from previous investigations were tested with and
without vegetation. The ability of each of the mixtures to support plant growth was
verified and compared through quantitative plant counts and qualitative observations
in the lab. The effect of the plant growth on the performance in terms of water
quality impacts and infiltration rate was assessed. The possibility of an additional
water quality impact of floating particles from the filtration media, seed mixture and
decaying plant material was observed in the lab.
78


5.2 BENCH SCALE TEST METHODOLOGY
Building on previous chapters, which defined the permissible filtration layer mixtures
and assessed the performance of the system without vegetation (bare soil), this
chapter adds vegetation. Testing was conducted in the same 38 cm (15 in) diameter
infiltrometers as previous tests. The columns were packed to replicate the current
recommendation of 45.7 cm (8 in) of filtration soil mix and 20.3 cm (8 in) of
aggregate under drain (Guo et al 2009). Sediment laden stormwater was added to the
top of the columns with bare soil creating a cake layer. The filtration layers were:
1. Control (15% peat and 85% sand)
2. CPS (7.5%compost, 7.5% shredded paper and 85% sand)
3. CPST (7.5% compost, 7.5% shredded paper, 8% tires and 77% sand)
Then grass seeds were germinated on top of the cake layer created from buildup of
sediment filtered out of the stormwater. The roots of the grass grew into the cake
layer and performance criteria were measured and compared to results without
vegetation (bare soil conditions).
The bench scale test consisted of the five steps in Figure 5-1 to compare un-vegetated
(bare soil) and vegetated conditions. Previous steps 1 and 2 provided data for the un-
vegetated condition. Step 3 began with germinating grass seeds in the cake layer
created from the sediment buildup in step 2. Grass was allowed to grow and the same
experimental procedures, adding stormwater, measuring flow rates and water quality
parameters were followed. The vegetation eventually died leading to additional
results with dead grass (without vegetation), steps 3 and 4. Grass seeds were again
planted on top of the cake layer with dead grass and additional information about the
effect on infiltration rate was collected.
79


Step 1- Initial Condition Bare Soil- No Vegetation
72-hour elevated outflow and clean water to measure infiltration capacity (fc) and calculate of
hydraulic conductivity
Lower the outflow and add stormwater for unsaturated field conditions
Top layers were excavated and sieve analysis was performed
Step 2 Duplicate Bare Soil Test
The top soil-mix layers were replaced
Begin with 72-hour lowered outflow and measure infiltration rate
Then lower the outflow and add stormwater for unsaturated field conditions
Step 3 Effect of Vegetation First Planting of Grass
Germinate grass seeds on top of cake layer
Measure restoration of infiltration rate with 72-hour lowered outflow
Then lower the outflow and add stormwater for unsaturated field conditions
Step 4 Dead Vegetation
Continue to add stormwater with sediment until the grass is choked and dies
Measure flow rates with dead grass
Step 5 Replant Vegetation Second Planting of Grass
Replant grass seeds on top of cake layer created in step 4
Then lower the outflow and add stormwater for unsaturated field conditions
Measure regeneration of flow rates
Figure 5-1 Steps for Bench Scale Testing
Stormwater was applied to the columns before and after the vegetation was growing.
The flow rate and water quality parameters were measured throughout the entire
experiment to compare treatments. Between each step of the testing an initial
condition for flow rate was measured by adding tap water at a consistent depth for 72
hours. The infiltration rate after 72 hours corresponds to the infiltration capacity in
Hortons infiltration equation (Horton 1993). Hortons estimate for infiltration rate
over time is based on the equation:
Equation 5-1
/(0 = fc -Kf0-fc)e(~k)<
Where f(t) is the infiltration rate at time t,f0 is the initial infiltration rate [L3/t] and/c
is the infiltration rate at field capacity [L3/t]. The k is a constant with units 1/time.
80


Synthetic stormwater was added during each step to measure filtering of contaminates
and create cake layer on top of the filtration layer. The saturated flow rate was again
monitored for 72 hours to conclude steps 1 and 2. Grass seeds were then planted on
the cake layer created during step 2 and germinated. Again 72 hour clean water flow
was conducted to compare the effect when the roots penetrated the cake layer.
Sediment laden stormwater was added after the grass was growing in step 3. The
stormwater killed the grass leading to step 4 and the cake layer thickened. A final
test, step 5, was conducted by again adding grass seeds directly on top of the dead
grass without disturbing the cake layer. Stormwater was again added to the top of the
system to complete step 5.
Mixing runoff water with additional sediment created the synthetic stormwater.
Runoff water from an outfall to the South Platte River in the City of Denver was
collected and transported to the lab. The water was spiked with additional sediment
to accelerate the clogging process in order to relate surface loading to field conditions
where the clogging process occurs over years. The sediment-laden stormwater was
applied to the columns with the same procedure throughout steps 1 through 5.
Sediment and water were mixed continuously while being pumped on top of the
column. Water quality samples were collected from the inflow and outflow.
Stormwater was applied until a cake layer was formed.
The seed mixture was the same mixture recommended by the UDFCD and used in the
batch test and recommended in Volume 3 Criteria Manual for the Denver
Metropolitan area (USWDCM 2001) presented in Table 5-1. Seeds were weighed in
relative amounts for the surface area of each column and presented in Table 5-2. A
discussion with horticulturalist at Botanic Gardens, Mark Fusco indicated the use the
grasses would adequately represent the germination and growth of the entire mixture.
Native seeds also require a cold period before germination, and therefore were
refrigerated for 30 days before sowing in the columns. The numbers of seeds were
also doubled to ensure good coverage.
81


-a*
Table 5-1 PLD vegetation seed mixture as prescribed in UDFCD criteria manual
COMMON NAME SCIENTIFIC NAME VARIETY PLSLbs per Acre Ounces per Acre
Sand bluestem Andropogon hallii Garden 3.5
Sideoats grama Bouteloua curtipendula Butte 3
Prairie sandreed Calamovilfa longifolia Goshen 3
Indian ricegrass Oryzopsis hymenoides Paloma 3
Switchgrass Panicum virgatum Blackwell 4
Western Wheatgrass Pascopyrum smithii Ariba 3
Little bluestem Scbizachyrium scoparium Patura 3
Alkali sacaton Sporobolus airoides 3
Sand dropseed Sporobolus cryptandrus 3
* Pasture Sage Artemisia frigida 2
Blue aster Aster laevis 4
* Blanket flower Gaillardia aristata B
* Prairie coneflower Ratibida columnifera 4
* Purple prairiedover Dalea (Petalostemum) purpurea 4
Sub-Totals: 27.5 22
Total lbs per acre: 28.9
(Source:USWDCM 2001)
Table 5-2 Amount and type of grass seeds planted in each column
Seed Name Amount (lbs/acre) seeds/lb container size (sq ft) Weight mg/column Seeds Planted mg/column
sideoats gramma 3 191,000 1.23 38.3 76.6
prarie sandreed 3 274,000 1.23 38.3 76.6
swtichgrass 4 270,000 1.23 51.0 102.1
western wheatgrass 3 110,000 1.23 38.3 76.6
little bluestem 3 260,000 1.23 38.3 76.6
sand dropseed 3 825,000 1.23 38.3 76.6
Grass seeds were spread on top of the cake layer, germinated and allowed to grow.
As shown in Figure 5-2 grow lights were installed above the columns and a timer was
set. The seeds were watered every other day and monitored for growth. The numbers
of plants in a 4 cm by 4 cm square were counted in three separate locations in each of
the columns. The numbers of plants were reported for day 20 and 25. No additional
plants were germinating by day 20 and by day 25 the plants had grown tall enough to
take the application of water. The effect of the vegetation on the infiltration rate was
82


Full Text

PAGE 1

SUSTAINABLE DESIGN OF URBAN POROUS LANDSCAPE DETENTION BASINS by Shauna Marie Kocman B.S., Environmental Engineering, Colorado State University, 2001 M.S., Civil Engineering, University of Colorado at Denver, 2006 A thesis submitted to the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Doctor of Philosophy Civil Engineering, Sustainable Urban Infrastructure Project 2010

PAGE 2

2010 by Shauna Marie Kocman All rights reserved.

PAGE 3

This thesis for the Doctor ofPhilosophy degree by Shauna Marie Kocman has been approved by Anu Ramaswami Rajagopalan Balagi April 16, 2010

PAGE 4

Kocman, Shauna Marie (Ph.D., Civil Engineering) Sustainable Design of Urban Porous Landscape Detention Basins Thesis directed by Professors James C.Y. Guo and Anu Ramaswami ABSTRACT Porous Landscape Detention basins (PLD) capture and filter stormwater from micro rainfall events while taking advantage of the intrinsic quality of plants to act as water treatment systems. The current design recommendations leave opportunity for the incorporation of waste symbiosis and holistic design concepts. This thesis investigates the physical, chemical and biological performance of the currently recommended PLD design and two waste-incorporated designs. The beneficial reuse of urban waste stream materials into sustainable filtration mixes is assessed and paired with environmental life cycle analysis. Results of waste screening tests indicate that the currently recommended 15% peat and 85% sand (control) may be replaced by two viable waste-incorporated mixes 1) Compost-Paper Sand (CPS) mix at 7.5%, 7.5% and 85%, respectively, and, 2) Compost-Paper-Sand Tires (CPST) mix at 7.5%, 7.5%, 76% and 8% respectively. Bench scale tests compared the water flow and water filtration capabilities of the CPS and CPST with the control first in un-vegetated and then in vegetated systems On-vegetated test results showed no significant difference between CPS and control in terms of infiltration rates, flow attenuation over time (with sediment loading of 6 kg/m2), and water quality tested for nutrients (TKN, N02+N03 and TP), pathogens (total coliforms) and total metals (copper, lead and zinc). Although without vegetation, the CPST did show less flow attenuation over time indicating a longer life span, this advantage disappeared in vegetated systems With vegetation, all three systems performed similarly in terms of flow attenuation over time (sediment loading of 13 kg/m2 equal to 5 years in an example PLD). Vegetation is an added benefit as it increases nutrient removal and the lifespan of the PLD. An example PLD is used to evaluate the lifespan, cost savings and environmental impact of installing one of the waste-incorporated mixes instead of the control. With vegetation, all three systems require replacement due to infiltration rates below the

PAGE 5

minimum 2.5 cm/hr in 14 to 16 years. The economic first cost analysis and environmental LCA of the example site shows that installation of the waste incorporated CPS mix saves $2,800 and 6.6 MTC02E. Vegetated PLD is successful at stonnwater treatment including the removal of sediment (93%-100% removal), nutrients (57%-98% removal), metals (83%-99% removal) and pathogens (87%-99% removal). Replacing the currently used peat and sand mixture with the waste-incorporated CPS mix is recommended based on environmental benefit and performance criteria. In locations which are highly sensitive to phosphorous, the peat and sand mixture should still be installed due to higher removal rates (98%). Because tires are buoyant and may overflow the PLD contaminating downstream waterways, more research and risk assessment is necessary before using shredded tires. This abstract accurately represents the content of the candidate's thesis. I recommend its publication. Signed __ James C.Y. Guo Signed Anu Ramswami

PAGE 6

DEDICATION PAGE I dedicate this thesis to my parents and my grandparents, who gave me an appreciation of learning and taught me the value of perseverance and resolve. I also dedicate this to my husband and my 3 children for their support and understanding while I was completing this thesis.

PAGE 7

ACKNOWLEDGMENT Thank you to my advisors, James C. Y. Guo and Anu Ramaswami, for their contribution and support to my research. Their insights and continual guidance was essential in completing this monumental task. I also wish to thank all the members of my committee for their valuable participation and insights. Gratitude goes to the babysitters that allowed me to study, work in the lab and write my dissertation my parents, my brother Daniel, Lauren, my cousins Jeremy and Cedar, Viola and others too numerous to mention. This research has been supported by funding from the Department of Education GAANN grant, ARCS Scholarship, CCHE Waste-to-Value grant, the Urban Drainage and Flood Control District and the Urban Water Research Institute. Thank you to Ken Mackenzie from UDFCD and Ben Urbonas from UWRI who have given their time and expertise to this research. Additionally, materials and time have been donated by the AcuGreen, Denver Botanical Gardens and the City and County of Denver Health Department.

PAGE 8

TABLE OF CONTENTS FIGURES ....................................................................................................... xvi TABLES ................................................................................................... xvii 1 SUSTAINABLE URBAN INFRASTRUCTURE .......................................... 1.1 SUSTAINABILITY .................................................................................... 1 1.2 URBAN INFRASTRUCTURE ................................................................... 1 1.3 STORMWATER MANAGEMENT ........................................................... 2 1.3.1 LANDSCAPE AND STORMWATER RUNOFF ...................................... 2 1.3.2 BMP FOR COLORADO UDFCD .............................................................. 3 1.3.3 PREVIOUS STUDIES ................................................................................ 5 1.4 OBJECTIVES .............................................................................................. 7 2 WASTE SCREENING AND ENVIRONMENTAL BENEFIT ............... s 2.1 INTRODUCTION AND OBJECTIVE ....................................................... 8 2.2 LITERATURE REVIEW ON MATERIALS SUBSTITUTION ................ 9 2.3 METHODOLOGY .................................................................................... 11 2.3 .1 LEACH TEST ........................................................................................... 12 2.3.2 FLOW RATE TEST .................................................................................. 13 2.3.3 GERMINATION TEST ............................................................................. 14 viii

PAGE 9

2.3.4 ENVIRONMENTALLCA ........................................................................ 15 2.4 RESULTS .................................................................................................. 17 2.4.1 LOCAL AVAILABILITY AND COST .................................................... 17 2.4.2 LEACHING CRITERIA ........................................................................... 19 2.4.3 FLOW RATES .......................................................................................... 22 2.4.4 CONFIRMATION TESTGERMINATION ............................................ 24 2.5 SCREENED BEDDING MATERIAL MIXTURE ................................... 25 2.6 GREENHOUSE GAS BENEFIT ............................................................... 27 2.7 CONCLUSION .......................................................................................... 30 3 INFILTROMETER DESIGN AND FLOW MODEL ............................... 32 3.1 INTRODUCTION ..................................................................................... 32 3.2 SURFACE STORAGE BASIN ................................................................. 33 3.3 SUBSURFACE FILTERING SYSTEM .................................................... 36 3.4 OPERATION OF POROUS LANDSCAPING BASIN ............................ 39 3.5 LABORATORY TESTS ........................................................................... 40 3.5.1 PREVIOUS DESIGN STUDIES .............................................................. .40 3.5.2 INFILTROMETER DESIGN AND TESTING ........................................ .41 3.6 DESIGN EXAMPLE AND SCHEMA TICS ............................................ .48 3.7 CONCLUSION .......................................................................................... 52 4 WASTE-INCORPORATED BENCH SCALE TESTBARE SOIL. .. 53 IX

PAGE 10

4.1 INTRODUCTION AND BACKGROUND .............................................. 53 4.2 BENCH SCALE TEST METHODOLOGY .............................................. 54 4.3 WATERQUALITYIMPACTS ................................................................ 60 4.4 CLOGGING EFFECTS ............................................................................. 68 4.5 DESIGN EXAMPLE ................................................................................. 73 4.6 CONCLUSION .......................................................................................... 76 5 BENCH SCALE TEST WITH VEGETATION .......................................... 78 5.1 INTRODUCTION AND BACKGROUND .............................................. 78 5.2 BENCH SCALE TEST METHODOLOGY ............................................. 79 5.3 VEGETATION .......................................................................................... 84 5.4 WATER QUALITY IMPACTS ................................................................ 89 5.4.1 NUTRIENT REMOVAL ........................................................................... 92 5.4.2 METALS REMOVAL ............................................................................... 96 5.4.3 PATHOGEN REMOVAL ......................................................................... 97 5.4.4 SUSPENDED PARTICLES ...................................................................... 99 5.5 CLOGGING EFFECTS ........................................................................... 103 5.6 CONCLUSIONS ..................................................................................... 108 6 CONCLUSIONS AND RECOMMENDATIONS ..................................... 110 APPENDIX A ................................................... ........................................... 112 REFERENCES ........................................................................................... 113 X

PAGE 11

FIGURES Figure 1-1 Picture of PLD Aurora, CO ................................................................................... 4 1-2 PLD Cross Section ................................................................................................ 4 2-1 Batch Test Setup for Leaching and Flow Rate .................................................... 13 2-2 Batch Test Setup and Outflow for Collection of Leachate and Flow Rate ......... 13 2-3 Picture of Batch Test Setup for Germination Rate. Grass Seeds Were Planted in Various Media Mixtures, Watered and Set by a Sunny Window ........................ 15 2-4 Reduction of Copper Leaching From Various Mixes After Multiple Flushes of Water. The Gray Area Indicated the Permissible Waste Mixture Passing the Metals Leaching Test ........................................................................................... 22 2-5 Effect of Varying Amounts of Compost and Paper on Flow Rate Shows Compost Increases Flow rate and Paper Slows the Flow Rate ............................ 23 2-6 Increase in Flow Rate is Related to the Amount of Shredded Tires ................... 24 2-7 Germination Tests Confirmed no Difference in the Ability of the Four Mixes to Support Vegetation .............................................................................................. 25 2-8 Picture of Control and Permissible Waste-Incorporated Media Mixtures for PLD ............................................................................................................................. 27 2-9 Net GHG Benefit of Installing the CPST Waste-Incorporated Mix Instead of the Business as Usual (Control) in an Example PLD ................................................ 29 2-10 Cost Savings to Installing the CPST WasteIncorporated Mix Instead of the Business as Usual (Control) in an Example PLD ................................................ 30 3-1 Layout of Porous Landscaping Basin .................................................................. 33 3-2 Stormwater Quality Control Volume for Porous Landscaping Basin Design ..... 35 Xl

PAGE 12

3-3 Illustration of Infiltrometer Operation under Saturated Conditions ................... .37 3-4 Soil Column Design with 2-Layered System and Lowered Outflow for Field Conditions ............................................................................................................ 42 3-5 Soil Column Design with Elevated Outflow for Saturated Conditions .............. .43 3-6 Infiltrometers Built in Laboratory ....................................................................... 44 3-7 Measurements for Soil and Water Levels Inside the Soil Column ..................... 44 3-8 Geotextile between Large Aggregate and Filter Layer Inside the Column ........ .45 3-9 Manometer Setup to be Placed inside the Column ............................................. .45 3-10 Variation of Infiltration Rates for Sample Columns (Control: Peat and Sand )46 3-11 Comparison between Observed and Calculated Filtering Thickness ................ 51 4-1 Setup of Triplicate Soil Columns and Barrels of Storm water in the Lab ............ 55 4-2 Variation of Infiltration Rates as the Filtration Mixtures Approach Saturation after 72 hours of Continuous Flow ...................................................................... 56 4-3 Outfall N-431E in the South Platte River Where Stormwater for the Experiment was Collected ....................................................................................................... 58 4-4 Plot of TKN Concentration in Water for Various Accumulated Sediment Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66 kg/m2. The Results Showed No Consistent Statistical Difference Between the Waste Incorporated Media Mixes (CPS, CPST) Versus the Control for Leaching or Filtering of TKN .................................................................................................. 60 4-5 Plot of Total Phosphorous Concentration in Water for Various Accumulated Sediment Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66 kg/m2. The Results Showed No Consistent Statistical Difference Between the Waste-Incorporated Media Mixes (CPS, CPST) Versus the Control for Leaching or Filtering of TP ............................................................................ 61 Xll

PAGE 13

4-6 Plot of Total Copper Concentration in Water for Various Accumulated Sediment Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66 kg/m2. The Results Showed No Consistent Statistical Difference Between the Waste-Incorporated Media Mixes (CPS, CPST) Versus the Control for Leaching or Filtering of Copper. ......................................................................................... 62 4-7 Plot of Total Coliform Forming Units Water for Various Accumulated Sediment Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66 kg/m2. The Results Showed that the Control (Peat and Sand) Leached the Most Coliforms. All Mixtures Showed Consistently High Pathogen Removal Rates 63 4-8 Picture of Floating Particles from the Filtration Mix. The PaperTire Deposit Shows How Density Stratification Occurs When Floating Particles Settle ........ 65 4-9 Plot of Density of Soil Samples in Depths 0-10 em of the Soil Column. Results Show Lighter Material in the Top Layer of the Soil Shows Density Stratification of Control ............................................................................................................. 66 4-10 Plot of Density of Soil Samples in Depths 0-10 em of the Soil Column. Results Show Lighter Material in the Top Layer of the Soil Shows Density Stratification ofCPS .................................................................................................................. 66 4-11 Plot of Density of Soil Samples in Depths 0-10 em of the Soil Column. Results Show Lighter Material in the Top Layer of the Soil Shows Density Stratification ofCPST ............................................................................................................... 67 4-12 Plot of Sieve Analysis of Samples at Various Depth (0-1 0 em) of Control (Peat and Sand Mixture). Results Show Small Particles in the Top Soil Layer Indicating Sediment Filtered out of the Storm water Accumulates on the Top 1 em ........................................................................................................................ 69 4-13 Plot of Sieve Analysis of Samples at Various Depth (0-1 0 em) of CPS Results Show Small Particles in the Top Soil Layer Indicating Sediment Filtered out of the Stormwater Accumulates on the Top 1 cm .................................................... 70 4-14 Plot of Sieve Analysis of Samples at Various Depth (0-1 0 em) of CPS Results Show Small Particles in the Top Soil Layer Indicating Sediment Filtered out of the Storm Water Accumulates on the Top 1 em ................................................... 71 xiii

PAGE 14

4-15 Plot of Reduced Infiltration Rate, fs, Normalized by fc (Horton's Constant Infiltration Rate) Versus the Accumulative Sediment Load for Various Media Mixtures. Results Show Reduction of Infiltration Rate with Accumulative Sediment Load ..................................................................................................... 73 4-16 Plot of Reduced Infiltration Rate, fs, Normalized by fc (Horton's Constant Infiltration Rate) Versus Time for Various Media Mixtures. Results Show Reduction of Infiltration Rate Over Time for Example PLD .............................. 76 5-1 Steps for Bench Scale Testing ............................................................................ 80 5-2 Picture of Grow Lights above Columns for Growing Grass in the Lab .............. 83 5-3 Picture of Un-Germinated Seeds Floating in Water above Soil Surface ............. 84 5-4 Plot of Plant Counts 25 Days After Planting the seeds in the Cake Layer. Results Indicate no Difference between the Treatment Groups ....................................... 85 5-5 Picture of Healthy Grass Growing in the Column 25 Days after 1st Planting ..... 86 5-6 Picture of Soil Surface after the 1st Planting of Grass was Choked and Died ..... 86 5-7 Picture of Grass in the Control (Peat and Sand) Column after 2 Months without Water. The Picture Shows the Stunted Growth of Vegetation in Control (Peat and Sand) After a Dry Period .............................................................................. 87 5-8 Picture of Grass in the CPS Column after 2 Months without Water. Healthy Vegetation in CPS after a Dry Period .................................................................. 87 5-9 Picture of Grass in the CPS Column after 2 Months without Water. Healthy Vegetation in CPS after a Dry Period .................................................................. 87 5-10 Average Number of Plants in the Columns with 2 Watering Schemes: Watering Every Other Day and without Water for 2 Months ............................................. 88 5-11 Plant Height after 2 Months without Water Shows CPS Supports Plant Growth through a Dry Period ........................................................................................... 88 5-12 Plot of Removal Rate for Nutrients and Metal with and without Vegetation. Results Show Increased Percent Removal of Contaminants with Vegetation .... 90 xiv

PAGE 15

5-13 Plot of TKN Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Vegetation does not Effect TKN Removal .......................................................... 93 5-14 Plot of N03+N02 Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Consistently Higher Concentrations in Outflow from the CPS than the Control Mix ....................................................................................................................... 94 5-15 Plot of Total Phosphorous Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Vegetation Increased the Removal Rate ofTP from the Control Mix .. 95 5-16 Plot of Percent Removal of Total Phosphorous for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Increased Removal Capacity for TP with Vegetative Conditions ......... 95 5-17 Plot of Total Copper Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Vegetation does not Effect Copper Removal.. ....................................... 96 5-18 Plot of Total Copper Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Breakthrough of Zinc in CPST Mixture ................................................ 97 5-19 Plot of Total Coliform Forming Units in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed All Mixes Were Continually Successful at Filtering Pathogens from Stormwater throughout the Experiment.. ............................................................. 98 5-20 Picture of Aoating Particles from the Filtration Mixture Plugging the Overflow in the Lab ........................................................................................................... 100 5-21 Picture of Light Particles which Overflowed a PLD ....................................... 100 5-22 Plot of the Amount of Particles (TSS) from the Various Filtration Mixes Which Were Found to Aoat in Clean Water. Results Indicate as Much as 2,000 mg/1 TSS May Overflow the PLD ............................................................................. 101 5-23 Aoating Particles in Water Samples of Overflow from the Soil Columns ..... 102 XV

PAGE 16

5-25 Particles on the Paper of Filtered Samples from Overflow after Grass was Growing. The Samples Show that Peat, Compost and Tires Particles Float and May Overflow the PLD ..................................................................................... 103 5-26 Plot of Infiltration Capacity of Various Media Mixtures Before and After Vegetation is Growing in the Cake Layer. Regeneration of Infiltration Capacity from Growth of Vegetation Shows the Control and CPS Benefited more than CPST .................................................................................................................. 104 5-27 Plot of Saturated Infiltration Capacity of Various Media Mixtures Before and After Vegetation is Growing in the Cake Layer. Results Show Increase in Saturated Infiltration Capacity after Grass was Germinated into the Cake Layer ........................................................................................................................... 105 5-28 Plot of Reduced Infiltration Rate, f5 Normalized by fc (Horton's Constant Infiltration Rate) Versus Accumulative Sediment Load for Various Media Mixtures, with and without Vegetation. Results Show Regeneration of Flow Rate in Control and CPS Mixes after Grass is Germinated in the Cake Layer .106 5-29 Plot of Reduced Infiltration Rate, f5 Normalized by fc (Horton's Constant Infiltration Rate) Versus Time for Various Media Mixtures, with and without Vegetation. Results Show Vegetation's Increases the Time to Clogging in the Control and the CPS Mixes ............................................................................... 1 07 XVl

PAGE 17

TABLES Table 1-1 Existing PLD research ........................................................................................... 6 2-1 Summary of literature on PLD waste stream replacements ................................ 11 2-2 Possible local waste streams to replace virgin materials for incorporation in PLD ............................................................................................................................. 12 2-3 PLD vegetation seed mixture prescribed in UDFCD criteria manual.. .............. .l4 2-4 Number of seeds planted in each batch test ........................................................ 15 2-5 Data sources to calculate GHG emissions impact.. ............................................. 17 2-6 Cost comparison of virgin and waste materials ................................................... 19 2-7 Leachate pH from control and waste incorporated media mixtures after 48 hours contact time .......................................................................................................... 19 2-8 Leachate pH from crushed concrete .................................................................... 20 2-9 Leachate pH from media mixtures on top of crushed concrete ........................... 20 2-10 Leaching results of various media mixtures indicate that three mixtures pass the leaching test ......................................................................................................... 21 2-11 Permissible ranges of waste-incorporated filtration mixes ............................... 26 2-12 GHG benefit in MTC02E of replacing business-as-usual (control) with a waste incorporated CPST mix ............................................................................. 28 3-1 Large column studies of systems similar to PLD ............................................... .41 3-2 Compaction and density of soil-mix sample ...................................................... .46 3-3 Variation of hydraulic heads measured at 5 stations .......................................... .47 XVll

PAGE 18

3-4 Hydraulic conductivity in filtration layer after 72 hours of water flow ............. .47 3-5 Observed case Studies for bio-retention medium (Li and Davis 2008a and 2008b) .................................................................................................................. 51 4-1 Average water quality results of stormwater outfall N-431E .............................. 58 4-2 Average event mean concentrations for Denver area .......................................... 59 4-3 Sediment loading to the soil column surface during the experiment .................. 64 4-4 Table of density of soil sample samples at various depth (0-1 0 em). Results show lighter material settled on the top soil layer resulting in density stratification ......................................................................................................... 67 4-5 Table of percent of particles passing 75 urn sieve for samples of the filtration layer at various depths (0-10 em) ........................................................................ 68 5-1 PLD vegetation seed mixture as prescribed in UDFCD criteria manual ............ 82 5-2 Amount and type of grass seeds planted in each column .................................... 82 5-3 Germination slows after 20 days as Sshown by the average number of plants 20 and 25 days after planting .................................................................................... 84 5-4 Table of contaminant removal rate comparing un-vegetated and vegetative conditions. Results show vegetation increases the nutrient removal rate ........... 90 5-5 Contaminant concentrations in inflow and outflow water through various media mixtures with and without vegetation compared to the EPA freshwater criteria ............................................................................................................................. 92 5-6 Percent removal of total coliform forming units without and with vegetation. The results show high percent removal of pathogens from stormwater .............. 98 5-7 Accumulative sediment loading to the soil surface during the experiments with and without vegetation ......................................................................................... 99 xviii

PAGE 19

1 SUSTAINABLE URBAN INFRASTRUCTURE 1.1 SUSTAINABILITY In a 1987 report to the United Nations, Gro Brundland defined sustainable development as "development that meets the needs of the present without compromising the ability of the future generations to meet their own needs" (WCED 1987). Sustainable urban infrastructure systems therefore are those that provide for the needs of urban residents in both the short and long term while also maintaining the environmental, social and economic resources. The combination of these resource parameters to represent sustainability was coined the Triple Bottom Line for businesses by John Elkington (Elkington 1998). At the 2002 World Summit on Sustainable Development in Johannesburg, South Africa, these same parameters were designated as People, Prosperity and the Planet (Gadepalle et al. 2007; UNEP 2002). Sustainable development is achieved by the use of green engineering principles. These principles help to enhance long-term sustainable infrastructure through design and use of products that minimize adverse impact on human health and the environment. The Green Engineering Principles (Anastas and Zimmerman 2003, EPA 201 0) include holistic design, conservation of natural resources, life-cycle thinking, use of inherently safe and benign materials, minimizing depletion and preventing waste, and implementation of innovative and culturally sensitive solutions. Waste symbiosis from local industrial byproducts, for example, is an innovative reuse and waste minimization technique that provides for sustainable development 1.2 URBAN INFRASTRUCTURE In 2009, urban areas served as home and employment centers to more than half of the world's population (UN 2007). By 2030 that number is expected to rise to 60% (UNEP 2007). In the U.S. urban areas as defined by the Census Bureau include areas with a population over 50,000 and urban clusters include 2,500 to 49,999 people (US Census Bureau 2000). In the US, over 79% of people live in urban areas (US Census

PAGE 20

Bureau 2000). Locally, of the 4.6 million people in Colorado, 3.7 million live in urban areas (USDA 2007). Population projections indicate a population increase of 65% in Colorado from 2000 to 2030. Of the expected 6.2 million people in Colorado in 2030, 5.3 million (85%) will live along the urbanized front range. Urbanized areas are characterized by population density and intensity of development and infrastructure. Urban infrastructure refers to the engineered systems that provide services of water, wastewater, energy and transport (of human goods and information) within an urban area (Ramaswami 2005). Sustainable development can be measured with performance, environmental, social equity and economic metrics. The focus of this thesis is to explore holistic and sustainable design of stormwater detention basins in Denver, Colorado and the arid Western region, addressing performance and environmental aspects of sustainable design. 1.3 STORMW ATER MANAGEMENT 1.3.1 LANDSCAPE AND STORMWATER RUNOFF The built environment will expand to support increasing urban populations. To highlight the accelerated rate of land use, it is estimated that in 2030 more than half of all the built environment in the U.S. will have been constructed since 2000 (Nelson 2004). New development changes the natural hydrologic character of the landscape. Large impervious areas, such as parking lots and roadways, increase the stormwater runoff volumes, reduce groundwater recharge and increase pollutant loads. Additionally, in an urban environment, the disturbance of land by construction can trigger the leaching of soil-bound nutrients (Smith 2005). In fact, the EPA indicates that "runoff from urban areas is the leading source of impairments to surveyed estuaries and the third largest source of water quality impairments to surveyed lakes (EPA 2004b)." Urban drainage in tum affects the water supply sustainability. For example, degraded water quality has forced five communities in Colorado to utilize aggressive reverse osmosis (RO) systems to treat drinking water (American Membrane Technology Association 2006). On the other hand, the city of New York was able to save millions of dollars through focusing their 2

PAGE 21

efforts on source water protection to avoid increased treatment capacity. (Appleton 2002). Quantity of Runoff: Dealing with stormwater is an essential function of urban infrastructure. Historically, the focus of stormwater management has been on public safety, acting as a flood control measure for a given hydrological area. Traditional stormwater conveyance and collection systems (concrete channels) have been a common approach to dealing with large volumes of runoff water from impervious areas. However, purely structural catchments areas such as constructed basins and detention tanks are successful to reduce the peak flow and runoff volume at the design or historic level, but provide little to no pollutant removal for frequent events (Federal Highway Administration 2006). Urban runoff solutions must include both quantity and quality controls, i.e. improving the on-site recharge to the local groundwater and reducing the pollutant loads at the source. Quality of Runoff: Since 1990, the focus of stormwater management has shifted to include water quality concerns. By adapting and maximizing the design of detention basins to enhance the treatment and infiltration of runoff, stormwater quality can be enhanced. A study by CH2M Hill found centralized conventional drainage design was unable to match the performance of the natural watershed (CH2M Hill 2001). Low impact development (LID) techniques take advantage of micro-scale approaches to mold the development of land to function similar to natural drainage systems, thus replicating ecosystem services which the open space would have performed (Sample and Heaney 2006). 1.3.2 BMP FOR COLORADO UDFCD Monitoring urban rainfall runoff and protecting water quality can be addressed through site-sensitive Best Management Practices (BMPs) which take advantage of structural basins and low impact development techniques. BMPs are site sensatie solutions based on location. The Urban Drainage Flood Control District (UDFCD) has been charged with assisting local governments in the Denver metropolitan area with multi-jurisdictional drainage and flood control issues. The UDFCD has developed design criteria for stormwater management for the semi-arid climate which 3

PAGE 22

includes comprehensive list of installation and maintenance instructions for BMPs specific to Denver metro area. Figure 1-1 Picture of PLD Aurora, CO In particular the porous landscape detention (PLD) is a design which the UDFCD promotes for small installations reducing the impact on developable land. The PLD is a constructed sedimentation facility intended to capture and filter stormwater for micro rainfall events taking advantage of the intrinsic quality of plants to act as water treatment systems (Guo 2007). The current design consists of a vegetated zone on top of a filtration mix underlain with large aggregate and drains. This design shown in Figure 1 is specified in Drainage Criteria Manual, Volume 3, Chapter 5.6 (USWDCM 2001 ). depth (12" max. ) Oerflow Pass Figure 1-2 PLD Cross Section 4

PAGE 23

The UDFCD criteria manual notes one specific disadvantage of the PLD is the potential for clogging. Additionally, the specific water quality benefits of the PLD in the Denver area are yet to be defined (USWDCM 2001 ). BMP design is still evolving and new research leads to upgrades in current recommendations. The current design recommendations leave opportunity for the incorporation of waste symbiosis and holistic design concepts. The current media mix design consists of peat, sand and gravel. Peat, which is imported, is very expensive ($130 per cubic yard at Paulino Gardens in 2009) and has associated environmental impacts of transportation and of peat mining. Local waste streams offer an opportunity for replacement of portions of the media. For example options for peat replacement include, compost, shredded paper and other organic waste stream materials (Tucker 2007). In addition the sub-layer may utilize waste stream materials such as recycled aggregate (McCambridge et al. 2004) and shredded tires (JaiTire 2008). However, as shown in Table 1-1 no previous study has evaluated the impact of waste materials on the performance and sustainability aspects of a PLD. 1.3.3 PREVIOUS STUDIES The PLD promoted by the UDFCD for the Denver Metro area is a vegetated detention basin which acts to treat stormwater. The PLD is also referred to as bioretention, biofiltration, rain garden and vegetated infiltration basins in other parts of the United States. PLD, or bioretention, was originally developed in Prince George's County, Maryland and has been used as a stormwater BMP since 1992 (Prince George's County Department of Environmental Resources (PGDER) 1993; USEPA 1999). PLDs have been proven to enhance stormwater by reducing peak runoff volumes, and by providing filtration of sediments and sorption of pollutants. (Davis et al. 2001; Davis et al. 2006; Davis et al. 2003; Hunt 2003; Hunt et al. 2006). The UDFCD adopted the recommendations as presented by the Prince George's County, Maryland with slight adjustments based on the hydrologic and geologic conditions common to Colorado (Guo and Urbonas 1996a; 1999). Table 1-1 presents studies completed to date of low impact development techniques similar to PLDs including bioretention areas and vegetated infiltration basins. 5

PAGE 24

Table 1-1 Existing PLD research Performance Sustainability Location Author (year) Scale Technique Physical-Chemical Infiltration Biological Waste Reuse Life Cycle (Caltrans Various BMPs to Visual N,MTSS, Life Cycle CA 2004) field retrofit sites observation fecal X X Cost coliform lab+ bioretention N (lab)+ M XPlanted Sorption PA (Hunt 2003) field mulch top layer balance TSS,BOD but no data X of metals and anoxic zone (field) NC (Sharkey lab+ bioretention with N 2006) field anaerobic zone X X X X bioretentionNewspaper, (Kim et al. lab anoxic zone and X N, turbidity X woodchips, X 2003a) various soil mixes samples sawdust, alfalfa bioretention grasses water (Sun, 2004) lab uptake of metals balance M evaluated for X X by grasses metals uptake (Davis 2007) field bioretention inflow and N,M, TSS X-planted outflow but no data X X MD (Hsieh, 2004; lab+ bioretention with N,M, TSS, Hsieh and field media mixes outflow 0/G, X mulch X Davis 2005) (Davis et al. lab bioretention flow rate M X-planted mulch Sorption 2001) but no data of metals (Davis et al. lab+ bioretention flow rate N X-planted 2006) field but nodata X X (Prince George's field LID techniques yes N X X X County 2000) (CH2MHill field LID techniques flow rate X X X Life Cycle WA 2001) Cost (Ames et al. lab retention basins inflow M,TSS, TPH 2001) with media mixes X X X (Siriwardene sediment transpon lifetime AU lab outflow TSS X X for 2007) in gravel filter clogging SystemNY (Appleton field various LID wide 2002) X X X X Economic savings (Guo and Urbonas field Various LID WQ(:_V X X X X co 1996b;2002) (Guo and field Infiltration bed WQ(:_V X X X X Hughes 2001) Note: X-not mcluded m the study; Mmetals, N-nutnents, TSStotal suspended sohds, BOD biological oxygen demand, 0/G-oil and gas, TPHtotal petroleum hydrocarbons WQVC-water quality control volume 6

PAGE 25

Since the 1990s the majority of the research on PLDs has been completed at the University of Maryland under Dr. Allen Davis (Davis 2007; Davis et al. 2001; Davis et al. 2006; Davis et al. 2003; 2004; Hsieh and Davis 2005a; 2005b; Hsieh et al. 2007; Kim et al. 2003; Sun and Davis 2007). No previous study has taken a combined look at performance criteria and environmental benefits of a waste incorporated enhanced PLD. Additionally, past studies have been focused on coastal areas with different weather patterns and hydrologic conditions, rather than that of the arid west. 1.4 OBJECTIVES The objectives of this thesis are: 1) to select the best waste material reuse for sustainable PLD sub-base system design (Chapters 1 and 2) 2) to quantify the life-cycle environmental benefits of waste reuse for stormwater PLD. (Chapter 2) 3) to investigate the operation of a 2-layered PLD (Chapter 1 and 3) 4) to investigate the impacts of waste materials and vegetation on performance of the PLD addressing (Chapter 4 and 5) 2a) infiltration capacity for on-site stormwater volume disposal, and 2b) effectiveness of contaminant removal for storm water quality enhancement. The objectives were completed through the following phases of work: Literature review and method development (Chapters 1-5) Waste material screening (Chapter 2) Environmental life cycle analysis (Chapter 2) Develop and test soil column design at the bench scale (Chapter 3) Model a 2-layered PLD flow from Bench scale test (Chapter 3) Bench scale test with waste materials and bare soil conditions (Chapter 4) Bench scale test with vegetation (Chapter 5) 7

PAGE 26

2 WASTE MATERIALS SCREENING AND ENVIRONMENTAL BENEFIT This chapter expands on the paper entitled "Waste Incorporated Sustainable Design of Stormwater Detention Basins 1) Waste Screening Tests" by Shauna M Kocman, Anu Ramaswami, James C. Y. Guo. The paper was submitted to the ASCE Journal of Environmental Engineering in April 2010. 2.1 INTRODUCTION AND OBJECTIVE Stormwater management is a site-sensitive solution requiring optimal design techniques for the geographic region and locally available construction materials. Consequently the associated economic and environmental benefits are closely tied to location also. The stormwater infrastructure chosen for this study, the porous landscape detention basin (PLD), is currently recommended for use in the Denver area (USWDCM 2001 ). The current design recommendations for materials leave opportunity for the incorporation of waste symbiosis and sustainable design concepts. The upper filtering layer is composed of a peat and sand mixture. Peat, which is imported, is very expensive and has associated environmental impacts of transportation and of peat mining (Cleary et al. 2005). Both large and small (sand) aggregates are virgin materials and carry associated impacts (Reiner 2007). To date there has been no comprehensive waste screening study to evaluate the use of various waste stream materials into PLDs and the resulting hydrologic performance and life cycle environmental impact. Local environmental benefits include reduction in peak runoff flows, improvements in water quality and reduction in pathogen counts. Life cycle environmental benefits include waste reuse and greenhouse gas benefits over the life cycle of a PLD, from materials extraction to installation. To the best of our knowledge only one study has investigated the use of waste materials as replacements in PLD. In 2003 Kim, Seagren and Davis investigated the use of various organic materials (paper, alfalfa, etc) for use as electron donor substrate for denitrification in the lower layers of a modified bioretention system but did not evaluate the impact on infiltration rates or environmental benefit (Kim et al. 2003). Other studies which incorporate leaf or mulch compost are not specifically focused on waste reuse and do not include compost created from a mix of organic 8

PAGE 27

waste materials. These studies have focused on the optimization of mixtures of native soil, sand, mulch and/or leaf compost mixtures based on measurements of infiltration rates and/or pollutant reduction (Ames et al 2001, Thompson 2008, Hsieh and Davis 2005, Bratieres 2008). The performance of field installations of bioretention areas with native soil, sand, and compost has been monitored and is reported in the literature,(Ames et al. 2001; Davis 2007; Hsieh 2004; Hunt et al. 2006). None of the above studies have evaluated the life cycle environmental benefit of offsetting virgin materials by incorporating waste reuse. The objective of this chapter is to find suitable waste materials and evaluate the environmental benefit (local and life cycle) of incorporating those materials. Possible waste materials through literature review are inventoried. Those materials were screened through performance tests (local availability, cost, leaching and infiltration rates) and confirmed by germination rate testing. Permissible waste-incorporated mixtures are defined and the environmental benefit is evaluated using greenhouse gas savmgs. 2.2 LITERATURE REVIEW ON MATERIALS SUBSTITUTION The Denver region was the focus of this study so the PLD design, materials, testing procedures and environmental benefit are based on this location. A list of possible materials was created based on literature review and exploration of local materials. The procedure for screening required that materials pass three criteria: 1) local availability and cost, 2) leaching and 3) infiltration rate. A last confirmation test on germination rates ensured the passing mixture(s) would support vegetative growth. The first step was to create a list of possible waste materials. While few studies have focused on incorporation of more than leaf/ mulch compost in PLDs, guidance can be gained from focusing on landscaping studies that have previously documented use of waste materials. An initial literature review shows aggregate replacement options in landscaping, include recycled aggregate, crushed glass, and shredded rubber tires (Moller and Leger 1998; Tang et al 2007). Peat replacement options include organics used as soil amendments such as compost, recycled paper, sawdust, and spent hops (Burgos et al. 2006; Castaldi et al. 2005; Davis and Wilson 2007Glenn et al. 2002; Grimes and Cooper 1999; Herwijnen et al. 2007; Kim et al. 2003; Motavalli and Discekici 2000; Molphy et al. 2001, Tucker 2007,). Further investigation into organics revealed a local source of compost utilizes various waste streams including those considered here, spent hops, saw dust, wood chips and biosolids (Yost 2008). Therefore this compost was considered as an avenue to capture more than one waste 9

PAGE 28

stream. Because sources of materials for compost production vary by time and location, Class A certified compost was chosen to ensure quality (Yost 2008). Some studies have shown incorporating waste stream materials may have unintended negative water quality impacts. For example, the leaching nutrients from compost are a concern (Burgos et al. 2006; Castaldi et al. 2005; Grimes and Cooper 1999; Herwijnen et al. 2007). Paper has shown to initially deplete garden soils of nitrogen so may possible offset the nutrients in compost (Glenn et al. 2002). Table 2-1 presents the potential waste stream materials for replacement based on literature review of landscaping applications. Additional information as whether data are available for leaching, life cycle analysis have been conducted and if the material is locally available in the Denver area is also included in Table 2-1. 10

PAGE 29

Table 2-1 Summary of literature on PLD waste stream replacements Waste Stream Author and Year Other Data LCAof Local Applications lA vailable f01 GHG? Availability Leachin11:? Aggregate Recycled (Clean Washington Landscaping as crushed glass Center 1996) groundcover or no no no mulch (Moller and Leger Topsoil mix yes no 1998) Shredded (EPA 2009, Tang Mulch, water no no yes Rubber et al. 2007) filtration Recycled (McCambridge et Landscape aggregate al. 2004) material and road no no yes base Substrate Media Compost (Burgos et al. Peat replacement 2006; Castaldi et al. 2005; Grimes yes no yes and Cooper 1999; Herwijnen et al. 2007) Office paper (Motavalli and Soil topdressing yes no yes Discekici 2000) Recycled (Glenn et al. 2002; Peat replacement Newspaper Kim et al. 2003; yes no yes Molphy et al. 2001) Spent hops (Tucker 2007) Soil amendment no no no and compost Pine needles (Pote and Daniel Groundcover no no no 2008) Sawdust (Davis and Wilson Soil amendment no no no 2007) And compost 2.3 METHODOLOGY Local landscape suppliers were contacted for availability and cost of possible materials. Landscape supply companies in Denver do not carry pine needles and do not recommend use of pine needles as a soil amendment due to their acidic nature. Availability of crushed glass in many locations is unreliable and transporting it is cost prohibitive in places where sand is readily available. The compost incorporated many waste stream organics, such as spent hops, saw dust, and wastewater residuals. Spent 11

PAGE 30

hops and sawdust were not readily available and were incorporated in the compost, so were not investigated separately. As a result, local waste materials choices were crushed recycled concrete, shredded paper, compost and shredded tires presented in Table 2-2. Table 2-2 Possible local waste streams to replace virgin materials for incorporation in PLD aste Stream Replacement Local Source Rubber Aggregate JaiTire Industries, AcuGreen !Recycled concrete Aggregate Oxford Recycling Center post Peat AI Organics Pffice paper Peat Waste Management !Recycled Newspaper Peat Denver Post newsprint and white waste Waste Management 2.3.1 LEACH TEST The narrowed list of possible materials was subjected to leach testing. The testing began with replacement of one virgin material with one waste material e.g., replacing peat with compost only. Subsequent experiments included combining various waste materials e.g., compost and paper mixture for peat replacement. Water quality samples were collected from 4 liter batches of the mixtures by adding deionized water to the top of each batch and collecting the outflow. Samples were analyzed for pH, nutrients and metals. A "sension 1 pH meter" was used to read the pH of the samples. Samples were sent to Metro Wastewater Reclamation District for analysis of total keldjal nitrogen (TKN), nitrate plus nitrate (N02+N03), total phosphorous (TP), and total metals. Figures 2-1 and 2-2 are photographs of the experimental set up for batch testing. 12

PAGE 31

Figure 2-1 Batch Test Setup for Leaching and Flow Rate Figure 2-2 Batch Test Setup and Outflow for Collection of Leachate and Flow Rate The minimum acceptable infiltration rate for a PLD is 2.5 centimeters per hour (crnlhr) (1 inch per hour (inlhr)) and the minimum filtration depth is 45.7 em (18 inches). Therefore the maximum contact time for stormwater filtering in a PLD is expected to be 18 hours just before failure. An exceedingly conservative 48 hour contact time was chosen the first leach test. A still conservative contact time of 18 hours was chosen for subsequent tests. Reduction in leaching was investigated through samples taken at the first and tenth flushes. Samples of leachate were analyzed at Metro Wastewater Reclamation District's certified laboratory for nutrients and metals. The results of material mixtures were compared to the EPA freshwater criteria and control, peat and sand leachate results (EPA 2006). 2.3.2 FLOW RATE TEST The same 4-liter batches of materials were then tested for flow rates. Water was added to the top and the outflow was measured with known volume and stopwatch. Flow rates were measured and compared to the control (peat and sand). Mixtures 13

PAGE 32

within a range of 80% to 120% of the control flow rate qualified for confirmation germination tests. 2.3.3 GERMINATION TEST After the completion of the flow rate and leaching tests, the new batches of the passing mixtures were created for germination tests. The germination tests were conducted with a mixture of native grass seeds in small batches of approximately 200 square centimeters of bedding material. Germination and growth rates were compared to the control (peat and sand). Grass seeds were counted in relative quantities recommended in Volume 3 Criteria Manual for the Denver Metropolitan area and mixed into the top 1/8 inch of media mix (USWDCM 2001). The recommended seed mixture is shown in Table 2-3. The surface area of the batch test would not allow for all 14 species of plants therefore only the available grasses were used. Table 2-4 presents the relative number of grass seeds were spread on the filtration media mixes and watered. Table 2-3 PLD vegetation seed mixture prescribed in UDFCD criteria manual COMMON NAME SCIENTIFIC NAME VARIETY PLS Lbs Ounces per Acre per Acre Sand bluestem Andropogon hallii Garden 3.5 Sideoats grama Bouteloua curtipendula Butte 3 Prairie sandreed Calamovilfa longifolia Goshen 3 Indian ricegrass Oryzopsis hymenoides Paloma 3 Switchgrass Panicum virgatum Blackwell 4 Western Wheatgrass Pascopyrum smithii Ariba 3 Little bluestem Schizachyrium scoparium Patura 3 Alkali sacaton Sporobolus airoides 3 Sand dropseecl Sporobolus ayptandrus 3 Pasture Sage Artemisia frigida 2 Blue aster Aster laevis 4 Blanket flower Gaillardia aristata 8 Prairie coneflower Ratibida columnifera 4 Purple prairiedover Dalea (Petalostemum) purpurea 4 SubTotals: 27.5 22 Totallbs per acre: 28.9 Source (USWDCM 2001) 14

PAGE 33

Table 2-4 Number of seeds planted in each batch test Amount Batch size seeds/ Seed Name (lbs/acre) seeds/lb (sq ft) Batch sideoats gramma 3 191,000 0.5 7 prarie sandreed 3 274,000 0.5 9 swtichgrass 4 270,000 0.5 12 western wheatgrass 3 110,000 0.5 4 little bluestem 3 260,000 0.5 9 sand dropseed 3 825,000 0.5 28 The containers in Figure 2-3, were set near a sunny window, watered and monitored daily for germination and growth. The number of plants germinated was tracked and reported on day 8, 10, 12 and 17. Germination rates did not increase between day 12 and 17 as shown. Therefore the day 17 plant counts were used to compare the mixtures' ability so support seed germination and plant growth. Figure 2-3 Picture of Batch Test Setup for Germination Rate. Grass Seeds Were Planted in Various Media Mixtures, Watered and Set by a Sunny Window 2.3.4 ENVIRONMENTAL LCA The final result testing lead to a multi-criteria permissible range of mixtures for incorporation in PLDs. Two sample mixtures in the mid-range of the permissible amounts were chosen to evaluate the environmental impact based on a design example site. The environmental benefit of incorporating the waste materials can be measured in greenhouse gas (GHG) emissions. The materials included in this study replaced virgin (sometimes foreign) materials with use of local waste materials. The impact of the virgin materials and the equivalent replacement are evaluated by offset of virgin material and GHG emissions. 15

PAGE 34

The EPA Waste Reduction Model (WaRM) and published GHG emissions specific to peat mining and transportation (EPA 2008, Cleary et al. 2005) and aggregate (Reiner, 2007) were combined to calculate the GHG savings. The boundaries considered for impact from materials were extraction to installation and included transportation. A design example was created to calculate the net benefit of a waste-incorporated PLD installed at a field test site. Peat, which is imported, has associated impacts of land use, fossil fuel combustion, and peat decomposition (ADAS UK Ltd and Enviros Consulting Ltd 2005). Peat bogs are drained before mining which increases decomposition and release of carbon dioxide and methane gas. The peat is then extracted, processed and transported long distances. Once the peat is used in soil applications it decomposes further releasing more GHGs. One life cycle analysis completed on Canadian Peat in 2005 estimated an emission factor of approximately 0.05 tons of carbon dioxide equivalents per ton of peat extracted in 2000 (Cleary et al. 2005). Transportation for aggregates, which are locally mined, is shorter than peat but scarcity in the Denver Metro area is a concern. The primary source of alluvial coarse aggregates in Denver is the South Platte River corridor. As sites near Denver are depleted, it is expected that the distance traveled for coarse aggregate will soon increase (Reiner 2007). As the travel distance to obtain virgin aggregate materials is expected to increase, energy use and cost will increase as well. Additionally, reuse of local materials minimizes the impact of disposal in a landfill. The waste materials and the replaced virgin materials (peat and sand) are presented in Table 2-5. Use of waste materials results in GHG savings (negative number) while use of virgin materials results in emissions of GHG (positive number). Offsetting one virgin material results in both the savings from the use of the waste material plus the savings of offsetting the virgin material. For example a ton of peat ( +.84 MTC02E) replaced by a ton of compost ( -.2 MTC02E) results in a net benefit of -1.04 MTC02E. 16

PAGE 35

Table 2-5 Data sources to calculate GHG emissions impact Material MTC02E Model Includes Literature impact per Source ton material Compost from -0.2 20 miles transport, sequestered (EPA 2008) mixed organics carbon and avoided landfill Paper -0.05 20 miles transport, avoided (EPA 2008) landfill Rubber -0.04 20 miles transport, avoided (EPA 2008) landfill Aggregate +.006 40 miles transport, processing (Reiner, 2007) and consumption Peat Moss +.84 2000 miles transport, processing, (Cleary et al. carbon release from mining 2005; EPA 2002; EPA 2004a) Note: MTC02E metnc ton carbon dioxide eqmvalents See Appendix A for more detail about the GHG emissions from peat The total peat produced in 2000 was 1.3 million tons with a total GHG emission, excluding transport, of 0.612 MTC02E (Cleary et al. 2005). Transportation was added based on EPA fuel economy standards of 5 miles per gallon for trucks. The EPA reported GHG emission factor for diesel fuel is 22.23 pounds per gallon (EPA 2002; EPA 2004 ). The average truckload of peat is 17.4 tons (Cleary et al. 2005). A total of .612 MTC02E per ton of peat plus transport from Canada to Denver (2,000 miles) creates .24 MTC02E per ton of peat is the total GHG cost of peat (EPA 2002; EPA 2004). 2.4 RESULTS 2.4.1 LOCAL AVAILABILITY AND COST The waste material screening criteria of availability/cost, leaching and flow rate, lead to a list of unsuitable materials and a range of optimal mixtures for the suitable materials. This study was designed for a location-sensitive solution for the Denver metro area, representing an urban area in the arid west. Following literature review, local suppliers were contracted for availability and cost. The possible materials were 17

PAGE 36

narrowed to compost from mixed stream of organics, paper, shredded tires and crushed concrete. The cost comparison for the chosen materials, compost, shredded paper, shredded tires, and crushed concrete, is presented in Table 2-3. All the waste materials are less than or equal to the cost of the virgin materials and therefore pass the cost screening criteria. Sand and aggregate are currently recommended in the design mixture and were obtained from Pioneer Sand (Golden, Colorado) and Sante Fe Sand and Gravel (Englewood, Colorado).The replacements considered for aggregates included recycled glass, shredded tires and crushed concrete. Availability of crushed glass in many locations is unreliable and transporting it is cost prohibitive in locations where sand is readily available. Locally, Coors Bottling collects the majority of available glass and melts it into new beer bottles. Some fines were screened off and donated from Coors Bottling. A source of surplus crushed glass is not readily available in the Denver Metro area. The City of Durango collects and sells glass cullet and the grinder was broken at the time of this study. Additionally the transportation to the Metro area is approximately 240 miles and would be expensive. Therefore, the glass cullet currently fails the availability and cost category for use in the Metro area. Shredded tires and crushed concrete are readily available. Acugreen (Denver, Colorado) donated shredded tires and provided a tour of the facility. Crushed concrete is recycled from many demolition projects and was donated from Oxford Recycling (Englewood, Colorado). The compost was obtained from A 1 Organics and incorporated many waste stream organics, such as spent hops, saw dust and wastewater residuals. Spent hops and sawdust were not reliably available and were incorporated in the compost, so were not investigated separately. Shredded waste paper is available from Waste Management. For the batch test waste paper was shredded from the newspaper and some office paper. Landscape supply locations do not carry pine needles and do not recommend use of pine needles as a soil amendment due to their acidic nature. The cost comparison for the remaining materials, compost, shredded paper, shredded tires, crushed concrete, is presented in Table 2-6. All the waste materials are less than or equal to the cost of the virgin materials and therefore pass the cost screening criteria. 18

PAGE 37

Table 2-6 Cost comparison of virgin and waste materials Virgin Material Replacement Density Cost Comparison (cost in $/cy) Material (lbs/cy material) ($/cy material) ($/cy replacement/ $/cy virgin) Peat AI Compost 1,030 lbs/cy $35 .27 ($170/cy) Peat Shredded 39lbs/cy $10 .08 ($170/cy) paper Sand Shredded Tires 2,000 lbs/cy $17 1 ($17/cy) Crushed 2,900 lbs/cy $11 .44 ($17/cy) concrete Note: lbs/cy = pounds per cubic yard = 0.59 kg/m3 2.4.2 LEACHING CRITERIA After screening for availability and cost, the waste materials were subjected to leach tests in order to ensure the protection of water quality. Leaching of nutrients and metals was found to be dependent on both the amount of the waste material in the mixture and contact time with water. Batch Tests for pH indicated that the organic material slightly acidifies the water (pH 6.09 -6.23 while crushed concrete drastically increases the pH ( 11-12). The results of the pH testing for are presented in Table 2-7. Table 2-7 Leachate pH from control and waste incorporated media mixtures after 48 hours contact time Material pH after 48 hours Peat and sand 6.23 Compost and sand 6.09 Paper and sand 6.12 Tires 6.78 Concrete 12.17 Concrete aggregate was quickly eliminated as the pH is very alkaline (11-12 standard units) even after many flushes of water. Table 2-8 presents the results of the pH batch test with 100% crushed concrete. 19

PAGE 38

Table 2-8 Leachate pH from crushed concrete Flush number Contact Time pH 1 48 hour 12.61 2 48 hour 11.03 4 48 hour 12.29 7 48 hour 12.43 8 continuous flow 11.13 10 continuous flow 11.53 Additional batch tests were constructed with media layers representing the filtration layer in the field installation to test the effect of recycled concrete below various soil mixes. The recommendation for layers in Volume 3 Criteria Manual is 30 em (12 inches) allowance for ponding water, 46 em (18 inches) of soil media and 20 em (8 inches) of aggregate under drain. In other words the water allowance is approximately 30% of the depth, soil media is 50% and the aggregate is 20%. Therefore the batch tests were constructed with 9 em (4 in) of water, 15 em (5 Y2 in) of soil media and 6 em (2 '12 in) of aggregate for a total of 30 em (12 in) depth. Three flushes (1, 2, 3) with a 24 contact time were measured. Deionized water was added the top of the soil media, left for 24 hours, drained and pH was measured. Three additional flushes (4, 5, 6) were continuous flow. The results of the testing are presented in Table 2-6. Table 2-9 Leachate pH from media mixtures on top of crushed concrete Soil Media on top of Crushed pH pH Concrete Flush Number Flush Number (continuous (24 hr contact) flow) I 2 3 4 5 6 15% Compost and 85%Sand 11.44 11.68 11.59 11.83 10.57 10.74 15% Peat and 85% Sand 12.06 11.90 11.94 11.55 11.61 10.70 16% Compost, 42% Tires, 11.53 11.73 11.47 10.83 10.50 10.47 85% Sand Pure shredded tires leach high concentrations of copper (68 ug/L) and zinc (13,500 ug/L) during 48 hours of contact time. Less contact time and a mixture of 50% crumb rubber and 50% sand reduces the concentrations significantly (23 ug/L copper and less than 20 ug/L zinc). After 10 subsequent flushes of water, both copper and zinc are reduced to non-detect levels ( <20 ug/L). Similarly, higher concentrations of 20

PAGE 39

compost and longer contact times increase the amount of nutrients and metals in the leachate. With a mixture of 15% compost and 85% sand (complete replacement of peat) the leachate contained 54 mg/L TKN, 31 mg/L TP and high values of copper (307 ug/L), lead (66 ug/L) and zinc (489 ug/L) after 48 hours of contact. Whereas after 18 hours of contact time and only 5% compost the nutrients ( 6.5 mg/L TKN and .41 mg/L TP) and metals (20 ug/L copper and lead and zinc both non-detect) were considerably lower. After ten flushes of water the nutrients and metals in all cases had been reduced by more than 30%. Table 2-10 presents the results of the leaching tests and whether the mixture passed this screening test after 10 flushes. Table 2-10 Leaching results of various media mixtures indicate that three mixtures pass the leaching test MEDIA Pass/ Flush Contact Nutrients (mg!L) Metals (u /L) Fail Number Time TKN Nitrate Total Cu Pb Zn (hours) +Nitrite p EPA Freshwater NR 10 0.055 13 65 120 Criteria (EPA 2006) Control 1st 48 11 0.15 0.27 14.6 ND(5) ND(20) 15% peat 85% sand Control 1st 18 4.4 0.59 0.42 4.5 ND(5) Control lOth 18 2.5 0.09 0.23 2.5 ND(5) ND (20) 15% compost 85% fail 1st 48 53.9 196 31.4 307 66.4 489 sand 15% paper 85%sand PASS 1st 48 ND(.3) 0.03 0.23 ND(2) ND(5) ND (20) 10% paper, 5% fail I st 18 6.5 45.5 0.41 19.7 ND(5) ND (20) compost, 85% sand PASS lOth 18 2.5 0.32 0.05 3.7 ND(5) ND (20) 5% paper, 10% fail 1st 18 12.5 35.4 1.94 34.3 ND(5) 24.3 compost, 85% sand fail lOth 18 4.1 0.06 2.01 13.2 ND(5) ND (20) 100% tires fail I st 48 43.9 0.72 0.23 68.3 ND(5) 13500 fail I st 18 7.4 0.41 0.09 22.4 ND(5) ND (20) 50% tires, 50% sand PASS lOth 18 2.8 0.05 0.09 4.1 ND(5) ND (20) PASS 1st 18 3.9 0.48 0.05 10.4 ND(5) ND (20) 10% tires, 90% sand PASS lOth 18 2.4 0.16 1.02 6.1 ND(5) ND (20) Note: Failed results are m 1tahcs. As presented in Table 2-10, compost and tires both leach metals, such as copper, and paper does not. Therefore, replacing only a portion of peat with compost and a portion of the sand with tires was considered to reduce the impact of leaching. Because the results indicate that metals and nutrients leaching decreases with flushes, mixtures which pass the tenth flush were considered for further evaluation. Figure 2-21

PAGE 40

4 presents the results of copper leaching after the first and tenth flush of various media mixtures. The shaded area represents the EPA fresh water quality standard of 13 ug/L copper (EPA 2006). Figure 2-4 indicates that a mixture ofless than 10% compost and less than half replacement of tires with sand must be used to attain the water quality standard within 10 flushes. Therefore, based on the water quality criteria, any mixture with less than 42% tires and less than 10% compost is permissible. Copper leaching 40 35 --------30 2. 20 -.. __ :I 15 0 10 5 0 0 2 4 6 8 flush number 10% paper, 5% 85% sand I 5% paper, 10% com ost, 85% sand b 50% tires, 50% san x 10% tires, 90% sand control ---... 10 12 Figure 2-4 Reduction of Copper Leaching From Various Mixes After Multiple Flushes of Water. The Gray Area Indicated the Permissible Waste Mixture Passing the Metals Leaching Test 2.4.3 FLOW RATES After leaching tests, paper, compost and tires remained as viable materials. Permissible mixtures of the materials based on leaching results were then compared to the flow rate. The results of the flow rates indicate that paper slows the infiltration while tires increase infiltration rates. An optimal range of paper and compost may be reached based on the comparison to the flow rate of the control. Figure 2-5 shows the permissible amount of compost in the mixture is about 6.5-11% and 4-8.5% paper based on flow rate alone. 22

PAGE 41

--------------------------, Pecent of Compost and Paper versus Row Rate 1.20 0 paper COIT1JOSI .. 1.00 -c f! g 0.80 :Cu.. ..2:::J 0.60 -c Gl o E 0.40 f! !: 0.20 u.. 0.00 0 1 2 3 4 5 6 7 8 9 10 1112 13 14 1516 percent material Figure 2-5 Effect of Varying Amounts of Compost and Paper on Flow Rate Shows Compost Increases Flow rate and Paper Slows the Flow Rate The shape, size and buoyancy of materials also affect the infiltration rate. Shredded tires, for example float in water. When water is added to tires only (100% tires), the tires float and the flow rate is basically free flow, as fast as the outlet allowed. This case was not considered. Figure 2-6 shows the increase in flow rate when replacing tires with sand. The increase in flow rate should be considered when mixing tires and paper, where tires can offset the reduction in flow rate from paper. The shaded area represents the permissible range of shredded tires in the mixture. 23

PAGE 42

0 Effect of Replacing Sand with Shredded Tires 10 20 30 40 percent material 50 Figure 2-6 Increase in Flow Rate is Related to the Amount of Shredded Tires 2.4.4 CONFIRMATION TESTGERMINATION The results of the previous three tests lead to permissible mixtures to be confirmed through germination test. The germination test was conducted with the following mixes: 1. Control: 15% peat and 85% sand 2. 10% compost, 5% shredded paper, and 85% sand 3. 5% compost, 10% shredded paper and 85% sand 4. 6% paper, 10% compost, 42% tires and 42% sand. As presented in Figure 2-7, the testing confirmed the success of germination of the four mixtures. The conclusion of the batch test germination rates was a mixture of between 5% to 10% compost and less than 42% tires will meet the germination test requirement. 24

PAGE 43

Germination Rate 14.0 >12.0 1--------...... ra r----OQ 10.0 '->! 8.0 I 6.0 +---------+---+-----+------+-----rt---::J c r--Z.!2 4.0 a.. 2.0 0.0 0 2 3 mix number mix #1 15% peat, 85% sand o mix #210% compost, 5% paper, 85% sand ::.:mix #3-5% compost, 10% paper, 85% sand 1 4 o mix #410% compost, 6% paper, 41% tires, 42% sand Figure 2-7 Germination Tests Confirmed no Difference in the Ability of the Four Mixes to Support Vegetation 2.5 SCREENED BEDDING MATERIAL MIXTURE The summary of the permissible bedding mixture is presented in Table 2-11 based on the three screening criteria, 1) cost/availability, 2) leaching, 3) flow rate, and confirmation germination tests. Results indicate that mixtures of compost, shredded paper, and shredded tires pass the screening criteria and confirmation tests. The three materials were available and cost effective. Leaching and flow rates determined the amount of compost, paper and tires in the mix. The permissible amount of compost is between 6.5-10%, shredded paper between 5-8.5% and shredded tires less than 10% of the complete mixture. 25

PAGE 44

Table 2-11 Permissible ranges of waste-incorporated filtration mixes Permissible amount of material in the mixture based Virgin Replacement on the following criteria: Permissible Cost/ Leaching Flow Rate 17 day Amount of Material Material Availability germination Material rate Compost 6.5-10% Pass 10% or 6.5-11% 5-10% (5-8% if Peat less* mixed with tires) Shredded paper Pass 15% or 4-8.5% 5-10% 5-8.5% less Sand Tires Pass 42% or <8% 42% or less 0-8% less Complete replacement of the organic portion (peat) is possible with a mixture of paper and certified compost. If the mixture contains only sand (not tires), a mixture of 6.5-10% compost and 5-8.5% shredded paper may be used. Due to leaching from both tires, the amount of compost must fall between 5-8% of the total mix if tires are also incorporated. If tires are incorporated in the mix to offset some sand, equal amounts of paper (7 .5%) and compost (7 .5%) would fall within the permissible amounts for the compost, paper, sand and tires scenario. A mixture of equal amounts of paper (7 .5%) and compost (7 .5%) falls within the permissible amount while obtaining benefit of each mixture. A mixture between the permissible amounts of each material creates a buffer for error in measurement while reaping the benefits of waste material replacement. The possible mixtures are shown in Figure 2-8. 26

PAGE 45

15% Peat 85% Sand 7.5% Compost 7.5% Paper 85% Sand 7.5% Compost 7.5% Paper 77% Sand 8% Tires Figure 2-8 Picture of Control and Permissible Waste-Incorporated Media Mixtures forPLD 2.6 GREENHOUSE GAS BENEFIT The environmental benefit of incorporating the waste materials was evaluated with greenhouse gas emissions. The benefit of waste materials (compost, paper, and shredded tires) and the offset virgin materials (peat and sand) are both accounted for in the net GHG benefit, (eg use of compost and paper must account for both the use of the material and the avoided peat). Therefore, if 1 ton of compost were to replace 1 ton of peat the benefit must include .2 MTC02E per ton of compost plus .84 MTC02E per ton of sand. A design example based on the GHG savings from replacing the virgin materials with waste materials is presented in Table 6. A field site on Auraria campus in Denver, Colorado was chosen to retrofit an older sand filter with a PLD. The footprint of the PLD is 213.7 square meters (2,300 square feet). The depth of the bedding material is 45.7 em (18 inches, 1.5 feet) so the volume of material required is 97.7 cubic meters (3,450 cubic feet, 127.7 cubic yards). Two options for bedding material are considered: the business as usual (BAU) currently recommended mixture of 15% peat and 85% sand by volume and the waste-incorporated mixture of 7.5% compost, 7.5% paper, 76% sand and 8% tires by volume. Bedding materials are generally sold by weight and the recommendation is based on volume therefore the density was measured. The density of the materials was evaluated in the lab and the total weight 27

PAGE 46

of each material in the example PLD of 97.7 cubic meters (127.7 cubic yards) is presented in Table 2-9. Table 2-12 GHG benefit in MTC02E of replacing Control with a_ Waste Incorporated CPST mix Mix Number Material %by Material Weight in GHG Total MTC02E Volume Example PLD MTC02E from the (tons) /ton installation of the [avoided tons of ExamplePLD virgin material] (+released -avoided) Business-As Peat 15 6.70 +0.84 Usual (control) Sand 85 146.54 +0.006 +6.51 wasteCompost 7.5 4.93 -0.2 incorporated [3.35 tons peat] PLD Paper 7.5 0.19 -0.05 -0.62 (CPST) [3.35 tons peat] Tires 8 10.22 -0.04 [15.52 tons sand] Sand 76 131.02 +0.006 As presented in Table 2-9, a total of 6.5 MTC02E is released from the use of peat and sand in the example PLD. The total GHG savings from installing the waste incorporated mixture is 0.6 MTC02E. The net benefit includes both the savings from the use of waste materials plus the GHG offset by avoiding the virgin materials. Additionally the benefit is location specific as 0.18 MTC02E of the 0.84 MTC02E are from transportation of Peat from Canada to Colorado. Additional information about the impacts of peat use, processing, transportation and degradation in soil are included in Appendix A. For example, the net benefit shown in Figure 2-9 is a total of7MTC02E. 28

PAGE 47

GHG Comparison of Two Media Mixes [i BAU Waste-lncorporatedl Net Benefit MTC02E Figure 2-9 Net GHG Benefit of Installing the CPST Waste-Incorporated Mix Instead of the Business as Usual (Control) in an Example PLD The total greenhouse gas benefit of offsetting all the peat with compost and paper and a portion of the sand with tires is 7 .I MTC02E for 98 cubic meters ( 128 cubic yards of material or 3,450 square feet of PLD surface area). Since the environmental impact of peat is high the total cost of all virgin materials (6.5 MTC02E) can be offset by replacing the peat with compost and paper (net benefit of 6.6 MTC02E). An additional benefit of 0.4 MTC02E is realized by offsetting a portion of the sand with crumb rubber. Similarly the cost benefit of offsetting the BAU with WI mix is attributed to the compost and paper. The cost of tires is the same as sand and does not contribute to the cost savings. Based on the example PLD, the cost savings to install the WI mix compared to the BAU is $2,800 as presented in figure 2-10. 29

PAGE 48

Cost Comparison for Two Media Mixes 6000 ,--------------, 5000 y; ;::4000 en 8 3000 ....J 2000 1000 0 I BAU Waste-Incorporated I Cost Savings $2800 Figure 2-10 Cost Savings to Installing the CPST Waste-Incorporated Mix Instead of the Business as Usual (Control) in an Example PLD 2.7 CONCLUSION Engineers and contractors may be inclined to change bedding material specifications when materials are unavailable or expensive. Randomly substituting materials into specifications may result in unknown impacts to the operation of the PLD in terms of water quality, infiltration rates and plant growth. The relative amount of material in the mixture, such as shredded tires, compost and paper, was found to effect the performance. For example, complete replacement of peat with compost initially would result in excess metals and nutrients leaching into the stormwater, thereby degrading water quality. Tire particles were found to leach high amounts of copper and zinc after long contact times. Therefore, tires and compost should be used correct percentages. The current recommendation of peat and sand may be used as well as the waste-incorporated mixture of compost and paper with less than 8% replacement with tires and a mixture of compost and paper. A variety of waste materials were selected and screened based on availability, cost, leaching, flow rate and germination rates. The screening process led to a list of unsuitable materials and range of permissible mixtures for the suitable materials. Two final mixtures were recommended for use in a waste-incorporated media for a PLD. The currently used 15% peat and 85% sand may be replaced by either of two mixtures. One option is to replace the peat with equal amounts of paper (7 .5%) and 30

PAGE 49

compost (7 .5% ). Another option is a mixture replacing both the peat and sand portions which would consist of7.5% paper, 7.5% compost, 77% sand and 8% tires. The greenhouse gas benefit of offsetting both peat and sand resulted in a total benefit of 7.13 MTC02E for an example PLD of 3,450 square feet of PLD surface area. Of the 7.13 MTC02E, the greatest portion of the savings (6.63 MTC02E) is realized from the replacement of the peat with compost and paper. Additionally, lighter materials such as organics (peat, compost, paper) and tires were found to be buoyant in water. Although this was not a screening criterion, the impact of buoyancy effects will be studied in further bench scale testing. Aoating particles in stormwater plus added particles from the bedding material may cause obstructions in outflow designs and possible impacts to downstream water quality. 31

PAGE 50

3 INFILTROMETER DESIGN AND FLOW MODEL 3.1 INTRODUCTION The porous landscaping detention basin (PLD) in Figure 3-1 reduces on-site storm water runoff volume and peak discharge using subsurface infiltration. Beneath the vegetated basin bottom is an aggregate sub-base that is typically divided into an upper filtering layer comprised of fine aggregate, and a lower reservoir layer comprised of larger aggregate. The geotextile fabric provides the separation between these two layers. Stormwater that is intercepted by the surface basin will be infiltrated into the subsurface reservoir where the seepage flow is filtered, stored, and gradually released into the perforated pipes that are tied into the downstream sewer manhole. The water detention process is usually assumed to begin with dry subsurface layers. During an event, all the aggregate voids are filled up with water before the seepage flow can be fully developed through the saturated medium. The infiltration rate on the basin bottom represents the inflow to the PLD system while the seepage rate through the subsurface medium represents the outflow. The operation of a PLD is controlled by either the infiltrating rate or the seepage rate, whichever is smaller (Guo 1998). If the subsurface seepage flow cannot sustain the infiltrating flow, the water mounding will be built up to balance the inflow and outflow rates. This phenomenon is manifested by standing water often observed in many water quality enhancement basins (Guo 2001). The similar phenomena were also observed in the subsurface drip irrigation system that has become a common method for the irrigation of field crops, trees, and landscaping. When the pre determined discharge of the emitter is larger than the soil infiltration capacity, water pressure at the dripper outlet increases and can become built up. This pressure buildup in the soil decreases the pressure difference across the dripper and, subsequently, decreases the trickle flow (Shani et al. 1996). Therefore, it is advisable that the subsurface geometry beneath the PLD sustains the continuity of flow. 32

PAGE 51

Perforated Pipe Sewer Outlet Figure 3-1 Layout of Porous Landscaping Basin The design goal of a PLD system is to seek the balance between the surface and sub surface flows under the available hydraulic head and required water quality control storage volume. This chapter presents an attempt to integrate the surface and subsurface hydrology and hydraulic processes together to design a PLD (Guo et al 2009). 3.2 SURFACE STORAGE BASIN A PLD is designed as an on-site storm water disposal facility. The storage volume of a PLD is often sized for the water quality control volume (WQCV) or equivalent to the flush volume. From previous studies, WQCV is approximately equal to 3to 4month event (Guo and Urbonas 1996). For convenience, WQCV is directly related to the local rainfall distribution. There are many recommended probabilistic distributions derived for complete rainfall data series, such as exponential distribution (Bedient and Huber, 1992), one-parameter Poisson distribution (Wanielista and Y ousef, 1993 ), and two-parameter model of Gamma distribution (Woolhiser and Pegram, 1979). In this study, the one-parameter exponential distribution is adopted to fit the frequency distribution of rainfall event depths (Guo 2002). The exponential distribution is described as: 33

PAGE 52

Equation 3-1 -p lp f(p) =-e m pm in whichf(p) is the frequency distribution for local rainfall depth, p. Integrating equation 3-1 yields the cumulative probability distribution as: Equation 3-2 F(O p P) = 1-e Pm in which F( P) is the non-exceedance probability for rainfall depth to be less than or equal to the design rainfall depth, P. Considering surface depression, a runoff producing rainfall depth can be converted into its runoff volume as: Equation 3-3 P0 = C(P-P;) in which Po is WQCV in mm per watershed, C is the runoff coefficient, and P; is the incipient runoff depth [mm]. As recommended, an incipient runoff depth of 2.5 mm is introduced to filter out extremely small rainfall events (Guo and Urbonas in 1996, Driscoll et al. in 1989). Normalizing equation 3-3 yields Equation 3-4 p Po p -=-+-' pm CPm pm in which Pm is the local average rainfall event-depth that can be found elsewhere (EPA 1986). Substituting equation 3-4 into equation 3-2 yields Equation 3-5 in which F(O$pSP0 ) is the probability to have an event that can be completely captured by the design water control volume, i.e. P0 In this study, equation 3-5 is termed the synthetic runoff capture curve that is normalized by local average rainfall event-depth, runoff coefficient, and runoff incipient depth. Re-arranging equation 3-5 yields: 34

PAGE 53

Equation 3-6 -P, C = 1-aeCPm v Equation 3-7 In which Cv is the runoff volume capture rate between zero and unity, and a= constant determined by incipient runoff depth. The value of a represents the watershed natural depression capacity. Figure 3-2 presents a set of normalized runoff capture curves produced using equation 3-6 with runoff coefficients of 0.4, 0.6, 0.8, 0. 9 and 1.0. It is noticed that the curvature of runoff capture curve increases when the runoff coefficient decreases. This tendency reflects the fact that the higher the imperviousness in a watershed, the less the surface depression and detention. As a result, the response of a watershed to rainfall is quick and direct. 1.00 0.90 0.80 .!! 0.70 0.60 :I 0.50 0.40 c 0.30 0.20 0.10 0.00 0.00 ----/ // I ///;. !J I lA /Iff 11 0.50 1.00 1.50 2.00 2.50 3.00 1-c=OAO --e=o.60 -e=o.eo --e--o.oo -C=l.O I Po/Pm (Basin Size/Average Rainfall Depth) Figure 3-2 Stormwater Quality Control Volume for Porous Landscaping Basin Design 35

PAGE 54

For a selected runoff volume capture rate, Figure 3-2 provides the required WQCV for a PLD basin as: Equation 3-8 vo = POA where V0 is the WQCV [m3 ] and A is the catchment area tributary to basin [m2]. Safety is always a concern when designing a PLD. Often the water depth in a PLD is set to be 15 to 30 em (6 to 12 inches). With a selected basin depth, the basin cross sectional area is determined as: Equation 3-9 A = Vo 0 y where A0 is the average cross section area, and Y is the basin depth. To enhance the infiltrating process, the basin bottom shall be on a flat to mild slope. 3.3 SUBSURFACE FILTERING SYSTEM Drain time is critically important to the operation of a PLD because it controls the sediment removal rate. Based on the urban pollutant characteristics, a drain time for the PLD is usually set to be between from 12 to 24 hours (USWDCM 2001). Figure 3-3 illustrates the flow through the two filtering layers including sand-mix and then gravel. The model is limited to saturated conditions as the infiltration rates through the filtration layers reached steady state through 2 filtering layers. This is not applied to the entire wetting cycle in soil saturation. Additionally, the ponding water is assumed to be constant depth. Under a constant head, the steady flow condition is derived as: Equation 3-10 f =VI =V2 In whichfis the infiltrating rate, and Vis the seepage flow velocity through each layer, The subscriptions "1" and "2" represent the variables associated with the sand mix and gravel layers, respectively. A saturated seepage flow through a medium is proportional to the energy gradient as: 36

PAGE 55

Equation 3-11 V -K dH1 I-I HI Equation 3-12 dH2 Vz = Kz-Hz Where K is the hydraulic conductivity, His the energy head, and dH is the energy loss. y r::::::::: :---. -_ -: l:: ::::::::::::::_ ------m H !-----------------------------[]--________ __ dH H2 --------------.: -- ---Water 1 :;.-: : __ ;__ _____________________________________________ BaseD -1\llanometer 0 Re:ldings Figure 3-3 Illustration of Infiltrometer Operation under Saturated Conditions In practice, the design infiltrating rate depends on the drainage nature of the selected soil-mix. With a pre-selected design infiltrating rate, the total filtering thickness for the two filtering layers is calculated as: Equation 3-13 D=JTd Where D is the total thickness for two filtering layers,/ is the infiltration rate, and Td is the PLD drain time. The fundamental challenge in PLD design is how to divide the 37

PAGE 56

total thickness between the two filtering layers because the layer thickness is directly related to the hydraulic gradients for seepage flow through the system. Equation 3-14 D=H1+H2 where H1 is the sand-mix thickness and H2 is the gravel layer thickness. As illustrated in Figure 3-3, the available hydraulic head for the PLD system is Equation 3-15 H=Y+D where Y is the water loading depth in PLD. In this study, the optimal performance of a PLD is defined by the infiltration flow and the subsurface thickness that allow the seepage flow to consume the hydraulic head available as: Equation 3-16 H =dH1 +dH2 Aided by equations 3-10, 3-11, and 3-12, the head losses through the two filtering layers are: Equation 3-17 f dH1 =-H1 Kl Equation 3-18 f dH2 =-H2 K2 Aided by equations 3-15, 3-16, 3-17 and 3-18, the optimal performance of a PLD is described as: Equation 3-19 (__L-1)-_!::__ .!!..J._+ K1 D =I inwhichf/K1 >1,andK2>K1 D (__L _L_) Kl K2 38

PAGE 57

Equation 3-20 H2=D-H1 Equation 3-19 is valid whenj/K1 > 1 and K2>K1 In other words, the infiltration rate is greater than the seepage rate and the sand-mix layer is above the gravel layer. Equations 3-19 and 3-20 are derived to be the guidance to divide the total required filtering thickness into two layers. 3.4 OPERATION OF POROUS LANDSCAPING BASIN The operation of a PLD is subject to clogging due to sedimentation in the basin. The pollutant deposit is often accumulated on top of the basin bottom and then diffused into the top layer of the sand-mix medium. When the infiltration rate decays on the basin bottom, the friction losses are reduced accordingly. Namely, Equation 3-21 dHI = f.,Hl Kl Equation 3-22 dHz = fsHz Kz In which.fs is the reduced infiltrating rate due to clogging effect. Clogging to a PLD system generally occurs as a thin, hard "cake" layer (0.2 to 0.5 em) of sediment on the basin bottom. As reported, migration of solids diffuses into the top 5 to 10 em of the sand-mix layer while the hydraulic conductivity in the lower sand-mix layer remains unchanged (Li and Davis 2008a and 2008b; Mays 2005). As illustrated in Figure 3-3, the reduced infiltration flow under a clogging condition may not completely consume the hydraulic head available. As a result, the residual pressure in the system, is calculated as: Equation 3-23 H, = Y +D-dH1 -dH2 39

PAGE 58

where H1 is the residual pressure head. If the cake layer presents an additional friction loss, the value of H1 will decrease; otherwise it represents the pressure built up in the filtering layers. The reduced infiltrating rate implies a prolonged drain time, or a period of standing water in the PLD, as: Equation 3-24 D T =--T w fs d in which T w is the increased drain time or period of standing water. As the basin bottom is gradually clogged, the infiltration flow rate continues to decrease while the residual pressure continues to increase. When the residual pressure is becoming to equal the available head, the system will about to cease functioning because the basin is completely plugged (Guo et al 2009). 3.5 LABORATORY TESTS 3.5.1 PREVIOUS DESIGN STUDIES Lab studies focusing on stormwater BMPs similar to the PLD have been limited. Additionally large column studies are uncommon due to many variables affecting infiltration in surface soils. The complexities of systems as they exist in the field are challenging to replicate and measure in lab conditions. Generally, in order to isolate one or several factors, column studies are completed with small diameter columns(< 25 em or <10 inches) in controlled lab conditions. Soil column studies focusing on physical characteristics for measurement of saturated hydraulic conductivity (Davis et al. 2001; Mays and Hunt 2005; Yang et al. 2004) are by up flow. Costs for this type of systems are cost prohibitive and researchers generally use one column. Additionally field conditions are best replicated by down flow. Studies focusing on unsaturated conditions generally use smaller diameter columns and add water from the top for down flow (Hsieh and Davis 2003; Siriwardene et al. 2007). Research completed by the team at the University of Maryland utilize small 40

PAGE 59

columns and growing boxes and apply synthetic stormwater to the top (Davis 2007; Davis et al. 2001; Davis et al. 2006; Davis et al. 2003). A large diameter column has the advantage of replicating field conditions and reducing the effects of sidewall flow (Corwin 2000; Hunt 2003). A few studies of soil characteristics have utilized larger columns (>30 em or > 12inch diameter) (Corwin and LeMert 1994; Sun 2004). Even fewer studies related to BMPs have utilized large diameter columns. The studies are presented in Table 1-3 below. Table 3-1 Large column studies of systems similar to PLD Author, Study Column Size Measurement equipment Year (Hunt 2003) column anoxic D= 30.5cm Outflow with graduated zone (12") cylinder and timer H=122 em (48") (Ames, Column for D=61 em Pressure transducer (flow rate) Inkpen et at. infiltration (24") Water content reflectometer 2001) rates of H= 91 em Temperature probe selected media (36") Data logger Sun 2004 uptake of D=31cm (12") Soil cores, vegetation and metals by H=31cm (12") water samples collected grasses Generally methods for these include adding water from the top, measuring inflow and outflow, and collecting water and or soil samples. Some studies also collect data about the moisture content and water suction in unsaturated conditions (Ames et al. 2001 ). None of the studies have collected comprehensive data about infiltration capacity, clogging and water quality. Ames et al had difficulty reproducing infiltration rates between tests (Ames et al. 2001) 3.5.2 INFIL TROMETER DESIGN AND TESTING The design for this study included a large diameter PVC column, pressure ports and a variable height outflow shown in Figure 1-3 and Figure 1-4. The large diameter cylinder (38 em) which was >600 times the particle size was used to reduce the influence of short circuiting. Pressure ports and manometers were installed to measure changes in pressure through the filtration soil-mix. Each column was equipped with an overflow to maintain constant head of 30 em (12 in) when necessary. The outflow at the bottom was designed to change depths from below the 41

PAGE 60

course aggregate to above the aggregate as in Figure 1-5. The elevated outflow served to saturate the filtration layers. 38" .--..l(._-----, 12 inches of water capacity M a n o m ete r s Figure 3-4 Soil Column Design with 2-Layered System and Lowered Outflow for Field Conditions 42

PAGE 61

I 38" .---"--------., 1 2 inche s of water capacity M a nom e ter s Figure 3-5 Soil Column Design with Elevated Outflow for Saturated Conditions Previous reports recommended that a soil-mix thickness of 30 to 60 em (or 12-24 inches) can effectively remove pollutants in stormwater (Hunt 2003; NCDENR 2007; Sun 2004; Toronto and Region Conservation Authority 2007). As reported, metals are removed in the top 20 to 45 em (8 to 18 inches) of soil-mix layer (Davis et al. 2003; Sun 2004; Winogradoff 2001). In practice, the soil-mix layer thickness is recommended to be 45 em (18 inches) to allow for both adequate pollutant removal and root zone for vegetation. In the laboratory, a 38-cm (15-inch) circular infiltrometer in Figures 3-4 and 3-5 was utilized to represent a section of the PLD A view of the inside of the column is presented in Figure 3-6. Soil sample columns were prepared to mimic the field conditions as closely as possible. As illustrated in Figure 3-5, the sample column in the infiltrometer is built with an upper filtration layer of soil-mix, a lower layer of gravel, and a perforated bottom drain. Thel.9 em (% -inch) crushed granite was spread in the bottom. ASTM C33 washed sand and Canadian peat were combined at the ratio of 85% sand and 15% peat by volume for the soil-mix layer. To evaluate the effect of the geotextile on the infiltration rates, Sample Column I had a geotextile layer separating the soil mix layer from the aggregate layer; and sample Columns J and F were constructed and tested without the geotextile. Figure 3-7 shows the geotextile set inside the soil column. 43

PAGE 62

Figure 3-6 Infiltrometers Built in Laboratory. Figure 3-7 Measurements for Soil and Water Levels Inside the Soil Column 44

PAGE 63

Figure 3-8 Geotextile between Large Aggregate and Filter Layer Inside the Column Manometers were installed at the ports seen in Figure 3-6 and 3-7. Inside the columns the manometers consisted of a hose barb, flexible tubing and a cover, as in Figure 3-9. The outside of the column flexible tubing allowed for a static pressure measurement. Figure 3-9 Manometer Setup to be Placed inside the Column Table 3-2 presents the compaction and density of the soil-mix samples prepared for infiltrometer tests. For this study, the filtering layers were structured with an upper 45-cm (18-inch) sand-mix layer and a bottom layer of 20-cm (8-inch) gravel. As illustrated in Figure 3-6, the total thickness of a sample column is set to be 65 em (26 inches). With a constant head of 30 em (12 inch), the total head applied to the infiltrating flow is 95 em (38 inches) as shown in Table 3-2. 45

PAGE 64

Table 3-2 Compaction and density of soil-mix sample Soil-mix Sample Density grams per cubic centimeter (pounds per cubic foot) Compaction Material Loose Soil-mix Compacted Soil-mix Ratio Sand 1.39 (86.79) 1.76 (1 09 .65) 1.26 Peat 0.42 (26.05) 0.84 (52.43) 2.01 85% Sand and 15% Pea 1.16 (72.31) 1.75 (109.22) 1.51 60 Test Flow Rates Over Time Column 11 'C' 50 .c r2 = 0.85 Column F e 40 u A Column J G) 30 ... 20 -Horton's 1 ;:: Estimate 1 0 10 0 0:00:00 24:00:00 48:00:00 72:00:00 hours:min:sec Figure 3-10 Variation of Infiltration Rates for Sample Columns (Control: Peat and Sand) Manometers were installed on the infiltrometer wall at 4 stations to measure the variation of static heads. The locations of manometers are expressed by the vertical distances above the bottom of the sample column as shown in Table 3-2. 46

PAGE 65

Table 3-3 Variation of hydraulic heads measured at 5 stations Column Infiltrating Locations of manometers above ground in em Sample Rate ID 66.0 61.0 27.9 22.9 entrance upper lower geotextile cm/hr sand layer sand layer sand layer layer Reading in manometers above ground in em Col F without geotex 22.9 97.8 95.3 31.1 22.9 Col I with geotex 24.3 96.5 94.0 29.2 22.9 Col J without geotex 25.9 96.5 94.6 48.6 22.9 Col F without geotex 8.8 97.8 96.2 27.9 22.9 Col I with geotex 12.4 96.5 94.6 27.9 22.9 Col J without geotex 13.5 96.5 95.3 30.5 22.9 Note: Col F without geotex IS the Soil Column F Without geotextile measured after 72 hr operation. Table 3-4 Hydraulic conductivity in filtration layer after 72 hours of water flow Column Sample ID Infiltration Rate Hydraulic Conductivity for Soil-mix cmlhr cmlhr Col F without geotex 8.8 4.9 Col I with geotex 12.4 6.1 Col J without geotex 13.5 7.3 The sample columns were tested with the "wet" initial condition by filling each column to 30.5 em (12 inches) of water and soaking the sample column overnight with the drain valves closed. The next morning the valve was opened and outflow measurements were recorded. The outflow rates from sample columns were recorded for 72 hours continuously. Figure 3-9 is the plot of the decayed infiltration rate through the sand mix layer over 72 hours. The best fitted Horton's infiltration equation for the soil mix tested is (Horton 1933): Equation 3-25 fr = 9.8 + 37.2e-o.J44t Where ft is the infiltration rate in cmlhr at time t in hours, and k = decay constant in 1/hour. Equation 3-25 has a correlation coefficient, r2 = 0.85. The differences in the 47

PAGE 66

infiltration decay among sample columns are attributed to sample preparations, compaction condition, and test operations. Soil compaction alone can have a significant impact on infiltration rates in sandy soils (Pitt et al. 1999). Although the impact of the geotextile fabric on the flow rate was not measurable through the sample columns, when the columns were deconstructed the soil-mix particles were found to migrate into the gravel layer. This visual observation has was also found by Haliburton and Wood. (Haliburton and Wood, 1982). Based on three soil column tests, the average final infiltration rate is found to be 11.6 crnlhr (or 4.6 inch/hr). The hydraulic conductivity coefficients were found to be 6.1 crnlhr (or 2.3 inch/hr) for soil mix and 64.0 cm/hr (or 25.2 inch/hr) for gravel. Results listed in Table 3-3, these values are within the range reported before (Schwartz and Zhang, 2003). Literature reports of hydraulic conductivities in bioretention systems are 1.3 to 15 centimeters per hour (crnlhr) (0.5 to 6 inches per hour (inlhr)) (Davis et al. 2001; Hunt 2003; Hunt and White 2001). During field tests of installed PLDs Hunt found a range of 1.3 to 3.3 crnlhr (0.5 to 1.3 inlhr). The minimum design criteria set by the UDFCD is 2.5 cm/hr (1 inlhr) (UDFCD 1999). 3.6 DESIGN EXAMPLE AND SCHEMATICS The PLD basin located in the City of Denver, Colorado is employed as an example to illustrate the design procedure. The catchment draining into this basin has a tributary area of 1.0 hectare (2.5 acre) and runoff coefficient of 0.60. As reported, the average event depth for the Denver metropolitan area is 1.0 em (or 0.4 inch) (EPA 1986). With a runoff volume capture rate of 80%, the WQCV for this basin is calculated as: Equation 3-26 -P; -o.25 a= e pm = e ----u> = 0.79 Equation 3-27 -Po -Po Cv =1-ae0:" =1-0.79e0 60xl.O =0.8 orP0=0.86cm Equation 3-28 V0 = P0A = 0.86 em xl.O ha =86m3 48

PAGE 67

From the laboratory test,fis 11.6 cmlhr (or 4.6 inchlhr), K1 is 6.1 cm/hr (or 2.4 inchlhr) for soil mix and Kz is 64.0 cm/hr (or 25.2 inch/hr) for gravel. Consider a drain time of 6 hours. The required dimension for the subsurface filtering system is calculated as: Equation 3-29 D = JTd = 11.6x6 = 69.6 em Equation 3-30 H = Y + D = 30.5 + 69.6 = 100.1 em or H1 = 50.1 em and H2 = 19.5 em Substituting the above dimension into equations 3-17 and 3-18 yields Equation 3-32 f 11.6 dH1 =-H1 =--X50.1=95.3em KI 6.1 Equation 3-33 f 11.6 dH2 =-H2 =--x19.5=3.5em K2 64.0 As expected, the total friction losses for the seepage flow through the two layers satisfy equation 3-15. This is the optimal performance for this PLD basin, according to equation 3-19. In practice, the infiltration rate at the site is estimated with uncertainties. Secondly, the infiltration rate is subject to the clogging effect. For comparison, the performance of this PLD is further assessed for the condition that the 49

PAGE 68

infiltration rate is reduced to 7.5 em/hr. The corresponding head losses are calculated as: Equation 3-34 dH fsHI 7.5x50.1 61 6 1--. em K1 6.1 Equation 3-35 dH fsH2 7.5xl9.5 2 3 2--. em K2 64.0 Equation 3-36 H, = H dH 1 -dH 2 = 100.161.6-2.3 = 36.2 em Equation 3-37 Tw = .!!._Td = 69 6 6.0 = 3.3 hr --an extended period of standing water. fs 7.5 Under the clogging condition, this PLD drains slowly. After the design drain time of 6 hours, standing water is developed in the basin. The residual pressure head, H1 is either absorbed in the clogging (cake) layer or built up in the soil medium. The above analysis was repeated for the bioretention medium studies (Li and Davis 2008a and 2008b). In the laboratory, the clean sand column was used as the subsurface medium to filter the solids in storm water. As listed in Table 3-4, a 0.1 to 0.6-cm cake layer was observed to form on the basin bottom. The top layer of the soil column is gradually clogged by the migrated solids. Table 3-4 summarizes the observed thickness of the top clogged layer. The clean sand column started with a hydraulic conductivity of 45 em/hr. After the top layer is clogged, the equivalent hydraulic conductivity coefficients for the soil column were measured as listed in Table 3-4. Consider that the clogged top layer is equivalent to a sand-mix layer and the bottom clean layer is equivalent to a gravel layer. Equations 3-19 were applied to 3-7 case studies to divide the total medium thickness into the top and bottom layers. Figure 3-10 presents good agreement between the calculated and observed thickness of the top layer. 50

PAGE 69

Table 3-5 Observed case studies for bio-retention medium (Li and Davis 2008a and 2008b) Infiltration Total Conductivity Coefficients Thickness Flow Filtering Top Bottom Equivalent Top Bottom Cake Rate Depth Layer Layer Layer Layer Layer f D Kl K2 Ke cm/h em cmlh cmlh em em em em 4.80 5.50 2.00 45.00 3.00 3.50 2.00 0-0.3 4.90 10.50 2.00 45.00 3.00 7.00 3.50 0-0.4 9.50 5.50 2.00 45.00 4.00 4.10 1.40 0-0.3 9.50 10.50 2.00 45.00 3.00 7.40 3.10 0-0.1 19.80 5.50 2.00 45.00 5.00 3.70 1.80 0-0.3 19.70 10.50 2.00 45.00 6.00 6.70 3.80 0-0.6 20.10 5.50 2.00 45.00 7.00 4.30 1.20 0-0.2 PLD SubBase Dimension i 0.80 I I ftl 0.70 ...J 0.60 1 0.50 ftl ... 0 IE 0.40 0 = 0 0.30 m :::1: 0.20 -N :::1: 0.10 0.00 0.00 0.20 0.40 0.60 0.80 1.00 i --Optimal Dimension H1/ H (Top Sand Layer) Figure 3-11 Comparison between Observed and Calculated Filtering Thickness 51

PAGE 70

3.7 CONCLUSION A PLD should be designed using the concept of hydrologic system to take both the surlace and sub-surlace flows into consideration. In this study, it is recommended that the PLD be sized for the on-site storm water quality control volume according to the selected drain time and target pollutant removal rate. The total filtering thickness underneath the PLD shall be determined by the selected drain time and infiltration rate. The filtering layers beneath a PLD shall be structured to completely consume the hydraulic head available in the system. The optimal dimension of the sub-base medium is found to be closely related to the design infiltrating and seepage rates. In this study, equation 19 was derived to provide the optimal thickness ratio between the sand-mix and gravel layers. The optimal sub-base dimension delivers the highest seepage flow rate for the unclogged condition. Equation 3-38 is numerically sensitive to f/K1 but not to f/K2 because the hydraulic conductivity coefficient of gravel is usually much higher than the infiltrating rate or the ratio, f/K2 which is numerically close to zero. For simplicity, the thickness for the sand-mix layer is approximated as: Equation 3-38 H1 =(1-.!.i_)D+.!.i_Y f f In practice, it is critically important that the ratio, f/K1. is properly selected to avoid undesirable prolonged standing water in the PLD. In this study, the infiltration rate for the soil-mix layer varies from 50 to 7.5 cmlhr (20 to 3 inchlhr). The final infiltration rate is approximately 7.5 to 12.5 cmlhr (3 to 5 inchlhr) after an operation of 72 hours. The hydraulic conductivity coefficient was varied within a small range through the soil-mix column. All these uncertainties are attributed to the residual pressure in the PLD system. As a common practice, perlorated pipes are installed in the subsurlace system. A sub-drain pipe creates an accelerated hydraulic gradient to collect the excessive water and to alleviate the build up pressure. 52

PAGE 71

4 WASTE-INCORPORATED BENCH SCALE TESTBARE SOIL This chapter expands on the paper "Waste Incorporated Sustainable Design of Stormwater Detention Basins 2. Bench Scale Tests" by Shauna M Kocman, James C. Y. Guo, Anu Ramaswami submitted to ASCE Journal of Environmental Engineering April2010. 4.1 INTRODUCTION AND BACKGROUND Sustainability as it relates to urban water must take into account both quality and quantity of water resources, as well as beneficial reuses of urban waste streams now and into the future. Low impact development (LID) techniques take advantage of micro-scale approaches to mold the development of land to function similarly to natural drainage systems, thus replicating ecosystem services which the open space would have performed (Sample and Heaney 2006). PLDs are designed based on water quality capture volumes from the contributing watershed to reduce peak flows and optimize sediment removal rates (Guo and Urbonas 1996b; Guo and Urbonas 2002). Studies have shown PLDs enhance stormwater by reducing peak runoff volumes, and by providing filtration of sediments and absorption of pollutants. (Davis et al. 2001; Davis et al. 2006; Davis et al. 2003; Hunt 2003; Hunt et al. 2006). The top layer of the PLD consists of a vegetated porous detention basin underlain by a drain.(USWDCM 2001). This chapter investigates the impact of varying the soil-mix filter layer based on the performance and behavior of the system. While many studies have been done on PLDs (Davis 2007; Davis et al. 2006; Davis et al. 2003; Guo et al. 2009; Hsieh and Davis 2005b; Hsieh and Davis 2005c; Kim et al. 2003; Sun and Davis 2007), traditional studies have focused on measuring infiltration rates and pollutant removal, but few have combined other performance characteristics such as sediment accumulation and pathogen removal rates. Although waste reuse is not a new concept, no previous study has taken a combined look at performance criteria and environmental benefits of a waste-incorporated enhanced PLD. 53

PAGE 72

4.2 BENCH SCALE TEST METHODOLOGY This research focuses on performance and sustainability of vegetated, waste incorporated PLD. The physical performance of the systems focuses on both hydraulic conductivity of soil layers and infiltration capacity of the system, both of interest for field design and maintenance recommendations. Additionally mass removal is an important measurement of water quality performance. The water quality portion focuses on contaminate removal. Fate and transport models include advection, dispersion, and diffusion equations which are complicated and not practical for field conditions. Field conditions include complexities such as root growth, metals sorption, and sedimentation which complicate the system. Therefore mass accounting will be used for effectiveness of contaminate removal. Infiltration capacity is described by Horton's infiltration equation (Horton 1933). As water is applied to the soil surface the infiltration rate reduces infiltration capacity by an exponential decay function as presented in Equation 4-1. Equation 4-1 f, = fc + (fofJe-kt Wheref, is the infiltration rate at timet [Ut],Jo is the initial infiltration rate [Ut], fc is the infiltration capacity or equilibrium infiltration rate after the soil has been saturated [Ut]. The variable k is the decay constant specific to the soil [ 1/t]. Over time the infiltration capacity depends on soil properties and assumes a maximum capacity. The infiltration capacity documented by Pit et alto depended heavily on compaction (Pit et al. 1999): "The effect of compaction on sandy soils is very large, reducing the infiltration rates by between 5 and 10 times." Through previous waste screening and batch testing permissible mixtures of waste materials to be incorporated into the PLD were defined. In this study the control and two permissible waste-incorporated mixtures were further evaluated through bench scale testing. The peat and sand mixture, currently recommended by the Urban Drainage Flood Control District for the Denver Metro area, served as a control (USWDCM 2001). The large (15 inch, 38 em) diameter infiltrometer designed for 54

PAGE 73

this study was utilized in the laboratory to represent a section of the recommended field installation of the PLD. The columns were packed to replicate the current recommendation of 45.7 em (18 in) of filtration media and 20.3 em (8 in) of aggregate under drain (Guo et al 2009). The under layers were consistent between all the columns including the geotextile, gravel layer, and PVC under drain. To avoid pressure back-up, the under drains were overdesigned to drain above the infiltration rate. The performance of the selected material mixtures in terms of infiltration and contaminate removal were measured and assessed. Stormwater was applied to the packed columns and monitored for performance criteria (flow rates, clogging, water quality parameters and sediment buildup). Triplicate columns in Figure 4-1 were created with filtration layers consisting of the following three mixtures: 1. Control (15% peat and 85% sand) 2. CPS (7.5%compost, 7.5% shredded paper and 85% sand) 3. CPST (7.5% compost, 7.5% shredded paper, 8% tires and 77% sand) Figure 4-1 Setup of Triplicate Soil Columns and Barrels of Stormwater in the Lab Duplicate bare soil bench tests were completed. The first consisting of a leaching and a filtering phase and the second a filtering phase only. Clean tap water was added to the system initially to measure for any leaching from the materials. Then synthetic stormwater was added to measure filtering of contaminates. The flow rate and water quality parameters were monitored throughout both the leaching and the filtering phases. During step 1 an initial saturated flow rate was measured by elevating the outflow of the soil columns and adding tap water for 72 hours.. The initial condition acts to saturate the column and measure any leaching of sediment, chemicals, or pathogens from the media mixture that may occur. As presented in Figure 4-2, after the columns were saturated the flow rates of triplicate samples of all three mixtures, except one outlier, approach 8.3 to 13 centimeters per hour (crnlhr) (3.3 to 5.2 inches 55

PAGE 74

per hour (inlhr)). The current design minimum requirement is 2.5 crnlhr (1 inlhr), which all three mixtures obtain. The current design minimum requirement is 2.5 crnlhr (1 inlhr), which all three mixtures obtain. An initial saturated flow rate for the first bare soil test with new media-mixture was measured during 72 hours of elevated outflow. The infiltration rates over the 72 hour duration are plotted in Figure 4-2 and follow Horton's infiltration model stated in equation 4-1 (Horton 1933). 45.0 40.0 35.0 30.0 Clean Water Infiltration Rate Control r2 = 0.90 CPS r2= 0.83 CPST r2 = 0.66 100.0 80.0 ....... 25.0 i 20.0 60.0 15.0 40.0 10.0 i _._ 20.0 5.0 .. .. 'lit 0.0 +-------.,-------or-------,......... 0.0 0 24 Control (Pea!,Sand) X Compos!, Paper, Sand, Tires Horton Curve for Compos!,Paper, Sand 48 Elapsed Time hours Compos!, Paper, Sand Horton Curve for Pea!,Sand 72 --Horton Curve for Compos!, Paper, Sand, Tires u Figure 4-2 Variation of Infiltration Rates as the Filtration Mixtures Approach Saturation after 72 hours of Continuous Flow Horton's estimate for infiltration rate over time is based on the equation: Equation 4-2 f(t) = fc +{fo-fc) e(-k)l Wheref(t) is the infiltration rate at time t,f0 is the initial infiltration rate [L3/t] andfc is the infiltration rate at field capacity [L3/t]. The k is a constant with units 1/time. 56

PAGE 75

The best fit equations for the infiltration rate of the three mixtures shown in Figure 42 are: Equation 4-3 Peat and Sand : fconrrot (t) = 9.9 +(50 -9.9)e<-o.o5>1 Equation 4-4 Compost, Paper, Sand :fcps (t) = 14.5 +(56 -14.5)e<-o.o9>1 Equation 4-5 Compost, Paper, Sand, Tires: fcpsr (t) = 9.1 +(35 -9.1) e<-O.D4)r Where/control (t) is the infiltration rate [cm/hr] of the control (peat and sand) mixture at timet. /cps (t) is the infiltration rate [cm/hr] at timet of the compost, paper and sand mixture and thefcpst (t) is the infiltration rate [cm/hr] at timet of the compost, paper, sand and tires mixture. In equations 4-3, 4-4 and 4-5 fo is expressed in cm/hr, fc is the infiltration rate at field capacity expressed in cm/hr, and t is the elapsed time in hours. The k is a constant with units 1/hr. Equations 4-3, 4-4 and 4-5 have correlation coefficients of r2= 0.90, 0.83, 0.66. After the conclusion of the initial "clean" water leaching phase, synthetic storm water was applied to the column in individual applications of 30 em (12-inch) depths to represent runoff events which fill the PLD. The stormwater used throughout the study was collected from an storm sewer outfall N-431E on the South Platte draining from Denver. Urban runoff was collected from this outfall in 30 gallon plastic drums and transported to the lab. This outfall carries approximately 2 cubic feet per second (cfs) dry weather flow and has a 475 cfs capacity for wet weather flows. The outfall drains 80% of the Lower Platte Valley Basin (4.47 square miles) consisting of mixed industrial, commercial and residential land uses. A picture of the outfall is presented in Figure 4-3. 57

PAGE 76

Figure 4-3 Outfall N-431E in the South Platte River Where Stormwater for the Experiment was Collected The summary of a 4 year record of dry weather sampling for outfall N-431E is presented in Table 4-1. The total suspended solids (TSS) results during the four years of sampling data have a mean 61 and a standard deviation of 190 mg/L. The average concentrations for the Denver area are presented in Table 1-9 for comparison (USWDCM 2001 ). After additional sediment is added and storm water is applied to the columns, an influent and effluent sample was collected and analyzed for TSS. Table 4-1 Average water quality results of storm water outfall N-431 E Sample Min Max Average pH 7.7 8.4 8.1 TSS (mg!L) <1 990 61 TKN(mg/L) >1.0 6.59 1.5 Nitrate +Nitrite (mg!L) >.5 5.66 3.0 Total P (mg!L) .08 1.77 .3 E-coli per lOOml 80 20,000 2,745 58

PAGE 77

Table 4-2 Average event mean concentrations for Denver area Sample Industrial Commercial Residential Undeveloped TSS (mg/L) 399 225 240 400 TDS (mg!L) 58 129 119 678 TKN(mg/L) 2.7 3.3 3.4 3.4 Nitrate +Nitrite 0.91 0.96 0.65 .5 (mg/L) Total P (mg/L) 0.42 0.43 0.65 0.4 Total Copper (ug/L) 84 43 29 40 Total Lead (ug!L) 130 59 53 100 Total Zinc 520 240 180 100 (ug/L) Stormwater was applied with the same characteristics and equivalent volumes as urban runoff defined by Guo and Urbonas (2002). The water was spiked with additional sediment to accelerate the clogging process in order to relate surface loading to field conditions where the clogging process occurs over years. Water quality samples were collected for the duration of the storm water applications. Water quality samples were collected and analyzed for pH, total suspended solids (TSS), total dissolved solids (TDS), pathogens, total keldjal nitrogen (TKN), nitrate plus nitrate (N02+N03), total phosphorous (TP), and total metals. The nutrient and metals were analyzed at Metro Wastewater Reclamation District's laboratory. The total metals sweep (Beryllium, Chromium, Manganese, Nickel, Copper, Zinc, Arsenic, Selenium, Molybdenum, Silver, Cadmium, Antimony) were analyzed by EPA 200.8 ICPMS methods. Total coliforms were measured as an indicator of concentration of pathpogens and were both analyzed in-house and at Industrial Labs by the membrane filtration method (Clesceri et al. 1998). The pH, TS, TSS, TDS and pathogens were analyzed at the Auraria Campus laboratory in Technology building following the Standard Methodology for Examination of Water and Wastewater guidelines (Clesceri et al. 1998). Two separate tests were conducted with stormwater. The first consisting of seven applications and the second with twelve applications of stormwater. After 7 stormwater applications and a total of 1.1 kg/m2 of sediment was accumulated, soil cores were collected. Then the top 10 em of filtration layer was removed and the second test was started. The soil core samples were analyzed for density stratification and particle size distribution. Samples of clean soil after compaction and dirty soil after stormwater 59

PAGE 78

applications were collected from each column at depths of 0-1 em, 4-6 em and 8-10 em. Soil cores were collected with a 1.45 em diameter tube. The samples were dried and weighed to calculate the density. Sieve analysis was conducted for particle size distribution by ASTM D-421 Standard Practice for Dry Preparation of Soil Samples for Particle-Size Analysis and Determination of Soil Constants (2009). 4.3 WATER QUALITY IMPACTS During the initial clean water test and subsequent stormwater applications, the systems were monitored for leaching and filtering of nutrients, metals, pathogens and suspended solids. All treatments initially leached total keldjal nitrogen (TKN) and total phosphorous (TP) during the clean water application and ultimately filtered TKN and TP from the system. The average removal rate of TKN was between 32% and 44% with an inflow of 1.7 mg/L and between 71% and 77% removal with 2.6 mg/1. Figure 4-4 presents the leaching of TKN during the application of clean water and then removal of TKN during the filtration of stormwater. Leaching and Filtering of TKN 0.5 0 Tap Water 0 Tap Water Final Stormwater 0 .33 Stormwater 2.66 Accumulative Sediment Load (kglm2 ) I control 11lCPS I gCPST 1 Figure 4-4 Plot ofTKN Concentration in Water for Various Accumulated Sediment Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66 kg/m2. The Results Showed No Consistent Statistical Difference Between the Waste-Incorporated Media Mixes (CPS, CPST) Versus the Control for Leaching or Filtering of TKN 60

PAGE 79

Initially all three mixtures leached TP and then filtered TP. Between 48% and 83% of TP was removed from an inflow of 0.62 mg/L and between 80% and 86% was removed from an inflow of 1.3 mg/L. The inflow and outflow concentrations are presented in Figure 4-5. Leaching and Filtering of Phosphorous 1 ... 0.8 11. j 0.6 -J--------------------F::-;:f-------1 ;;1-------1 Tap Water 0 Tap Water Final Stormwater 0 .33 Stormwater 2.66 Accumulative Sediment Load (kglm2 ) Dlnflow I conlroll l';ICPS Figure 4-5 Plot of Total Phosphorous Concentration in Water for Various Accumulated Sediment Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66 kg/m2. The Results Showed No Consistent Statistical Difference Between the Waste-Incorporated Media Mixes (CPS, CPST) Versus the Control for Leaching or Filtering ofTP. All the treatments filtered copper, even from the "clean" tap water. Copper may leach into tap water from copper pipes in a building (Denver Water 2009). The removal rates varied between 45% and 81% during an inflow of 21 ug/L and between 93% and 95% with the application of 113 ug/L. The difference in removal rate between the control and the two treatments (CPS and CPST) is not statistically significant. The inflow and outflow concentrations of copper in ug/L are presented in Figure 4-6. 61

PAGE 80

Filtering of Copper 270 mg/L Cu .. 113 mg/L Cu ' 50 45 :::1 40 "a, 35 2. 30 Dlnflow ... Ql Control a. 25 a. 0 cPS (J 20 ]i I!ICPST 0 15 1-10 5 0 Tap Water Tap Water Final Stormwater Stormwater 0 0 .33 2.66 Accumulative Sediment Load (kg/m2 ) Figure 4-6 Plot of Total Copper Concentration in Water for Various Accumulated Sediment Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66 kg/m2. The Results Showed No Consistent Statistical Difference Between the WasteIncorporated Media Mixes (CPS, CPST) Versus the Control for Leaching or Filtering of Copper. During the clean water application total coliforms were initially leached from the systems. The control (peat and sand) leached the most total coliforms (950 cfu) while the compost, paper and sand mixture added the least (90 cfu) to the outflow. Visual observation in the lab confirmed the compost did have high microbial activity but testing showed few coliforms. During the composting process heat is generated and pathogens are killed. Certified compost must meet strict pathogen limits before sale (US Composting Council 2009; Yost 2008). The runoff water was high in total coliforms (5,700 to 60,0000 cfu). Additionally, one stormwater application was spiked with bird droppings (26,000 cfu) and the filtering capacity of the systems was measured with initial and final outflow samples. The results indicate a high filtering capacity for total coli forms as presented in Figure 4-7. Removal rates varied between 88% and 92% with the application of 5,700 cfu and between 95% and 99% with application of 60,000 cfu. 62

PAGE 81

Pathogen Leaching and Filtering 60,000 cfu + ' '5" cfu :2. 10000 !! ..... iE .----------, 0 inflow 6000 I Filtering Control J 1 3000 r--i :: =:I m fll [i tap water final tap water storm water storm water storm water 0 0 .33 .64 2.66 Accumulative Sedfmant Load (kglm2 ) Figure 4-7 Plot of Total Coliform Forming Units Water for Various Accumulated Sediment Loads Including Tap Water (No Sediment) and Stormwater with 0.33 and 2.66 kg/m2. The Results Showed that the Control (Peat and Sand) Leached the Most Coliforms. All Mixtures Showed Consistently High Pathogen Removal Rates Suspended sediment is a contaminant also and a goal of PLD is to filter total suspended solids (TSS). The removal rate of solids from the systems varied between 93% and 100% during test 1 and test 2. The sediment loaded in the lab was increased to accelerate the clogging process and then related to sediment loads in stormwater in the field. As presented in Table 4-3, a total of 1.08 kg/m2 at an average of .15 kg/m2 per application were filtered through the columns during test 1 and a total of 6.06 kg/m2 at an average of .5 kg/m2 per application during test 2. 63

PAGE 82

Table 4-3 Sediment loading to the soil column surface during the experiment Number of Test 1 Test 2 Storm water Average Standard Accumulative Average Standard Accumulative Applications Deviation Load Deviation Load TSS (kwm2 ) TSS (kg/m2 ) 1 0.10 O.ot 0.10 0.16 O.ot 0.16 2 0.08 0.01 0.19 0.33 0.02 0.49 3 0.07 0.00 0.26 0.18 0.01 0.66 4 0.07 0.01 0.33 0.33 0.02 0.99 5 0.05 0.02 0.38 0.50 0.02 1.49 6 0.26 0.05 0.64 0.61 0.02 2.10 7 0.44 0.02 1.08 0.55 0.02 2.66 8 0.58 0.01 3.24 9 0.56 0.01 3.80 10 0.57 0.05 4.36 11 0.86 0.08 5.23 12 0.83 0.05 6.06 During testing in the lab it became evident that an additional water quality impact may come from the filtration mixture. As seen in the screening tests the lighter particles, such as tires and organics, float. When mixed with sand, the particles are trapped in the lower layers by the heavier particles and will not float. When the top layer is subject to mixing from rain the lighter particles are dislodged and float to the top, as shown in Figure 4-8. Floating material in stormwater, such as leaves, twigs and cigarette butts combine with the light materials in the filtration mix and may be carried downstream in the event of overflow or plug the outlet. 64

PAGE 83

Figure 4-8 Picture of Floating Particles from the Filtration Mix. The PaperTire Deposit Shows How Density Stratification Occurs When Floating Particles Settle With each filling of the basin, light materials float and become incorporated with the sediment from the stormwater. Density stratification occurs and the upper layer becomes less dense than the underlying layers. Density tests conducted on soil samples at 3 depths indicate that the top 0-1 em of soil become less dense than lower layers after application of stormwater. Figures 4-9, 4-10 and 4-11 show the difference in the density of the clean and dirty layers at different depths. The top layers of the final dirty mixtures (0-1 em) are less dense than mixture at lower depths. 65

PAGE 84

Peat and Sand Density Stratification 1.80 ,.------------------, i 1.60 s 1.40 t t 1.20 1.00 i ---... Ill 0.80 0 0.60 ii 0.40 0.20 --.r--I-.. 0.00 +---..-----..-----,.------,.------1 0 2 4 6 8 10 Depth of Soli (em) Figure 4-9 Plot of Density of Soil Samples in Depths 0-10 em of the Soil Column. Results Show Lighter Material in the Top Layer of the Soil Shows Density Stratification of Control Compost, Paper and Sand Density Stratification 1.80 .----------------, 8 1.60 +--------........ s 1.40 i 1.20 .. = 1.00 +------------------1 111080+-----------------1 0 >-0.60 +-----------------1 0.40 +-----------------1 .. 0.20 0.00 0 2 4 6 8 10 Depth of Soli (em) Figure 4-10 Plot of Density of Soil Samples in Depths 0-10 em of the Soil Column. Results Show Lighter Material in the Top Layer of the Soil Shows Density Stratification of CPS 66

PAGE 85

Compost, Paper, Sand and Tires Density Stratification 1.80 .-----------------, 1.60 1.40 H 1.20 --i--1.00 1-c 0.80 - -8 0.60 0.40 0.20 f 0.00 0 2 4 6 8 10 Depth of Soil (em) Figure 4-11 Plot of Density of Soil Samples in Depths 0-10 em of the Soil Column. Results Show Lighter Material in the Top Layer of the Soil Shows Density Stratification of CPST As the water infiltrates, sediment is accumulated in the matrix of the media, creating a cake layer. This accumulation on the top layers reduces the infiltration rates. It was observed that this cake layer includes both stormwater sediment and light particles. The cake layer formed by the sediment accumulation and lighter material is shown from the density stratification and the sieve analysis. As presented in Table 4-4, the top layer (0-1cm) is consistently lighter than the lower layers, indicating lighter particles floated and settled into the top layers. Table 4-4 Table of density of soil sample samples at various depth (0-10 em). Results show lighter material settled on the top soil layer resulting in density stratification Control Compost, Paper, Compost, Paper, Sand Sand, Tires Soil Sample Depth Density (glee) 0-1 em 1.0 1.2 1.0 4-6 em 1.3 1.5 1.5 8-10 em 1.5 1.5 1.6 67

PAGE 86

4.4 CLOGGING EFFECTS The clogging process can be measured from reduction in infiltration rates and observed from sediment accumulation in the filtration layers. Sediment accumulates on top of the filtration mixture creating a cake layer. Collecting soil cores from the columns and a sieve analysis shows evidence of sediment accumulation on the top 1 em of filtration layer. Storm water particles are small and the increase in particles smaller than 75 urn after application of stormwater indicates that the sediment accumulated in the top layers. The sieve analysis as shown by the particle size distribution plots, Figures 4-12, 4-13, and 4-14, show the change in smaller particles in the top 0-1 em. The change is significant when viewing the percent of particles passing the 75um sieve as presented in Table 4-5. Table 4-5 Table of percent of particles passing 75 urn sieve for samples of the filtration layer at various depths (0-10 em) Control Compost, Paper, Compost, Paper, Sand Sand, Tires Soil Sample Depth Percent Passing 75 urn Sieve 0-1 em 3.2 3.0 3.0 4-6 em 0.7 1.0 1.6 8-10 em 0.8 0.8 0.9 68

PAGE 87

l f j 0.01 I i I I I I Control Sieve Analysis of Fine Aggregates I I I I I I Tl I I f.' .! ' I l I l I / i "l I /' '] I / I i / I / I I I _/ I Jl : i 0.1 Sieve Size (mm) 1---o.1 clean -+-4-s clean 11-10 clean .. m 0-1 dirty --4-s dirty -+-11-1o dirty 1 I I I ; I I I I i I I I i ' l l I ' : I i I I I I ' I I I l i I I ' I i I I I : I 10 Figure 4-12 Plot of Sieve Analysis of Samples at Various Depth (0-10 em) of Control (Peat and Sand Mixture). Results Show Small Particles in the Top Soil Layer Indicating Sediment Filtered out of the Stormwater Accumulates on the Top 1 em. 69

PAGE 88

Control Sieve Analysis ot Fine Aggregates I i -----------90. 0 --------------, --_ I --_ 5!1-o-:_ -'c_ ___ : -__ _.._ ___ -- 0 i i f I I j I jT f + : I I i -I f --:--; ___ --+ ---+.:_ ------------------------1.. !--20. i 0 .01 0 1 10 Sieve Size (mm) 1------0-1 clean 4 6 clean B--10 clean Q-1 dirty -11-4-6 dirty ---8--10 dirty I Figure 4-13 Plot of Sieve Analysis of Samples at Various Depth (0-10 em) of CPS. Results Show Small Particles in the Top Soil Layer Indicating Sediment Filtered out of the Storm water Accumulates on the Top 1 em. 70

PAGE 89

I I f s . -I I 1-1-fe. .. ------\--! f--1---1-. ... --.. l3b. t----1---. f1i"'' ..... -1-. b.< 0.01 Compost, Paper, Sand and Tires Sieve Analysis of Fine Aggregates I ; ..1 .. .. l.rv+--+-+-+ Ill... -.!' .. t-t--I .. . ... . -.. -------------ly ---... 'I ---------f---fl .. ------.. -t--\-.. -e -+------I i ---. --1 -. -0.1 Sieve Size (mm) 1---o-1 clean --+--4-6 clean 8 clean 0 dirty -11-4-6 dirty --+--8-10 dirty I 10 Figure 4-14 Plot of Sieve Analysis of Samples at Various Depth (0-10 em) of CPS. Results Show Small Particles in the Top Soil Layer Indicating Sediment Filtered out of the Storm Water Accumulates on the Top 1 em. Sediment accumulation on the top layer causes a reduction in infiltration rate. The change of infiltration rate was measured throughout the continuous application of stormwater. For mathematic convenience, the reduced infiltration rate,fs, is normalized by fc the Horton's constant infiltration rate defined in equation 4-3, 4-4 and 4-5. The accumulative sediment load, Ls. is expressed as weight of sediment per . kg/ 2 umt area m m. 71

PAGE 90

The decay of infiltration is plotted in Figure 4-15 for various sub-base mixtures. The relationship can be depicted by an exponential decay function between fife and Ls [kg/m2]. Equation 4-6 Peat and sand : fs = -4.449e-o.I3611, fc control Equation 4-7 C d d fs __ 2 927e-O.I369/, ompost, paper an san : -fc cps Equation 4-8 Compost, paper, sand and tires: fs = 2.8303e -0 08141 fc cpst Wherefs is the infiltration rate [crnlhr] after accumulative sediment load L5,[kg/m2 ] andfc is Horton's constant infiltration rate [crnlhr]. The subscripts, control, cps and cpst, indicate the media-mixture associated with each equation. Where/' th J c control lS e constant infiltration rate [ crnlhr] of the control (peat and sand) mixture. fc cps is the infiltration rate [crnlhr] of the compost, paper and sand mixture and thefccpst is the infiltration rate [crnlhr] of the compost, payer, sand and tires mixture. Equations 4-6, 4-7 and 4-8 have correlation coefficients r = 0.89, 0.93, 0.92 respectively. 72

PAGE 91

Clogging Rate 6 Control r2 = 0.89 0.93 CPST =0.92 0 0.00 5.00 10.00 15.00 Acummulative TSS load (kg/m2) Peat and Sand Compost, Paper, Sand x Compost, Paper, Sand, Tires Figure 4-15 Plot of Reduced Infiltration Rate, fs, Normalized by fc (Horton's Constant Infiltration Rate) Versus the Accumulative Sediment Load for Various Media Mixtures. Results Show Reduction of Infiltration Rate with Accumulative Sediment Load The minimal infiltration rate according to the local stormwater criteria for PLD is 2.5 cm/hr (1 in/hr). Therefore the system is considered clogged whenfs becomes less than 2.5 em/hr. 4.5 DESIGN EXAMPLE As indicated in Figure 4-15, infiltration through the filtering layers is closely related to the accumulative amount of sediment loaded onto the infiltrating bed. The design example intends to illustrate how to interpret the accumulative sediment load into the basin's operation. The annual sediment yield generated from the tributary area can be estimated by the annual mean event concentration of sediment and the annual runoff volume as: 73

PAGE 92

Equation 4-9 V = CPA,ributary Equation 4-10 L5 = C5V In which Vis the annual runoff volume in [L3], Cis the runoff coefficient, Pis the annual rainfall depth in [L], A tributary is the watershed tributary area [L2], Cs is the mean sediment concentration [MIL\ and Ls is the annual sediment load [M/L3]. In practice, the PLD is designed to intercept a portion of the tributary area. As a result, the annual loading to the PLD is estimated as: Equation 4-11 Atributary a=---=--APLn Equation 4-12 B5 = C5aCP In which a is the ratio of the tributary area intercepted by PLD, Apw is the surface area of the PLD [L2 ] and Bs =annual unit-area sediment load to the basin of the PLD [M/L2/year]. Figure 4-15 shows the decay of the infiltration rate with respect to the accumulative sediment load, Ls, which can be converted into the PLD's service years as: Equation 4-13 Bs Where N is the number of PLD's service years, Ls is the accumulative sediment load into laboratory tested PLD [MIL2 ] and Bs is the annual accumulative sediment load into the basin of interest [M/L2/year]. Equation 4-13 assists the engineer to convert Figure 4-15 into any PLD' s service years for an investigation of the life-cycle operation. For example, a PLD in Denver is designed based on the local requirements (USWDCM 2001) using the average annual rainfall in Denver, and average sediment 74

PAGE 93

concentration observed in Colorado (Doefer and Urbonas 1993). The example PLD has a surface detention capacity up to a water depth of 0.305 m (12 inches). It will capture and treat runoff from a parking lot. The ratio (a ) of the parking lot area as the contributing watershed to the PLD is 20 to 1. The event mean sediment concentration (Cs), TSS, in runoff from commercial areas in Colorado is recorded as 240 mg/L. Annual precipitation (P) in Denver area is .4 meters (15.4 in). Aided by equation 12 with Cs = 240 mgiL, a=20, C=0.9 for parking lot, and P=0.4 m, the annual unit-area sediment load to the example PLD is calculated as: Equation 4-14 Bs = (240mg I /)(20)(.9)(.4m) = 1.728kg I m2 The accumulative sediment load (L5 ) on the x-axis in Figure 4-15 can then be converted into years of service for the example PLD. Figure 4-16 presents the reduction in infiltration over time for the example PLD. The decay of infiltration rate fs down to 2.5 cm/hr (1 inlhr) varies for each sub-base mixture based on the Horton's constant infiltration rate fc For this example, the PLD is considered clogged at f /fc=2.51fc As shown in Figure 4-2 and equations 4-2, 4-3 and 4-4,/c is 9.9 cmlhr for the control (peat and sand), 14.5 crn/hr compost, paper and sand and 9.1 cm/hr compost, paper, sand and tires. Therefore the PLD is considered clogged in Figure 4-16 where fife is 2.519.9 for the control, 2.5114.5 for compost, paper and sand and 2.519.1 for compost, paper, sand and tires. 75

PAGE 94

5 ra 4 c -....,;; 0 == -3 == 0 0 ;;: ;;: 'C 2 G) 0 ra 0 ... :;::: ::::J 1 ra ra a: Cl) 0 ... ... ... ...... [it.t. ... ... ...... + "+.'+.. -* l' .. .. ... --, ., 0 5 Clogging Rate Control r = 0.89 CPS r2 = 0.93 CPST r=0.92 ....... ..; ----.. .. -. -.. ---10 Time (years) 15 .a. Peat and Sand Compost, Paper, Sand + Compost, Paper, Sand, Tires Figure 4-16 Plot of Reduced Infiltration Rate,.fs, Normalized by fc (Horton's Constant Infiltration Rate) Versus Time for Various Media Mixtures. Results Show Reduction of Infiltration Rate Over Time for Example PLD 4.6 CONCLUSION The infiltration tests indicated that improvements in water quality result in sediment accumulation on the PLD's bottom, and clogging through the sub-base filtering media over time. All three mixtures filtered TKN, phosphorous and copper from the system during the stormwater applications. Removal rates for TKN, TP, and Copper varied from 32% to 77%, 48% to 86%, and 48% to 95% respectively. The pathogen removal rates were between 88% and 99%. Sediment removal rates remained high throughout the two tests, 93% to 100%. Accumulation of sediment creates a cake layer on top of the filtration layer. Sieve analysis indicated the cake layer formed in all three treatments was on the top surface up to 1 em. Additionally, the cake layer consists of both solids from storm water and floating particles from the PLD's media mixture. 76

PAGE 95

Individual components of media perform differently and mixing may alter the characteristics of the system. Shredded tires float and are hydrophobic. When mixed with sand the layers become compacted and are held in place by the sand but tires from the top layer still float. Dry organic materials such as peat and compost are buoyant as well. Shredded paper floats initially and will sink as they become absorb water. These materials may affect the cake layer, the outlet and possibly the downstream waterways. A model has been developed to approximate the clogging of a PLD in the field based on sediment loading. The model can be used to estimate the life span and maintenance of the PLD. A design example indicates clogging occurs when the infiltration rates is reduced to the minimal 2.5 cmlhr (1 inlhr) which occurs in 11 to 16 years. Replacement of the top 10 em of material is suggested to renew the infiltration rate. 77

PAGE 96

5 BENCH SCALE TEST WITH VEGETATION 5.1 INTRODUCTION AND BACKGROUND The PLD is designed to treat stormwater through a vegetated basin and soil-mix filtration layer (USWDCM 2001). The previous chapter defined the possible media mixes for the filtration layer to achieve stormwater treatment goals. This chapter investigates the impact of vegetation on the performance of the system. The goals of this chapter are to: 1. Ensure the ability of the filtration mixture to support plant growth. 2. Evaluate the impact of vegetation on performance and 3. Evaluate the possibility of the PLD as a source of water quality impact. Plants have been shown to reduce contaminates in water and soil through degradation, uptake and other biochemical mechanisms (Langergraber 2005; Munch et al. 2005; Schoonover et al. 2005; Sirivedhin and Gray 2006; Vidon and Hill 2004). Many engineered systems such as wetlands and phytoremediation function successfully due to the intrinsic treatment capacity of plants (Elodie et al. 2009; Gottschall et al. 2007; Huett et al. 2005; Kohler et al. 2004; Syversen 2005) Additionally, the roots of plants create macropores in the soil, creating water pathways. Both the treatment capacity and the infiltration rate are affected by the addition of plants in the PLD. Permissible filtration soil-mixtures from previous investigations were tested with and without vegetation. The ability of each of the mixtures to support plant growth was verified and compared through quantitative plant counts and qualitative observations in the lab. The effect of the plant growth on the performance in terms of water quality impacts and infiltration rate was assessed. The possibility of an additional water quality impact of floating particles from the filtration media, seed mixture and decaying plant material was observed in the lab. 78

PAGE 97

5.2 BENCH SCALE TEST METHODOLOGY Building on previous chapters, which defined the permissible filtration layer mixtures and assessed the performance of the system without vegetation (bare soil), this chapter adds vegetation. Testing was conducted in the same 38 em (15 in) diameter infiltrometers as previous tests. The columns were packed to replicate the current recommendation of 45.7 em (8 in) of filtration soil mix and 20.3 em (8 in) of aggregate under drain (Guo et al 2009). Sediment laden stormwater was added to the top of the columns with bare soil creating a cake layer. The filtration layers were: 1. Control (15% peat and 85% sand) 2. CPS (7.5%compost, 7.5% shredded paper and 85% sand) 3. CPST (7.5% compost, 7.5% shredded paper, 8% tires and 77% sand) Then grass seeds were germinated on top of the cake layer created from buildup of sediment filtered out of the storm water. The roots of the grass grew into the cake layer and performance criteria were measured and compared to results without vegetation (bare soil conditions). The bench scale test consisted of the five steps in Figure 5-1 to compare un-vegetated (bare soil) and vegetated conditions. Previous steps 1 and 2 provided data for the un vegetated condition. Step 3 began with germinating grass seeds in the cake layer created from the sediment buildup in step 2. Grass was allowed to grow and the same experimental procedures, adding stormwater, measuring flow rates and water quality parameters were followed. The vegetation eventually died leading to additional results with dead grass (without vegetation), steps 3 and 4. Grass seeds were again planted on top of the cake layer with dead grass and additional information about the effect on infiltration rate was collected. 79

PAGE 98

Step 1-Initial Condition -Bare Soil-No Vegetation 72-hour elevated outflow and clean water to measure infiltration capacity lfc) and calculate of hydraulic conductivity Lower the outflow and add stormwater for unsaturated field conditions Top layers were excavated and sieve analysis was performed Step 2 -Duplicate Bare Soil Test The top soil-mix layers were replaced Begin with 72-hour lowered outflow and measure infiltration rate Then lower the outflow and add stormwater for unsaturated field conditions Step 3 -Effect of Vegetation -First Planting of Grass Germinate grass seeds on top of cake layer Measure restoration of infiltration rate with 72-hour lowered outflow Then lower the outflow and add stormwater for unsaturated field conditions Step 4 -Dead Vegetation Continue to add stormwater with sediment until the grass is choked and dies Measure flow rates with dead grass Step 5 -Replant Vegetation -Second Planting of Grass Replant grass seeds on top of cake layer created in step 4 Then lower the outflow and add stormwater for unsaturated field conditions Measure regeneration of flow rates Figure 5-l Steps for Bench Scale Testing Stormwater was applied to the columns before and after the vegetation was growing. The flow rate and water quality parameters were measured throughout the entire experiment to compare treatments. Between each step of the testing an initial condition for flow rate was measured by adding tap water at a consistent depth for 72 hours. The infiltration rate after 72 hours corresponds to the infiltration capacity in Horton's infiltration equation (Horton 1993). Horton's estimate for infiltration rate over time is based on the equation: Equation 5-1 J(t) = fc +(fo-fc)e(-k)t Wheref(t) is the infiltration rate at time t,f0 is the initial infiltration rate [L3/t] andfc is the infiltration rate at field capacity [L3/t]. The k is a constant with units 1/time. 80

PAGE 99

Synthetic stormwater was added during each step to measure filtering of contaminates and create cake layer on top of the filtration layer. The saturated flow rate was again monitored for 72 hours to conclude steps 1 and 2. Grass seeds were then planted on the cake layer created during step 2 and germinated. Again 72 hour clean water flow was conducted to compare the effect when the roots penetrated the cake layer. Sediment laden stormwater was added after the grass was growing in step 3. The stormwater killed the grass leading to step 4 and the cake layer thickened. A final test, step 5, was conducted by again adding grass seeds directly on top of the dead grass without disturbing the cake layer. Stormwater was again added to the top of the system to complete step 5. Mixing runoff water with additional sediment created the synthetic stormwater. Runoff water from an outfall to the South Platte River in the City of Denver was collected and transported to the lab. The water was spiked with additional sediment to accelerate the clogging process in order to relate surface loading to field conditions where the clogging process occurs over years. The sediment-laden stormwater was applied to the columns with the same procedure throughout steps 1 through 5. Sediment and water were mixed continuously while being pumped on top of the column. Water quality samples were collected from the inflow and outflow. Stormwater was applied until a cake layer was formed. The seed mixture was the same mixture recommended by the UDFCD and used in the batch test and recommended in Volume 3 Criteria Manual for the Denver Metropolitan area (USWDCM 2001) presented in Table 5-1. Seeds were weighed in relative amounts for the surface area of each column and presented in Table 5-2. A discussion with horticulturalist at Botanic Gardens, Mark Fusco indicated the use the grasses would adequately represent the germination and growth of the entire mixture. Native seeds also require a cold period before germination, and therefore were refrigerated for 30 days before sowing in the columns. The numbers of seeds were also doubled to ensure good coverage. 81

PAGE 100

.. .:;, Table 5-1 PLD vegetation seed mixture as prescribed in UDFCD criteria manual COMMON NAME SCIENTIFIC NAME VARIETY PLS Lbs Ounces per Acre per Acre Sand bluestem Andropogon hallii Garden 3.5 Sideoats grama Bouteloua curtipendula Butte 3 Prairie sandreed Calamovilfa longifolia Goshen 3 Indian ricegrass Oryzopsis hymenoides Paloma 3 Switchgrass Panicum virgatum Blackwell 4 Western Wheatgrass Pascopyrum smilhii Ariba 3 Uttle bluestem Schizachyrium scoparium Patura 3 Alkali sacaton Sporobolus ain.>ides 3 Sand dropseed Sporobolus cryptandrus 3 Pasture Sage Artemisia frigida 2 Blue aster Aster laevis 4 Blanket flower Gaillardia aristata 6 Prairie coneflower Ratibida columnifera 4 Purple prairiedover Dalea (Petalostemum) purpurea 4 Sub-Totals: 27.5 22 Totallbs per acre: 28.9 (Source:USWDCM 2001) Table S-2 Amount and type of grass seeds planted in each column container Amount size Weight Seeds Planted Seed Name (lbs/acre) seedsllb (sq ft) mg/column mg/column sideoats gramma 3 191,000 1.23 38.3 76.6 prarie sandreed 3 274,000 1.23 38.3 76.6 swtichgrass 4 270,000 1.23 51.0 102.1 western wheatgrass 3 110,000 1.23 38.3 76.6 little bluestem 3 260,000 1.23 38.3 76.6 sand dropseed 3 825,000 1.23 38.3 76.6 Grass seeds were spread on top of the cake layer, germinated and allowed to grow. As shown in Figure 5-2 grow lights were installed above the columns and a timer was set. The seeds were watered every other day and monitored for growth. The numbers of plants in a 4 em by 4 em square were counted in three separate locations in each of the columns. The numbers of plants were reported for day 20 and 25. No additional plants were germinating by day 20 and by day 25 the plants had grown tall enough to take the application of water. The effect of the vegetation on the infiltration rate was 82

PAGE 101

measured by saturating the columns for 72 hours and then adding stormwater in the same method as the previous bare soil steps. The system was monitored for physical and chemical performance plus biological parameters of plant counts and qualitative growth rates. Figure 5-2 Picture of Grow Lights above Columns for Growing Grass in the Lab A portion of the grass seeds did not germinate and were found to float. Un germinated seeds, dead vegetation as well as light particles in the PLD filtration media mix with trash carried in the storm water. Parts of the filtration media include peat, bark and tire particles float. During large storm events these particles may overflow causing downstream impacts or plugging the outlet. The amounts of particles in the overflow during various steps of the experiment were measured. Samples were collected by skimming the surface of the water and analyzing for TSS. Photographs were taken as confirmation of visual observations in the lab of the makeup of the floating particles. Figure 5-3 shows grass seeds floating in the water. 83

PAGE 102

Figure 5-3 Picture of Un-Germinated Seeds Floating in Water above Soil Surface 5.3 VEGETATION The ability of each mixture to support plant growth was evaluated with quantitative plant counts and qualitative observations. To take advantage of the benefits of vegetation in the PLD the filtration media must first support plant growth. Additionally the negative impact of high sediment loads and standing water in the PLD on plant growth was observed when the grass was choked out and died. The average number of plants in a 4 em square on day 20 and 25 is shown in Table 5-3. The results indicate no further germination of plants occurred after day 20. Table 5-3 Germination slows after 20 days as shown by the average number of plants 20 and 25 days after planting Average Number of Plants in 4 square em Filtration Mixture Planting 1 Planting 2 Day Day Day Day 20 25 20 25 Control 6 6 4 6 Compost, Paper, Sand 11 11 5 6 Compost, Paper, Sand, Tires 9 10 5 6 As shown in Figure 5-4 no statistical difference in the number of plants on day 25 is seen in either the first or second growth phases (step 3 and 5). The ability of each mixture to germinate and support plant growth is statistically similar. 84

PAGE 103

--I Number of Plants 25 Days After Planting 20 Ill 18 r::: 16 ca c:: 14 0 ... 12 Q) .c 10 E ::::J z 8 Q) 6 C) I!! 4 Q) >
PAGE 104

Figure 5-5 Picture of Healthy Grass Growing in the Column 25 Days after 1st Planting Figure 5-6 Picture of Soil Surface after the 1st Planting of Grass was Choked and Died After step 5 (the second germination of grass seeds) was performed the grass was left for 2 months without water and continued to grow in all columns. The growth of the plants in the control (peat and sand) mixture was noticeably impaired compared to the plants in the other two treatments. A photo of the plants after a long dry spell is presented in Figure 5-7, Figure 5-8 and Figure 5-9. 86

PAGE 105

Figure 5-7 Picture of Grass in the Control (Peat and Sand) Column after 2 Months without Water. The Picture Shows the Stunted Growth of Vegetation in Control (Peat and Sand) After a Dry Period Figure 5-8 Picture of Grass in the CPS Column after 2 Months without Water. Healthy Vegetation in CPS after a Dry Period Figure 5-9 Picture of Grass in the CPS Column after 2 Months without Water. Healthy Vegetation in CPS after a Dry Period 87

PAGE 106

Figure 5-10 presents the average plant count in the columns after the 2 months without water. The average height of the plants is presented in Figure 5-11. 20 18 Cll 16 a: 0 14 .. 12 .c 10 E :I 8 z Gl 6 g) I! 4 "" 2 0 Number of Plants During Testing 25 Days after 1st Planting 25 Days Alter 2nd 2nd Planting after 2 Planting months without water I Control liJ Compost, Paper, Sand Compost, Paper, Sand, Tires I Figure 5-10 Average Number of Plants in the Columns with 2 Watering Schemes: Watering Every Other Day and without Water for 2 Months Average Plant Height 50.0 .-----------------------, 45.0 -j---------------------------1 E .2. 40.0 +-----------1------------l i, 35.0 ;! 30.0 +---------1 'lii 25.0 20.0 :: 15.0 10.0 "" 5.0 0.0 +---_.l_ ___ Figure 5-11 Plant Height after 2 Months without Water Shows CPS Supports Plant Growth through a Dry Period 88

PAGE 107

Although no statistical difference could be detected in the average number of plants per column, the average height of the plants indicates that the compost, paper, sand mixture support growth through dry spells better than the control (peat and sand). 5.4 WATER QUALITY IMPACTS The water quality impacts investigated include both contaminant removal and possible loading due to floating particles in the overflow. The systems continued to show good contaminate removal during both bare soil steps 1 and 2 and vegetated steps 3 (first planting). Contaminate removal included nutrients, metals, pathogen and sediment load. The negative impact of possible loading due to particles introduced from the PLD into the overflow was evaluated based in the amount and make-up of floating particles. All treatments filtered nutrients and metals in bare soil conditions and with vegetation. For both mixtures metals removal rates remained above 92% with vegetation. The removal rate of TP and TKN in the compost, paper, sand tires mix was statistically greater with vegetation than without. However nitrate removal was variable and a consistent trend was not observed. The zinc removal rate initially increased from 86% to 95% with the growth of grass but reduced to 83% after an additional 2. 7 kg/m2 of sediment load was added. The average percent removal rates are presented in Table 5-3 and Figure 5-11. 89

PAGE 108

Table 5-4 Table of contaminant removal rate comparing un-vegetated and vegetative conditions. Results show vegetation increases the nutrient removal rate Accumulative Removal Rate Sediment Cin Cout /Cout Test N02 kg/m2 TKN +N03 tTotal P Cu Pb Zn 1 Bare Soil .33 44% 0% 83% 45% 70% 87% Control 2 Bare Soil 2.65 77% 769o 86% 95% 95% 98% 3 Vegetation 7.00 87% 85% 98% 98% 97% 99% 3 Vegetation 9.7 80% 69% 98% 92% 97% 98% 1 Bare Soil .33 32% -2% 48% 66% 70% 88% CPS 2 Bare Soil 2.65 71% 65% 80% 93% 95% 98% 3 Vegetation 7.00 77% 57% 85% 95% 97% 99% 3 Vegetation 9.7 81% 39% 91% 93% 97% 99% 1 Bare Soil .33 35% 27% 61% 81% 70% 86% CPST 2 Bare Soil 2.65 71% 82% 85% 93% 95% 88% 3 Vegetation 7.00 79% 76% 92% 97% 97% 95% 3 Vegetation 9.7 74% 82% 92% 92% 97% 83% Average Percent Removal of Contaminants 1.00 c:: G) 0.80 U-.... ca G) > 0.60 11. 0 G) E 0.40 ICnQ) !a: 0.20 G) >
PAGE 109

Both removal rates of contaminates as well as outflow concentrations are important to measure the filtering capacity of the PLD system. The three treatments filtered nutrients and metals from the system as presented in Table 5-5. The inflow and outflow concentrations can be compared to the EPA freshwater criteria for maximum allowable in-stream contaminate concentrations (EPA 2006). Both inflow and outflow concentrations of N03+N02 was below the standard of 10 mg/L of N03. Total Phosphorous outflow concentration in all three filtration mixtures and both bare soil and vegetated conditions consistently met the criteria. With a few exceptions the systems continuously filtered metals from the stormwater to below EPA freshwater standards. Only during the last test the average copper concentration in the outflow of the each mixture, (14.0, 13.2 and 14.0 ug/L) slightly above the criteria of 13 ug/L. The lead concentrations in the outflows were consistently non-detect (5 ug/L) well below the standard of 65ug/L. Although the highest inflow concentration of zinc was 2410 ug/L, only the last outflow from the compost, paper, sand and tire mix (412.7 ug/L) was above the standard of 120 ug/L of zinc. 91

PAGE 110

Table 5-5 Contaminant concentrations in inflow and outflow water through various media mixtures with and without vegetation compared to the EPA freshwater criteria Accumulative Nitrate Sediment Load TKN +Nitrite Total P Cu Pb Test (kg/m2 ) (mg/L) (mg/L) (mg!L) (u_&!!-) (ug!L) Zn (ug/L) 1 Bare Soil 0.33 1.70 3.56 0.63 21.1 16.7 169.0 2 Bare Soil 2.65 2.60 4.43 1.30 113.0 100.0 955.0 Inflow Tap Water 5.92 0.50 0.08 ND(.03) 9.7 ND(5) ND(20) 3 Vegetation 7.00 4.20 4.66 1.98 217.0 195.0 2380.0 3 Vegetation 9.70 4.30 2.34 1.68 179.0 170.0 2410.0 EPA Freshwater Criteria (EPA 2006) NR 10.00 0.06 13.0 65.0 120.0 1 Bare Soil 0.33 0.95 3.58 0.11 11.65 ND(5) 22.3 2 Bare Soil 2.65 0.60 1.08 0.19 6.10 ND(5) ND(20) Outflow Tap Water 5.92 0.37 0.20 ND(.03) ND (2) ND(5) ND(20) Control 3 Vegetation 7.00 0.04 ND(20) 0.53 0.72 4.87 ND(5) 3 Vegetation 9.70 0.87 0.73 0.04 14.00 ND(5) 40.0 1 Bare Soil 0.33 1.15 3.63 0.33 7.25 ND(5) ND(20) Outflow 2 Bare Soil 2.65 0.77 1.56 0.26 7.83 ND(5) ND(20) CPS Tap Water 5.92 0.45 0.23 0.09 ND(2) ND(5) ND(20) 3 Vegetation 7.00 0.97 2.02 0.29 10.80 ND(5) 21.4 3 Vegetation 9.70 0.83 1.43 0.15 13.23 ND(5) ND(20) 1 Bare Soil 0.33 1.11 2.59 0.25 3.95 ND(5) 24.0 2 Bare Soil 2.65 0.77 0.80 0.20 8.23 ND(5) 114.6 Outflow Tap Water 5.92 0.37 0.24 0.07 2.13 ND(5) 50.9 CPST 3 Vegetation 7.00 0.87 1.11 0.16 7.37 ND(5) 110.8 3 Vegetation 9.70 1.10 0.42 0.14 14.07 ND(5) 412.7 Note: Samples that failed the EPA criteria are marked in bold 5.4.1 NUTRIENT REMOVAL The difference in average inflow and outflow concentration of TKN did not vary significantly between bare soil test and the vegetated test for each of the three treatments and is presented in Figure 5-13. The total concentration ofTKN in the 92

PAGE 111

outflow from the compost, paper, sand and tires mix was slightly greater than the outflow of the control (peat and sand) during the first test with vegetation. Filtering of TKN 5.-------------------------------------------. 4.5 +-----------------------------------------------1 4+---------------------------+ 3.5 .-. Bare Soil o Inflow I .. 2 5 I Control z I 2 IICPST 1.5 0.5 0 Stormwater Stormwater Tap Water Stormwater Stormwater .33 2.65 5.92 7.00 9.70 Accumulative Sediment Load (kg/m2 ) Figure 5-13 Plot ofTKN Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Vegetation does not Effect TKN Removal The inflow concentration of N03+N02 was below the criteria of 10 mg/L of N03. After the application of 0.33 kg/m2 of accumulative sediment load the control (peat and sand) and compost, paper, sand mixture were not filtering any N03+N02 out of the water. During the last test with vegetation the outflow from the compost, paper, and sand mixture was greater (1.43 mg/L) than the control (peat and sand) of .73 mg/L. The outflow from the compost, paper, sand and tires mixture was less (0.42 mg/L) than either of the other two treatments. The outflow concentrations remained low after the addition of vegetation to the system as shown in Figure 5-14. 93

PAGE 112

Filtering of N02 Plus N03 4 3.5 .. 3 0 z 2.5 II) ..2 2 Q. 0 1.5 z 0.5 0 Vegetation Stormwater Stormwater Tap Water Stormwater Stormwater .33 2.65 5.92 7.00 9.70 Accumulative Sediment Load (kg/m2 ) Dlnflow control IDCPST Figure 5-14 Plot of N03+N02 Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Consistently Higher Concentrations in Outflow from the CPS than the Control Mix Total Phosphorous was filtered from the stormwater before and after vegetation as presented in Figure 5-15 TP removal initially increased in the bare soil test from between 48% to 83% to 80% to 86% with the addition of 2.3 kg/m2 of sediment and reduction in flow rate. Then after the growth of grass the removal rate increased and remained between to 91% to 98% as presented in Figure 5-16. Although high removal rates were achieved for all mixes the concentration of TP in the outflow was consistently lower for the control (peat and sand) than other two treatments. The average outflow concentration of the control (peat and sand) met the EPA standard of 0.55 mg/L during the growth of the vegetation. The outflow concentrations in all treatments did not vary statistically between the bare soil and the vegetated test. 94

PAGE 113

Filtering of Phosphorous 2.----------------------.=------------. 1.8 1.6 11olnflow I 1.4 j 1::'1 0.8 iliiCPST 1-0.6 ::J-----1 '-----------" o.4 H;:::t------+a----,----------+ 0.2 0 Stormwater Stormwater Tap Water Stormwater Stormwater .33 2.65 5.92 7.00 9.70 Accumulative Sediment Load (kglm2 ) Figure 5-15 Plot of Total Phosphorous Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Vegetation Increased the Removal Rate of TP from the Control Mix iii > 0 E Gl a: -c Gl u ... Gl a. Percent Removal of Total Phosphorous 120% 100% 80% 60% 40% 20% 0% 1.33 2.65 7.00 9.70 Accumulative Sediment Load (kg/m2) Control! 0CPS CPST Figure 5-16 Plot of Percent Removal of Total Phosphorous for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Increased Removal Capacity for TP with Vegetative Conditions 95

PAGE 114

5.4.2 METALS REMOVAL After the first addition of sediment which lowered the flow rate, metals removal increased and remained between 88% and 99%. Accumulative sediment loading and flow rates affect the filtering of metals as well as vegetation. Except for the last test, copper was consistently filtered to below the stormwater criteria through all three treatments ass shown in Figure 5-17. Cu 100 90 80 70 2. 60 .. !. 50 Cl. 0 (.) 40 ]i 30 0 1-20 10 0 Stormwater Stormwater Tap Water Stormwater Stormwater .33 2.65 5.92 7.00 9.70 Accumulative Sediment Load (kglm2 ) Figure 5-17 Plot of Total Copper Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Vegetation does not Effect Copper Removal The removal rate of zinc remained high throughout the testing of bare soil and vegetation. The inflow and outflow concentrations are presented in Figure 5-18. Breakthrough of zinc occurred into the compost, paper, sand, tires mix after the addition of 2.66 kg/m2 of accumulative sediment. The last test the outflow concentration of zinc was approximately three and a half times the EPA criteria of 120 mg/L. 96

PAGE 115

2380 Zn 2410 ug/L Zn "'" .. 1000 8' 900 v -.::. 8: BOO v :-... I are Soil I I 700 I I Dlnllow 1 2.. 600 -._. u T control c 500 v ,. N f.? l ICPS 400 ::::: 'iii 11 0 f.? : IIICPST 300 1-f.? t?. 200 it '"" -,:,-.+, "" ..L 1:?. 100 r-< 1J ... tJ 0 .-. 0 '<' Stormwater Stormwater Tap Water Stormwater Stormwater .33 2.65 5.92 7.00 9.70 Accumulative Sediment Load (kg/m2 ) Figure 5-18 Plot of Total Copper Concentration in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed Breakthrough of Zinc in CPST Mixture 5.4.3 PATHOGEN REMOVAL Pathogens were removed from the stormwater before and after the growth of vegetation. As shown in Table 5-6 an average of 88% to 99% cfu were removed under bare soil conditions and between 87% to 99.8% pathogens were removed after vegetation. The two highest concentrations of pathogens in the inflow were 60,000 cfu with bare soil and 50,000 cfu with vegetation and the removal rate remained above 94%. The removal rate when the stormwater inflow was 5,700 cfu the removal rate varied between 88% ad 92%. The outflow concentrations are presented in Figure 5-19. 97

PAGE 116

Table 5-6 Percent removal of total coliform forming units without and with vegetation. The results show high percent removal of pathogens from storm water Total Colony Forming Units (cfu) Inflow 5,700 26,000 60,000 20,000 50,000 Percent Removal with Percent Removal in Bare Soil Vegetation Control 92.0% 97.7% 99.0% 90.7% 99.8% CPS 88.0% 98.6% 94.7% 97.5% 99.5% CPST 90.1% 99.2% 98.4% 87.3% 99.2% Pathogen Filtering 60,000cfu 50,000 + .. : S" ,: 10,000 Ill p i .. 9,000 "2 ::I I I Cll 6,000 . c:: e 7,000 Dinflow .. I I 0 control IL 6,000 >-Soil I .5 0 5,000 1::3 cPST 1 u .5 4,000 r-Ill E 3,000 f.. g 2,000 0 u 1,000 d!l ll k h 0 Stormwater Stormwater Stormwater Stormwater Stormwater .33 .64 2.66 7.00 9.70 Accumulative Sediment Load (kg/m2 ) Figure 5-19 Plot of Total Coliform Forming Units in Water for Various Accumulated Sediment Loads (0.33 to 9.70 kg/m2) with and without Vegetation. The Results Showed All Mixes Were Continually Successful at Filtering Pathogens from Stormwater throughout the Experiment 98

PAGE 117

5.4.4 SUSPENDED PARTICLES Suspended solids are carried in stormwater and filtered through the PLD system through settling and filtration. As presented in Table 5-7, a total of 1.08 kg/m2 at an average of .15 kg/m2 per application were loaded to the columns during step 1 and a total of 12.69 kg/m2 at an average of .6 kg/m2 per application during step 2 through 5 with both bare soil and vegetation. Table 5-7 Accumulative sediment loading to the soil surface during the experiments with and without vegetation First Bare Soil Test (Step I) Second Test with Bare Soil and then Vegetation (Step 2-5) Soil Standard Accumulative Standard Accumulative Conditions Average deviation Load Average deviation Load TSS (kg/m2 ) TSS (kg/m2 ) 0.10 0.01 0.10 0.16 O.oJ 0.16 0.08 0.01 0.19 0.33 0.02 0.49 0.07 0.00 0.26 0.18 O.oJ 0.66 0.07 0.01 0.33 0.33 0.02 0.99 0.05 0.02 0.38 0.50 0.02 1.49 0.26 0.05 0.64 0.60 O.oJ 2.09 Bare Soil 0.44 0.02 1.08 0.55 0.02 2.65 0.48 0.03 3.13 0.56 0.01 3.69 0.54 0.02 4.23 0.86 0.08 5.10 0.83 0.05 5.92 0.53 0.05 6.45 0.55 0.05 7.00 Grass 1.03 0.01 8.03 0.82 0.14 8.85 0.85 0.22 9.70 0.76 0.06 10.46 Dead Grass 0.62 0.04 11.08 1.00 0.02 12.09 0.00 0.00 12.09 New Grass 0.60 0.03 12.69 99

PAGE 118

During previous testing light particles from the filtration mixture were observed floating on the stormwater and in the overflow outlet as in Figure 5-20. The floating particles may come from the light portions of the filtration soil-mixture such as organics and tires. Additionally, vegetation may add un-germinated grass seeds and dead vegetation to the floating material. Large suspended solids in stormwater, such as leaves, twigs and cigarette butts combine with the light materials in the filtration mix and may be carried downstream in the event of overflow or plug the outlet. Figure 5-21 shows trash which was combined with dead grass and overflowed the PLD. Figure 5-20 Picture of Floating Particles from the Filtration Mixture Plugging the Overflow in the Lab Figure 5-21 Picture of Light Particles which Overflowed a PLD Total suspended solids samples collected in the lab indicate that the rainwater disrupting clean bare soil can float as much as 2,000 mg/L of suspended solids. Figure 5-22 presents the concentration if floating particles, when tap water is added to the following conditions; 1) clean bare soil, 2) after the cake layer is formed and 3) after grass has been germinated in the cake layer. The cake layer forms a crust which 100

PAGE 119

holds the floating particles in the filtration mix in place. When the cake layer has been disrupted by grass roots the light particles float again. -----Floating Particles 3000-r------------------, 2000 ,-.-Co-nt-ro-,11 .. 1500 1::1 CPS 1000 CPST I 500 0 Clean Bare Soil Bare Soil with Cake Layer With Grass Figure 5-22 Plot of the Amount of Particles (TSS) from the Various Filtration Mixes Which Were Found to Float in Clean Water. Results Indicate as Much as 2,000 mg/1 TSS May Overflow the PLD Although the TSS concentration [mg/L] in the overflow from the three soil-mixes is statistically similar, the particles are different substances. The floating material is related to the nature of the filtration media and the vegetation growing in the PLD. The overflow from each of the mixtures was filtered and photographs of the filtration paper are presented in Figure 5-23, 5-24, 5-25 The particles include light portions of the filtration mix (eg. peat, compost and shredded tires) plus un-germinated seeds and small pieces of dead grass. Figure 5-23 shows the solids filtered from the overflow water in the clean bare soil. Peat, compost particles and tire particles can be seen on the filtration papers in Figure 5-24. Figure 5-25 shows the make up of the floating particles after grass seeds are planted and vegetation is growing. 101

PAGE 120

Figure 5-23 Floating Particles in Water Samples of Overflow from the Soil Columns Peat and Sand Compost Paper and Sand Compost Paper, Tires and Sand Figure 5-24 Particles on the Paper of Filtered Samples of Overflow from Clean Bare Soil. The Samples Show that Peat, Compost and Tires Particles Float and May Overflow the PLD 102

PAGE 121

Compost Paper Sand Compost Paper, Tires Sand Figure 5-25 Particles on the Paper of Filtered Samples from Overflow after Grass was Growing. The Samples Show that Peat, Compost and Tires Particles Float and May Overflow the PLD. The overflow risk and downstream water quality must be taken into account when designing the PLD and utilizing waste incorporated mixtures. For example the waste screening test and the water quality tests indicated zinc leaching from tires. 5.5 CLOGGING EFFECTS The experiment was conducted in steps 1 through 5 to test the effect of vegetation on the system. The bare soil tests were steps 1 and 2. The top layers of the filtration media from step 1 were replaced with clean soil and step 2 began by measuring the initial infiltration rate. Synthetic stormwater was then added until a cake layer was formed and reduced infiltration rate was measured. Step 3 began with the spreading seeds on the cake layer and allowing the seeds to germinate with the roots penetrating the cake layer. The regeneration of flow rate due to germination of plants on cake layer was measured. Stormwater was again added to the top until the plants were choked and died in step 4. A definite decrease in infiltration rate was measured after the death of the plants as the macro-pores were filled in. Step 5, again seeds were germinated without disrupting the cake layer and the regeneration in flow was measured in step 5. The 72 hour clean water test showed Horton's infiltration capacity of the system, Figure 5-25. Water was added to the system for 72 hours between each step of the 103

PAGE 122

test. The initial decrease in infiltration rate in the clogged bare soil and the subsequent regeneration of the infiltration rate due to germination of grass in the cake layer is presented in Figure 5-26. The increase in infiltration capacity lfc) indicates the regeneration of flow capacity due to the grass roots penetrating the cake layer and creating macro-pores. Effect of Vegetation on Infiltration Capacity (f c) 20.-------------------------------------. ::. 12 -t---1-1 .; 10 a: 8 6 u:: 4 2 0 +-'---'"-'.:..:..:'--Peat and Sand Compost, Paper, Sand Compost, Paper, Sand, Tires I D Clean Bare Soil rJ Clogged Bare Soil 1 After Germination of Grass Dead Grass Figure 5-26 Plot of Infiltration Capacity of Various Media Mixtures Before and After Vegetation is Growing in the Cake Layer. Regeneration of Infiltration Capacity from Growth of Vegetation Shows the Control and CPS Benefited more than CPST 104

PAGE 123

c--Ill u .c Ill 0 a. Ill 1110 Clc s 0 c -8 1! ... = IIIIC D. .5 ---------------------Effect of Vegetation on Infiltration Capacity (f c) 100%.------------------T--------------, 80%+---------------------60%+----------------------40%+----------------0% -20% ___________ I i Clogged Bare Soil After Germination of Grass Figure 5-27 Plot of Saturated Infiltration Capacity of Various Media Mixtures Before and After Vegetation is Growing in the Cake Layer. Results Show Increase in Saturated Infiltration Capacity after Grass was Germinated into the Cake Layer The unsaturated infiltration rate was measured as stormwater was applied to the top of each column. Figure 5-27 presents the ratio of the infiltration rate during each stormwater application,.fs, to the initial Horton's infiltration rate,Jc. The infiltration rate decays as sediment is built up on the bare soil and the flow rate is regenerated after the germination of grass seeds. A second reduction inflow rate is seen as the vegetation is choked and sediment builds up on bare soil. 105

PAGE 124

Regeneration of Flow Rate 6.00 .------------------------., Bare Soil ---------:----:-::-:----:::-1 -- 1 Grass Dead Grass 0.00 -t----,-------,--------,---,---,----,--------j 0.00 2.00 4.00 6.00 8.00 1 0.00 12.00 14.00 Acummulative TSS load (kg/m2) Peat and Sand Compost, Paper, Sand x Compost, Paper, Sand, Tires Figure 5-28 Plot of Reduced Infiltration Rate,fs, Normalized by fc (Horton's Constant Infiltration Rate) Versus Accumulative Sediment Load for Various Media Mixtures, with and without Vegetation. Results Show Regeneration of Flow Rate in Control and CPS Mixes after Grass is Germinated in the Cake Layer The minimal design infiltration rate according to the local criteria is 2.5 cm/hr (linlhr). The example PLD from Chapter 4 can be used to evaluate the increase in life span of the PLD with vegetation. The decay of infiltration rate Is down to 2.5 cm/hr (1 inlhr) varies for each sub-base mixture based on the Horton's constant infiltration ratefc In Figure 5-29, the PLD is considered clogged atfl!c=2.51fc The trend lines in Figure 5-29 have correlation coefficients of r2 = 0.89, 0.93, 0.82 for the Control, CPS and CPST mixes without vegetation. The second set of trend lines for the condition with vegetation have correlation coefficients of r2 = 0.95, 0.96, 0.84 for the Control, CPS and CPST mixes respectively. 106

PAGE 125

Clogging Rate 4 .,--------------------------------------------, 1 -----Without vegetation Control r = 0.89 CPS r2= 0.93 CPST r =0.92 With Vegetation Control r = 0.95 CPS r=0.96 CPST r =0.R4 0 5 Peat and Sand + Compost, Paper, Sand, Tires 10 Time (years) 15 Compost, Paper, Sand t:, Peat and Sand with Veg o Compost, Paper, Sand with Veg x compost, Paper, Sand, and Tires Figure 5-29 Plot of Reduced Infiltration Rate, fs, Normalized by fc (Horton's Constant Infiltration Rate) Versus Time for Various Media Mixtures, with and without Vegetation. Results Show Vegetation's Increases the Time to Clogging in the Control and the CPS Mixes The regeneration of infiltration rate after the first germination is 54% (Control), 76% (compost, paper, sand) and 40% (compost, paper, sand, tires) increase in flow rate. The second germination of seeds results in a 235% (Control), 96% (compost, paper, sand) and 8% (compost, paper, sand, tires) increase in flow rate from the infiltration rate after the grass was choked. The regeneration of flow rate in Figure 2-28 equates to a longer lifespan of the PLD. The vegetation had the greatest effect on the peat and sand mixture and the least effect on the compost, paper, sand and tires mixture. Clogging occurs in 15 years for the control with vegetation, 14 years for the CPS mix and 16 years for the CPST mix. 107

PAGE 126

5.6 CONCLUSIONS The porous landscape detention (PLD) is designed to treat stormwater through a vegetated basin and soil-mix filtration layer. Previous chapters defined possible media-mixes for the filtration layer to achieve stormwater treatment goals. This chapter evaluated those filtration mixes capacity to support plant growth the possible water quality impacts due to the filtration mixture in the PLD and the vegetation were investigated. The vegetation's effect on performance of the PLD was assessed. The three treatments 1. peat and sand, 2. compost, paper, sand and 3. compost, paper, sand and tires) supported germination and vegetation growth similarly. The waste incorporated mixes (compost, paper, sand and compost, paper, sand and tires) sustained the plant growth through the 2 month dry period. The growth of the vegetation was stunted in the control (peat and sand) mixture without water. The vegetation did not reduce the treatment capacity of the PLD and in some cases increased the benefit. The outflow concentration of nutrients from the control mixture was consistently less than the other two treatments. The control reflected the most nutrient removal capacity. The outflow concentration and percent removal of TKN through the control (peat and sand) was positively effected with vegetation. The nitrite plus nitrate concentration in the outflow from all three mixes consistently met the EPA in-stream criteria. The concentration of nitrite plus nitrate in the outflow from the control (peat and sand) was lower with vegetation. The vegetation had the greatest effect on TP in the control mixture. The TP concentration in the outflow from the control (peat and sand) mix with vegetation consistently met the EPA in-stream criteria and had improved removal rate. The outflow from the control (peat and sand) mix with bare soil condition and both the outflow from the other two mixes with bare soil and with vegetation were 3 to 6 times the EPA limit for TP. The total metals (Cu, Pb, Zn) removal was consistently high (88% to 99%) with bare soil and vegetated conditions. The concentrations in the outflow met the EPA limits within one standard deviation, except for the last sample result for zinc in the compost, paper, sand and tires mix. The outflow concentration of zinc slowly increased in the mix the tires and the last result was three and a half times the EPA limit. 108

PAGE 127

Pathogen removal is of interest because pathogens in runoff water contaminate rivers and expose the public health risks. The bench scale test consistently removed 88% to 99.8% of the total coliforms from the inflow water. No difference in removal rate or concentrations was detected in bare soil and vegetated conditions or between the three mtxes The possibility of the PLD acting as a source of water quality impact is a concern. Previous chapters evaluated the leaching of contaminates in the filtered storm water flowing out the bottom of the PLD. This chapter investigated the potential for contamination of the overflow water in the case of a large event. Samples from the overflow water indicated that light material from the filtration mix such as peat, compost and tire particles mix with dead vegetation and grass seeds. The shredded paper became saturated and sank before floating to the top of the water. Up to 2,000 mg/1 TSS were found in the overflow which may clog the outlet or enter the downstream waterways. Vegetation positively affects the infiltration capacity and regenerates the infiltration in the PLD. The regeneration of flow rates were observed with both 72 hour saturated conditions and with unsaturated field conditions. The effect of vegetation on the 72 hour saturated conditions was similar between all three mixes. Under unsaturated field conditions the infiltration rate was regenerated from 54% to 235% for the control (peat and sand), from 76% to 96% for the compost, paper, sand and 8% to 40% for the compost, paper, sand and tires mix. Clogging is defined as the time which the infiltration rate is reduced to less than the minimum 2.5 em/hr. The trendlines for the clogging rate under bares soil and vegetative conditions indicates that vegetation would save 2 to 4 years for the control (peat and sand) and compost, paper, sand mixes. The trendline for the compost, paper, sand, tires mixture indicates that grass would not affect the lifespan of the PLD .. 109

PAGE 128

6 CONCLUSIONS AND RECOMMENDATIONS This research evaluated the beneficial reuse of urban waste stream materials into sustainable substrate mixes for effective functioning of the PLD. The best waste materials for the sustainable sub-base system design were screened and the environmental life cycle analysis (LCA) for a waste-incorporated design was completed. The impact of the waste materials and vegetation on the performance of the system was evaluated in bench scale tests with bare soil and vegetative conditions. The filtration mixes capacity to support plant growth the possible water quality impacts due to the filtration mixture in the PLD and the vegetation were investigated. The vegetation's effect on performance of the PLD was assessed. A large diameter infiltrometer was designed and tested to simulate a flow system through a two-layered PLD. A model for the optimal dimensions for the sub-base in the two layered design indicates that the layers should be divided to consume the hydraulic head available. The ratio, f/K1 should be selected to avoid undesirable prolonged standing water in the PLD. The depth of the filtration mix must be selected based on drain time and infiltration rate. The thickness of the sand mix layer, H1 can be calculated by equation 6-l. Equation 6-1 H =(1-EL)D+ELY I f f The current recommendation of a sand-mix filtration layer on top of a larger gravel layer creates accelerated hydraulic gradient drawing water through the top layer. The goal of the lab research was to find suitable waste replacements for these sub base layers. Through screening and confirmation tests the currently recommended 15% peat and 85% sand mixture was compared to two waste-incorporated mixes 1) 7.5% compost, 7.5% paper, 85% sand and 2) 7.5% compost, 7.5% paper, 77% sand and 8% tires. By installing a WI mix (7.5% compost, 7.5% paper, 77% sand and 8% tires) instead of the BAU (peat and sand) in an example PLD of 98 cubic meters, a net GHG benefit of 7 MTC02E would be realized. Additionally, $2,800 would be saved by installing the WI mix. All of the cost savings and most (90%) of the GHG benefit is from offsetting the peat with compost and paper. 110

PAGE 129

Bench Scale lab tests were conducted to compare the control (peat and sand) to the WI mixes. Initial results without vegetation indicated that removal rates and outflow concentrations of nutrients, metal and pathogens was similar in all three treatments. Reduction in infiltration rate was compared to accumulative sediment load to assess the clogging process. A PLD is assumed to be clogged when the infiltration rate is reduced to the local design minimum of 2.5 em/hr. A design example based on lab results indicated that clogging occurs in 11 to 16 years. The WI mix with tires has almost 50% greater lifespan than the control (peat and sand). Additional testing with vegetation was conducted. Vegetation increases the nutrient removal capacity and regenerates a reduced infiltration rate. Total phosphorous was lowest in the control (peat and sand) mix. The impact of vegetation was to extend the life of the PLD up to 4 years for the control (peat and sand) and the compost, paper and sand mixture. Results indicated that grass would not affect the clogging rate over time of the compost, paper, sand and tires mixture. Grass seeds were germinated in each mixture and growth rates were measured. All three mixes supported germination and plant growth with regular water. The WI compost, paper, and sand mix supported plant growth best during a dry period. Light particles in the filtration mix were found to be buoyant, floating and settling with each water application. The settling of the particles creates density stratification in the top soil layers and the cake layer consists of both sediment and the lighter particles. Additionally, the potential for the buoyant filtration media to become a source of water quality impact was investigated. In a large storm event the overflow may carry a mix of light particles in from the filtration mix ( eg. peat, compost and tire particles), dead vegetation and grass seeds. Up to 2,000 mg/1 TSS were found in the overflow which may clog the outlet or enter the downstream waterways. Experiments showed that the vegetated PLD is successful at stormwater treatment the following recommendations are based on the laboratory results It is recommended that PLDs should be installed to remove sediment, nutrients, metals and pathogens. Vegetation is an added benefit as it increases nutrient removal and the lifespan of the PLD. A 2-layered design should be used for the PLD and the optimal dimension should be calculated based on the characteristics of the filtration mix. Replacing the currently used peat and sand mixture with the waste-incorporated compost, paper, and sand mix is recommended based on environmental benefit and performance criteria. In locations which are highly sensitive to phosphorous, the peat and sand mixture should 111

PAGE 130

still be installed due to higher removal rates. Although tires increase the lifespan of the PLD, more research and risk assessment is necessary before using tires due to the possibility of contaminating downstream waterways. Although the study was focused on materials and designs for Colorado, the results are generally applicable to any area. The 2-layered design model is suggested to protect groundwater from possible contamination and to increase the life of the PLD. The materials selected, compost and paper are readily available any place with organic waste streams. Additionally the recommendations for the permissible amounts of waste materials should be followed. One recommendation is to mix paper with compost and to avoid possible leaching of nutrients from compost. Second, although tires increase the lifespan of the PLD, they should not be used to treat stormwater. 112

PAGE 131

APPENDIX A ENVIRONMENTAL IMPACTS FROM PEAT GHG Emissions from Use of Peat Mining and Use 1 1 Transportation Mining and Use, including release of GHG from degradation, is 0.66 MTC02E Transoortina oeat from Canada to Colorado is 0.18 MTC02E 113

PAGE 132

REFERENCES ADAS UK Ltd, and Enviros Consulting Ltd. (2005). "Monitoring of peat and alternative products for growing media and soil improvers in the UK 2005." report prepared for UK Department for Environment Food & Rural Affairs,. American Membrane Technology Association. (2006). "Interactive membrane water treatment facilities map." Stuart, FL. Ames, K. C., Inkpen, E. L., Frans, L. M., and Bidlake, W. R. (2001). "Preliminary Assessment of Infiltration Rates and Effects on Water Quality of Selected Infiltration Media for Use in Highway Runoff Retention Basins in Washington State." W. S.D. o. Transportation, ed. Anastas, P. L., and Zimmerman, J. B. (2003). "Through the 12 Principles of Green Engineering." Environmental Science and Technology, 37(5), 95-101A. Appleton, A. F. (2002). "How New York City used an ecosystem strategy carried out through an urban-rural partnership to preserve the pristine quality of its drinking water and save billions of dollars." Forest Trends, New York City, Tokyo. CH2M Hill. (2001). "Pierce Country low impact development study." Bellevue, Wash. Cleary, J., Roulet, N. T., and Moore, T. R. (2005). "Greenhouse Gas Emissions from Canadian Peat Extraction, 1990-2000: A Life-cycle Analysis." AMBIO: A Journal of the Human Environment, 34(6), 456-461 Clesceri, L. S., Greensberg, A. E., and Eaton, A. D. (1998). Standard Methodology for the Examination of Water and Wastewater, American Public Health Association, Washinton, DC. Corwin, D. L. (2000). "Evaluation of a simple lysimeter-design modification to minimize sidewall flow." Journal of Contaminant Hydrology, 42(1), 35-49. Corwin, D. L., and LeMert, R. D. (1994). "Construction and evaluation of an inexpensive weighing lysimeter for studying contaminant transport." Journal of Contaminant Hydrology, 15(1-2), 107-123. 114

PAGE 133

Davis, A. P. (2007). "Field Performance of Bioretention: Water Quality." Environmental Engineering Science, 24(8), 1048. Davis, A. P., Shokouhian, M., Sharma, H., and Minami, C. (2001). "Laboratory study of biological retention for urban storm water management." Water Environment Research, 73( 1 ), 5-14. Davis, A. P., Shokouhian, M., Sharma, H., and Minami, C. (2006). "Water Quality Improvement through Bioretention Media: Nitrogen and Phosphorus Removal." Water Environment Research: A Research Publication of the Water Environment Federation, 78(3), 284. Davis, A. P., Shokouhian, M., Sharma, H., Minami, C., and Winogradoff, D. (2003). "Water Quality Improvement through Bioretention: Lead, Copper, and Zinc Removal." Water Environment Research: A Research Publication of the Water Environment Federation, 75(1), 73. Elkington, J. ( 1998). Cannibals with Forks: The Triple Bottom Line of 21st Century Business. Elodie, P., William, F. H., Daniel, E. L., Ryan, A. S., and Robert, A. B. (2009). "Field Study of the Ability of Two Grassed Bioretention Cells to Reduce Storm Water Runoff Pollution." ASCE, 505-510. EPA. (2002). "Updating Fuel Economy Estimates in MOBILE6.3." Radiation, Washington DC. EPA. (2004a). "EPA emissions factor brochure. Unit Conversions, Emissions Factors, and Other Reference Data." Washington DC EPA. (2004b ). "Polluted Runoff (Nonpoint Source Pollution) Factsheet." Office of. Water. Washington DC EPA. (2006). "National Recommended Water Quality Criteria" Office of Water (May 1, 2009). EPA. (2008). "WAste Reduction Model (WARM)." (May 1, 2009). 115

PAGE 134

EPA (2009). A Scoping-Level Field Monitoring Study of Synthetic Turf Fields and Playgrounds. National Exposure Research Laboratory Office of Research and Development, ed. Washington DC EPA (2010)."EPA Green Engineering Program." . March 8, 2010 Federal Highway Administration. (2006). "Water quality managment of highway runoff course number 1404 7." US Department of Transportation, ed. Gadepalle, V. P., Ouki, S. K., Van Herwijnen, R., eacute, and Hutchings, T. (2007). "Immobilization of Heavy Metals in Soil Using Natural and Waste Materials for Vegetation Establishment on Contaminated Sites." Taylor & Francis, 233-251. Gottschall, N., Boutin, C., Crolla, A., Kinsley, C., and Champagne, P. (2007). "The role of plants in the removal of nutrients at a constructed wetland treating agricultural (dairy) wastewater, Ontario, Canada." Ecological Engineering, 29(2), 154-163. Guo, J. C., Kocman, S., and Ramaswami, A. (2009). "Design of Two-Layered Porous Landscaping Detention Basin." Journal of Environmental Engineering. 135(12). 1268-1274 Guo, J. C. Y. (2007). "Stormwater Detention and Retention systems." Journal of Urban Water Management. Guo, J. C. Y., and Hughes, W. (2001). "Runoff Storage Volume for Infiltration Basin." ASCE J. of Irrigation and Drainage Engineering, 127(3). Guo, J. C. Y., and Urbonas, B. (1996a). "Maximized Detention Volume Determined by Runoff Capture Rate." ASCE J. of Water Resources Planning and Management, 122(1). Guo, J. C. Y., and Urbonas, B. (1996b). "Maximized Detention Volume Determined by Runoff Capture Ratio." ASCE J. of Water Resources Planning and Management, 122(1), 33-39. Guo, J. C. Y., and Urbonas, B. (2002). "Runoff Capture and Delivery Curves for Storm Water Quality Control Designs." ASCE J. of Water Resources Planning and Management, 128(3). 116

PAGE 135

Horton, R. E. (1933). "The role of infiltration in the hydrologic cycle." Transactions, American Geophysical Union, 14, Hsieh, C.-h. (2004). "Engineering bioretention for treatment of urban storm water runoff," Doctor of Philosophy Dissertation, University of Maryland, College Park, Maryland. Hsieh, C.-h., and Davis, A. P. "Multiple-Event Study of Bioretention for Treatment of Urban Storm water Runoff." Diffuse Pollution Conference Dublin. Hsieh, C.-h., and Davis, A. P. (2005a). "Evaluation and Optimization of Bioretention Media for Treatment of Urban Storm Water Runoff." Journal of Environmental Engineering, 131 ( 11 ), 1521. Hsieh, C. H., and Davis, A. P. (2005b). "Evaluation and optimization ofbioretention media for treatment of urban storm water runoff." Journal of Environmental Engineering, 131 (11 ), 1521-1531. Hsieh, C. H., and Davis, A. P. (2005c). "Multiple-event study of bioretention for treatment of urban storm water runoff." Water Science and Technology, 51(34), 177-181. Hsieh, C. H., Davis, A. P., and Needelman, B. A. (2007). "Bioretention column studies of phosphorus removal from urban storm water runoff." Water Environment Research, 79(2), 177-184. Huett, D. 0., Morris, S. G., Smith, G., and Hunt, N. (2005). "Nitrogen and phosphorus removal from plant nursery runoff in vegetated and unvegetated subsurface flow wetlands." Water Research, 39(14), 3259-3272. Hunt, W. F. (2003). "Pollutant Removal Evaluation and hydraulic characterization for bioretention stormwater treatment devices," Pennsylvania State University, University Park, PA Hunt, W. F., Jarrett, A. R., Smith, J. T., and Sharkey, L. J. (2006). "Evaluating Bioretention Hydrology and Nutrient Removal at ThreeField Sites in North Carolina." Journal of Irrigation and Drainage Engineering, 132(6), 600-608. Hunt, W. F., and White, N. M. (2001). "Designing rain gardens/ bioretention areas." North Carolina State University, Raleigh, NC. 117

PAGE 136

Kim, H., Seagren, E. A., and Davis, A. P. (2003). "Engineered Bioretention for Removal of Nitrate from Storm water Runoff." Water Environment Research, 75(4), 335-367. Kohler, E. A., Poole, V. L., Reicher, Z. J., and Turco, R. F. (2004). "Nutrient, metal, and pesticide removal during storm and nonstorm events by a constructed wetland on an urban golf course." Ecological Engineering, 23(4-5), 285-298. Langergraber, G. (2005). "The role of plant uptake on the removal of organic matter and nutrients in subsurface flow constructed wetlands: a simulation study." Water Science and Technology, 51(9), 213-223. Mays, D. C., and Hunt, J. R. (2005). "Hydrodynamic aspects of particle clogging in porous media." Environmental Science and Technology, 32(2), 577-584. McCambridge, W., Camougis, G., and Recycle Technology, L. (2004). "Recycling And Beneficial Uses of Asphalt, Brick and Concrete (ABC) Rubble." NEW ENGLAND'S ENVIRONMENT, 10(6). Munch, C., Kuschk, P., and Roske, I. (2005). "Root stimulated nitrogen removal: only a local effect or important for water treatment?" Water Science and Technology, 51 (9), 185-192. NCDENR. (2007). "NCDENR Stormwater Best Management Practices Manual." D. o. W. Quality, ed. Nelson, A. C. (2004). "Toward a New Metropolis: The Opportunity to Rebuild America." The Brookings Institution Metropolitan Policy Program, Washington, DC. Pit, R., Lantrip, J., Harrison, R., Henry, C. L., Xue, D., and O'Connor, T. P. (1999). "Infiltration Through Disturbed Urban Soils and Compost-Amended Soil Effects on Runoff Quality and Quantity." 0. o. R. a. D. USEPA, ed., USEPA, Washington DC. Pitt, R., Lantrip, J., Harrison, R., and O'Connor, T. P. (1999). "Infiltration Through Disturbed Urban Soils and Compost-Amended Soil Effects on Runoff Quality and Quantity." USEPA, Washington, D.C. Prince George's County Department of Environmental Resources (PGDER). ( 1993). "Design Manual for Use of Bioretention in Storm water Management." 118

PAGE 137

Division of Environmental Management and Watershed Protection Branch, eds., Landover, MD. Ramaswami. (2005). "Engineering Sustainable Urban Infrastructure." Sustainability Science and Engineering, M.A. Abraham, ed., Elsevier, 419-441. Sample, D. 1., and Heaney, J.P. (2006). "Integrated Management of Irrigation and Urban Storm-Water Infiltration." Journal of Water Resources Planning & Management, 132(5), 362. Schoonover, J., Williard, K., Zaczek, J., Mangun, J., and Carver, A. (2005). "Nutrient Attenuation in Agricultural Surface Runoff by Riparian Buffer Zones in Southern Illinois, USA." Agroforestry Systems, 64(2), 169-180. Sharkey, L. J. (2006). "The Performance of Bioretention Areas in North Carolina: A Study of Water Quality, Water Quantity and Soil Media," North Carolina State University, Raleigh. Sirivedhin, T., and Gray, K. A. (2006). "Factors affecting denitrification rates in experimental wetlands: Field and laboratory studies." Ecological Engineering, 26(2), 167-181. Siriwardene, N. R., Deletic, A., and Fletcher, T. D. (2007). "Clogging of stormwater gravel infiltration systems and filters: Insights from a laboratory study." Water Research, 41(7), 1433-1440. Smith, J. (2005). "Attenutation of Nitrate in the Sub-surface Environment." U. E. Agency, ed., Rio House. Sun, X. (2004). "Dynamic Study of Heavy Metal Fates in Bioretention Systems," University of Maryland, College Park, Maryland. Sun, X. L., and Davis, A. P. (2007). "Heavy metal fates in laboratory bioretention systems." Chemosphere, 66(9), 1601-1609. Syversen, N. (2005). "Effect and design of buffer zones in the Nordic climate: The influence of width, amount of surface runoff, seasonal variation and vegetation type on retention efficiency for nutrient and particle runoff." Ecological Engineering, 24(5), 483-490. 119

PAGE 138

Toronto and Region Conservation Authority. (2007). "Performance Evaluation of Permeable Pavement and a Bioretention Swale ",Seneca College, King County, Ontario. Tucker, M. F. (2007). "Great Beer at a California Breweing Company, Breweries grow on Zero Waste Initiatives,." In Business, 29(2), 18. UDFCD. (1999). "Urban Storm Drainage Criteria Manual's Volume 3-Best Management Practices ", Denver, CO 80211. UN. (2007). "State of the World Population 2007." UNFPA. UNEP. (2002). "World Summit on Sustainable Development." Johannesburg. UNEP. (2007). "World Population Prospects: The 2006 Revision and World Urbanization Prospects: The 2005 Revision." Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat, ed. Urban Drainage and Flood Control District (UDFCD). (1999). "Urban Storm Drainage Criteria Manual's Volume 3 Best Management Practices ", Denver, CO 80211. US Census Bureau. (2000). "United States Census 2000." USDA. (2007). "Colorado Fact Sheet." U. D.o. Agriculture, ed. USEPA. (1999). "Storm Water Technology Fact Sheet Bioretention." Office of Water, ed. USWDCM (2001). Urban storm drainage criteria manual (USWDCM), volume 3, Best management practices, published by Urban Drainage and Flood Control District, Denver, Co. Vidon, P., and Hill, A. R. (2004). "Denitrification and patterns of electron donors and acceptors in eight riparian zones with contrasting hydrogeology." Biogeochemistry, 71(2), 259. WCED. (1987). "Our Common Future." Oxford Univ. Press, New York:. 120

PAGE 139

Winogradoff, D. A. (2001). "The Bioretention Manual, 2001 Update." M. Programs & Planning Division Department of Environmental Resources Prince George's County, ed. Yang, H., Rahardjo, H., Wibawa, B., and Leong, E.-C. (2004). "A Soil Column Apparatus for Laboratory Infiltration Study." Geotechnical Testing Journal, 27(4). Yost, B. (2008). "A-1 Organics." April19 2008, ed., Eaton, CO. 121