I-42_7.27.2018_Summer 2018 FMD Sampling Report Memo with Appendices

ENVIRONMENTAL
Duke LAW anA POLICY CLINIC
Ryke Longest, Director • Michelle Nowlin, Supervising Attorney
Box 90360 • Durham, NC 27708-0360
Telephone: [phone] G9 • Toll Free: (888)-600-7274 -Fax: [phone]
To: Ryan Lavinder, Facilities Management Department
From: Christine Gerbode (Intern), Talia Sechley and Nancy Lauer (Fellows)
Re: Summer 2018 Campus Litter Sampling Results
Executive Summary
In conjunction with Duke University's Facilities Management Department (FMD), the Duke Environmental
Law and Policy Clinic ("the Clinic") conducted targeted in -stream litter surveys during the summer of 2018.
This investigation aimed to document how litter moves through stormwater drainage channels on and
around the Duke University West Campus, and focused on stormwater channels flowing into and out of
the Reclamation Pond (Rec Pond) and Storm Water Assessment and Management Park (SWAMP) areas.
In total, six stream segments were sampled according to methods described in a previously developed
litter survey protocol (appended to this document). Each segment was sampled to gain baseline data,
then sampled again following a rain event. These sampling events provided information regarding the
amount and composition of trash in Duke's waterways, as well as information regarding how those
metrics change in response to rainfall.
Surveyors, including Clinic staff and FMD summer interns, collected nearly 1,500 pieces of litterfrom these
streams between June 7t" and July 5t". Litter was categorized for analytical purposes as one of seven
"types": hard plastic, soft plastic, foam, metal, glass, sports equipment, and "other." Further analysis
focused on "floatable" (soft plastic and foam) and "sinkable" (glass and metal) litter types, reflecting likely
differences in transport mechanism.
Major observations from the study:
• Significant numbers of litter items of a variety of types were observed in sampled streams,
during both baseline and post -rain sampling.
• Soft plastic trash (including plastic bags, films, and wrappers) made up nearly half of all litter
observed cumulatively; medical and laboratory waste also made up a noteworthy volume of
observed litter in streams adjacent to the Rec Pond.
Major conclusions and recommendations:
• Floatable litter appears to be more abundant and more mobile in the Campus watersheds,
providing a more urgent target for pilot structural and non-structural litter controls.
• Potential areas worth future investigation include the laboratory and hospital complexes on
the north side of campus, as well as the irrigation intake on the Duke Golf Course.

Table of Contents
ExecutiveSummary.......................................................................................................................................1
Introduction..................................................................................................................................................
3
Contextof Project.....................................................................................................................................3
Goalsof Summer 2018 Sampling..............................................................................................................3
Site Selection — Rec Pond and SWAMP inflows/outflows........................................................................4
Reclamation Pond (Rec Pond)...............................................................................................................4
Storm Water Assessment and Management Park (SWAMP) Sites.......................................................4
Methodology.................................................................................................................................................
6
Buildingon the 2017 Litter Protocol.........................................................................................................6
Results...........................................................................................................................................................
7
Summary and Aggregate Comparisons.....................................................................................................7
Post -Rain Litter Accumulation..............................................................................................................8
A "Steady State" of Transect Trash?.....................................................................................................
9
Floatablevs. Sinkable Trash..................................................................................................................9
QualitativeSummary..........................................................................................................................12
Summary of Rec Pond Trends.................................................................................................................14
DataSummary.....................................................................................................................................15
Potential Sources of Litter...................................................................................................................15
Caveats and site -specific difficulties...................................................................................................17
Summaryof SWAMP Trends...................................................................................................................18
DataSummary.....................................................................................................................................19
Potential Sources of Litter...................................................................................................................19
Caveats and site -specific difficulties...................................................................................................20
Discussion....................................................................................................................................................
21
OverallTrends.........................................................................................................................................21
Foodbrands........................................................................................................................................
21
Link to Trash Flow Beyond the Duke Campus.....................................................................................21
Limitations of Protocol Implementation.................................................................................................21
Turbidity..............................................................................................................................................
21
PersonnelConsistency........................................................................................................................22
Limitationsof Study Design....................................................................................................................22
GeneralConclusions....................................................................................................................................23
KeyFindings............................................................................................................................................
23
Broader Implications of Trash Control on the Duke Campus.................................................................23
The Problem with Plastic Pollution.....................................................................................................
23
LocalRoots of a Global Problem.........................................................................................................24
Duke Can Lead the Way in Durham....................................................................................................24
Recommendations for Future Follow-Up...............................................................................................
25
Appendix A: Updated Litter Protocol Document
Appendix B: Rainfall Data
Appendix C: Sampling Data Tables
2

Introduction
Context of Project
During June and July of 2018, as part of ongoing collaboration between the Duke Environmental Law and
Policy Clinic ("DELPC" or "the Clinic") and Duke University's Facilities Management Department ("FMD")
interns from both groups conducted a series of surveys of litter in stream and stormwater channels on
and near Duke's West Campus and associated properties. This project complements and expands upon
previous Clinic efforts focused on characterizing the on -land sources and transport of marine plastics and
other anthropogenic debris; these efforts, in turn, support a long-term goal of developing practicable
means of assessing and minimizing the volume of urban trash transported to those marine environments
through action at the local and municipal level.
Following the Clinic's pilot implementation of a litter surveying protocol at various sites along Ellerbe
Creek during Summer 2017, the Clinic and FMD discussed mutual priorities for a follow-up project focused
specifically on areas of the Duke Campus and the Sandy Creek watershed. This memo is intended to
describe the goals, scope, and methodology of the resulting project, to report the results of this field
sampling, and to discuss major conclusions and potential opportunities for further study.
/ :fkP2v3
Caul-: �
RP2c 2"
RP2c 1
Reclamation
/ T- RP2A 3
/✓]r^i"�2A.,P26 2�
C SC1 1 RP26_1
t_SILT _x
Bryan
Center
� Fuqua
nrv0
A.
1 r.
f
S SWAMP
s
_Duke'Golf Retention
Course Pond`',.
" GL1 1
s ,
Figure 1: Yellow boxes represent
individual sampling
transects included in the 2018 summer study. Pink shading
has been applied to property owned by Duke.
Goals of Summer 2018 Sampling
The intent of this project was to provide insight into
trash movement through two major drainage areas of
Duke's property (Fig. 1), and the effectiveness of trash
control measures already in place. These areas — the
West Campus Reclamation Pond (Rec Pond) and Storm
Water Assessment and Management Park (SWAMP)' —
each contain retention ponds that are cleaned of litter
on a periodic basis; this project explored whether
inflows and outflows to these areas appear to carry
significantly different volumes of litter, thereby gauging
to what extent the in -pond litter removal is effective.
Sampling sites were selected on the basis of their status
as either inflow or outflow channels to these two areas,
as described in further detail below.
The study also aimed to provide insight into how rain
events influence the movement of trash in these
streams. Each site was sampled twice, once during
baseline conditions ("dry") and a second time following
at least one storm event with at least 0.1" of
precipitation ("post rain"). Because the Clinic's sampling
protoco12 involves the removal of trash from the
' A detailed description of the development and history of this project are available at
https://nicholas.duke.edu/wetland/swampi.htm
z See updated 2018 Litter Sampling Protocol document, attached as Appendix A.
3

sampled transects, this design allowed surveyors to roughly estimate the amount of "new" trash present
within a sampled transect after rainfall of known quantity occurred. These data provide insight into the
rate at which trash accumulates, and the mechanisms via which trash travels, through these areas of the
Sandy Creek watershed.
Site Selection — Rec Pond and SWAMP inflows/outflows
Reclamation Pond (Rec Pond)
am,
The Rec Pond is a landscaped
F-bsP�vl p
stormwater retention pond
located north of Towerview Rd.
�9r w
Reclamation between Circuit Dr. and Erwin
Pond Rd. The pond collects culverted
inflows from three stormwater
channels draining into a small
forebay (Fig. 2). Because these
inflow segments are relatively
short and few in number, it was
possible to survey all inflow
channels into the pond, as well
Figure 2. Overview of stormwater channels entering the Rec Pond, as mapped by the as the single outflow channel at
National Hydrography Dataset (NHD). Arrows indicate direction of flow. the western end of the pond. In
selecting this site, the team
postulated that a comparison of litter levels observed in streams entering the pond to litter levels
observed in the outflow stream would provide a sense of whether current litter control practices are
successfully capturing a significant portion of the litter moving through this system.
Existing trash removal at the Rec Pond:
Bimonthly litter collection is contracted by FMD for the Rec Pond forebay, which is designed to collect
litter and debris. Contractors have been directed to track the amount of trash cleared from this area in
terms of weight or number of bags needed to collect litter.
Storm Water Assessment and Management Park (SWAMP) Sites
The SWAMP complex is an area of
forested wetland southwest of the
intersection of Cameron Rd. and Duke
University Rd. This area has been
Duke Golf recontoured in phases in an effort to
Course restore a more natural hydrologic flow
SWAMP
Irrigation Detention regime to an existing degraded channel; a
Intake Pond number of wetland cells have been
Golf Course created as well. This complex is fed by a
1 Water Feature
� major channel draining from another
created wetland northeast of Cameron
Figure 3 Overview of stormwater and stream channels in the broader
SWAMP area and nearby Duke Golf Course, as mapped by NHD. Arrows Blvd., as well as many smaller tributaries
indicate the direction of flow in each stream. feeding this main stream from nearby
4

residential, commercial, and other property types (Fig. 3). Most of these inflow channels ultimately drain
into a small dammed reservoir (the "SWAMP pond"). The outflow from this dammed pond drains into a
stream running between fairways of the Duke Golf Course, where some water is diverted into an irrigation
intake feeding the campus water features and providing water for groundskeeping. Because the inputs to
the SWAMP system were more numerous and more complex than in the Rec Pond system, the team
elected to focus on only the main inflow and outflow channels to the SWAMP system, and selected survey
sites accordingly.
Existing trash removal procedures of the SWAMP and Duke Golf Course:
The SWAMP pond is also cleaned on a monthly basis by workers contracted by FMD. A Duke Golf Club
manager' communicated to the Clinic that trash consisting largely of plastic bottles is periodically
removed from four locations around the course where golf cart paths cross streams. A significant amount
of both organic debris and anthropogenic trash also gets caught in the irrigation intake, which could
influence what kinds and volumes of trash appear downstream in the golf course. No trash collection had
been done in the selected stream segment prior to this survey since at least March of 2018.4
' Personal communications with Sadler Stowe, June 19th and July 17th, 2018.
41d.
S

Methodology
Building on the 2017 Litter Protocol
Sampling was conducted according to the sampling protocol developed during Summer 2017.s Each
stream segment was divided into three 30-meter litter collection transects, interspersed with 30-meter
buffer zones. Segments to be sampled were selected based on a combination of factors, including
accessibility from a road or other crossing/easement. With a few exceptions described in later sections of
this report, all transects within each studied segment were sampled twice:
1. A baseline or "dry" sample was conducted following a period of at least four days without a rain
event of 0.1" or more, as recorded by the USGS rain gauge at Maureen Joy Charter School.'
2. A post -rain or "wet" sample was conducted no more than two days after a subsequent rain event.
Ideally, this sample should represent only the "new' trash transported to a given sampling site as
a result of the rain event, as baseline sampling should have removed the majority of trash initially
present.
Observed litter was characterized as one of seven major "types": hard plastic, soft plastic/films, glass,
metal, foams (primarily Styrofoam, but also other foams), sports equipment, and "other." These types
were further subdivided into additional categories in an effort to identify potential major sources of trash
for each stream; specific brands were recorded when visible. Observed litter was removed, with the
exception of heavy or strongly secured items (for example, vehicle tires, construction/engineering
materials like bricks and pipes, exposed geotextile fabrics, buried cables, etc.). Items not removed during
dry sampling were recorded, and were not counted during post -rain sampling.
Table 1. Summary of impairment classification
criteria, based on number of litter items observed.
(Average) Number
of Litter Items
Impairment
Classification
0-10
........ ......... ......... ..._.........
None/Very Light ......
11-25
Light
........ ......... ......... ..._.....
26 — 50
......
Moderate
........ ......... ......... ..._.....
51-100
High
101 - 200
............................................
Very High
>200
Based on these values, the Clinic assigned an
impairment classification to each transect, according to
the criteria presented in Table 1. After tabulating the
litter collected in each transect, the Clinic calculated the
average number of litter items per stream segment
according to the method prescribed in the Sampling
Protocol document. In addition, we compared how
impairment classifications differ between baseline and
post -rain samples, both for individual transects and for
segment averages.
5 See updated Litter Sampling Protocol document, attached as Appendix A.
' See rain gage data relevant to the reported sampling events in Appendix B, attached. USGS data for this site is
available at https://waterdata.usgs.gov/nwis/uv?cb 00045=on&format=gif stats&site no=355852078572045.
11

Results
Summary and Aggregate Comparisons
Across all sites, a total of 1,456 individual pieces of litter were collected (Table 2). All seven types of
trash were found in each sampled area. Overall, soft plastic was the most commonly found item,
accounting for nearly half of all trash collected (710 items); the remaining litter was composed of trash
categorized as (in decreasing order of count) "other," "glass," "metal," "Styrofoam," "hard plastic," and
"sports equipment" (the smallest category by far). Stream transects contained an average of 44.1 ± 8.1
(standard error) pieces of litter. Tables summarizing litter counts across all stream segments and
transects, before and after rain, are included as Appendix C.
The composition of trash within each study area (Rec Pond vs. SWAMP) differed substantially between
the two areas. Rec Pond -adjacent stream segments were dominated by soft plastic and other floatables
and contained only a single piece of sports equipment, while trash from the SWAMP -adjacent stream
segments was composed in larger part of glass and other sinkable trash. A table summarizing litter
composition and trends observed across all sampled litter, as well as pie charts showing the difference in
composition between areas, are presented below (Table 2; Fig. 4).
Table 2. Summary of all litter collected during Summer 2018 on -campus sampling.
Type of Litter Hard Plastic Soft Plastic Styrofoam Metal Glass Sports Equip.
Category Subtotal
...................................................................................................-........................
119
710
Average# /30m
...................................................................................................-........................
3.6
21.5
Std. Error
0.55
6.09
All Sampled Litter Composition
Other
Sports , AUL
Equi,
Glass
11%
Metal
8%
Sty
Hard
stic
ift Plastic
49%
121
........................-............................................._............................
124
155
29
3.7
........................-............................................._............................
3.8
4.7
0.9
0.68
1.01
1.77
0.55
r
198 1456
...................................................
6.0 44.1
1.32 8.08
Sports Rec Pond Litter Composition
Equipment Other _ Hard
0% 15% Plastic
Glasses 8%
3%
Metal
7% Styrofoam Soft Plastic
11% 56%
Other SWAMP Litter Composition
9% Sports Hard Plastic
Equipment_ 7%
8% ` Soft Plastic
25%
Glass 0 `Styrofoam
° Metal 2%
14%
Figure 4. (Left) Compositional breakdown of all sampled litter, in terms of the seven types noted by surveyors; (right) compositional breakdown
for all Rec Pond- and SWAMP -adjacent stream segments, as labeled.
7

A direct comparison of average transect litter composition and volume, split by sample area, also shows
that a larger number of items was found in the average Rec Pond -adjacent transect than in the average
SWAMP -adjacent transect (Fig. 5).
40.0
35.0
v
30.0
U U
.a
ai v 25.0
V)
c
o `��° 20.0
x�
0 15.0
b bn M
v 10.0
Q 5.0
0.0
■ Rec Pond Segments
■ SWAMP segments
Hard Plastic Soft Plastic
Styrofoam Metal
Glass
Sports
Type of litter
Equipment
Figure 5. Average number of each type of litter observed across all sampled
30-m transects.
i=
Other
Post -Rain Litter Accumulation
One focus of this investigation was the rate at which litter accumulates within the sampled stretches of
the stream. Due to resource limitations, this study was not designed in a manner that allowed us to
calculate detailed accumulation rates. However, by comparing the volume of "new," post -rain trash
collected after a given transect had been cleared during baseline sampling, some trends become visible
(Fig. 6).
40
Q 35 Dry Sample Average ■ Wet Sample Average
Ln
v +�
30
U U
v 25
Q c
o E 20
E 15
°1
en 0
� r^ 10 T T T
> 5 1 11 1
Q 0 z z i MTN i I
Hard Plastic Soft Plastic Styrofoam Metal Glass Sports Other
Type of Litter Equipment
Figure 6. Average transect litter composition for baseline ('dry) and post -rain ("wet") samples.
Specifically, in all transects, large amounts of new trash accumulated in the several days between baseline
and post -rain sampling. A comparison of the average composition of baseline samples and of post -rain
samples illustrates that while less metal, glass, and sports equipment were observed in post -rain samples,
the amount of Styrofoam and soft plastic was generally comparable (Fig. 6). The implications of this trend
are discussed further in the section below.
M

A "Steady State" of Transect Trash?
Because trash varies widely in its material composition, shape, and density, certain types of trash may
travel and accumulate at different rates in different areas of a watershed, and may be transported through
the watershed by different mechanisms. Glass fragments, for example, were often found incorporated
into sand bars and rocky areas, seemingly subject to classical riverine transport forces, while plastic films
tended to float, and to snag on vegetation or other larger debris. In light of this strong role that a transect's
specific morphology appears to play in allowing some types of litter to accumulate preferentially over
time, it is reasonable to expect that each transect might reach a sort of "steady state" carrying capacity
of typical trash content (i.e., a "normal" maximum volume and composition of trash present in the
transect, depending on the transect's propensity for catching certain types of litter, given a continuous
input of all litter types from upstream). While this "steady state" model may not apply equally to all trash
types (given the diversity of items falling under certain trash categories, such as "other" or "sports
equipment"), it is a valuable concept for discussing the movement and accumulation of more
homogeneous categories of litter, as described below.
Floatable vs. Sinkable Trash
The Clinic analyzed the percentage of litter observed in each transect that fell into two categories related
to their transport potential. All items classified as "soft plastic" or "foams" were group together as
"floatable" items; "sinkable" items are all items classified as "glass" or "metal." Because items described
as "sports equipment," "hard plastic," and "other" vary widely in terms of their transport potential, these
categories of litter were excluded from the floatable/sinkable analysis.
With the exception of segment 11132A, which contained an unusually dense area of soft plastic litter
trapped behind a log jam, the percentage of trash collected at each segment classified as "floatable" was
consistently higher during post -rain sampling than during baseline sampling (Fig. 7).
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
■ Baseline Floatable % Post -Rain Floatable %
Figure 7. Percentage of average litter content classified as 'floatable,"showing change between baseline and post -rain samples
for each stream segment. Note that only RP2A shows a decrease in floatable percentage.
01

The percentage of "sinkable" trash also decreased at the majority of sampled sites between baseline and
post -rain sampling (Fig. 8).
70%
60%
50%
40%
30%
20%
10%
0%
Dry Wet Dry Wet Dry Wet Dry Wet
SWMP1 GC1 RP2A RP2B
■ Baseline Sinkable Post -Rain Sinkable
[---*
F-I.
Dry Wet
Dry Wet
RP2C
C_SC1
Figure 8. Percentage of average litter content classified as "sinkable," showing change between baseline and post -rain samples
for each stream segment. Note that only RP2A and RP2B show a minor increase in sinkable percentage.
These two trends are displayed in terms of percentage change between sampling events for each stream
segment in the graph below; as seen in the case of segment RP213, an increase in floatable trash proportion
does not necessarily imply a decrease in sinkable trash proportion.'
20%
15%
10%
5%
0%
-5 %
-10%
-15%
-20%
-25%
-30%
■ Change in Floatable % ■ Change in Sinkable %
SWMP1 GC1 RP2A RP2B RP2C C_SC1
Stream Segment
Figure 9. Change in percentages of floatable and sinkable litter between baseline and post -rain samples for each segment.
' Future groups utilizing this survey methodology might consider analyzing all individual litter items as 'likely
floatable' or 'likely sinkable' during the survey process, to improve the granularity of this kind of analysis.
10

The implication of these trends in floatable and sinkable composition are that floatable trash types may
be entering the watershed in greater abundance, and/or moving more easily through stream systems,
allowing each transect to return to its normal "steady state" carrying capacity of soft plastics and
Styrofoam at a faster rate than sinkable items are replenished. In terms of absolute counts of floatable
litter, 4 out of 6 sampled sites had at least half as much floatable trash observed during post -rain sampling
as was observed and collected during baseline sampling (Fig. 10), suggesting that this carrying capacity is
quickly reached even after only a few days. This finding suggests floatable items as a more urgent target
for future trash control measures.
Change in Floatables Count Between Baseline and Post
Rain Sampliniz
M
M 140%
121%
-120%
� a
O E 100%
v <n 78% 79%
-a c) 80%
a 60% 50%
00 40%
40%
c
i 20%
0 13%
0% -
SWMP1 GC1 RP2A RP2B RP2C C_SC1
Figure 10. Percentage change in counted floatable items between baseline and post -rain sampling.
11

Qualitative Summary
A wide variety of litter types were found across the surveyed transects. Major subcategories are
summarized in Figure 11.
Tires q 5 Pieces
`t of lathing
Ceramic 7
fragments
Bricks 1 5
Lab-styLe
gloves 44
Medical./Lab
textiles 3 5
Drumstick 1
Glass
bottles
136
Fragments
of glass
Aluminum L
cans V
Fragments of
aluminum foil 19
Golf
Tricycle 1 23 balls
Shoe
Field
ft
2 hock
balls
Litter
Collected
(1,456 items) ..
Other metal �7 16 Heavy
L equipment
items fragments 1114
10 Fragments of
packing
materials
Eyedropper
83
Fragments of
styrofoa m
19 Styrofoam
cups
Plastic
bottles
Plastic
2-6 fragments
20 straws and
utensils
1 Plastic cups
and dishes
Lids and
caps
r� Cigarette
J butts
169 Plastic
bags
-'} Pieces
4L of tape/
flagging
6
Cigarillo)
wrappers
87
'Other food
wrappers
42Chip/candy
wrappers
321
Fragments of
plastic film
Figure 11. Graphic display of major litter types and number of each found during Summer 2018 sampling. Noteworthy sub -types
of each litter category are displayed around primary types; these secondary totals may not add up to the full total for each type
category.
While not all items collected displayed identifiable brands, the brands of those that did were recorded.
Table 3 summarizes the majority of observed brands by informal item type. Those brands with a strong
known or suspected association with a particular location on the Duke Campus are highlighted in bold.
12

Table 3. Summary of major brand types and specific brands identified during Summer 2018 sampling. Names highlighted in bold indicate
brands with a known or suspected connection to a specific on -campus location.
Budweiser (and Light), Natural Ice
Coca-Cola (or Diet Coke),
(and Light), Busch Beer, Founder's
Mountain Dew, Big K Soda (Black
Beer, St. Ives, Olde English, Stroh's
Cherry), Dr. Pepper, Sprite, AMP
•
Beer, Keystone (and Light), Miller,
Energy (PepsiCo), StarbucksDoubleshot,
Schlitz
Sunkist, Pepsi,
Member's Mark, Welch's Soda
Reese's Peanut Butter Cup, Twix,
Flavor Ice, Milky Way, Rice Krispie
Biscuitville, Cookout, Dominoes,
• Treat, Snickers, Nabisco, Butterfinger,
• • • Popeye's, Chik-Fil-A, McDonald's,
Tictac, Fun Dip, Bob's Sweet Stripes,
Subway, Red Mango
Hershe 's Bar, Creamsicle, Skittles
Slim Jim, Nutri-Grain, Go-gurt, Kraft,
Smartfood, Nabisco Graham
Titleist, Nike, Calloway,
• • • • Crackers, Frito Lay Sunflower Seeds,
• • Bridgestone, TaylorMade, Strata,
Quaker Oats, Nature Valley Granola
Maverick
Bar, Lance Peanut Butter Crackers
Doritos, Cheetos, Sun Chips, Lay
• -
Gatorade, Capri Sun, Kool-aids,
Funyuns
Jammer, Nestle (Water), Kirkland
, H2O, Aquafina, Disani, Perrier
Febreeze Air Freshener, Spic-n-
• -
Span Floor Cleaner,
• - - - Kroger, Sam's Club, North Carolina
Westinghouse Security
• - - Education Lottery, Circle K
Electronics, EcoLab, Pyramex,
see 0 Vita Hume Potting Soil/Peat
The fast food brands highlighted in the table above are those with known on -campus locations, either in
the Duke Medical Center or Bryan Center. Titleist golf balls are sold in the pro shop at the Duke Golf Club,
according to the most recent update of the shop's website. The Vita Hume potting soil bag was collected
in segment RP2B, directly downstream of the campus greenhouse complex, and is suspected to have come
from this area.
13

Summary of Rec Pond Trends
The map below shows the results of baseline ("dry") and post -rain ("wet") sampling for each transect in
the Rec Pond study area (Fig. 12). Each transect is represented as a box approximately 30m in length,
oriented in the direction of the mapped stream channel. Transect boxes are filled with a color
representing the impairment classifications determined by baseline sampling; each box is also outlined in
a color representing the impairment classifications determined by the results of post -rain sampling. Note
that transects C_SC1_3 (in the outflow segment) and RP2A_1 (in the central inflow segment) were only
sampled post -rain, as discussed in more detail below. Consequently, these transects likely contained trash
during wet sampling that would have been removed during dry baseline sampling, had it occurred.
Results of Dry Sampling
Light (11-25)
Moderate (26-50)
High (51-100)
_ Extremely High (250+)
Results of Post -Rain Sampling
Very Light (0-10)
Light (11-25)
Moderate (26-50)
High (51-100)
Very High (101+)
Sandy Creek & Tribs
2r
Reclamation
Pond
C SC1 1
C_SC 1 _2
C SC1 3
RP2c 3
r RP2c_2
i
�i RP2c —1
RP2A_
RP2A 2
F�P2A 1 —
RP2AB
RP2B_2
RP2B_1
13
Figure 12. Summary map of baseline and post -rain sampling results in the Rec Pond area. Colored fill represents baseline
impairment classification, while the colored outline represents post -rain impairment results.
Key Observations:
0
• Inflow stream segments had significantly more trash present than the outflow stream during
both baseline and post -rain sampling.
• Post -rain litter levels were similar or higher than baseline levels within some individual upstream
transects, but were generally lower on average within stream segments overall, and lower at all
previously sampled outflow sites. This suggests a comparatively slow accumulation of litter in
14

these outflow segments, which might in turn suggest that trash removal in the Rec Pond forebay
is capturing a significant amount of trash. However, it was not possible within this survey design
to distinguish between trash originating from the Rec Pond itself and trash originating via direct
runoff from Towerview Rd., complicating the previous conclusion.
Data Summary
Table 4 summarizes the average per-transect litter types and impairment classifications for each
sampled segment, for both baseline and post -rain sampling.
Table 4. Summary of average litter findings for each sampled Rec Pond stream segment, before and after rain.
Segment Sample Hard Soft Styrofoam Metal Glass Sports Other Total Floatable Sinkable Segment
Plastic Plastic Equipment Impairment
Dry
6.5
119.5
3.0
6.5
0.0
0.5
21.0
157.0
78%
4%
RP2A
Wet
6.7
51.3
4.7
2.7
1.7
0.0
15.0
82.0
68%
5%
High
Dry
3.3
15.3
6.0
3.0
2.0
0.0
10.7
40.3
53%
12%
Moderate
RP2B
Wet
2.0
12.3
6.7
3.7
2.3
0.0
3.7
30.7
62%
20%
Moderate
Dry
6.7
20.3
8.0
6.0
2.0
0.0
7.0
50.0
57%
16%
High
RP2C
Wet
2.7
15.3
6.7
2.7
0.3
0.0
2.7
30.3
73%
10%
Moderate
C_SC1
Dry
3.5
6.5
2.0
1.5
1.0
0.0
2.0
16.5
52%
15%
Light
(outflow)
Wet
3.0
7.3
3.0
1.3
0.7
0.0
1.0
16.3
63%
12%
Light
With the exception of segment RP2A-which contained a natural dam created by fallen woody debris that
had collected hundreds of items of litter before baseline sampling occurred -the proportion of floatable
trash was higher in every segment during post -rain sampling. This supports the notion that floatable items
may be more abundant and/or mobile than sinkable items in the watershed, and may therefore be
replenished in a given transect more quickly.
Potential Sources of Litter
Campus Laboratories and Hospitals
Figure 13. Map of stormwater drains (purple) draining into the top of segment
RP2A, as shown in the Geocortex mapping system.
The site where the greatest density
of litter was observed was at the top
of stream segment RP2A, where a
log jam caused -100 pieces of litter
to accumulate within five days of
dry sampling (after approximately
1.25" inches of rain). A significant
portion of the litter observed in the
stream segments flowing west
beneath Circuit Dr. consisted of
laboratory textiles, disposable
gloves, disposable shoe covers,
surgical masks, and other
disposable protective equipment
15

likely utilized in a laboratory or hospital setting. There are numerous natural science buildings immediately
east and north of segments RP2A and RP213. Three culverts (two large and one small) feed into the stream
at the top of RP2A (Fig. 13).
One of the two main culverts
directing stormwater into the
stream contained significantly
more trash and debris than the
other (Fig. 14). This difference
may be due to the relatively high
lip blocking the lower edge of the
culvert on the right, which may
retain water and debris in the
pipe. It is therefore unclear
whether one of these pipes is fed
by a network of stormwater
drains that take in significantly
Figure 14. Photo of the two major culverts draining into the top of segment RP2A. more trash than the other, or
Note the presence of debris in the culvert on the right. whether the culvert on the right
merely retains more of the trash
that moves through it. Regardless, a significant volume of trash appears to be transported through one or
both of these culverts.
sC �•- According to the stormwater pipe maps
Hudson E�
ea�!i Building available on FMD, s GeoCortex mapping
Duke Children'
MSRB Fbspildl Heahh system, these two larger culverts connect to
Can rIII ancillary
Center, Ifni r Bldg
1 s IMion Vi-riuin stormwater pipes draining a number of
Fac i lit
RP2C e�rch physical science and laboratory buildings,
ParA I puke Medicine
��n arch naline
Pavilion
;r h H. o„�
Kb rLk
Bldg uikli including but not limited to the French
/ Gancer Science Center, the Physics building, the
QenGe
e Elect vine men ` � tau eC1 Levine Science Research Center, the
Laser La Researc
Center Engineering School buildings, Environment
ka
TUN ��
RP2A B' Hall, the Free Electron Laser Lab, and a
g„Ln sksgld{i IL Cente number of clinical and research complexes
RP2Bf e C across Research Dr. in the Duke Medical
r
ica
9 P-r"In Center (Fig. 15). Further discussion with
Idc park � E.
Garage t� �� Ryan Lavinder at FMD indicated that these
Bryan Studen u 1 11,` u, drainage channels connect all the way across
Erwin Rd. into the broader Medical Center
Figure 15. Geocortex map of stormwater drains (purple) throughout the
laboratory and hospital complexes on the north side of West Campus, complex. The Clinic conducted a brief
which drain into stream segments RP2A and RP28. preliminary excursion in late June to
examine some of the roll -off and waste
disposal sites associated with the buildings nearest to the sampled stream segments, in an effort to see if
any obvious breaches of normal trash disposal practices were occurring, but no obvious single sources
were discovered. Further study of this extensive complex and its trash disposal practices by a future intern
16

might illuminate potential areas or processes in which additional structural or non-structural trash
controls could be valuable in reducing litter.
Roadway Sources
Samplers observed that the transect C_SC1_3
appears to receive trash not only from the
upstream transects of this segment, but also via
direct runoff from Erwin Rd. This direct trash
transport from runoff appears to be partially
prevented on the opposite side of the Erwin
culvert by chain link fencing. However, in all
transects sampled, it is likely that some portion of
sampled trash was transported by direct runoff
from areas adjacent to —and not upstream of— the
sampling site in question.
Caveats and site -specific difficulties
The short length of the stream channels studied in
these areas was not wholly conducive to as -is
implementation of the Clinic's three -segment
sampling protocol. For example, the stream
segments labeled RP2A and RP2B converge just
Figure 16. A view of a mass of organic debris mixed with plastic before the location where a third transect would
trash visible at the edge of the Erwin Rd. bridge%ulvert have been sampled for segment RP2B (note the
spanning segment C SC1. This debris appears to wash down the transect segment labeled RP2AB). At the same
slope from the roadway.
time, this close physical arrangement does not
allow for a full 30-meter buffer zone between this
final "joined" transect and what would otherwise have been the final transect of segment RP2A. In part
due to questions related to the propriety of taking a sample using this odd geometry, an initial dry sample
was not taken for RP2A_3 (see Fig. 12). For analytical purposes, RP2AB was considered the third transect
of segment RP2B.
Similarly, the stream segment flowing out of the Rec Pond (C_SC1) was not long enough to place three
transects of appropriate length with full 30 m buffers. The final transect in this segment (C_SC1_3) was
added on the outflow of the stream just beyond a box culvert running under Erwin Rd., but was not initially
sampled during baseline sampling for the rest of the segment due to perceived hydrological barriers
between the sites (specifically, little water flows between stream transects through the box culvert during
dry conditions, due to the fact that the culvert floor is elevated above the stream bed). However, a sample
was taken at this transect during post -rain sampling; while this transect was included in calculating the
averages for the C_SC1 wet sample, it represents a baseline sample perhaps more comparable to the dry
sample for this site.
17

Summary of SWAMP Trends
The map below shows the results of baseline ("dry") and post -rain ("wet") sampling for each transect in
the Rec Pond study area (Fig. 17). As in Figure 12, each transect box is filled with a color representing the
results of dry sampling; each box is also outlined in a color representing the results of wet sampling.
Transect GC1_1 was not sampled post -rain due to site access issued described below.
Results of Dry Sampling
I Light (11-25)
F_ Moderate (2
High (51-10
_ Extremely F
Results of Post
Very Light (f
Light (11-25
Moderate (2
High (51-10
Very High (,
Sandy Cree
GC1 1
GC 12
�- GC1
VMP1 1
\i
1► I M
Figure 17. Summary map of baseline and post -rain sampling results in the SWAMP area. Colored fill represents baseline
impairment classification, while the colored outline represents post -rain impairment results.
Key Observations:
• More litter was present at the downstream segment (GC1) than at the upstream segment
(SWMP1) during baseline sampling; however, more litter was present at SWMP1 than at GC1
during post -rain sampling.
• Observed litter in these transects may be skewed by a combination of contracted and incidental
litter collection occurring upstream of each site — at a created wetland complex across Cameron
Blvd., and at the Golf Course irrigation intake.
• Litter was likely undercounted during post -rain sampling due to large increases in turbidity at
these sites following rain.
18

Data Summary
Table 5 summarizes the average per-transect litter composition and overall impairment classifications for
each sampled segment, for both baseline and post -rain sampling.
Table 5. Summary of average litter findings for each sampled SWAMP stream segment, before and after rain.
Segment Sample Hard Soft Foam Metal Glass Sports Other Total Floatable Sinkable Segment
Plastic Plastic Equipment Impairment
Dry 3.3 7.3 0.7 2.3 18.0 0.7 2.7 35.0 47% 44% Moderate
SWMP1
Wet 1.3 6.3 0.0 0.3 6.0
0.0 2.0 16.0 51% 37% Light
GC1 Dry 3.7 11.7 1.3 13.3 17.3 7.7 5.0 60.0 27% 45% Heavy
(outflow) Wet 0.7 1.7 0.0 0.7 1.0 1.0 1.3 6.3 37% 26% Very Light
Potential Sources of Litter
SWAMP Area
Litter in this area consisted of relatively little floatable material, and is also clearly the site of extensive
scientific study (likely by faculty and students involved with the Duke Wetland Center). Some of the items
found in this stretch of stream appeared to be defunct or broken scientific equipment (pipes, wooden
platforms, flagging tape, etc.) It is also unclear whether any minor litter removal is carried out by students
working in the area that might have skewed this segment's sampling results.
Golf Course
Figure 18. Old steel Budweiser can, likely dating to the 1980s or earlier
based on out -of -production pull -tab style opening and can graphics.
The Golf Course site (GC1) presented a
minor puzzle in that a large percentage of
the trash collected at this site appeared
to be extremely old (and mostly sinkable)
items. It istherefore not likelyto be highly
representative of current patterns of
trash transport and collection. For
example, a significant number of the beer
cans found wedged among the rip -rap at
site GC1 _1 during dry sampling appeared
to be steel cans dating to the 1980s or
even 1970s, based on the out -of -
production pull -tab style openings and
similarly dated graphics (Fig. 18). These
heavy cans were mostly found wedged
among the rip -rap; while the periodic
cleaning of the course has included this
relatively difficult -to -access area of the
19

stream,$ these cans may have been passed over many times due to their obscured position buried among
sharp rocks, in many cases partially or fully submerged beneath the water.
Litter items observed in this segment that appear to better represent modern trash flows included
significant numbers of golf balls, other sports equipment (including what appeared to be the wheel and
windshield of a golf cart), sports drink bottles, glass fragments, and pieces of equipment or debris
evidently related to groundskeeping and civil engineering (bricks, eroded concrete pipe segments, etc.).
Caveats and site -specific difficulties
While the proportion of floatable and plastic trash recovered during sampling was lower in this area than
in the Rec Pond -adjacent stream segments, the Duke Golf Course manager reports9 that significant
amounts of floatable trash (including hard -plastic drink bottles) are removed from the water crossings
during periodic cleaning. Furthermore, he reports that significant amounts of floatable trash and organic
debris are found in the irrigation intake immediately upstream of the sampled Golf Course segment. This
may indicate that the composition of trash observed in the golf course segment was skewed by the
disproportionate removal of floatable trash and plastics upstream.
Access to the sites on the golf course proved more difficult than anticipated. While the vicinity of the
sample transects was easily reached by golf cart, the channel itself is difficult to enter and exit. Site access
for GC1_1 was gained during dry sampling by climbing down a rip -rap -covered slope, and much of the
transect itself is underlain by rip -rap. Due to the extremely sharp nature of these rocks, and given that the
knee-high rubber boots worn by the sampling team were not appropriate for walking on rough terrain,
this area became unsafe to access after water levels rose just prior to wet sampling. Consequently, this
transect was not sampled a second time. Other transects at the site had to be reached by climbing down
into the highly eroded stream channel from elsewhere along its steep banks, or in some cases by climbing
out using fallen trees; pockets of greater depth prevented the teams from simply walking along the
streambed from one end of the site to the other.
Both the golf course segment and upstream SWAMP channel segment also became largely opaque with
sediment in the wake of the rain event that triggered post -rain sampling. As a result, it was largely
impossible to see the bed of the stream channels for these post -rain samples, whereas the same channels
were extremely clear during baseline sampling. This discrepancy likely led to an undercount of litter
present in these streams after rain occurred.
s Personal comm. with Sadler Stowe, supra 3.
s Personal comm. with Sadler Stowe, supra 3.
20

Discussion
Overall Trends
Key Observations:
• The percentage of floatable trash observed at all but one site was higher during post -rain
sampling, which fits with the assumption that floatable trash is transported in greater
abundance or with greater ease through the watershed.
• On average, the SWAMP sites had a much lower percentage of floatable trash than Rec Pond
sites; this could be because water moves through this system at a higher volume, creating a
higher -energy deposition regime and therefore a proportionately higher rate of deposition of
heavy, sinkable items. It may also be related to trash collection by the irrigation intake on the
Golf course or upstream at the created wetland across Cameron Blvd.
Food brands
Most of the identified brands and food packaging did not directly link observed trash to a known on -
campus source — for example, few of the fast food or convenience store brands identified in these
watersheds seemed to be located in the immediate vicinity near which they were found. A small amount
may have originated from restaurants at Bryan Center or within the hospital complex (e.g., Red Mango
and Subway -brand litter) or could have plausibly come from some of the campus cafes or vending
machines that sell these brands (e.g., candy and pre -packaged snacks available at Twinnie's).10 This lack
of immediate geographic correlation suggests that much of this food -related trash may be brought to
campus by students, employees, or other visitors to Duke and its hospital system, rather than generated
exclusively from items purchased on -site. It could also reflect litter from the areas immediately off campus
transported by unchannelized stormwater runoff.
Link to Trash Flow Beyond the Duke Campus
Beyond what has been described in this report, additional litter samples taken at City of Durham water
quality sampling sites downstream of the Duke Golf Course show that golf balls are being carried as far as
Sandy Creek Park across Highway 15/501.11 In light of the present study's findings that floatable trash
appears to be more abundant and/or more mobile in these stream systems, it is reasonable to assume
that floatable trash is leaving the Duke campus as well.
Limitations of Protocol Implementation
A number of limitations in design or execution of the sampling protocol became evident over the course
of this pilot study, as described below.
Turbidity
As mentioned in the discussion of the SWAMP sites, post -rain turbidity presented a sampling challenge at
some sites, likely leading to an undercount of trash during these sampling events. A potential solution to
11 While a number of the fast food brands that appeared on campus do have locations on Hillandale Rd. near the
junction with 15-501, these locations are largely in the Ellerbe Creek watershed (i.e., across Highway 147 from West
Campus), and are therefore considered unlikely to have been transported to the sampled sites by stormwater runoff.
11 More information about these additional City -related sampling sites and results can be provided upon request.
21

this issue during future sampling might be to wait for a longer period following rainfall to ensure that
water levels and clarity have returned to a more baseline level prior to taking the second sample.
Personnel Consistency
Over the course of the summer, sampling was conducted by five different samplers, in groups as small as
two and as large as four. Because the pre- and post -rain samples for some transects were conducted by
different teams, it is possible that variation in how individual team members reported or recorded litter,
or other personal factors related to human error (e.g., eyesight, attention to detail, height, etc.), may have
skewed sampling results in unpredictable ways.
Limitations of Study Design
Stream morphology: As noted above, litter presence/absence in a given stream transect or segment is
highly dependent on local factors, and while a high amount of observed litter does likely indicate that
significant amounts of trash are transported to/through the sampled site by natural processes, a low
amount of observed trash does not necessarily mean that only minimal trash passed through a site. In
future sampling endeavors, it might be worth exploring means of incorporating stream shape and
substrate types (such as rocks or vegetation that are likely to catch or snag certain types of trash) into the
data collection process. Alternatively, future surveyors interested in gathering data on the amount of litter
passing through a given transect might explore options related to outfall netting or other in -stream
collection devices.
Seasonal Differences: As all sampling took place during the academic year summer, the majority of the
Duke student body was absent from campus and the surrounding neighborhoods. This could mean our
data reflects skewed, seasonal litter generation patterns, likely resulting in underreporting of certain types
of trash that are more abundant during the academic year.
Completeness of Inflow/Outflow Assessment: Due to a lack of personnel and resources, the Clinic had to
select a limited subset of the tributaries flowing into and out of the SWAMP for sampling over the course
of the summer. A more complete picture of how litter moves through this system could be acquired
through sampling of more of these streams.
22

General Conclusions
Key Findings
• A significant volume of litter appears to be moved by stormwater in the streams feeding the water
features on the Duke Campus.
• Floatable items appear to make up the largest volume of trash items moving into the Rec Pond
drainage system, and based on anecdotal evidence12 a significant volume of litter in the SWAMP
system as well. This suggests that these items might warrant special targeting by future
investigations and potential structural or non-structural trash controls.
• Based on the relative absence of trash in the Rec Pond outflow stream, trash collection in the Rec
Pond does appear to reduce the amount of trash that exits the system into the broader Sandy
Creek watershed. However, current measures do not capture 100% of trash, and significant
amounts of trash still sit in the waterways throughout campus. This conclusion should also be
taken within the context of the study design limitations discussed above.
• The West Campus science buildings and the nearby hospital complex appear to generate
significant amounts of trash that end up in campus streams; investigating trash disposal
procedures and potential litter sources in these areas of campus might provide valuable insight
on opportunities for preventing litter from entering waterways.
• Anecdotal evidence from the Duke Golf Course manager suggests that plastic bottles also move
in significant numbers into the SWAMP complex, and downstream in the Golf Course irrigation
intake.
Broader Implications of Trash Control on the Duke Campus
The Problem with Plastic Pollution
Widespread plastic pollution has gained significant attention in the past few years, from local awareness
and cleanup campaigns13 to the discussion of international treaties." The fundamental reasons for
concern about plastics are their longevity and their ubiquity: plastics do not fully decompose or break
down on a human timescale, and they are disposed of in enormous quantities around the globe every
day. Once plastic trash enters the environment, it may remain there for hundreds or thousands of years;
meanwhile, more plastic is continually produced to replace these discarded items.
Perhaps the most dramatic visible sign of the global plastic problem (beyond the litter typically present in
local waterways and beaches) is the relatively recent15 discovery of massive caches of plastics in areas far
from normal human activity, and far from the points of plastic use or production: the central circulatory
gyres of major ocean basins. The Great Pacific Garbage Patch, for example, is a floating debris zone located
in the open sea between Hawaii and California. This zone, while varying widely in density, is now estimated
" Personal comm. with Sadler Stowe, supra 3.
13 E.g., work done locally by organizations like Don't Waste Durham.
14 E.g., the Oceans Plastics Charter supported this year by the majority of G7 nations and the EU.
11 Captain Charles Moore of the environmental organization Algalita appears to be one of the first people to report
an alarming density of plastic debris in the Northern Pacific gyre, which he encountered during a boat trip in 1997.
More information available on the organization's history page, available at http://www.aIgaIita.org/about-
algalita/history .
23

to be 1.6 million square miles in area.16 The patch is estimated17 to contain well over a trillion pieces of
plastic of various sizes, weighing on the order of 80 to 100 thousand tons. Even in far lower concentrations
in near -shore and coastal zones, plastics and other debris can have devastating effects on marine and
coastal fauna including birds, whales, seals, and turtles, which may mistake plastics for food or become
fatally entangled in trash.18
However, the international community has increasingly begun to investigate the understudied dangers of
microplastics as well.19 These tiny fragments of plastic (typically considered plastic particles of 5mm width
or less) can result from a partial breakdown of larger pieces of plastic, and may be so small that they are
not visible to the human eye. Microplastic particles have been found in both well -trafficked and remote
water bodies around the world,20 where they can leach bioaccumulative toxins — or even be incorporated
directly into the tissues of animals, including humans who eat plastic -contaminated seafood. The
consequences of widespread plastic pollution clearly include not only aesthetic problems and ecological
harm, but potentially serious direct impacts to human health as well.
Local Roots of a Global Problem
The connection between local litter and marine debris is intuitive in concept, though arguably not intuitive
in scale. This summer's surveys demonstrate that floatable trash is abundant and mobile in the campus
stormwater drainage systems, and likely into the broader Cape Fear watershed. Trash generated on the
Duke campus, particularly highly mobile floatable items like plastics and foams, are likely to flow toward
Sandy Creek, New Hope Creek, and through increasingly large waterways ultimately draining to the
Triangle area's major drinking water reservoirs and to the marine environment.
While the amount of litter present on the Duke campus is minute in comparison to the total volume of
plastic pollution present globally, this massive worldwide volume of litter is nonetheless significantly
derived from cumulative action (and inaction) at small-scale and local levels across the country and around
the world. Estimates from 201621 suggest that roughly 80% of the plastic that ends up in marine
environments originates on land. It is reasonable to assume, therefore, that partial solutions to this
massive problem can also be developed through actions at the local level, magnified by implementation
across many areas. Duke's contribution to the global plastic problem may be small, but the potential exists
for Duke's efforts to help generate effective local trash control solutions that could have an impact far
beyond the campus's borders.
Duke Can Lead the Way in Durham
The Clinic is working with the City of Durham on potential means of addressing trash at the municipal
scale. Part of the Clinic's motivation for seeking to partner with FMD on understanding and controlling
16 See reports from Ocean Cleanup, which spearheaded the most extensive survey ever conducted of the zone.
17 Id.
18 Ryan P.G. (2015) A Brief History of Marine Litter Research. In: Bergmann M., Gutow L., Klages M. (eds) Marine
Anthropogenic Litter. https://doi.org/10.1007/978-3-319-16510-3 1 .
19 Id.
20 Barnes, D. K. A., Galgani, F., Thompson, R. C., & Barlaz, M. (2009). Accumulation and fragmentation of plastic debris
in global environments. Philosophical Transactions of the Royal Society B: Biological Sciences, 364(1526), 1985-1998.
http://doi.org/10.1098/rstb.2008.0205
21 Eunomia. "Plastics in the Marine Environment." June 2016. Available for download at
http://www.eunomia.co.uk/reports-tools/plastics-in-the-marine-environment/.
24

trash moving through Duke's campus is the potential for lessons learned on the hyperlocal scale to be
applied eventually to the broader Durham region, or even in other interested cities. Duke's participation
in developing effective means of studying and controlling its own litter, therefore, could directly
contribute to broader solutions adopted at the municipal level and beyond.
Recommendations for Future Follow -Up
• Investigation of litter generation upstream of RP2A — science lab buildings and hospital complex.
• Investigation of trash in the Golf Course irrigation intake, and its associated overtopping
problems, might make for an interesting case study.
Additional follow-up projects discussed with FMD include:
• Potential pilot study of an intake at the created wetland upstream of site SWMP1 to study flow
rate of trash into this area, using a structural trash capture device — to be discussed further before
start of Fall semester?
• Educational and behavior -change campaigns including periodic required education or training for
FMD or groundskeeping employees, as well as campaigns targeted at students and other
customers in areas serviced by Duke Dining.
• Establishment of metrics to gauge campus progress on trash control over time.
[Ryan to insert additional notes from 7/25 Meeting]
25

Assessing Litter Loading in Urban Streams:
Litter Sampling Protoco
Prepared by
Duke ENVIRONMENTAL
LAW ,wel POLICY CLINIC
U N I V E R S I T Y
July 2018

Duke Environmental Law and Policy Clinic July 2018
Tableof Contents....................................................................................................................1
Introduction...........................................................................................................................2
Litter loading in urban waterways.................................................................................................................2
Why survey litter in streams?........................................................................................................................3
Aboutthis protocol........................................................................................................................................4
Methodology..........................................................................................................................5
Siteselection..................................................................................................................................................5
Preparingto sample.......................................................................................................................................6
Samplingprocess...........................................................................................................................................7
Follow-up sampling........................................................................................................................................9
Equipmentlist..............................................................................................................................................10
AnalyticalMethods...............................................................................................................11
Assessing the level of litter impairment......................................................................................................11
Identifyingbrands and logos.......................................................................................................................14
References and Resources.....................................................................................................15
Appendix................................................................................................................................1
Samplingsheet...............................................................................................................................................1
Example analysis spreadsheet.......................................................................................................................3
Examplecalculations.....................................................................................................................................4
Litter Sampling Protocol for Urban Streams 1

Duke Environmental Law and Policy Clinic July 2018
Figure 1. Signage at Beaver Creek, a wetland area in Durham, NC preserved by the Ellerbe Creek Watershed Association.
LITTER LOADING IN URBAN WATERWAYS
The problem of plastics and other non -biodegradable litter in the environment has come to the forefront
as one of the most tangible and nefarious impacts of humans on the natural world. As of 2017, plastics
have been found in all of the global ocean's major gyres, and in some of the world's most remote locations
including Henderson Island in the South Pacific,' and the Mariana Trench.'
Plastics have also shown up in the stomachs of seabirds, sea turtles, fish, and marine mammals including
whales.' Often, animals mistake plastic items for food, and emerging research shows that the chemicals
found in some plastics may in fact attract animals to them.' The physical presence of plastics and other
litter in the environment can also disrupt habitats and alter natural processes such the flow of rivers.
Most litter that ends up in the ocean actually originates far inland, and is carried to the marine
environment via stormwater that flows into streams and rivers, and ultimately, the ocean. In fact, an
estimated 80% of marine debris stems from on -land sources.5
Urban areas are important sources of plastics and other litter into local waterways — litter collects near
streets, houses, and businesses, and is easily carried into streams during storm events. Once in the
waterway, this litter degrades the aesthetic quality of the environment, disrupts important
hydrogeological processes, and threatens human and environmental health. Plastics can also break down
Litter Sampling Protocol for Urban Streams 2

Duke Environmental Law and Policy Clinic July 2018
into smaller pieces, called microplastics, which can infiltrate drinking water and invade ecosystems, posing
an even greater risk to health.6
Fortunately, there are tools available to help stem the tide of litter into urban waterways. All stormwater
that passes through municipal stormwater systems (called "MS4s") is monitored and regulated via a
permitting process administered under the Clean Water Act. Although litter is not often regulated under
this system, certain cities, including Honolulu, New York, and cities throughout California, have developed
litter reduction provisions and implemented them in their stormwater permits.?
The cities that have successfully adopted litter reduction provisions into their stormwater permits and
implemented plans to reduce litter in their stormwater system all started by measuring and documenting
the amount of litter present in their waterways. Building an understanding of the extent of the litter
problem is an essential first step in convincing stormwater managers and other authorities that litter is
present in amounts that threaten water quality and impair ecosystems.
The litter survey protocol outlined in this document provides a simple, low-cost method by which citizens,
advocacy groups, and local authorities can measure and record the level of litter impairment in their local
streams. Establishing this baseline understanding of the amount, type, and location of litter in urban
waterways is critical for developing workable solutions to the persistent, ever-expanding problem of litter
pollution in our environment.
WHY SURVEY LITTER IN STREAMS?
Determining the amount, type, and distribution of litter in urban waterways is necessary in order to
demonstrate whether a litter problem is present. In addition to increasing stormwater managers'
understanding of the extent of litter pollution, litter surveys provide important information about the
baseline load of litter in local waterways. These data are essential during the process of developing and
implementing controls on the sources of stormwater litter. For example:
• Understanding the spatial distribution of litter and identifying locations where litter accumulates
helps cities develop targeted litter controls to deploy within a stream, thereby saving costs.
• Estimating the amount of litter in a waterway provides important baseline or "business -as -usual"
data, which can then be compared to litter levels following implementation of control methods
to determine their effectiveness.
Documenting the baseline litter level has been a key component of successful efforts to develop litter
reduction provisions in certain urban areas, including around the Anacostia and throughout California.' In
these regions, stormwater managers needed to know what the starting conditions were so that they could
develop plans to reduce the amount of litter in their waterways and track progress towards their litter -
reduction goals.
Litter Sampling Protocol for Urban Streams 3

Duke Environmental Law and Policy Clinic July 2018
Using a simple, replicable litter survey methodology is necessary so that data can be compared between
cities and across time. The protocol outlined in this document can be easily implemented in streams of
various sizes and provides a cost-effective method of estimating baseline litter loads in urban waterways.
ABOUT THIS PROTOCOL
The Duke Environmental Law and Policy Clinic (ELPC) conducted a pilot project during Summer 2017 to
determine a simple, replicable, and cost-effective method of assessing baseline litter loading into an urban
waterway. This survey methodology can be implemented by two or three people in the field, requires very
little specialized equipment, and provides data that can be easily analyzed to inform the development of
litter reduction plans.
The protocol outlined in the following section is based on survey techniques that were pioneered in the
Anacostia ("Stream Trash Indexing System") and in the San Francisco Bay area ("Rapid Trash Assessment
Protocol").' These techniques center around two common elements:
• The use of 100-foot (-30-meter) transects within which litter is collected
• Pre -determined impairment scores to quantify litter levels in the stream, e.g.:
Number of Litter Items
per 100ft/30m (avg.)
0-10
10.1-25
25.1-50
>50.1
Impairment Score
None/very light
Light
Moderate
High
The survey methodology developed by ELPC built off this foundation, adding three key elements:
• Recording the types of litter collected
• Keeping track of identifiable brands or logos
• Cleaning -up the stream!
ELPC's method provides a simple way of quantifying the baseline load of litter and determining potential
litter "hotspots." In addition, this protocol allows surveyors to track the distribution of certain types of
litter within a stream, and keep a record of the brands and logos on litter items to identify the businesses
or companies that bear responsibility for this litter upstream.
Implementing a consistent methodology across litter assessment efforts is important to build an
understanding of patterns in litter loading into streams both inland and at the coast.
Litter Sampling Protocol for Urban Streams 4

Duke Environmental Law and Policy Clinic
July 2018
Duke Environmental Law & Policy Clinic 0 0,25 0.5 1 Mile
Summer 2017
Figure 2. Example sampling map, showing the location of selected sampling sites along a waterway. Inset
illustrates the location of three 30m transects (TR1-TR3) at Site 1.
SITE SELECTION
The following site selection criteria should be used to identify appropriate sites for sampling:
• Easily accessible - Sites should be easily accessible by foot, and should be located on public or
publicly -accessible land (unless permission is specifically granted by the land owner to access a
stream).
• Minimum stream length of 150m - Sites should include at least 150m of continuous, accessible
stream. For this protocol, three 30m sampling transects are surveyed, interspersed with two 30m
"buffer" transects. See Figures 2 and 3.
Litter Sampling Protocol for Urban Streams

Duke Environmental Law and Policy Clinic
July 2018
At least three sampling transects should be surveyed at each site if possible. See Figures 2 and 3. This
replication is necessary to ensure that an "average" level of litter can be calculated for each site. See
Analytical Methods, below.
Road passing
over creek
within this area
Figure 3. Illustration of site layout and sampling methodology.
PREPARING TO SAMPLE
Prior to sampling:
1. Map sampling sites. This can be done using a detailed map, Geographic Information System
(GIS) software, or by visiting sites in the field to identify appropriate transects.
- Note: The litter survey is not intended to target the most litter -impaired portions of the
waterway, but instead build an understanding of the actual baseline litter levels in the
stream. Therefore, sites should be selected "blindly" without knowledge about the level of
litter present in the stream.
- To determine an appropriate starting point for the sampling transects, consider selecting a
point in the stream where a road or a bridge crosses the stream, or where the stream is
blocked (e.g., by a dam or a natural log boom. These places may impair accessibility, so it is
easier to start the transects on one side or the other of blockages). See Figure 3.
2. Visit sampling sites. After selecting sites using GIS or another method as outlined above, visit the
field sites in -person to ensure that they are accessible and free of hazards. Be on the lookout for
steep banks, slippery/sharp rocks, deep water, or other hazards that may make sampling
dangerous or impossible. If sampling will not be possible at the selected site, scope out
upstream or downstream locations to find an appropriate site.
3. Gather field equipment and sampling sheets. See Equipment List, below.
Litter Sampling Protocol for Urban Streams 6

Duke Environmental Law and Policy Clinic
July 2018
4. Organize sampling crew. At least two people should be available for sampling. One acts as the
litter collector, the other acts as the note -taker. Three people are ideal — that way, two people
can collect litter simultaneously (one on each side of the stream), and one person can record
data.
SAMPLING PROCESS
On the day of sampling, follow the protocol outlined below:
1. Travel to the site that has been selected for sampling.
2. Fill out the site information on the sampling form. See Appendix.
3. Map out the first 30m transect at the site:
- Mark the beginning of the transect with flagging tape or sampling marker poles.
- To map the 30m length, either estimate the distance by pacing the stream (a large step by an
adult is around 1m), or measure the distance using a measuring tape, pre -measured rope, or
handheld GPS unit.
- Mark the end of the transect with flagging tape on the banks or a marker in the stream.
4. Collect all litter (to the extent possible) from the center, sides, and 1m up each bank of the stream
within the 30m sampling transect.
- Do not survey litter within the buffer transect — although you can still pick it up and dispose of
it! See Box 1.
- If the stream is too deep to sample by foot (e.g., with hip or chest waders), use a canoe or
kayak to access the center of the stream.
Box 1. Buffer transects
"Buffer transects" refer to the transects that fall in-between the sampling transects from which you will
collect litter. Interspersing sampling transects with buffer transects allows the survey protocol to cover
more area of the stream without adding to the surveying burden. By covering more ground, this survey
method allows data -collectors to extrapolate site -wide trends from a relatively short stream segment.
5. As each litter item is collected, the sampler should state the type of litter (See Box 2), and the
note -taker should record the number of litter items and type of litter on the sampling sheet. See
Appendix.
6. If logos or brands are present on litter items, take note of those in the areas provided on the
sampling sheet.
Litter Sampling Protocol for Urban Streams 7

Duke Environmental Law and Policy Clinic
July 2018
7. Place the litter item in the trash bag and move on.
- Note: Some litter items may be too large to collect. In this case, take a picture of the item if
possible, and note the type of litter and any associated brands/logos on the collection sheet
before moving on. If stream will be sampled a second time (e.g., pre- and post -rain sampling),
make a note about these items to avoid double -counting them during the second visit.
- Note: It is good practice to take pictures of interesting, unlikely, or exceptionally large pieces
of litter. It is also recommended that you take pictures of areas where trash has accumulated
in the stream (potential "hotspots") or blockages formed of litter items. Photographs of these
provide evidence of the physical impediment formed by the presence of trash in waterways,
or of the natural morphological features that promote heavy litter accumulation. See Figure 4.
8. Repeat for the remaining transects at the site.
9. Dispose of trash. Depending on the goals of the litter survey, samplers may also wish to weigh the
collected trash prior to disposal.
Box 2. Categories of Litter— During sampling, keep a record of the type of litter that is collected:
Plastic film (single -use plastic bags, chip and candy wrappers, bottle labels, other fragments of
film such as tape)
Hard plastic (water bottles, caps and lids, disposable cutlery, miscellaneous hard plastic items)
Styrofoam & other foam (take-out containers, packing peanuts, disposable cups)
Metal (beer cans, fragments of aluminum foil)
Glass (glass bottles, fragments of glass)
Other material (materials generally found in lower volumes than those noted above, e.g., rubber,
ceramic, fabric, string, latex/nitrile gloves)
Sports equipment & other large items (generally composed of a mixture of materials, e.g., soccer
ball, shoe, swimming goggles, bicycles, shopping carts
Because litter is not always present as intact items (e.g., an entire plastic bag or glass bottle), samplers
may wish to include subcategories that specify whether the item was "whole" or a "fragment" for each
category.
Litter Sampling Protocol for Urban Streams

Duke Environmental Law and Policy Clinic
July 2018
Figure 4. (Left) example of a large and unlikely piece of debris found in an urban stream. This item would be classified
as "Sports Equipment" in Box 2. (Right) Example of a location of litter accumulation within an urban stream.
FOLLOW-UP SAMPLING
Depending on local conditions and the goal of the litter survey, data -collectors may want to conduct
follow-up sampling.
• If the goal of the litter survey is to establish a connection between stormwater runoff and litter
levels in local waterways, data -collectors can conduct litter surveys following rainfall events and
compare these levels to control data that is collected during a dry spell. Be aware that high
turbidity in the immediate wake of a storm event can reduce visibility of litter, potentially skewing
results in these instances. Raised water levels following a storm can also change access
conditions, so proceed with care.
• If the goal is to monitor changes in litter levels over the course of a year (e.g., across the seasons),
data -collectors can conduct sampling at three-month intervals throughout the calendar year.
• Data -collectors may also wish to record other information about local water quality, ecological
health or hydrodynamic function, as appropriate. If the goal of sampling is to demonstrate the
water quality impacts of plastics, data -collectors may want to collect samples and have them
tested for the presence of microplastics or other plastic byproducts that can impair
environmental and human health (e.g., Bisphenol A (BPA), styrene trimer).11
Litter Sampling Protocol for Urban Streams 9

Duke Environmental Law and Policy Clinic
July 2018
EQUIPMENT LIST
Required
• Heavy-duty gloves
• Tall, waterproof boots
• Heavy-duty trash bags
• Clip board
• Sampling sheets on waterproof paper
• Extra waterproof paper for field notes
• Pencils
• Flagging tape or markers
• Rope or measuring tape
• Camera
Recommended
• Water proof gloves or glove inserts
• Insect repellent
• First -aid supplies including bandages,
antibacterial ointment, snake bite kit, and
poison ivy salve
• Chest or hip waters for medium -deep water
• Canoe or kayak for deep water
• Trash picker/grabber
• GPS unit or camera with location services
enabled, for marking field location
• Hat, sunscreen, sunglasses
Litter Sampling Protocol for Urban Streams 10

Duke Environmental Law and Policy Clinic
July 2018
ANALYTICAL METHODS
Duke lriwonmental Law F. Policy Chnrc
Summer 1017 0 0 25 0 5 l Mile
Figure S. Map of the Ellerbe Creek in Durham, NC, showing the location, volume and type of litter items collected
during ELPC's Summer 2017 pilot study.
ASSESSING THE LEVEL OF LITTER IMPAIRMENT
Determining the level of litter impairment involves a simple analysis to estimate the average number of
litter items present in a "typical" 30m transect at each site. This analysis also accounts for the variability
present in the data. As shown in the table below, if transect 1 (TR1) has 15 items, transect 2 (TR2) has 47
items, and transect 3 (TR3) has 30 items, the average number of items per 30m transect is 30.6, but this
value does not take into consideration the fact that the actual number of items ranged from 15 to 47.
Calculating the statistical variation around these values, known as standard error, provides an estimate of
the accuracy with which the average value represents the true average number of litter items per 30m
transect at a given site. An example of this calculation is provided below for a theoretical "Site 1."
P
Litter Sampling Protocol for Urban Streams 11

Duke Environmental Law and Policy Clinic July 2018
First, organize the litter data into a table:
Transect Number of Items Total
Number Plastic Film Hard Plastic ... (add across)
Site 1 - TR1
10
5
(add across) 15
3
...
Site 1 - TR2
0
33
14
... 47
v
Site 1 - TR3
18
12
... 30
Total (add down)
61
31
... 92
Average/30m
20.3
10.3
... 30.6
Standard
11.7
4.7
... 16.0
Deviation
Standard Error
6.7
2.7
9.2
• Calculate the total number of each type of litter item at site 1 (plastic film, hard plastic, etc.) by
summing all transects ("add down"). Site 1 has 61 plastic film items total, or20.3 plastic film items
on average per 30m transect.
• Calculate the total number of items in each transect by summing across the types of litter ("add
across"). Transect 1 at Site 1 has 15 items total.
• Add the total number of items found at each transect together to find the total number of litter
items present at Site 1. Site 1 has 92 litter items total.
• To find the average number of items per 30m transect, divide the total number of litter items at
Site 1 by the number of transects (3). 9213 = 30.6
• This is the number that will be used to determine the impairment score. See below.
Number of Litter Items
per 100ft/30m (avg.) Impairment Score
0-10 None/very light
10.1-25 Light
25.1-50 Moderate
50.1-100 High
100.1-200 Very High
>200.1 Extremely High
Litter Sampling Protocol for Urban Streams 12

Duke Environmental Law and Policy Clinic
July 2018
Based on this table, the litter pollution at Site 1 would be classified as "moderate." However, this
determination does not take into account the fact that one of the transects at Site 1 (TR1), would have
been classified as "light" since it only had 15 items. Therefore, a measure of variability (how spread out
the values are) should be included when reporting the average number of litter items at each site. To do
this, calculate standard error from the standard deviation, as illustrated below using the example of Site
1 in the table above.
Standard deviation - the easiest way to calculate this is by using the stdev function in
Excel. Using Site 1 as an example:
=stdev(15,47,30) = 16.0
There are also online calculators available to help determine standard deviation, for
example http://www.calculator.net/standard-deviation-calculator.html.
The formula for calculating standard deviation by hand is provided in the Appendix.
Standard error - again, the easiest way to calculate standard error is in Excel, using
the following equation:
=((stdev)/-V-n-)
Where "stdev" is the standard deviation calculated above, and "n" is the number of
transects (in this case 3):
=(16)/(V-3) = 9.2
So, the standard error for the average number of litter items collected at Site 1 is 9.2. Therefore, the
average should be presented as 30.6 ± 9.2. Reporting it this way indicates that although the average level
of litter fell within the "moderate" impairment category, individual transects fell above and below this
value.
Standard error can be represented on graphs using error bars. For example, the graph below shows the
average number of plastic film and hard plastic items found at Site 1, along with the standard error.
Litter Sampling Protocol for Urban Streams 13

Duke Environmental Law and Policy Clinic
July 2018
30
V
V)
N
C 25
L
VC
C
Q
L 20
aj
n
cN
C
Y_ is
w
O
L
10
i
v
so 5
L
Q
0
Plastic film
20.3±6.7
IDENTIFYING BRANDS AND LOGOS
Hard plastic
10.3 ± 2.7
Brands and logos can be reported in different ways, depending on the goal of the sampling and the types
of brands identified. For example, data -collectors can split brands and logos into different types of
products (e.g., fast food restaurants, beverage companies, beer brands), or by parent company (e.g.,
PepsiCo owns many brands, including Gatorade, Cap'n Crunch, and Smartfood popcorn). If there are
certain local establishments that data -collectors would like to draw attention to, these brands can be
highlighted, especially if litter from these businesses is accumulating in discrete "hotspots."
Litter Sampling Protocol for Urban Streams 14

Duke Environmental Law and Policy Clinic
July 2018
1 Ed Young, A Remote Paradise Island is Now a Plastic Junkyard, The Atlantic (May 15, 2017), https://www
.theatlantic.com/science/archive/2017/05/a-remote-paradise-island-is-now-a-plastic-junkyard/526743/.
z Plastic Pollution Coalition, Pollution Found in the Most Remote Park of the World Ocean,
plasticpollutioncoalition.org (February 15, 2017), http://www.plasticpollutioncoalition.org/pft/2017/2/
15/pollution-found-in-the-most-remote-part-of-the-world-ocean
3 Seabirds: Chris Wilcox, Erik Van Sebille & Britta Denise Hardesty. (2015). Threat of plastic pollution to
seabirds is global, pervasive, and increasing, PNAS 112: 11899-11904.; Marine mammals: Charles James
Moore. (2008). Synthetic polymers in the marine environment: A rapidly increasing, long-term threat,
Environmental Research 108: 131-139.; Sea turtles: Rita Mascarenhas, Robson Santos & Douglas
Zeppelini. (2004). Plastic debris ingestion by sea turtle in Paraiba, Brazil, Marine Pollution Bulletin 49: 354-
355.; Fish: Christinana M. Boerger, Gwendolyn L. Lattin, Shelly L. Moore & CharlesJ. Moore. (2010). Plastic
ingestion by planktivorous fishes in the North Pacific Central Gyre, Marine Pollution Bulletin 60: 2275-
2278.
4Austin S. Allen, Alexander C. Seymour & Daniel Rittschof. (2017). Chemoreception drives plastic
consumption in a hard coral, Marine Plastic Bulletin 124: 198-205.
s Chris Sherrington, Plastics in the Marine Environment, Eunomia (June 2016), http://www.eunomia.co.
uk/reports-tools/plastics-in-the-marine-environment/.
6 Amy Fox, Lindsay McGairty & Margaux Bergen, Global Scientific Study Finds Microscopic Plastic Fibers
Contaminating Tap Water, Orb Media & Edelman (September 2017), https:Horbmedia.org/sites/default
/files/Orb%20Media%20M icroscopic%20Plastics/20Press%20Release%20SEP%205%202017. pdf.
Honolulu: City and County of Honolulu Department of Environmental Services, Authorization to
Discharge under the National Pollutant Discharge Elimination System (Permit No. HI S000002, 2015) at 32,
http://www.honolulu.gov/rep/site/dfmswq/dfmswq_docs/NPDES_permit_2015.pdf.; New York City:
New York State Department of Environmental Conservation, State Pollutant Discharge Elimination System
(SPDES) Discharge Permit (Permit No. NY-0287890, 2015) at 28, http://www.dec.ny.gov/docs/water_
pdf/nycros4permit.pdf.; California: California Environmental Protection Agency Division of Water Quality,
Amendment to the Water Quality Control Plan for the Ocean Waters of California to Control Trash and
Part 1 Trash Provisions of the Water Quality Control Plan for Inland Surface Waters, Enclosed Bays, and
Estuaries of California, State Water Resources Control Board (April 7, 2015), https://www.waterboar
ds.ca.gov/water issues/programs/trash control/docs/01_final sed.pdf
8 Anacostia: John Galli & Kathy Corish, Anacostia Stream Trash Surveying Methodology and Indexing
System, Anacostia Trash Workgroup (May 19, 1998), https://www.anacostia.net/Archives/download/
TrashSurveyProtocol.pdf.; California: Geoff Brossau, Tracking CA's Trash: On -land Visual Assessments, Bay
Area Stormwater Management Agencies Association (March 21, 2017), http://basmaa.org/An noun
cements/tracking-cas-trash-on-1and -visual-assessments.
9 See Anacostia Stream Trash Surveying Methodology and Indexing System, supra note 8, at 2 ("Stream
Trash Indexing System"); Surface Water Ambient Monitoring Program (SWAMP), A Rapid Trash
Assessment Method Applied to Waters of the San Francisco Bay Region: Trash Measurement in Streams,
California Water Boards (April 2007), https://www.waterboards.ca.gov/sanfranciscobay/docs/swampt
Litter Sampling Protocol for Urban Streams 15

Duke Environmental Law and Policy Clinic
July 2018
hrashreport.pdf. (Note: the methodology pioneered in the Rapid Trash Assessment (RTA) Method is more
complex than that used in the Anacostia Stream Trash Indexing System (STIS), likely because the rivers
being surveyed in the San Francisco Bay region are larger and more polluted than the tributaries of the
Anacostia. Notably, the RTA method involves scoring 100m stream segments on qualitative and
quantitative levels of trash, as well as on estimates of water quality. The STIS method is based solely on
the number of litter items present in the stream.
io Hadley Leggett, Toxic Soup: Plastics could be leaching chemicals into ocean, Wired (August 19, 2009),
https://www.wired.com/2009/08/plasticoceans/.
Litter Sampling Protocol for Urban Streams 16

Duke Environmental Law and Policy Clinic
July 2018
SAMPLING SHEET
(See next page)
Litter Survey Protocol for Urban Streams Al

Duke Environmental Law and Policy Clinic
July 2018
Date:
Time:
Location:
GPS Coordinates:
Right bank (start):
Left bank (start):
Right bank (end):
Left bank (end):
Names of data -collectors:
Transect no.
Litter type
Number of items
Brands/logos identified
Litter Survey Protocol for Urban Streams A2

Duke Environmental Law and Policy Clinic
EXAMPLE ANALYSIS SPREADSHEET
Site Name:
July 2018
Number of Litter Items
Transect
Sports
Totals
Styrofoam &
Other
Number
Plastic Film
Hard Plastic
Metal
Glass
Equipment,
(add across)
Other Foam
Material
etc.
TR1
TR2
TR3
Total
(add down)
Average/30m
Standard
Deviation
Standard
Error
Litter Survey Protocol for Urban Streams A3

Duke Environmental Law and Policy Clinic
EXAMPLE CALCULATIONS
July 2018
Transect
Number
Plastic Film
Number of Items
Hard Plastic
Total
(add across)
Site 1 - TR1
10
5 ... (add across)
15
3
Site 1 - TR2
O
-0
33
14 ...
47
0
m
V
Site 1 - TR3
18
12 ...
30
Total (add down)
61
31 ...
92
Average/30m
20.3
10.3 ...
30.6
Standard
11.7
4.7 ...
16.0
Deviation
Standard Error
6.7
2.7
9.2
Calculating standard
deviation (SD) for total number of litter items at Site 1 (above):
SD — Y^V (X,
— µ)z
Where: SD = standard deviation
N1 i
N = number of samples (number of transects)
µ = sample mean
SD = �3 1 1 [(15 — 30.6)2 + (47 — 30.6)2 + (30 — 30.6)2]
SD = 2 [-15.62 + 16.42 + (-0.6)2] =Ti(12.67) = 256.34 = 16.01
Calculating standard error (SE) for total number of litter items at Site 1:
SE _ SD Where: SD = standard deviation
N = number of samples (number of transects)
16.01
SE _ - = 9.2
Litter Survey Protocol for Urban Streams A4

Appendix B: Rainfall Data
All rainfall measurements were taken from the United States Geologic Survey's Maureen Joy Charter
School precipitation gauge. This gauge is located at 1955 W Cornwallis Rd. at a former Maureen Joy
Charter School facility, now operating as the Carter Community Charter School.
t7
oo rfdgeRG
CI, RAINGA13E AT MAUREEN JOY
'0CHARTER SCHOOL NR DURHAM
Last Data Update:
2018-07-F7 11:30:00 ED7
Carter Cam mun i. 15 }dour Total:
Charter Sc ICA
inches
[7
Lerner Jewish
01 Cornmunity Day School
Levin Jewish
Ccmmunity Center 19
Cornwallis
Road Parr
Figure 1. USGS rain gauge map showing the Maureen Joy Charter School NR Durham gauge, located at what is now Carter
Community Charter School at 1955 W. Corwallis Rd.
This data is available online at the following address:
https://waterdata.usgs.gov/nwis/uv?cb 00045=on&format=gif stats&site no=355852078572045
Baseline sampling was conducted after more than 4 days without a major rain event registering at the
USGS gauge. Post -rain sampling was conducted within 1-2 days of such an event. For the purposes of this
study, a major rain event sufficient to trigger post -rain sampling was defined as 0.1" of rain or more.
Appendix B - 1

Rec Pond Samples
Baseline sampling for stream segments RP2A, RP2B, RP2C, and C_SC1 was conducted on June 7t" and 811
Less than 0.1" of rain fell on June 8t"; the first major rain event after baseline sampling was considered to
be on June 101", when approximately 1.25" of rain was recorded at the gauge.
1.40
1.00
ti
C
• 4 0.60
a=
�A
.C6 0.40
C]
0
L
iL 0.20
0.00
USGS 355852078572045 RAIMGAGE AT MAUREEN JOY CHARTER SCHOOL HR OURHAM
Jun Jun Jun Jun Jun Jun Jun Jun
07 08 09 10 11 12 13 14
2018 2018 2018 2018 2018 2018 2018 2018
---- Provisional data Subject to Revision ----
Figure 2. Rainfall recorded at Maureen Joy Charter School USGS gauge, June 7" through 14t".
Appendix B - 2

SWAMP Samples
Baseline sampling for stream segments SWMP1 and GC1 was conducted on June 191n and June 201n. A
major rain event did occur on June 27tn, but due to personnel limitations, post -rain sampling was not
conducted at this time. Subsequently, another major rain event (-0.6") occurred on July 4tn. Post -rain
sampling was subsequently conducted on July 5tn
USGS 355852078572045 RAIHGAGE AT MAUREEN JOY CHARTER SCHOOL HR DURHAM
2.5
t
1.5
a
0.0
Jun Jun Jun Jun Jun Jun Jul Jul Jul
19 21 23 25 27 29 01 03 05
2018 2018 2018 2018 2016 2018 2018 2016 2018
---- Provisional Data Subject to Revision ----
Figure 3. Rainfall recorded at Maureen Joy Charter School USGS gauge, June 191h through July 51n
Appendix B - 3

Appendix C: Sampling Data Tables
Rec Pond Sites
30-
Sample
Sample
Rainfall
meter
Hard
Soft
Sports
Segment
Time
Styrofoam
Metal
Glass
Other
Total
Date
(inches)
Transect
Plastic
Plastic
Equipment
(start)
Name
-
-
C SC1 3
-
-
-
-
-
-
-
-
6/08/18
11:09
C_SC1_1
4.0
7.0
3.0
3.0
0.0
0.0
3.0
20.0
6/08/18
11:35
C_SC1_2
3.0
6.0
1.0
0.0
2.0
0.0
1.0
13.0
Average
3.5
6.5
2.0
1.5
1.0
0.0
2.0
16.5
C_SC1
6/12/18
9:25
C_SC1_1
1.0
10.0
2.0
0.0
0.0
0.0
1.0
14.0
6/12/18
9:45
1.2
C_SC1_2
1.0
5.0
0.0
0.0
2.0
0.0
0.0
8.0
6/12/18
10:05
C_SC1_3
7.0
7.0
7.0
4.0
0.0
0.0
2.0
27.0
Average
3.0
7.3
3.0
1.3
0.7
0.0
1.0
16.3
-
-
RP2A 1
-
-
-
-
-
-
-
-
6/07/18
10:15
RP2A-2
0.0
38.0
0.0
4.0
0.0
0.0
9.0
51.0
6/08/18
9:46
RP2A-3
13.0
201.0
6.0
9.0
0.0
1.0
33.0
263.0
Average
6.5
119.5
3.0
6.5
0.0
0.5
21.0
157.0
RP2A
6/13/18
9:17
RP2A-3
10.0
64.0
4.0
6.0
3.0
0.0
18.0
105.0
6/13/18
9:45
1.2
RP2A-2
8.0
54.0
7.0
2.0
0.0
0.0
21.0
92.0
6/13/18
10:05
RP2A-1
2.0
36.0
3.0
0.0
2.0
0.0
6.0
49.0
Average
6.7
51.3
4.7
2.7
1.7
0.0
15.0
82.0
6/07/18
10:45
RP2AB
4.0
34.0
6.0
6.0
2.0
0.0
24.0
76.0
6/08/18
8:15
RP2B_1
3.0
3.0
8.0
2.0
0.0
0.0
4.0
20.0
6/08/18
8:55
RP2B_2
3.0
9.0
4.0
1.0
4.0
0.0
4.0
25.0
Average
3.3
15.3
6.0
3.0
2.0
0.0
10.7
40.3
RP213
6/13/18
10:23
RP2AB
0.0
12.0
0.0
2.0
0.0
0.0
7.0
21.0
6/13/18
8:30
1.2
RP2B_1
2.0
6.0
6.0
7.0
6.0
0.0
2.0
29.0
6/13/18
8:46
RP2B_2
4.0
19.0
14.0
2.0
1.0
0.0
2.0
42.0
Average
2.0
12.3
6.7
3.7
2.3
0.0
3.7
30.7
6/07/18
8:10
RP2C_3
2.0
12.0
9.0
1.0
2.0
0.0
5.0
31.0
6/07/18
8:45
RP2C_2
9.0
31.0
12.0
9.0
2.0
0.0
15.0
78.0
6/07/18
9:25
RP2C_1
9.0
18.0
3.0
8.0
2.0
0.0
1.0
41.0
Average
6.7
20.3
8.0
6.0
2.0
0.0
7.0
50.0
RP2C
6/12/18
8:16
RP2C_3
2.0
29.0
7.0
3.0
0.0
0.0
6.0
47.0
6/12/18
8:41
1.2
RP2C_2
1.0
8.0
2.0
2.0
1.0
0.0
1.0
15.0
6/12/18
9:05
RP2C_1
5.0
9.0
11.0
3.0
0.0
0.0
1.0
29.0
Average
2.7
15.3
6.7
2.7
0.3
0.0
2.7
30.3
Appendix C - 1

SWAMP Sites
Segment
Sample
Date
Sample
Time
(start)
Rainfall
30-meter
Transect
Name
Hard
Plastic
Soft
Plastic
Styrofoam
Metal
Glass
Sports
Equipment
Other
Total
6/20/18
9:14
SWMP1_1
4.0
5.0
0.0
3.0
47.0
2.0
4.0
65.0
6/20/18
9:25
SWMP1_2
1.0
8.0
1.0
1.0
7.0
0.0
3.0
21.0
6/20/18
9:37
SWMP1_3
5.0
9.0
1.0
3.0
0.0
0.0
1.0
19.0
Average
3.3
7.3
0.7
2.3
18.0
0.7
2.7
35.0
SWMP1
7/05/18
10:06
SWMP1_1
1.0
2.0
0.0
0.0
13.0
0.0
3.0
19.0
7/05/18
10:16
0.68 SWMP1_2
2.0
9.0
0.0
1.0
1.0
0.0
2.0
15.0
7/05/18
10:28
SWMP1_3
1.0
8.0
0.0
0.0
4.0
0.0
1.0
14.0
Average
1.3
6.3
0.0
0.3
6.0
0.0
2.0
16.0
6/19/18
8:14
GC1_1
6.0
16.0
3.0
33.0
15.0
4.0
4.0
81.0
6/19/18
9:14
GC1_2
2.0
14.0
0.0
4.0
37.0
1.0
2.0
60.0
6/19/18
9:45
GC1_3
3.0
5.0
1.0
3.0
0.0
18.0
9.0
39.0
Average
3.7
11.7
1.3
13.3
17.3
7.7
5.0
60.0
GC1
7/05/18
8:40
GC1_1
1.0
0.0
0.0
0.0
1.0
0.0
1.0
3.0
7/05/18
9:00
0.68 GC1_2
1.0
2.0
0.0
2.0
2.0
0.0
2.0
9.0
7/05/18
9:17
GC13
0.0
3.0
0.0
0.0
0.0
3.0
1.0
7.0
Average
0.7
1.7
0.0
0.7
1.0
1.0
1.3
6.3
Appendix C - 2