-
extracted text (extracttext:extracted_text)
-
The Effects of Marine Reserves on Regional Groundfish Diversity within the San
Juan Archipelago, Washington
by
Kwasi Addae
A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2013
©2013 by Kwasi Addae. All rights reserved.
This Thesis for the Master of Environmental Studies Degree
by
Kwasi Addae
has been approved for
The Evergreen State College
by
________________________
Dr. Dina Roberts
Member of the Faculty
________________________
Date
ABSTRACT
The Effects of Marine Reserves on Regional Groundfish Diversity within the San
Juan Archipelago, Washington
Kwasi Addae
Abstract
Groundfish populations in the greater Puget Sound region have experienced
intense declines as the result of past commercial and recreational fisheries. In
recent decades mitigation efforts have involved the utilization of Marine Protected
Areas, such as marine reserves. Marine reserves have been shown to support
previously stressed groundfish populations by prohibiting the harvest of targeted
species within their boundaries. Marine reserves may also influence the
population structure and biodiversity of target species outside the protected
boundaries. This study investigates the regional effects of established marine
reserves on the biodiversity of groundfish within the San Juan Archipelago,
Washington. Fishery – Independent survey data was provided by the Washington
Department of Fish and Wildlife to analyze three groundfish groups: family
Gadidae (Cod), family Hexagrammidae (Lingcod and Greenlings) and the genus
Sebastes (Rockfish). Diversity levels for the three species groups were analyzed
using two diversity indices for temporal and spatial variations. Habitat and depth
preference was also examined to determine what species benefit from the
established marine reserves. Species normally associated with complex rocky
substrate were shown to significantly prefer that habitat. Thus suggesting that
established marine reserves are appropriately located for these targeted species,
and the implementation of new reserves should be considered. Significant
temporal variations in diversity levels were observed over the eight sampled
years, with a decrease in mean diversity levels. Spatial variations in diversity
were also observed in all three species groups, accurately describing changes in
groundfish population structures throughout the region. Areas of significantly
high or low diversity, however, had no correlation to the proximity of an
established reserve. The changes in groundfish population structure were unable
to be determined due to uncertainties in other variables. Due to the importance
placed on marine reserves within this ecosystem the significant variations in
biodiversity levels warrants continued monitoring.
Table of Contents
List of Figures
v
List of Table
vi
Acknowledgements
vii
Chapter 1: Literature Review
Fisheries Management
2
The Ecological Consequences of Fishing
6
Marine Protected Areas
10
Ecology of MPAs
15
MPA Design
21
Chapter 2: Manuscript
26
Introduction
Historical Fisheries Exploitation
29
Characteristics of Study Species
30
Study Area
34
Source of Data
35
Occurrence Rates and Temporal Variation
36
ROV: Species Composition and Habitat
40
Mapping and Spatial Analysis
42
Statistical Analysis
44
Trawl Frequency of Occurrence
46
Diversity and Evenness by Year
48
ROV: Habitat Preferences
52
Spatial Analyses
59
Methods
Results
iv
Discussion
Frequency of Occurrence
65
Measuring Diversity
67
Temporal Variations
70
Spatial Distribution
71
Biodiversity Monitoring
75
Significance of this Study
76
Continued Use of MPAs
78
Chapter 3: Conclusion
References
81
Appendix A
91
Appendix B
94
Appendix C
97
Appendix D
101
Appendix E
104
iv
List of Figures
Figure 1. Response ratios to reserve protection for six trophic groups
13
Figure 2. Response ration to reserve protection by maximum length
14
Figure 3. Relationship between recruitment and spawning stock
16
Figure 4. Estimated dispersal distance as a function of propagule duration 18
Figure 5. Models of population replenishment
19
Figure 6. Proper reserve placement for a reproductive stock
23
Figure 7. Equations for Shannon’s and Simpsons diversity indices
39
Figure 8. Clustering and dispersion example
43
Figure 9. Mean diversity value by index
50
Figure 10. Percentage of substrate by depth stratum
55
Figure 11. Species group frequency of occurrence per substrate type
57
Figure 12. Species group frequency of occurrence per habitat complexity 58
Figure 13. Species group occurrence by depth stratum
59
Figure 14. Lummi “Hot Spot” and San Juan “Cool Spot”
62
Figure 15. Cypress Island “Hot Spot ” and Non-Significant grouping
64
v
List of Tables
Table 1. WDFW marine reserves in the San Juan Archipelago
34
Table 2. Number of survey per year and sampling method
37
Table 3. Species observed by species group
38
Table 4. ROV Survey: observed species by species group
42
Table 5. Frequency of occurrence and abundance of each species group
47
Table 6. Variation of diversity values between sampling methods
48
Table 7. Shannon’s and Simpsons diversity: significant year pairings
49
Table 8. Significant year pairings by sampling method
51
Table 9. Correlation of depth to diversity values by sampling method
52
Table 10. ROV: Species group frequency of occurrence
53
Table 11. Habitat Substrate description
57
Table 12. Habitat Complexity level description
58
Table 13. Global Moran’s I: Diversity values
60
Table 14. Species abundance for the Lummi “Hot Spot”
61
Table 15. Global Moran’s I: Species per species group
64
vi
Acknowledgements
This research was supported by faculty of The Evergreen State College,
notably: Dr. Dina Roberts, Dr. Martha Henderson Director of the Master of
Environmental Studies program and Dr. Gerardo Chin-Leo. Data support was
provided by staff of the Washington Department of Fish and Wildlife: Dr. Dayv
Lowry, Robert Pacunski, James Selleck, and James Beam. Thanks to my
housemates, friends, and family who provided encouragement and motivation
throughout this experience.
Chapter 1:
Literature Review
Marine ecosystems comprise one of the largest, most dynamic and least
understood environments on the planet. Covering two-thirds of Earth’s surface,
oceans and seas consists of complex interactions between organisms, habitats, and
external forces to form diverse ecosystems (NRC 1998). Prior to the evolution of
Homo sapiens the biotic factors (e.g. species assemblages, predator prey
interactions, and primary production) and abiotic factors (e.g. ocean currents,
upwelling, and weather) were the primary governing forces (Roughgarden et al.
1998). However, humans should now be considered another factor influencing the
function and structure of marine ecosystems (NRC 1998). For several millennia,
human population growth, coastal development, industrialization, and more
recently the indirect impacts of climate change and ocean acidification have
increasingly put pressure on the marine environment and marine resources
(Hilborn and Hilborn 2012, Cooley and Doney 2009). One important and
widespread anthropogenic stressor is the direct exploitation and overharvest of
marine organisms, namely large fish.
Humans have generally been slow to reverse the effects of overfishing,
often waiting until fisheries threatened collapse (Myers and Worm 2003).
Historically, due to the vast expanse of our oceans, there was a belief that the
oceans represented an inexhaustible resource and source of food (Pauly and
Watson 2003). Overcapitalization of global fishing efforts, and unregulated open
access fisheries are only a few of the causes that have lead to the rapid decline of
1
global fisheries. There are, however, communities of researchers, policy makers,
management agencies, and fishing industry stakeholders who have been aware of
the overfishing issues for decades (NRC, 2004). Analyzing national and global
fishing data has led many scientists and managers to look for ways to offset the
damaging ecological effects of overfishing. Understanding the true magnitude of
overfishing may not be possible, because most of the depletion occurred during
early periods of exploitation, typically before data were collected on the fishery
(Myers and Worm, 2003). Mitigating the effects of overfishing and maintaining a
viable fishery, simultaneously, is challenging but necessary to provide support
coastal communities dependent on fish and to maintain marine ecosystem
function. In response to the growing concern to the state of fisheries, management
agencies have begun exploring alternative management practices to restore and
maintain these declining resources. The growing utilization of Marine Protected
Areas (MPAs) has been cited as a strategy positively affecting populations of
marine organisms (Soble and Dahlgren 2004, Palumbi 2001).
Fisheries Management
The problem of overfishing has been of interest to researchers, fisherman
and governments well before the modernization of fishing fleets and modern
fisheries management practices (Pinnegar and Engelhard 2008). Prior to modern
fisheries and management practices, community based management was the
normal means of managing a local fishery (Hilborn and Hilborn 2012). This put
the fishermen in direct control of their local resource. The conservation needs of
the resource were met in part through the intuition and knowledge of the local
2
peoples responsible for the harvest (Johannes et al. 2000). As fishing fleets have
modernized, so have management practices. In recent decades an increased
demand for fishery resources on areas well outside of coastal communities has led
to a departure from community based fisheries management (Jentof et al. 1998).
Overarching governmental agencies have adopted management roles with varying
degrees of success (Scheiber 2002). This has led to the identification of modern
fisheries management practices as the primary contributors to the depletion of
high trophic level large fish.
Most management practices use a single species approach, focused on
maintaining the harvest goal of a maximum sustainable yield (MSY) for a
particular target species. The contradiction between “maximum” and
“sustainable” make this concept controversial (Heneman 2002). The sustainable
harvest level of a species is a biological reference point that allows for
recruitment to replace the individuals of a species removed by fishing. In many
fisheries, analyses of stock assessments and catch data from previous years are
used in calculations for setting seasonal MSY; these calculations cannot possibly
account for fluctuations inherent with a dynamic population. This approach
maximizes short-term profits by assuming knowledge of a maximum sustainable
harvest level. An incentive is then placed on fishing directly to the maximum
yield. However, reaching MSY is not economically cost effective. Other
reference points such as, Optimal Yield (OY), have been employed concurrently
with MSY. The Magnuson –Stevens Fisheries Conservation Act (1976) defines
“optimum” as the amount of fish that “will provide the greatest overall benefit to
3
the Nation… taking into account the protection of marine ecosystems” [16 U.S.C
180(28)(A)]. Optimal Yield, unlike MSY, considers ecosystem, social, and
economic variables accounting for changes in fish stocks, optimizing the fishery
to maximize profit (Goldberg 2002). Thus OY should be much lower than MSY.
Failures with OY occur when governing agencies such as The National Marine
Fisheries Service (NMFS) permit fisheries management plans that set OY at
MSY, negating any positive effects of a lower catch (Goldberg 2002).
Other management strategies have focused on a “bottom up” approach that
targets individual fishing communities and fishermen, similar to past fisheries
management strategies (Jentof et al. 1998). In the United States, Individual
Fishing Quotas (IFQs) or Community Fishing Quotas (CFQs) assign a given
amount of the total allowable catch to licensed quota holders (NRC 1998). These
quota-based systems are generally established after a fishery has reached an
overfished status, and give exclusive rights to catch and sell (and to sell the right
to catch and sell fish) to those who have been most effective at catching fish
(Allison 2002). These methods do not target the cause of overfishing as limiting
catches with optimal yields and IFQs/CFQs only works when the society in which
the fishery is based is tolerant of the shifting management practices.
Unfortunately these management practices cannot mitigate against ecosystem
damaging effects caused by overfishing, nor have the capacity to be effective
when they are set under an overall maximum/optimum yield quota and thus do
not offer a solution (Macinko and Hennessey 2002).
4
Solutions to the overfishing crisis have been sought after using
reinventions of current fisheries management practices, however, these do not
address the root cause of the problem (Rosenberg 2003). Commonly utilized
modern fisheries management practices are not without scientific support. The
testing of alternative hypotheses and sensitivity analyses determine where
uncertainties in parameter estimations are likely, thus influencing managing
agencies decisions (NRC 2004). The problem is these same analyses now show
the true state of marine fisheries as one of political and/or fisheries
mismanagement and dwindling stocks. Although a solution to the overfishing
seems to be simple –a reduction in fishing efforts-, the approach to this is mainly
political with wide reaching ecological and socio-economic implications
(Rosenberg 2003). In response, a shifting trend towards adaptive ecosystem
based management has been observed through the use of alternative management
strategies such as MPAs, within the fisheries. Within an ecosystem based
management scheme, the science departs from analyzing a single population, and
looks for changes in the environment that would affect that population of interest
(Boehlert 2002). Within the fisheries, biotic and abiotic factors such as habitat
structure, biodiversity, and species interactions are examined giving scientists an
overwhelming amount of information to better inform alternative management
needs like the justification of more conservative management practices like the
use of MPAs. This information has also lead to better analysis methods that
examine the ecological, biological and social consequences of the overfishing
problem.
5
Analyzing the problems associated with overfishing begin with identifying
where population exploitation is occurring. Accomplishing this for individual
species requires distinguishing distinct population segments from one another and
how these segments interact. These units are called unit stocks, an idealized
discrete entity with its own demographics, and fate (Waldman 2005). Modern
fisheries science uses several techniques in the identification of fish stocks. These
vary in effectiveness, and the use of these techniques has increased as technology
has advanced and population dynamics theories have become integral components
of modern fisheries assessments for effective fisheries management (Begg and
Waldman 1999).
Fisheries managers often examine an exploited fishery as a simple system
of inflow, stock, and outflow. When the inflow (recruitment rate) is greatly less
than the outflow (harvest rate), the stock (harvestable fish) cannot be sustained
and will decrease over time. In this simple model, for a fishery to be sustainable,
the harvest rate must be small enough to allow for fluctuations in the recruitment
rate, and recruitment may be influenced by multiple factors, including mortality,
ENSO (El Nino Southern Oscillation), and other changing ocean conditions
(Bakun and Broad 2003). This would mean a comparatively small harvest rate.
Such a rate would likely not meet current economical demands of a fishery, even
if that rate were more ecologically sustainable.
The Ecological Consequences of Fishing
Marine ecosystems are as complex in function and structure as they are
diverse. In addition to proper stock identification, the simplest way to describe
6
the biota in such a diverse ecosystem is by assessing the feeding interactions
among the inhabiting organisms (Pauly et al. 2002). All organisms in a marine
ecosystem, ranging from benthic invertebrates to large apex predators, can be
represented by their trophic level. Trophic levels represent the number of steps an
organism is removed from primary production organisms such as algae and
bacteria (TL=1), and generally higher trophic levels are characterized by an
increase in body size, especially for piscivorous (fish eating) species (Pauly et al.
2002). Larger commercially important fish have mean trophic levels that range
from 3.0 to 4.5 (Pauly et al. 2002). By examining changes in mean trophic level,
it is possible to analyze the population structure for a given locality or habitat.
Other means of assessing population structure can be achieved through measuring
and monitoring the diversity including species richness and evenness of an area
(Magurran 2004). Both means of addressing population structure become
important when considering the effect fishing has on an ecosystem.
Fundamental causes for many of the changes in global marine ecosystems
have been attributed to overexploitation of several different fisheries (Tetreault
and Ambrose 2007). The act of harvesting fish via current fishing methods
removes a desired size or specific species (target species) from a local ecosystem,
thus effectively removing them from the food web. Most fisheries generally
target large, predatory high-trophic level fish species, as these are the species
most desired for human consumption (Myers and Worm 2003). These species are
generally long-lived slow growing fish that once mature, play intricate roles as
predators in their marine habitats. However, large fish are not always large.
7
Throughout various life stages these high trophic level organisms are preyed upon
by a vast variety of organisms (Pauly et al. 2002). So, the removal of high trophic
level fish not only alters predator-prey relationships changing the fish community
structure, but also the feeding ecology and mean trophic level of an entire
ecosystem (Sumaila et al. 2000, Pauly et al. 2002). What has been observed is a
global decline in mean trophic levels, correlating to the removal of large fish from
marine ecosystems (Pauly et al. 2002).
Other ecosystem damaging effects come from the gear used in certain
fisheries. A common method used to fish demersal species is bottom trawling.
Bottom trawling techniques consist of dragging a large net along the sea floor,
thus indiscriminately catching any organisms in the path of the net. The first
problem with these techniques is the incidental catch of non-target species, or bycatch. Trawling for one species often impacts many other species; however, bycatch is produced in nearly all forms of fishing (Palumbi 2001). Perhaps the most
devastating, long-lasting effect of bottom trawling is the plowing of the substrate
by the net. This action destroys critical habitat necessary for healthy benthic
organism communities by reducing the complexity of the sea floor, thus
eliminating microhabitats utilized by benthic organisms (including juvenile fish)
(Sumaila et al. 2000). These benthic communities are often highly productive,
comprised of low trophic level organisms, and form the base of marine ecosystem
food webs (Pauly et al. 2002).
The effects of fish harvesting can be observed at multiple scales, from the
individual species level, to impacts on populations, and at the ecosystem level.
8
Often, changes in community structure are the result of overfishing targeted
fisheries. Each species within the ecosystem may respond differently to changes
in population structure. One common effect seen is an increase in non-target
species populations, due to the lack of predation or competition (Myers and
Worm 2003). This may seem beneficial for those non-target species, but in
reality it can be detrimental. Because the food webs of a marine ecosystem are so
interwoven, a population increase of a previously suppressed species may lead to
a sudden crash of that population when the ecosystem cannot support increased
numbers (Pauly et al. 2002). Examples of this form of community restructuring
have been seen in both oceanic billfish and groundfish populations (Myers and
Worm 2003).
Another effect of fishing often observed is changes in fish physiology.
Since fisheries tend to select larger, fast-growing individuals from the fish
population, they run the risk of altering the genetic information thus changing the
evolutionary characteristics of that population (Pauly et al. 2002). Fishing can
therefore select against fish with slow maturation because these fish would not
have an opportunity to reproduce before being harvested. Research into this
phenomenon is limited, but research has shown the prevalence of early maturation
in targeted fish species (Kurlansky 1997). Earlier maturation may allow a female
fish to spawn sooner in life increasing the spawning potential; larvae from young
spawners, however, may experience lower survivability than larvae from older
spawning fish, negating any positive affect of increased spawning potential
(O’Farrell and Botsford 2006).
9
By recognizing the negative ecological impacts of overfishing, fisheries
managers can take action as needed to stem the problem. As mentioned above,
some forms of action have come by means of alternative management practices.
Though increasing in popularity, the application of Marine Protected Areas as
tools for conservation and fisheries management is still a relatively new strategy
(Sobel and Dahlgren 2004).
Marine Protected Areas
Marine reserves, a form of a Marine Protected Area (MPA), were
developed based on the idea of eliminating or limiting extraction of fish or any
other natural resource within the reserve boundaries (Palumbi 2001). Often
referred to as “No-Take” MPAs, Marine Reserves are receiving global attention
from fisheries managers, environmental groups, ecologist and various government
agencies as a means of conserving marine organisms and restoring depleted fish
stocks (Micheli et al. 2004). These reserves operate by protecting local fish stocks
as a management tool to augment or stabilize regional fisheries yields (Palumbi
2001). By virtue of their properties, no-take reserves also protect the ecosystem
functions of the habitat within their boundaries (Micheli et al. 2004).
Within the MPA category, there are several kinds of Marine Reserves.
Each type of reserve utilizes different management strategies to reach different
goals. This diversity of strategies and goals can be beneficial for management
agencies and biologists in that a reserve can be designed to meet specific
management needs. To maximize the benefits of this flexibility, it is necessary to
identify the specific conservation needs of a particular ecosystem in order to
10
effectively achieve protection. Often the predominate conservation goals of a
marine reserve stress the desire for protecting biological attributes of marine
ecosystems, such as providing critical habitat for a single or several species,
maintaining high biomass and species diversity, providing dispersal points for
propagules, and establishing refuges from fishing (Allison et al. 1998). Individual
reserves are unique in their topography and biodiversity. Where one is effective,
another may fail because each situation has unique qualities (Palumbi 2001). To
understand the overarching benefits offered by marine reserves this literature
review presents a broad comparison across reserves in different ecosystems.
From a fishery perspective marine reserves are a promising management
tool. They offer a fundamentally different type of protection not seen in
traditional fisheries management practices (Allison et al. 1998). By restricting the
access to critically important habitat, such as fish nurseries and spawning grounds,
marine reserves specify locations that can and cannot be fished (Hilborn et al.
2004). Protecting a population from fishing pressure often allows fish population
structure to be governed by natural mortality instead of fishing mortality (Allison
et al. 1998). Marine reserves worldwide have been shown to produce drastic
increases in biomass of species that are heavily fished outside reserve boundaries
(Palumbi 2004). What often accompanies this increase in biomass is a significant
increase in species diversity and population density. Even as benefits may vary
by geographic location, the general trend of increased abundance, biomass, and
diversity have been documented for a variety of targeted species (Micheli et al.
2004). A recent meta-analysis confirmed the benefits that marine reserves have
11
on fish populations, and concluded that average abundances of target fish inside
reserves were 3.7 times higher than outside reserve populations (Mosqueira et al.
2000).
Similar to fishing, protection within a reserve has the potential to modify
the community structure within a marine ecosystem. An increase in biomass of
predatory species within a reserve equates to an increase in predation on smaller
mid-trophic level species, which are often times not targeted by a fishery. In
general, non-target species do not demonstrate the same response (increase in
abundance, biomass, and diversity) from the protection of a marine reserve as
species targeted by fisheries (Figure 1) (Micheli et al. 2004). This is not to imply
that no benefits occur to non-target species. As previously discussed, fishing
practices have negative effects on the ecosystem as a whole, thus affecting more
than just the targeted fishery stock. Marine reserves offer protection to non-target
species by preventing their capture as by-catch and by reducing habitat
degradation that occurs during fishing (Allison et al. 1998).
12
Figure 1. Response ratios (ln R) vs. duration of protection for each of six trophic
groups. A statistical significant temporal trend was found only in piscivores.
Figure Source: Micheli et al. 2004
A large body of literature provides supportive evidence for the
development and implementation of marine reserves; however, individual species
response to protection varies depending on family association, trophic level,
whether or not the fish were the target of a fishery (level of exploitation), and
most importantly body size (Mosqueira et al. 2000). A strong correlation between
positive response and species with large body size has been shown (Figure
2)(Mosqueira et al. 2000). Moreover, Micheli et al. (2004) conducted a metaanalysis of 20 studies looking at the effects of reserves on community structure.
Their analysis showed that omnivores and other mid-trophic level species
responded poorly to reserve protection. This response was most likely explained
by the increase in high-trophic level organisms reported in the study (Palumbi
13
2004). Species benefiting most from protection are the same species most
susceptible to the effects of fishing: large, high-trophic level, long-lived, and slow
to mature predatory fish (Mosqueira et al. 2000).
Figure 2. Response Ratios by maximum length groups for (a) all species, (b)
Species that are target of fishing and (c) species that are not fished. Figure
source: Mosqueira et al. 2000
This has direct implications for marine planners establishing marine
reserves to conserve a fishery. Though marine reserves may initially reduce the
yield of surrounding fisheries by limiting their access to fish, the long-term
ecological benefits may ultimately benefit the fishery (Hilborn et al. 2004,
Sumaila et al. 2000). Maintaining a stable protected population of large fish
14
within a marine reserve may be extremely beneficial to the regional recovery of a
previously exploited fishery. Maximum body-size usually correlates to life
history parameters such as age at maturity, growth and reproductive output
(Mitcheli et al. 2004). As we have examined, these are all parameters selected
against when the species is targeted by a fishery. Larger body size also positively
correlates with fecundity and LEP (lifetime egg production), which increases
reproductive capacity (Plumbi 2004). Simply put, the larger the fish, the more
eggs produced, the higher recruitment rate, the greater the growth in abundance of
large fish in a region.
Ecology of MPAs
By allowing large fish to congregate, marine reserves ultimately support
more eggs, more larvae, and thus more adult fish to supply a neighboring fishery.
Three major underlying biological factors that make this possible: Lifetime Egg
Production (LEP), larval dispersal, and spillover. The first, Lifetime Egg
Production, is the number of eggs produced by a recruit over the course of its
lifetime (Botsford et al. 2009). Within a fishery stock LEP is directly correlated
with body size. Thus individual small-bodied early maturing fish selected against
in a fishery would produce substantially fewer eggs throughout their life
(O’Farrell and Botsford 2006). Lifetime egg production becomes very important
in quantifying the recruitment rate for a fishery. If the value of LEP drops below
the critical replacement value (the number of spawning recruits), the fishery will
destabilize and collapse (Botsford et al 2009). The longer female fish live, the
more clutches they bare, the higher the recruitment. Though, within a non-fished
15
population LEP is extremely difficult to estimate (Botsford et al. 2009). As to be
expected, an increase in fishing mortality decreases the LEP, hence researchers
rely on a ratio (Fractional Lifetime Egg Production or FLEP) between fished and
non-fished LEPs to assess the spawning potential ratio (SPR) of a fishery
(Botsford et al. 2009). For a fishery to be sustainable FLEP needs be 0, a ratio
representing no change in egg production between a fished and non-fished
population. This is what a marine reserve provides, an area where the spawning
stock enabled to produce the most recruits (see Figure2.). Current management
schemes push to use a FLEP of 0.4, allowing a female to only produce 40% of her
LEP before being harvested. As figure 3 shows, an F of 40% diminishes the
spawner-recruit relationship to levels lower than a non-fished population (F=0).
Increased LEP becomes an important byproduct of protection through a marine
reserve.
Figure 3. The relationship between recruitment and the spawning stock for a
hypothetical fish population. The intersection of replacement line and the
spanner-recruit curve show the change in equalibria as the fishing mortality rate
(F) is increased and lifetime egg production is diminished. FCRT: Replacement
16
line associated with critical replacement threshold F40%: Replacement line that
represents the fishing mortality rate that reduces LEP to 40% FREP: Replacement
line that best fits the data F100% (F=0): Replacement line for a non-fished stock.
Figure source: Botsford et al. 2009
The second factor, larval dispersal, can be thought of as propagule
diffusion from a central location, or marine reserve. Larval dispersal varies from
species to species but can highly determine the potential replenishment of a
fishery stock outside reserve boundaries and determine the self-sustainability of
species within a reserve (Planes et al., 2008). The greater the distance larvae
disperses for a target species propagates outside a reserve could lead to greater
distribution of a species. In a marine environment the dispersal distance is
positively correlated to pelagic larval duration (PLD), or time spent in the water
column as larva (Figure 4) (Bostford et. al. 2009, Shanks. 2009). Though
statistical models show that species with longer pelagic larval stages are able to
disperse propagules a greater distance than ones whose larval stages are short, the
real world outcomes are much more complex. Dispersal results in complex
spatial patterns that reflect an interaction of flow patterns that vary in space and
time with pattern in survival and behavior (Botford et al. 2009).
17
Figure. 4 Estimated dispersal distance plotted as a function of propagule duration.
Dashed line is the best fit to the data. Open circle represent animal populations.
Closed circles represent plant populations. A significant correlation between
propagule duration and dispersal is show for these data. Figure source: Shanks et
al. 2003
In a study by Carr and Reed (1993) four conceptual patterns of population
replenishment via larval dispersal were distinguished (Allison et al. 1998). These
four patterns were then organized along varying axes: distance of propagule
transport relative to the scale of reserve and number of population replenishment
sources (Fig 5). Their study shows the various ways in which larval dispersal
might occur, each four having its own benefit to species that propagate in that
method. Populations that exhibit short distance dispersal maintain the ability to
be self-replenishing, whereas long distance dispersal populations rely heavily on
the influence of a single source population. Whether the source population is a
larval stock, single or multiple active breeding populations, or a single isolated
breeding population depends on the species of fish. Though each four of the
scenarios could be applied to the life history traits of several target fisheries, a
common observation of our large bodied, long-lived fish with high PLD is
18
propagation via long dispersal distance - either from a larval stock that serves
several small populations, or a single breeding population whose propagules
support other populations. This occurs when migratory or seasonal breeders
congregate to form a single large spawning group leaving a larval population that
exists absent of the adult population (Fig 5D), or when only one population is
healthy enough to promote breeding (Figure 5C). Again, this is highly variable
by species, genera and family (Mosqueira et al. 2000.) For example, some
members of the rockfish genus Sebastes spp., have low site fidelity, breed in large
colonies, and leave propagules to disperse (Figure 5D) while others exhibit high
site fidelity, breed in small populations to which the propagules disperse freely
between (Figure 5B) (Allison et al. 1998). Knowing which populations support
the reproductive success of a species will ultimately affect the placement of
marine reserves by asking the question, “What is the spatial structure of the
populations?” and “How demographically connected are these populations?”
Figure 5. Models of population replenishment: Patterns are distinguished by the
distance of propagule dispersal and the number of local propagule sources for a
given local population. Ellipses represent isolated adult populations. Bold lines
indicate high recruitment rates within or between isolated adult populations.
Broken lines indicate low recruitment rates. Figure source: Allison et al. 1998
19
The third factor is the poorly understood movement of large adult fish to
areas outside the reserve. This is know as the “spillover” effect and occurs when
there are high quantities of large fish within the boundaries of a reserve (Palumbi
2004). With dense population condition existing within a marine reserve
competition between cohorts and other species may force fish to venture into
unprotected waters. Conservation goals benefit from minimum spillover outside
of reserves, while fisheries enhancement may benefit from high spillover (Roberts
2000). As movement rates increase, yield can also increase as fish spend less
time in protected reserves (Botsford et al. 2009). A study conducted by
McClanahan and Mangi (2000) monitored fish catch per unit effort (CPUE) from
the edges of Kenyan marine reserves. Their analysis showed mass, size, and
species diversity decreased the farther samples were from the reserve boundary
(Palumbi 2004). Another study by Russ and Alcala (1989) analyzed fishing yield
in the Philippines before and after the collapse of a marine reserve due to illegal
harvesting. This study stated that fishing was 25% more productive when the
reserve was in place. When spillover benefits the yield of fishery neighboring a
reserve, it also reduces the replacement and buildup of spawning stock inside the
reserve. This results in less sustainability and less recruitment (Botsford et al.
2009). The benefit to a fishery from a reserve is dependent on the species being
managed and appears to be highly variable across fish families even when
mobility is similar (Palumbi 2004).
20
MPA Design
Most fish targeted by fisheries go through four distinct life stages: eggs,
larvae, juvenile and adult (Pauly et al. 2002). At each life stage different habitats
may be required, and because marine reserves provide spatially explicit
protection, informed reserve design can protect each life stage. The efficiency of
a reserve is greatly enhanced if the design is scientifically sound (Allison et al.
1998). This may seem obvious, but opponents of marine reserve implementation
argue that there is scientific uncertainty over optimal reserve locations based on
habitat (Roberts 2000). For a reserve to benefit a stressed fishery, it should
overlap spatially with essential habitats required by that species. This means
protecting enough critical habitat by incorporating all aspects of the habitat
utilized by that particular species which in turn requires knowledge of that species
life history, including the requirements of each life stage (Allison et al. 1998).
Larval dispersal and adult movements (spillover) are important to the
replenishment of a target species and support of a fishery. Depending on the
goals set forth by management agencies, reserve designer will incorporate these
factors, such as (restate here) into the planning phase. Proper reserve design and
establishment would be one that retains sufficient offspring to sustain its own
population while also exporting the majority to replenish fishing grounds (Roberts
2000). The effects of successful reserve design may also have ecological benefits
by increasing regional diversity. Protecting these “source” areas whereby species
recruitment is higher than mortality allows for migration out of the reserve, and
increases the potential for sustainable harvesting.
21
Creating reserves that protect source populations or essential habitat
require huge amounts of effort from stakeholders, including scientists and
resource managers, policy makers, and the local fishing community. Necessary
collaboration may involve cooperative research to work towards the management
goal set forth by invested parties. (NRC 2004). A common argument from within
this group is centered on reserve size and configuration, asking: “What is better,
several small reserves or one large reserve?” The answer to this question depends
on the intended goal of the reserve. Reserves implemented for fisheries
replenishment may require scaling larger than reserves implemented for the
conservation of a particular species. Figure 5 conveys what is known of how
several small reserves would work in areas as compared to the effectiveness of
one large reserve. Individual breeding populations with short dispersal will best
benefit from the local protection of a small reserve. Several small reserves would
protect these species because of their high retention of larvae (Roberts 2000).
Large reserves would be best suited to protect species with little local retention of
larvae or key habitats such as natal/nursery grounds, feeding grounds, migratory
routes, etc. (Allison et al. 1998, Roberts 2000). Determining the proper
placement and number of areas or reserves will ultimately govern the success of
MPAs. Figure 6 illustrates how a reserve design can suit a population of fish with
long distance dispersal and little propagule retention outside of the source
population. Each reserve site will have to be treated as its own entity, however,
there is the possibility for connectivity between reserve sites.
22
Figure 6. An example of proper reserve placement: Single reserve (shaded box)
established on a single-resource population. The reserve protects the reproductive
source population. Source: Allison et al 2008.
The analysis of the distribution of biological resources in relation to the
physical environment is a challenge that must be met for the successful
designation of reserve area (Kracker 1999). The appropriate allocation of space is
highly dependent on the understandings of the biological processes within that
geographic region, the ecology of the organisms targeted for protection, and the
goal of the reserve. Current ecologically based strategies involved in the optimal
reserve area selection are often based on specified biodiversity and habitat criteria
(Berglund et al. 2012). The relationship of these two criteria can be examined in
studies focusing on aspects of species distribution, evenness, and richness as it
relates to essential fish habitat (EFH) and the implications that habitat has on
population dynamics for targeted species. Research has demonstrated a high
correlation between biodiversity and quality of habitat connectivity (Berglund et
al. 2012). Specifically, certain species may disperse throughout a region if they
are able to translocate from one protected area to another, by way of high quality
habitat.
This notion gives rise to what is commonly referred to as “Networked
MPAs”. Networked MPAs are a series of marine reserves that work in
conjunction with one another to support the movement and migration of fish
23
within a region. Scenarios where networked reserves are expected to be
successful are when target species exhibit home range behavior (Moffit et al.
2009). For many high-trophic level large fish (such as Sebates), the required
habitat of a home range changes with maturity. Different habitat requirements are
highly correlated to each life stage, from larvae to mature adult fish (Palsson et al.
2009). These “recruitment pathways” may include a successional use of many
types of benthic habitat (Buckley 1997). Changing habitat needs equate to
differing home range behavior at each life stage. For several species of the
Sebastes sp., mature individuals demonstrate high site fidelity, prefer benthic
rocky habitats, and rely on pelagic larvae dispersal to propagate throughout a
region. Networked reserves have demonstrated their efficiency when specific
habitats along a species “recruitment pathways” are targeted (Eisenhardt 2001).
In the past two decades, ecological modeling and marine spatial planning
utilizing tools such as GIS have aided in the creation and appropriate placement
of reserves (Wrights and Heyman 2008.) Researchers are also applying habitat
modeling methods based on remote sensing techniques to better understand the
needs of species in certain habitats, what habitats have the greatest abundance,
and why species occur where they occur (Valavanis et al. 2008). These efforts
directly affect the successful implementation of marine reserves, networked
reserves, or any other form of MPA.
The techniques utilized in the ecological modeling and spatial planning of
marine reserves can also be adapted for analyzing the effects for established
marine reserves. Incorporating the biological needs of target species, the known
24
ecological benefits of marine reserve protection, and technologies to model and
predict species distribution can have wide ranging implications for assessing the
regional effects of marine reserves. In order to understand the variation and
effectiveness of MPAs, in Chapter 2, I examine the ways in which the ecology of
marine reserves affects the surrounding benefit ecosystems region, and fish
populations stressed by previous overfishing.
25
Chapter 2: Manuscript
Introduction
Globally, overfishing is thought to be the primary factor in the ecological
and biological deterioration of coastal ecosystems (Jackson et al. 2001, Worm et
al 2006). The marine ecosystems of North America’s largest Pacific coast estuary,
the Puget Sound, are no different. Decades of overfishing, and loss of marine
habitat to anthropogenic stressors such as development and pollution
contamination have negatively affected most populations and stocks of Puget
Sound groundfish (Rice 2007, Tsao et al 2005). The viability of these fisheries
has been diminished due to historical exploitation to levels that have now
influenced the structure and function of the current Puget Sound marine
ecosystem (Williams et al. 2010). An understanding of trends and changes in
current groundfish populations are needed to effectively manage and conserve the
species within this degraded ecosystem.
Despite historical exploitation, in recent decades there has been a
growing interest in reforming fisheries management and recovering threatened
stocks (Williams et al. 2010). This interest has focused on examining ways to
rebuild depleted stocks of Rockfish (genus Sebastes), and lingcod (family
Hexagrammidae), high-trophic level groundfish that have been adversely affected
by recreational and commercial fishing. An early commercial fishery restriction,
critical to the recovery of the stocks, was the 1989 prohibition of trawling in the
major North and South basins of the Puget Sound (Rice 2007, WAC 220-48-015).
Later regulations adopted by the Washington Fish and Wildlife Commission
26
included the Puget Sound Groundfish Policy to ensure the conservation of habitat
and ecosystems used by groundfish to maintain healthy populations of these
species (Palsson et al. 1998). Recent management actions specific to Rockfish
have targeted the recreational fishery with the implementation of “bag limits” or
daily catch quotas, Rockfish Recovery Zones, and the petitioning to list selected
species under the Endangered Species Act (Biological Team Review NMFS
2008).
With an increase in the reform of fisheries management regulations that
specifically target groundfish, managing agencies (WDFW) implemented the
utilization of Marine Protected Areas in 1990 (Van Cleve et al. 2009). As a
current form of mitigation, MPAs are geographically established areas that
prohibit the harvest of fish. A positive global consensus is building for the use of
MPAs for conservation of degraded marine habitats (Allison et al. 1998,
Mosquera et al. 2000, Palumbi 2004). However, data and studies have been highly
species and region based, demonstrating that target species react positively to the
protection offered by MPAs placed in certain habitats (Van Cleve et al. 2009).
This study investigates the regional effect MPAs have on the diversity of
groundfish within the San Juan Archipelago, and assesses habitat utilization by
target groundfish species. As with any natural resource management tool,
continued analysis is often needed to ensure its proper utilization. MPAs are no
exception; analyzing the regional effects of established MPAs on marine
organisms has led to a greater understanding of their function and implications
with respect to their intended goals. Biodiversity surveys and ecological studies
27
often focus on areas with high concentrations of plant and animal diversity such
as intact reserves and protected areas with low levels of human intervention
(Chazdon et al. 2009). This is also the case for Puget Sound and San Juan MPAs
where studies of biodiversity have focused attention on areas within reserve
boundaries (Tuya et al. 2000). But how do these surveys and studies translate to
areas outside reserve protection? Within Washington, limited studies have been
conducted to understand the effects MPAs have on surrounding fish communities.
Additional research is necessary for the continued implementation of marine
reserves, such as examining regional biodiversity levels and population status for
groundfish at larger spatial scales. By addressing regional biodiversity patterns,
scientists can understand how MPAs function as tools for fisheries management
and as a means for restoring ecosystem function and reverse the past effects of
overfishing (Palsson 2001). The goals of this study are to (1) describe how MPAs
affect regional biodiversity levels, and species occurrences through out time and
over distance by hypothesizing that samples nearer to MPAs will have
significantly different levels in diversity than samples further away, and (2)
determine species by species group composition in various habitats to describe
species preferences for habitats not targeted by MPA, hypothesizing that species
groups will have significant preference for depth and substrate.
By using biodiversity as a metric, this study addresses spatial and temporal
variations in measured diversity levels. It is hypothesized that spatial and
temporal variables will have a significant effect on levels of diversity throughout
the San Juan region, potentially due to the influence of marine reserves.
28
Observing and quantifying change across these ecological scales will inform
managing agencies in the implementation and proper utilization of marine
reserves. Analyzing habitat preference by species family group is an effort to
validate current MPAs selection and implementation strategies, as knowing that
many agencies target specific habitats for protection.
Historical Fisheries Exploitation
The history of harvesting marine fish from the Puget Sound extends to the first
peoples to inhabit the region. Similar to the global trend, Puget Sound fisheries
have suffered from the distinct effects of increase fishing pressure (Myers and
Worm, 2003). Though several species, including groundfish species, are
commercially harvested within the Puget Sound, few have received more
attention than members of the Salmonid family and Sebastes family. The six
species of Pacific salmon are an intricate aspect of the regions ecosystem,
economy, and cultural identity. However, these species have also been a source
of contention for people in the region. The 1974 federal court case United States
v. Washington, commonly referred to as the “Boldt Decision”, upheld
Washington’s Treaty Tribes rights to 50% of the total allowable catch and
responsibility as co-managers of the salmon fishery (Wilkinson, 2000). This was
not only a tide turning event for fisheries management in Washington state, but
also an event that would have unexpected ecological consequences. In response
to the 50/50 allocation of salmon between native and non-native fishers, WDFW
urged non-native fishers to target the “abundant” groundfish stocks of the region
(Dinnel et al., 2003).
29
Though commercial groundfish fisheries began in the 1920’s, the catch
rates for Rockfish species and other groundfish species experienced a rapid
decline in catches landed that began in the early1970’s (West 1997). In the early
years of the commercial rockfish fisheries, rockfish were often caught as nontargeted by-catch of the Cod (Gadidae sp.) fishery. As market demand grew for
rockfish and fishing vessels were restricted to the United States waters the Puget
Sound in response to the Magnuson Stevens Act of 1976, commercial catch
landings for rockfish increased into the 1980’s (Williams et al. 2010). This
depletion rate of stocks increased rapidly in conjunction with the increased fishing
effort that began in the 1970’s (WDFW, 2011). Currently several rockfish species
and stocks have been categorized as “critical” or “fully utilized”, and three
species have been ESA (Endangered Species Act) listed by the National Marine
Fisheries Service (West 1997, Williams 2010, Biological Review Team NMFS,
2009).
Characteristics of Study Species
Groundfish, a non-taxonomic grouping composed of several different
species, have been the focus of decades of research. The examination of
groundfish within this analysis groups species into three species groups. Members
of this category include Rockfish (genus Sebastes), Lingcod (family
Hexagrammidae), and Pollock, Pacific Cod, Tomcod, and Hake (family
Gadidae). With the exception of some Gadids, groundfish are long-lived, slowgrowing, and late-maturing high-trophic level predatory fish (Pacunski et al.
2013). They function both as predators and prey in the complex trophic-web of
30
the Puget Sound (WDFW 2011). Species can be found in nearly all habitats and
depths, a trait that creates difficulty for management, and the perfect scenario for
a commercial and recreational fishery. Several members of the Sebastes genus,
and the Lingcod O. elongates, are targeted within the recreational fishery. Other
species, mostly member of Sebastes are easily caught as by-catch in the salmon
trolling fishery. Management of these species has relied heavily on continuous
scientific analyses. Surveys of population dynamics, distributions, species
compositions and abundance have contributed to a wealth of knowledge pertinent
to Puget Sound groundfish. Unfortunately, the primary cause for these studies has
been in response to the continued depletion of groundfish stocks throughout the
Puget Sound.
Researching the ecological benefits of Puget Sound and San Juan marine
reserves has been accomplished by comparing species populations within reserve
protection to populations outside, in similar habitat. Studies have demonstrated
the benefits of protection include: increases in fish size, total biomass, and species
richness or biodiversity (Palsson 2003 West, 1997). These results have been
replicated for several groundfish species over several managed reserves.
Conducting this research is a necessary step in establishing and monitoring
marine reserves (Palsson 2001); continuous analysis is needed when assessing a
dynamic ecosystem such as the marine ecosystems of the Puget Sound. While the
measurable and observable benefits of marine reserves are well published, a push
towards the creation of networked systems of reserves has proceeded (Moffit et
al. 2009, WDFW MPA Work Group 2009). However, thorough analyses
31
assessing the broader ecological impacts of marine reserves must be conducted.
This is necessary to address gaps in management, monitoring, and evaluation,
ultimately leading to a lack of current understandings.
Building upon the scientific studies, agency recommendations, and
shifting trends towards ecosystem based management practices, the Washington
Department of Fish and Wildlife (WDFW) has contributed several resources to
the investigation and implication of MPAs within the San Juans Islands. These
established MPAs are herein referred to as marine reserves, selected for in this
analysis because of the level of protection employed. The goals of these reserves
are to protect and conserve target species by prohibiting harvest for groundfish
and restricting harvest for other organisms within their boundaries. Of the 22
established reserves managed by WDFW, five occur within the San Juan
Archipelago (Van Cleve et al. 2009). As well as WDFW, The Washington
Department of Natural Resources (DNR) manages two reserves within the
geographic range of this study that employ similar protection levels to support
similar goals.
The analysis of the Trawl and ROV datasets provided by Washington
Department of Fish and Wildlife comprised a two-part study to determine the
regional affects of marine reserves in the San Juan Archipelago. The first, the
analysis of eight years of trawl surveys in the San Juan Archipelago, employed a
unique methodology of using diversity indices to quantify temporal and spatial
variations in levels of diversity of high trophic level ground fish throughout the
region. The second, an analysis of ROV survey transects determined species
32
composition by species group per habitat type and depth. These parameters can
be used to assess species expected occurrence rates when habitat data is absent.
The utilization of both datasets allowed for a comprehensive examination of the
biotic and abiotic factors that affect the implementation of marine reserves in the
San Juan region.
33
Methods
Study Area
All data used for this study was collected within the waters surrounding
the San Juan Islands (Appendix A-1). Located off the northwest coast of
Washington State, the San Juan Archipelago is comprised of over 450 islands,
rocks, and tidally exposed reefs. Depths range from the shallow intertidal to areas
over 200m. The regions distinct geomorphology has created an intricate network
of straits, channels, and passages between landmasses with complex underwater
landscapes. This complexity equates to an abundance of potential habitat for
rockfish and other benthic species (Tilden 2004). Compared to other regions of
the Puget Sound, the San Juan Archipelago has a significant amount of rocky
habitat (Palsson et al. 2009). These waterways represent the San Juan Basin, a
distinct Sub-Basin within the North Puget Sound Basin commonly referred to by
fisheries manager (Williams et al. 2010). Within the San Juan Archipelago, the
Washington Department of Fish and Wildlife manages five San Juan Marine
Preserves (Table 1) (Appendix A-2).
Table 1. WDFW Marine Reserves. UML = Uniform Multiple Use, ProRec Recreational harvest prohibited, ResCom - Commercial harvest restricted, ResAll
– All harvest restricted. Source:Van Cleve et al. 2009).
Preserve Name
Acreage Year
Protection
Harvest
Established
Level
Restrictions
Argyle Lagoon MP 13.00
1990
UML
ProRec/ResCom
False Bay MP
94.70
1990
UML
ResAll
Friday Harbor MP 427.20 1990
UML
ResAll
Shaw Island MP
432.50 1990
UML
ResAll
Yellow and Low
187.20 1990
UML
ResAll
Islands MP
34
Source of Data
The Washington Department of Fish and Wildlife’s Marine Fish Division
(WDFW) collects and compiles vast amounts of fisheries related data for the
Puget Sound and adjoining waters. As per WDFWs mission statement to protect
the fish and wildlife resources of the state, the collection of biological and
ecological data of native fish species is a primary goal. For fisheries related
issues within Washington State data collection on various projects have
contributed to a wealth of information. The data collection methods used by
WDFWs Marine Fish Division are similar to that of other agencies. Data
collection includes the following methods: dive surveys, research trawl surveys,
intertidal habitat surveys, and Remotely Operated Vehicle (ROV) surveys. Each
method has its benefits for specific analyses. WDFW has used these survey
methods to analyze abundance, diversity, species composition, and to gather
biological information for several marine organisms.
Washington Department of Fish and Wildlife’s Marine Fish Program
provided the datasets for the analyses. Two datasets were selected for this study.
The first contained trawl survey data collected from 1987 through 2012 and was
utilized for the regional biodiversity analysis. The second dataset contained ROV
survey data from the 2008 and 2010 survey years and was utilized for the habitat
utilization analysis. The trawl and ROV datasets differ by collection methods and
spatial scales. These differences allow for separate analyses to be conducted.
Prior to acquiring the data, both datasets were sorted by species of interest.
Because this analysis focuses on commercially important high-trophic level fish,
35
all species other than members of Sebastes, Gadidae, and Hexigrammidae were
filtered from the dataset. Trawl and ROV surveys were conducted at differing
times through out the region. Trawl surveys used in this analysis were conducted
during May and June. The ROV survey regime sampled the San Juan
Archipelago independent of season, with data collected over 42 days (Pacunski, et
al. 2013). Data for mapping was provided by WDFW through personal contact.
These data consisted of Washington’s MPA inventory geodatabase compiled by
Van Cleve and colleagues (2009).
Occurrence Rates and Temporal Variation
Sampling methodology and survey techniques for the Trawl Surveys used
are detailed extensively in WDFW’s Trawl Survey Field Plan and Manual
(Palsson et al. 2002). The general practice for these trawl surveys consisted of
station selection, catch processing, and sub-sampling for biological sample
collections. An established protocol for station selection ensured that the
regionally based and station based systematic surveys were implemented without
bias (Palsson et al., 2002). For each survey site, data was collected on location
and duration of tow, beginning and end depths, species caught, abundance, and
sex/length frequencies.
Data auditing was needed to remove errors from the trawl dataset.
Because the trawl database contained data on regions outside the San Juan
Archipelago, data were selected from the database after establishing filter
parameters. Post-filtering yielded a dataset specific to the San Juan region. After
this process was completed, the total number of target species was identified for
36
each year within the dataset. Within each year, individual samples transects were
identified by haul sequence number.
Data for the biodiversity analysis were compiled using surveys from 2001,
2004, 2006, and 2008 - 2012. These survey years represented all years data were
collecting in the San Juan region. Prior to 2007, trawl surveys were regionally
based, so not all years surveyed had data collected in the San Juan region
(WDFW personnel, personal communication). From 1987 to 2007, only three
surveys were present for years 2001, 2004, and 2006. In 2008, WDFW changed
their data collection strategy to a Puget Sound wide station survey. This
implemented sampling by station, and not by region. Under this new collection
strategy each station would be sampled twice per year. The two separate
sampling strategies affected that amount of samples collected per year (Table 2).
The regionally based sampling strategy averaged 39 sites per year. The station
based sampling strategy averaged 11.5 sites per year. In 2011, 27 sites were
sampled, though still under the station based sampling method, 2011 eleven was
analyzed with years of similar sample sizes. A total of 189 sample sites were
utilized in this analysis.
Table 2. Years with number (n) of Survey Trawls by sampling method. 1In 2011
technically station based. Analyzed with Regional Sampling due to sample size.
Regional Based Sampling
2001: 40
2004: 35
2006: 41
2011: 271
Station Based Sampling
2008: 12
2009: 12
2010: 12
2012: 10
Species group demographics were calculated for the trawl dataset. For
these analyses, the 14 species observed were compiled into their most similar
37
taxonomic ‘species’ groups: Sebastes, Gadidae, and Hexigrammidae. These three
groups consisted of the total counts for all target species observed over the course
of the trawl surveys. (Table 3) For each year, frequencies of occurrence rates
(%FO) were compared via ANOVA and Chi-squared contingency methods.
Count data for %FO rates were log-natural transformed. To test the effects of
depth on species group composition, the average depth per sample was first
categorized into two depth stratums, Shallow <36.6m, Deep >36.6m. Nontransformed count data for number of species per family present in each sample
were utilized in the analysis of specis group occurrence rates by depth stratum.
This data were incorporated into the spatial analysis.
Table 3. Species observed by species group
Sebastes (genus)
Gadidae (family)
Copper Rockfish, Greenstriped
Rockfish, Puget Sound
Rockfish, Quillback Rockfish,
Redstripe Rockfish, and
Redbanded Rockfish
Walleye Pollock,
Pacific Cod,
Pacific Tomcod,
Hake (Pacific
Whiting)
Hexagrammidae
(family)
Lingcod, Kelp
greenling, and
WhiteSpotted
Greenling,
Biodiversity was calculated for each sample within each year using the
Shannon-Weiner and Simpsons D diversity indices (Figure 7) (Magurran, 1998.
Keith 2005). These indices measure the species richness and evenness and are
commonly used in ecology to describe demographic relationships between
organisms within a given sample. A simple abundance calculation (number of
species divided by total number of individuals) would demonstrate species
richness but does not allude to the distribution of species throughout the sample,
or to the commonness relative to other species. Though similar, the ShannonWeiner and Simpsons D indices utilize different methods for describing the
38
richness and evenness of a given community of biological sample. First, both
indices make no assumption as to the underlying species abundance distribution,
and work well with small sample sizes, such as in this study (Magurran 2004).
Second, both indices are calculated from the log natural transformation of the
abundance data, creating proportional data of the individual per species by
sample. The differences in the indices are seen in their utilization and
interpretation of data. The Shannon-Weiner assumes that individuals are
randomly sampled from an infinitely large community and that all species are
represented in the sample (Margurran 2004). Using these assumptions, the
equation calculates evenness values to detail species evenness within samples.
Though the Shannon-Weiner Index is inclusive of evenness, as separate evenness
calculation was also conducted. The Simpson D index incorporates evenness and
richness and this index is also considered robust when working with small
samples, however, it is weighted towards the most abundant species, assuming
that the most abundant species is of greater importance to the ecosystem (Greene
1975).
Shannon-Weiner:
H =-Σpi ln pi
EH = H/Hmax =H/lnS
Simpson D:
D = Σ p i2
D = 1/Σpi2
(Reciprocal representation for analysis)
Ed = D/Dmax
Figure 7. Shannon-Weiner Diversity and Evenness Equation Simpson’s D
Diversity and Evenness Equation
Evenness, for both indices is represented on a scale between 0 and 1,
where 1 equals complete equitability or evenness amongst the sample. Both
39
indices were utilized in the comparison of means to total catch per year. Unlike
the frequency occurrence analysis, the years sampled under the regional sampling
scheme were incorporated to other, separate from the years under the station
based sampling scheme.
The temporal analyses conducted consisted of three separate tests
designed to examine when significant changes in diversity were occurring. These
test were to compare diversity and evenness values:
•
Between all years sampled.
•
Between years by sampling method.
•
Within year by depth stratum.
These analyses were conducted using a non-parametric test for comparison, and
resampling ANOVAs. Diversity indices values were also examined against
abiotic variables of depth (average depth of sample location) and trawl tow
length to assess possible correlations.
ROV: Species Composition and Habitat
The ROV surveys conducted in the San Juan Archipelago focused on
determining the habitat preference for species of Sebastes and other target ground
fish. Data for the habitat utilization analysis were compiled from the 2010 ROV
dataset. Though both the 2008 and 2010 datasets were present, the 2008 data had
recently been reported in Pacunski et al. (2013). The 2008 results would be used
to reference the methods of the analysis for the 2010 dataset. Detailed
methodology for the deployment and sampling protocol used can be found in
Pakunski et al. (2013).
40
The ROV survey data for 2010 complied observations made over a total of
139 sample segments. For each segment, the ROV ran approximately 1km,
capturing species and habitat structure observations via video camera. The
process of video recording as the ROV traveled along the bottom simulated a fish
net like “capture event” as if the fish were captured in a trawl net. The video was
then later analyzed and all species were recorded with their location and habitat
type along each segment.
Data auditing for this data set consisted of truncating the data to include
only significant habitat variables, location, sample ID, species groups, and species
counts. The significant habitat variables consisted of values for the complexity
(structure), and substrate.
The ROV data was utilized to determine the relationship between habitat
complexity and species group occurrence. The target species within the ROV
data were the same as those for the trawl data. Because these species are often
associated with structured and complex habitat, such as rocks, outcroppings,
boulders, slopes, and substrate depressions, an analysis of species occurrence by
habitat complexity was used to determine any habitat preference observed within
the sampled population (Pacunski et al. 2013). Habitat categories for substrate,
complexity and relief were compared, jointly and independently, to counts of
species abundance. Similar to the analysis for the trawl dataset, the 16 species
(including all unidentified but counted species) observed were compiled into their
most similar taxonomic ‘family’ groups: Sebastes, Gadidae, and Hexigrammidae.
41
These three groups consisted of the total counts for all target species observed
during the 2010 ROV survey (Table 4).
Table 4. Species by species group
Sebastes (genus)
Canary Rockfish, Copper Rockfish,
Greenstriped Rockfish, Puget Sound
Rockfish, Quillback Rockfish,
Vermillion Rockfish, Yelloweye
Rockfish, Yellowtail Rockfish, and
Rockfish Unidentified.
Gadidae (family)
Gadidae Unidentified,
Pacific Cod
Hexagrammidae (family)
Lingcod, Kelp greenling,
Painted Greenling, WhiteSpotted Greenling,
Hexagrammid Unidentified.
Mapping and Spatial Analysis:
ArcGIS 10.0 was used in the spatial analysis of all sampled points. Data
from the Trawl dataset were compiled in a concise format, and imported to the
geodatabase containing the shapefiles for all of Washington State’s MPAs. Trawl
sample locations and latitude/longitude were converted to degree decimal degree
prior to georeferencing. All points were projected into
NAD_HARN_1983_Washington State Plane_South, to match the projection of
the MPA shapefiles. MPAs were selected from the geodatabase by location (the
San Juan Archipelago), and managing agency (WDFW and DNR). Both state
agencies manage MPAs as functional marine reserves that prohibit the take of
groundfish and other organisms.
Diversity Hot Spots were determined using the Hot Spot Analysis “GetisOrd Gi*”. For these analyses data from all eight survey years were examined as
one set. The total 188 sample sites were used to generate the Diversity Hot Spot
map (one sample had incorrect latitude/ longitude and was discarded). Data
selected for this analysis tool were the four values of diversity and evenness. The
tool calculates the Getis –Ord Gi* statistic for each feature in the dataset (these
being the sample location points). The distance from each point in the feature
42
was measured to the 10 closest neighbors as determined by the nearest-k setting.
This tool then mapped how clusters of similar diversity values were spatially
distributed throughout the surveyed area. Based on the sum of attribute value for
a spatially distinct group, the tool calculates the probability of that summation
occurring randomly within the total feature dataset. Feature points with high
values created hot spots when surrounded by feature points of similar high values.
The outputs of this tool are z-scores and p-values that denote significance.
Because of the limits of the data collected, the hot spot analysis could not be used
to create a model for predicting diversity values at other locations, such as similar
methods that utilize Kriging (Bolstad 2008). Presence of Hot and Cool spots were
then analyzed by comparing species composition and abundance between the
samples comprising the hot spots.
The Spatial Autocorrelation tool (Global Moran’s I) was also employed to
detect clustering of diversity values. The null hypothesis for this tool states that
feature values are randomly distributed across the study area. The outputs of this
tool are z-scores, and p-values. Unlike the Getis –Ord Gi* tool the Global
Moran’s I tool does not locate regions within the data where significant values are
clustered (Figure 8).
Figure 8 Example of dispersion and clustering (ESRI.com)
43
Distance from sample location to nearest marine reserve was calculated
using the “Near” proximity tool. This gave the Euclidian distance from each
sample location to its closest marine reserve. In most instances, the measured
distance between reserve and sample point was a relatively straight line that did
not cross over land. However, a small number of sample points did. Rather than
assume a least cost path of possible fish travel around an obstruction, the
determination to utilize Euclidian distance was validated by referenced literature
(Tuya et al. 2000). Due to the low number of samples fitting this scenario, the
analysis was not likely skewed. The acquired distance from sample point to
nearest reserve was then correlated to the values of diversity and evenness.
Presence/Absence data for number of species per species group was also
spatially analyzed. The primary tool utilized for this analysis was the Global
Moran’s I, to detect any clustering of samples with high counts of species
occurrences. Species counts per sample were analyzed using all years for
Sebastes, Gadidae, and Hexagrammidae.
Statistical Analysis
All statistical analyses were run and tested using JMP 9.0 (SAS Software).
Diversity indices variation by sampling method, and depth stratum were analyzed
using the Kolmogorov-Smirnov (KS-test) for non-parametric data. The analysis
for index variation over time with regards to sampling method was conducted
using similar non-parametric tests. First a resampling ANOVA was utilized to
determine if variances were occurring throughout the sampled years. Comparison
test of all year parings utilizing the Wilcoxon method was then run to detect
44
which year groupings had significant variation in diversity indices values. To
determine the effects of two abiotic factors (depth, and trawl tow distance) an
analysis of multivariate data employing Pearson’s R Pairwise correlations were
used. Depth was also categorized and examined for significant effects on
diversity indices values using a Chi-Square contingency test.
Habitat variables affecting species presence and composition were
analyzed using a series of Chi-Square contingency and ANOVA tests. These tests
examined Frequency of Occurrence (FO%) for the three species groups in
different depth and habitat categories. For the trawl dataset, FO% was analyzed
using a chi-square analysis and an ANOVA. Data were log-natural transformed
prior to the ANOVA analysis.
45
Results
Trawl Frequency of Occurrence
Over the course of the eight surveyed years, a total of 11,749 observations
were made of 14 target species. These species were positively identified to the
lowest possible taxonomic grouping; for this survey identification to species was
possible. The 6 species of rockfish (Sebastes) were: Copper rockfish (S.
caurinus), Greenstriped rockfish (S. elongates), Puget Sound rockfish (S.
emphaeus), Quillback rockfish (S. maliger), Redstriped rockfish (S. proriger), and
Redbanneded rockfish (S. babcocki). Three species of Hexagrammidae were
observed: Lingcod (Ophiodon elongates), Kelp greenling (Hexagrammos
decagrammus), and Whitespotted greenling (Hexagrammos stelleri). Four
species of Gadidae were observed: Pacific Cods (Gadus macrocephalus), Walleye
Pollock (Theragra chalcogramma), Hake/Pacific Whitting (Merluccius
productus), and Pacific Tomcod (Microgadus proximus). Species frequencies of
occurrence rates (%FO) were determined for each year of the trawl surveys. The
three species groups differed significantly in their occurrence rates (Likelihood
ratio ChiSquare DF =14, 839.33, p <0.0001, Pearsons ChiSquare DF=14,
810.187, p <0.0001). Of the 11,749 observations members of the Gadidae family
constituted the vast majority, 92.3% (10,813), of observed fish for all eight years,
of which the majority of these observations were Walleye Pollock. The family
Hexagrammidae constituted 5.40%, and members of Sebastes made up 2.56% of
the total observed population. Looking at each year separately, similar significant
proportions can be seen (Table 5). Similarly, significance was demonstrated
46
between the abundance of each species group (ANOVA DF=2, F=17.371, p <
0.0001).
Table 5. Frequency of Occurrence and Abundance rates for the three target
species groups. Total row represents percentage of total observations and
abundance.
Frequency of Occurrence (%FO) and Abundance
Year
Gadidae
Hexagrammidae
Sebastes
2001
81.92
2519
13.59
418
4.49
138
2004
94.55
1595
2.19
37
3.26
55
2006
89.25
1328
5.24
78
5.51
82
2008
98.55
542
0.55
3
0.91
5
2009
96.82
1067
3.18
35
0.00
0
2010
98.73
931
0.95
9
0.32
3
2011
97.38
2231
2.40
55
0.22
5
2012
97.88
600
0.00
0
2.12
13
TOTAL %
92.03
10813
5.40
635
2.56
301
The relationship between depth stratum and species group examined the
number of species per group that occurred in each sample. For the genus Sebastes
and the family Hexagrammidae, the most species caught in a single sample was 3.
The most caught for Gadidae was 4. The distribution for number of species
present in each sample was shown to be significant by depth stratum only for the
families Hexagrammidae and Gadidae (respectively: Likelihood ratio ChiSquare
68.80 and 25.364, p<0.0001). Sebastes species were caught in 65 of the 189
samples, of which the number of species per sample was not significantly
dispersed between the two depth stratums. Species of Hexagrammidae were also
caught in 65 of 189 samples, however the number of species present in each
sample was shown to increase in the shallow depth stratum. For this family, 68%
of single species samples and 100% of the samples containing two or more
species occurred in the shallow depth stratum. Gadidae were present in all but 7
of the samples throughout the eight surveyed years. Significance for the deep
47
depth stratum was shown with 47% of single species samples, 63% of dual
species samples, 76% of three species samples, and 90% of four species samples
occurring in deep depth stratum.
Diversity and Evenness by Year
Diversity and evenness were compared for each survey year using two
diversity index values. The Shannon-Weiner index (H), and the Simpsons D (D)
index along with their corresponding evenness calculations were used in
determining each samples indices value and mean indices values for each year
surveys took place. Prior to the statistical analyses, a comparison of sampling
methods utilized to collect diversity data was done. Of the total 189 samples, 143
were under the regional sampling scheme, and 46 were collected under the station
sampling method. Diversity and evenness values differed significantly by
collection method for all but the Simpsons D index (Table 6). This analysis
examined the variability in mean H, D, Eh, and Ed, by sampling methodology
employed by WDFW throughout the data collection years.
Table 6. Variation of diversity and evenness values between sampling methods.
Kolmogorov-Smirnov test for significance. *Denotes significance.
Diversity
Indices
Shannon-Weiner
(H)
Shannons
Evenness (Eh)
Simpsons D
(D)
Simpsons
Evenness (Ed )
Regional
Mean
0.5547
Station
Mean
0.4014
KS
Statistic
0.1084
D Max
deviation
0.2526
P value
0.0235*
0.1884
0.1706
0.0988
0.2303
0.0498*
1.6137
1.3746
0.0818
0.1906
0.1593
0.1083
0.0457
0.11475
0.2674
0.0138*
48
Analyses for testing the Shannon (H) and Simpsons (D) diversity and
evenness (Eh, and Ed respectively) values for significance for all years sampled
were conducted utilizing a resampling analysis of variance (ANOVA) method,
and a Wilcoxon Ranked-Sums test. Significance in both Shannon’s and Simpsons
values (p <0.0001) was found when testing all years together using the resampling
ANOVA. Though, this method did not allude to where the significance was
occurring, it was possible to examine the decrease in mean diversity and evenness
values over all years surveyed. Mean diversity values were shown to fluctuate
over 11-year time span of the data collected (Figure 9). The lowest mean
diversity values were observed in 2008 for both diversity indices. Mean values
were shown to increase from 2008 to 2012, however, they did not surpass the
initial diversity values observed in 2001. A non-parametric comparison of pairing
using the Wilcoxon method was applied to each diversity and evenness index.
Significance for the Shannon and Simpson’s diversity indices was found in year
pairings between 2001 and all other sampled years except 2006 and 2012. The
corresponding evenness indices exhibited similar significance with the exception
of Simpson’s evenness; this value’s significance was found between several year
pairings (Table 7).
Table 7. Shannon and Simpsons Diversity Indices: Only observed significant year
pairings between all year groupings.
Diversity
Index
Shannons
(H)
Simpsons
(D)
Year
pairings
2001-‐2004
2001-‐2008
2001-‐2009
2001-‐2010
2001-‐2011
Z
score
-‐2.7615
-‐3.0737
-‐2.9538
-‐1.9873
-‐2.2818
p
value
0.0058
0.0021
0.0031
0.0496
0.0225
Z
score
-‐3.0005
-‐2.3031
-‐2.5412
-‐2.0959
-‐2.0324
p
value
0.0027
0.0423
0.0011
0.0361
0.0421
49
Mean
Diversity
and
Evenness
Valuse
+-‐
se
Diversity
Values
by
Index
2
1.5
1
0.5
0
2001
2004
H
2006
2008
2009
2010
Years
Surveyed
D
Eh
2011
2012
Ed
Figure. 9 Mean Values for Shannon-Weiner Index (H, Eh), and Simpsons D (D,
Ed) With standard error.
To exclude sampling bias, years were categorized and separated by
sampling method and similar sample size. Years 2001, 2004, 2006, and 2011
comprised the regional based sampling method or large sample and were analyzed
separately from years 2008, 2009, 2010, and 2012 using a Wilcoxon RankedSums test. The ranked sums analysis one-way chi-square approximation for the
years under the regional sampling showed diversity and evenness value varied
significantly by sampling year, with the exception of Simpsons evenness (EH) (H:
ChiSquare 8.50409, DF = 3, p = 0.0361*, Eh: ChiSquare 7.7032, DF = 3, p =
0.0526, D: ChiSquare 8.8345 DF=3, p = 0.0316*, Eh: ChiSquare 23.4963, DF = 3,
p < 0.0001*). Values for these years varied in significance, depending on the
pairing of years compared. A Nonparametric comparison for each pair using the
Wilcoxon method examined which year pairing demonstrated the most
significance. Of the six possible year combinations, the diversity and evenness
values for the 2001 – 2011 and 2001 – 2004 were the only years to demonstrate
50
significant values. (2001 and 2004 respectively: H: Z = -2.28181, p = 0.0225, Z =
-2.76148, p = 0.0058, Eh: Z = -2.07641, p = 0.0379, Z = -2.80587, p = 0.0050, D:
Z = -2.03248, p = 0.0421, Z = -3.00059, p = 0.0027, Ed: Z = -3.25852, p = 0.0011,
Z = -4.77580, p < 0.0001). Year pairing 2001-2006 showed no significance for
any index value except for Ed (Z = 2.55546, p = 0.0106).
Years 2008, 2009, 2010, 2012 comprised the station based sampling
method. Due to this change in sampling methodology only 12 samples per year
were collected. (Table 8.) The ranked sums analysis one-way chi-square
approximation showed no significant variance in the values of both indices in
diversity and evenness. Further examination using the nonparametric comparison
for the 6 possible year pairings also showed no significant variance in the values
of both indices in diversity and evenness (all p values greater than 0.05).
Table 8. Shannon’s and Simpsons Diversity: Significant year pairings when
analyzed by sampling method (P<0.05)
Regional
Station
2001*
2004*
2008
2009
2001
2006
2008
2010
2001*
2011*
2008
2012
2004
2006
2009
2010
2004
2011
2009
2012
2006
2011
2012
2010
The two abiotic variables addressed were trawl tow distance and depth.
As per the sampling protocol, the target tow distance ranged from 0.2 to 0.4
nautical miles. Tow length was observed to have no linear relation to total catch
numbers, and was shown to have no significant correlation to diversity and
evenness values (Pearson’s R, H:-0.0397, Eh: -0.045, D: -0.0406, Ed: -0.0037).
51
Depth, when categorized by depth stratum (shallow < 36.6m, and deep
>36.6m), showed significant variation by in all was determined for all diversity
and evenness values (Wilcoxon, H: p = 0.0012, Eh p < 0.0001, D: p = 0.0054, Ed:
p 0.0005). A median test for non-parametric data was also performed and showed
similar significant values for the four variables. Depth was determined if to
exhibit significant negative relationship, when correlated to the diversity indices
values by sampling method.
Similarly, when analyzed as a continuous factor, depth was shown to have
significant negative correlation to both indices values when analyzed per
sampling method. Within the regional sampling scheme, the Simpsons D index
for evenness (Ed)) was the only value to have a non-significant correlation to
depth. Within the station sampling scheme, the Simpsons D diversity index (D)
was the only value show non-significance (Table 9).
Table 9. Pairwise correlation: Depth to Diversity Indices by Sampling method*
Denotes significance.
Depth
Region
Station
Correlation
Diversity
Pairwise Corr. (r) p value
Pairwise Corr. (r) p value
Index
H:
-0.3200
<0.0001*
-0.3115
0.0035*
Eh::
-0.3221
<0.0001*
-0.3903
0.0073*
D:
-0.3070
<0.0001*
-0.2100
0.1612
Ed:
-0.0694
=0.4101
-0.3160
0.0324*
ROV: Habitat Preference
During 2010 WDFW’s conducted 139 ROV transects were conducted
throughout the San Juan Island Archipelago. Several abiotic variables were
determined during the analysis of the 2010 ROV dataset. These variables
included: depth, substrate composition, substrate relief, and substrate complexity.
52
Species compositions were compared to each of these variables with varying
degrees of significance. Of the 17 species observed, 13 were positively identified
to species. Canary rockfish (S. pinniger), Copper rockfish (S. caurinus),
Greenstriped rockfish (S. elongates), Puget Sound rockfish (S. emphaeus),
Quillback rockfish (S. maliger), Vermillion rockfish (S. miniatus), Yelloweye
rockfish (S. ruberrimus) and Yellowtail rockfish (S. flavidus) were the positively
identified members of the Sebastes genus. The Lingcod (Ophiodon elongates),
Kelp greenling (Hexagrammos decagrammus), Painted greenling (Oxylebius
pictus), and White-spotted greenling (Hexagrammos stelleri) were the positively
identified members of the Hexagrammidae family. Unidentified rockfish,
unidentified red rockfish and unidentified Hexagrammidae were categorized and
counted separately. All observed gadoids were either Pacific Cods (Gadus
macrocephalus), or were categorized as a family group, Gadidae.
Because of these varying levels in taxonomic ordering, three subgroups:
Gadidae, Sebastes, and Hexagrammidae, were used and comprised the
summations of all observation for the appropriate species. Members of the
Sebastes group comprised 45.98% (1407) of the 3060 species observations.
Gadidae observations comprised 45.75% (1400), while Hexagrammidae
comprised 8.27% (253) of all observations (Table 10).
Table 10. Species Group Frequency of Occurrence for all samples.
Species Group
Frequency of Occurrence
Total Observation
(%FO)
Sebastes
45.98
1407
Gadidae
45.75
1400
Hexagrammidae
8.27
253
TOTAL
100.00
3060
53
Relationships between species abundance/occurrence, substrate variables,
and depth were evaluated using Chi-Square contingency analyses. Depth stratum
was determined using a “count if” function that categorized depth into two
classes, Deep (>36.6m) and Shallow (<36.6m) (Pacunski et al, 2013). Eight
substrate categories: boulder, cobble, gravel, shell/shell hash, mud, pebble,
bedrock, and sand were determined by visual measurements of the substrate when
species were encountered (Table 11). The analysis of substrate occurrence and
depth showed seven of the eight substrates occurred in both deep and shallow
depth stratums, with gravel being the exception and not present in any of the
shallow observations. Of the eight categories, bedrock and sand contributed to
60.42% of the encountered substrate, with 90.19 % of bedrock and 95.12% of
sand observed in the deep stratum (Likelihood ratio ChiSquare 100.665 p
<0.0001, Pearsons 106.347, p <0.0001) (Figure 10).
54
45.00%
38.59%
40.00%
% of Total(Habitat)
35.00%
30.00%
23.83%
25.00%
20.00%
15.00%
12.08%
11.07%
10.00%
7.38%
4.03% 3.02%
5.00%
0.00%
0.00%
B
C
G
H
M
P
R
S
Substrate
Shallow
45.00%
40.00%
38.28%
% of Total(Habitat)
35.00%
30.00%
25.00%
23.29%
20.00%
15.00%
10.00%
10.79%
8.76% 8.37%
5.61%
4.17%
5.00%
0.72%
0.00%
B
C
G
H
M
P
R
S
Substrate
Deep
Figure. 10 Habitat Substrate percentages by depth stratum.
55
All three species groups demonstrated a significant affinity for substrate
type (Pearsons ChiSquare, DF =2, 2022.16, p <0.0001). Gadidae observations
were made in seven of the eight substrate categories. No members of this species
group were observed while the ROV was over a boulder substrate. In most cases,
the Gadidae group comprised the highest percentage of total species observation
for the seven substrates it was present. The exception to this was the bedrock
substrate category, where only 6.7% of the species observed in this substrate were
members of the Gadidae species group. Gadidae most frequently occurred in the
sand substrate with 40.2% of all Gadidae observations made while the ROV was
over sand substrate. The Hexagrammidae group showed the most preference for
bedrock substrate with 51.0% of Hexagrammidae observation occurring in this
substrate. Members of the Sebastes group were observed most frequently on
boulder and bedrock substrate, and comprised 92.8% and 82.3% of the total
observations for this boulder and bedrock substrate. Of the total 1407 Sebastes
observations, the majority 68.6% occurred on bedrock substrate. The fewest
rockfish observations were made in shell hash and mud substrate, collectively
comprising 0.7% of the total Sebastes observations (Figure 11).
56
Percetage
of
Species
Group
Observations
Frequency
of
Occurence
(%FO)
by
Substrate
Type
80.0
60.0
40.0
Sebastes
20.0
Gadidae
0.0
Hexagrammidae
Substrate
Type
Figure 11. Species group frequency of occurrence per substrate
Table 11. Habitat Substrate Description
Code
Substrate type
B
boulder
C
cobble
G
gravel
H
shell/shell hash
M
mud
P
pebble
R
bedrock
S
sand
Grain Size
0.25-3.0 mm
64-256 mm
2-4 mm
<0.06 mm
2-64 mm
continuous
0.06-2 mm
Habitat Complexity was also shown to significantly relate to species
presence (Pearsons ChiSquare, DF=7, 2315.08 p <0.0001). The Gadidae family
was on simple flat habitat through 97.8% of the observations. Hexagrammidae
and Sebastes however, were observed to preferred habitat with more structure:
49.2% of Hexagrammid observations and 65.5% of Sebastes observations were
made in low but structured habitat (Figure 12). In medium to high complexity
habitat, Sebastes accounted for 85%-95% of all fish observations.
57
Percentage
of
Observations
by
Species
group
Frequency
of
Occurence
by
Habitat
Complexity
100.00
80.00
60.00
Sebastes
40.00
Gadidae
20.00
Hexagrammidae
0.00
Simple
Low
Medium
High
Habitat
Complexity
Figure: 12 Species group occurrences per Habitat Complexity
Table 12. Complexity of Benthic Substrate
Complexity Description
Simple
simple (flat/hummocky w/no crevices)
Low
low (very few crevices, vert. relief 0.5 to 2 meters)
Medium
medium (more than a few crevices, vert. relief >2 meters)
High
high (lots of crevices, Steep slope or wall)
Depth stratum was also shown to be a significant abiotic variable in
determining species occurrence rates (Pearsons ChiSquare 589.172, p < 0.0001).
Over 90% of all observations were made within the deep stratum. Gadidae and
Sebastes demonstrated higher occurrence rates, with 96.30% and 91.89% of the
observed species in these groups occurring in the deep stratum. Though present
in the deep stratum, the Hexagrammids showed a preference for shallower depths.
52.57% of observed Hexagrammids occurring in the shallow depth stratum.
Hexagrammidae were the only species group to be higher in abundance in the
shallow depth stratum than the deeper, however, this abundance did not allude to
a significant preference for the shallow depth stratum (Figure 13).
58
46.74%
48.91%
% of Total(Species)
% of Total(Species)
17.39%
38.13%
44.48%
4.34%
Species
Species
Species
Gadidae
Hexagrammid
Sebastes
Species
Deep
Gadidae
Hexagrammid
Sebastes
Shallow
Figure 13 Species group occurrence rates per depth stratum.
Though species group occurrence rates differed significantly between
depth strata, an analysis of variance tests for species group composition between
depth strata was shown to be nearly significant (ANOVA, DF = 33, F = 2.3585, p
=0.0661). Though non-significant (ANOVA, DF = 1, F = 3.8014, p = 0.0613),
the variable for depth stratum was shown to have the most significant influence
on species counts. An example, and the only significant interaction for the depth
analysis, was made between the species group Hexagrammidae and the deep
Depth Stratum (ANOVA, t = -2.12, p = 0.0427).
Spatial Analysis
Spatial analyses for clustering of species and diversity “hot-spots” were
run using ArcGIS 10.0 software (ESRI). The Spatial Autocorrelation Global
Moran’s I test examined the dispersion of Shannon and Simpson’s diversity and
evenness values. All diversity and evenness values, except the Simpsons
59
evenness, demonstrated significant clustering across the survey area, against the
null hypothesis of the test stating no significant clustering (Table 13).
Table 13. Results of the Spatial Autocorrelation Global Moran’s I: All Diversity
and Evenness values significantly clustered except for the Simpsons Evenness
(Eh). *Denotes significance.
Diversity Index
Moran’s Index
Z-score
p-value
H:
0.3018
5.9452
<0.0001*
EH:
0.0995
2.0864
0.0369*
D:
0.1766
3.5825
0.0003*
ED:
0.0830
1.7827
0.0746
The Getis-Ord Gi* Hot Spot analysis spatially assigned significance to
sample points with similar diversity levels as their 10 closest neighbors. “Hot
Spots” and “Cool Spots” of diversity values are seen throughout the surveyed
region. However, the majority of samples were shown to not exhibit significant
levels of diversity when compared to their neighbors by this analysis. The
analysis of the Shannon’s index values demonstrated significant grouping of
diversity value for 68 (36%) of the 188 samples. Within these 68 samples 38
were significantly low values (GiZScores < -2.0, p < 0.05) and 30 were
significantly high values (GiZScores > 2.0, p < 0.05) (Appendix B-1). The
Simpsons index demonstrated significant groupings of samples 32 (17%) of the
188 samples, with 8 being significantly low values (GiZScores < -2.0, p < 0.05)
and 24 being significantly high values (GiZScores > 2.0, p < 0.05) (Appendix B2).
Both the Shannon’s and Simpson Diversity indices showed a significant
hot spot in the NE region of the surveyed area, nearest to Lummi Island (Mean
Shannon’s Diversity: 0.7571, mean Simpsons diversity: 1.8764) (Figure 14)
(Appendix C-1). The “Lummi” hot spot was comprised of 13 samples. Sample
60
represented all years surveyed except 2012. Four of the 13 samples were in the
deep depth stratum, while 9 samples were in the shallow depth stratum. Ten
species were observed in these samples from all three species groups, with total
abundance 665 fish. Because presences a single Quilback and Copper rockfish
were detected only in only two samples, these species were categorized as
Sebastes sp. for graphical representation (Table14). Members of the Gadidae
family contributed the most to abundance with 545 fish, however samples
dominated by Gadidae, were observed to have two or more species present, thus
reflecting in diverse samples (Appendix D -1).
Table 14. Species abundance for the13 Samples of the “Lummi” Hot Spot and 12
Samples of the “San Juan” Cool Spot
Species:
Lummi Abundance
San Juan
Abundance
Lingcod
3
6
Kelp Greenling
1
0
Pacific Cod
45
78
Pacific Tomcod
236
25
Hake
22
1
Walleye Pollock
309
821
Puget Sound rockfish
2
1
Sebastes spp.
2
6
Whitespotted Greenling
45
0
A noticeable clustering of “Cool” areas was observed in close proximity to
the San Juan Island Friday Harbor Marine Reserve (Appendix C-2). This
grouping consisted of 12 samples: 7 significant (GiZScores < -2.0, p < 0.05) and 5
non-significant samples (4 of which were very close to significant, p = 0.0502).
Due to the proximity of the samples to a large marine reserve, an examination of
the species composition was conducted. The samples represented all years
surveyed except 2004. Ten of the 12 samples were in the deep depth stratum,
61
while 2 samples were in the shallow depth stratum. Similar to the Lummi hot
spot, the “San Juan” Cool spot had species from all 3 species groups. Mean value
of the Shannon’s index was 0.313. Mean value of the Simpsons index was 1.2132
(Figure 14). The proportions at which species comprised samples were highly
dominated by a single species, Walleye Pollock. In total, 8 species were observed
with a total abundance of 938 fish caught over the 12 samples. No observations
of Whitespotted or Kelp Greenlings were made. Green and Redstripe Rockfish
were observed in 3 samples and were categorized as Sebastes spp. (Appendix D –
2)
Mean
Diversity
values
+-‐
SE
Lummi
Island
Hot
and
San
Juan
Cool
Spot
1.87636
2.50
2.00
1.50
1.21371
0.75715
1.00
0.50
0.31342
H
D
H
D
0.00
Lummi
Island
'Hot
Spot'
and
San
Juan
'Cool
Spot'
Figure 14. A comparison of the Lummi Island “Hot Spot” and San Juan Island
“Cool Spot”. Mean Shannon’s and Simpsons diversity value represented. See
Map in Appendix C for reference.
A similar comparison was conducted between two regions surrounding
Cypress Island (Appendix C-3). The first region was a “hot spot” area observed
off the SW shoreline of Cypress Island (Mean Shannon’s Diversity: 0.85652,
mean Simpsons diversity: 2.1829) (Figure 15). Samples represented years 2001,
62
2004, and 2006. Ten samples were shown to be significant (GiZScores > 2.0, p <
0.05), 3 of which were within the Cypress Island Aquatic Reserve. A total of 299
fish were observed between all ten samples. Though members of the Gadidae
family were the most abundant group, the majority of samples had more than one
species of Gadidae present, similar to the Lummi hot spot (Appendix D-3). All
but one of the ten samples was in the shallow depth stratum.
The second region was a grouping of non-significant samples North of
Cypress Island. Eleven samples were taken in this region, with a total of 831 fish
observed. Six of the samples occurred in the shallow depth stratum, while 5
occurred in the deep depth stratum. These 11 samples were not assigned any
significance during the running of the Getis –Ord Gi* test. Mean diversity values
for the 11 samples were less than the Cypress hot spot (Shannon’s Diversity:
0.4661, mean Simpsons diversity: 1.55406) (Figure 15). An examination of the
species composition showed samples dominated by Walleye Pollock (Appendix
D-4).
63
Mean
Diversity
values
+-‐
SE
Cypress
Island
Hot
Spot
and
Non-‐SigniNicant
Area
3.00
2.1829
2.50
2.00
1.50
1.55406
0.85652
1.00
D
H
0.4661
0.50
H
D
0.00
Cypress
Island'Hot
Spot'
and
Cypress'Non-‐SigniNicant'
Figure 15. A comparison of the Cypress Island SW “Hot Spot” and Cypress
Island N “Non-Significant spot”. Mean Shannon’s and Simpsons diversity value
represented. See Map in Appendix C for reference
Using Spatial Autocorrelation Global Moran’s I it was also possible to
determine that locations of samples with 0, 1, 2, 3, or 4 species per sample were
significantly clustered. This test was conducted for each species group, all of who
were shown to exhibit significant clustering by number of species present in each
sample (Table 15). This analysis method did not take into account abundance,
however, it did illustrate where more species were occurring for each species
group (Appendix E-1 to E-3).
Table 15. Results of the Spatial Autocorrelation Global Moran’s I: Significant
cluster patterns of samples with 0 to 4 species per species group.
Species Group
Moran’s Index
Z-score
p-value
Sebastes
0.124807
2.5376
0.011158
Gadidae
0.198056
3.9328
<0.0001
Hexagrammidae
0.336231
6.630993 <0.0001
64
Discussion
Frequency of Occurrence
Frequency of occurrence rates between the ROV data and Trawl data
differed drastically. No significant test were done between the two, however, the
occurrence frequencies were measured for the same species groups. The vast
majority, over 90% of observed species within all but one of the years of the trawl
surveys were members of the Gadidae family. In contrast, species of Gadidae
only made up 46% of the total observation for the 2010 ROV survey. This
variability in frequencies can be attributed to the utilization of two very different
survey techniques (number of samples collect and location of samples).
Depending upon depth, the trawl net opening would range between 8 and 14
meters, capturing the fish within its path (Palsson and colleagues 2009). The
ROV submersible has a much smaller “capture” area, this being represented by
the calculated transect width observed by the mounted camera (Pacunski et al.
2013). ROV transect width varied between 1 and 3 meters (Pacunski et al 2013).
Habitat and depth preferences were demonstrated for all species groups.
Within each species group, members of Gadidae, Hexagrammidae, and Sebastes
exhibit similar life traits (Love et al. 2002). For example, several species of
Sebastes are solitary and prefer nearshore rocky habitat, while other species
school, and are observed within a greater depth range ie: the Puget Sound
Rockfish (Yates 1988). The variation in life history traits of individual species
per family group was not examined for this analysis. By focusing on the general
65
traits like habitat and depth preference for each species group, it was possible to
observe significant variations between the three species groupings.
Several significant observations were made between species groups the
environmental variables analyzed in the 2010 ROV survey data. The observations
made with these analyses are supported by reviewed literature, and the basic
ecology of the three species groups studied (Pacunski et al. 2013, Palsson et al.
2009, Love et al. 2002). Within this study, two species groups, the genus
Sebastes and family Hexagrammidae are composed of species that are associated
with complex, structured, rocky habitat. The substrate and complexity categories
observed were similarly distributed in both depth strata, though over 90% of the
observations were made in the deep depth stratum. Species group composition
between the two depth strata was shown to be significantly different, with the
only preference for the shallow stratum being by members of the Hexagrammidae
family. The observed species within the Hexagrammidae family (Lingcod, Kelp
greenling, and Whitespotted greenling) are commonly found at shallower depths
(Yates 1998). In both depth strata, observations of any fish were most frequently
made while the ROV was over the bedrock substrate. Bedrock substrate offers
levels of habitat structure preferred by several rockfish (Sebastes) species
(Palsson et al. 2009). This was evident by the high frequency of Rockfish
observations, and the over all percentage of total organism observations while the
ROV was over the bedrock substrate. The family Gadidae represented roughly
46% of the total fish observation of the 2010 ROV survey. The four species
observed during the survey exhibit similar life history traits, such as school and
66
broadcast spawning (West, 1997). The data represented in this study
demonstrated that these fish prefer habitat with little complexity and structure as
the Gadidae were most frequently observed above a sandy substrate in the deep
depth stratum. The results of these occurrence rates for these three in the different
environmental variables lend to the interpretation of the diversity data analyzed in
conjunction with ROV data.
Measuring Diversity
The utilization of the Shannon’s and Simpsons diversity indices
represented a different approach to measuring the groundfish communities of the
San Juan Archipelago. The data used to compile and calculate these index values
are data used to measure the relative abundance, species composition, and
biological characteristics of key groundfish species (Palsson and colleagues
2009). As for the use of diversity indices, much care should be taken when
selecting which index to use (Yoccoz et al. 2001). The diversity indices chosen
for this analysis represent two that are commonly used to in the ecological studies
of community structure (Greene 1975, Lande 1996). The Simpsons index was
utilized as a measurement the probability that any two individuals within a sample
are different species. As a measurement of species diversity for a given number
of species, S, within a community (each sample) the index has a maximum value
equal to S when all species are equally present (Lande 1996). Similarly, the
Shannon’s index measure diversity for a given number of species, S, and reaches
a maximum value, the log-natural of S, when all species are equally frequent in
the community (Lande 1996). The Simpsons and Shannon’s index work well with
67
relatively small samples. The only samples that represented complete evenness,
while containing no diversity were samples containing only a single species.
These were kept in the analysis because these samples were representative of area
that could be dominated by heavy abundance of a single species. This
information is useful in understanding the distribution characteristic for some
groundfish species.
Measurements of the diversity values for both indices were low when
compared to the maximum potential value based on the number of species present
in each sample. The community structure between the three species groups is
such that the natural abundance of selected species would make for seemingly low
diversity. Gadidae have historically occurred at a much higher abundance than
Sebastes and Hexagrammidae with in the greater Puget Sound region (Palsson et
al. 1997, 50 CFR Parts 223 and 224. 2000). Although diversity values may be
mathematically low in comparison to the diversity indices maximum potential, the
significance of a sample with higher diversity were representative of low counts
of Gadidae species, namely Walleye Pollock, and higher abundance of Sebastes
and Hexagrammidae species.
Addressing temporal variation in the diversity of targeted groundfish
species is important for understanding factors that may affect community
structure. In this case, several species of groundfish composed a community of
fish that have been targeted by conservation efforts in the form of marine reserves
(Palsson et al. 2003). While significant temporal variations were observed in
diversity and evenness values between the eight sampled years, within-year
68
variation in diversity and evenness values was only shown significant in one year,
suggesting that changes diversity levels occur slowly. This slow change was also
demonstrated by the lack of significance in diversity values between the years
surveyed under the station based sampling method employed by WDFW. Though
these years consisted of smaller samples, they were surveyed consecutively,
allowing for a more sensitive means of detection for any regional change in
diversity. The observed significant variation in diversity values occurred over
time spans from 3 to 10 years from 2001. A significant decrease in mean
diversity was demonstrated between 2001, the first year surveyed, and all years
except 2006 and 2012. This downward trend in mean diversity values accurately
describe changes in community structure over time and space, yet detecting the
cause for such change becomes difficult when analyzing the community in this
fashion (Greene, 1975).
During the data collection process, WDFW changed sampling methods
from regional based sampling to station based (WDFW personal correspondence).
This affected the effort and number of samples collected in the San Juan
Archipelago each surveyed year and elapsed time between surveyed years.
Approximately 76% of the samples analyzed for the study were collected under
the regional sampling method. Thus the majority of diversity and evenness values
for this analysis were acquired from samples collected under the regional
sampling method. Testing the variability in diversity and evenness values by
sampling method showed that all diversity values, with the exception of the
Simpsons D index, were affected by the sampling strategy employed. However,
69
since diversity value variations were shown significant in year pairing between
the two methods, it is possible to attribute this variation to change in population
structure over time more so than sampling method.
Temporal Variation
The temporal variations observed in diversity and evenness values were
indicative of the rate at which community structure changes occur within the
ecosystem. However, an increase in time did not necessarily correspond to
greater significant variation. The most significant variation in diversity values
was observed between 2001 and 2008, with 2008 representing the lowest mean
diversity values recorded. 2008 also had the lowest number of total fish
abundance, and was the first year of the station based sampling method. Changes
in abundance are incorporated into the calculations for each diversity index;
however, this is assuming that all species have an equal probability of being
encountered (Yoccoz, et al 2001). With the survey returning to similar locations
each year, species encountered on a previous year had an equal probability of
occurring in subsequent years.
Similar to the significance of depth to occurrence frequencies, depth was
shown to greatly contribute to change in species composition and was negatively
correlated to species diversity. For the diversity values, the shallower samples
exhibited higher diversity and evenness levels. Several species of the three
groups: Sebastes, Gadidae, and Hexagrammidae inhabit the water column at
various depths (Yates, 1988). The most abundant species, Walleye Pollock, are
schooling member of the Gadidae family. These fish inhabit various depths,
70
though are often found in high abundance in deeper regions of the water column
(West, 1997). Because of this characteristic, Pollock were often the dominant
species in deep stratum samples, resulting in the relatively low diversity values
for deep stratum samples.
A counterpart to the temporal variation in diversity values, the spatial
distribution of diversity values was also shown to be indicative of sample
locations that repeatedly had significantly high or low levels of diversity. The
observations made with in this analysis are also reflective of areas that experience
more intense sampling. With the change from regional to station based sampling
in 2008, the survey vessel, F/V Chasina trawled in as close to the same location
(station) to minimize the variation in sea floor and habitat to provide more
powerful inter-annual comparisons (Palsson and colleagues 2009).
Spatial Distribution
The geographic distribution of significantly high and low diversity
groupings showed no influence by proximity to a marine reserve. However,
within this study it is not possible to conclude that marine reserves have no effect
on the diversity values. Descriptive comparisons between significantly “hot”
clusters and significantly “cool” clusters were made to describe the species
composition within these differing areas. The first two groups to be compared
were a “hot” grouping of samples near Lummi Island and a “cool” grouping of
sample very close to WDFW’s San Juan Friday Harbor Marine reserve. It was
shown that the cool area samples were dominated by a single gadoid species,
Walleye Pollock. Though the locations of the samples also were in close
71
proximity to an established marine reserve, the samples were taken in the deep
depth stratum favored by the Gadidae family. The compared “hot” grouping of
samples consisted of members of Gadidae as well, however, were species were
more evenly distributed throughout the samples. These samples were also
predominantly taken in the shallow depth stratum (West 1997).
Of the 188 samples, three fell within the boundaries of the Cypress Island
Aquatic Reserve, managed by the Washington Department of Natural Resources.
This reserve prohibits the harvest for all species of groundfish, and has been
shown to support a vast diversity of intertidal and nearshore fish species (WFC
2011, DNR 2007). The diversity values for the three samples were significantly
high, however, the reason for this may be confounded by several factors,
including the protection status of the area where these samples were taken. As
was shown, diversity values had a negative correlation with and increase in depth.
Shallower samples had higher levels of diversity due to the increased presence of
multiple species from all three species groups (Coleman et al. 1997). The three
samples within the reserve boundaries were statistically similar to seven more
samples, just outside the reserve boundary. In total, nine of the ten sample points
were in the shallow depth stratum. Though the Gadidae species group
demonstrated to prefer the deep depth stratum was present in these samples, their
presence was represented by 3 - 4 species being observed in some samples.
The temporal and spatial distribution of samples by species group gave a
visual representation for regions of the study area that continually had abundance
of more than one species within each species group. Groupings of samples with 1
72
– 2 species of Sebastes were concentrated in a few locations. Notably one
location was within close proximity to the San Juan Friday Harbor marine
reserve. This reserve, managed by WFDW, restricts the harvest of groundfish and
protect habitat suitable for Sebastes species (Van Cleve et al. 2009). It is possible
that the occurrence of more than 1 species per sample in this area was influenced
by the presence of fish using the habitat protected by this marine reserve.
This study demonstrates the significant spatial and temporal variation in
diversity values and to shifts in the species composition within the region.
Assessing diversity on a regional level for an ecosystem such as the San Juan
Archipelago is an important step in monitoring the groundfish community
variations. Spatial and temporal trends for the groundfish species represented here
are have been poorly understood throughout the Puget Sound region (Williams
2010). These trends are an important means of ecosystem monitoring (Palsson et
al. 2003, West 1997). However, the challenge in determining whether the shifts in
trends of diversity species composition, and habitat preference are related to the
implementation of marine reserves, are representative of natural shifts in
community structure, or are being influenced by a combination of factors, makes
assessing cause difficult. Further evaluation in the variation of diversity value
could be aided by testing potential confounding variables separately. For
example, research using similar diversity calculations, multivariate habitat
modeling and marine spatial planning has been used to map potential marine
reserve locations (Keith 2005, Douvere 2008). The potential for such work in the
San Juan Archipelago exist. Though few studies have been done to assess
73
regional effect of marine reserves, reserve provide an important mitigation
strategy for selected groundfish species (Tuya 2000). Combining diversity
information, from studies like the one presented here, to future studies of
abundance and habitat utilization will add to the understanding of groundfish
communities within the San Juan Archipelago.
74
Chapter 3: Conclusion
Biological Monitoring
Measuring and identifying patterns in biodiversity can be a powerful tool
in assessing an ecosystems capacity to provide for the organisms it harbors
(Primback 2010). Monitor diversity in areas where reserves have the potential to
augment the biological community structures outside the protected area is
especially needed to assess the ecological response to these management
strategies. However, accomplishing this task requires monitoring that collects a
vast amount of necessary information. Largely the objectives of monitoring
biological diversity can be assigned to two categories: scientific and management
(Yocozz et al. 2001). The data utilized for the analyses of this study were
collected in efforts to meet both these objectives.
As is common in fisheries management, ecological research is often
accomplished by utilizing any suitable data source (NRC, 2004). The
Washington Department of Fish and Wildlife have conducted trawl surveys of the
Puget Sound since 1987. These surveys have proven to be an invaluable fisheryindependent indicator for groundfish population abundance (Palsson and
colleagues 2009). Prior to these surveys, population data was derived from catch
landing reports (recreational and commercial), and was often not representative of
true population status for several species. In 1990, WDFW implemented the use
of marine reserves as a means to mitigate declining stocks of groundfish,
primarily species of the Sebastes genus. Efforts to protect these groundfish
species via the utilization of marine reserves are dependent on how the reserves
75
ecologically function. Tasked with managing five of the marine reserves
examined in this study, WDFW relies on biological surveys to detect changes in
population and community structure within their marine reserves. In years
following the implementation of several marine reserves, WDFW revised their
Puget Sound Groundfish Management Plan, adding emphasis to ecosystem-based
management and the conservation of biodiversity within the groundfish stocks
(Palsson et al. 1998).
Significance of this Study
The selection of the San Juan Archipelago region for this study was
important for two main reasons. First, this region has a history of established
marine reserves that have been subjected to several studies, thus data on past
species abundance was available (Van Cleve and colleagues 2009, Eisenhardt
2001). Second, the San Juan region has been previously subject to intense
recreational and commercial fisheries. Thus, marine reserves would potentially
have a greater benefit for species in this region over areas that were not subjected
to as intense of fishery. Though studies have examined diversity, abundance, and
community structure for groundfish populations between reserves and nonprotected waters within the San Juan Archipelago (Palsson et al., 2000), none
have focused on what large scale regional effects the marine reserves have on the
surrounding non-protected areas.
To hypothesize that a marine reserve would have an outward effect on its
surrounding environment requires that there be priori knowledge of how that
protected area functions in that ecosystem. WDFW surveys of marine reserves
76
within the San Juan Archipelago have demonstrated benefits such as increased
abundance, and size of groundfish within their boundaries (Palsson et al. 2003).
In addition to these two factors, larval dispersal can also greatly increase from
within marine reserves. The combination of these factors create was had been
called the “Reserve Effect” theory, stating that a protected population can be use
to supplement a non-protected population through larval dispersal and adult
spillover (Allison et al. 1998). The reserve effect may be highly species specific.
For example Black rockfish (Sebastes melanops), as species of Sebastes that
congregate over rocky habitat, were shown to exhibit lower levels of larval
dispersal than previously assumed by modeling, thus may not contribute progeny
to area outside of reserves (Miller and Shanks 2004). Similar larval dispersal
studies have observed this characteristic in other Sebastes species (Buonaccorsi et
al. 2002). The detection of reserve effect is often hampered by a lack of
knowledge about biotic and abiotic factors the level of biological and ecological
knowledge for the species. Within the San Juan Archipelago, referencing marine
reserves as larval distribution points has demonstrated significant variance in
dispersal potential base on locations of reserve and influences of surface currents
(Engie and Klinger 2007). Other abiotic factors such as El Niño Southern
Oscillation, thermocline layers, and ocean acidification need to be clearly
investigated as factors affecting the role of marine reserves for groundfish
communities (Sato and Wyllie-Echeverria 2004).
The approach of using the Shannon-Weiner and Simpson’s diversity
indices represented in this study allowed for trends in diversity and species
77
evenness to be observed over time, for three distinct families of groundfish. The
observed results of this study were only sensitive to the fact that changes in
community structure had occurred, but could not determine the cause for the
observed changes. An understanding of a “biological timeline” for the species of
concern may give clues as to how the community structure of these fish change
through time, and how these species will react to protection. With the exception
of the Gadidae family, the species examined in this study (Hexagrammidae and
Sebastes) exhibit slow growth, late maturity, and occur in relatively few numbers
through out the San Juan region (Love et al. 2002, Pacunski et al. 2013). These
fish are also high trophic level organisms and once mature are not subjected to
much predation. Using the methodology of this study to monitor temporal and
spatial variations for these species represents an important way to analyze trends
in their community structure.
This study emphasizes the use of quantifying biodiversity as a means of
detecting population trends over time and space. There are several other
techniques for monitoring community structures; however, biodiversity is often
associated with the health of the habitat. In an era where managing agencies such
as WDFW are shifting efforts towards ecosystem-based management, surveys that
monitor the broad scale changes within an ecosystem will be utilized.
The Continued use of MPAs
Though no observable changes in biodiversity could be attributed the
marine reserves of the San Juan Archipelago, I believe the continued utilization of
reserves should be a priority of management agencies. As is often the case with
78
several management strategies, stakeholders and invested peoples expect
relatively quick results. Within the context of this study, the five WDFW
managed reserves examined have only been established for 23 years (Van Cleve
et al. 2009). In some cases since these reserves were since 1992 collecting
population data (Palsson et al. 2003). As these reserves were implicitly designed
to protect rockfish and other fish associated with rocky habitat, the long-term
affect of their protective status may yet to be seen due to these species long life
and slow maturation characteristics. This may be evident by the lack in increase
in rockfish Sebastes sp. densities since surveys in the mid 1990’s were conducted,
however, Lingcod densities have significantly increased during this time (Palsson
et al. 2003). Variations like this demonstrate the need for continued observation
within these marine reserves, while also focusing attention to area outside the
protected boundaries.
The continued utilization and implementation of marine reserves in
Washington State is not without its own hindrances. Currently there are 11
different agencies, managing 127 MPAs within the State (Van Cleve et al. 2009).
The differences inherent to the diversity of management practices and goals lend
itself to disorganization between agencies in charge of the MPAs. However,
preliminary data from a survey sent to 57 stakeholders and invested peoples
(including managing agencies) suggest that the majority of correspondents believe
that MPAs can: be an effective too to conserve and manage marine resources in
the Puget Sound, and believe that a network of marine reserves should be
established within the State (Hanlon 2013, unpublished). Positive response to
79
marine reserve implementation is an encouraging sign that attention is being
focused towards the protection and conservation of groundfish species.
80
Reference
Allison, D.L. 2002. Problems with U.S. Ocean Governance and Institutional
Structures: The Impact on Waters, Fish, and Fisheries in the U.S.
Exclusive Economic Zone. In Managing Marine Fisheries in the United
States. Proceedings of the Pew Ocean Commission Workshop on Marine
Fishery Management
Allison, G.W., Lubchenco, J., Carr, M.H. 1998. Marine Reserves are Necessary
but not Sufficient for Marine Conservation. Ecological Applications. 8:1.
79-92
Babcock, R. C., S. Kelly, N. T. Shears, J. W. Walker, and T. J. Willis. 1999.
Changes in community structure in temperate marine reserves. Marine
Ecology Progress Series 189:125–134.
Bakun, A., & Broad, K. 2003. Environmental 'loopholes' and fish population
dynamics: comparative pattern recognition with focus on El Nino effects
in the Pacific. Fisheries Oceanography, 12(4‐5), 458-473.
Bargmann, G., W. Palsson, C. Burley, D. Friedel, and T. Tsou. (n.d.). Puget
Sound Rockfish Conservation Plan (PSRCP) and Final Environmental
Impact. Canadian Journal of Fisheries and Aquatic Sciences 61:2499–
2510.
Begg, G.A., Walmand, J.R. 1999. An holistic approach to fish identification.
Fisheries Research. 43: 35-44.
Begg, G.A., Friedland, K.D., Pearce, J.B. 1999. Stock identification and its role in
stock asesmnet and fisheries management: an overview. Fisheries Science
. 43: 1-8.
Berglund, M., M. Nilsson Jacobi, and P. R. Jonsson. 2012. Optimal selection of
marine protected areas based on connectivity and habitat quality.
Ecological Modeling 240:105–112.
Berkeley, S. A., C. Chapman, and S. M. Sogard. 2004. Maternal age as a
determinant of larval growth and survival in a marine fish, sebastes
melanops. Ecology 85:1258–1264.
Biological Review Team. 2009. Preliminary Scientific Conclusions of the Review
of the Status of 5 Species of Rockfish: Bocaccio (Sebastes paucispinis),
Canary Rockfish (Sebastes pinniger), Yelloweye Rockfish (Sebastes
ruberrimus), Greenstriped Rockfish (Sebastes elongatus) and Redstripe
Rockfish (Sebastes proriger) in Puget Sound, Washington. Northwest
Fisheries Science Center. National Marine Fisheries Service.
81
Boehlert, G.W. 2002. Improving Science in Marine Fishery Mangement: Looking
at Other Disciplines for Strategies to Develop New Models. Proceedings
of the Pew Ocean Commission Workshop on Marine Fishery Management
Bostad, P. 2008. GIS Fundamentals: A first Text on Geographic Information
Systems 3rd Edition. Eider Press, White Bear Lake, Minnesota, USA.
Botsford, L.W. 2009. Connectivity, Sustainability, and yield: Bridging the Gap
Between Conventional Fisheries Management and Marine Protected Areas.
Rev Fish Biol Fisheries. 19: 69-95
Buckley, R.M. 1997. Substrate Associated Recruitment of Juvenile Sebates in
Artificial Reef and Natural Habitats in Puget Sound and the San Juan
Archipelago, Washington. Washington Department of Fish and Wildlife
Buonaccorsi, V. P., Kimbrell, C. A., Lynn, E. A., & Vetter, R. D. (2002).
Population structure of copper rockfish (Sebastes caurinus) reflects
postglacial colonization and contemporary patterns of larval dispersal.
Canadian Journal of Fisheries and Aquatic Sciences, 59(8), 1374-1384.
Carr, M.H., Reed, D.C. 1993. Conceptual Issues Relevant to Marine Harvest
Refuges: Examples from Temperate Reef Fishes. Canadian Journal of
Fisheries and Aquatic Science. 50: 2019 - 2028
Chazdon, R. L., Peres, C. A., Dent, D., Sheil, D., Lugo, A. E., Lamb, D., ... &
Miller, S. E. 2009. The potential for species conservation in tropical
secondary forests. Conservation Biology, 23(6), 1406-1417.
Cheng, H., Niles, C., Palsson, W., Stick, K., Wallace, F. 2010. Washington
Contribution to the 2010 Meeting of the Technical Sub-Committee (TSC)
of the Canada-US Groundfish Committee. Washington Department of
Fish and Wildlife
Coleman, N., Gason, A.S, Poore, G.C. 1997. High Species Richness in the
Shallow Marine Waters of South-East Australia. Marine Ecology Progress
Series. 154: 17-26
Committee on Ecosystem Management for Sustainable Marine Fisheries. National
Research Council. 1999. Sustaining Marine Fisheries. National Academy
Press. Washington D.C. USA
Cooley, S. R., & Doney, S. C. 2009. Anticipating ocean acidification's economic
consequences for commercial fisheries. Environmental Research Letters,
4(2), 024007.
82
Cunningham, K. M., M. F. Canino, I. B. Spies, and L. Hauser. 2009. Genetic
isolation by distance and localized fjord population structure in Pacific
cod (Gadus macrocephalus): limited effective dispersal in the
northeastern Pacific Ocean. Canadian Journal of Fisheries and Aquatic
Sciences 66:153–166.
Dahl, R., C. Ehler, and F. Douvere. 2009. Marine Spatial Planning, A Step-byStep Approach toward Ecosystem-based Management. IOC Manuals and
Guides 53.
Dinnel, P., M. McConnell, I. Dolph, J. Ramaglia, and M. Sato. (n.d.). Rocky Reef
Bottomfish Reserves For Skagit County, Washington?
Douvere, F. 2008. The importance of marine spatial planning in advancing
ecosystem-based sea use management. Marine Policy 32:762–771.
Essington, T. E., A. H. Beaudreau, and J. Wiedenmann. 2006. Fishing through
marine food webs. Proceedings of the National Academy of Sciences of
the United States of America 103:3171–3175.
Eisenhardt, E. 2001. A marine preserve network in San Juan Channel: Is it
working for nearshore rocky reef fish. Scho of Aquatic and Fishery
Sciece, University of Washington.
Engie, K., & Klinger, T. (2007). Modeling passive dispersal through a large
estuarine system to evaluate marine reserve network connections.
Estuaries and coasts, 30(2), 201-213.
Foley, M. M., B. S. Halpern, F. Micheli, M. H. Armsby, M. R. Caldwell, C. M.
Crain, E. Prahler, N. Rohr, D. Sivas, M. W. Beck, M. H. Carr, L. B.
Crowder, J. Emmett Duffy, S. D. Hacker, K. L. McLeod, S. R. Palumbi,
C. H. Peterson, H. M. Regan, M. H. Ruckelshaus, P. A. Sandifer, and R. S.
Steneck. 2010. Guiding ecological principles for marine spatial planning.
Marine Policy 34:955–966.
Goldberg, M.B. 2002. Optimal Yield: A Goal Honored in the Breach. Proceedings
of the Pew Ocean Commission Workshop on Marine Fishery
Management
Greene, C. S. 1975. A comparison of diversity indices. Coastal Water Research
Project. Annual report for the year ended, 30.
Gustafson R.G., W.H. Lenarz, B.B. McCain, C.C. Schmitt, W.S. Grant, T.L.
Builder, and R.D. Methot. 2000. Status review of Pacific Hake, Pacific
Cod, and Walleye Pollock from Puget Sound, Washington. U.S. Dept.
Commer., NOAA Tech. Memo. NMFS-NWFSC - 44, 275 p.
83
Hanlon, E. 2013. Perception of Marine Protected Areas in the Puget Sound.
Masters Thesis. The Evergreen State College
Herneman, B. 2002. Federal Fisher Laws: New Model Needed to Sustain
Fisheries and Ecosystems. Managing Marine Fisheries in the United
States. Proceeding of the Pew Oceans Commission Workshop on Marine
Fishery Management
Hilborn, R. et al. 2004. When can marine reserves improve fisheries
management? Ocean and Coastal Management. 47: 197-205.
Hildreth, R. 2002. U.S. and International Fisheries Law: The Role of
Sustainability, Biodiversity Protection, Externality Internalization, and
Precaution. In Managing Marine Fisheries in the United States.
Proceedings of the Pew Ocean Commission Workshop on Marine Fishery
Management
Hilborn, R. Hilborn, U. 2012. Overfishing: What Everyone Needs to Know.
Oxford University Press. New York, NY, USA.
Hirzel, A. H., J. Hausser, D. Chessel, and N. Perrin. 2002. Ecological-Niche
Factor Analysis: How To Compute Habitat-Suitability Maps Without
Absence Data? Ecology 83:2027–2036.
Jackson, J. B., Kirby, M. X., Berger, W. H., Bjorndal, K. A., Botsford, L. W.,
Bourque, B. J., ... & Warner, R. R. 2001. Historical overfishing and the
recent collapse of coastal ecosystems. science, 293(5530), 629-637.
Jentoft, S., McCay, B. J., & Wilson, D. C. 1998. Social theory and fisheries comanagement. Marine Policy, 22(4), 423-436.
Johannes, R.E., Freeman, M.R, Hamilton, R.J. 2000. Ignore Fishers Knowledge
and Miss the Boat. Fish and Fisheries. 1:257-271
Keith, C.M. 2005. GIS Modeling Potential Marine Protected Areas in the
Northwest Atlantic via Biological and Socioeconomic Parameters. Masters
Thesis. Oregon State University, Corvallis, OR.
Kracker, L. M. 1999a. The Geography of Fish: The Use of Remote Sensing and
Spatial Analysis Tools in Fisheries Research. The Professional
Geographer 51:440–450.
Kurlansky, M. 1997. Cod: A Biography of the Fish That Changed the World. The
Penguin Group, New York, Ny.
84
Lande, R. 1996. Statistics and Partitioning of Species Diversity, and Similarity
Among Multiple Communities. Nordic Society Oikos. 76: 5-13
Love, M. S., Yoklavich, M., & Thorsteinson, L. K. (2002). The rockfishes of the
northeast Pacific. University of California Press.
Lubchenco, J., S. R. Palumbi, S. D. Gaines, and S. Andelman. 2003. Plugging A
Hole In The Ocean: The Emerging Science Of Marine Reserves.
Ecological Applications 13:3–7.
Macinko, S., Hennessey, T. 2002. 11 Questions and some Partial
Answers/Thoughts. Proceedings of the Pew Ocean Commission Workshop
on Marine Fishery Management
Magurran, A. E. (2004). Measuring biological diversity. Blackwell, Maldan, MA
Melvin, E. F., and J. K. Parrish. 2001. Seabird bycatch. Trends, roadblocks and
solution. University of Alaska Sea Grant, Fairbanks, Alaska.
Micheli, F. et al. 2004. Trajectories and Correlates of Community Change in NoTake Marine Reserves. Ecological Applications. 14(6): 1709-1723
Miller, J. A., & Shanks, A. L. (2004). Evidence for limited larval dispersal in
black rockfish (Sebastes melanops): implications for population structure
and marine-reserve design. Canadian Journal of Fisheries and Aquatic
Sciences, 61(9), 1723-1735.
Mills, M. L., and B. MacDonald. 2004. Ecoregional Conservation Planning in the
Marine Environment. 2003 Georgia Basin/Puget Sound Resarch
Conference Proceedings.[np]. Feb 2004.
Moffitt, E. A., L. W. Botsford, D. M. Kaplan, and M. R. O’Farrell. 2009. Marine
reserve networks for species that move within a home range. Ecological
Applications 19:1835–1847.
Mosqueira, I. et al. 2000. Conservation benefits of marine reserves for fish
populations. Animal Conservation. 4: 321-332
Myers, R., Worm, B. 2003. Rapid Worldwide Depletion of predatory Fish
Communities. Letters to Nature. 423
National Resource Council. 1999. Sustaining Marine Fisheries. National
Academy Press, Washington D.C, USA
National Research Council. 2004. Cooperative Research in the National Marine
Fisheries Service. National Academy Press, Washington D.C, USA
85
National Research Council. 2004. Improving the Use of the "Best Scientific
Information Available" Standard in Fisheries Management. National
Academy Press, Washington D.C, USA
O'Farrell, M.R., Botsford, L.W. 2006. The Fisheries Management Implications of
Maternal-Aged Dependent Larval Survival. Canadian Journal of Fisheries
and Aquatic Science. 63(10): 2249 - 2258
Pacunski, R., Palsson, W., Greene, H. 2013. Estimating fish abundance and
community composition on rocky habitats in the San Juan Islands using a
small remotely operated vehicle. Washington Department of Fish and
Wildlife.
Palsson, W.A., Northrup, T.J., Baker, M.W. 1998. Puget Sound Groundfish
Management Plan. Washington Department of Fish and Wildlife
Palsson, W.A. 2001. The Development of Criteria for Establishing and
Monitoring No-take Refuges for Rockfish and Other Rocky Habitat Fishes
in Puget Sound. Washington Department of Fish and Wildlife
Palsson, W., R. E. Pacunski, and T. R. Parra. 2003. Time will tell: long-term
observations of the response of rockyhabitat fishes to marine reserves in
Puget Sound. Georgia Basin/Puget Sound Research Conference
Proceedings.
Palsson, W. A., J. Beam, S. Hoffmann, P. Clarke, T. W. Droscher, and D. A.
Fraser. 2004. Fish without borders: trends in the status and distribution of
groundfish in the transboundary waters of Washington and British
Columbia. Proceedings of the 2003 Georgia Basin/Puget Sound Research
Conference.
Palsson, W.A., Tsou, T., Bargmann, G., Buckley, R.M., West, J., Mills, M.,
Cheng, Y., Pacunski, R. 2009. The biology and assessment of Rockfishes
in Puget Sound. Washington Department of Fish and Wildlife.
Palumbi, S. R. 2001. The ecology of marine protected areas. In Marine
Community Ecology, M. Bertness, S. Gaines, and M. Hay, Eds. Sinauer
Press, Sunderland, MA pp 509-530.
Palumbi, S.R. 2004. Marine Reserves and Ocean Neighborhoods: The Spatial
Scale of Marine Populations and Their Management. Annual Reviews of
Environmental Resource. 29:31–68
Pauly, D. et al. 2002.Towards Sustainability in World Fisheries. Nature. 481:689695
86
Pauly, D., Watson, R. 2003. Counting the Last Fish. Scientific American. 289: 4247
Pereira J., and R. Itami. 1991. GIS-based habitat modeling using logistic multiple
regression- A study of the Mt. Graham red squirrel. Photogrammetric
Engineering and Remote Sensing 57:1475–1486.
Pinnegar, J.K., Engelhard, G.H. 2008. The 'Shifting Baseline' Phenomenon: A
Global Perspective. Review of Fish Biology Fisheries. 18: 1-16
Primback, R.B. 2010. Essentials of Conservation Biology 5th Edition. Sinauer
Associates, Massachusetts, USA
Roberts, C.M. 2000. Selecting Marine Reserve Locations: Optimality Versus
Opportunism. Bulletin of Marine Science. 66(3): 581-592
Roberts, C.M, et al. 2001. Effects of Marine Reserves on Adjacent Fisheries.
Science. 294, 1920
Rosenberg, A. 2003. Managing to the Margins: The Overexploitation of Fisheries.
Frontiers in Ecology. 1:2 102-106
Rice, C.A. 2007. Evaluating the Biological Condition of Puget Sound. Doctor of
Philosophy Dissertation. University of Washington. Seattle, WA.
Rice, C. A., J. J. Duda, C. M. Greene, and J. R. Karr. 2012. Geographic Patterns
of Fishes and Jellyfish in Puget Sound Surface Waters. Marine and
Coastal Fisheries 4:117–128.
Sato, M., Wyllie-Echeverria, T. 2004. Best Science and Best Management for
Rockfish and Lingcod in the San Juan County. Prepared for the San Juan
County Marine Resource Committee
Scheiber, H.N. 2002. Bringing the Community Back In: The Next Step in Fishery
Management. Proceedings of the Pew Ocean Commission Workshop on
Marine Fishery Management
Shanks, A. L. 2009. Pelagic Larval Duration and Dispersal Distance Revisited.
The Biological Bulletin 216:373–385.
Shanks, A.L. et al. 2003. Propagule Dispersal Distanceand the Size and Spacing
of Marine Reserves. Ecological Society of America. 13(1):s159-s169
Sobel, J. A., & Dahlgren, C. P. (2004). Marine reserves: A guide to science,
design, and use. Washington, D.C: Island Press.
87
SOS2012_all_110112.pdf.
Sumaila, U.R, et al. 2000. Addressing Ecosystem Effects of Fishing Using Marine
Protected Areas. ICES Journal of Marine Science. 57:752-760
Swain et al. 2005 Stock Identification Methods: Applications in Fisheries Science.
Cadrin, S. Friedland, K. Waldman, J. EDS) Elsevier Academic Press.
Burlington, MA, USA
Tetreault, I., Ambrose, R.F. 2007. Temperate Marine Reserves Enhance Targeted
but Not Untargeted Fishes in Multiple No-Take MPAs. Ecological
Applications. 17(8) 2251-2267
Tilden, J. 2004. Marine Geology and Potential Rockfish Habitat in the
Southwestern San Juan Islands, Washington. Masters Thesis. California
State University Monterey Bay. Moss Landing Marine Laboratories.
Tsao, C.F., Morgan, L.E., Maxwell, S. 2005. The Puget Sound/Georgia Basin
Region Selected as a Priority Conservation Area in Baja California to
Bering Sea Initiative. Proceedings of the 2005 Puget Sound Georgia Basin
Research Conference. Marine Conservation Biology Institute
Tuya, F. C., M. L. Soboil, and J. Kido. 2000. An assessment of the effectiveness
of Marine Protected Areas in the San Juan Islands, Washington, USA.
ICES Journal of Marine Science: Journal du Conseil 57:1218–1226.
Valavanis, V. D., G. J. Pierce, A. F. Zuur, A. Palialexis, A. Saveliev, I. Katara,
and J. Wang. 2008. Modelling of essential fish habitat based on remote
sensing, spatial analysis and GIS. Hydrobiologia 612:5–20.
Van Cleve, FB, G Bargmann, M Culver, and the MPA Work Group. 2009.
Marine Protected Areas in Washington: Recommendations of the Marine
Protected Areas Work Group to the Washington State Legislature.
Washington Department of Fish and Wildlife, Olympia, WA
Washington Administrative Code. 220 - 48 -015.
Washington State Department of Natural Resources. 2007. Cypress Island
Comprehensive Management Plan.
Waldman, J. 2005. Stock Identification Methods: Applications in Fisheries
Science. (Cadrin, S. Friedland, K. Waldman, J. EDS) Elsevier Academic
Press. Burlington, MA, USA
88
Walters, S., H. Cornell, N. Hamel, E. Knudsen, J. Lombard, and C. Steward.
(n.d.). Science synthesis in support of ecosystem-based management: The
Puget Sound Science Update.
West, J.E. 1997. Protection and Restoration of Marine Life in the Inland Waters
of Washington State. Puget Sound/Gorgia Basis Environmental Report
Series:6
West, J. E., R. M. Buckley, and D. C. Doty. 1994. Ecology and Habitat Use of
Juvenile Rockfishes (Sebastes spp.) Associated with Artificial Reefs in
Puget Sound, Washington. Bulletin of Marine Science 55:344–350.
Washington Department of Fish and Wildlife. 2011. Final Puget Sound Rockfish
Conservation Plan. Policies, Strategies, and Actions Including Preffered
Range Actions.
Wild Fish Conservancy. 2011. Cypress Island Aquatic Reserve. Pilot Nearshore
Fish Use Assessment. Technical Report. Wild Fish Conservancy
Northwest
Wilkinson, C. 2000. Message from Frank's Landing: a Story of Slamon, treaties,
and the Indian Way. University of Washington Press, Seattle, Washington,
USA
Williams, G.D., Levin, P., Palsson, W.A.2010. Rockfish in Puget Sound:
Ecological History of Exploitation. Marine Policy. 34:1010-1020
Worm, B., Barbier, E. B., Beaumont, N., Duffy, J. E., Folke, C., Halpern, B. S., ...
& Watson, R. (2006). Impacts of biodiversity loss on ocean ecosystem
services. science, 314(5800), 787-790.
Wright, D.J., Heyman, W.D. 2008. Introduction to the Special Issue: Marine and
Coastal GIS for Geomorphology, Habitat Mapping, and Marine Reserves.
Marine Geodesy. 31: 1-8
Yates, S. 1988. Marine Wildlife of the Puget Sound, the San Juans, and Strait of
Georgia. The Globe Pequot Press, Chest, Connecticut, USA
Yoccoz, N.G., Nichols, J.D., Boulinier, T. 2001. Monitoring of Biological
Diversity in Space and Time. Trends in Ecology and Evolution. 16:8 446453
50 CFR Parts 223-224.2000. Endangered and Threatened Species: Puget Sound
Population of Hake, Pacific Cod, and Walleye Pollock. Federal Register.
65:227
89
90
Appendix A
91
A-1. Study Area
92
A-2. WDFW and DNR Marine Protected Area Locations
93
Appendix B
94
B-1. Shannon’s Diversity Map: Significant Grouping
95
B-2. Simpsons Diversity Map: Significant Grouping
96
Appendix C
97
C-1. Location of the Lummi Island Hot Spot: Area of significantly high diversity
98
C-2. Location of the San Juan Island Cool Spot: Area of significantly low
diversity
99
. C-3 Location of the Cypress Island Hot and Non-Significant Spots
100
Appendix D
101
D-1. Species composition for the Lummi Island “Hot Spot”
D-2. Species Composition for the San Juan Island “Cool Spot”
102
D-3. Species composition for the Cypress Island SW area “Hot Spot”
D-4. Species Composition for the Cypress Island Northern Non-Significant group
103
Appendix E
104
E-1. Sebastes: occurrence by number of species present per sample
105
E-2. Hexagrammidae: occurrence by number of species per sample
106
E-3. Gadidae: occurrence by number of species per sample
107