-
Identifier
-
Thesis_MES_2019_CallowayM
-
Title
-
Nereocystis Luetkeana (Bull kelp) in South Puget Sound: Stressor Impacts on the Health of Native Floating Kelp Canopies
-
Date
-
May 2019
-
Creator
-
Calloway, Maxwell
-
extracted text
-
NEREOCYSTIS LUETKEANA (BULL KELP) IN SOUTH PUGET SOUND:
STRESSOR IMPACTS ON THE HEALTH OF NATIVE FLOATING KELP
CANOPIES
by
Maxwell D. Calloway
A thesis submitted in partial fulfillment of the requirements for the degree of Master of
Environmental Studies
May 2019
i
© Maxwell D. Calloway. All rights reserved.
ii
This Thesis for the Master of Environmental Studies Degree
by
Maxwell D. Calloway
has been approved for
The Evergreen State College
by
_________________________________
Erin Martin
Member of the Faculty
_________________________________
[Date]
iii
ABSTRACT
Nereocystis luetkeana (bull kelp) in South Puget Sound: Stressor Impacts on the Health
of Native Floating Kelp Canopies
Maxwell D. Calloway
Bull kelp (Nereocystis luetkeana), is a critical, habitat-forming, floating seaweed found
along Salish Sea’s shorelines. East of the Strait of Juan de Fuca, Nereocystis is the sole
floating canopy forming seaweed. Unfortunately, its abundance in Puget Sound is declining
for unknown reasons. Little research exists on Nereocystis within Puget Sound, an inland
sea with distinct environmental conditions and stressors from outer coast environments. In
addition, stressors often interact synergistically in marine environments, compounding
negative impacts more than would be predicted from single stressors alone. To assess the
causes of decline in the Puget Sound, this study monitored Nereocystis blade length and
plant density at four sites spanning a north to south gradient within the South Puget Sound.
In addition, several common Puget Sound stressors were assessed including temperature,
nitrate availability and densities of the native kelp crab, Pugettia producta. Generally,
stress intensity followed a geographic gradient, increasing in the more interior waters of
the South Puget Sound. High temperatures and crab densities were correlated with
significant declines in blade length although no significant interaction between the two was
observed. In addition, the bed exposed to the highest temperatures and crab densities was
characterized by significantly lower densities of plants than any other bed selected for
density monitoring. This study also documented the loss of a surface canopy in the South
Puget Sound. Given that ocean temperatures are predicted to continue to rise over the next
decades, it is likely that Puget Sound Nereocystis forests will continue to decline unless
conservation and recovery actions are implemented.
TABLE OF CONTENTS
Introduction.........................................................................................................................1
Literature Review ................................................................................................................3
Puget Sound: Place and Processes ........................................................................................... 4
Nereocystis lifecycle ................................................................................................................... 6
Abiotic controls: Environmental conditions and physical forces .......................................... 9
Photosynthetic performance...................................................................................................................9
Temperature .........................................................................................................................................10
Nutrients ...............................................................................................................................................12
Biotic controls in the SPS ........................................................................................................ 13
Washington State kelp............................................................................................................. 15
Methods .............................................................................................................................18
Site Selection ............................................................................................................................ 18
Mean Water Column Temperature ....................................................................................... 21
Nitrate availability ................................................................................................................... 22
Bed-wide and Nereocystis blade length monitoring .............................................................. 22
Analysis ..................................................................................................................................... 23
Results................................................................................................................................25
Survey Timing .......................................................................................................................... 25
Site differences ......................................................................................................................... 27
Plant Density ........................................................................................................................................27
Blade Length ........................................................................................................................................29
Temperature .........................................................................................................................................32
Pugettia producta densities ..................................................................................................................35
Nitrate Concentrations .........................................................................................................................37
Interactions Between Temperature, Crab Density and Blade Length................................ 39
Discussion..........................................................................................................................43
Plant and bed condition .......................................................................................................... 43
Environmental and biotic stressors........................................................................................ 44
Devil’s Head, substrate availability and bed area ................................................................ 47
Methods Assessment ................................................................................................................ 48
Continued monitoring ............................................................................................................. 49
Conclusion.........................................................................................................................49
References .........................................................................................................................51
Appendix ............................................................................................................................57
iv
LIST OF FIGURES
Figure 1: Map of the Puget Sound showing major basins (Gelfenbaum et al. 2006) ......... 5
Figure 2: South Puget Sound and study sites. ................................................................... 20
Figure 3: Site monitoring design....................................................................................... 21
Figure 4. Plant densities at each site (black line = median, white diamond = mean). ...... 27
Figure 5. Monthly plant densities at each site (mean ± SE). ............................................ 29
Figure 6: Blade length at each site (black line = median, white diamond = mean). ......... 30
Figure 7: Average monthly blade lengths of Nereocystis individuals at each site (mean +
SE)..................................................................................................................................... 31
Figure 8: Average monthly integrated water column temperatures.................................. 33
Figure 9: Crab densities at each site (black line = median, pink diamond = mean) ......... 36
Figure 10: Average monthly crab densities at each site (mean ± SE). ............................. 36
Figure 11: Growing season (May–September) water nitrate concentrations at study sites
(black line = median, white diamond = mean). ................................................................ 38
Figure 12: Monthly growing season nitrate concentrations (µM) at two depths (0.25 m
and 4 m) at study sites. ...................................................................................................... 38
Figure 13: Blade length as a function of water temperature with regression trend line ... 40
Figure 14: Blade length as a function of crab densities with regression trend line .......... 40
Figure 15: Squaxin Island blade length as a function of temperature with trend line of
predicted values from regression analysis. ....................................................................... 42
LIST OF TABLES
Table 1. Plant density per square meter as a function of site: Sample size, mean, standard
deviation, standard error, median, first quartile, third quartile and interquartile range. ... 27
Table 2. Blade length (m) as a function of site: Sample size, mean, standard deviation,
standard error and 95% confidence interval for ................................................................ 30
Table 3. Temperatures (°C) as a function of survey month and site with Games-Howell
pairwise post hoc differences*: Sample size, mean, standard deviation and maximum
integrated water column .................................................................................................... 34
Table 4. Crab densities as a function of site with Games-Howell pairwise post hoc test pvalues: Sample size, mean, standard deviation, standard error......................................... 35
Table 5. Nitrate concentrations as a function of site: Sample size, mean standard
deviation, standard error and 95% confidence interval .................................................... 37
v
Table 6. Nereocystis blade length, temperature and crab densities at all sites: Multiple
linear-regression with quadratic polynomial treatment of temperature, and interaction
between crab densities and average temperature .............................................................. 39
Table 7. Nereocystis blade length, temperature and crab densities at Squaxin Island:
Multiple linear-regression with quadratic polynomial treatment of temperature, and
interaction between crab densities and average temperature ............................................ 42
Table 8. Plant density per square meter as a function of site and survey date: sample size,
mean, standard deviation, standard error, median, first quartile, third quartile and
interquartile range of ......................................................................................................... 57
Table 9. Blade length (m) as a function of site and survey month: sample size, mean,
standard deviation, standard error, median and 95% confidence interval ........................ 58
Table 10. Crab densities as a function of site and survey month: sample size, mean,
standard deviation, standard error, median, first quartile, third quartile and interquartile
range .................................................................................................................................. 59
Table 11. Nitrate concentrations (µM) as a function of survey month and depth at each
site: sample size, mean, standard deviation, standard error and 95% confidence interval 60
vi
ACKNOWELDGMENTS
This project owes its success to the absolutely amazing network of mentors, friends
and peers in Olympia. First and foremost, I would like to thank my advisory committee Dr.
Erin Martin of the Evergreen State College, Helen Berry of the Washington Department of
Natural Resources Submerged Aquatic Vegetation program and Dr. Tom Mumford
(former WA-DNR). I am indebted to their expertise, guidance and trust in my ability to
complete a field-based project with a shoestring budget, bubble-gum and scotch tape. I owe
special thanks to Helen Berry whose collaboration made the field component of this project
possible. Thanks also to John Withey, and Bart Christiaen and Pete Dowty of WA-DNR
for their willingness to brainstorm on statistical methods. Thank you to Erin Martin for
believing that a scruffy dirt-bag with a degree in literature could make the transition to a
career in science and for being patient as the writing component lasted longer than
expected. And, of course, thank you to my wonderful Olympia family: Wolf and Rainier
Mithrandir, and Tasa Anderson for reminding me to always look at the world with the
curiosity and wonder of a child; John and Alexandria Messina for letting me steal their
kayak for extended periods of time with no expectation that it, or I would ever come back
but always suggesting we cook dinner, relax and enjoy house-plants on a front porch on
summer evenings when I did; John Messina, Averi Azar, Keegan Michael Curry, Kirsten
Miller, Julian Wischniewksi and Simon Young for their willingness to let me whisk them
to kelp beds in a tiny dinghy with a small motor with a bad habit of over-heating. Thank
you also to Kyle Grosten for volunteering his boat and time for field support. Finally,
special thanks to the kelp forests at Salt Creek State Park for inspiring me to study kelp in
the first place and Dr. Tim Quinn of the Washington Department of Fish and Wildlife for
confirming that this whole crazy adventure was a good idea in the first place.
vii
INTRODUCTION
The bull kelp, Nereocystis luetkeana (Mertens) Postels & Ruprecht (hereafter
Nereocystis) is a monotypic, annual, kelp (order Phaeophyceae, class Laminariales) found
along the temperate, eastern Pacific coast from southern California to Alaska (Druehl &
Clarkston 2016). Nereocystis, along with the perennial giant kelp Macrocystis pyrifera
(hereafter Macrocystis), form dense floating canopies throughout rocky, subtidal
environments in the Pacific Northwest (PNW). Mixed canopies in Washington state occur
along the outer coast and into the Strait of Juan de Fuca, but Nereocystis is the only floating
canopy forming species found in the Puget Sound (Berry 2017).
Anecdotal accounts from local residents, tribes and management agencies suggest
that Nereocystis canopies in the South Puget Sound (SPS) – the southernmost basin of the
Puget Sound incorporating all waters south of the Tacoma Narrows (Fig. 1) – are declining
in abundance and linear extent. These accounts have been confirmed by recent analyses of
historical survey data conducted by the Washington Department of Natural Resources
(WA–DNR) that document a 67 percent decrease in Nereocystis canopy linear extent from
1873 to 2017 (Berry 2018). In light of these considerable losses of SPS Nereocystis
canopies, understanding causes for observed declines is paramount if conservation and
restoration measures are to be successfully implemented. Unfortunately, no data exists on
the response of Puget Sound bull kelp to common stressors such as temperature, nutrient
availability and grazing pressure – three parameters known to significantly effect kelp
populations elsewhere in North America and abroad (Steneck et al. 2002).
Assessing the impacts of stressors on kelp can be difficult as the majority of kelp
species live out their entire lives in subtidal waters, requiring the use of SCUBA surveys
1
(Gabrielson et al. 2012). Even surveys of Macrocystis often require subtidal observations
as all specialized reproductive blades (sporophylls) occur at the base of the plant (Schiel &
Foster 2015). Nereocystis is unique in that it produces all of its reproductive tissue along
vegetative blades that grow only at the surface, making surface assessments of stressor
impacts comparatively less labor intensive than with other kelp species (Druehl &
Clarkston 2016). This also means that surface monitoring of blade parameters (length,
weight, proportion of blades with reproductive sorus present, etc.) encompasses a majority
of individual biomass.
This study monitored Nereocystis canopy density, blade length, temperature, nitrate
concentrations and abundances of a common SPS grazer (the northern kelp crab, Pugettia
producta), in order to assess if temperature, nutrient availability and grazing pressure are
associated with decreased blade lengths and bed densities. Oceanographic conditions in the
Puget Sound display a strong geographic gradient with temperature decreasing, and
nutrient concentrations and salinity increasing from south to north (Berry 2018). Given
these known geographic gradients, we sought to answer two questions: Do environmental
conditions, Nereocystis canopy density and plant condition differ between study sites?
And, are differences in temperature, nutrient concentrations and grazing pressure
correlated with differences in Nereocystis blade length? We hypothesized that
environmental conditions, canopy density and blade length would differ along a geographic
gradient with the most stressful conditions, shortest blade lengths and lowest plant densities
occurring in the southernmost interior reaches of the SPS. We also hypothesized that high
temperatures, low nitrate availability and dense aggregations of kelp crabs would be
correlated to shorter blade lengths.
2
LITERATURE REVIEW
The ‘true’ kelps (order Phaeophyceae, class Laminariales) form large, biogenic,
coastal habitats across the globe. These underwater forests serve as foundations for a
stunning diversity of marine flora and fauna, engineering conditions that increase overall
biodiversity. The largest species found in the northeast Pacific – Macrocystis and
Nereocystis – grow tens of meters tall forming a dense floating canopy that slows water
movement and reduces the amount of light reaching the forest floor (benthos).
Nereocystis is the only floating canopy forming kelp found in the Puget Sound and
has been recently identified as critical habitat for the federally endangered Boccaccio
(Sebastes paucispinis) and Yelloweye Rockfish (Sebastes ruberrimus) (National Marine
Fisheries Service 2017). These forests are also crucial foundations for Puget Sound food
webs that indirectly support healthy salmon and ocra populations (Bertocci 2015). Kelps
have also been identified as sentinel species, sensitive to temperature increases and water
quality degradation (Steneck et al. 2002). Recent documented losses to Puget Sound bull
kelp forests are cause for concern and identifying causes and consequences of these losses
is a high priority to federal and local resource managers, and non-profits.
The Puget Sound is unique from an oceanographic perspective, composed of a
number of fjordal basins and subbasins separated by narrow and shallow constriction
points. Little is known about kelp in the context of this unique environment (Ebbesmeyer
et al. 1988). However, the large body of available research on closely related species in
similarly oceanographic climates provides important insight into the kelp response to biotic
and abiotic stress (Schiel & Foster 2015; Steneck et al. 2002; Dayton 1985).
3
This literature review synthesizes available information on Nereocystis, Puget
Sound kelp species and ecosystems as well as responses to stress from closely related kelp
from other geographic and oceanographic contexts. It begins with a brief overview of the
physical conditions common to the Puget Sound before providing a summary of the
Nereocystis lifecycle and discussing the response of Nereocystis and other kelp species to
common abiotic and biotic stressors. Finally, it concludes with a summary of available
Nereocystis trend and distribution data and ongoing monitoring efforts in the Puget Sound.
Puget Sound: Place and Processes
Puget Sound is an estuary made up of four glacially scoured basins (fjords)
connected by shallow, constricted passages: The Central Puget Sound, Whidbey Basin,
South Puget Sound (SPS) and the Hood Canal (Ebbesmeyer et al. 1988; Fig. 1). The
shallow waters of Admiralty Inlet separate the Strait of Juan de Fuca from the Puget Sound
proper (waters inland of Admiralty Inlet) and Hood Canal. Admiralty Inlet also forms the
junction between the Hood Canal, Central Puget Sound and Whidbey Basin. The Tacoma
Narrows serves as the boundary between the Central and South Puget Sound.
The Central and South Puget Sound are considered well mixed systems, but the
waters of the Hood Canal and Whidbey basin can become stratified due to less water
mixing and large inputs of freshwater (Williams et al. 2001). A majority of water mixing
occurs at constriction points due to turbulent, vertical mixing of the water column as water
from deep basins is forced into shallow waters as it attempts to exit the basin of origin. As
a result of this mixing, approximately 50 percent of water in a given basin is retained
resulting in residence times of three to six months for dissolved materials (Ebbesmeyer et
al. 1988).
4
Oceanographic conditions in the Central and South Puget Sound follow a distinct
geographic gradient with waters becoming cooler and more nutrient rich as one travels
north. These gradients are even detectible within basins – water temperatures were 2 to 3
°C cooler in the Tacoma Narrows, at the entrance to the South Puget Sound,m than the
waters of Budd Inlet near Olympia, WA (Berry et al. 2019).
Figure 1: Map of the Puget Sound showing major basins (Gelfenbaum et al. 2006)
5
While environmental trends on a climatological scale explain long-term variation
in kelp trends in the Strait of Juan de Fuca (Pfister et al. 2017), a review of the global kelp
literature found that local environmental conditions explained local forest variation best
(Krumhansl et al. 2016). This characteristic gradient of environmental conditions in the
Puget Sound make it an ideal location for natural experiments to investigate the role of
different environmental conditions on Nereocystis persistence and condition.
Nereocystis lifecycle
Like most kelp, Nereocystis exhibits a heteromorphic life history (Schiel & Foster
2006). This simply means that kelp alternate between macroscopic and microscopic
lifestages. The Nereocystis lifecycle begins when an adult plant (diploid sporophyte)
releases billions of zoospores (haploid spores) from special patches on the blades known
as sori (singular: sorus) into the water. Nereocystis spore release follows a distinct diel
pattern occurring during the few hours before and after sunrise and begins with sorus
abscission resulting from cell necrosis around the sorus perimeter (Amsler & Neushul
1990). Kelp generally require hard substrates and zoospores will attach readily to both
consolidated bedrock and unconsolidated gravel, in some cases kelp spores even attach to
other macroalgae (Dayton 1985). Once attached, Nereocystis spores germinate into male
and female gametophytes (in as quickly as one week, although the longevity in nature of
the gametophyte plants is not known) that reproduce sexually to produce microscopic,
germling sporophytes (after approximately three weeks). Following this initial recruitment
juvenile sporophytes grow rapidly throughout the entirety of the growing season (Maxell
& Miller 1996).
6
Generally, kelp spores fall close to the parent plant, although spores can be carried
further (up to several kilometers) depending on local current conditions and the depth at
which spores are released (Gaylord et al. 2002). Considering that Nereocystis sorus
production occurs at the water surface, it is likely that spore dispersal distances for
Nereocystis may be significantly larger than for other species. Additionally, adult plants
broken free from the benthos can form floating rafts capable of successfully producing
viable spores (Rothäusler et al. 2009).
For all Nereocystis and other annual species as well as perennial species that recruit
according to predictable seasonal patterns, there is evidence that microscopic forms remain
dormant or overwinter until conditions are favorable for reproduction and growth (Carney
& Edwards 2006). Generally, evidence points to gametophytes being the most common
lifestages capable of overwintering although there is some evidence that spores and
germling sporophytes may as well. This is especially the case for Nereocystis and other
annuals as adult plants are often totally absent for a portion of the year. This study focuses
on the response of adult Nereocystis sporophytes to common stressors and from this point
forward Nereocystis will refer to the adult sporophyte stage unless otherwise specified.
Nereocystis produces sori on apical blades that float near the surface, whereas sori
of Macrocystis sporophytes are found on specialized blades called sporophylls near the
benthos (Druehl & Clarkston 2016). This means that all Nereocystis sorus production
occurs in the first few meters of the water column. As a result, surface conditions likely
exert more influence over the reproductive potential of Nereocystis than other kelp species.
Nereocystis is often described as a ruderal and early successional species and it
quickly recruiting to new substrate following disturbances ( Suskiewicz 2010; Dayton et
7
al. 1992). This has allowed for insight into the response of Nereocystis to a variety of
disturbances as full successional cycles can be manipulated and observed over the course
of a few years—a stark contrast to terrestrial forest research which requires observations
over decades (Duggins 1980). However, the alternation of heteromorphic generations
characteristic of the Nereocystis life-cycle complicates attempts to understand the full
impacts of environmental and biological stress as environmental requirements may differ
between microscopic and macroscopic lifestages (Steneck et al. 2002).
Nereocystis forests, like all kelp forests, engineer resilient foundations for complex,
diverse and productive nearshore ecosystems (Bertness et al. 2014; Hurd et al. 2014). Due
to its annual nature, many Nereocystis canopies are often completely absent from late fall
to early spring with some, in the more oceanic waters of the Strait of Juan de Fuca persisting
for one to two years (Maxell & Miller 1996; personal observations). Variation in canopy
extent and density is influenced by a combination of environmental and biological factors.
Broadly, interactions between environmental conditions, physical forces and biotic
interactions work to constrain or promote kelp forest structure (Pfister et al. 2017; Steneck
et al. 2002). These interactions are often complex and further complicated by the
synergistic effects of multiple stressors (Crain et al. 2008).
Kelp forests have also played a large role in providing evidence for biotic controls
on ecosystems, specifically top-down trophic effects from keystone predators. Most well
documented are the interactions between sea otters (Enhydra lutis) and urchins
(Strongylocentrotus spp.) (see Steneck et al. 2002). However, this classic dynamic likely
does not pertain to Puget Sound kelp forest as urchin populations aren’t as robust as in the
Strait of Juan de Fuca and along the outer coast.
8
Abiotic controls: Environmental conditions and physical forces
Photosynthetic performance
Nereocystis requires light in order to carry out photosynthesis, grow and reproduce
(Hurd et al. 2014; Dayton 1985), and lack of sufficient light is one stressor that can
negatively affect Nereocystis. Kelp occupy the photic zone, the area of the nearshore
environment where light penetrates to the benthos and photosynthesis occurs at variable
rates over a range of irradiances (Hurd et al. 2014).
At compensation irradiances, photosynthetic rates match cellular respiration
(Wiencke et al. 2006). The compensation level of irradiance for kelp is generally 2 to 11
µmol m-2 s-1 PAR (0.01 to 0.05 percent of surface light on a sunny day) (Hurd et al. 2014).
Even though growth may still occur below compensation irradiances, the ability of any
kelp lifestages to deal with additional stressors is compromised. In addition, sporophyte
sorus production, spore germination, gametophyte reproduction and germling saprophyte
growth may be delayed or impeded at such low irradiances (Carney & Edwards 2006;
Vadas 1972). For example, germling Macrocystis sporophytes delay growth between 2 to
3 µmol m-2 s-1 (Carney & Edwards 2006). Similarly, Vadas (1972) found that only 1 percent
of Nereocystis gametophytes attained fertility when exposed for three weeks to irradiances
of approximately 2 µmol m-2 s-1 .
Photosynthetic rates increase with increasing irradiance until maximum
photosynthetic rates are attained at saturation irradiances. At this point, any increase in
irradiance returns little to no increases in photosynthetic rates. Saturation irradiances for
kelp fall between 150 to 250 µmol m-2 s-1 (Hurd et al. 2014). In laboratory cultures,
9
Nereocystis gametophytes and germling sporophyte growth rates peaked between
approximately 15 to 30 µmol m-2 s-1, similar to the critical levels needed to induce germling
sporophyte growth in Macrocystis (20 to 30 µmol m-2 s-1) (Carney & Edwards 2006; Vadas
1972).
While light is essential for photosynthesis, photo-inhibition can occur when high
irradiances and UV exposure lead to cellular damage and tissue death. Photo-inhibition for
kelp in general occurs between 850 and 1000 µmol m-2 s-1 with microscopic stages being
especially susceptible to UV damage (Swanson & Druehl 2000). However, phototolerance, like many traits, is species specific and no data exists for any Nereocystis
lifestage.
Temperature
On a global scale, kelp species are distributed along latitudinal temperature
gradients (Hurd et al. 2014; Bartsch et al. 2008; Lüning & Freshwater 1988). Individual
species have unique optimal temperature ranges that can also differ between alternative
lifestages ( Hurd et al. 2014; Harley et al. 2012; Dayton 1985). Temperature optimums can
be difficult to fully describe as individual species can adapt and acclimate to temperature
stress to different degrees (Lind & Konar 2017; Dayton 1985). However, trends show
similar optimal ranges for species with similar distributions.
Other cold-temperate kelp species related to Nereocystis in the genus Laminaria
can survive and reproduce from 0 to 18 °C but grow optimally in the range of 5 to 15 °C
(Bartsch et al. 2008). Similarly, while Nereocystis sporophytes can survive at a range of 1.5 °C to 18 °C (Lüning & Freshwater 1988), Maxell & Miller (1996) found that Puget
10
Sound Nereocystis stipe and blade growth rates peaked in concert with summer time high
temperatures of 13.5 °C.
Power plants in California with outflow pipes terminating in coastal waters are
required to comply with rigorous environmental monitoring practices. As a part of
regulatory compliance, laboratory experiments were undertaken to understand Nereocystis
response to temperature (Springer et al. 2010). Temperatures exceeding 18 °C resulted in
total mortality of adult sporophytes (Tera Corp. 1982). However, even exposure to slightly
elevated temperatures of 15.9 °C resulted in 25 percent mortality of adult sporophytes over
approximately a month (Tera Corp. 1982). The authors attributed this mortality to the
inability of Nereocystis to repair physical damage from handling during flume experiments.
These laboratory results were corroborated by Schiel et al. (2004) after an analysis of
community changes over ten years in response to water temperature increase of 3.5 °C in
a kelp forest adjacent to a powerplant outflow.
Increases in water temperature coincided with a 90 percent loss of the midcanopy
species Pterygophora californica and a 97 percent loss of Nereocystis in California (Schiel
et al. 2004). However, loss of adult sporophytes is not antithetical to kelp persistence. As
mentioned, kelp is resilient to disturbance and microscopic lifestages show the ability to
remain dormant until conditions favor growth (Edwards 2000).
Similarly, Wernberg et al. (2010) explored the relationship between disturbance,
temperature and canopy recovery along a naturally occurring latitudinal temperature
gradient along the west Australian coast. Temperatures ranged from between 2 to 4 °C
between sites allowing the researchers to explore the effects of predicted sea surface
temperature rise on the Australian canopy species Ecklonia radiata. While temperature did
11
not affect canopy cover or total biomass of beds, it did significantly impact recruitment
following disturbance. This suggests that temperature tolerances of microscopic lifestages
are significantly lower than those of adult plants and that even small increases in oceanic
temperatures over the long-term could reduce overall kelp resiliency.
Outside of optimal temperature ranges, photosynthetic performance can be
augmented by increasing pigment content, reaction centers, and protein complexes
allowing for higher maximum photosynthetic rates and yield (Bartsch et al. 2008).
However, maintaining positive growth in the face of temperature stress may make adult
sporophytes more susceptible to additional disturbances (Wernberg et al. 2006). This is in
part due to the fact that photosynthetic rates can only be increased so much before
respiration and biomass loss outpaces photosynthetic yields (Hurd et al. 2014).
For example, Krumhansl et al. (2014) used field observations of kelp detrital
production across seasonal temperature variations to model detrital production response to
predicted climate change. They found that as temperatures warm, kelp will provide
increasing levels of detritus until the point at which tissue degradation outpaces growth.
Similarly, Rothäusler et al. (2009) monitored experimental rafts of Macrocystis along a
latitudinal temperature gradient on the Chilean coast. Rafts kept at temperatures of 12 to
19 °C grew and reproduced successfully while those kept at temperatures > 20 °C began
to degrade and did not produce reproductive sori.
Nutrients
Seasonal and geographic variations in nutrient availability have the potential to
influence the health and productivity of Nereocystis (Schiel & Foster 2006; Dayton 1985).
12
Like other marine aquatic vegetation and kelp species, Nereocystis is nitrogen limited
(Hurd et al. 2014; Dayton 1985). In laboratory investigations of nutrient uptake rates, Ahn
et al. (1998) found that Nereocystis showed preference for nitrate over ammonia as nitrate
uptake rates increased linearly with nitrate additions while ammonium uptake rates peaked
at around 10 µM of ammonium. This preference for nitrate holds true for other closely
related kelp species such as Macrocystis and kelp species in the genus Lamiaria Sensu
Lato (Schiel & Foster 2015; Bartsch et al. 2008).
The majority of nutrient transport into the Puget Sound occurs via deep water
influent of oceanic waters through the Strait of Juan de Fuca (Khangaonkar et al. 2018).
Transport of nutrients from deeper, oceanic waters to surface waters is a slow process but
the Puget Sound is, overall, a well-mixed system, thanks to regular vertical mixing
occurring at shallow constriction points between basins (Ebbesmeyer et al. 1988).
Generally, it is accepted that nitrate availability is inversely related to temperature (Dayton
1985). Along the outer coast and Strait of Juan de Fuca, this is because nutrient upwelling
regimes are often interrupted during periods of high temperatures. Whether this
relationship holds true for all reaches and embayments of the Puget Sound is unknown an
may only be an issue for shallower and less mixed reaches.
Biotic controls in the SPS
Anecdotal evidence from local residents, tribal members and environmental
professionals document increases in the numbers of the kelp crab Pugettia producta in the
central and SPS. These accounts often go hand in hand with observations of canopy loss in
the same regions. Recent research by Dobkowski (2017) confirms P. producta grazing
13
preference for Nereocystis over other locally abundant kelp and specifically a preference
for juvenile sporophytes over adults, but does not document P. producta impacts to
Nereocystis recruitment or growth. In California, multiple year classes of kelp crabs (as
determined by comparisons of carapace size) are found coexisting in floating canopies
(Hines 1982). Similar observations made at SPS Nereocystis forests documented little
variation in carapace size with nearly all crabs starting the season as juveniles and maturing
over the course of the summer (Berry et al. 2019). This may suggest that kelp crabs recruit
annually with the appearance of Nereocystis. Little is known regarding adult kelp crab
habitat preference in the Puget Sound or what effect adult populations may have on early
kelp recruitment.
There is evidence in the literature that environmental conditions can interact with
relative grazer abundances to influence the mortality and health of kelp (Rothäusler et al.
2009; Duggins et al. 2001). Duggins et al. (2001) investigated the role of current flow on
grazer abundance of the molluscan gastropod Lacuna vincta and Nereocystis mortality in
the San Juan Islands of Washington state and found that even minor damage to Nereocystis
stipes significantly reduced their breaking strength when exposed to high current velocities.
Pfister & Betcher (2017) observed similar patterns of mortality associated with wave action
and grazer damage to stipes of the upper subtidal species Pleurophycus gardneri along the
coast of Tatoosh Island off the coast of Washington. However, interactions between
grazing and other kelp stressors is not limited to current velocity alone.
As discussed previously, temperature stress can reduce the resiliency of kelp to
additional stressors by negatively impacting physiological ability to repair damaged tissue
( Krumhansl, K. A. et al. 2014; Harley et al. 2012). To investigate the interactions of grazer
14
stress and temperature Rothäusler et al. (2009) observed the impact of temperature and
grazing on biomass of floating Macrocystis rafts along the Chilean coast. At intermediate
temperatures of 16 to 19 °C, rafts without grazers maintained or increased biomass while
rafts exposed to grazers steadily lost biomass suggesting that temperature stress reduces
overall resiliency of adult sporophytes.
Washington State kelp
Our current understanding of the state of Puget Sound kelp resources is largely
restricted to floating Nereocystis canopies as they can be easily surveyed from boats,
aircraft and satellites. Little information exists regarding distributions of the remaining 20+
Puget Sound kelp species in Puget Sound (Mumford 2007). Traditional ecological
knowledge from tribes and local residents, citizen-science surveys, and analysis of
historical data points to significant declines in the extent of Nereocystis forests throughout
the Puget Sound (Palmer-McGee 2019; Berry 2018; Berry et al. 2005).
Local residents encountered at docks and boat launches throughout the state often
offer personal accounts of losses to Nereocystis forests (personal communication).
Whether individuals encountered are recreational boaters, avid beachcombers,
management agency officials or tribal members, most recall a greater abundance and area
of Nereocystis canopies in the 1970’s and 1980’s than today. When asked about possible
explanations, responses vary from increases in kelp crab abundance and losses to important
fisheries species to effects of shoreline development and climate change.
Comparisons of long-term aerial photography of the north Olympic Peninsula to
kelp surveys from 1911 and 1912 document decreases in kelp canopy extent and area
15
around Dungeness Spit, Protection Island, and Port Townsend (Pfister et al. 2017). Similar
comparisons between 2016 aerial photography and 2006 remote sensing data document a
36 percent decrease in Nereocystis canopy area in the San Juan Islands (Palmer-McGee
2019).
In the South Puget Sound (SPS), recent and comprehensive analysis of historical
kelp surveys, navigation charts, and incidental kelp observations from habitat and other
target species surveys show clear changes to Nereocystis canopy extent (Berry 2018).
Linear extent of kelp canopies and area have declined approximately 68 percent in the SPS
(Berry 2018). The Washington Department of Natural Resources (WA-DNR) is currently
working on analyzing historical data in order to parse out Nereocystis canopy trends for
the Central Puget Sound but the Puget Sound Restoration Fund (PSRF) has already
documented a total loss of Nereocystis canopies around Bainbridge Island and from the
Kitsap Peninsula around Jefferson Head (PSRF, personal communication).
Comprehensive inventories of all kelp species found in Puget Sound are
nonexistent. WA-DNR has taken first steps toward understanding the full extent and area
of Puget Sound Nereocystis canopies by identifying and mapping all beds in the SPS during
the summer of 2017. Similar inventories are slated for the Central Puget Sound during
summer of 2019 (Berry, personal communication). In addition, WA-DNR is undertaking
the first assessment of non-floating kelp extent using towed underwater video taken in the
summer of 2018 of the entire shallow, subtidal King County coast. WA-DNR has also
monitored declines in forest area, depth extent, and kelp health at the Tucksel Point
Nereocystis forest at Squaxin Island from 2013 to present (Berry 2017).
16
In addition to WA-DNR’s efforts, the Northwest Straits Commission (NWSC) and
seven county Marine Resource Committees (MRCs) have implemented protocols for
citizen-science kayak mapping of Nereocystis canopies (Bishop, n.d.). Surveys conducted
in 2014 of kelp area in the Strait of Juan de Fuca, Smith and Minor Islands, Cypress Island,
and Cherry Point all documented decreases in canopy area. Kelp canopy area recovered in
2015 and 2016 in well-mixed areas, but not at the more sheltered site at Cherry Point. In
addition, the Snohomish County MRC has documented losses of several beds and declines
in remaining beds between Edmonds and Mukilteo (personal communication).
Current assessments of the extent and abundance of kelp canopies are, so far, rarely
paired with monitoring of key abiotic and biotic conditions within associated kelp forests.
Local conditions play a significant role in the large interannual variability of kelp forest
extent and abundance, and Puget Sound encompasses a number of sub-basins and reaches
with distinct environmental conditions (Krumhansl et al. 2016; Ebbesmeyer et al. 1988).
Thus, causes for declines in one region may not be applicable to Puget Sound as a whole.
This study is an early attempt to document differences in forest condition in
response to key stressors known to exert significant influences on the persistence and health
of kelp forests elsewhere. Surveys adapted existing WA-DNR methods for monitoring
plant condition and density at sites in the Puget Sound and paired these observations with
assessments of environmental conditions and grazer abundance. Pairing observations of
forest condition with records of environmental and community interactions may be used to
identify priority stressors at local sites through the Puget Sound and develop mitigation and
conservation strategies to protect remaining Nereocystis canopies and understory
assemblages.
17
METHODS
Site Selection
Monitoring was conducted between 2017 and 2018 in the South Puget Sound (SPS),
Washington. The SPS is the shallowest and most inland portion of the Puget Sound and is
separated from the Central Puget Sound by the Tacoma Narrows (Fig. 1). As is the case in
much of the Puget Sound, all sites were characterized by high current velocities (often
exceeding 100 cm sec-1 during maximum tidal exchanges) tied to large tidal shifts (often
>3m).
A total of four sites were selected for monthly temperature and nitrate monitoring
(Fig. 2). The sites were distributed south to north from Budd Inlet, Olympia, Washington
to Salmon Beach, Tacoma, Washington in order to encompass a majority of SPS
Nereocystis forests.
From south to north the sites were: Tucksel Point located at the southern tip of
Squaxin Island (lat. 47°10'0.49"N, long. 122°53'34.54"W); Devils Head on the southern
Kitsap Peninsula (lat. 47° 9'58.42"N, long. 122°45'37.11"W); Day Island south of Titlow
Beach in the Tacoma Narrows (lat. 47°14'21.78"N, long. 122°33'52.00"W) and Salmon
Beach, north of the Tacoma Narrows bridge (lat. 47°17'34.58"N, long. 122°31'48.00"W).
Three sites (Squaxin Island, Devil’s Head and Day Island) were initially selected
for intensive monitoring Nereocystis individuals and bed density during the usual growing
season (May to September). Intensive monitoring protocols were conducted at Salmon
Beach in July and September 2018 after the Devil’s Head canopy failed to appear in June.
At all intensive monitoring sites, two along-shore transects were used to establish 15 sets
of paired plots within previously surveyed 2017 bed perimeters. Paired plots consisted of
18
a near-shore (shallow) and off-shore (deep) plot to attempt to encompass total depth strata
of each bed while covering as much of previous year’s bed area as possible (Fig. 3). Three
monitoring points were established along the offshore side of each bed at a depth of -7
MLLW and revisited monthly from May 2017 to September 2018. Casting locations were
independent from the 15 paired plots established for intensive monitoring. Casting stations
were located offshore of the north, central and southern sections of each Nereocystis forest
(Fig. 3).
19
Figure 2: South Puget Sound and study sites.
20
Figure 3: Site monitoring design.
Mean Water Column Temperature
Monitoring of water temperature occurred monthly in partnership with the
Washington Department of Natural Resources (WA-DNR) Nearshore Habitat Program. All
sites were surveyed on the same day, within two hours before and after solar noon.
Temperature and salinity were measured using a weighted SonTek Castaway®CTD instantaneous data sonde. The sonde relies on flow-through sensors to log
instantaneous temperature and salinity data while free-falling through the water column at
a rate of approximately 1 m s-1. After each cast, data was quality checked in the field using
the sonde’s real-time data display. In the event of irregularities in cast depth profiles, the
21
sonde was recast to ensure data quality. Integrated water column temperatures were
averaged across all three casting sites.
Nitrate availability
Monthly field filtered water samples were collected from the central monitoring
point at each site for analysis of nitrate concentrations. An acid washed 60 mL syringe with
an attached 0.45 µm cellulose acetate filter was filled with water directly from a Van Dorn
sampler. A small amount of water was filtered through the syringe to rinse the syringe and
syringe filter before rinsing an acid washed 60 mL high density polyethylene (HDPE)
bottle with filtrate. The bottle was then filled with filtrate before being placed immediately
in a cooler on ice and transported to the Evergreen State College laboratory where they
were frozen (-10˚ C) for later transport to the University of Washington’s Marine
Chemistry Lab for total dissolved nutrient analysis using continuous flow automated
colorimetric analysis (Knap et al. 1996).
In March only one sample was taken from each site from a depth of -4 m MLLW
due to the relatively low water temperatures observed during sampling. From May to
September two samples were taken from each site at depths of -0.25 m MLLW and -4 m
MLLW in order to detect if nitrate concentrations on the surface differed from water
column in order to assess possible effects of seasonal thermal stratification.
Bed-wide and Nereocystis blade length monitoring
Monitoring of Nereocystis and kelp crab (Pugettia producta) densities and
individual Nereocystis blade length began in May when adult plants were first visible at
22
the water surface during daytime low tides (≤ 0 m MLLW). Monitoring continued monthly
during spring tide windows with water levels ≤ 0 m MLLW. All bed and individual
monitoring occurred one hour before and after predicted low tide during the slack tide
window to minimize interference from current velocities during the survey process. The
15 paired plots established within previously mapped 2017 bed perimeters were used to
monitor Nereocystis and kelp crab (Pugettia producta) densities, and the selection of one
Nereocystis individual selected per point for morphometric measurements.
At each point, demarcated kayak paddles were used to delineate a 3.79 m2 circular
quadrat. The paddle was positioned at stomach level with the middle of the paddle serving
as the center of the circular quadrat. Plants directly under the observer could not be seen
but those under the stern and prow of the kayak could be easily counted. Within each
quadrat, all individual Nereocystis bulbs were enumerated, as well as any kelp crabs present
in the plot (regardless of associated substrate).
The Nereocystis individual closest to the off-shore side of the kayak was selected
for individual blade length analysis at each point. At points with density counts of zero, the
closest plant within 4 m of the point was selected. If no plant fell within this extended
radius, no individual was measured for blade length. Blade length was recorded only for
the longest blade present on an individual. If the longest blade was more than 0.5 m longer
than the second longest blade, the second longest blade was measured.
Analysis
All data were analyzed using the R statistical programing language (Version 3.5.1,
R Core Team 2018). Plant and crab densities per m2 were calculated by dividing raw plant
23
count data by the area of the sampling quadrat. Plant density and crab density were
positively skewed and contained a large number of zeros. Temperature and nitrate
concentration data were non-normally distributed with no obvious skew to the data. Oneway ANOVAs were used to test for differences in plant density, crab density, water
temperature and nitrate concentrations despite deviations from normality due to the
robustness of the test in the face of non-normally distributed data (Schmider, E. et al. 2010).
Plotted residuals from the one-way ANOVAs investigating differences in mean
plant density between sites and between sites through time were skewed right and an
investigation of the models reveled significant heteroscedasticity in the data as determined
using a Bruesch-Pagan test from the lmtest package (Zeileis, A. & T. Hothorn 2002). Due
to the inclusion of zeros in the data set, a Tukey’s power transformation from the
rcompanion package was selected over a logistic or Box-Cox transformation (Mangiafico,
S. 2019).
One-way ANOVAs using the transformed density data rectified issues of nonnormally distributed residuals and heteroscedasticity for mean plant density through time
but not for comparisons of mean plant density between sites. As a result, a Welch’s
ANOVA with Games-Howell post hoc pairwise comparisons (biostat package; Gegzna, V.
2018) were used to determine differences in plant density between sites while a one-way
ANOVA with Tukey’s power transformed plant density data was used to determine
differences between sites through time.
Similarly, one-way ANOVAs using both crab density and Tukey’s power
transformed crab density data suffered from non-normally distributed residuals and
heteroscedasticity while transformed temperature and nitrate concentration data suffered
24
from heteroscedasticity only. Thus, a Welch’s ANOVA with Games-Howell post hoc test
was used to determine differences in average crab density and nitrate concentration
between sites as well as differences in water temperature between sites through the season.
Due to the low sample sizes for nutrient concentrations and high month-to-month
variability in crab densities, no analyses were preformed to investigating changes in
nutrient concentration or crab densities from month to month.
Differences in mean blade length between sites and across time was investigated
using a one-way ANOVA with a post hoc Tukey’s test used to assess pairwise differences.
A linear regression with a quadratic polynomial treatment of temperature was used
to assess the effect of temperature, crab densities and nitrate availability on blade length
across all sites. However, considering the significant differences in explanatory and
response variables between sites, individual, site specific models were developed in
addition.
RESULTS
Survey Timing
Surveys of bed density and plant morphometrics began on 20 May 2018 at Squaxin
Island and 31 May 2018 at Day Island after the first mature sporophytes recruited to the
surface. A surface canopy was absent at Devil’s Head on 18 May 2018 and 26 June 2018
and, as a result, Devil’s Head was excluded from bed density and individual Nereocystis
blade condition analyses.
Despite the decision to discontinue canopy surveys at Devil’s Head, we felt that it
was important to conduct in-water observations in order to ascertain if the absence of a
25
surface canopy was due to an absence of Nereocystis sporophytes, or if a sub-surface
canopy existed. A subsurface canopy was encountered during a snorkel survey of the
southern half of the known 2017 bed area but only five plants were observed. With no
surface canopy at Devil’s Head, surveys began at Salmon Beach on 26 July 2018 in order
to ensure three sites for statistical comparisons of bed density and individual Nereocystis
blade condition.
No surface canopy at Devil’s Head was observed during temperature and water
nutrient sampling on 18 July 2018 or while passing by Devil’s Head on the way to Salmon
Beach on 26 July 2018. A SCUBA survey was undertaken on 13 August 2018 at Devil’s
Head to determine if the subsurface canopy had persisted through July, but a visual
encounter survey conducted along a transect bisecting the 2017 bed perimeter
longitudinally found no Nereocystis.
Salmon Beach is located in the Tacoma Narrows, the shallow and narrow
constriction separating the SPS from the Central Puget Sound (Ebbesmeyer, C. C. et al.
1988). This area is characterized by high tidal currents and is far from public boat launches
making kayak surveys contingent on motorized boat support. No Salmon Beach survey
was conducted in August due to a lack of boat support.
No blades were found on any of the Nereocystis individuals surveyed on 24 August
2018 at Squaxin Island. As a result, surveys were discontinued, and no survey was
conducted in September.
26
Site differences
Plant Density
Plant densities differed significantly between sites as determined by Welch’s
ANOVA (F2,127.25=31.9, p < 0.001). Densities at Squaxin Island were significantly lower
(M = 0.57 m-2, SE = 0.09) than at either Day Island (M = 2.26 m-2, SE = 0.24 , p < 0.001)
or Salmon Beach (M = 1.96 m-2, SE = 0.26, p < 0.001) (Table 1, Fig. 4). Mean aggregate
plant densities were similar at Day Island and Salmon Beach.
Table 1. Plant density per square meter as a function of site: Sample size, mean,
standard deviation, standard error, median, first quartile, third quartile and
interquartile range.
Site
Squaxin Island
Day Island
Salmon Beach
n
64
108
50
M
0.57
2.26
1.96
SD
0.73
2.55
1.81
SE
0.09
0.24
0.26
MED
0.26
1.06
1.32
Q1
0.00
0.53
0.79
Q3
0.79
3.43
2.57
IQR
0.79
2.90
1.78
Figure 4. Plant densities at each site (black line = median, white diamond = mean).
27
Decreases in plant density were also observed between months over the course of
the growing season (one-way ANOVA, F10,211= 9.43, p < 0.001), with plant densities
declining at all sites each month (Table 8, Fig. 5). A Tukey’s HSD post hoc test was run to
determine differences between sites during the same month and within sites from month to
month.
Density surveys in May at Day Island captured initial high density (M = 5.8 m-2,
SE = 1.16) recruitment to the canopy characteristic of Nereocystis (Dobkowski et al. 2019)
followed by a significant decrease in June (M = 2.35 m-2, SE = 0.5, p < 0.001). Densities
at Salmon Beach were slightly higher than those at Day Island for months where data is
available, and it is likely that the higher seasonal average plant density observed at Day
Island is the result of the initial high densities observed in May. Squaxin Island consistently
had significantly lower densities compared to Day Island and Salmon Beach for all months
with available data (p ≤ 0.01 for all pairwise comparisons). Finally, Squaxin Island bed
density in August (M = 0.3 m-2, SD = 0.1) was significantly lower than September bed
density at Salmon Beach (M = 1.54 m-2, SE = 0.27 , p = 0.003).
28
Figure 5. Monthly plant densities at each site (mean ± SE).
Blade Length
Overall, mean growing season blade lengths differed significantly between sites
(one-way ANOVA, F2, 171 = 32.07, p < 0.01) as well as throughout the growing season
(one-way ANOVA, F8, 165 = 18.06, p < 0.01). Average blade length at Salmon Beach (M =
3.34 m, SE = 0.16) was significantly longer than both Squaxin Island (M = 2.37 m, SE =
0.15, p < 0.001) and Day Island (M = 2.06 m, SE = 0.08, p < 0.001) as determined by a
Tukey’s HSD post hoc test (Table 2, Fig. 6). Average blade length at Squaxin Island was
the shortest of all sites but not significantly shorter than at Day Island.
29
Table 2. Blade length (m) as a function of site: Sample size, mean, standard deviation,
standard error and 95% confidence interval for
Site
Squaxin Island
Day Island
Salmon Beach
n
38
89
49
M
2.37
2.06
3.34
SD
0.91
0.76
1.10
SE
0.15
0.08
0.16
95% CI
2.08 – 2.66
1.90 – 2.22
3.03 – 3.64
Figure 6: Blade length at each site (black line = median, white diamond = mean).
Monthly average blade length differed significantly between sites during the 2018
growing season (one-way ANOVA, F9,166 = 16.18, p < 0.01). A Tukey’s HSD post hoc test
was used to determine differences between sites on the same month and within sites
between months. Blades lengths at both Squaxin and Day Island increased at the beginning
of the season, hitting peak length in June and July respectively before beginning significant
declines (Table 9, Fig. 7). Early season blade lengths recorded in May and June at both
sites did not differ significantly from one another. However, by July, Squaxin Island blade
30
length (M = 1.69 m, SE = 0.18) was significantly shorter than at Day Island (M = 2.68, SE
= 0.12, p = 0.007) and Salmon Beach (M = 3.51 , SE = 0.17 , p < 0.001 ).
Observations of longest blade lengths at Squaxin Island occurred in June (M = 2.99
m, SE = 0.15) with significant declines occurring in July (M = 1.69 m, SE = 0.13, p <
0.001) followed by a total loss of blades on all plants surveyed in August. At Day Island
this general trend was delayed by one month, with peak blade lengths observed in July (M
= 2.68 m, SE = 0.12) followed by significant declines in August (M = 1.83 m, SE = 0.16, p
= 0.04) and further declines September (M = 1.34 m, SE = 0.13, p = 0.63).
Figure 7: Average monthly blade lengths of Nereocystis individuals at each site (mean +
SE).
Average blade lengths at Salmon Beach were significantly longer than those at
Squaxin Island (p < 0.001) and Day Island (July: p = 0.02; August: p < 0.001) for months
31
where data is available (Table 9, Fig. 7). In addition, while average blade length was lower
in September than July, the decline was not significant.
Temperature
Monthly average integrated water column temperatures differed significantly
between sites during the 2018 growing season (Welch’s ANOVA, F19,474.42 = 34007 , p <
0.001). A Games-Howell post hoc test was used to determine differences between sites
during each month. Temperature trends at all sites monitored showed predictable seasonal
increases during the beginning of the summer, peaking near the summer solstice before
declining in late summer (Table 3, Fig. 8). Average monthly temperatures increased along
a north to south gradient with Salmon Beach having significantly lower temperatures than
all other sites in June (Table 3, p < 0.001 for all pairwise comparisons). Squaxin Island was
exposed to significantly higher temperatures than all other sites every month during the
growing season (Table 3, p < 0.001 for all pairwise comparisons). Waters at Devil’s Head
were significantly warmer than at Day Island for all months during the growing season (p
< 0.001) except July (Table 3).
32
Figure 8: Average monthly integrated water column temperatures
33
Table 3. Temperatures (°C) as a function of survey month and site with Games-Howell
pairwise post hoc differences*: Sample size, mean, standard deviation and maximum
integrated water column
May
June
July
August
September
MAX
1
2
n
M
SD
1. Squaxin Island
91
11.91
0.39 13.30
2. Devil's Head
91
10.76
0.24 11.51
-1.15
3. Day Island
89
10.42
0.36 11.18
-1.49 -0.34
4. Salmon Beach
83
11.02
0.34 11.55
-0.88 0.27
1. Squaxin Island
79
13.72
0.37 14.58
2. Devil's Head
73
11.97
0.42 13.47
-1.76
3. Day Island
88
11.31
0.16 11.76
-2.41 -0.66
4. Salmon Beach
81
10.74
0.06 10.88
-2.98 -1.23
1. Squaxin Island
69
16.20
0.05 16.35
2. Devil's Head
91
13.53
0.22 14.02
-2.67
3. Day Island
88
13.50
0.33 14.01
-2.69 -0.02**
4. Salmon Beach
104 12.24
0.05 12.44
-3.96 -1.29
1. Squaxin Island
81
15.87
0.11 15.97
2. Devil's Head
75
14.80
0.29 15.16
-1.06
3. Day Island
68
14.22
0.06 14.37
-1.65 -0.58
4. Salmon Beach
113 13.91
0.06 14.05
-1.96 -0.90
1. Squaxin Island
52
15.06
0.03 15.11
2. Devil's Head
25
14.12
0.05 14.28
-0.93
3. Day Island
60
13.67
0.03 13.74
-1.39 -0.45
4. Salmon Beach
25
13.49
0.03 13.58
-1.56 -0.63
3
0.61
-0.57
-1.27
-0.31
-0.18
* All pairwise comparisons p < 0.001
** Non-significant difference
34
P. producta densities
Field observations of crab density were marked by extremely high variation (Table
4, Fig. 9). Results from a Welch’s ANOVA showed significant differences in crab density
between sites (F2,113.67 = 27.7, p < 0.001). A Games-Howell post hoc tests showed
significantly lower densities at Salmon Beach (M = 0.03 crabs m-2, SE = 0.01 than Squaxin
Island (M = 0.36 crabs m-2, SE = 0.11, p < 0.001) and Day Island (M = 0.38 crabs m-2, SE
= 0.05, p = 0.01, Table 4). Average summer crab densities did not differ significantly
between Squaxin Island and Day Island.
Crab densities also showed distinct trajectories at each site through time, and the
seasonal patterns were different at each site (Table 10, Fig. 10). Highest recorded crab
densities peaked at Squaxin Island during July (M = 0.84 crabs m-2, SE = 0.32) and then
declined in August (M = 0.51 crabs m-2, SE = 0.23). In contrast, at Day Island crab densities
steadily increased until September (M = 0.75 crabs m-2, SE = 0.11, Table 10). Salmon beach
had the lowest observed crab densities. However, all densities per m2 were low rarely
exceeding one crab per m2 (Table 10, Fig. 10).
Table 4. Crab densities as a function of site with Games-Howell pairwise post hoc test pvalues: Sample size, mean, standard deviation, standard error
Site
1. Squaxin Island
2. Day Island
3. Salmon Beach
n
64
108
50
M
0.36
0.38
0.03
SD
0.86
0.50
0.10
SE
0.11
0.05
0.01
Games-Howell post hoc
p-value
1
2
0.98
0.01
5.73e10-10
35
Figure 9: Crab densities at each site (black line = median, pink diamond = mean)
Figure 10: Average monthly crab densities at each site (mean ± SE).
36
Nitrate Concentrations
Observations of nitrate concentration (µM) during the growing season documented
declines in nitrate availability at Squaxin Island over the course of the summer while nitrate
availability remained relatively consistent at the remaining three sites (Table 11, Fig. 12).
Nitrate availability differed significantly as a function of site (one-way ANOVA, F3, 22 =
43.13, p < 0.001) with average seasonal nitrate concentrations significantly lower at
Squaxin Island (M = 3.43 µM, SE = 0.89, p < 0.001 for all pairwise comparisons) than all
other sites as determined by a Tukey’s post hoc test (Table 5, Fig. 11). Nitrate
concentrations at Squaxin were consistently below 5 µM from June until September in both
surface waters (-0.25 m MLLW) and at a depth of -4 m MLLW; highest growing season
nitrate concentrations were observed in May at -4 m MLLW (7.9 µM) however, surface
nitrate concentrations were still below 5 µM (Table 11, Fig. 12). At all other sites, months
and depths nitrate concentrations were above 10 µM with the exception of Devil’s Head in
June where concentrations dipped to 8.8 µM at a depth of -4 m MLLW(Table 11, Fig. 12).
Table 5. Nitrate concentrations as a function of site: Sample size, mean standard deviation,
standard error and 95% confidence interval
Site
1. Squaxin Island
2. Devil's Head
3. Day Island
4. Salmon Beach
n
8
8
6
4
M
3.43
11.93
13.63
13.81
SD
2.50
1.57
2.04
1.31
SE
0.89
0.55
0.83
0.66
95% CI
1.70 – 5.17
10.84 – 13.02
12.00 – 15.26
12.52 – 15.10
37
Figure 11: Growing season (May–September) water nitrate concentrations at study sites
(black line = median, white diamond = mean).
Figure 12: Monthly growing season nitrate concentrations (µM) at two depths (0.25 m and
4 m) at study sites.
38
Interactions Between Temperature, Crab Density and Blade Length
During field observations, blade length decreased as densities of P. producta
abundances increased (Fig. 14). Observed reductions in blade length also coincided with
highest recorded temperatures at both Squaxin Island and Day Island (Fig. 13). The
combination of these two common stressors was negatively correlated with lengths,
together explaining approximately 20 percent of observed variations in blade length (F4,165
= 11.09, p < 0.01, Table 6). Temperatures above ~12 °C were associated with significant
decreases in blade length (Fig. 13). Crab density was also negatively correlated with blade
length; however this may be due to the small number of high crab densities observed and
large variation in blade lengths associated with low crab densities (Fig. 14).
Table 6. Nereocystis blade length, temperature and crab densities at all sites: Multiple
linear-regression with quadratic polynomial treatment of temperature, and interaction
between crab densities and average temperature
(Intercept)
Mean Temperature
Mean Temperature2
Crab density (m-2)
Crab density : Mean temperature
F (4, 165) = 11.09, Adj. R2 = 0.19, p = 5.46 e –8
B
-12.90
2.50
-0.10
-2.69
0.15
SE B
4.61
0.71
0.03
1.24
0.08
t
-2.80
3.56
-3.71
-2.16
1.77
p
0.006
0.001
0.0003
0.03
0.08
39
Figure 13: Blade length as a function of water temperature with regression trend line
Figure 14: Blade length as a function of crab densities with regression trend line
40
Distinct differences in blade length, crab densities and water temperature between
all sites included in the study suggested sites specific differences influencing Nereocystis
blade length at all sites. As a result, blade lengths at each site were modeled as a function
of water temperature and crab densities at each site individually. Of the three models, water
temperature and crab densities only explained a significant amount of the variation
measured in blade length at Squaxin Island (Table 17). As with the all-site inclusive model,
the Squaxin Island regression showed a significant correlation between high temperatures
and decreases in blade length, however no significant correlation was demonstrated
between crab densities and blade length (Table 7, Fig.15).
41
Table 7. Nereocystis blade length, temperature and crab densities at Squaxin Island:
Multiple linear-regression with quadratic polynomial treatment of temperature, and
interaction between crab densities and average temperature
B
SE B
t
p
14.7
(Intercept)
-43.39
6
-2.94
0.006
Mean Temperature
6.70
2.04
3.29
0.003
Mean Temperature2
-0.24
0.07
-3.45
0.002
Crab density (m-2)
-11.49
8.80
-1.31
0.20
Crab density : Mean temperature
0.69
0.54
1.276
0.21
F (4, 31) = 12.13, R2 = 0.56, p = 4.75 e –6
Figure 15: Squaxin Island blade length as a function of temperature with trend line of
predicted values from regression analysis.
42
DISCUSSION
Losses to kelp forest canopies and serious negative impacts to individual plant
health in the face of temperature, grazing and nutrient stress are well documented in the
literature; Nereocystis forest canopies in the SPS appear to be no different. High
temperatures and P. producta presence were correlated with decreases in blade length in
this study. Furthermore, the kelp canopy exposed to the highest stress, Squaxin Island, also
had the lowest canopy density of any of the forests monitored during this study, with the
exception of Devil’s Head—which never formed a floating canopy during the summer of
2018.
Plant and bed condition
Salmon Beach stands out as the healthiest forest in the SPS. While plant densities
at Salmon Beach were similar to those observed at Day Island, the Salmon Beach forest
had the highest densities of Nereocystis, and blade length was longest throughout the
growing season (Table 6). Additional monitoring conducted by the Washington
Department of Natural Resources (WA-DNR) also found that Salmon Beach Nereocystis
had significantly larger bulbs and less physical damage than plants at Squaxin Island (Berry
et al. 2019).
Previous observations of SPS Nereocystis canopies document declines in overall
blade growth rates during the late summer and are similar to measured decreases in total
blade length observed in this study at Day Island and Squaxin Island (Maxell & Miller
1996). Like other photosynthetic autotrophs, seaweeds rely on environmental cues for the
timing of biological processes and it is possible that observed declines to blade length
43
during the late season may be tied to such an environmental trigger. However, studies into
blade senescence in Macrocystis found that biological cues tied to tissue age were better at
explaining rates of blade loss than environmental cues (Rodriguez et al. 2013).
Considering that Nereocystis is an annual species that senesces in the autumn, it is
possible that reduced blade lengths in late summer are merely the product of normal annual
cycles, however no research to date has been done on the mechanisms associated with
Nereocystis tissue senescence and growth rates. Regardless, late summer blade length
differed by site along a clear north to south gradient suggesting that differences in
environmental conditions and biotic interactions may explain increased blade tissue loss in
more southerly regions.
Environmental and biotic stressors
Average water column temperatures at the Squaxin Island forest remained above
15 °C from June until October (Table 6). In contrast, temperatures at Day Island and
Salmon Beach never rose above 14.2 °C during the summer. Temperatures observed at all
sites in this study were below the generally agree upon stress threshold of 17 °C. However,
25 percent of Nereocystis sporophytes held in artificial flumes at 15.9 °C died after one
month due to an inability to recover from physical damage, suggesting reduced resiliency
at temperatures near 16 °C (Tera Corp. 1982). Furthermore, concurrent monitoring of
surface water temperatures within the Squaxin Island canopy itself during 2018 found a
temperature gradient spanning 17 °C to 20 °C – well in the range known to significantly
impact Nereocystis sporophyte resiliency (Berry et al. 2019).
44
High temperatures at Squaxin Island coincided with average nitrate concentrations
below 5 µM from June to August (Table 11). In contrast, nitrate concentrations at Day
Island and Salmon Beach never dropped lower than 12.5 µM. Limiting nitrogen thresholds
for Nereocystis are unknown but Macrocystis requires 1 to 3 µM of total inorganic nitrogen
to maintain healthy growth (Schiel & Foster 2015). Nitrate concentrations at Squaxin
dipped below 1 µM only in August (M = 0.54 µmol, 0.25 m = 0.21 µmol, 4 m = 0.86 µmol,
Table 11). However, Squaxin Island water temperatures remained above 15 °C from July
through September and declines in Nereocystis photosynthetic performance at
temperatures above 15 °C in conjunction with low nitrate concentrations have been
documented in the literature (Wheeler et al. 1984). Furthermore, declines in Chilean
Macrocystis forests have been observed at temperatures above 16 °C even when nitrogen
concentrations are greater than 3 µM (Schiel & Foster 2015).
While this may be interpreted to mean that the Squaxin Island Nereocystis forest
suffers from a lack of nitrate, it is more likely that seasonal algal blooms depleted summer
water nitrate concentrations as has been documented in other regions of the Puget Sound
(Khangaonkar et al. 2018). Surface water nutrients concentrations (top 12 m of water
column) are only a fraction of concentrations encountered in deeper, oceanic inflows and
nutrient transport to shallow water is a slow process (Khangaonkar et al. 2018). Instead,
summer surface water nutrient concentrations are dominated by waste water treatment
effluent (81 percent of summer time land based nitrogen inputs to surface waters;
Mohamedali et al. 2011). Such artificially high nutrient levels may increase the frequency
and severity of summer algal blooms, hogging nutrients that would otherwise be used by
macroalgal communities.
45
As with temperature and nutrient stress, grazing stress was significantly more
intense at Squaxin Island than other sites. Average P. producta densities at Squaxin Island
peaked in July and declined sharply in August in tandem with the total loss of vegetative
blades on all plants surveyed that month. Crab density at Day Island continued to increase
during the entire growing season while blade length declined. In contrast, crab densities at
Salmon Beach were significantly lower than at Squaxin Island and Day Island and blade
length did not decline significantly in the late summer.
High temperatures and crab densities were correlated with decreases in blade
length. However, low crab densities coincided with a large range of blade lengths and
instances of high crab densities were relatively rare in this survey. Additionally, the
negative correlation between high temperatures and blade length may also be explained by
normal seasonal trends in Nereocystis blade growth which decline in late summer when
temperatures are the highest.
Regardless, environmental conditions (temperature and nutrient concentrations)
at Salmon Beach and Day Island were similar and yet Day Island plants experienced
significant declines in blade length during the late summer while Salmon Beach plants did
not. Observed reductions in blade length declines at Day Island coincided with increasing
crab densities, suggesting that grazing pressure may cause significant damage to
Nereocystis blade length in the SPS.
The low nutrients, high temperatures and abundance of kelp crabs make the
continued persistence of the Squaxin Island forest surprising. Investigations into
Nereocystis genetics in the Salish Sea revealed that Nereocystis allelic diversity decreases
as one moves south from Admiralty Inlet towards the SPS with the lowest allelic richness
46
found at Squaxin Island (Gierke et al. 2018). This low genetic diversity may stem from
adaptation to stressful conditions, but it may also be due to inbreeding depression —
Squaxin Island is southernmost Nereocystis forest in the Salish Sea. However, if the
Squaxin Island forest is indeed specially adapted to high stress conditions, it may be
possible to develop temperature resilient restoration stock for future recovery and
enhancement efforts currently underway in the Puget Sound.
Devil’s Head, substrate availability and bed area
Environmental conditions at Devil’s Head were similar to Day Island in that they
occupied an intermediary position between Squaxin Island and Salmon Beach but, unlike
Day Island, was more similar to Squaxin Island in terms of temperature (Fig. 6, Fig. 9).
Despite this, the few scattered Nereocystis individuals encountered on snorkel surveys in
June failed to reach the surface and had all disappeared two months later.
Macroalgae requires ample hard substrate on which to anchor and remain stationary
in areas of high currents (Dayton 1985). Unlike the outer coast and Western Strait of Juan
de Fuca, characterized by large swaths of rocky reefs composed of consolidated bedrock
substrates in the Puget Sound are dominated by unconsolidated pebble, cobbles and
boulders (Ebbesmeyer et al. 1988). All plants encountered during June snorkel surveys at
Devil’s Head appeared to be lodged in large stands of understory Saccharina latissima
(sugar kelp) and substrate observed consisted almost exclusively of fine sediment. Lack of
available substrate combined with the low-density of 2018 recruits may point to an
interaction between total bed area and substrate availability in determining forest density
and long-term persistence but requires further study.
47
Methods Assessment
This study relied on kayak surveys for data collection. This survey method not only
allowed for monitoring of target parameters but afforded an opportunity to test the efficacy
of low cost and accessible methods for use in monitoring Nereocystis forests with the help
of citizen science initiatives. Ecological data, especially in the marine realm, can be
difficult and expensive to obtain. As a result, most research relies on short-term (one to
five year) studies to generalize about incredibly dynamic and complex systems (Krumhansl
et al. 2016). Kelp forests, in particular, are characterized by a high degree of variation over
time. Much of this short-term variation is due to local variations in environmental and
biotic conditions (Krumhansl et al. 2016). However, on larger scales, variation in kelp
forests follows climatolgical scale alterations in environmental conditions (Pfister et al.
2017). Thus, low cost, easily implemented monitoring protocols could allow for more
consistent assessments of Puget Sound Nereocystis forests over longer time scales.
Some of the data collected during this study violated assumptions for standard
statistical tests in ways that are commonly encountered during ecological studies.
Specifically, both plant density data and crab counts were over-dispersed due to a large
number of zeros recorded during surveys. In some cases, zero-inflated data sets reflect the
reality of the study population, as in the case of rare species. However, neither Nereocystis
nor P. producta were rare in SPS kelp forests. Instead, the data suggests that in future
monitoring more plots need to be incorporated or that another sampling protocol, perhaps
across shore transects, should be selected.
48
Continued monitoring
More than half of the linear extent of Nereocystis forests in the SPS have been lost
since the 1890’s (Berry, personal communication). However, reasons for declines are
unknown, making conservation and recovery efforts difficult to implement effectively.
Sites for conservation and restoration actions can be identified from available data but a
lack of information regarding kelp distributions and stressor impacts may impede the
reestablishment of persistent annual canopies. Finally, the SPS is characterized by the
highest temperatures and lowest nutrient availabilities of all Puget Sound basins (Berry et
al. 2019). Continued monitoring of known kelp stressors and plant condition at SPS sites
has the potential to help identify priority stressors for mitigation actions and also provide
insight into the response of Puget Sound Nereocystis forest response to climate change.
CONCLUSION
The results from this study represent a first step towards identifying key stressors
driving losses in the SPS but pinpointing causes for declines will require further monitoring
and study. Each major basin of the Puget Sound is characterized by different oceanographic
conditions meaning that causes for declines may differ by region. Regardless, the stressors
monitored in this study remain high on the list of possible causes for kelp declines in the
Puget Sound and elsewhere (Steneck et al. 2002).
Low densities of adult sporophytes and rapid, early season loss of Nereocystis blades at
Squaxin Island is consistent with much of the literature regarding kelp responses to a
changing climate (Harley et al. 2012). The temperatures observed at Squaxin Island were
well within the accepted upper thermal range (15 to 20 °C) for Nereocystis. However,
49
temperatures at Day Island were more similar to those at Salmon Beach and yet significant
late summer losses to blade length were only observed at Day Island. This suggests the
strong possibility that kelp crab activity on adult Nereocystis sporophytes may exert a
strong influence over late season blade biomass at SPS kelp forests, especially as
temperatures continue to increase. These negative impacts are likely amplified at sites, such
as Squaxin Island, where additional temperature and nutrient related stress interact
synergistically with grazing damage. While these results confirm the suspicions of the
author and other nearshore management and restoration professionals, it is difficult to
understand whether grazing pressure in the SPS has increased over time due to a lack of
data.
Nereocystis forests in the Puget Sound form critical habitats that provide
foundations for food-webs that support healthy populations rockfish, salmon and orca.
Given the recent documented Nereocystis losses in the SPS and other basins, identifying
causes for canopy losses is essential for the effective future management.
Future studies should focus on investigating and monitoring key stressors in
Nereocystis forests across multiple basins. In addition, future research would benefit by
focusing on recording blade lengths of multiple blades on a single individual, monitoring
blade growth rates, estimating sorus production rates and assessing individual fecundity.
More detailed surveys of kelp crab abundances on a plant-by-plant basis (as opposed to per
m2) may provide better understanding into the effect of crab grazing on adult Nereocystis
sporophyte condition. Finally, pairing nutrient concentration monitoring with chlorophyll
monitoring could provide insight into the effect of algal blooms on nutrient availability to
large kelp species.
50
REFERENCES
Ahn, Okhyun, Royann J. Petrell, and Paul J. Harrison. 1998. Ammonium and Nitrate
Uptake by Laminaria Saccharina and Nereocystis Luetkeana Originating from a
Salmon Sea Cage Farm. Journal of Applied Phycology Volume 10(4) pages 333–
340.
Amsler, C D, and M Neushul. 1990. Diel Periodicity of Spore Release from the Kelp
Nereocystis Luetkeana (Mertens) Postels et Ruprecht. J. Exp. Mar. Biol. Ecol.
Volume 134(2) pages 117–127.
Bartsch, Inka, Christian Wiencke, Kai Bischof, Cornelia M. Buchholz, Bela H. Buck, Anja
Eggert, Peter Feuerpfeil, et al. 2008. The Genus Laminaria Sensu Lato: Recent
Insights and Developments. European Journal of Phycology Volume 43(1) pages
1–86.
Berry, Helen. 2017. Assessment of Floating Kelp at Squaxin Island in 2013, 2014, and
2016. Draft. Washington State Department of Natural Resources.
Berry, Helen. 2018. Kelp Status and Trends in Puget Sound. Powerpoint Presentation
presented at the Kelp Recovery Plan Workshop: Kelp Ecosystem Services,
Stressors and Coverage Trends, Seattle, WA.
Berry, Helen D., Max Calloway, and Julia Ledbetter. 2019. Bull Kelp Monitoring in South
Puget Sound in 2017 and 2018. Nearshore Washington Department of Natural
Resources Habitat Program Aquatic Resources Division.
Berry, Helen D., Thomas F. Mumford Jr, and Peter Dowty. 2005. Using Historical Data to
Estimate Changes in Floating Kelp (Nereocystis Luetkeana and Macrocystis
Integrifolia) in Puget Sound, Washington. No page number in Proceedings of the
2005 Puget Sound George Basin Research Conference. Puget Sound Action Team,
Olympia, Washington. Vol. 9.
Bertness, Mark D., John F. Bruno, Brian R. Silliman, and John J. Stachowicz, eds. 2014.
Marine Community Ecology and Conservation. Sinauer Associates, Inc.
Bertocci, I., R. Araújo, P. Oliveira, and I. Sousa-Pinto. 2015. REVIEW: Potential Effects
of Kelp Species on Local Fisheries. Edited by Henrik Österblom. Journal of
Applied Ecology Volume 52(5) pages 1216–26.
Bishop, Emily. n.d. A Kayak-Based Survey Protocol for Bull Kelp in Puget Sound, 26.
Carney, Laura T., and Matthew S. Edwards. 2006. Cryptic Processes in the Sea: A Review
of Delayed Development in The Microscopic Life Stages of Marine Macroalgae.
Algae Volume 21(2) pages 161–168.
51
Crain, Caitlin Mullan, Kristy Kroeker, and Benjamin S. Halpern. 2008. Interactive and
Cumulative Effects of Multiple Human Stressors in Marine Systems. Ecology
Letters Volume 11(12) pages 1304–15.
Dayton, Paul K. 1985. Ecology of Kelp Communities. Annual Review of Ecology and
Systematics Volume 16(1) pages 215–245.
Dayton, Paul K., Mia J. Tegner, Paul E. Parnell, and Peter B. Edwards. 1992. Temporal
and Spatial Patterns of Disturbance and Recovery in a Kelp Forest Community.
Ecological Monographs Volume 62(3) pages 421–45.
Dobkowski, Katie. 2017. The Role of Kelp Crabs as Consumers in Bull Kelp Forests—
Evidence from Laboratory Feeding Trials and Field Enclosures. PeerJ Volume
5(May) pages e3372.
Dobkowski, Katie A., K. Darby Flanagan, and Jessica R. Nordstrom. 2019. Factors
Influencing Recruitment and Appearance of Bull Kelp, Nereocystis Luetkeana
(Phylum Ochrophyta). Edited by A. Buschmann. Journal of Phycology Volume
55(1) pages 236–44.
Druehl, Louis D., and Bridgette E. Clarkston. 2016. Pacifc Seaweeds: A Guide to Common
Seaweeds of the West Coast. Second. BC Canada: Harbour Publishing.
Duggins, David, James E. Eckman, Christopher E. Siddon, and Terrie Klinger. 2001.
Interactive Roles of Mesograzers and Current Flow in Survival of Kelps. Marine
Ecology Progress Series Volume 223pages 143–155.
Duggins, David O. 1980. Kelp Beds and Sea Otters: An Experimental Approach. Ecology
Volume 61(3) pages 447–53.
Ebbesmeyer, Curtis C., Jack Q. Word, and Clifford A. Barnes. 1988. Puget Sound: A Fjord
System Homogenized with Water Recyclced. No page number in Hydrodynamics
of Estuaries, edited by Bjorn Kjerfve. CRC Press.
Edwards, Matthew S. 2000. The Role of Alternate Life-History Stages of a Marine
Macroalga: A Seed Bank Analogue? Ecology Volume 81(9) pages 2404.
Gabrielson, P.W., S.C Linstrom, and C.J. O’Kelly. 2012. Keys to the Seaweeds and
Seagrasses of Southeast Alaska, British Columbia, Washington and Oregon.
Phycological Contribution No. 8, PhycoID, Hillsborough, NC. Iv + 192 pp.
Gaylord, Brian, Daniel C. Reed, Peter T. Raimondi, Libe Washburn, and Stephen R.
McLean. 2002. A Physically Based Model of Macroalgal Spore Dispersal in the
Wave and Current-Dominated Nearshore. Ecology Volume 83(5) pages 1239–51.
52
Gegzna,
Vilmantas.
2018.
biostat: Routines
for Basic
(Bio)Statistics
(https://gegznav.github.io/biostat/ version R package version 0.0.18).
Gelfenbaum, Guy, Tom Mumford, Jim Brennan, Harvey Case, Megan Dethier, Kurt Fresh,
Fred Goetz, et al. 2006. Coastal Habitats in Puget Sound: A Research Plan in
Support of the Puget Sound Nearshore Partnership. Technical Report 2006-1. Puget
Sound Nearshore Partnership.
Gierke, Lily, Bobby San Miguel, Tom Mumford, and Filipe Alberto. 2018. Genetic
Structure of the Bull-Kelp Nereocystis Luetkeana in the Salish Sea. presented at the
Salish Sea Ecosystem Conference, Seattle, WA, April 6.
Harley, Christopher D. G., Kathryn M. Anderson, Kyle W. Demes, Jennifer P. Jorve,
Rebecca L. Kordas, Theraesa A. Coyle, and Michael H. Graham. 2012. Effects of
climate change on global seaweed communities. Journal of Phycology Volume
48(5) pages 1064–78.
Hines, Anson H. 1982. Coexistence in a Kelp Forest: Size, Population Dynamics, and
Resource Partitioning in a Guild of Spider Crabs (Brachyura, Majidae). Ecological
Monographs Volume 52(2) pages 179–98.
Hurd, C.L., P.J. Harrison, K. Bischof, and C.S. Lobban. 2014. Seaweed ecology and
physiology. Second. Cambridge University Press.
Khangaonkar, Tarang, Adi Nugraha, Wenwei Xu, Wen Long, Laura Bianucci, Anise
Ahmed, Teizeen Mohamedali, and Greg Pelletier. 2018. Analysis of Hypoxia and
Sensitivity to Nutrient Pollution in Salish Sea. Journal of Geophysical Research:
Oceans Volume 123(7) pages 4735–61.
Krumhansl, Kira A., Jean-Sébastien Lauzon-Guay, and Robert E. Scheibling. 2014.
Modeling Effects of Climate Change and Phase Shifts on Detrital Production of a
Kelp Bed. Ecology Volume 95(3) pages 763–74.
Krumhansl, Kira A., Daniel K. Okamoto, Andrew Rassweiler, Mark Novak, John J. Bolton,
Kyle C. Cavanaugh, Sean D. Connell, et al. 2016. Global Patterns of Kelp Forest
Change over the Past Half-Century. Proceedings of the National Academy of
Sciences of the United States of America Volume 113(48) pages 13785–90.
Lind, Alyssa C., and Brenda Konar. 2017. Effects of Abiotic Stressors on Kelp Early LifeHistory Stages. ALGAE Volume 32(3) pages 223–33.
Lüning, Klaus, and Wilson Freshwater. 1988. Temperature Tolerance of Northeast Pacific
Marine Algae. Journal Of Phycology Volume 24(3) pages 310–515.
Mangiafico, Salvatore. 2019. rcompanion: Functions to Support Extension Education
Program Evaluation (version R package version 2.0.10).
53
Maxell, B. A., and K. A. Miller. 1996. Demographic Studies of the Annual Kelps
Nereocystis Luetkeana and Costaria Costata (Laminariales, Phaeophyta) in Puget
Sound, Washington. Botanica Marina Volume 39(1–6) .
Mohamedali, T., M. Roberts, B. Sackmann, and A. Kolosseus. 2011. Puget Sound
Dissolved Oxygen Model Nutrient Load Summary for 1999-2008. 11-03–057.
Washington State Department of Ecology.
Mumford, Thomas F. 2007. Kelp and Eelgrass in Puget Sound: Fort Belvoir, VA: Defense
Technical Information Center.
National Marine Fisheries Service. 2017. Rockfish Recovery Plan: Puget Sound / Georgia
Basin Yelloweye Rockfish (Sebastes Ruberrimus) and Bocaccio (Sebastes
Paucispinis). Seattle, WA: National Marine Fisheries Service.
Palmer-McGee, Casey. 2019. A Decade of Dissaperance: Kelp Canopies in Samish
Traditional Territory. Web site accessed March 19, 2019.
Pfister, Catherine A., Helen D. Berry, and Thomas Mumford. 2017. The Dynamics of Kelp
Forests in the Northeast Pacific Ocean and the Relationship with Environmental
Drivers. Edited by A. Randall Hughes. Journal of Ecology, December.
Pfister, Catherine A., and Samuel P. Betcher. 2017. Climate Drivers and Animal Host Use
Determine Kelp Performance over Decadal Scales in the Kelp Pleurophycus
Gardneri (Laminariales, Phaeophyceae). Journal of Phycology.
R Core Team. 2018. R: A Language and Environment for Statistical Computing. Vienna,
Austria: R Foundation for Statistical Computing.
Rodriguez, Gabriel E., Andrew Rassweiler, Daniel C. Reed, and Sally J. Holbrook. 2013.
The Importance of Progressive Senescence in the Biomass Dynamics of Giant Kelp
(Macrocystis Pyrifera). Ecologicial Society of America.
Rothäusler, Eva, Iván Gómez, Iván A. Hinojosa, Ulf Karsten, Fadia Tala, and Martin Thiel.
2009. Effect of Temperature and Grazing on Growth of Macrocystis Spp.
(Phaeophyceae) along a Latitudinal Gradient. Journal of Phycology Volume 45(3)
pages 547–59.
Schiel, David R., and Michael S. Foster. 2006. The Population Biology of Large Brown
Seaweeds: Ecological Consequences of Multiphase Life Histories in Dynamic
Coastal Environments. Annu. Rev. Ecol. Evol. Syst. Volume 37pages 343–372.
———. 2015. The Biology and Ecology of Giant Kelp Forests. Oakland, California:
University of California Press.
54
Schiel, David R., John R. Steinbeck, and Michael S. Foster. 2004. Ten Years of Induced
Ocean Warming Causes Comprehensive Changes in Marine Benthic Communities.
Ecology Volume 85(7) pages 1833–39.
Schmider, Emanuel, Matthias Ziegler, Erik Danay, Luzi Beyer, and Markus Bühner. 2010.
Is It Really Robust?: Reinvestigating the Robustness of ANOVA Against
Violations of the Normal Distribution Assumption. Methodology Volume 6(4)
pages 147–51.
Springer, Yuri P., Cynthia G. Hays, Mark H. Carr, and Megan R. Mackey. 2010. Toward
Ecosystem-Based Management of Marine Macroalgae—The Bull Kelp,
Nereocystis Luetkeana. Oceanography and Marine Biology Volume 48pages 1.
Steneck, Robert S., Michael H. Graham, Bruce J. Bourque, Debbie Corbett, Jon M.
Erlandson, James A. Estes, and Mia J. Tegner. 2002. Kelp Forest Ecosystems:
Biodiversity, Stability, Resilience and Future. Environmental Conservation
Volume 29(04) .
Suskiewicz, Matthew S. 2010. Effects of Competition and Dispersal on the Recruitment of
the Annual Kelp Nereocystis Luetkeana. Masters Thesis, San José, California,
U.S.A.: San José State University.
Swanson, A. K., and L. D. Druehl. 2000. Differential Meiospore Size and Tolerance of
Ultraviolet Light Stress within and among Kelp Species along a Depth Gradient.
Marine Biology Volume 136(4) pages 657–64.
Tera Corp. 1982. Compendium of Thermal Effects Laboratory Studies: Diablo Canyon
Power Plant. Volume 1. Tera Corporation.
Vadas, R L. 1972. Ecological Implications of Culture Studies on Nereocystis Luetkeana.
Journal of Phycology, no. 8pages 196–203.
Wernberg, Thomas, Mads S. Thomsen, Fernando Tuya, Gary A. Kendrick, Peter A. Staehr,
and Benjamin D. Toohey. 2010. Decreasing Resilience of Kelp Beds along a
Latitudinal Temperature Gradient: Potential Implications for a Warmer Future:
Climate and Resilience of Kelp Beds. Ecology Letters Volume 13(6) pages 685–
94.
Wernberg, Thomas, Mathew A. Vanderklift, Jason How, and Paul S. Lavery. 2006. Export
of Detached Macroalgae from Reefs to Adjacent Seagrass Beds. Oecologia Volume
147(4) pages 692–701.
Wheeler, W. N., R. G. Smith, and L. M. Srivastava. 1984. Seasonal Photosynthetic
Performance of Nereocystis Luetkeana. Canadian Journal of Botany Volume 62(4)
pages 664–70.
55
Wiencke, Christian, Michael Y. Roleda, Ansgar Gruber, Margaret N. Clayton, and Kai
Bischof. 2006. Susceptibility of Zoospores to UV Radiation Determines Upper
Depth Distribution Limit of Arctic Kelps: Evidence through Field Experiments.
Journal of Ecology Volume 94(2) pages 455–63.
Williams, G.D., R.M. Thom, J.E. Starkes, J.S. Brennan, J. P. Houghton, D. Woodruff, P.L.
Striplin, M. Miller, M. Pedersen, A. Skillman, R. Kropp, A. Borde, C. Freeland, K.,
and McArthur, V. Fagerness, S. Blanton, and L. Blackmore. 2001. Reconnaissance
Assessment of the State of the Nearshore Ecosystem: Eastern Shore of Central
Puget Sound, Including Vashon and Maury Islands (WRIAs 8 and 9). Seattle, WA:
King County Department of Natural Resources.
Zeileis, Achim, and Torsten Hothorn. 2002. Diagnostic Checking in Regression
Relationships (https://CRAN.R-project.org/doc/Rnews/ version).
56
APPENDIX
Table 8. Plant density per square meter as a function of site and survey date: sample size,
mean, standard deviation, standard error, median, first quartile, third quartile and
interquartile range of
Month
Site
n
May
Squaxin Island
Jun.
Jul.
Aug.
Sep.
M
SD
SE
MED
Q1
Q3
IQR
13 0.87
0.74
0.21
0.79
0.26
1.06
0.79
Day Island
10 5.80
3.67
1.16
4.75
3.49
9.23
5.74
Salmon Beach
--
--
--
--
--
--
--
Squaxin Island
19 0.67
0.98
0.23
0.26
0.00
0.92
0.92
Day Island
25 2.35
2.50
0.50
1.06
0.53
3.17
2.64
Salmon Beach
--
--
--
--
--
--
--
Squaxin Island
17 0.47
0.58
0.14
0.26
0.00
0.79
0.79
Day Island
26 2.22
2.46
0.48
1.06
0.53
3.49
2.97
Salmon Beach
24 2.42
2.11
0.43
1.85
0.79
3.17
2.37
Squaxin Island
15 0.30
0.37
0.10
0.00
0.00
0.53
0.53
Day Island
24 1.68
1.72
0.35
1.06
0.46
2.31
1.85
Salmon Beach
--
--
--
--
--
--
--
--
Squaxin Island
--
--
--
--
--
--
--
--
Day Island
23 1.28
1.47
0.31
0.53
0.40
1.58
1.19
Salmon Beach
26 1.54
1.39
0.27
1.06
0.79
2.31
1.52
--
--
57
Table 9. Blade length (m) as a function of site and survey month: sample size, mean,
standard deviation, standard error, median and 95% confidence interval
Month
Site
n
M
SD
SE
95% CI
May
Squaxin Island
2
2.21
0.13
0.10
2.02 – 2.39
Day Island
14
2.11
0.46
0.12
1.87 – 2.35
Salmon Beach
--
--
--
--
--
Squaxin Island
19
2.99
0.64
0.15
2.70 – 3.27
Day Island
17
2.38
0.65
0.16
2.07 – 2.69
Salmon Beach
--
--
--
--
--
Squaxin Island
17
1.69
0.73
0.18
1.35 – 2.04
Day Island
20
2.68
0.53
0.12
2.44 – 2.91
Salmon Beach
23
3.51
0.81
0.17
3.18 – 3.84
Squaxin Island
--
--
--
--
--
Day Island
18
1.83
0.69
0.16
1.52 – 2.15
Salmon Beach
--
--
--
--
--
Squaxin Island
--
--
--
--
--
Day Island
20
1.34
0.59
0.13
1.08 – 1.60
Salmon Beach
26
3.18
1.30
0.25
2.68 – 3.68
Jun.
Jul.
Aug.
Sep.
58
Table 10. Crab densities as a function of site and survey month: sample size, mean,
standard deviation, standard error, median, first quartile, third quartile and interquartile
range
Month
Site
n
May
Squaxin Island
June
July
August
September
M
SD
SE
MED
Q1
Q3
IQR
13 0
0
0
0
0
0
0
Day Island
10 0
0
0
0
0
0
0
Salmon Beach
--
--
--
--
--
--
--
Squaxin Island
19 0.06
0.11
0.03
0
0
0
0
Day Island
25 0.21
0.40
0.08
0
0
0.26
0.26
Salmon Beach
--
--
--
--
--
--
--
Squaxin Island
17 0.84
1.31
0.32
0.26
0
1.06
1.06
Day Island
26 0.26
0.29
0.06
0.26
0
0.26
0.26
Salmon Beach
24 0.02
0.07
0.02
0
0
0
0
Squaxin Island
15 0.51
0.90
0.23
0
0
0.66
0.66
Day Island
24 0.49
0.63
0.13
0.26
0
0.79
0.79
Salmon Beach
--
--
--
--
--
--
--
--
Squaxin Island
--
--
--
--
--
--
--
--
Day Island
23 0.75
0.53
0.11
0.79
0.40
1.06
0.66
Salmon Beach
26 0.04
0.12
0.02
0
0
0
0
--
--
59
Table 11. Nitrate concentrations (µM) as a function of survey month and depth at each
site: sample size, mean, standard deviation, standard error and 95% confidence interval
Month
Site
n M
SD
May
Squaxin Island
2 5.82
2.96 2.10 1.71 – 9.92
Devil's Head
Day Island
June
Squaxin Island
Devil's Head
Day Island
July
Squaxin Island
Devil's Head
Salmon Beach
August
Squaxin Island
Devil's Head
September
Day Island
Salmon Beach
2 13.45
2 13.31
2 2.70
2 10.56
2 12.54
2 4.69
2 12.50
2 12.92
2 0.54
2 11.21
2 15.06
2 14.70
SE
95% CI
0.18 0.13 13.20 – 13.70
3.72 2.63 8.16 – 18.46
1.59 1.13 0.49 – 4.90
2.49 1.76 7.11 – 14.01
0.04 0.03 12.49 – 12.58
0.08 0.06 4.58 – 4.79
0.24 0.17 12.17 – 12.83
1.27 0.90 11.16 – 14.68
0.46 0.33
- 0.10 – 1.17
0.95 0.67 9.90 – 12.52
0.49 0.35 14.38 – 15.73
0.64 0.46 13.80 – 15.59
Depth (m)
Nitrate
0.25
3.72
4
7.91
0.25
13.32
4
13.58
0.25
15.94
4
10.68
0.25
1.57
4
3.82
0.25
12.32
4
8.80
0.25
12.51
4
12.56
0.25
4.63
4
4.74
0.25
12.33
4
12.67
0.25
13.82
4
12.02
0.25
0.21
4
0.86
0.25
10.54
4
11.88
0.25
14.71
4
15.40
0.25
15.15
4
14.24
60
