-
extracted text
-
AN EXAMINATION OF THE TOXICOLOGICAL EFFECTS
OF THE LEACHATE FROM BIORENTENTION SYSTEM
SUBSTRATES ON ZEBRAFISH, DANIO RERIO
by
Maria G. Redig
A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2015
©2015 by Maria G. Redig. All rights reserved.
This Thesis for the Master of Environmental Studies Degree
by
Maria G. Redig
has been approved for
The Evergreen State College
by
________________________
Erin Martin, Ph. D.
Member of the Faculty
________________________
Date
ABSTRACT
An Examination of the Toxicological Effects of the Leachate from
Bioretention System Substrates on Zebrafish, Danio rerio
Maria G. Redig
Stormwater pollution is a major concern in urbanized areas with an increased amount of
impervious surfaces. These surfaces accumulate contaminants from vehicles and other
sources. During rain events, these chemicals are washed into drains, ditches, ponds,
wetlands and streams nearby. Bioretention structures are the most commonly used type of
stormwater conveyance feature. They are composed of mixtures of sand, bark, and
compost. Compost used in bioretention systems binds with pollutants and removes them
from the stormwater. Recent studies have also shown that in certain situations composts
have the potential to leach heavy metals, adding to the stormwater pollution issue. In this
study, gravel and soil were tested to determine if they were adding to the contaminant load
when used in bioretention systems. The toxic response of the leachate was measured using
Zebrafish as a freshwater model organism. By testing leachate of clean water flushed
through bioretention systems, it was found that extremely high levels of heavy metals were
leached from the substrates; however, as more water was flushed through, these metal
concentrations drastically decreased, indicating that conditioning the bioretention systems
is an appropriate means to reducing heavy metal load on aquatic systems. When
contaminated stormwater runoff was flushed through the bioretention soils and the gravel,
the metal levels decreased. Zinc was reduced by 75%, copper by 44%, nickel by 11%, lead
by 56% and cadmium by 33%. Metal concentrations alone are not an adequate indication
of ecological health. Dissolved organic carbon, pH and other cations present in the water
all affect the toxicity of metals. The leachate from the clean water passing through the
bioretention structures was found to be not toxic to Zebrafish embryos even though the
metal concentrations were elevated. Stormwater runoff displayed a toxic effect to the
Zebrafish by causing reduced eye size and an enlarged periventral area, indicating blood
pooling near the heart. Treating the stormwater through a bioretention system decreased
this toxic response. This shows that even though the substrates used in this experiment
leached elevated levels of heavy metals, it was not enough to cause toxicity. Ultimately,
the soils were successful in the removal of metals and subsequently the toxicity of the
stormwater, however this was after conditioning of the composts, which may prove to be
a necessary component to the use of composts in bioretention systems.
Table of Contents
1.
Introduction……………………………………………………...01
2.
Literature Review………………………………………………………..04
2.1
Stormwater Concern………………………………………….. 04
2.2
Stormwater Contaminants……………………………………. 06
2.2.1
2.2.2
2.3
Effects to Aquatic Organisms…………………………………12
2.3.1
2.3.2
2.4
Metal Toxicity……………………………………………… 12
PAH Toxicity………………………………………………. 16
Stormwater Management……………………………………... 18
2.4.1
2.4.2
Low Impact Development…………………………………..19
Bioretention Systems………………………………………. 20
2.5
Metals in Bioretention Substrates…………………………….. 22
2.6
Effectiveness of Bioretention………………………………… 24
2.6.1
2.6.2
3.
Heavy Metals………………………………………………. 06
Polycyclic Aromatic Hydrocarbons………………………... 08
Previous Studies……………………………………………. 24
Current Research…………………………………………… 26
Thesis research…………………………………………………..29
3.1
Methods………………………………………………………. 30
3.1.1
3.1.2
3.1.3
3.1.4
3.1.5
3.2
Bioretention Column Set-up……………………………….. 30
Biorentention Treatment and Leachate Collection………… 35
Highway Stormwater Runoff………………………………. 37
Chemical Analysis…………………………………………. 38
Zebrafish Toxicity Testing…………………………………. 39
Analysis………………………………………………………. 43
3.2.1
3.2.2
Image and Video Analysis…………………………………. 43
Statistical Analysis…………………………………………. 44
iv
3.3
Results………………………………………………………....45
3.3.1
3.3.2
3.3.3
3.3.4
3.3.5
4.
Discussion………………………………………………………. 66
4.1
Leaching of Metals…………………………………………… 66
4.1.1 Uncontaminated Water…………………………………….. 66
4.1.2
5.
Conventional Chemistry Results……………………………45
Metal from Uncontaminated Water Treatment…………….. 51
Metals from Treating Stormwater Runoff…………………..55
Toxicological Results- Uncontaminated Water……………. 57
Toxicological Results- Stormwater Runoff………………... 60
Stormwater Runoff………………………………………….70
4.2
Toxicological Effects…………………………………………. 72
4.3
Other Research……………………………………………….. 76
4.4
Ecological Implications………………………………………. 78
Conclusion……………………………………………………… 80
References……………………………………………………………... 83
Appendices
Appendix A: Metal Analysis Values………………………………… 88
Appendix B: Substrate Concentration by Metal Type……………….. 90
v
List of Figures
Figure 1. Estimated PAH Total Release into Puget Sound…………………………………09
Figure 2. Stormwater Runoff Toxicity Test: Detox and Cardiac Injury Genes……………. 27
Figure 3. Construction Design of Bioretention System Columns…………………………. 34
Figure 4. Stages of Zebrafish Embryo Development……………………………………….41
Figure 5. Zebrafish Image Analysis Measurements………………………………………...44
Figure 6. Conventional Chemistry Results: pH……………………………………………. 46
Figure 7. Conventional Chemistry Results: Alkalinity…………………………………….. 47
Figure 8. Conventional Chemistry Results: Calcium……………………………………….48
Figure 9. Conventional Chemistry Results: Magnesium…………………………………... 49
Figure 10. Conventional Chemistry Results: Dissolved Organic Carbon…………………. 50
Figure 11. Conventional Chemistry Results: Total Organic Carbon………………………. 51
Figure 12. Metals: BSM Column Treatment of Clean Water……………………………… 52
Figure 13. Metals: BSM plus Gravel Treatment of Clean Water…………………………...53
Figure 14. Metals: Gravel Treatment of Clean Water………………………………………54
Figure 15. Metals: Bioretention Treatment of Stormwater Runoff…………………………56
Figure 16. Clean Water Treatment Toxicity Test: Zebrafish Lengh………………………..58
Figure 17. Clean Water Treatment Toxicity Test: Zebrafish Eye Area…………………….58
Figure 18. Clean Water Treatment Toxicity Test: Zebrafish Periventral………………….. 59
Figure 19. Clean Water Treatment Toxicity Test: Zebrafish Pericardial Area……………..59
Figure 20. Clean Water Treatment Toxicity Test: Zebrafish Heart Rate………………….. 60
Figure 21. Stormwater Treatment Toxicity Test: Zebrafish Length……………………….. 62
Figure 22. Stormwater Treatment Toxicity Test: Zebrafish Aye Area……………………. 62
Figure 23. Stormwater Treatment Toxicity Test: Zebrafish Periventral Area……………...63
Figure 24. Stormwater Treatment Toxicity Test: Zebrafish Pericardial Area……………... 63
Figure 25. Stormwater Treatment Toxicity Test: Zebrafish Heart Rate…………………… 64
Figure 26. Stormwater Treatment Toxicity Test: Zebrafish Blood Pooling……………….. 65
vi
List of Tables
Table 1. Total Metal Release into the Puget Sound Basin………………………………….07
Table 2. Pollutant Reductions from LID Approaches………………………………………25
Table 3. Metal Concentration Limits for Compost………………………………………… 32
Table 4. Mineral Aggregate and Gravel Gradation Size Distribution………………………33
Table 5. Clean Water Treatment Toxicity Test: ANOVA Results………………………… 57
Table 6. Stormwater Treatment Toxicity Test: ANOVA Results…………………………..61
Table 7. Gene Expression Results…………………………………………………………..77
vii
Acknowledgements
There are a few key people that made this research possible. I would first like to thank
Erin Martin, my thesis reader. She was not only a collaborator and reviewer of this work,
but she really went over and beyond to be supportive and understanding of the hardships
life can throw at you. She helped me up until the last minute of turning this thesis in.
I would also like to thank Jenifer McIntyre. Jenifer helped design this research and
allowed me to work out of her lab spaces at the Washington State University Green
Stormwater Infrastructure Facility in Puyallup, Washington as well as the Northwest
Fisheries Science Center at the National Oceanic and Atmospheric Administration in
Seattle. She allowed me to collaborate on other research that coincided with this study
and helped me present my research at professional conferences. I can say with certainty
that this research would not have been possible without her.
And finally, I would like to thank my loving husband, Jonathan Redig. He was the real
reason I went back to school and followed my passions. He knew what I could
accomplish before I ever did. I am so grateful that I was finally able to believe in my
potential as he did. The accomplishment of the completion of this thesis is shared
between him and all my family and friends that supported me along the way.
viii
1.
INTRODUCTION
Stormwater is responsible for a majority of water pollution in urban waterways,
even more so than pollution from sewage treatment plants and industrial wastewaters
(Joerger, 2008). Pollutants that land on roadways eventually get washed away by rain,
becoming stormwater, and have the potential to end up in rivers, lakes, and marine
environments. As of 2008, there were 43 marine species in the Puget Sound which were
at risk, endangered or even threatened with extinction (Joerger, 2008). These species
includes orcas, groundfish, abalone, salmon and marine birds. One of the leading reasons
of this threat to viable populations of species is due to stormwater contamination
(Joerger, 2008). The severity of the ecosystem effects of the large toxic mixture of
pollution are still being studied, but it is safe to say that the effects are detrimental and
stormwater needs to be studied further. Heavy metals and polycyclic aromatic
hydrocarbons (PAHs) are of particular concern to the Puget Sound region due to the
effects they have on the fish populations (Department of Ecology, 2011).
Stormwater conveyance features, such as bioretention structures, are meant to be
a means of contaminant removal before the stormwater enters larger bodies of water,
such as the ocean, however, recent data has shown that compost and soils used in
bioretention structures, may leach metals (Cambier et al., 2014)(Kaschl, Römheld, &
Chen, 2002)(Page, Harbottle, Cleall, & Hutchings, 2014). This has not been studied
extensively, but if bioretention systems do leach metals, they would have the potential to
add to the metal contamination in the stormwater. Though, this relationship is not clear,
as soil and compost has also been shown to bind with and remove metals by adsorption
1
(Good, O’Sullivan, Wicke, & Cochrane, 2012)(Geronimo, Maniquiz-Redillas, Tobio, &
Kim, 2014)(McIntyre et al., 2014).
Soils also bind with polycyclic aromatic hydrocarbons (PAHs) (McIntyre et al.,
2014). This property is one reason which makes bioretention systems successful in
treating stormwater. McIntyre et al. (2014) found that stormwater treated through
bioretention structures resulted in a 95% reduction in total PAH concentration. The
concern is that bioretention systems may not only be successfully removing PAHs and
other organic contaminants from the stormwater, but they also may be adding heavy
metals.
In order to better analyze the impacts of bioretention systems, the scholarly
literature was reviewed, focusing on stormwater contaminants, toxicity to marine
organisms and the use of bioretention systems. The literature showed a general lack of
research in the performance of soils and the ability of bioretention systems to act as either
stormwater pollution treatment or pollutant generators by leaching metals. This thesis
research addressed the concern of metal leaching in bioretention systems by analyzing the
leachate when clean water was flushed through them. Metal concentrations were
measured initially, as well as over time, as more water was flushed through. This tested a
potential to condition the soils and reduce leachate metal concentrations. The analysis
was compared to the metal removal efficiency of the bioretention systems when
contaminated stormwater runoff was flushed through.
The next part of this research assessed the toxicity of the leachate from the
bioretention systems. Zebrafish were used as a model freshwater aquatic organism.
2
Toxicity endpoints such as hatch rate, growth and mortality were analyzed to test if the
metals in the leachate had a toxic effect. To assess the toxicity potential for other
contaminants leaching through the stormwater, such as PAHs, heart development was
also measured, including pericardial area, periventral area and heartbeat. While metals
can also cause heart defects, those toxic endpoints would have better signified a response
to organic pollutants in the stormwater itself (Incardona, Collier, & Scholz, 2004).
By comparing the metal concentrations in the leachate to the toxicity to aquatic
organisms, this thesis was designed to determine if bioretention systems are generating
contamination and if that pollution is toxic enough to potentially cause an ecological
impact. Furthermore, by assessing which toxic endpoints were observed (i.e. growth
versus heart defects) if could be determined if the toxicity to the zebrafish was likely
caused by metals leaching from the soils, or pollutants from the stormwater itself making
it through the bioretention structures.
3
2.
LITERATURE REVIEW
Stormwater pollution in a rising concern in urban waterways. It is incredibly
difficult to regulate, being that it is a broad category, encompassing many different
pollutants from a variety of sources. There are numerous studies on the toxic impact of
the individual pollutants, but not many on the mixtures found in an actual urban
environment. State and local governments are attempting to combat stormwater pollution
with green stormwater infrastructure (GSI) techniques. Some of these techniques are new
and lacking in research on the effectiveness of them.
2.1
STORMWATER CONCERN
There are two categories of stormwater pollution: point sources and non-point
sources. The Clean Water Act defines point source pollution as “any discernible,
confined and discrete conveyance, including but not limited to any pipe, ditch, channel,
tunnel, conduit, well, discrete fissure, container, rolling stock, concentrated animal
feeding operation, or vessel or other floating craft, from which pollutants are or may be
discharged. This term does not include agricultural storm water discharges and return
flows from irrigated agriculture” (EPA, 2015). Point source pollution is regulated within
the individual states by National Pollutant Discharge Elimination System (NPDES)
permits (EPA, 2015). Non-point source pollution is much more difficult to control and
regulate. Unlike pollution from a point source, non-point source pollution can come from
many sources. Non-point source pollution encompasses all the contaminants that end up
on the ground, waterways or in the atmosphere that cannot be tracked to a single source
4
(EPA, 2015). It is, basically, all the stormwater pollution that is not point source
pollution. With rainfall flowing over the ground, the contaminants get carried away and
deposited into larger water bodies. These pollution sources can include fertilizers,
pesticides, oils, fuels, sediments off of construction sites or agricultural areas, nutrients
and bacteria or even pollution from atmospheric deposition (EPA, 2015). The initial
discharge of these contaminants may be far below the levels to be concerned with, but
they accumulate and create toxic mixtures that are difficult to control or treat. It is nonpoint source pollution that is the main threat to waterways and aquatic organisms
(Department of Ecology, 2011).
Another issue with stormwater is the amount of water that urban areas have to
control. With the rapid increase in development, more of the land is urbanized and
covered with impervious surfaces. Major storm events can cause too much water for city
stormwater drains to handle. A large amount of rain can overflow the stormwater
conveyance features, allowing for large, fast flushes of the contaminants to enter larger
water bodies in a short time period, rather than to recharge the groundwater. When
overflowed, the stormwater goes untreated, dumping a mixture of toxic contamination
into the waterways. This results in a very high concentration of contaminants in a short
period of time. A stormwater conveyance feature is anything that moves stormwater in a
system. Many are designed to allow for slow infiltration into the groundwater.
Stormwater conveyance features that allow for infiltration are a means of treatment of the
stormwater as well.
5
2.2
STORMWATER CONTAMINANTS
Stormwater contaminants include any contaminant that collects on the land and is
flushed into a stormwater conveyance features. Commonly this includes metals from
vehicles and rooftops, polycyclic aromatic hydrocarbons (PAHs) from fuel drips,
refueling vehicles and fossil fuel combustion, and pesticides and other chemicals from
farming applications (Norton, Serdar, Colton, Jack, & Lester, 2011). Heavy metals and
PAHs are the largest concern in stormwater pollution due to the high concentrations and
the fact that they are difficult to control (Norton et al., 2011).
2.2.1
HEAVY METALS
Metals can enter the aquatic environment in a variety of ways. Metals get washed
into the stormwater drainage systems off of fields, roads, and parking lots. Heavy metals
that are the largest threat to aquatic species include copper, zinc, cadmium, mercury, lead
and arsenic (Norton, Serdar, Colton, Jack, & Lester, 2011). According to Dengler and
Brasino (2007), runoff, from cars in particular, is a huge source of metals pollution.
Copper, cobalt, cadmium, barium, aluminum, lead, nickel and zinc are all used in brake
pads and tires on vehicles (Scholz et al., 2011). Furthermore, copper is used in building
materials (roofs), and is common in agriculture and individual home pesticides (Norton et
al., 2011). The amount of copper from pesticide applications and the release of copper
from roofing materials are not as easy to quantify. The source of these metals vary in
each particular area. In this thesis, Puget Sound will be a focus, due to the amount of
research that has been conducted in this watershed. Table 1, below, shows the amount of
6
metals and their main sources in the Puget Sound, Washington and demonstrates the
issues with stormwater by showing that surface run-off delivered by stormwater is a
major source of these toxic metals (Norton, Serdar, Colton, Jack, & Lester, 2011).
Table 1. Department of Ecology total metal release into the Puget Sound Basin catchment
(metric tons/year) by the major sources and the total load that ends up in the Puget Sound
water by the major pathways (Norton et al., 2011).
Much of the metals that are deposited into a catchment area end up in lakes and
streams. Table 1 shows some differences in the total release to Puget Sound Basin and the
total load to the Puget Sound. With the exception of arsenic, all the metals have a higher
release than a load to the Sound. This indicates that some metals are left to accumulate in
7
streams, lakes, or near shore areas. Stormwater pollution in freshwater systems is a major
issue. Copper and zinc are particularly high in this region. One common main sources of
copper and zinc is off of vehicles and ends up in the Puget Sound from surface runoff
(Norton et al., 2011).
2.2.2
POLYCYCLIC AROMATIC HYDROCARBONS
Polycyclic aromatic hydrocarbons (PAHs) are another concern with stormwater
pollution. They are simply a combination of two or more fused aromatic rings and have
hundreds of different forms. PAHs are found in different mixtures due to a variety of
sources. Some are naturally occurring as there is a low level of background concentration
but the largest source is from human activity, mostly from the use of fossil fuels (Rand,
1995) (Newman & Unger, 2003) (Istenič, Arias, Matamoros, Vollertsen, & Brix, 2011).
The PAHs can be from petroleum spills and leaks, creosote oil, burning of organic
material such as brush fires, wastewater from refineries, as well as municipal and
industrial effluents (Rand, 1995). PAHs are in fossil fuels in varying concentrations.
Fossil fuels can collect on the roadways from cars, oil, fuel spills, pesticides or industrial
practices. Stormwater run-off flushes the petroleum products off the roadways and into
the stormwater systems (Norton et al., 2011). The amount released into Puget Sound
watershed from each source is depicted in Figure 1 below. The majority of PAHs are
from the incomplete combustion of organic material in woodstoves and fireplaces. Since
the PAHs are not very volatile they sorb to particulates. This results in atmospheric
8
deposition on the land and roadways. The next highest sources are from the use of
creosote followed by gasoline.
Figure 1. Estimated PAH total release into the Puget Sound watershed. Values shown are
in thousands of kg/year (Norton et al., 2011).
9
PAHs are lipophilic, meaning that they are found more in oil than in water. In a
stormwater conveyance feature, this means that they will sorb to soils, sediments or oil,
rather than staying in the water. Once in Puget Sound, they will be found bound to
sediment. This also means that once ingested in an organism they will have an affinity for
the fats and can accumulate up the food chain (Norton et al., 2011).
There are high molecular weight PAHs and low molecular weight PAHs. Low
molecular weight PAHs tend to be in elevated concentrations in fossil fuels, whereas high
molecular weight PAHs are from the incomplete combustion of those fossil fuels or other
organics, like wood (Norton et al., 2011). (Newman & Unger, 2003). Combustion in an
engine is not an efficient process; it leads to a high concentration of incompletely
combusted fuel (Newman & Unger, 2003). High molecular weight PAHs are more of a
concern. They are larger and are more lipophilic as well as less volatile (Istenič, Arias,
Matamoros, Vollertsen, & Brix, 2011). This makes them less water soluble and form a
stronger affinity for soils and sediments. The Clean Water Act listed the following 16
PAHs as being priority pollutants:
10
Low Molecular Weight PAHs (LPAHs)
- Acenaphthene
- Acenaphthylene
- Anthracene
- Fluorene
- Naphthalene
- Phenanthrene
High Molecular Weight PAHs (HPAHs)
- Benzo(a)anthracene*
- Benzo(a)pyrene*
- Benzo(b)fluoranthene*
- Benzo(k)fluoranthene*
- Benzo(g,h,i)perylene
- Chrysene*
- Dibenzo(a,h)anthracene*
- Fluoranthene
- Indeno(1,2,3c,d)pyrene*
- Pyrene
* Probable human carcinogens (cPAHs) by EPA (Norton et al., 2011)
Levels of PAHs are difficult to measure because there are so many types. In the
Puget Sound area, based on limited data, it is estimated that the freshwater concentrations
range from 0.1 – 1.0 µg/L, with the marine waters being slightly higher (Norton et al.,
2011). The sediment, however, is where the PAHs accumulate. It is estimated that the
freshwater and marine sediments in the Puget Sound area are 100- 1,000 µg/kg with an
approximation of about ten-fold higher in urban bays (Norton et al., 2011). These levels
would be comparable to other urban environments.
PAHs can degrade naturally and can even be broken down and metabolized by
organisms so the environmental concentrations varies widely and depend on many factors
such as sunlight, medium accumulated in (soil, water pavement, etc), or even presence
and abundance of microbes (Rand, 1995). There is a natural process of photodegradation
over time, but this strong affinity for soils makes them difficult to breakdown naturally,
11
as they would on the surface of water. This makes the concentration in sediment much
higher than in surface soils (Norton et al., 2011). They also biodegrade over time when
not attached to the soil, but because of the huge variety of PAH compounds and the
different half-lives, this process is highly variable and experimental data ranges anywhere
from a 2 day half-life to a 1.9 year half-life (“Technical Factsheet on: POLYCYCLIC
AROMATIC HYDROCARBONS (PAHs),” n.d.). Microorganisms can assist in the
degrading of the PAH compounds and actually use the PAHs as a carbon and energy
source forming non-toxic products, such as cell biomass, carbon dioxide and water (Atlas
& Cerniglia, 1995) (Lundstedt et al., 2007). This process is known as bioremediation.
2.3
EFFECTS TO AQUATIC ORGANISMS
Metals and PAHs have both been found to be toxic to aquatic organisms. This is a
concern in the freshwater streams and lakes where the stormwater initially gets
discharged. Levels of toxicity very between metal type, PAH type and mixture
concentrations. In this thesis, metal concentrations will be analyzed against the toxic
response measured, but PAH toxicity endpoints will also be considered.
2.3.1
METAL TOXICITY
Most of the metals that are of concern can be found naturally occurring within the
aquatic environment. They become harmful due to the increased concentration from
human activity (Newman, & Unger, 2003). There are a number of ways that metals are
12
toxic to fish, including adversely affecting the immune system and liver, the olfactory
system and the ability to regulate ion uptake, transport and kidney function (Rand, 1995).
Metals are usually found in ionic form or as simple compounds readily available for
uptake and accumulation (Rand, 1995). Copper and zinc are necessary for cellular
function (Rand, 1995). These essential elements are only beneficial for ionoregulation at
low levels. At higher levels they become toxic; this effect is known as hormesis
(Newman & Unger, 2003). The threshold value for the essential metals is different
among species and continues to be studied for fish.
The metallothionein protein, in the fish liver, binds to the metal ions for
detoxification purposes (Newman & Unger, 2003). This process reduces the overabundance of essential and non-essential metals. Metallothionein can lead to a metal
resistance by the liver by producing more of the protein and making the threshold for
toxicity higher. But the production of more metallothionein could also mean less
production of other detoxifying proteins, causing the fish to be susceptible to injury by
other contaminants (Newman & Unger, 2003).
Besides metallothionein, which is a detoxifying protein, there are also
environmental stress proteins that are produced in response to heavy metal exposure.
There are a wide array of proteins that are rapidly produced as a defense against an
environmental stress, such as that caused by toxic metals (Rand, 1995). They weaken the
fish and decrease the function of the immune system making disease, cancer, and gene
mutations more common (Rand, 1995). This is known as immunosuppression and is a
reduction in the fish defense mechanism making them more susceptible to pathogens
(Rand, 1995).
13
One way in which immunosuppression may occur is through exposing fish to
metal ions that disrupt and alter ionoregulation in fish (Scholz et al., 2011). The metal
ions bind to the gill surface leading to damage of the gills themselves as well as the
blocking of essential sodium and potassium ion uptake (Landis & Yu, 2004). This effect
is observed with many metals because of the common positive charge of a metal ion in
water. The sodium/potassium ATPase channel would normally be used for sodium and
potassium uptake and regulation, but because of the common positive charge of the metal
ions, the transport channel can be blocked by the accumulation of metals. This simple
metal accumulation on the gill surface can directly impact the health of the fish and cause
the development of tumors, as well as impair other biological functions which require the
ions that are being blocked (Rand, 1995).
Salmon in particular are sensitive to accumulations on the gill surface due to the
fact that they transition from saltwater to freshwater fish in order to spawn in freshwater
streams. This adjusts how the gill surface functions in terms of osmoregularity and the
regulation of sodium and potassium uptake (Scholz et al., 2011). Metal accumulated on
the gill surface in a saltwater fish may be more of a functional problem once it transitions
into a freshwater fish. This also affects the general ability to acclimate to the freshwater
system (Scholz et al., 2011). The gill surface is not the only site that metal acts on. Heavy
metals also alter the olfactory system of the fish, which affects the sense of smell.
The effects of copper on the olfactory system of juvenile Coho salmon have been
studied in lab experiments. In a study done by McIntyre et al. (2012), Coho salmon were
exposed to copper and then allowed to be in the same tank as predator fish. They
demonstrated that the “alarm response was absent in prey fish…exposed Coho were
14
unresponsive to their chemosensory environment, unprepared to evade nearby predators,
and significantly less likely to survive an attack sequence” (McIntyre, Baldwin,
Beauchamp & Scholz, 2012). The olfactory system is very important in salmon because
they use the sensory neuron output to get information about the surrounding environment,
in this case the predators nearby. In juvenile salmon, the normal response to prey in the
environment is to become motionless. In freshwater at low copper concentrations of 2 to
20 µg/L, copper ions can block the sensory receptor neurons and inhibit the response to
prey in the environment. In concentrations above 20 µg/L, the copper can cause cell
death of the olfactory receptor neurons (McIntyre et al., 2012).
The copper disrupts and distorts the neuron output and the fish becomes confused
and cannot detect predators nearby (McIntyre et al., 2012). The prey is more
disadvantaged than the predator because trout (the predator) are visual hunters so the
olfactory disruptor does not affect the ability to hunt. The olfactory system is not only
used for avoidance of predators, it is essential for the recognition of family and for the
synchronization for spawning salmon to find mates ( McIntyre et al., 2012). The fish also
use olfactory chemical cues for their migration pattern as forms of memory to return to
the stream they are from to spawn, as well as determine their habitat quality (Baldwin et
al., 2003). The inhibitions of the olfactory system has the potential to be detrimental to
salmon species. Because of the common charges of metal ions and the similarities in the
fish olfactory system, a similar effect could occur with other fish dependent on sense of
smell, as well as with other metals. The effect on the olfactory system has also been
studied in Chinook Salmon, Rainbow Trout, Brown Trout, Fathead Minnow, Colorado
Pikeminnow and Tilapia (Sandahl et al., 2007).
15
Juvenile salmon in rivers out of the Puget Sound catchment were showing
neurotoxic effects at levels as low as 2 µg/L copper (McIntyre et al., 2012), and
exhibiting a lack of predator avoidance behavior at levels as low as 0.7 µg/L (Norton et
al., 2011). Spawning salmon and juvenile salmon have been shown to avoid point sources
of copper contamination, but in a situation such as stormwater pollution, where there is
not a point source, the fish are unable to avoid the contamination (Sandahl et al., 2007).
Norton et al. with the Department of Ecology reported near shore (freshwater) and off
shore (marine) concentrations of copper in the Puget Sound are at an average of 5 µg/l
and 2 µg/l respectively (2001). Freshwater streams in California were found to have
dissolved copper concentrations ranging from 3.4 – 64.5 µg/l. These levels are
representative of freshwater in an urban watershed area (Sandahl et al., 2007). This shows
how metal concentrations in urban watersheds are high enough to cause a toxic effect.
2.3.2
PAH TOXICITY
The other main contaminant of concern is polycyclic aromatic hydrocarbons
(PAHs). High levels of PAHs are a concern for human and ecological health (Sun, Liu,
Jin, & Gao, 2013). PAHs range in potency and toxicity, and are most harmful by causing
DNA damage, some being carcinogens, mutagens, or even teratogens (Rand, 1995). This
means there is an increase in cancer to exposed aquatic organisms, they cause mutations
to occur within the cells of an organism at all stages of life, and they can cause
deformities to the forming embryos through the mother’s exposure (Rand, 1995).
16
Most PAHs are toxic by the process of biotransformation. They are readily
absorbed across lipid membranes and accumulate in the fats, gills, skin and digestive tract
of aquatic organisms (Rand, 1995). The body’s reaction to this lipophilic nature is an
attempt to metabolize the compounds by detoxification. Certain detox genes are produced
in response to PAH ingestion and can be measured in different areas of an organism to
indicate PAH exposure. Enzymes, such as cytochrome P450 monooxygenase, are
produced in response to the elevated levels of PAHs. PAHs in the body induce the gene
cytochrome cyp1A for production of this enzyme (Rand, 1995). In summary, the enzyme
is produced and catalyzes a series of reactions which break down the PAHs and make the
components water-soluble. This can make a chemical more or less toxic. In the case of
PAHs, it is a process which detoxifies the contaminant from the body, but the byproducts
can be harmful. Benzo (a) pyrene, for example, is not toxic on its own. The hydrolysis of
benzo (a) pyrene creates a byproduct known as benzo (a) pyrene diol epoxide. This
byproduct is extremely carcinogenic and can form covalent bonds with DNA, resulting in
DNA point mutations, ultimately causing cancer (Rand, 1995). In fish, the process of
biotransformation is most commonly observed in the liver, creating liver tumors.
Other PAHs in the body promote the production of free radicals and interfere with
the normal function of coping with oxidative stress (Newman & Unger, 2003). This
process can occur near DNA and cause DNA damage leading to cancer and other
genotoxic effects such as mutations (Newman & Unger, 2003). DNA damage and
mutations can be an indication of PAH exposure. If the rate at which a contaminant can
be detoxified from the organism is lower than the rate of accumulation, than PAHs will
17
accumulate in the lipids of the organisms. This creates the potential of bioaccumulation
and biomagnification as the PAHs move up the food chain (Rand, 1995).
PAHs are also considered immunotoxic (toxic to the immune system) as well as
cardiotoxic (disrupts normal heart function). They have been shown to result in
inflammatory gill damage (Rand, 1995). When the immune system is required and the
fish has inflamed gills, the antibody production is drastically decreased. This makes the
fish vulnerable to other contaminants and diseases. Other studies on fish embryos show a
decrease in size and development rate, specifically in the head and eye sizes of the fish
(Incardona et al., 2004). PAHs are cardiotoxic by causing deformities in the heart of fish
embryos. Incardona et al. (2004) measured an increase in the amount of fish embryos
with arrhythmia (uneven heart beats), pericardial edema (fluid accumulation in the heart
area), circulatory stasis (lack of blood flow) and unlooped hearts (lack of distinct heart
chambers). This shows that the development, as well as the function, of the heart is
effected by PAH exposed in the embryonic stages (Incardona et al., 2004).
2.4
STORMWATER MANAGEMENT
Nationwide, understanding of stormwater pollution is of growing importance
(Ormond, Mundy, Mary Weber, & Friedman, 2010). There is a push for innovative
solutions to the non-point pollution problem. Updated development planning is
implemented to control the flow of the water from over-whelming the city stormwater
infrastructure. Another goal is to decrease anthropogenic impacts by containing the
contaminants coming off the highways, parking lots, roofing and other developed areas.
18
The development approaches are known as Best Management Practices (BMP), Low
Impact Development (LID), or Green Stormwater Infrastructure (GSI). They aim for
development of cost effective and sustainable stormwater solutions (Ormond et al.,
2010).
2.4.1
LOW IMPACT DEVELOPMENT
Low impact development (LID) is an approach to the stormwater control problem
which attempts to use structures and nature functionally for stormwater control and
pollution reduction. The idea is to improve landscapes which would already be used or to
increase the perviousness of surfaces (EPA, 2015a). LID ultimately attempts to transform
stormwater from being viewed as a waste source to more of a resource. The goals of LID
are, according to the EPA (2015a), to maintain groundwater quality and recharge, to
reduce stormwater pollutant loads, to protect streams and channels, to prevent overbank
flooding, and to safely control extreme floods.
One of the major issues with increased urban areas and impervious surfaces is that
all the water that naturally falls on the ground concentrates on the roadways and ends up
in storm drains flowing back to the rivers and ocean. As a result of this the groundwater
is left with little recharge. Another issue is that when larger storm events happen, the
stormwater pipes can get overloaded and cause a flooding and overflow of the sewage
system. This is extremely detrimental because in many cases it can lead to untreated
water being diverted directly into the water bodies, creating a larger toxic control
problem. By using LID principles home owners, as well as municipalities, can harness
19
stormwater to recharge the groundwater as well as reduce the contaminant load to larger
waterways. “Applied on a broad scale, LID can maintain or restore a watershed's
hydrologic and ecological functions” (US EPA, 2015). A few of the LID approaches are
bioretention ponds, rain gardens, vegetated rooftops, rain barrels, grassed channels and
permeable pavements (EPA, 2015).
2.4.2
BIORETENTION SYSTEMS
Bioretention systems are landscape designs that are used to control stormwater
runoff. Ideally, they are shallow depressions with shrubs, trees and grasses planted in
them or areas with no plants, covered with gravel or bark (Dietz & Clausen, 2005). They
are placed adjacent to roads and parking lots and are used to divert stormwater and road
runoff. They can also be used on residential areas to capture rainfall from individual
homes and rooftop runoff. According to Hinman (2009), with the WSU extension
campus, “Bioretention is one of the most common applied and adaptable integrated
management practices in the low impact development approach”.
The two main types of bioretention systems are rain gardens and bioretention
ponds. The premise behind a rain garden is to allow for slow ground filtration of
stormwater so that the water from the rooftops and streets will flow in and permeate the
garden to be re-entered into the groundwater supply. This helps to recharge the
groundwater, securing and controlling the freshwater supply. They also act as a pollution
control technique for the treatment of stormwater. The water can filter through the soil
column allowing the contaminants in the stormwater to be retained by adsorption or taken
20
up by plants in the gardens (Dietz & Clausen, 2006). In this way, the rain garden can act
as a natural way to provide pollutant treatment (Dietz & Clausen, 2006).
Bioretention ponds are very similar to rain gardens with one exception. Rain
gardens are not meant to have any standing water, while bioretention ponds are. A rain
garden is a sort or bioretention area in that is designed to retain the contaminants and
allow for water to pass through, whereas, a bioretention pond is designed for larger
quantities of overflow water from a road system. These ponds are a standing pool of
stormwater (EPA, 2015).
The ponds are designed to capture the stormwater and allow for the slow settling
of the contaminants to the bottom of the pond. The ponds usually have an overflow
designed to divert untreated water into a stream or channel or into another stormwater
feature. One issue with bioretention ponds is that they tend to have an extremely high
concentration of stormwater contaminants, specifically copper and zinc (WiumAndersen, Nielsen, Hvitved-Jakobsen, & Vollertsen, 2011). This does indicate that the
pollutants are being stopped from entering larger streams and water bodies, but it also
poses a concern because these ponds create mini contaminated ecosystems for many
different species. Retention ponds are usually constructed in highly urbanized areas
where natural habitat is difficult to access. These ponds create a place for birds, insects
and amphibians to survive and breed.
Both rain gardens and bioretention ponds help lessen the load on other stormwater
conveyance features to avoid the burden of heavy rain and storm events. Collectively
they will be referred to as bioretention systems for the remainder of this thesis. While
21
bioretention systems are very valuable in the absorption of stormwater contaminants and
lessening the impact of flow in storm events, there is concern of the effectiveness in
treatment of contaminants in stormwater runoff.
2.5
METALS IN BIORETENTION SUBSTRATES
Some studies show that bioretention systems aid in the binding capacity and
removal of metals while other studies show that the composts in these systems can leach
metals (McIntyre et al., 2014)(Geronimo et al., 2014)(Page et al., 2014)(Good et al.,
2012). Composts are known to have high concentrations of metals. Much of this metal is
naturally occurring, but land use can also add to metals in the composts (Cambier et al.,
2014). Using compost from agricultural lands in particular can have elevated levels of
metals due to the use of different fertilizers and pesticides (Cambier et al., 2014). Metals
in the soils can do three things: accumulate in the soil and create a larger environmental
issue, leach from the soil and contaminate the groundwater, or be taken up by plants and
potentially enter the food web (Kaschl et al., 2002). The purpose of a bioretention system
is to allow for the accumulation of the metals in the soils and ultimate retention of the
metals. If the metals are leaching out at a greater rate than they are being retained, then
they are adding to stormwater pollution.
In a study of different stages of compost production, it was found that compost
can leach out metals regardless of what stage of decomposition they are in (Page et al.,
2014). Metal concentrations in dry compost were initially measured. Deionized water
was allowed to pass through the compost and the resulting heavy metal extractability was
22
measured. The metals that leached out were highest for nickel and zinc, with 13.4 % and
29% of the metals leaching out respectively. This study did, however, find that the water
soluble forms of the metals were in very low concentrations. Most of the metal leached
out of the compost were in complexes and not in free ionic form (Page et al., 2014). In
this study pH was also measured. The leachate water became more basic from flushing
through the compost, from 5.64 to 6.5. This indicates a retention of H+ in the compost
(Page et al., 2014).
Another study found elevated levels of copper, zinc, cadmium and lead in the
compost leachate at 37, 259, 0.21 and 4.5 mg/l respectively (Zhao, Lian, & Duo, 2011).
In a separate study of compost over time, the compost was allowed to age for 2-10 years.
All samples were similar in the leachate of metals. This leachate was found to contain
elevated levels of metals, however the only metal leaching out of the compost was zinc
(Cambier et al., 2014). Municipal solid waste composts are measured with higher levels
of heavy metals than background soil concentrations and there is concern of
contaminating groundwater (Zhao et al., 2011). Multiple experiments show pH as the
leading factor in metal leaching. Metal leaching is effected by pH by processes outlined
in the cation exchange capacity and the nature of soils to prefer certain cation over others.
In a more acidic environment, hydrogen ions could displace other metal ions, resulting in
metal leaching (Cambier et al., 2014). One study found that copper, zinc, nickel and
chromium were all leached out of compost, but at levels below drinking water standards.
There was a correlation with both pH and dissolved organic matter and the amount of
copper leached (Kaschl et al., 2002). All of the studies mentioned show leaching of zinc
and other metals out the composts. The concentrations vary widely. All of these studies
23
were done with clean water being flushed through compost. There is still a question of
what happens when the compost is merely a component of the soil as well as what
happens when contaminated stormwater is flushed through the soils.
2.6
EFFECTIVENESS OF BIORETENTION
Bioretention systems have been shown to be effective at controlling the flow of
water off of impervious surfaces as well as treating the stormwater. This ultimately
reduces the potential to overflow stormwater conveyance features, which could cause
large flushes of highly concentrated stormwater into larger bodies of water. Contradictory
to studies finding the leaching of metals, many studies find that composts, used in
bioretention systems, are effective at retaining metals by absorption.
2.6.1
PREVIOUS STUDIES
The Northwest Cascade Project from 2004 through 2006 was aimed at using a
combination of LID approaches in an average neighborhood and found that the pollutant
load from stormwater was significantly decreased with LID technology (Table 2) (EPA,
2012) . Total copper was reduced by 83%, total zinc by 76%, total lead by 90% and
motor oil, which contains a majority of the PAHs was reduced by 92% (EPA, 2012).
Another similar study found that when synthetic stormwater (water spiked to
replicate stormwater contaminants) was leached through compost, the compost retained
93% of the copper, 88% of the zinc and 97% of the lead. Based on this data the relative
24
sorption affinity of the composts were lead > copper > zinc (Seelsaen, McLaughlan,
Moore, & Stuetz, 2007). At the beginning of the study, the pH of the compost was
adjusted with hydrochloric acid, HCl, to be around 5. This was done to load the soils with
hydrogen ions and test the efficiency of the soils. This compost was also found to leach
very high concentrations of dissolved organic carbon (DOC) (Seelsaen et al., 2007).
Table 2. Pollutant mass loading reductions from a combination of LID approaches done
by The Northwest Cascade Project from 2004 through 2006 in Seattle, Washington (EPA,
2012).
McIntyre et al. (2014) performed a study on zebrafish embryos, using them as
biological indicators of the effectiveness of green stormwater infrastructure treatment, in
this case, bioretention systems. The fish were tested with untreated stormwater runoff as
well as runoff that had been treated by soil filtration. As expected, the untreated
stormwater was highly contaminated with PAHs and resulted in an array of heart
25
conditions, as well as reduced growth, reduced eye size and swim bladder inflation
(McIntyre et al., 2014). The bioretention treatment of the stormwater, by flushing through
a bioretention system, was successful in reducing nearly all developmental toxicity
(McIntyre et al., 2014). The concentrations of dissolved metals were reduced by 99% for
zinc, 72% for copper, 31% for nickel, 91% for lead, and 95% for cadmium. The PAHs
were reduced 95% by the treatment (McIntyre et al., 2014).
2.6.2
CURRENT RESEARCH
There is research currently being conducted by Jenifer McIntyre and her
colleagues at the Washington State University (WSU) Green Stormwater Infrastructure
(GSI) Facility in Puyallup, Washington studying bioretention treatment. They have set up
columns with bioretention soil medium and gravel to study the effectiveness of
bioretention treatment. Some column have plants planted on top of the bioretention
systems and some have a mulch layer instead. They allowed collected stormwater runoff
to flush through the soil columns and captured the leachate to do toxicity tests with
zebrafish on gene expression. The preliminary results of their study show significant
upregulation of detox (cyp1a) and cardiac injury genes in the fish exposed to highway
runoff (Figure 2) (McIntyre et al., unpublished). These are genes are produced in excess
when an organisms is exposed to environmental stressors. Each gene indicated a different
kind of environmental stress. The cardiac injury genes are an indication of pollutant
exposure causing cardiotoxicity. Detox genes, such as cyp1a, provide metabolic
protection against contaminants such as high molecular weight PAHs (McIntyre et. al,
26
unpublished). Metallothionein (mt2) is the gene upregulated for detox due to exposure to
metals. An upregulation would be observed with an increase in contaminants in the
system.
Figure 2. Zebrafish upregulation of detox and cardiac injury genes from exposure to
untreated stormwater runoff and stormwater runoff treated through bioretention columns
with and without plants (McIntyre et. al, unpublished).
Filtering the runoff through the bioretention cells significantly reduced expression
of those genes, but some still showed significant upregulation compared to controls. The
cyp1a gene was still significantly elevated after treatment of the stormwater runoff.
Metallothionein (mt2) was significantly reduced in the treatment with and without plants.
This indicated that it is more probable that PAHs rather than metals were responsible for
the toxic effects observed (Figure 2). The results from this test indicate that there is still a
27
toxic response from the bioretention system leachate of the treated stormwater. The
leachate from columns themselves were never tested with clean water being flushed
through. There is not conclusive evidence that the toxic effects of the leachate where
from the stormwater that had passed through or if it was from contaminants leaching out
of the soils.
These examples all used toxic stormwater runoff to measure retention capability
in the soils. The bioretention systems in all of those studies were shown to remove a
majority of the metals and PAHs, but this was only when the stormwater was
contaminated to begin with. Soils and composts have already been shown to have metal
in them. For a thorough analysis on bioretention systems it is important to consider what
can leach out of the material on its own.
Bioretention systems may be primarily used to treat stormwater, but in some
applications they are used to control the flow of water. Residential areas use rain gardens
primarily to allow for water filtration to recharge the ground water. This prevents some of
the flow onto impervious surfaces. In this situation, the water that flushes through the
retention systems is not significantly contaminated compared to stormwater runoff in an
urban setting. There is research that suggests in certain situations, metals could be
leaching out of the compost. However this has not been extensively studied in the
application of bioretention systems. It is possible that there are contaminants from the
stormwater which pass through the bioretention systems, but it is also possible that the
substrates themselves in the bioretention systems are leaching enough metal to cause a
toxic effect
28
3.
THESIS RESEARCH
There is still question of the toxicological impact of bioretention systems on
aquatic organisms. This thesis research was designed to answer the following research
questions:
-
Could the substrates in bioretention systems be a source of metal pollution? If so,
is it from the soil mixture or the gravel?
-
As more water is passed through a soil column, does the amount of metals in the
leachate decrease?
-
What is the biological response of the contaminant mixture generated from
bioretention systems to Zebrafish?
-
Does the pollutant removal efficiency of bioretention systems outweigh the
pollution generating behavior observed?
This research will help to better understand the contaminant removal potential of
different substrates. It will also add to the understanding of the pollution generating
potential of bioretention systems. Overall, this data could be used to determine if
bioretention systems are an adequate treatment for stormwater or if they could be adding
to stormwater pollution by leaching metals into aquatic systems.
29
3.1
METHODS
To address the research questions outlined above, bioretention system columns
were set up in a greenhouse at the Green Stormwater Infrastructure Facility (GSI) at the
Puyallup Extension Campus of Washington State University (WSU) in Puyallup,
Washington. Clean water was leached through the columns at varying times from August
2014 through January 2015. There were first eight separate flush tests, each a week apart,
then there was a conditioning period for the columns, where a large amount of water was
flushed through in a 3 week period. After the conditioning with the clean water,
stormwater runoff was collected and treated through though the columns. The stormwater
runoff collection and leach was done in February 2015. The leachate of the individual
tests were measured for heavy metals. Zebrafish toxicity tests were conducted at the
Northwest Fisheries Science Center of the National Oceanic and Atmospheric
Administration (NOAA) facility in Seattle, Washington. A comparative analysis was
done between the leachate of the clean water and the leachate of the collected stormwater
runoff.
3.1.1
BIORETENTION COLUMN SET-UP
The soil columns were designed to replicate a bioretention structure in
compliance with recommendations for flow control and pollutant removal (Palmer, Poor,
Hinman, & Stark, 2013) (Hinman, 2009). A typical design for a bioretention system, as
recommended by the Department of Ecology and the Seattle Municipal Stormwater
Code, includes a mineral aggregate drainage layer at the base, followed by a bioretention
30
soil media (BSM), with either mulch, plants, or a ponding zone on the top (Hinman,
2012). To allow for a better understanding of the BSM and gravel layers, the top layer
was not included in this study. Plants were eliminated given that they do not significantly
improve removal of contaminants (Palmer et al., 2013). This research compared the
leachates from two substrates used in bioretention systems, the bioretention soil medium
(BSM) and the gravel drainage layer.
Three separate columns were set up. The first column was filled with all
bioretention soil medium (BSM). The next column was filled with the gravel
representative of a drainage layer in a rain garden. The final column was both BSM and
gravel. The bioretention soil medium (BSM) was a composition of 60% mineral
aggregate, 15% compost, 15% finely shredded cedar bark and 10% drinking water
treatment residuals (WTR) (Palmer et al., 2013). This ratio is accepted by the Department
of Ecology from bioretention system stormwater treatment (Hinman, 2009). The WTRs
used were amorphous aluminum hydroxides from the Anacortes Water Treatment Plant
in Anacortes, WA. They are produced when aluminum sulfate is added to water for
treatment. The WTR are a byproduct of flocculation and are obtained in the precipitate
that forms. They were dried after use in the drinking water treatment and sieved to
remove large clumps. When added to the BSM mixture they were a fine grained material
(Palmer et al., 2013). Since they are composed of aluminum and hydrogen, and this study
is not measuring aluminum, there is no concern of them being a source of the metals
measured. The compost used was from an all organic compost company, Cedar Grove, in
Seattle, WA. The compost was Type 1 feedstock, which was derived from materials such
as yard, garden, wood, agricultural residuals and pre-vegetative food wastes (Palmer et
31
al., 2013)(Hinman, 2009). The compost used in this mixture was in compliance with
Washington Administrative Code (WAC) 173-350 which requires the organic matter
content to be between 45 and 60 percent, have a pH between 5.5 and 8, a carbon:nitrogen
ratio between 20:1 and 25:1, and less than one percent of manufactured inert materials
(Hinman, 2009). There is also metal regulations for this compost. The metals had to meet
the requirements in Table 3, below, in order to be in compliance with WAC 173-350.
Table 3. The limits in metal concentration (mg/kg dry weight) for compost to be in
compliance with WAC 173-350 (Hinman, 2009).
Metal
Arsenic
Cadmium
Copper
Lead
Nickel
Zinc
Limit
≤ 20 ppm
≤ 10 ppm
≤ 750 ppm
≤ 150 ppm
≤ 210 ppm
≤ 1400 ppm
For the gravel substrate a 3/4" Seattle Type 26 sandy gravel was used. The
gradation size distribution for this gravel layer as well as the BSM mineral aggregate are
within the Department of Ecology guidelines as shown in Table 4 (Hinman, 2009). This
was the same BSM and gravel used in the research by Palmer et al. (2013) and McIntyre
et al. (2014).
32
Table 4. BSM mineral aggregate and gravel layer aggregate gradation size distribution
(Palmer et al., 2013).
Sieve Size
BSM Mineral Aggregate
Gravel Layer Aggregate
3/8 inch
U.S. No. 4
U.S. No. 10
U.S. No. 40
U.S. No. 100
U.S. No. 200
3/4 inch
1/4 inch
U.S. No. 8
U.S. No. 50
U.S. No. 200
Percent Passing
100
95-100
75-90
25-40
4-10
2-5
100
30-60
20-50
3-12
0-1
According to the Bioretention Soil Mix Review and Recommendation for Western
Washington, a BSM in the ratio used for this experiment would have a highorganic
matter content and a high cation exchange capacity (Hinman, 2009). Theoretically the
mixture of materials used for this BSM should have a cation exchage capacity of ≥ 5
meq/100 grams of dry soil.
The columns were constructed in polyvinyl chloride (PVC) pipes measuring 10.2
cm in diameter. They were open on the top and had a plastic screen on the bottom to
avoid the loss of substrate material. The columns were placed on a wooden frame with
plastic funnels under them, draining into 4 L amber glass carboys to catch the leachate.
The PVC columns and funnels were rinsed with DI water before filling with the
substrates. Each column was filled to the 50 cm mark with the testing material. This
amount was adequate because the majority of the PAHs are sequestered in the top few
33
centimeters of the bioretention systems and the majority of metals in the top 20 cm (Hunt,
Davis, & Traver, 2012).
For the column with both BSM and gravel, the first 16 cm were gravel and the
remaining 34 cm was BSM for a total of 50 cm. This ratio was proportionally
representative of previous experiments (Palmer et al., 2013). The construction design is
depicted below in Figure 3. After each 10 cm of substrate was added, the material was
gently compressed by shaking and pressing down on it. This was done to avoid the
creation of any voids or inconsistency in the amount of material. More material was
added to ensure an equal 50 cm of total substrate for each column after compression.
Figure 3. Construction design of the three bioretention substrate columns.
34
3.1.2
BIORENTENTION TREATMENT AND LEACHATE COLLECTION
A 1 g/L Instant Ocean® Sea Salt solution was made using deionized water which
had also been filtered through a Brita and a Millipore filter. This solution is the same as
the embryonic medium solution used for culturing the laboratory zebrafish and replicates
previous zebrafish toxicity tests (McIntyre et al., 2014). It is a freshwater solution made
with the essential minerals needed for zebrafish survival. When doing a toxicity test, it is
important to be sure that the response measured is from the manipulated variable, in this
case, the contaminants in the leachate. Using an embryonic medium solution decreases
the chance that the response is due to the control water being different than solution the
zebrafish is accustomed to. The same solution was used for each treatment column,
which was pumped into the columns at a flow rate of 25 mL/min to replicate Palmer et al.
(2013) and McIntyre et al. (2014). Each tube was attached to an irrigation head made
with plastic disks and hypodermic needles to allow for drips similar to an average rainfall
(0.05 mm/s). This is the rate at which the water would infiltrate into the substrates and
was standardized for the size of the column used. The irrigation heads were fastened
above the individual columns.
For each water collection period, three liters were allowed to drip into the
columns. The amount of water retained in soil due to saturation was about one liter.
After approximately two hours, two liters of leachate was collected in each corresponding
glass carboy. In between each individual leach test, the carboys were rinsed three times
with DI water to avoid cross-contamination. The glass carboys were shaken to
homogenize prior to distribution into the separate analysis containers and storage jars.
The caps were composed of Polytetrafluoroethylene (PTFE)-lined, polypropylene.
35
For the dissolved metal, calcium, magnesium, hardness, and alkalinity (no headspace)
tests, plastic bottles were used. These samples were unpreserved. For the dissolved
organic carbon, and pH tests, amber glass bottles were used. This was also unpreserved.
An amber glass bottle was also used for the total organic carbon tests. This sample was
preserved with sulfuric acid (H2SO4). All samples were placed on ice and transported for
analysis within 24 hours. For future toxicity testing, four to six amber glass jars with at
least 150 ml of leachate were frozen within 24 hours. The column flush was completed a
total of eight times. There was at minimum one week between each flush and collecting
period where the columns were allowed to completely dry out to mimic natural rainfall
events. These eight flushes were all measured as outlined in the chemical analysis
section and samples were frozen for each flush.
Next, the columns were flushed with more water to test if the retention medium
could be conditioned prior to use to decrease the concentration of contaminants in the
leachate. This portion of the experiment was not used for the toxicity tests. It was only
done to test for the conditioning potential. In order to replicate the practices in the
experiment by Palmer et al. in 2013, water from a garden hose was used rather than the
embryonic medium as before. This was the source of water for that test. Each column
received a flush of 34 L over a two week period using the same water distribution
methods as before. After the 34 L flush of tap water, three additional liters of the tap
water was dripped through and collected for chemical analysis.
36
3.1.3
HIGHWAY STORMWATER RUNOFF
Stormwater runoff was collected from Highway 520 in Seattle, Washington in
February 2015 on the second day of a rain event. This portion of the highway is rated at
60,000 annual average daily traffic (AADT), meaning that an average of 60,000 vehicles
use the bridge each day and is considered a high volume urban highway (McIntyre,
unpublished). There is an elevated portion of the highway with a downspout routed into
the National Oceanic and Atmospheric Administrative (NOAA) Northwest Fisheries
Science Center (NWFSC) parking lot. The stormwater was collected into a large stainless
steel cistern. The stormwater is not acidic enough to be concerned with the stainless steel
leaching metals. This portion of the highway has guardrails and no plants or soil. This is
important because the only chemicals present in any stormwater collected would be from
the roadway. There is no obvious potential from contamination from pesticides, lawn
care, roof materials or other stormwater pollution besides what comes off vehicles or
from asphalt. The main contaminants of concern in these samples are heavy metals and
PAHs.
A three liter sample of the collected stormwater was flushed through each of the
bioretention columns and two liters of leachate was collected in the same procedure as
described above. This was done a few hours after collection to avoid any breakdown of
PAHs in the samples as recommended by previous research (McIntyre et al., 2014). The
leachate from the treatment was collected and frozen or sent for analysis within a few
hours.
37
3.1.4
CHEMICAL ANALYSIS
All samples were transported within 24 hours to Analytical Resource Incorporated
(ARI) Laboratory in Tukwila, Washington. ARI conducted all chemical analysis using
US Environmental Protection Agency (EPA) methods or EPA approved in-house
methods. Three water samples were measured for the conventional water chemistry
parameters: 1) The water from the 8th flush with the Instant Ocean® Sea Salt solution, 2)
the final conditioning flush with tap water and 3) the stormwater runoff through the
bioretention treatments. These tests consisted of measuring calcium (In-house method
6010C), magnesium (In-house method 6010C), alkalinity (In-house method SM 2320),
dissolved organic carbon (DOC) (EPA method 9060), total organic carbon (TOC) (EPA
method 9060), and pH (EPA method 150.1). All leachate samples, the untreated the
Instant Ocean® Sea Salt solution (1 g/L), untreated tap water, and the untreated
stormwater runoff, were measured for total arsenic, chromium, copper, lead, nickel, and
zinc (In-house method 200.8) using an inductively coupled plasma mass spectrometer
(ICP-MS). PAH measurements were not taken, but samples from the stormwater runoff
leach test were preserved with methylene chloride and frozen for possible testing in the
future. The first four leachate tests were also analyzed for cadmium and silver. The levels
were below detection so this analysis was not continued for the remainder of the
uncontaminated water treatment, but analysis was continued for the stormwater runoff
test.
38
3.1.5
ZEBRAFISH TOXICITY TESTING
The biological portion of this research was conducted to determine if there was a
toxic response to aquatic organisms from the leachate. Two separate toxicity tests were
done. The first toxicity test was with the leachate from the Instant Ocean® Sea Salt
solution treated through the BSM plus gravel column. Leach number two, four and eight
were all tested against a control. The second toxicity test was with the stormwater runoff.
The untreated stormwater runoff, the leach through the BSM, the leach through the
BSM/gravel and the leach through the gravel column were all tested against a lab control.
The lab control was the same composition as the embryonic medium used in the
uncontaminated water tests. The methods used are in replication of a previous experiment
(McIntyre et al., 2014). Zebrafish were used because as embryos they are transparent, so
the developmental changes and heart function can be visually measured. They were also
used because the formation of their heart and other organs happens within hours. A toxic
response can be measured with a relatively short exposure. Wild type (AB) zebrafish
were tested. The zebrafish were cultured and spawned at the National Oceanic and
Atmospheric Administration Northwest Fisheries Science Center (NOAA NWFSC) in
Seattle, Washington.
The day before egg collection, two separate containers each containing three
males and three females were set up. The morning before collection the water was
changed to ensure that the old eggs were disposed of. The Zebrafish were then allowed to
spawn and two to three hours later the eggs were collected, rinsed and put into a petri
dish in the incubator. The water samples were taken from the freezer and placed in a
warm water bath for thawing.
39
Once the eggs were between the eight cell and high stage cell stages, (1.25-3.5
hours post fertilization) based on the guidelines in Stages of Embryonic Development of
the Zebrafish, they were picked through to get the healthiest looking ones for the
experiment (Kimmel, Ballard, Kimmel, Ullmann, & Schilling, 1995). The eggs closest to
the ideal pictures were chosen (Figure 4, Images C-H). Eggs were chosen that had even
cell division and were symmetrical. Each petri dish was given 15 eggs by collection with
a glass pipette. There were three replicates for each treatment and the control. The control
used was an embryo rearing medium made with 1 g/L Instant Ocean® Sea Salt. The
residual water was removed and the eggs were dosed with 10 milliliters of leachate water
(or control water) per petri dish. The dishes were covered and the positions were
randomized before placement in the incubator at 28.5 °C.
40
Figure 4. Stages of Zebrafish embryo development. A: Two-cell stage (0.75 hr). B: Fourcell stage (1 hr). C: Eight-cell stage (1.25 hr). D: Sixteen-cell stage (1.5 hr). E: Thirty-two
cell stage (1.75 hr). F: Sixty-four cell stage (2 hr). G: 256-cell stage (2.5 hr). H: High
stage (3.5 hr). I: Transition between high and oblong stages (3.5 hr). (Kimmel, Ballard,
Kimmel, Ullmann, & Schilling, 1995)
After 24 hours, the water was replaced with the same initial water (i.e. treatment
or control water) and any dead eggs were removed. Survival count was measured by
counting the number of dead embryos per replicate. The dish was placed back in the
incubator until the following day. At the 48 hour post exposure point, the petri dishes
were analyzed in random order. Based on the previous work by McIntyre et al. (2014), 48
hours was sufficient for the responses measured; the heart is developed enough at this
41
point to determine toxicity. Zebrafish pigmentation begins to form after 48 hours making
it difficult to visualize heart defects and see the heart development. The hatch rate was
measured as the amount of embryos per replicate that had hatched from the chorion. At
48 hours this was an insignificant measurement and was not included in the analysis.
All embryos that were still unhatched were manually dechorionated by using
tweezers to rip the chorion without injuring the embryo. Once all the embryos were
dechorionated, the dirty water was pipetted out and replaced with clean embryonic
medium water. A few drops of tricaine methanesulfonate (250 µg/L MS-222) was added
to the dishes to anesthetize the embryos to avoid twitching and movement during
imaging.
A daub of 3% methylcellulose was put on the bottom of a plastic petri dish and
spread to be as flat as possible with minimal bubbles. The embryos were then removed
with a dropper and put on top of the methylcellulose. Excess water was removed from the
surface. One at a time, the embryos were shifted away from each other and oriented in the
same direction. They were shifted on their sides facing the left with the eyes stacked,
creating uniformity in the imaging. Each embryo was analyzed under a Nikon SMZ800
stereomicroscope (Nikon Instruments Inc., Melville, NY) and imaged with a Fire-I 400
digital camera (Unibrain Inc., San Ramon, CA). A five second video was also taken of
the periventral area by focusing the microscope on the heart. This was repeated for each
replicate and each treatment group as well as the control.
42
3.2
ANALYSIS
3.2.1
IMAGE AND VIDEO ANALYSIS
The still images were measured using ImageJ 1.48v software (National Institutes
of Health). These measurements included the total length of the embryo, which is the sum
of the length from the tail through the notochord to the ear plus the length straight
through the middle of eye to the end of the embryo (Figure 5A). Consistent use of this
method normalized measurements for kinks and head angle between the different
embryos. The eye area was measured by tracing the circumference the eye (Figure 5B).
The periventral and pericardial area was also measured. The periventral area consists of
the area with the heart as well as any area of fluid accumulation under the yolk sac
(Figure 5D). The pericardial area is the area where the heart and surrounding fluid are
(Figure 5C). This excludes the area under the yolk sac where the blood initially flows
before entering the heart chamber. Since all the measurements were done in the ImageJ
program at the same microscope zoom setting (2X for length and 6.3X for heart
measurements), the images will all be compared in pixels rather than a metric unit. The
videos were used to analyze heart function and determine the borders of the periventral
and pericardial area. Periventral blood pooling was identified when there was blood
pooled in the periventral area that did not get pumped into the pericardial area where the
heart was. The heart rate was counted in each 5 second clip and multiplied to get beats
per minute.
43
A
B
C
D
Figure 5. Zebrafish image measurements. A) Length, B) Eye Area, C) Pericardial area,
and D) Periventral area.
3.2.2
STATISTICAL ANALYSIS
For the toxicological results, all statistics were done using JMP Pro 11.0.0 (SAS
Institute Inc.). The statistical means of each replicate and measurement (i.e. length, eye
area, pericardial area) were analyzed to compare the treatments to the control. The
measurements were first tested for normality and equal variance using Shapiro-Wilk and
Levene’s tests respectively. None of the treatments met the assumptions of both the
Shapiro-Wilk and Levene’s test using a p-value of 0.05. Since the ANOVA test is
considered to be robust in the analysis of environmental tests it was still used to remain
consistent with previous Zebrafish toxicity tests (McIntyre et al., 2014). After the
ANOVA, a Dunnett’s post-hoc analysis was done to test the statistical difference
44
compared to the control group. For the second toxicological study with the stormwater, a
Dunnett’s post-hoc analysis was also done with the stormwater as the control to test the
differences between the stormwater and the treatments through the bioretention
substrates.
3.3
RESULTS
The results for this thesis research are broken up into conventional chemistry
results, metal concentrations in the leachate when uncontaminated water was treated
through bioretention systems, metal concentration in stormwater before and after
treatment and toxicological results. These results will be analyzed further in the
discussion section.
3.3.1
CONVENTIONAL CHEMISTRY RESULTS
Alkalinity, pH, magnesium, calcium, total organic carbon (TOC) and dissolved
organic carbon (DOC) were all measured in the following three tests: 1) the leachate of
the 8th water flush with the Instant Ocean® Sea Salt solution, 2) the leachate after the
conditioning period with the tap water, and 3) the stormwater runoff leachate.
The pH in the leachate of the 8th Instant Ocean® Sea Salt solution being treated
through the bioretention columns did not change very much. Before treatment the
solution was 7.39. It lowered when filtered through either of the columns, with the lowest
being the leachate from the BSM column at a pH of 7.07. The pH of the BSM plus gravel
and the gravel only column were both 7.28. For the conditioning treatment, the initial pH
45
before treatment was 7.35, which was very similar to the 8th leach test. With treatment
through the BSM, BSM plus gravel and gravel only column, the pH decreased to 7.19,
7.14 and 7.18 respectively. The untreated stormwater runoff had a pH of 7.1. Treatment
of the stormwater runoff had an opposite effect on pH as the treatment on
uncontaminated water. With treatment through the bioretention columns the pH
increased to 7.35 through the BSM column, 7.61 through the BSM plus gravel column
and 7.81 through the gravel only column (Figure 6).
pH
8
pH
7.6
Leach # 8
7.2
After Conditioning
Stormwater
6.8
6.4
Before
BSM Only
Treatment
BSM +
Gravel
Gravel Only
Figure 6. pH of the 8th leach test with uncontamianted water, leachate after conditioning
and treatment of stormwater runoff initially and after being treated through the columns
for the conventional water chemistry parameters. Error bars are not depicted but are ±
0.01 which is the reporting limit.
The alkalinity increased with treatment through the bioretention columns for both
the 8th uncontaminated water flush as well as the leachate of the uncontaminated water
46
after the conditioning period. In both tests, it increased more in the treatment through the
BSM and the BSM plus gravel columns compared to the gravel only column. The
alkalinity of stormwater runoff was much higher than the uncontamianted water before
treatment, at 52 mg/L CaCO3, whereas the uncontamianted water for the 8th flush and the
flush after conditioning was only 3.7 and 3.9 mg CaCO3/L respectively. The treatment of
stormwater runoff through the BSM or the BSM plus gravel decreased the alkalinity,
while the treatment through the gravel increased it (Figure 7).
Alkalinity
60
mg/L CaCO3
50
40
Leach # 8
30
After Conditioning
20
Stormwater
10
0
Before BSM Only
Treatment
BSM +
Gravel
Gravel Only
Figure 7. Alkalinity in mg/L CaCO3 of the 8th leach test with uncontamianted water,
leachate after conditioning and treatment of stormwater runoff initially and after being
treated through the columns for the conventional water chemistry parameters. Error bars
show ± 1 mg/L CaCO3 which is the reporting limit.
The initial conentration of calcium was higher in the stormwater runoff than the
uncontaminated Instant Ocean® Sea Salt solution. Treatment of the 8th leachate test with
47
uncontamianted water through the bioretention columns increased the calcium. The
highest concentration was from the treatment through the gravel only column. This was
the same pattern with the leachate of the test after the conditioning. The concentrations of
the calcium after conditioning were much higher than the 8th leachate test. An opposite
pattern was observed for the treatment of stormwater runoff. Calcium decreased with
treatment of stormwater runoff with the treatment through the gravel only column being
the lowest (Figure 8).
Calcium
80
mg Ca 2+ /L
60
Leach # 8
40
After Conditioning
Stormwater
20
0
Before
BSM Only
Treatment
BSM +
Gravel
Gravel Only
Figure 8. Calcium concentration in mg Ca2+/L of the 8th leach test with uncontamianted
water, leachate after conditioning and treatment of stormwater runoff initially and after
being treated through the columns for the conventional water chemistry parameters. Error
bars are not depicted but are ± 0.05 mg Ca2+/L which is the reporting limit.
48
The magnesium in the Instant Ocean® Sea Salt solution before treatment for the
8th leachate test as well as the test after the conditioning period was much higher than the
untreated stormwater runoff. The 8th leachate test magnesium concentration decreased
with treatment through the bioretention columns while the concentrations after the
conditioning period as well as the treatment of stormwater runoff stayed relatively stable
(Figure 9).
Magnesium
35
mg Mg 2+ / L
30
25
20
Leach # 8
15
After Conditioning
10
Stormwater
5
0
Before
BSM Only
Treatment
BSM +
Gravel
Gravel Only
Figure 9. Magnesium concentration in mg Mg2+/L of the 8th leach test with
uncontamianted water, leachate after conditioning and treatment of stormwater runoff
initially and after being treated through the columns for the conventional water chemistry
parameters. Error bars are not depicted but are ± 0.05 mg Mg2+/L which is the reporting
limit.
Initially there was a very low concentration of dissolved organic carbon (DOC) in
the Instant Ocean® Sea Salt solution in the 8th leach test as well as the test after the
conditioning period. With treatment during the 8th leach test through the BSM and BSM
plus gravel columns this concentrations increased greatly. After the conditioning period
49
the concentration did not change as drastically with treatment through the columns. The
stormwater runoff initially had a much higher concentration of DOC comparred to the
other uncontaminated water flushes. Following the same pattern as the 8th leach test, the
concentrations increased with treamtent through the BSM and the BSM plus gravel
columns. None of the tests showed a change in DOC with treatment through the gravel
column (Figure 10).
DOC
30
mg/L DOC
25
20
Leach # 8
15
After Conditioning
10
Stormwater
5
0
Before
BSM Only
Treatment
BSM +
Gravel
Gravel Only
Figure 10. Dissolved organic carbon concentration in mg DOC /L of the 8th leach test
with uncontamianted water, leachate after conditioning and treatment of stormwater
runoff initially and after being treated through the columns for the conventional water
chemistry parameters. Error bars are ± 1.5 mg DOC/L which is the reporting limit.
In the treatment of the Instant Ocean® Sea Salt solution in the 8th leachate test
there was an increase in total organic carbon (TOC) in the BSM and BSM plus gravel
treatments but not the gravel alone. This was the same pattern as observed in the DOC
50
measurements. After the conditioning period there was very low concentrations of TOC.
The untreated stormwater runoff had very high concentrations of TOC. The
concentrations degreased with treatment through the BSM and the BSM plus gravel
columns (Figure 11).
TOC
60
mg/L TOC
50
Leach # 8
40
After Conditioning
30
Stormwater
20
10
0
Before BSM Only
Treatment
BSM +
Gravel
Gravel
Only
Figure 11. Total organic carbon concentration in mg TOC /L of the 8th leach test with
uncontamianted water, leachate after conditioning and treatment of stormwater runoff
initially and after being treated through the columns for the conventional water chemistry
parameters. Error bars are ± 1.5 mg TOC/L which is the reporting limit.
3.3.2
METAL FROM UNCONTAMINATED WATER TREATMENT
In the treatment of the Instant Ocean® Sea Salt solution through the bioretention
soil medium (BSM) only column there was a general trend of an increase in metal
concentration in the first few flushes. The zinc and copper concentrations were the
highest, peaking around 35 µg/L for zinc and 31.7 µg/L for copper. The first eight flushes
were completed when 24 liters had passed through the column. The conditioning period
51
was completed when 61 liters had been flushed through the columns. After conditioning
all of the metal concentrations were under 10 µg/L which was lower than the values of
the first flush through the treatment column. The values for the metal analysis as well as
the detection limits are in Appendix A. A graph of the metal concentrations are shown in
Figure 12. Cadmium and silver were excluded because the levels were below detection.
BSM
60
Concentration (µg/L)
50
40
30
20
10
0
0
10
20
30
40
50
60
Liters Flushed
Arsenic
Chromium
Copper
Lead
Nickel
Zinc
Figure 12. Metal concentrations in µg/L in the leachate of the bioretention soil medium
(BSM) during the first eight flushes and the measurements after the conditioning period
shown as a function of how many liters were flushed through the column.
The increase in metal concentration in the bioretention soil medium plus the
gravel layer follows the same pattern as the bioretention soil medium only column. There
52
was a general increase in the first few flush periods followed by a decline in metal
concentrations. Again the zinc and copper concentrations were the highest in the leachate.
Zinc peaked at 28 µg/L. The copper was much higher in the BSM plus gravel column
compared to the BSM only column, peaking at 60.1 µg/L. After the conditioning period
at 61 liters, all metal concentrations were under 10 µg/L just as in the BSM only column
(Figure 13).
BSM + Gravel
60
Concentration (µg/L)
50
40
30
20
10
0
0
10
20
30
40
50
60
Liters Flushed
Arsenic
Chromium
Copper
Lead
Nickel
Zinc
Figure 13. Metal concentrations in µg/L in the leachate of the bioretention soil medium
(BSM) plus the gravel layer during the first eight flushes and the measurements after the
conditioning period shown as a function of how many liters were flushed through the
column.
The gravel layer treatment did not result in an overall increase in metal
concentrations of the leachate, however it is important to note that the concentrations of
arsenic, copper and chromium all increased as more water was flushed through and then
53
decreased in the end of the conditioning. The arsenic concentration at the end of the
conditioning period was higher than the first flush for the gravel layer and lower than the
first flush for the BSM and BSM plus gravel treatments. For the chromium, copper, lead,
nickel and zinc, the metal concnetrations at the end of conditioning were lower than the
leachate of the first flush for all treatments. The gravel alone layer produced little above
detection for lead and nickel and was below detection for zinc. The metal leaching from
the gravel column was never over 10 µg/L (Figure 14).
Gravel
60
Concentration (µg/L)
50
40
30
20
10
0
0
10
20
30
40
50
60
Liters Flushed
Arsenic
Chromium
Copper
Lead
Nickel
Zinc
Figure 14. Metal concentrations in µg/L in the leachate of the gravel only during the first
eight flushes and the measurements after the conditioning period shown as a function of
how many liters were flushed through the column.
54
3.3.3
METALS FROM TREATING STORMWATER RUNOFF
The concentrations of metals in the untreated stormwater runoff were the highest
for zinc and copper with 420 µg/L and 107 µg/L respectively. The levels for each metal
in the untreated stormwater runoff are shown in Figure 14. When stormwater runoff was
passed though the different treatment columns most had a reduction in concentration.
Arsenic was the only metal that increased as it was treated and this was only observed in
the BSM plus gravel treatment column. The arsenic, chromium and nickel concentrations
were lowest in the gravel leachate and highest in the BSM plus gravel leachate. The
copper, lead and zinc were all lowest in the BSM leachate indicating the retention of the
highest concentration in the soil mixture (Figure 15).
55
Figure 15. Metal concentrations in µg/L of the leachate of stormwater runoff treated
through the different experimental columns. Stormwater indicates untreated stormwater
runoff. Values are in µg/L. Error is ± the reporting limit. For arsenic this is 0.2 µg/L, for
chromium it is 0.5 µg/L, for copper it is 0.5 µg/L, for lead it is 0.1 µg/Lm for nickel it is
0.5 µg/L and for zinc it is 4 µg/L.
56
3.3.4
TOXICOLOGICAL RESULTS- UNCONTAMINATED WATER
The second, fourth and eighth leachates of uncontaminated water through the
BSM plus gravel layer were used for toxicity analysis. The zebrafish tested did not show
a significant difference from the control in any of the parameters measured with a p value
of 0.05 (Table 5). The measurements for length, eye area, periventral area, pericardial
area and heart rate are shown in Figures 16-20.
Table 5. Results from one-way ANOVA for leachate of uncontaminated water passing
through different treatment columns for measurements of length, eye area, periventral
area (PVA), pericardial area (PCA) and heart rate. F ratios for comparison between
averages of replicates and different columns. No values were statistically significant
(p < 0.05).
Length Eye Area
F ratio
P Value
0.92
0.4738
2.6949
0.1166
PVA
PCA
Heart Rate
0.707
0.5743
0.6617
0.5984
1.0503
0.4219
57
Length
960
Pixels
940
920
900
880
860
840
Control
#2
#4
#8
BSM + Gravel Column Treatment
Figure 16. Zebrafish lengh in pixels measured at 48 hours. Zebrafish embryos exposed to
Instant Ocean® Sea Salt solution (control) treated through bioretention soil medium plus
gravel for the 2nd, 4th and 8th leach test. Results shown with ± 1 Standard Error.
Eye Area
Square Pixels
50000
45000
40000
35000
30000
Control
#2
#4
#8
BSM + Gravel Column Treatment
Figure 17. Zebrafish eye area in square pixels measured at 48 hours. Zebrafish embryos
exposed to Instant Ocean® Sea Salt solution (control) treated through bioretention soil
medium plus gravel for the 2nd, 4th and 8th leach test. Results shown with ± 1 Standard
Error.
58
Periventral Area
Square Pixels
31000
27000
23000
19000
15000
Control
#2
#4
#8
BSM + Gravel Column Treatment
Figure 18. Zebrafish periventral area in square pixels measured at 48 hours. Zebrafish
embryos exposed to Instant Ocean® Sea Salt solution (control) treated through
bioretention soil medium plus gravel for the 2nd, 4th and 8th leach test. Results shown
with ± 1 Standard Error.
Pericardial Area
Square Pixels
24000
20000
16000
12000
Control
#2
#4
#8
BSM + Gravel Column Treatment
Figure 19. Zebrafish pericardial area in square pixels measured at 48 hours. Zebrafish
embryos exposed to Instant Ocean® Sea Salt solution (control) treated through
bioretention soil medium plus gravel for the 2nd, 4th and 8th leach test. Results shown
with ± 1 Standard Error.
59
Beats Per Minute
Heart Rate
180
170
160
150
140
130
120
110
100
Control
#2
#4
#8
BSM + Gravel Column Treatment
Figure 20. Zebrafish heart rate in beats per minute measured at 48 hours. Zebrafish
embryos exposed to Instant Ocean® Sea Salt solution (control) treated through
bioretention soil medium plus gravel for the 2nd, 4th and 8th leach test. Results shown
with ± 1 Standard Error.
3.3.5
TOXICOLOGICAL RESULTS- STORMWATER RUNOFF
For the stormwater runoff that was collected, the leachate from each treatment
columns was used for toxicity testing, as well as the untreated stormwater itself. The
results of the one way ANOVA are in Table 6. There was no significant difference in the
length, eye area or heart rate of the zebrafish when exposed to stormwater or treated
stormwater. (Figure 21, 22 and 25). For the eye size, while the overall test did not show
significance, the results of the Dunnett’s post hoc analysis did show that the untreated
stormwater runoff had significantly smaller eyes than the control water (P=0.0197)
(Figure 22). The untreated stormwater runoff and the treatment with only the gravel layer
60
had a significantly larger periventral area compared to the control. This was based on the
Dunnett’s post hoc analysis comparing to the control with a p value of 0.0199 for the
untreated stormwater runoff and a p value of 0.0237 for the gravel layer compared to the
control (Figure 23). There was no significant difference in the pericardial area of the
different treatments when compared to the control (Figure 24).
Table 6. Results from one-way ANOVA for leachate of stormwater runoff passing
through different treatment columns for measurements of length, eye area, periventral
area (PVA), pericardial area (PCA) and heart rate. F ratios for comparison between
averages of replicates and different columns. Values in bold indicate statistical
significance (p < 0.05).
F ratio
P Value
Length
Eye Area
PVA
PCA
Heart Rate
0.5282
0.718
3.3382
0.0556
4.6721
0.0219
4.8256
0.0199
1.4797
0.2967
61
Length
900
Pixels
880
860
840
820
800
Control
Gravel
BSM
BSM +
Gravel
Runoff
Treatment
Figure 21. Zebrafish length in pixels measured at 48 hours. Zebrafish embryos exposed to
untreated stormwater runoff (runoff) treated through the different bioretention substrate
columns. Instant Ocean® Sea Salt solution was used for the control. Results shown with
± 1 Standard Error.
Eye Area
Square Pixels
40000
35000
*
30000
25000
Control
Gravel
BSM
BSM +
Gravel
Runoff
Treatment
Figure 22. Zebrafish eye area in square pixels measured at 48 hours. Zebrafish embryos
exposed to untreated stormwater runoff (runoff) treated through the different bioretention
substrate columns. Instant Ocean® Sea Salt solution was used for the control. Results
shown with ± 1 Standard Error. The * indicates significance (p<0.05) from the results of
the Dunnett’s post hoc analysis comparing each mean to the control.
62
Periventral Area
Square Pixels
30000
*
*
25000
20000
15000
10000
Control
Gravel
BSM
Treatment
BSM +
Gravel
Runoff
Figure 23. Zebrafish periventral area in square pixels measured at 48 hours. Zebrafish
embryos exposed to untreated stormwater runoff (runoff) treated through the different
bioretention substrate columns. Instant Ocean® Sea Salt solution was used for the
control. Results shown with ± 1 Standard Error. The * indicates significance (p<0.05)
from the results of the Dunnett’s post hoc analysis comparing each mean to the control.
Pericardial Area
Square Pixels
22000
19000
16000
13000
10000
Control
Gravel
BSM
BSM +
Gravel
Runoff
Treatment
Figure 24. Zebrafish pericardial area in square pixels measured at 48 hours. Zebrafish
embryos exposed to untreated stormwater runoff (runoff) treated through the different
bioretention substrate columns. Instant Ocean® Sea Salt solution was used for the
control. Results shown with ± 1 Standard Error.
63
Beats Per Minute
Heart Rate
160
150
140
130
120
110
100
Control
Gravel
BSM
BSM +
Gravel
Runoff
Treatment
Figure 25. Zebrafish heart rate in beats per minute measured at 48 hours. Zebrafish
embryos exposed to untreated stormwater runoff (runoff) treated through the different
bioretention substrate columns. Instant Ocean® Sea Salt solution was used for the
control. Results shown with ± 1 Standard Error.
The increased periventral area in this study was correlated with the presence of
blood pooling under the yolk sac of the zebrafish. This was not a measurement that was
taken in the analysis, so the reporting of this is only looking at observational trends.
There was no statistics done on this and there was not a complete count. The untreated
stormwater runoff had a high percentage of zebrafish which appeared to have blood
pooling under the yolk sacs (Figure 26A). When treated through the gravel column, the
phenotypic response was still present (Figure 26B). Treatment through the BSM column
and the BSM plus gravel column did not result in as many zebrafish with blood pooling
(Figure C and D).
64
A
B
C
D
Figure 26. Representative images of zebrafish of each treatment group used in the
stormwater leaching toxicity analysis. Images taken at 48 hours post fertilization. A)
Untreated stormwater runoff, B) Stormwater runoff treated through the gravel column, C)
Stormwater runoff treated through the BSM plus gravel column, and D) Stormwater
runoff treated through BSM column. Arrows point to blood pooling under the yolk sac.
65
4.
DISCUSSION
This thesis was designed to analyze the potential for bioretention systems to leach
enough metals to cause an ecological impact. It was first found that the soils used in the
bioretention systems in this study leached very high concentrations of metals. This
indicated that soils can be a source of metal pollution. With conditioning of the soils, by
flushing uncontaminated water through them, the concentrations were reduced. To
address the ecological impact, toxicity tests were performed. The results of the toxicity
tests show that the bioretention systems themselves do not cause an ecological impact,
within the parameters of this experiment.
4.1
LEACHING OF METALS
The bioretention systems were flushed with eight treatments of uncontaminated
water before the conditioning portion of the experiment. After the conditioning with
uncontaminated water, stormwater runoff was flushed through the columns. The metal
concentrations in each step of this process with will be analyzed in greater detail in order
to get a better understanding of the processes which were taking place within the soils.
4.1.1 UNCONTAMINATED WATER
Both of the columns containing bioretention soil medium (BSM and BSM plus
gravel) had an initial increase in metal concentration. It was predicted that the first flush
would have the highest metal concentrations. In this experiment the highest metal
66
concentrations were found between the second and fourth flush treatments. Conventional
water chemistry measurements (alkalinity, pH, DOC, etc.) were not taken until the eighth
water flush, so it is not definite what caused this pattern.
The cation exchange capacity (CEC) can be used to potentially explain the
affinity of cations to soils. By simple definition, the numeric value of the cation exchange
capacity of a soil is the concentration of negatively charged sites present in the soils and
available to adsorb exchangeable cations (Sikora & Moore, 2014). The higher the CEC,
the more cations the soil can adsorb. As more cations are added, the soil becomes less
negative and the CEC goes down, making the soils less available to bind with the inflow
of cations (Sikora & Moore, 2014). At this point of saturation, the cations begin to
exchange on the binding sites based on affinity. In this thesis research, it appears that the
soils were not fully saturated with cations at the beginning, resulting in a lower leaching
of metals in the first few flushes. The relative affinity for the cations in average soil
outlined in the cation exchange capacity is in the order as follows:
Al3+ > H+ > Ca2+ > Mg2+ > K+ = NH4+ > Na+ (Chapin, Matson, & Vitousek, 2011).
Since our inflow water was slightly basic and did not contain much calcium
(Figure 6 and 8), there may not have been much exchange initially. In this research the
inflow water had elevated concentrations of magnesium, between 25 and 30 mg/L
(Figure 9). Based on the cation exchange capacity, the soils would have retained this
magnesium. The cation exchange capacity (CEC) further explains the soil’s affinity for
metal ions. The soils could have contained high levels of metals from the beginning, most
likely from the compost, based on the compost regulations outlined in the methods
67
section (Seelsaen et al., 2007). The Irving-Williams series compares heavy metal affinity,
some of the metals are not always found in soils.
Based on the Irving-Williams series, the soils have a stronger affinity for the
following heavy metals, listed in order of affinity:
Pb2+ > Cu2+ > Ni2+ > Co2+ > Zn2+ > Cd2+ > Fe2+ > Mn2+ > Mg2+ (Seelsaen et al., 2007)
All of the heavy metals in that sequence have a higher affinity than magnesium. By the
eight water flush, the bioretention soils began to have a retention of magnesium (Figure
9). This is also when the leaching of the metals was far less than at the peak. This
indicates that the soils were no longer saturated with metals.
As predicted, the results of the conditioning portion of the experiment show a
decline in metals as more water is passed through. The water used to flush was consistent
through the first eight flushes; there was no substantial difference in mineral input or pH.
The conditioning period was performed with tap water that was not tested for
conventional chemistry parameters; however the flush after the conditioning period was
performed with the same Instant Ocean® Sea Salt solution as the first eight flushes.
Overall, the flushes appeared to rinse the columns out. The concentrations of metals after
the conditioning of the BSM and the BSM plus soil column were all under 10 µg/L but
appeared to be still declining (Figure 12 and 13). Other research could be used to
determine if the concentrations could have become lower still by extending the
conditioning period longer. As the declining slowed, it could be assumed that the soils
had available negative charge to then begin retaining metals, as seen in with the
application of stormwater.
68
The cation exchange capacity of soils and the ability for sorption of metals is
highly dependent on pH (Lanno, 1998). The water used to flush the columns was slightly
basic. The hydroxides added into the soil could have formed complexes with the organic
matter resulting in the release of organic matter and hydrogen ions (Lanno, 1998). The
hydroxides would, in turn, react with the metal already present in the soil to form
hydrolyzed metal complexes in the forms of MeOH+ where the Me would indicate a
metal ion (Lanno, 1998). In addition to the DOC that would have been leached by just
rinsing water through, this could have increased it. This in part explains why initially the
columns were leaching out organic matter (Figures 10 and 11), and why the pH was
decreased in the leach compared to the inflow water (Figure 6).
Based on this limited information from this research and the soil properties
outlined with the cation exchange capacity it can be concluded that the soils retained
magnesium and hydrogen ions in the first few flushes and, after a while, became oversaturated with cations (Lanno, 1998). This could have altered the soil chemistry and
changed the CEC of the soil (Sikora & Moore, 2014). Eventually the over-saturation
would lead into a leaching of cations, which would be the metals already present in the
soils (Sikora & Moore, 2014).
Another potential explanation has to do with the rate of inflow of the water. After
the initial flush of water the columns could have been more compressed from the water
addition. Even though the columns were allowed to dry, the compression would have
remained. The compact soils would have slowed the flush down, allowing for more
metals to be removed from the substrates because the residence time in the water column
would have been longer, however this research did not take the appropriate
69
measurements to confirm this theory. These are just a few potential explanations. To
study this in greater detail, soil samples could have been taken and analyzed for metals as
well as water chemistry parameters could have been taken with each flush, rather than at
the end of the experiment.
4.1.2
STORMWATER RUNOFF
Based on this principle, by the end of the conditioning period, the CEC would
have been higher than the initial CEC, which could explain why metals were retained in
the soils when the stormwater was flushed through (Figure 15). When stormwater runoff
was flushed through these pre-conditioned columns the metal concentrations in the
leachate were drastically reduced from the untreated stormwater measurements. Since
soil samples were never analyzed, it is only predicted that the conditioning of soils
resulted in much lower concentrations of metals than at the start of the experiment. This
means that there would have been more negative sites available in the soils and a higher
CEC. At that point, the capacity of the soils to retain metal ions would have been much
higher. The soil columns then retained metals from the stormwater runoff. This response
supports previous research and the over goal of stormwater bioretention structures.
McIntyre et al. found that the same bioretention soil medium with the gravel layer
reduced zinc by 99%, copper by 72%, nickel by 31%, lead by 91% and cadmium by 95%
(2014). This thesis research found that the same BSM plus gravel column resulted in a
reduction in zinc by 75%, copper by 44%, nickel by 11%, lead by 56% and cadmium by
33% in the treated stormwater runoff. The study by McIntyre et al. (2014) did not pre70
condition the soils and had a higher reduction in metal concentrations. The stormwater
runoff used by McIntyre at al. (2014) was similar to this thesis research with an average
pH of 6.9, compared to this research with a pH of 7.1. As far as the soil columns go,
conditioning of the soils also reduces the total organic carbon in the soils (Figure 11),
since McIntyre et al. (2014) did not condition the soils, if can be assumed that the soils
would have had a higher organic matter content. The stormwater used in that study also
contained much higher DOC ranging from 25 to 400 mg/L, whereas this thesis research
used stormwater with a DOC of 17 mg/L (Figure 10). This could have bound to more
metals passing through (as explained above), which would have supported the findings of
a higher reduction in metal concentrations.
A study on a summary of multiple low impact development approaches as
described in the literature review of this thesis, found a 76% reduction in zinc, a 83%
reduction in copper and a 90% reduction in zinc (EPA, 2012). Other research on
bioretention systems found a reduction in zinc of 88%, 93% for copper and 97% for zinc
(Seelsaen et al., 2007). Neither of those two studies reported the pH or DOC so it is
difficult to compare what could have cause the difference in the results. None of these
studies used pre-conditioned soils and they all used stormwater which was more
contaminated than this thesis study (McIntyre et al., 2014)(EPA, 2012)(Seelsaen et al.,
2007).
71
4.2
TOXICOLOGICAL EFFECTS
In the treatment of the stormwater runoff, an increased periventral area in the
untreated stormwater runoff as well as in the stormwater that was passed through the
gravel layer was measured. This effect was not observed in the BSM or the BSM plus
gravel treatments. This indicated that the BSM is the substrate responsible for reducing
that toxic effect by decreasing contamination concentrations.
For the toxicity tests, the leachate from the uncontaminated water for the soil plus
gravel column was tested. Zebrafish were exposed to the second, fourth and eighth leach.
Out of these three samples, the highest metal concentrations were found in the second
and fourth leach, with copper concentrations over 60 µg/L in the second leach and zinc
concentrations at 26 µg/L in the fourth leach. There was no toxic effect from this leachate
in the parameters measured. There was a toxic response in the untreated stormwater
runoff and the treated stormwater through the gravel layer. In the untreated stormwater
runoff, the copper was 107 µg/L and the zinc was 420 µg/L. In the stormwater treated
through the gravel column the copper was 52.6 µg/L and the zinc was 105 µg/L.
The only metal that is higher in the treated stormwater compared to the treated
uncontaminated water is zinc (Figure 15). The zinc concentration in the stormwater
treated through the soil plus gravel column was 101 µg/L. This is very close to the zinc
concentration in the stormwater passing through the gravel treatment. There are three
possible explanations for what was causing the toxic effects observed. The first
explanation is that the zinc was causing the toxic effect and other water chemistry
parameters present (i.e., organic matter) in the soil leachate were providing a protection
against the zinc toxicity, as outlined in the biotic ligand model (Rand, 1995). The second
72
is that other contaminants, such as PAHs were leaching through the columns and causing
the toxicity, this would have been removed with the BSM. The third, and most plausible
possible explanation is that it was a mixture of both of these that caused a toxic response
in the zebrafish.
The toxicity of metals is highly dependent on water chemistry (Bergman &
Dorward-King, 1996). In freshwater fish this concept is known as the biotic ligand model
(Rand, 1995). The premise behind the biotic ligand model is that the fish has a toxic site
of action, known as the biotic ligand, this is usually the ion transport channels. Free metal
ion in the water can bind with the biotic ligand and block essential cation exchange
(Rand, 1995). Certain water paraemters effect this action, such as the concentration of
dissolved organic carbon (DOC), the pH and other cations present in the water (Rand,
1995). The biotic ligand model shows that metal toxicity can be predicted with freshwater
fish by measuring the metal ions, the other cations, such as calcium, magnusium and pH,
and the dissolved organic carbon. High metal ions would indicate a greater chance of a
toxic effects, but when there is also a high DOC or high competing cations in the
solution, the toxic effect can be mitigated (Rand, 1995).
Dissolved organic carbon (DOC) in the water can bind with and form complexes
with metal ions. Since metal in the ionic form is the most toxic due to the binding
potential on the organisms, this makes the metals less bioavailable to an organism and
therefor less toxic (Rand, 1995). This is all also affected by pH, since that is basically a
presence of either hydroxide compounds or hydrogen ions (Rand, 1995). The presence of
hydroxides can also bind with metals and form complexes in a basic solution. In an acidic
solution there is the the presence of hydrogen ions, which are positively changed, just like
73
the metal ions (Bergman & Dorward-King, 1996). The hydrogen binds on the biotic
ligand and, in a sense neatralizes it, so that site is no longer availabel for the metals to
bind to. The same action is observed with the presence of calcium and magnesium in the
water. They act as competition for binding sites to an organisms.
In this research, the results of the conventional chemistry show that metals were
highest in the treatments that did show a toxic effect. The pH increased with treatment of
the stormwater through the bioretention columns while it decreased when the
uncontamianted water was flushed though the treatment columns (Figure 6). A decreased
pH would indicate more hydrogen ions, which would provide toxicity protection (Rand,
1995). The calcium and magnesium concentrations were both higher in the treated
uncontaminated water than in the treated stormwater, this would also provide for
protection against toxicity (Rand, 1995). For the stormwater, the calcium decreased upon
treatment and the magnesium increased but only slightly (Figures 8 and 9). The DOC in
the treatment of stormwater runoff was higher in the treatments containing soils leaching
22-23 mg/L dissolved carbon. This was lower than the observed DOC in the stormwater
treated through the gravel layer, which had a DOC of 17.6 mg/L. This supports the
toxicity observed based on the biotic ligand model and the higher DOC forming
complexes with the metal ions (Rand, 1995).
Looking only at metal concentration in the leachate of the uncontamianted water
passing through the columns, a toxic response would have been expected (Linbo,
Baldwin, McIntyre, & Scholz, 2009). Because of the high levels of calcium, magnesium
and DOC this effect could have been reduced. A toxic effect in the untreated stormwater
was measured as small eye size. This was not observed in the leachate from the
74
uncontamianted water even though the metal concentrations were very high (Figure 22).
This was also not observed in any of the treatments of the stormwater runoff. This
indicated that the treatement was sucessful in reducing that toxic effect. Based on the
biotic ligand model, the zinc and copper ions present could have been responsible for the
toxicity in the stormwater and the stormwater treated through the gravel column, but the
differences were not that drastic so further research would be needed to make this
conclusion. There is more than just metals in stormwater pollution. Since the only
leachate which showed a toxic response was with stormwater runoff, the investigation of
other pollutants is needed.
This research only measured metals. If metals were responsible for the toxicity, it
would also be predicted that the other parameters such as length and hatch rate would
have been effected (Linbo et al., 2009). These measurements were not significantly
different than the control. The increased periventral area is an indication of blood pooling
outside the heart chamber and would not be a typical response due to metal exposure
(Incardona et al., 2004). This response is more indicative of exposure to an organic
compound such as a PAH. Stormwater collected in the same locations from 2011 to 2012
contained total PAH concentrations ranging from 4 to 10 µg/L and was shown to be toxic
to zebrafish (McIntyre et al., 2014). The samples from McIntyre et al. (2014) also had
elevated metals, so the cause of the toxicity can only predicted based on comparison with
the toxic response to the concentrations measured. In this thesis research, it is also
plausible that the PAHs were attributing to at least some of the toxicity. In order to be
positive, further research would be needed which would include measurements of how
the soil treatments affect PAH measurements. The toxicity observed in the untreated
75
stormwater and the leachate from the gravel column was most likely due to a
combination of both metals as well as other contaminants present in the stormwater.
4.3
OTHER RESEARCH
A sample of the leachate of the uncontaminated water being treated through the
BSM and the BSM and gravel column was used by the researchers at the WSU Puyallup
GSI Facility for a qPCR (quantitative polymerase chain reaction) (McIntryre et al.,
unpublished). This is a test on zebrafish to analyze gene expression of cardiac injury
genes and detoxification genes when exposed to contaminants. The leachate was used in
the toxicity tests measuring cardiac injury genes as previously discussed. This was done
specifically to focus on what was causing upregulation of cardiac injury genes in their
original study. If the toxicity was due to the bioretention systems, than it would be
expected that the toxicological effects could be observed in the samples of the
uncontaminated water flushed through the bioretention soils. However, if the PAH’s or
other contaminants from the stormwater was the reason for the observed response in the
previous studies, than this would not be observed with the substrate leachate alone. If the
bioretention systems were not successful in retaining the contamination from the
stormwater runoff, than there would be a similar response to the untreated stormwater
runoff. Because stormwater contains other contaminants besides metals, which are more
toxic (i.e., polycyclic aromatic hydrocarbons), the toxic response of zebrafish would be
76
decreased with bioretention treatment of contaminated stormwater due to removing other
contaminants thereby, decreasing leachate toxicity but not completely removing toxicity.
However, because we are suspecting that metals could be added with treatment, we
would expect some toxic response from that as well.
They used frozen samples from their previous test of runoff leached through the
BSM plus gravel columns, they then compared the leachate from this research. They
tested the leachate with a zebrafish qPCR assay done on groups (25-30) of zebrafish
embryos at the age of 52 hpf (hours past fertilization) and measured for detox genes
(cyp1a) which would be increased with exposure to PAHs and other planar aromatic
hydrocarbons. They also tested for cardiac-specific (nppa, tbx5a) or cardiac-related (ilk)
genes which could be an indication of other contaminants as well, not only PAHs. An
upregulation would be seen with an increase in the gene due to the body producing more
of it to detoxify the body.
Table 7. Quantitative polymerase chain reaction (qPCR) assay of the filtered runoff with
and without plants and the control runoff of the uncontaminated water treated through the
BSM and BSM with gravel columns. All values have a base of 2, so any upregulation is
double the value indicated. The control water groups were leachate from this thesis
research (McIntyre et al., unpublished).
77
The results of this study matched what our study concluded. There was significant
upregulation in filtered highway runoff in the BSM and gravel column with plants for
cyp1a and nppa (Table 7) (McIntyre et al., unpublished). For the filtered highway runoff
with no plants, there was significant upregulation of the cyp1a, tbx5a and ilk genes. In
the leach water from this research, with the uncontaminated water going through the
columns, there was only a significant upregulation of the nppa gene in the BSM only
column. This could easily be explained by the fact that nppa is also sensitive to osmotic
changes, so different osmolarity of the filtered runoff compared to the control water could
explain that upregulation (McIntyre et al., unpublished). There was no upregulation in
the other genes for the tests with the clean water running through the bioretention
treatments, supporting no toxicity from the soil columns themselves. This indicates that
the effects observed in the treated runoff studies were from the stormwater coming
through the columns rather than the columns themselves generating enough
contamination to cause toxicity. This is the same conclusion that was drawn from this
thesis research and supports the findings.
4.4
ECOLOGICAL IMPLICATIONS
In the application of bioretention systems, there is no requirement to condition the
soils prior to use. As shown in this research, conditioning can drastically reduce the metal
concentrations leaching out. Even with no toxicity observed from the soils observed,
78
conditioning the soils in a controlled situation should still be conducted. Bioretention
systems are used as a means of removing contaminants from the stormwater. This
research found that soils have the potential to leach out high levels of metal
concentrations. This can add to stormwater pollution. This effect was observed when
uncontaminated water was flushed through the soil medium. Metals have the ability to
accumulate in an aquatic environment. Even though this study did not show a toxic effect
from the metals leached out of the soil, in other aquatic environments with other water
chemistry parameters they could.
Copper has been shown to be neurotoxic to freshwater fish in levels as low as
11.5 µg/L (McIntyre et al., 2014). Copper has also been shown to cause blocking of
olfactory sensory neurons in salmon in freshwater at concentrations as low as 2 µg/L
(Sandahl et al., 2007). The copper concentrations leaching in this thesis research were as
high as 60 µg/L at one point. While they did not remain that high, after leaching the
ending concentration was still around 5 µg/L. These levels are high enough to have a
potential toxicological response. With the accumulation potential, this amount of metal
leaching has the potential to create larger ecological concerns.
79
5.
CONCLUSION
This research found that bioretention soil medium leaches very high
concentrations of metals. This is most likely due to the compost that is present in the
mixture. The highest concentrations in the leachate were zinc and copper. This is also the
highest metal concentrations permitted in compost. As more water was rinsed through the
soil, the metal concentration decreased, implying that conditioning the soils can reduce
the metal load of the leachate. This research has shown that there was no toxic effect
from the leachate coming out of soils alone. This was attributed to calcium, magnesium,
pH and dissolved organic carbon providing a protection against the metal ions to the
freshwater fish studied. In an aquatic environment all these factors could change. This
study was not an accurate representation of changing dynamics in a freshwater system. It
was purely used as a comparative analysis of toxicity to stormwater runoff, versus
toxicity of bioretention system leachate.
Zinc and copper have been shown to adversely affect freshwater fish and are
considered contaminants of concern in stormwater pollution. Stormwater retention
structures are meant to be a means of removal of contaminants in the stormwater, but if
they are also adding to metal pollution, they may need to be studied further. Bioretention
systems have been shown to leach metals when uncontaminated water is applied, but
retain metals when contaminated stormwater runoff is applied. If used in situations where
a mixture of stormwater pollution is being treated, the overall effectiveness of
bioretention systems, according to this study as well as previous research, is a reduction
of pollutants.
80
Based on the results of this study, bioretention systems are a successful treatment
for stormwater pollution. Further studies should be conducted on the use of the composts
in the systems. The compost used in this study leached high concentrations of metals, it
may be worth studying, to examine other compost types and test if there is a difference in
metal leachate. This study also showed a drastic reduction in metal concentration in the
leachate due to a conditioning of the soils by flushing them with water. Based on this, it
seems that conditioning in a controlled manner, should be a practice more commonly
used prior to application of composts in bioretention systems.
81
82
References
Brandenberger, J. M., Crecelius, E. A., & Louchouarn, P. (2008). Historical Inputs and
Natural Recovery Rates for Heavy Metals and Organic Biomarkers in Puget Sound
during the 20th Century. Environmental Science & Technology, 42(18), 6786–6790.
Balades, J.-D., Legret, M. & Madiec, H. (1995). Porous pavements: pollution
management tools. Water Science and Technology. 32 (1), 49–56.
Bergman, H., & Dorward-King, E. (1996). Reassessment of Metal Criteria for Aquatic
Life Protection. SETAC Press.
Chapin, F. S., Matson, P., & Vitousek, P. (2011). Principles of Terrestrial Ecosystem
Ecology (2nd ed.). Springer.
Collins, K., Hunt, W., & Hathaway, J. (2008). Hydrologic Comparison of Four Types of
Permeable Pavement and Standard Asphalt in Eastern North Carolina. Journal of
Hydrologic Engineering, ASCE. 13 (12), 1146.
Das, K. C., & Kirkland, J. T. (2008). Quantification of water extractable contaminants
from food waste and biosolids blends at different stages of composting. Compost Science
& Utilization. 16(3), 200–206.
Department of Ecology. (2011, May). Toxics in surface Runoff to Puget Sound. State of
Washington Department of Ecology.
Dietz, M. E., & Clausen, J. C. (2005). A Field Evaluation of Rain Garden Flow and
Pollutant Treatment. Water, Air, and Soil Pollution, 167(1-4), 123–138.
EPA. (2009). Pervious Concrete Pavement. Retrieved from
http://cfpub.epa.gov/npdes/stormwater/menuofbmps/index.cfm?action=browse&Rbutton
=detail&bmp=137
EPA. (2012). Effectiveness of Low Impact Development: Proven LID Technologies Can
Work for Your Community. Retrieved February 21, 2015, from
http://water.epa.gov/polwaste/green/upload/bbfs5effectiveness.pdf
EPA. (2015). Growth and Water Resources. Watershed Academy Web. US EPA.
Retrieved February 28, 2015, from
http://cfpub.epa.gov/watertrain/moduleFrame.cfm?parent_object_id=133
EPA. (2015b). Low Impact Development (LID). Retrieved February 28, 2015, from
http://water.epa.gov/polwaste/green/
83
EPA. (n.d.) Technical Factsheet on: POLYCYCLIC AROMATIC HYDROCARBONS
(PAHs). EPA. http://www.epa.gov/ogwdw/pdfs/factsheets/soc/tech/pahs.pdf
Färm, C., & Waara, S. (2005). Treatment of stormwater using a detention pond and
constructed filters. Urban Water Journal. 2(1), 51–58.
Foran, J. A., Hites, R. A., Carpenter, D. O., Hamilton, M. C., Mathews-Amos, A., &
Schwager, S. J. (2004). A survey of metals in tissues of farmed Atlantic and wild Pacific
salmon. Environmental Toxicology and Chemistry. 23(9), 2108–2110.
Geronimo, F. K. F., Maniquiz-Redillas, M. C., Tobio, J. A. S., & Kim, L. H. (2014).
Treatment of suspended solids and heavy metals from urban stormwater runoff by a tree
box filter. Water Science & Technology. 69(12), 2460.
Good, J. F., O’Sullivan, A. D., Wicke, D., & Cochrane, T. A. (2012). Contaminant
removal and hydraulic conductivity of laboratory rain garden systems for stormwater
treatment. Water Science & Technology. 65(12), 2154.
Grebel, J. E., Mohanty, S. K., Torkelson, A. A., Boehm, A. B., Higgins, C. P., Maxwell,
R. M., … Sedlak, D. L. (2013). Engineered Infiltration Systems for Urban Stormwater
Reclamation. Environmental Engineering Science. 30(8), 437–454.
Hinman, C. (2009). Bioretention Soil Mix Review and Recommendations for Western
Washington. Puyallup Extension Station, Puyallup, Washington.
Hunt, W. F., Davis, A. P., & Traver, R. G. (2012). Meeting Hydrologic and Water
Quality Goals through Targeted Bioretention Design. Journal of Environmental
Engineering. 138(6), 698–707.
Incardona, J. P., Collier, T. K., & Scholz, N. L. (2004). Defects in cardiac function
precede morphological abnormalities in fish embryos exposed to polycyclic aromatic
hydrocarbons. Toxicology and Applied Pharmacology. 196(2), 191–205.
Istenič, D., Arias, C. A., Matamoros, V., Vollertsen, J., & Brix, H. (2011). Elimination
and accumulation of polycyclic aromatic hydrocarbons in urban stormwater wet detention
ponds. Water Science & Technology. 64(4), 818.
Joerger, S. (2008, May). Stormwater Regulation in Puget Sound. Stormwater. 36–46.
Kamo, M., & Nagai, T. (2008). An application of the biotic ligand model to predict the
toxic effects of metal mixtures. Environmental Toxicology and Chemistry. 27(7), 1479–
1487.
Landis, W. G. (2004). Introduction to environmental toxicology: impacts of chemicals
upon ecological systems (3rd ed.). Boca Raton: Lewis Publishers.
84
Lanno, R. (1998). Contaminated Soils: From Soil-Chemical Interactions to Ecosystem
Management. SETAC Press.
LeFevre, G. H., Hozalski, R. M., & Novak, P. J. (2012). The role of biodegradation in
limiting the accumulation of petroleum hydrocarbons in raingarden soils. Water
Research. 46(20), 6753-6762.
Li, L. Y. (2006). Retention Capacity and Environmental Mobility of Pb in Soils along
Highway Corridor. Water, Air, and Soil Pollution. 170(1-4), 211–227.
Linbo, T. L., Baldwin, D. H., McIntyre, J. K., & Scholz, N. L. (2009). Effects of water
hardness, alkalinity, and dissolved organic carbon on the toxicity of copper to the lateral
line of developing fish. Environmental Toxicology and Chemistry. 28(7), 1455–1461.
Lundstedt, S., White, P. A., Lemieux, C. L., Lynes, K. D., Lambert, I. B., Öberg, L., &
Tysklind, M. (2007). Sources, fate, and toxic hazards of oxygenated polycyclic aromatic
hydrocarbons (PAHs) at PAH-contaminated sites. AMBIO: A Journal of the Human
Environment, 36(6). 475–485.
Malins, D. C., Anderson, K. M., Stegeman, J. J., Jaruga, P., Green, V. M., Gilman, N. K.,
& Dizdaroglu, M. (2006). Biomarkers Signal Contaminant Effects on the Organs of
English Sole (Parophrys vetulus) from Puget Sound. Environmental Health Perspectives.
114(6), 823–829.
McIntyre, J. K., Baldwin, D. H., Beauchamp, D. A., & Scholz, N. L. (2012). Low-level
copper exposures increase visibility and vulnerability of juvenile coho salmon to
cutthroat trout predators. Ecological Applications. 22(5), 1460–1471.
McIntyre, J. K., Davis, J. W., Incardona, J. P., Stark, J. D., Anulacion, B. F., & Scholz,
N. L. (2014). Zebrafish and clean water technology: Assessing soil bioretention as a
protective treatment for toxic urban runoff. Science of The Total Environment. 500-501,
173–180.
McIntyre, J. K., Edmunds, R. C., Redig, M. G., Mudrock, E. M., Davis, J. W., Incardona,
J. P., Stark, J. D., & Scholz, N. L. (unpublished). Confirmation of stormwater
bioretention treatment effectiveness using molecular indicators of cardiovascular toxicity
in developing fish.
Meyer, J. S., & Adams, W. J. (2010). Relationship between biotic ligand model-based
water quality criteria and avoidance and olfactory responses to copper by fish.
Environmental Toxicology and Chemistry.
Newman, M. C., & Unger, M. A. (2003). Fundamentals of ecotoxicology (2nd ed.). Boca
Raton, FL: Lewis Publishers.
85
Norton, D., Serdar, D., Colton, J., Jack, R., & Lester, D. (2011, November). Control of
Toxic Chemicals in Puget Sound. Washington State Department of Ecology.
Ormond, T., P. E., Mundy, B., Weber, M., A.S.L.A., & Friedman, Z. (2010). LID Meets
Permaculture: Sustainable Stormwater Management in the Mountains of Western North
Carolina. ASCE.
Page, K., Harbottle, M. J., Cleall, P. J., & Hutchings, T. R. (2014). Heavy metal leaching
and environmental risk from the use of compost-like output as an energy crop growth
substrate. Science of The Total Environment. 487, 260–271.
Peraza, M. A., Ayala-Fierro, F., Barber, D. S., Casarez, E., & Rael, L. T. (1998). Effects
of Micronutrients on Metal Toxicity. Environmental Health Perspectives. 106, 203.
Petrell, R. J., & Gumulia, A. (2013). Saturated and unsaturated flow through sloped
compost filter beds of different particle sizes. Water Science & Technology. 67(11),
2406.
Quan, Q., Dong, L., Li, J. K., Shen, B., & Jin, C. X. (2014). Application of Bio-retention
Hydrologic Performance Tool for Urban Runoff Pollutants Removal. Nature
Environment & Pollution Technology. 13(3).
Rand, G. M. (1995). Fundamentals of aquatic toxicology: effects, environmental fate, and
risk assessment (2nd ed.). Washington, D.C: Taylor & Francis.
Reish, D. J., Oshida, P. S., Mearns, A. J., Ginn, T. C., & Buchanan, M. (2001). Effects of
Chemicals on Microorganisms. Water Environment Research. 73(Supplement 1), 1581–
1657.
Pitt, R., Field, R., Lalor, M., & Brown, M. (1995). Urban Stormwater Toxic Pollutants:
Assessment, Sources, and Treatability. Water Environment Federation. 67(3), 260–275.
Quan, Q., Dong, L., Li, J. K., Shen, B., & Jin, C. X. (2014). Application of Bio-retention
Hydrologic Performance Tool for Urban Runoff Pollutants Removal. Nature
Environment & Pollution Technology. 13(3).
Sandahl, J. F., Baldwin, D. H., Jenkins, J. J., & Scholz, N. L. (2007). A Sensory System
at the Interface between Urban Stormwater Runoff and Salmon Survival. Environmental
Science & Technology. 41(8), 2998–3004.
Schell, W. R., & Nevissi, A. (1977). Heavy metals from waste disposal in central Puget
Sound. Environmental Science & Technology. 11(9), 887–893.
Scholz, N. L., Myers, M. S., McCarthy, S. G., Labenia, J. S., McIntyre, J. K., Ylitalo, G.
M., & Collier, T. K. (2011). Recurrent Die-Offs of Adult Coho Salmon Returning to
Spawn in Puget Sound Lowland Urban Streams. PLoS ONE. 6(12), e28013.
86
Sébastian, C., Barraud, S., Gonzalez-Merchan, C., Perrodin, Y., & Visiedo, R. (2014).
Stormwater retention basin efficiency regarding micropollutant loads and ecotoxicity.
Water Science & Technology. 69(5), 974.
Seelsaen, N., McLaughlan, R., Moore, S., & Stuetz, R. M. (2007). Influence of compost
characteristics on heavy metal sorption from synthetic stormwater. Water Science &
Technology. 55(4), 219.
Sikora, F. J., & Moore, K. P. (2014). Soil test methods from the southeastern United
States. Southern Cooperative Series Bulletin, (419).
Teng, Z. & Sansalone, J. (2004) In-situ partial exfiltration of rainfall- runoff-II: particle
separation. J. Environ. Eng. ASCE 130.
Tucker, M. (2008). Compost in the Mix for Storm Water Management. BioCycle. 42–45.
Wium-Andersen, T., Nielsen, A. H., Hvitved-Jakobsen, T., & Vollertsen, J. (2011).
Heavy metals, PAHs and toxicity in stormwater wet detention ponds. Water Science &
Technology. 64(2), 503.
87
Appendix A: Metal Analysis Values
Raw data from the analytical analysis of total metal concentrations (µg/L) for the Instant
Ocean® Sea Salt solution though each treatment column, flush 1 through 8 (3 L through
24 L) and the final concentration measured after conditioning of the columns (61 L). The
last column is the reporting limit (RL). All values in bold are above the reporting limit.
Before Treatment
3L
0.5
Arsenic
0.1
Cadmium
0.5
Chromium
0.5
Copper
0.1
Lead
Nickel
0.6
0.2
Silver
4
Zinc
6L
0.5
0.1
0.5
2.1
0.1
0.5
0.2
27
9L
NA
NA
NA
NA
NA
NA
NA
NA
12 L
NA
NA
NA
NA
NA
NA
NA
NA
15 L
0.5
NA
0.6
1.7
0.1
0.5
NA
12
18 L
0.6
NA
0.5
1.9
0.1
0.5
NA
6
21 L
0.8
NA
0.5
1.2
0.1
0.5
NA
5
24 L
0.6
NA
0.5
0.9
0.1
0.5
NA
5
61 L
0.5
NA
0.5
13.4
0.1
0.5
NA
5
RL
0.5
0.1
0.5
0.5
0.1
0.5
0.2
4
BSM Treatment
3L
Arsenic
3.2
Cadmium
0.3
Chromium
3.8
Copper
22
Lead
4.4
Nickel
10.7
0.2
Silver
Zinc
19
6L 9L
7.4
7.3
0.2
0.2
11
13
31.6 31.7
8.1
9.6
24.1 25.4
0.2
0.2
33
35
12 L
6.3
0.2
13
24.8
7.3
20.8
0.2
29
15 L
4.6
NA
7.2
19.7
6.3
14.1
NA
22
18 L
5.3
NA
6.5
17.8
6.0
13
NA
20
21 L
4.3
NA
5.4
12.7
3.7
10
NA
18
24 L
3.7
NA
4.6
9.8
2.6
8.5
NA
12
61 L
2.5
NA
2.0
4.6
1.1
4.5
NA
7
RL
0.5
0.1
0.5
0.5
0.1
0.5
0.2
4
BSM + Gravel Treatment
3L 6L 9L
Arsenic
3
8.1
6
0.1
0.1
0.1
Cadmium
Chromium
2.8
8.3
10
Copper
52.5 60.1 54.4
Lead
3.5
5.0
6.4
Nickel
8.3 16.8 18.3
0.2
0.2
0.2
Silver
Zinc
11
20
23
12 L
5.9
0.1
11
54.3
6.9
19.6
0.2
26
15 L
4
NA
9.0
45
5.6
13.5
NA
18
18 L
5.3
NA
7.0
40.3
5.4
13.9
NA
17
21 L
4.7
NA
6.8
30.8
3.6
11.5
NA
15
24 L
3.8
NA
5.0
24.3
2.6
8.9
NA
11
61 L
1.8
NA
1.0
5.7
0.5
2.7
NA
4
RL
0.5
0.1
0.5
0.5
0.1
0.5
0.2
4
88
Gravel Treatment
3L
0.5
Arsenic
0.1
Cadmium
0.5
Chromium
Copper
4.1
0.1
Lead
Nickel
0.8
0.2
Silver
4
Zinc
6L
1.4
0.1
0.6
4.5
0.1
1.0
0.2
4
9L
1.6
0.1
1.0
5.8
0.1
0.8
0.2
4
12 L
1.1
0.1
0.5
3.2
0.1
0.6
0.2
4
15 L
0.8
NA
0.5
3.1
0.1
0.8
NA
4
18 L
1.8
NA
0.6
3.2
0.1
1.0
NA
4
21 L
4.4
NA
1.6
7.0
0.1
1.1
NA
4
24 L
7.2
NA
2.6
9.2
0.1
1.1
NA
4
61 L
1.1
NA
0.5
2.8
0.2
1.1
NA
4
RL
0.5
0.1
0.5
0.5
0.1
0.5
0.2
4
89
Appendix B: Substrate Concentrations by Metal Type
Metal concentrations measured in each of the different bioretention substrate columns as
a function of the number of liters of Instant Ocean® Sea Salt solution that had been
flushed through.
90
