The Evergreen State College Institutional Repository

Wolff_DMESthesis2011.pdf

Media

Part of Site Selection for Eel Grass: A Model for Puget Sound Replanting Projects

extracted text (extracttext:extracted_text)

SITE SELECTION FOR EELGRASS: A MODEL
FOR PUGET SOUND REPLANTING PROJECTS

By
Daniel Wolff

A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2011

© 2011 by Daniel Wolff. All rights reserved.

This Thesis for the Master of Environmental Study Degree
By
Daniel Wolff

has been approved for
The Evergreen State College
by

____________________________________
Judith Bayard Cushing, PhD.
Member of the Faculty
Computer Science and Ecology Informatics

____________________________________
Date

ABSTRACT

A Site Selection Model for Eelgrass (Zostera marina)
Replanting Projects in the Puget Sound

Daniel Wolff

Eelgrass (Zostera marina) is an important nearshore habitat in
ecosystems around the world. In recent years extensive meadows have
been threatened or destroyed by anthropogenic pressures consistent with
development and increasing human population, resulting in experimental
replanting projects that have met with mixed success. Replanters have
reached consensus that site selection is a crucial first step in such projects.
Site selection for eelgrass is often based on a prioritization matrix that
takes into account parameters such as wave energy and substrate.
Models of this kind have been used with success on the east coast of the
US, but not to date with replanting efforts in Washington State Puget
Sound. The Washington State Department of Natural Resources (WADNR) is currently inventorying eelgrass stocks and has created extensive
Geographic Information Systems (GIS) files detailing Puget Sound aquatic
conditions including substrate, shoreline modification, exposure, overwater
structures, and vegetation. The research described in this thesis used the
WA-DNR GIS files to produce maps of Puget Sound highlighting areas
that appear to be good candidates for eelgrass replanting projects. These
sites were chosen by substrate, exposure class, degree and type of
shoreline modification, and proximity to existing eelgrass beds. It was
found that of the total shoreline under consideration, 19.02% was already
populated with continuous eelgrass beds, and 5.55% was suitable for
replanting. These maps are preliminary and the recommended sites
require field testing before any replanting projects are commenced.

CONTENTS
LIST OF FIGURES
LIST OF TABLES
ACRONYMS AND ABBREVIATIONS USED
ACKNOWLEDGEMENTS

vi
vii
viii
ix

I. INTRODUCTION

1

II. BACKGROUND

3

SIGNIFICANCE OF EELGRASS
CAUSES OF EELGRASS DECLINE
EELGRASS REPLANTING TECHNIQUES
HISTORICAL EELGRASS PRESENCE IN PUGET SOUND
ESTIMATIONS OF CURRENT EELGRASS EXTENT
PREVIOUS SITE SELECTION MODELS
LIMITING FACTORS FOR EELGRASS GROWTH
Current and seed dispersal
Wave energy
Light and exposure
Temperature and salinity
Substrate
III. METHODS
AN EELGRASS SITE SUITABILITY MODEL
GEOGRAPHIC INFORMATION SYSTEMS (GIS)
SHOREZONE AND SVMP
The ShoreZone Inventory
PSAMP and SVMP
MANIPULATION OF DATA SETS
Limiting the ShoreZone data set to the regions
included in the SVMP data set
Data Normalization
FACTORS CONSIDERED FROM SHOREZONE
Eelgrass presence
Substrate type
Shoreline modification
Exposure class
Overwater structures
CREATION OF SITE SELECTION OUTPUT
Applying the model parameters

3
4
5
6
7
8
9
10
10
11
13
13
15
15
16
17
17
19
20
20
21
22
22
23
25
26
27
30
33

iv

IV. RESULTS
MODEL OUTPUT OF POTENTIAL REPLANTING SITES
Central Puget Sound
North Bremerton
Hood Canal
Quilcene-Poulsbo
North Puget Sound
Bellingham area
San Juan Islands – Strait of Juan de Fuca
San Juan Islands
Saratoga Passage – Whidbey Basin
Everett area

34
34
36
37
38
39
40
41
42
43
44
45

V. CONCLUSIONS AND DISCUSSION

46

REFERENCES

51

v

LIST OF FIGURES
FIGURE

TITLE

PAGE

1

Estimated depth profiles for Puget Sound eelgrass
based on 2002-2004 DNR Submerged
Vegetation Monitoring Project data

12

2

The six zones determined by the DNR’s Submerged
Vegetation Monitoring Project

20

3

Percent continuous eelgrass presence (all regions)
by substrate type

24

4

Bainbridge Island marina showing overwater
structures

29

5

Bainbridge Island marina as represented by GIS
model

29

6

ArcGIS 9.3 ‘Select by Attributes’ tool

32

7

Priority areas for Central Puget Sound

36

7a

Priority areas for North Bremerton

37

8

Priority areas for Hood Canal

38

8a

Priority areas for Quilcene-Poulsbo

39

9

Priority areas for North Puget Sound

40

9a

Priority areas for Bellingham area

41

10

Priority areas for San Juan Islands – Straits

42

10a

Priority areas for San Juan Islands

43

11

Priority areas for Saratoga Passage – Whidbey Basin

44

11a

Priority areas for Everett area

45

12

Area in Bremerton showing model-selected
‘optimal’ site (blue)

47

13

Erlands Point

48
vi

LIST OF TABLES

TABLE

TITLE

PAGE

1

An example of some of the data available for
each of the 7365 ShoreZone units

18

2

Normalized data indicating eelgrass preference
for protected exposure class

27

3

Continuous eelgrass and optimal replanting area
by region

35

vii

ACRONYMS AND ABBREVIATIONS USED

GIS

Geographic Information Systems

GPS

Global Positioning System

MLLW

Mean Lower Low Water

NAIP

US Department of Agricultural National Agriculture Imagery
Program

PSAMP

Puget Sound Assessment and Monitoring Program

SAV

Submerged Aquatic Vegetation

SVMP

Submerged Vegetation Monitoring Project

WA-DNR

Washington State Department of Natural Resources

WDFW

Washington Department of Fish and Wildlife

viii

ACKNOWLEDGEMENTS

I am indebted to the assistance and encouragement of several
members of the WA-DNR’s Nearshore Habitat Program. Helen Berry,
Pete Dowty and Jeff Gaeckle all gave generously of their time and energy
to support my work. Helen welcomed me into the program as a volunteer,
made sure I had everything I needed, and suggested several projects I
could develop using their data. Pete’s work with eelgrass suggested the
idea that eventually became this thesis and he was tireless in providing
me with whatever data I requested (as well as helping me navigate the
Department’s servers). Jeff’s experience with previous eelgrass replanting
projects and site selection models proved invaluable and his humor
helped me believe the project was possible. I remain touched and
impressed at the lengths these three busy people went to in encouraging
a student’s work.
I would also like to thank my reader, Judith Cushing, for her early
enthusiasm for the work and many detailed corrections, both of a scientific
and a grammatical nature. I always enjoyed my meetings with her.
Finally I would like to thank my wife, Leslie Wolff, for completing her
thesis the previous year. Having observed the experience at first hand I
was able to approach my own with slightly less fear.

ix

I. INTRODUCTION

Zostera marina, familiarly known as eelgrass, is the most common
and widespread species of approximately 60 seagrasses found in nearshore coastal environments around the world. Seagrasses form meadows
via horizontal rhizomes and grow leafy shoots vertically into the water
column. All seagrasses with the exception of surfgrass (Phyllospadix
torreyi, which adheres to rocks) grow in soft substrates, which they
stabilize with their rhizomes. These meadows play a significant role as
habitat and food for many species, including commercially important fish
and crustacean species. Eelgrass meadows are in worldwide decline due
to anthropogenic stress factors.
As eelgrass is sensitive to many stressors associated with the
development that has taken place since European settlement, the
Washington State Department of Natural Resources (WA-DNR) has made
it a priority to steward the health of Puget Sound by using eelgrass as one
of the top five indicator species. Although there is little data concerning
historical eelgrass extent, researchers suspect that eelgrass habitat has
been degraded by development, agricultural run-offs, increased boat
traffic, changing oceanic and climate conditions, and introduction of
invasive species (Thom et al. 2008). Debilitated eelgrass meadows affect
the survival of many species that depend on eelgrass for food, shelter, or
habitat and represents a significant threat to the health of Puget Sound.

1

Attempts to restore eelgrass meadows have met with mixed
success as its habitat is complex and affected by many different factors.
Given that the reason for the cause of the original loss is known and has
been corrected, the scientific community has generally concluded that
restoring eelgrass meadows is possible but that restoration efforts are
hampered by large knowledge gaps. One of the main factors that will
determine the success of any restoration attempt is an understanding of
what natural conditions influence eelgrass success and how these
condition can be used to predict suitable habitat for restoration. As
Fonseca et al. (1998) suggest in Guidelines for the Conservation and
Restoration of Seagrasses, replanting is not technically complex but
“planting will not succeed unless managers appreciate and emphasize the
extreme importance of site selection.”
This work uses geographic data from WA-DNR and Geographic
Information Systems (GIS) to prioritize potential replanting locations in the
Puget Sound with the goal of increasing the chance for replanting success.
Section II, Background, discusses current and historic eelgrass extent,
factors limiting eelgrass growth and survival, and previous replanting
efforts and site selection models. Section III, Methods, describes a model
developed using WA-DNR GIS data to select optimal sites for restoration
based on a number of specific metrics (variables that characterize a place).
The maps generated by this model are presented in Section IV, Results,
and the limitations of the model are discussed in Section V, Conclusions.

2

II. BACKGROUND

SIGNIFICANCE OF EELGRASS

Eelgrasses and seagrasses provide extensive nearshore habitat
around the world. In Washington State eelgrass provides nursery and
spawning habitat for Pacific herring and salmon, feeding and foraging
habitat for waterbirds, and also acts to improve water quality and prevent
erosion by stabilizing sediment. Eelgrass meadows are also a sink for
nutrients and shelter for many valuable species such as Dungeness crab
(Dowty et al. 2010).
Although nature should not be viewed solely in terms of its uses for
humankind, eelgrass meadows hold a significant economic value.
Eelgrass meadows are an aquatic net primary producer, providing food for
the marine environment and the secondary production of fish species, and
thus ultimately sustain humans. An Australian study found that, in terms of
catch reduction, the loss of just 16% of seagrass in one fishing block (an
area corresponding loosely to 1° latitude and longitude) resulted in an
economic loss of A$235,000 per year, and that the relationship would
conceivably arrive at a ‘catastrophic’ point if habitat loss continued
(McArthur and Boland, 2006). In short, even relatively small losses of
eelgrass and seagrass habitat can have significant impacts on the species

3

composition of the nearshore habitat, and negatively affect the fishing
industry.
Washington State recognizes the significance of eelgrass meadow
health and affords meadows special protection through the Department of
Fish and Wildlife (WDFW). In 2010 the Puget Sound Partnership created
the Dashboard of Ecosystem Indicators intended to estimate the health of
the Puget Sound by monitoring twenty species, of which eelgrass was one
of the top five (Puget Sound Partnership 2010). The Washington State
Department of Natural Resources subsequently recommended that the
Puget Sound Partnership adopt a target of a 20% increase in eelgrass
meadow areas by 2020. Replanting is one of the strategies recommended
to achieve this goal.

CAUSES OF EELGRASS DECLINE

With the increase in human population on the shorelines of the
Puget Sound it is inevitable that nearshore resources become stressed.
Due to relatively high light requirements eelgrasses thrive in shallow
nearshore waters which makes them especially vulnerable to damage by
human activities (Fonseca et al. 1998). Logging and agriculture has led to
increased runoff, siltation, increased turbitidy, and loss of water quality
that all restrict eelgrass growth. Other activities such as the dredging and
filling required to maintain shipping lanes and the construction of coastal

4

armoring and overwater structures such as docks and bridges also impact
on eelgrass habitat.
Aside from anthropogenic influences eelgrass also suffers from a
periodic wasting disease, first documented in the 1930s when it virtually
eliminated the species in the North Atlantic. This disease is caused by a
pathogenic strain of Labyrinthula and has been isolated in the Puget
Sound, although it has yet to cause a mass dieoff in that area (Short et al.
1987). The slowness of the recovery of North Atlantic eelgrass indicates
that natural recruitment of eelgrass does not keep pace with population
mortality that can occur very rapidly (Fonseca et al. 1998).

EELGRASS REPLANTING TECHNIQUES

The basic rationale behind replanting is to adjust the ratio of
recruitment to mortality and thus effect net eelgrass population growth.
Replanting eelgrass, however, is not a simple matter. Eelgrass is aquatic
and replanting is typically an expensive and labor-intensive process
involving boats and SCUBA divers. Shoots are fragile and very
susceptible to physical damage and must be kept wet during the entire
period between collecting and replanting, which should ideally take place
on the same day (Fonseca et al. 1998).
Existing methods for eelgrass replanting typically involve fastening
collected shoots to the substrate with either bamboo staples or temporary

5

metal frameworks. When this is performed by SCUBA the cost of
replanting increases significantly. Restoration dollars, with eelgrass as
with many species, are scarce, and it is important to maximize their value
by doing everything possible to ensure replanting success. Good site
selection is thus a crucial first step.

HISTORICAL EELGRASS PRESENCE IN PUGET SOUND

Eelgrass restoration in the Puget Sound differs from restoration
efforts in some other parts of the world (notably the Eastern US coast) due
to a comparative paucity of data concerning historical eelgrass meadow
extent. As Dowty et al. (2010) point out, historical abundance is an
appropriate reference point for setting management targets for future
abundance, as well as suggesting whether management practices should
focus on restoration of lost meadows or protection of remaining vegetation.
Since eelgrass is affected by a wide range of stressors it seems
reasonable that increased human activity in the Puget Sound has resulted
in eelgrass loss from the factors considered above; however, limitations in
the historic record make the extent of this hypothetical loss difficult to
estimate.
The earliest records of eelgrass in the Sound are from 19th century
hydrographic charts and are limited by the scale of the charts and the fact
that eelgrass, not being a navigational aid or of economic importance, was

6

not of great concern at the time (Thom & Hallum 1990). In 1962-3 Ron
Phillips used divers and boats to conduct a survey of eelgrass density at
107 sites throughout the Puget Sound. In 1974 he estimated that 9% of
the lower Puget Sound photic zone below mean lower low water (MLLW)
was covered with eelgrass (Phillips 1984).

ESTIMATIONS OF CURRENT EELGRASS EXTENT

In 1998, Bailey et al. used a probability model, based on 325
randomly selected sites along 3715 km (2303 mi) of Puget Sound
shoreline, to estimate that 23.4% of the total shoreline was then vegetated
with eelgrass (Bailey et al. 1998). This report covers an area similar to this
thesis but also includes the South Puget Sound, where eelgrass is known
not to occur for reasons of tidal range (Section: “Light and exposure”).
Because of this tidal range, the South Puget Sound was eliminated as a
location suitable for eelgrass restoration in this thesis.
The WA-DNR Submerged Vegetation Monitoring Project (SVMP),
discussed in greater detail below, was established in 2000 and has since
conducted a yearly survey of eelgrass extent. The 2009 report (which
contains the 2008 data) estimate a Sound-wide eelgrass area of 22,800 ±
4,500 ha (± 95% CI) for the zones covered in the report (Gaekle et al.
2009).

7

PREVIOUS SITE SELECTION MODELS

Thom et al. (2008) summarized and evaluated all previous
restoration efforts in the Pacific Northwest and included a synthesis of
what researchers have learned about the process to improve project
success. They found it difficult to summarize the relative performance of
the more than 30 projects due to the wide variety in replanting techniques,
project size, performance criteria, duration of monitoring, and project goals.
Most projects were conducted as mitigation to compensate for shoreline
development, and in all cases areas replanted shrunk in subsequent years,
resulting in a net loss of habitat (Fonseca et al. 1998). Thom et al.
concluded eelgrass restoration science is hampered by large knowledge
gaps and that good site selection was of extreme importance.
The Judd et al. (2009) research into eelgrass restoration on the
Lower Columbia River estuary used baseline in-situ field tests combined
with satellite observations to determine ambient habitat conditions. Their
measurements covered salinity, temperature, current velocity, light
availability, wave energy, and desiccation to predict the suitability of an
area for eelgrass replanting. Based on this model five areas were planted
with eelgrass. One year later two of the five sites had good survival rates,
two had poor survival, and one had total eelgrass loss. They concluded
that this 40% survival rate represented reasonable success by restoration

8

standards, although the labor required to obtain the site selection
measurements was significant.
On the east coast of the United States Short et al. (2002)
developed a site selection model and later generated a CD-ROM that
takes field data entered by the researcher and produces a Geographic
Information Systems (GIS) map of recommended suitable areas. The
parameters considered were historical eelgrass distribution (from maps),
current eelgrass distribution, proximity to natural eelgrass beds, sediment,
wave exposure, water depth, and water quality. When using this model,
replanting efforts showed a 62% success rate which was approximately
double that reported by previous restoration efforts in the area. As with
Judd et al., the model requires field tests be made as a necessary starting
point. This model is not applicable to the west coast due to the
comparative paucity of data concerning historical eelgrass extent.
Although the model produced by Short et al. inspired the current work, this
thesis differs in that it begins with existing GIS data sets which are to be
supplemented later with field tests once inappropriate sites have been
eliminated as a first step.

LIMITING FACTORS FOR EELGRASS GROWTH

To evaluate the success of any eelgrass restoration project, we first
must understand the health and succession of eelgrass in an undisturbed

9

state. At one location we might see a lush eelgrass meadow, but note that
a nearby location with apparently the same depth range, water
temperature, salinity, substrate, and turbidity, is barren. Before deciding
that the second location is ideal for a replanting project, we must pose the
question as to why eelgrass did not naturally recruit to this site.

Current and seed dispersal. Studying seed bank patterns in
Chesapeake Bay, Harwell and Orth (2002) found that the number of viable
seeds showed high variability both between and among zones sampled,
with seeds found in sites not displaying any Z. marina shoots as well as in
mixed species and Z. marina monospecific sites. The number of
reproductive shoots was also highly variable, probably due to different
local environmental conditions (Harwell 2002).
In their 2009 study of eelgrass restoration in the lower Columbia
River estuary, Judd et al. found that eelgrass distribution may be limited
by poor seed distribution, particularly in areas with a pronounced current
(such as a river estuary). In other words, for eelgrass to colonize an area,
water currents must be capable of dispersing the seeds to that area.
Variability in local water currents may be one reason why a site that
otherwise seems suited for eelgrass presence is unvegetated.

Wave energy. Eelgrass exists within a specific range of wave
energy and tidal current speed (Murphy and Fonseca, 1995). Eelgrass

10

seems to prefer a certain amount of water mixing to obtain oxygen and
nutrients and thus avoids completely protected coves and bays. On the
other hand, water energy is correlated with sediment stability (Fonseca et
al. 1998) and at a certain threshold eelgrass will not be able to survive due
to erosion and shoot burial.

Light and exposure. Eelgrass habitat is constrained to a depth
gradient that represents at its upper boundary the likelihood of exposure to
desiccation at low tide, and at its lower boundary light attenuation in the
water column. Variations in tidal range and light availability account for the
variety exhibited in eelgrass range between different species and sites
(Krause-Jensen et al. 2000, Herb and Stefan 2003). This range is usually
determined at the upper boundary by the mean lower low water (MLLW)
mark of the Puget Sound’s two tides. Phillips found that while vegetative
growth was observed from 1.8 m above MLLW to 30 m deep, the optimum
range for reproductive and vegetative activity was from MLLW to 6.6. m
below (Phillips 1984). A more detailed survey by the WA-DNR’s ongoing
Submerged Vegetation Monitoring Project (SVMP) has found that
eelgrass depth range varies throughout the Sound (Fig. 1).

11

Fig. 1. Estimated depth profiles for Puget Sound eelgrass
based on 2002-2004 Submerged Vegetation Monitoring Project data

Not all locations in Puget Sound are suitable for eelgrass growth.
Southern Puget Sound is generally considered unsuitable habitat for
eelgrass due to a large tidal range which at one extreme exposes eelgrass
to desiccation and at the other extreme reduces light availability to
unacceptable levels (Dowty 2011).
Overwater structures such as bridges, docks, piers, floats, and
miscellaneous buildings cover large sections of shoreline in the Puget
Sound. Any part of the shoreline that is in permanent shadow from an
overwater structure will be unsuitable for photosynthesis and thus
unsuitable for eelgrass growth.

12

Temperature and salinity. As with depth, eelgrass thrives best
within a specific range of temperatures and salinities. A study of eelgrass
restoration in the Columbia River estuary, where during low-flow
conditions salinity intrusion can occur many miles upstream, found that
eelgrass experienced optimal conditions within the salinity range 10-30 ppt
(Judd 2009). This agrees with Phillips (1972) which showed that any
salinity lower than 10 ppt resulted in stunted growth. Eelgrass will not grow
in persistent fresh water (Phillips 1974).
Eelgrass species tolerate a wide variety of temperatures worldwide,
from -6°C (21.2°F) in Alaska to around 27°C (80.5°F) in the Gulf of
California, Mexico; however, there is some evidence that specific
genotypes evolve with different temperature requirements determined by
location. Thus temperature may affect the availability of transplants; the
optimal temperatures for reproductive growth in the Puget Sound occurs in
the temperature range 6°C – 12.5°C (42.8°F – 54.5°F) (Phillips 1984).

Substrate. Substrate is a strong factor influencing eelgrass
success. All seagrasses, with the exception of surfgrass (Phyllospadix,
which attaches to rocks) grow in unconsolidated substrates ranging from
gravelly sand to fine muds and silts, with a general preference towards
finer particle sizes (Kenworthy et al. 1977). Depth of sediment is also a
factor: bedrock too near the surface (which might be exposed by currents)

13

limits the distribution of some seagrasses (Fonseca et al. 1998). For this
reason eelgrass tends to prefer a fairly protected level of wave exposure.
Substrate can also be a success factor for a restoration project.
There are several approaches to anchoring the transplanted shoots to the
substrate which have been used with varying degrees of success. Kopp
and Short (2001) found in a study in New Bedford, MA, that a technique
where eelgrass rhizomes were ‘stapled’ to the substrate with bamboo was
less successful than a method where transplants were secured to the
ocean floor by a metal frame for a period of one month. It was
hypothesized that burrowing fauna such as crabs, which use eelgrass for
shelter, dislodged the bamboo-stapled rhizomes from the loose sediment.

14

III. METHODS

AN EELGRASS SITE SUITABILITY MODEL

Puget Sound is a large and complex estuarine system of many
interconnected waterways and significant variability in depth, tidal range,
substrate, and development. In the current work potential eelgrass
replanting sites were selected by examining available Geographic
Information Systems (GIS) data on these variables and constructing a set
of tables that included only those locations that matched a specific set of
criteria.
This thesis aims to demonstrate how GIS can be used as a first
step in selecting areas for eelgrass restoration in Puget Sound. As
indicated above, a great many factors influence eelgrass success, and
available GIS data sets do not enough information to select a site without
additional field experimentation. It is unlikely that such a model could exist,
given the sheer area under consideration and the necessary
simplifications of a tabular data set. However, where pertinent information
has been recorded, this information can be used to eliminate a great many
sites on the basis of unsuitable substrate, for example. In this way the
work differs from Short and Burdick’s computerized site selection model
for the New Bedford Massachusetts area, which requires inputting field
measurements to calculate site suitability. Recall from Section II that Short
15

and Burdick produced a program where the users enter field
measurements for salinity, turbidity, and fetch, and then see a GIS map
which highlights likely eelgrass restoration sites.
In this thesis, the GIS output is based solely on the existing
ShoreZone data set produced by the Nearshore Habitat Program of the
WA-DNR, and requires field measurements of unconsidered factors (such
as salinity and turbidity) be taken after the fact. In both models the aim is
the same: to prioritize likely eelgrass restoration locations in order to
facilitate restoration decisions that must be made with limited budget and
resources.

GEOGRAPHIC INFORMATION SYSTEMS (GIS)

Geographic Information Systems (GIS) are a powerful tool for
manipulating and analyzing spatial information. The great advantage of
GIS is that data sets containing multiple attributes can be presented with
reference to geographic locations. It is comparatively straightforward,
therefore, to create a prioritization matrix based on a set of defined
parameters and link the output to a map which can be easily visually
interpreted.

16

SHOREZONE AND SVMP

This thesis builds its site-selection model primarily on data available
from two publically available WA-DNR GIS data sets: the ShoreZone
Inventory and the annual SVMP report. Data from the two data sets were
combined in ArcGIS 9.3 and analyses and statistics were performed with
Microsoft Excel. The SVMP data set is used only to delineate the regions
in Puget Sound that are considered for this thesis. All of the data on local
conditions, substrate, eelgrass presence, and shoreline modification are
contained within the ShoreZone Inventory.

The ShoreZone Inventory. The Nearshore Habitat Program of the
WA-DNR has produced a large GIS data set, known as ShoreZone, which
contains an inventory of Washington’s saltwater shorelines from the
Canadian border to the Columbia River. This data set, compiled from data
gathered over the period 1994-2000, contains information concerning
shoreline morphology, substrate, wave exposure, and biota.
The ShoreZone Inventory divides the saltwater shore of
Washington into 7365 individual units of approximately 0.5 miles in length
where the primary geomorphology is consistent. A unit might be thus
classified as a gravel beach, and abut a unit classified as a mud flat (or
another gravel beach). The longest unit is 2.38 miles, while the shortest is
59 feet (Berry et al. 2001).

17

Each unit is characterized as possessing one of fifteen shoreline
types, seven substrate types, and six wave exposure classes. Further
information on each unit includes percentage of anthropogenic
modification, primary, secondary, and tertiary kinds of modification, and
degree of presence of eelgrass, seagrass, surfgrass, kelp, sargassum,
dunegrass, and saltmarsh. Table 1 presents an example, in tabular form,
of some of the data an individual unit might contain.
Attribute

Unit

Unit ID

2646

Length (ft)

1563.668

Shoreline Type

Sand flat

Substrate Type

Sand

Shoreline Modification

90%

Primary Modification

Wooden bulkhead

Secondary Modification

Rip Rap

Tertiary Modification

None

Exposure Class

Semi-protected

Surfgrass

Absent

Eelgrass

Continuous

Kelp

Absent

Sargassum

Absent

Dunegrass

Absent

Salt Marsh

Absent

Table 1. An example of some of the data available
for each of the 7365 ShoreZone units

18

PSAMP and SVMP. In 2000, as part of its work with the multiagency Puget Sound Assessment and Monitoring Program (PSAMP), the
Nearshore Habitat Program of the WA-DNR created the Submerged
Vegetation Monitoring Project, or SVMP. The intention of the SVMP is to
monitor and track the health and extent of eelgrass using a statistically
robust sampling design and underwater videography (Graekle et al. 2009).
The SVMP provides both Sound-wide and regional data, dividing the
Sound into six zones: North Puget Sound, San Juan Islands- Strait of
Juan de Fuca, Central Puget Sound, Hood Canal, Saratoga PassageWhidbey Basin, and Southern Puget Sound (Fig. 2).
Because tidal ranges and light availability vary within Puget Sound,
each of the six SVMP zones under consideration demonstrates a different
depth gradient where eelgrass is found. Southern Puget Sound, for
example, due to extreme tide changes which desiccate eelgrass at low
tide and place it outside the range of light it requires at high tide, does not
support eelgrass except in very rare circumstances (Dowty 2011). The
ShoreZone data list no eelgrass presence at all for the South Puget Sound.

19

Fig. 2. The six zones determined by the DNR’s
Submerged Vegetation Monitoring Project

MANIPULATION OF DATA SETS

Limiting the ShoreZone data set to the regions included in the
SVMP data set. As previously stated, the ShoreZone data set includes all
of Washington’s saltwater shorelines from the Canadian Border to the
20

mouth of the Columbia River. In order to provide statistically meaningful
data for eelgrass presence in the Puget Sound alone, all data referencing
regions outside of Puget Sound were eliminated from the ShoreZone
inventory. This was done with ArcGIS 9.3 by clipping the ShoreZone data
to areas that fell within the regions delineated by the SVMP data set. After
the irrelevant regions had been clipped from the ShoreZone Inventory,
6460 of the original 7365 units remained for analysis. All further analysis
was performed on the ShoreZone data and no other data from the SVMP
data set was required.

Data normalization. Before performing any statistics on the
ShoreZone features of eelgrass presence or absence, substrate, wave
exposure, and shoreline modification, we accounted for the relative
occurrence of each feature. For example, a key indicator for eelgrass
success is substrate. ShoreZone lists seven substrate categories (more
on this below) but each substrate is not equally abundant in Puget Sound.
When correlating Continuous eelgrass presence to substrate (to
determine if eelgrass shows a significant presence for a substrate type)
the ratio was based off the relative abundance of the substrate and not the
actual number of counts of that particular type.

21

FACTORS CONSIDERED FROM SHOREZONE

Not all data in ShoreZone was considered in the construction of this
model. Due to different substrate/ habitat requirements, surfgrass,
sargassum, kelp, and dune grass were not considered in competition for
space with eelgrass and were eliminated. Shoreline type was considered
less important than substrate type and was not considered. Pete Dowty at
the WA-DNR had previously combined eelgrass presence data with beach
width and found no clear association (personal communication, January
2011). In the end, it was decided that eelgrass success was to be
predicted using substrate type, percent and type of shoreline modification,
and wave exposure class.

Eelgrass presence. The ShoreZone inventory contains three
categories of eelgrass presence: Continuous, Patchy, and Absent. For this
study only Continuous eelgrass presence was considered, because it was
the strongest way to relate eelgrass health with the factors considered
(substrate, shoreline modification (extent and type), and exposure class).
Analysing the Absent category produces much the same data, only
inversely. That is to say, Continuous eelgrass is strongly correlated to a
sandy substrate and very weakly to a rocky substrate. Absent eelgrass
shows a strong correlation to a rocky substrate and a weak correlation to a

22

sandy substrate. It was decided to only use data that reflected eelgrass
success, as indicated by Continuous eelgrass presence.
The presence of Continuous eelgrass is used in two ways: to
determine existing correlations between eelgrass and environmental
factors, and to determine which locations have no need of replanting.
Clearly, if a unit shows Continuous eelgrass presence, restoration is not
required.
Proximity to existing Continuous eelgrass beds is also a factor to
consider in selecting a site for replanting. In their site-selection model for
eelgrass transplantation in the northeastern US, Short et al. include a
buffer of 100 m from natural eelgrass beds to insure that transplanting is
taking place outside an area that would otherwise be naturally revegetated
by seed dispersal (Short et al. 2002). Local field tests may indicate that
such seed dispersal is not possible due to water currents, but that data
must be collected in the field and is outside the scope of this model. It
should also be considered that proximity to donor beds can be an
important factor in the actual practical work of obtaining donor eelgrass
shoots, which should ideally be done on the same day as replanting to
prevent desiccation.

Substrate Type. Substrate type is the single strongest factor
included in the ShoreZone data set for eelgrass success. To some extent
this is self-explanatory: as stated above, all seagrasses with the exception

23

of surfgrass grow in unconsolidated substrates ranging from gravelly sand
to fine mud and silts. The ShoreZone data set includes seven substrate
classes. They are, ranging from coarsest to finest: Man-Made; Rock;
Gravel; Rock, Gravel, and Sand; Gravel and Sand; Sand; Mud and Fines.
By correlating the occurrence of Continuous eelgrass presence with the
occurrence of each kind of substrate and then normalizing for percentage
of each substrate type, eelgrass was shown to statistically demonstrate a
strong preference for Sand substrate type (Fig. 3). Therefore, only shore
units that had a Sand substrate class were considered as potential sites
for eelgrass replanting (even if, in the real world, it might have been
possible to replant in several of the other substrate types).
35.00%

30.00%

25.00%

20.00%

15.00%

10.00%

5.00%

0.00%
man-made

rock

gravel

rock, gravel,
and sand

gravel and
sand

sand

mud and fines

Fig. 3. Percent continuous eelgrass presence (all regions) by substrate type

24

Shoreline Modification. ShoreZone contains both ordinal and
nominal data concerning shoreline modification. The percent modification
per unit and the primary, secondary, and tertiary type of modification is
listed. The categories considered are Boat Ramp, Concrete Bulkhead,
Landfill, Rip Rap, Sheet Pile, Wooden Bulkhead, or None, presenting
some difficulties for analysis as the percent modification has different
significance for each type of modification, and the types of modification
vary widely in quantity. For example, in the North Puget Sound zone, there
are one hundred and sixty-four counts of rip rap (armoring) and only four
boat ramps. Continuous eelgrass occurs at fifty-nine of the sites with rip
rap (36%) but at three of those four with boat ramps (75%). This does not
imply that eelgrass shows a preference for boat ramps over rip rap.
The sample size can be increased by considering all regions
simultaneously, but even then chi-square tests show that eelgrass
demonstrates no significant preference for percentage of shoreline
modification. The data are just too general. Yet it is known that eelgrass
cannot survive in places shadowed by overwater structures (which inhibit
photosynthesis). Additionally, it has been established that coastal
armoring negatively affects habitat on sandy beaches by increasing
erosion and reducing beach width (Dugan et al., 2008). It is reasonable to
avoid areas that have been too heavily modified for replanting projects.
For the purposes of this model, a shoreline modification percentage of
35% or less was considered preferable. This does not take into account

25

type of shoreline modification (e.g. rip rap, boat ramps, etc) as the
available data is too broad to be of significance. It seems reasonable,
though, that a lower level of shoreline modification implies less general
anthropogenic disturbance and traffic.

Exposure Class. ShoreZone classifies all of the Puget Sound in
terms of six levels of exposure class: Very Exposed, Exposed, SemiExposed, Semi-Protected, Protected, and Very Protected. The class for
any unit was calculated by combining an exposure model that computed
fetch characteristics with wave exposure data determined on site by a
geomorphologist. In the 6460 units encapsulated by the SVMP zonal
regions, none are classified as Very Exposed, only one as Exposed, and
only 3 as Semi Exposed (all Very Exposed regions lie in the ShoreZones
on the Pacific Coast, which are out of the range of the SVMP regional
zones). Thus, for all intents and purposes, all of Puget Sound can be
considered to fall under the classifications of Semi Protected, Protected,
and Very Protected and severe wave energy can be discarded as a
limiting factor for this model.
With regards to eelgrass requiring a certain level of water
movement to facilitate nutrient and oxygen mixing and seed dispersal, the
data show eelgrass demonstrating an approximately three times greater
preference for Semi-Protected and Protected zones over Very Protected

26

(table 2). As before, this was normalized by figuring in the relative
abundance of each exposure class.

TOTAL COUNTS

ALL REGIONS

EELGRASS %

CLASS %

EXPOSURE CLASS

eelgrass

class

OF TOTAL

OF TOTAL

very exposed

0

0

0.00%

0.00%

exposed

0

1

0.00%

0.02%

semi exposed

3

253

1.19%

3.92%

semi protected

365

1814

20.12%

28.08%

protected

845

3447

24.51%

53.36%

very protected

68

945

7.20%

14.63%

Table 2. Normalized data indicating eelgrass preference
for protected exposure class

Overwater structures. Overwater structures such as bridges,
docks, piers, floats, marinas, floating homes, and miscellaneous buildings
cover large sections of shoreline in the Puget Sound. Any part of the
shoreline that is in permanent shadow from an overwater structure will be
unavailable for photosynthesis by eelgrass and thus unsuitable for
replanting. The WA-DNR has made available a shapefile of all overwater
structures in the Puget Sound at their GIS Data Centre website 1. This file
was digitized from 3-foot resolution color orthophotos taken between 2002
and 2006 by either the WA-DNR or the United States Department of

1

http://fortress.wa.gov/dnr/app1/dataweb/dmmatrix.html, retrieved January 2011

27

Agricultural National Agriculture Imagery Program (NAIP). The data is
classified so that overwater structures can be referenced based on
structure type or size.
Not all overwater structures will eliminate the possibility of eelgrass
replanting. Small, recreational family-use docks do not shade much area
and may even be helpful to a replanting team. Large industrial docks, on
the other hand, not only shade large areas but indicate substrate
disturbance and a hazard for replanters and should be eliminated. To
illustrate, Fig. 4 shows a Bainbridge Island marina. The overwater
structures are clearly visible. Fig. 5 provides the same information
depicted as a GIS output of the ShoreZone data. The GIS map shows that
both locations where continuous eelgrass is present (deep in the marina
and just outside the mouth) have relatively light or absent overwater
structures. If this area becomes a candidate for replanting, we would first
have to find units where there is currently no eelgrass and a sandy
substrate. These two locations are shown in blue in Fig. 5. However, it is
clear from both the photograph and the GIS map that these locations are
in areas of dense construction, and are thus unsuitable for replanting. Of
the other locations where eelgrass is absent, none have suitable
substrates. This marina is thus an unsuitable location for eelgrass
replanting.

28

Fig. 4. Bainbridge Island marina showing overwater structures. Image
courtesy of Google Earth, retrieved 3/01/11

Fig. 5. Bainbridge Island marina as represent by GIS model

29

The overwater structures data suffer from the same limitations as
the shoreline modification nominal data: they are too broad in nature to be
a strong indicator for eelgrass presence. The types of overwater structures
listed are Bridge, Building, Buoy/Float, Dock/Pier, Fill or Other. The size of
any listed structure is given in acres, hectares, and square feet and varies
from about 35 to 260,000 square feet. Attempting to correlate Continuous
eelgrass to both structure type and size would probably be unnecessarily
complex and fruitless. However, as seen in the Bainbridge Island marina
images above, it seems reasonable to assume that large docks over a
shore unit imply less light availability and more anthropogenic traffic and
disturbance. The data includes a “Complexity” category, which estimates
dock usage based on the size of the structure. Simple docks are
interpreted to mean small docks for family or recreational use, whereas
Complex docks are for community, commercial, or industrial use. For this
model, shore units with a Complex dock presence were eliminated for
consideration for replanting.

CREATION OF SITE SELECTION OUTPUT

ShoreZone data on Continuous eelgrass presence was correlated
to the factors of substrate type, exposure class, and shoreline modification
(extent and type). This was done by joining the eelgrass attribute table to
each of the above layers in order. The Select by Attributes tool was used

30

to create a selection linking Continuous eelgrass presence with each of
the above criteria both by individual SVMP region and all the regions as a
whole. While the entire region-wide data was sufficient to discover any
correlation, individual zones were considered for practical purposes and
the ability to practically represent the data on a map.
When the attribute tables were linked, simple if-then statements
were used in the Select by Attributes tool to return a count of units that fit
the criteria. For example, for the North Puget Sound SVMP zone, the
following statement was used to return a value of Continuous eelgrass
presence in areas with between 5% and 35% shore modification:

“eelline.EELGRASS” = “CONTINUOUS” AND
“shoremod.SM_TOT_PCT” > = 5 AND “shoremod.SM_TOT_PCT” < 36

where “eelline.EELGRASS” is the column in the eelgrass attribute table
that contains the values of Continuous, Patchy, or Absent, and
“shoremod.SM_TOT_PCT” is the column in the shoreline modification
attribute table that contains the percentage of total modification. The
previous statement returned 21 records out of 591, indicating that out of
591 North Puget Sound ShoreZone Units where shoreline modification
was between 5% and 35%, 21 out of 591 (3.55 %) showed continuous
eelgrass presence. Fig. 6 shows an image of the tool.

31

Fig. 6. ArcGIS 9.3 ‘Select by Attributes’ tool as used
to restrict records to those fitting the selection criteria

Continuous eelgrass presence for each of 6460 units was
correlated in this fashion to the following criteria from ShoreZone:

SUBSTRATE TYPE: Man-Made; Rock; Gravel; Rock, Gravel, and Sand;
Gravel and Sand; Mud and Fines

EXPOSURE CLASS: Very Exposed; Exposed; Semi-Exposed; SemiProtected; Protected; Very Protected

32

SHORELINE MODIFICATION (PERCENT): <5; 5-35; 36-65; 66-95; >95

Applying the Model Parameters. The criteria for determining
optimum sites for eelgrass replanting were as follows:
•

Absent existing eelgrass beds

•

Sand substrate

•

Shoreline modification >36%

•

Protected exposure class

•

No Complex large docks in the shore unit

A formula with these criteria was applied to every one of the 6460
ShoreZone shore units and the output was returned as ‘Optimal Sites.’
This is represented visually in the maps of each region given in Section IV.

33

IV. RESULTS

MODEL OUTPUT OF POTENTIAL REPLANTING SITES

The maps below were produced when the model was applied to the
five SVMP zones in question. Sites selected as optimal for potential
replanting are highlighted in bright blue. These sites show an ideal
substrate, low levels of modification, absence of commercial or otherwise
large overwater structures, no current Continuous eelgrass, and a
Protected exposure class.
Areas that display current continuous eelgrass presence are
highlighted in bright red for comparison. Aside from giving a visual
impression of the extent of eelgrass in the zone and the consequent need
(or lack of need) for replanting, proximity to existing eelgrass beds may be
a factor when eliminating potential replanting sites, as dense local
eelgrass might eliminate a site on the basis that natural recruitment is
likely. Conversely, replanting sites too far away from existing eelgrass may
cause logistical difficulties when it comes to obtaining donor shoots, which
should for preference be done on the day of the transplanting to prevent
shoot desiccation. For clarity, narrative interpreting each map has been
incorporated into the captions for each map.
Since the zone maps are of a fairly small scale, close-up maps of
areas of particular interest in each zone are included. These larger scale
34

maps include place names in order to aid location, though any mapping
program (such as Google Earth) should be sufficient to locate the
replanting zones.
Table 3 below summarizes the relative abundance of eelgrass by
region and the area available for replanting according to the model output.
Note that this represents a linear measure of eelgrass presence along the
shoreline and not an estimation of entire area colonized, and so cannot be
compared with the estimates of eelgrass area in Section II.

Length
Percent
Continuous
Continuous
Eelgrass
Eelgrass
(ft)

Length
Available
(ft)

Percent
Available

Region

Length of
shoreline
(ft)

CPS

3,856,624

523,304

13.57%

211,514

5.48%

HDC

1,284,400

440,463

34.29%

207,834

16.18%

NPS

1,314,357

438,784

33.38%

68,137

5.18%

SJS

3,551,901

537,441

15.13%

101,239

2.85%

SWH

1,786,296

303,704

17.00%

66,379

3.72%

All
Regions

11,793,578

2,243,695

19.02%

655,104

5.55%

Table 3. Continuous eelgrass and optimal replanting area by region

35

Fig. 7. Priority areas for Central Puget Sound
Since the southern area of this map has a high density of ‘optimal’ sites,
one must question why eelgrass has not naturally recruited to the area. Due to the
length of the inlets, it may be that the tidal range of the area is too extreme. This
could be determined by supplemental field tests, but in the meantime it might be

36

better to focus on a location with more of a balance between absent and
continuous eelgrass presence, such as the Bremerton area at the middle of the
zone. A close up of the area follows.

Fig. 7a. Priority areas for North Bremerton

37

Fig. 8. Priority areas for Hood Canal
Again we see a large open area in the southern part of Hood Canal which
appears to be ideal for eelgrass, as it may well be (the opposite shoreline
demonstrates considerable eelgrass presence). Of more interest is the northern
section where we see a good mix of available sites and donor sites.

38

Fig. 8a. Priority areas for Quilcene-Poulsbo
There is a good mix here of eelgrass presence and absence although care
should be taken to avoid replanting in areas that field tests indicate will likely
experience natural recruitment.

39

Fig. 9. Priority areas for North Puget Sound
North Puget Sound is already well populated with eelgrass and may not
need restoration. The dominance of continuous beds indicate natural recruitment
is likely. Some areas around Bellingham could pose potential sites.

40

Fig. 9a. Priority areas for Bellingham area
Large stretches near Gooseberry Point and Sand Point are good
candidates for eelgrass replanting. The sites near the Nooksack River mouth may
not be suitable due to the influx of fresh water lowering local salinity to
unacceptable levels.

41

Fig. 10. Priority areas for San Juan Islands - Strait of Juan de Fuca
In this map the chief area of interest is the San Juan Islands, which
experience a high degree of anthropogenic disturbance. There are no sites at all
along the top of the Olympic Peninsula.

42

Fig. 10a. Priority areas for San Juan Islands
There are relatively few available sites on the San Juan Islands, implying
that eelgrass has successfully recruited to all suitable areas already. Replanting
efforts should perhaps be directed elsewhere.

43

Fig. 11. Priority areas for Saratoga Passage - Whidbey Basin
Certain parts of this area are densely populated with eelgrass, leaving few
sites with ideal conditions. Again, natural recruitment might be the best option.
However, the area around Everett has some potential sites.

44

Fig. 11a. Priority areas for Everett area
Mission Beach, Kayak Point and Bretland all have stretches of
approximately one mile that could be good sites for replanting.

45

V. DISCUSSION AND CONCLUSIONS

This work prioritizes locations in the SVMP regions of Puget Sound
as the most suitable for further investigation for eelgrass replanting
projects. These areas are recommended to practitioners as areas where
transplanting resources should be focused. Because this model is based
on a single, albeit complex, data set, it cannot make any claim of
completeness without supplemental field data. Nevertheless, it provides a
valuable starting point by narrowing down the vast length of Puget Sound
shoreline into a few particularly likely locations. At this point, field research
as well as the common sense of the restoring team must come into play.
Prioritization models have been used with success in other parts of
the country, but Puget Sound still lacks a unified model. As with many
environmental restoration efforts, reestablishing eelgrass beds is a
political, economic, and scientific endeavor. Eelgrass restoration is a labor
intensive and expensive process, usually requiring many workers and
SCUBA divers, and restoration dollars are limited. What dollars are
available must be spent wisely. An effective prioritization model for
restoration site selection must be developed so that resources can be
efficiently distributed.
The output of the model demonstrates the location of sites that,
according to five basic but important parameters, appear to have nearideal conditions for eelgrass growth. The question that must follow is, why

46

has eelgrass not colonized these areas naturally? Only in some of the
cases can the absence of eelgrass be explained by anthropogenic
disturbance. Other limiting factors must be considered, such as
inappropriate tidal ranges. For example, fig. 12 shows a map of Erlands
Point in Bremerton where the ‘optimal’ site the model has selected for
eelgrass replanting is highlighted in blue. Although it is not in the range of
this map, continuous eelgrass grows in the area, indicating good general
conditions. Fig. 13 is a photograph of the same area. A quick visual
examination reveals this inlet is extremely shallow and thus experiences a
tidal range that would most likely result in eelgrass desiccation. It must be
eliminated from the model. Field experience is the best way to judge the
locations the model selects as ‘optimal.’ In this case, an inlet is rejected in
favor of a more open stretch of coastline.

Fig. 12. Area in Bremerton showing model-selected ‘optimal’ site (blue)

47

Fig. 13. Erlands Point. Image courtesy of Google Earth. Retrieved 3/11/11.
Image taken 7/9/07.

The findings of the model are preliminary at best and, as the above
example demonstrates, require field tests to determine their real-world
suitability in terms of tidal range, temperature, salinity, turbidity, as well as
accessibility, proximity of donor beds, and the terms of the local coastal
management plan. Additionally, as Thom et. al conclude, it is necessary to
understand the reason behind the initial absence of eelgrass and correct it
(Thom et al. 2008). The best current source for field test guidelines is the
Guidelines for the Conservation and Restoration of Seagrasses in the
United States and Adjacent Waters by Fonseca et al. (1998).

48

The ShoreZone Inventory is extensive but it was never intended to
be superior to site-specific surveys (Berry et al. 2001). Most of the
information contained within the data set was collected by helicopter using
video imagery with locational information (GPS). A geomorphologist and a
marine ecologist aboard the helicopter recorded continuous data on the
physical and biological features of the shoreline. The wave exposure was
estimated by combining the observed geomorphological data with a
computer model that returned a modified effective fetch based on the GIS
data of the shoreline characteristics. Eelgrass presence (classed as
Continuous, Patchy, or Absent) was determined by the marine ecologist
based on the aerial video. As such, the data produced is of a relatively
low-resolution status. As Berry et al. put it in The Washington State
ShoreZone Inventory User’s Manual, when determining what features
were included in the data, one should ask “Could I have seen the feature
from the window of a helicopter traveling at 60mph and 300 feet above the
ground?” (Berry et al. 2001)
Additionally, Puget Sound is a dynamic environment and the
ShoreZone data set is at least ten years old. A certain amount of shoreline
development, not included in the data set, will inevitably have occurred.
Beaches and sandbars shift over time, particularly in response to
shoreline armoring (Dugan et al. 2008). Eelgrass itself is not confined to
the lines delimited by a data set but recruits over areas naturally (and
existing beds die off).

49

GIS is a powerful tool for manipulating spatial data, but, like any
other system, is only as good or up to date as the data input to it. While
the ShoreZone data set is not continually updated, one of the great
advantages of using GIS in creating models of the kind described in this
thesis is that further surveys and updated data can be integrated into
existing maps as those data are completed and made available. Any two
data records that contain a common attribute, such as overlapping
geographical location, can be related in GIS and an analysis performed.
The model reported in this thesis could be made more effective by adding
tidal range or turbidity data, for example. As an ever-increasing wealth of
GIS data becomes available freely online from government agencies, a
restoration team could produce a study similar to this one by using similar
data and using this thesis as an exemplar. In that case, I would first
recommend checking what data are available from the local Department of
Natural Resources or equivalent body.
The Nearshore Habitat Program of the WA-DNR is currently (as of
2011) at work on a more comprehensive survey of eelgrass meadows
using underwater videography, but this work is not expected to be
completed for several years. Once it is complete we will have detailed
knowledge of the extent of existing eelgrass meadows and can plan our
restoration efforts accordingly, but that should not prevent us from acting
now to restore this crucial habitat using the data we have.

50

REFERENCES

SOFTWARE AND DATA
Berry, H.D., Harper, J. R., Mumford, Jr., T. F., Bookheim, B. E., Sewell, A.
T., and L.J. Tamayo. 2001. The Washington State ShoreZone
Inventory User’s Manual. Nearshore Habitat Program, Washington
State Department of Natural Resources, Olympia, WA.
ESRI. 2008. ArcGIS Desktop: Release 9.3. Environmental Systems
Research Institute, Redlands, CA.
Short, F. T. and D. M. Burdick. 2005. Eelgrass Restoration Site Selection
Model. CD-ROM and Manual. CICEET, University of New
Hampshire, Durham, NH.
Nearshore Habitat Program. 2001. The Washington State ShoreZone
Inventory. http://fortress.wa.gov/dnr/app1/DataWeb/dmmatrix.html,
retrieved 10/10/2010. Washington State Department of Natural
Resources, Olympia, WA.
2007. Overwater structures in marine waters of Washington State
(shapefile). Washington State Department of Natural Resources,
Aquatics Division, Olympia, WA.

JOURNAL ARTICLES
Dugan, J. E., Hubbard, D. M., Rodil, I. F., Revell, D. L. and S. Schroeter.
2008. Ecological effects of coastal armoring on sandy beaches.
Marine Ecology 29 (Suppl. 1): 160-170.
Harwell, M. C. and Orth, R. J. 2002. Seed bank patterns in Chesapeake
Bay eelgrass (Zostera marina L.): a bay-wide perspective.
Estuaries 25, no.6, part A: 1196-1204.
Herb, W. and H. Stefan. 2003. Integral growth of submersed macrophytes
in varying light regimes. Ecological Modelling 168: 77-100.
Kenworthy, W. J., and M. Fonseca. 1997. Reciprocal transplant of the
seagrass (Zostera marina L.) Effect of substrate on growth.
Aquaculture 12:197-213.
51

Krause-Jensen, D., Middelboe, A. L., Sand-Jensen, K. and P. B.
Christensen. 2000. Eelgrass Zostera marina growth along depth
gradients: upper boundaries of the variation as a powerful
predictive tool. Oikos 91, no. 2: 233-244.
McArthur, L. C. and J. Boland. 2006. The economic contribution of
seagrass to secondary production in South Australia. Ecological
Modelling 196: 163-172.
Murphey, P. L., and M. S. Fonseca. 1995. Role of high and low energy
seagrass beds as nursery areas for Penaeus duorarum in North
Carolina. Marine Ecology Progress Series 121: 91-98.
Short, F. T., Davis, R. C., Kopp, B. S., Short, C. A., Burdick, D. M. 2002.
Site-selection model for optimal transplantation of eelgrass Zostera
marina in the northeastern US. Marine Ecology Progress Series
227: 253-267.
Short, F. T., Muehlstein, L., and D. Porter. 1987. Eelgrass wasting disease:
cause and recurrence of a marine epidemic. The Biological Bulletin
173: 557-562.

TECHNICAL REPORTS.
Bailey, A., Berry, H. and B. Bookheim. 1998. Probability-based estimation
of nearshore habitat characteristics. Proceedings of Puget Sound
Research ’98 Conference, Seattle, WA. Washington State
Department of Natural Resources, Aquatic Resources Division.
Dowty, P. 2011. Depth-based estimates of potential eelgrass area in
Greater Puget Sound. Nearshore Habitat Program, Aquatic
Resources Division, Washington State Department of Natural
Resources.
Dowty, P., Berry, H. and J. Gaeckle. 2010. Developing indicators and
targets for eelgrass in Puget Sound: a science assessment.
Nearshore Habitat Program, Aquatic Resources Division,
Washington State Department of Natural Resources.
Fonseca, M. S., Kenworthy, W. J., and G. W. Thayer. 1998. Guidelines for
conservation and restoration of seagrass in the United States and
adjacent waters. NOAA/NMFS Coastal Ocean Program and
Decision Analysis Series, No. 12. NOAA Coastal Ocean Office,
Silver Spring, Maryland.

52

Gaekle, J., Dowty, P., Berry, H. and L. Ferrier. 2009. Puget Sound
Submerged Vegetation Monitoring Project: 2008 monitoring report.
Nearshore Habitat Program, Aquatic Resources Division,
Washington State Department of Natural Resources.
Judd, C., Roegner, C., Thom, R., Vavrinec, J., Borde, A., Yang, Z.,
Woodruff, D., Zhang, J. 2009. Eelgrass enhancement and
restoration in the lower Columbia River estuary. Pacific Northwest
National Laboratory, prepared for the U. S. Department of Energy.
Kopp, B. S., and F. Short. 2001. Status report for the New Bedford Harbor
Eelgrass Habitat Restoration Project, 1998-2001. NOAA Damage
Assessment and Restoration Program.
Phillips, R. C. 1984. The ecology of eelgrass meadows in the Pacific
Northwest: a community profile. U. S. Fish and Wildlife Service.
FWS/OBS-84/24.
Puget Sound Partnership. 2010. Strategic science plan. Prepared by the
Puget Sound Partnership Science Panel.
Thom, R., and L. Hallum. 1990. Long-term changes in the areal extent of
tidal marshes, eelgrass meadows and kelp forests of Puget Sound.
Final report to EPA, EPA 910/9-91-005, Fisheries Research
Institute, University of Washington, Seattle, WA.
Thom, R. Gaeckle, J., Borde, A., Anderson, M. Boyle, M., Durance, C.,
Kyte, M., Schlenger, P., Stutes, J. Weitkamp, D., Wyllie-Echeverria,
S. and S. Rumrill. 2008. Eelgrass (Zostera marina L.) restoration in
the Pacific Northwest: Recommendations to improve project
success. Pacific Northwest National Laboratory, prepared for the U.
S. Department of Energy.

53