-
extracted text (extracttext:extracted_text)
-
THE RELATIONSHIP BETWEEN El NIÑO SOUTHERN
OSCILLATION AND LEVELS OF PARALYTIC SHELLFISH
POISONING PRESENT IN WASHINGTONS MARINE WATERS
By
Randy N. Lumper
A Thesis: Essay of Distinction
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Study
The Evergreen State College
June 2008
i
This Thesis for the Master of Environmental Study Degree
by
Randy Lumper
has been approved for
The Evergreen State College
by
________________________
Member of the Faculty
________________________
Date
ii
2008 by Randy Lumper. All rights reserved.
iii
THE RELATIONSHIP BETWEEN El NIÑO
SOUTHERN OSCILLATION AND LEVELS OF PARALYTIC
SHELLFISH POISONING PRESENT IN WASHINGTONS MARINE
WATERS.
Randy Lumper
Can large scale weather phenomenon such as El Nino Southern
Oscillation be used to predict when levels of paralytic shellfish poisoning
will be higher, thus mitigating their harmful impacts? This thesis examines
the relationship between levels of Paralytic Shellfish Poisoning (P.S.P.)
present in Washington State’s marine waters and large scale weather
phenomena or Oceanic circulation patterns such as the El Niño Southern
Oscillation (ENSO). It questions whether large scale climatic conditions
such as those produced by ENSO affect blooms of algae, such as
Alexandrium catenella. Once we have a better working knowledge of the
ecology of A. catenella we can better predict and monitor outbreaks of
P.S.P., thus reducing the amount of human illnesses and economic
hardships associated with P.S.P. The P.S.P. data is compared to sea
surface temperatures departures from the normal in Niño region 3.4, and
localized weather parameters to see if any significant relationship exists.
As indicated by the results of this study P.S.P. levels in Washington State
are not influenced by the phases of ENSO.
i
Table of Contents
List of Figures ....................................................................................................... v
List of Tables ....................................................................................................... vi
Chapter 1 Introduction .......................................................................................... 1
1.1
Background and Problem Description .................................................... 2
1.2
Thesis Statement and Layout ................................................................. 5
Chapter 2 Paralytic Shellfish Poisoning and Alexandrium Catenella Life Cycle .... 7
2.1
The Life Cycle of Alexandrium Catenella and the Physical Processes of
its Environment................................................................................................ 10
2.2
Turbulence............................................................................................ 14
2.3
Precipitation and Runoff ....................................................................... 17
2.4
Salinity .................................................................................................. 18
2.5
Water Temperature .............................................................................. 19
Chapter 3 Large Scale Oceanic Circulation Patterns Such as ENSO Affect
Regional Weather, and Drive Small Scale Physical Processes .......................... 23
3.1
The El Niño Southern Oscillation (ENSO) ............................................ 23
3.2
La Niña ................................................................................................. 26
3.3
ENSO Influence on Pacific Northwest Climate and Weather ................ 27
Chapter 4 Climate Changes Impacts to Pacific Northwest Weather and ENSO . 30
Chapter 5 Data Sources ..................................................................................... 34
5.1
ENSO Index.......................................................................................... 34
5.2
WDOH Database .................................................................................. 35
Chapter 6 Analysis and Results .......................................................................... 40
6.1 Air temperature.......................................................................................... 47
6.2 Wind .......................................................................................................... 49
6.3 Precipitation............................................................................................... 51
6.4 Multiple liner regression analysis ............................................................... 52
Chapter 7 Implications of This Study .................................................................. 57
7.1
Regulatory Agencies and/or Groups ..................................................... 59
7.2
Shellfish Industry .................................................................................. 60
7.3
Recreational Harvest and Tribal Ceremonial/Subsistence Harvesting . 63
Chapter 8 Conclusion ......................................................................................... 65
References: ........................................................................................................ 68
iv
List of Figures
Figure 1: Alexandrium catenella Photo: by Jan Rimes. ............................................. 1
Figure 2: The world wide distribution of P.S.P. during 1970 and 2000,Glibert
2005.................................................................................................................................... 3
Figure 3: Meroplanktonic life cycle of the dinoflagellate species Alexandrium
(Anderson, 2005). ............................................................................................................ 9
Figure 4: Mean and anomalies of sea surface temperature (SST) from 1986 to
present, showing El Niño events left 1986-1987, 1991-1992, 1993, 1994 and
1997 and La Niña events right in 1985 and 1995
(www.pmel.noaa.gov/tao/elNiño/la-Niña-story.html). ............................................... 25
Figure 5: a (left)-b (right): Box-and-whisker plots showing the influence of ENSO
on October-March (a) temperature and (b) precipitation (1899-2000). For each
plot, years are categorized as cool (La Niña), neutral (ENSO neutral), or warm
(El Niño). For each climate category, the distribution of the variable is indicated
as follows: range of values (whiskers); median value for the phase category
(solid horizontal line); regional mean for all categories combined (dashed
horizontal line); 75th and 25th percentiles (top and bottom of box). Areaaveraged Climate Division data are used for temperature and precipitation
(C.I.G., 2008). ................................................................................................................. 27
Figure 6: Chemical structure of Saxitoxin .................................................................. 36
Figure 7: ENSO over time based on a threshold of +/- 0.5ºC which is calculated
from a 3 month running mean of sea surface temperature anomalies or
“departures from normal” in the Niño 3.4 region which is based on the 1971-2000
base period. An El Niño or La Niña event occurs when the threshold of +/- 0.5ºC
is met or exceeded for a minimum of 5 consecutive months. ................................. 41
Figure 8: P.S.P. levels for the period 1/1/1977-12/31/2007, over time.................. 41
Figure 9: P.S.P. levels equal to or less then 5,000µg STX/100g over time .......... 42
Figure 10: 3 month moving average of P.S.P. µg STX/100g levels over time ..... 43
Figure 11: Area plot of 3 month running means departures from normal for both
ENSO and P.S.P. levels in Washington States marine waters. ............................. 44
Figure 12: Linear Regression of Log of 3 month moving average vs. ENSO Index
........................................................................................................................................... 46
Figure 13: Linear Regression of Log of 3 month moving average vs. mean daily
air temperature (C) ......................................................................................................... 49
Figure 14: Linear Regression of Log of 3 month moving average vs. max daily
wind speed ...................................................................................................................... 49
Figure 15: Linear Regression of Log of 3 month moving average vs. average
daily wind speed ............................................................................................................. 51
Figure 16: Linear Regression of Log of 3 month moving average vs. daily
precipitation amount (inches) ....................................................................................... 52
Figure 17: Estimated West Coast production of farmed oysters, clams, mussels
and geoduck, 2000
http://www.psat.wa.gov/Programs/shellfish/fact_sheets/economy_ web1.pdf. .... 58
v
List of Tables
Table 1: Multiple regression of the log of the 3 month moving average vs. the
ENSO index, daily air temperature, max daily wind speed, daily wind speed, and
daily precipitation............................................................................................................ 53
Table 2: Correlation output among the Independent variables. .............................. 54
Table 3: Multiple regression of the log of the 3 month moving average vs. the
ENSO index, daily air temperature, daily wind speed, and daily precipitation
amounts. .......................................................................................................................... 55
Table 4: Multiple regression of the log of the 3 month moving average vs. the
ENSO index, daily air temperature, and daily wind speed. ..................................... 56
vi
Chapter 1
Introduction
Figure 1: Alexandrium catenella Photo: by Jan Rimes.
Millions of microscopic plant cells, commonly known as algae, thrive in
nearly every drop of seawater. Algae (phytoplankton) are the primary energy
producers in the ocean, forming the base of the marine food chain. These
single-celled plants photosynthesize and multiply, creating a bloom that feeds
everything from fellow microbes to larger fish species. Primary production is the
synthesis and storage of organic molecules during the growth and reproduction
of photosynthetic organisms such as algae. Seasonal phytoplankton dynamics
are most often characterized by low biomass and primary productivity during
winter-spring seasons of high river flow, followed by a slow (2-3 months)
accumulation of biomass during summer (Alpine, 1992). Typically during late
spring and early summer the dominate phytoplankton is diatoms, a species that
thrives in un-stratified environments and is unable to migrate below the
1
thermocline when nutrients become depleted in the surface layer. Once nutrients
become depleted in the surface layers, more motile species such as
dinoflagellates become dominate, due to their ability to migrate below the
thermocline and carry on photosynthesis when other algae are dying off. Among
the thousands of species of algae in the sea, a few dozen produce toxins and
pose a formidable natural hazard. In Washington’s Puget Sound basin the
dinoflagellate Alexandrium catenella (Figure 1) is one such species.
This thesis looks at the relationship between the dinoflagellate A. catenella
and the climate phenomenon know as El-Nino Southern Oscillation (ENSO).
ENSO, a global-scale weather phenomenon influences local climate, which in
turn influences the physical properties of the environment where A. catenella, is
found. This thesis asks the question, is there a relationship between the phases
of ENSO, and levels of paralytic shellfish poisoning present in Washington States
marine waters? If a relationship exists between the phases of ENSO or a
parameter of A. catenellas environment and the paralytic shellfish poisoning
levels produced by A. catenella, then ENSO can be used to predict when blooms
of Paralytic shellfish poisoning are likely to occur. If this is the case then
regulators of the shellfish industry may be able to use this information to mitigate
the economic and health impacts from blooms A. catenella and the paralytic
shellfish poisoning that it produces.
1.1
Background and Problem Description
2
Early records, partially based on local native lore passed down by word of
mouth, suggest that harmful algal blooms have been present along the coast of
the United States for hundreds of years (Trainer, 2003), but actual historic
scientific data are non-existent.
Figure 2: The world wide distribution of P.S.P. during 1970 and 2000, Glibert 2005.
“Historically, P.S.P. has been known in the Pacific Northwest and
Alaska for centuries. Records of P.S.P. events date back as early as
June 15, 1793 (Vancouver 1798), when a member of Captain George
Vancouver’s exploration team died after eating contaminated mussels
3
harvested in the uncharted coastline of what is now known as British
Columbia (Trainer, 2003)”.
Washington State’s only three fatalities due to paralytic shellfish poisoning were
recorded in 1942 near the entrance to the Strait of Juan de Fuca, since then
there has been no illnesses resulting in death.
Harmful algal blooms have been known throughout the world and
recorded history, but the nature of the problem has changed considerably over
the past several decades, both in extent (Figure 2) and in public perception.
Over the past three decades, the occurrence of harmful or toxic algal incidents
has increased in many parts of the world, both in frequency and in geographic
distribution (Van Dolah, 2000). This increase in occurrence and awarness has
public officials and citizens both calling for a greater scientific understanding of
the biological and physical properties associated with harmful algal blooms.
There are many theories to what sets the stage for large scale blooms of
A. catenella and other harmful algal blooms around the world.
“One cause for the increasing frequency of harmful algal blooms
seems to be nutrient enrichment-that is, too much “food” for the algae.
Just as the application of fertilizer to lawns can enhance grass growth,
marine algae grow in response to nutrient inputs from domestic,
agricultural, and industrial runoff (Anderson, 2004)”.
Though nutrient enrichment increases the amount of available food for A.
catenella and other harmful algal blooms, it does not necessarily allow the
blooms to persist for longer durations or increase the frequency of blooms interannually. Recent studies suggests that not only do nutrients play a role, but that
physical process may be important in creating a favorable environment for large
scale blooms, and may be what is enabling blooms to persist in the environment
4
longer than historically recorded. Currently, not much is known about how
physical processes such as wind, precipitation, salinity, and water temperature,
limit or enhance the duration and frequency of harmful algal blooms. However, a
growing body of laboratory, field, and theoretical work suggests that the
dynamics of harmful algal blooms and their impacts on other organisms are
frequently controlled not only by physiological responses to local environmental
conditions as modified by trophic interactions, but also by a series of interactions
between biological and physical processes occurring over an extremely broad
range of temporal an spatial scales (Donaghay, 1997). Some of these physical
processes are not yet defined at the appropriate scale, yet may be crucial in the
formation of harmful blooms (Gentein, 2005). Physical factors affecting A.
catenella will be even harder to quantify in the future as we live in a world of
changing climate.
1.2
Thesis Statement and Layout
Do the phases of ENSO have an influence on paralytic shellfish poisoning
levels in Washington State? It is undetermined at this time if this is the case,
however, by performing statistical analysis of paralytic shellfish poisoning levels
in Washington States marine waters versus ENSO it may be possible to
determine if such a relationship exists.
The layout of this thesis was designed in the following order to provide a
logical path to answering the question above. The first two chapters break down
5
the life cycle of A. catenella, and how the physical properties of its environment
influence this life cycle. Chapter 3 looks at how the different phases of ENSO
affect local weather parameters and patterns, and thus the physical properties of
A. catenella environment. Chapter 4 discusses how climate change influences
ENSO and local weather. Chapter 5 describes the data sources used for this
study. Chapter 6 includes a statistical analysis of ENSO and associated weather
patterns versus paralytic shellfish poisoning levels in Washington State and the
results obtained from this analysis. The following chapter, Chapter 7, is a
discussion of what the implications of this study may have for individuals and
industries impacted by paralytic shellfish poisoning and those who regulate them.
Finally, Chapter 8 states the conclusions and discusses the implications that a
relationship between ENSO and paralytic shellfish poisoning levels in
Washington State may have and how this study may be strengthened in the
future.
6
Chapter 2
Paralytic Shellfish Poisoning and
Alexandrium Catenella Life Cycle
A. catenella is an armored, marine, planktonic dinoflagellate associated
with toxic Paralytic Shellfish Poisoning (P.S.P.). A. catenella occurs either as
single cells or as a chain of cells (Figure 1). A. catenella produces an array of
chemically similar neurotoxins that are responsible for P.S.P., which are
transmitted via tainted shellfish. Neurotoxins are so named because they disrupt
nerve impulses. P.S.P. toxin accumulates in marine animals that feed either
directly on toxic phytoplankton or on consumers of toxic phytoplankton. Bivalve
shellfish concentrate biotoxin’s when they filter toxic phytoplankton (A. catenella)
out of the water while feeding.
“These consumers include zooplankton, bivalve shellfish, predatory
marine snails, crabs, fish, birds, and marine mammals. Mass mortalities
among other shellfish-eating animals including birds, fur seals, foxes,
sea otters, and humpback whales have been traced to P.S.P.
(Determan, 2003)”.
These toxins can make their way up the food chain and end up affecting
humans, other mammals, fish and birds. On a global basis, almost 2,000 cases
of human poisonings are reported per year, with a 15% mortality rate (Van Dolah,
2000). “Human illness and death are the primary impact of Harmful algal
blooms, but effects on other wildlife are also important. Some fish kills due to
7
Harmful algal blooms can be spectacular is size, with millions of fish and millions
of dollars lost to local economies (Glibert, 2005)”. In the Puget Sound the
seasonal occurrence of A. catenella follows a scenario like this and the one in
Figure 3. Wintertime rain storms, rivers, and storm water runoff carry nutrients
into Puget Sound from uplands and watersheds. Strong winds mix the
freshwater with nutrient rich water from the open sea. Typically sunlight is the
most dominate factor limiting growth in the winter, however strong winds and
colder water temperatures may have an affect (Determan, 2003). In early spring,
winds become lighter reducing mixing, and the sun warms the water raising
water temperatures.
This causes the water column to stabilize and vertical mixing to slow down
or stop. A second influx of nutrients is then introduced in the form of snowmelt,
which increases the flow of most Puget Sound Rivers. The addition of these
abundant dissolved nutrients in a stable water column lead to blooms of many
phytoplankton species in the surface waters (Determan, 2003).
These blooms can be so dense that they can color the water red, brown,
green etc; this condition is called a “red tide”, and is often thought to be
associated with P.S.P., which is a common misconception. A. catenella may be
present in shellfish at dangerous levels, long before the water becomes
discolored (Determan, 2003). By mid to late summer, surface water is frequently
depleted of nutrients and many species of phytoplankton die back.
8
Figure 3: Meroplanktonic life cycle of the dinoflagellate species Alexandrium (Anderson,
2005).
However, two flagella enable A. catenella to vertically migrate to the
surface during the day and to depth at night. Because of this, A. catenella is able
to journey to deeper water via its flagella, where nutrients remain in abundance
(Trainer, 2003). Thus when nutrients become depleted at the surface during
summer A. catenella can simply move down to where nutrients are still
9
abundant, giving it an advantage over non-flagellated phytoplankton such as
diatoms. A. catenella then returns to the surface layer to carry on
photosynthesis. As a result of this A. catenella is able to bloom longer typically
until late autumn. At this time water temperatures begin to cool off and winds
speeds increase, inducing vertical mixing that leads to the breakdown of the
stratification of the water column, while sunlight starts to become too dim thus
limiting or shutting down the blooms production. At which time A. catenella may
form resting cysts that settle to the bottom awaiting the return of more favorable
growing conditions (Figure 3).
2.1
The Life Cycle of Alexandrium Catenella and the Physical
Processes of its Environment
The physical properties of aquatic environments play a fundamental role in
driving the dynamics of planktonic communities. The physical properties of an
aquatic environment not only shape the structure of the pelagic zone but also
affect the biological processes of phytoplankton, such as A. catenella (figure 1),
in many direct and indirect ways. In order to understand how the physical
properties of A. catenella environment affect its biological processes, it is
important to understand how physical processes such as turbulence,
precipitation/runoff, salinity and water temperature influence the aquatic
environment in which A. catenella thrives.
10
The environment most favorable to blooms of A. catenella is a nutrient rich
heavily stratified body of water. Generally, dinoflagellates thrive in stratified
waters because of their motility which enables it to move to nutrient-rich areas
within the water column (Trainer, 2003). Stratification of the water column can be
found in upwelling systems, coastal embayments, estuaries, and in retentive
zones such as gyres in the open oceans. Typically a stratified environment is
composed of layers of water, which represent natural boundaries to
phytoplankton. The top layer is known as the mixed layer and consists of a thin
band of water that extends from the ocean’s surface to the thermocline and can
be up to several tens of meters thick.
“The mixed layer is the boundary layer between the atmosphere and
the Deep Ocean, and results from solar radiation and surface cooling.
Winds help mix the heat through the layer by generating waves, inducing
Langmuir circulation, producing convection, and generating turbulence”
(Gentien, 2005).
The thermocline represents the boundary between the surfaced mixed
layer and more dense subsurface cold waters. It is produced by sudden changes
in water temperature with depth and is the area of greatest and most rapid
temperature change with depth in the oceans. The rapid changes in
temperatures that produce the thermocline, also creates density changes in the
seawater over the same depth range; such a zone of density change is known as
the pycnocline.
“The seasonal thermocline is an important physical barrier in the
ocean, separating the surface mixed layer for the deeper water. As the
transition region between the nutrient-poor, well-lit surface layer (mixed
layer), and the darker, nutrient-rich deeper water, the thermocline plays a
role in determining the biological properties of the water column”
(Sharples, 2001).
11
In order for a body of water to become stratified, to a degree that would
support blooms of A. catenella, certain physical processes in the environment
must be maintained for extended periods of time. In Washington State’s Puget
Sound Basin, stratification occurs on a seasonal time scale. In early spring
winds become lighter; this reduction in wind is associated with reduced
turbulence, which in turn reduces the mixing occurring in the water column. At
the same time increased solar radiation raises the water temperatures and in turn
increases the mixed layer’s thickness. Reduced amounts of precipitation
influences the water column's surface layer by allowing the water temperatures at
the surface to warm increasing the depth of the mixed layer, as well as reducing
the amount of nutrients available in the surface layer. These processes together
cause the water column to stabilize and vertical mixing to slow down or stop. A
second influx of nutrients is then introduced in the form of snowmelt, which
increases the flow of most Puget Sound Rivers, this colder more nutrient rich
waters will settle to the bottom layer below the thermocline due to its relative
density. Understanding how these environmental drivers shape the aquatic
environment is paramount in understanding how they affect the different life
stage processes of A. catenella.
“The complexities of Alexandrium blooms in dynamic coastal or
estuarine systems are far from understood. One common characteristic
of such phenomena is that physical forcings (environmental drivers) play
a significant role in both bloom dynamics and the patterns of toxicity.
The coupling between physics an biological “behavior” such as
swimming, vertical migration, or physiological adaptation holds the key
for understanding these phenomena, yet this is perhaps where our
knowledge of the genus is weakest” (Anderson, 1998).
12
The best way to understand how different environmental drivers affect
different stages of the life cycle of A. catenella is to look at how each
environmental driver affects the processes associated with different stages in the
life cycle of A. catenella. In order to do this we need to briefly describe the life
cycle of A. catenella in a way that can easily be related to the physical properties
of its environment. For the purpose of this paper I am going to break the life
cycle of A. catenella into four stages: dormancy, germination, growth, and
termination. These are just general categories and will be discussed more in
depth, in relation to individual drivers, later.
Dormancy pertains to the stage when resting cysts of A. catenella lay
dormant buried in sediments on the ocean floor. The duration of this interval,
which is generally considered a time for physiological maturation, varies
considerably among dinoflagellate species and can be anywhere from 12 hours
to 6 months” (Anderson, 1998). They can remain this way for years unless
disturbed by physical or natural forces. When oxygen is present and conditions
are right germination may proceed. Germination typically occurs during spring
and summer when warmer water temperatures and increased irradiance
stimulate the cyst to break open. After the cyst breaks open a swimming cell
emerges. Within a few days of emerging this cell will then reproduce by simple
division. Growth will continue if favorable conditions persist and the cell will
continue to divide, reproducing exponentially. A single cell of A. catenella can
produce several hundred cells within a few weeks, add the fact that there are
usually several hundred cysts and you have the makings of a bloom (Anderson,
13
2005). The termination of A. catenellas growth usually occurs when nutrients
have been depleted in the water column, and or temperatures have dropped
below an optimal range. At this time A. catenella forms a gamete, when two
gametes join a zygote is formed which then turns into a resting cyst. The cyst
then falls to the ocean floor where it awaits in the sediment for the return of more
favorable growing conditions. We can now look at how the physical properties of
its environment affect the various life stages of the dinoflagellate A. catenella.
2.2
Turbulence
It has long been noted that blooms of A. catenella typically occur during
periods of weak winds, which creates less turbulence in the water column.
Turbulence has been described as a “puzzling blend of order and disorder”, but
really turbulence is a property of motion, not of the fluid, and two of its
characteristics are randomness and diffusivity (Estrada, 1998). Not much is
know about the affects of turbulence on cysts dormancy and germination.
However, oxygen may help to stimulate germination and high levels of turbulence
could supply that to the sediment. According to Smayda, (1997) turbulence can
negatively influence dinoflagellate blooms by three mechanisms: physical
damage; physiological impairment; and behavioral modification. Turbulence
affects the growth of A. catenella in four important ways. First, essential nutrients
are transported from below the thermocline to the euphotic zone by turbulence,
negating the need for A. catenella to swim below the thermocline for food, thus
14
modifying its behavior. Second, turbulence helps to suspend A. catenella in the
euphotic zone which saves its energy for cell division because it does not have to
expend energy to stay in the euphotic zone, which is another example of
modified behavior. Third, turbulence may instead transport A. catenella from the
surface to dark sterile waters where it may not survive, an important mechanism
of loss thru physiological impairment. Fourth, strong turbulence can cause
physically damage and destroy cells of A. catenella, sometimes terminating
blooms all together.
Some laboratory experiments in which dinoflagellates were grown under
strong agitation suggest, in general, a negative effect of turbulence on
dinoflagellate cell division and growth (Lee Karp-Boss, 2000). Thus during
periods of sustained strong winds, blooms of A. catenella are slower to develop
and may even experience a decline in population. Sullivan et al. studied the
affects of small scale turbulence on the division rate and morphology of A.
catenella. After performing two replicate experiments Sullivan found the same
results. The division rate of A. catenella was hardly affected by turbulence in the
range of ε (ε = turbulence dissipation rate) approximately 10−8 to10−5 m2 s−3, but
as ε increased above that value there was a 15–20% reduction in division
(Sullivan et al., 2003). In the highest turbulence intensity treatment, the cross
sectional area (CSA) of A. catenella increased approximately 20% (Sullivan et
al., 2003). Sullivan also found that in his unstirred control treatments, 80–90% of
the A. catenella population was found as single cells, indicating that during
periods of low turbulence that A. catenella does not form chains. As the
15
turbulence intensity increased, the number of cells in chains increased (up to 16
cells per chain), at the same time single cells decreased to less then 20% of the
total A. catenella population (Sullivan et al., 2003). However, in the very highest
turbulence intensity treatment, there was a reduction in the percentage of longer
cell chains, and 4 cell chains became more common then 8 cell chains (Sullivan
et al., 2003). Thus according to Sullivan the number of A. catenella cells per
chain were markedly affected by turbulence. Lab experiments are often thought
of as a poor replicate of natural conditions. But Sullivan compared his lab results
with some in situ samples in the East Sound of Washington and found a similar
pattern.
“In 1997, A. catenella populations in East Sound, WA were present at
very high cell concentrations (approx. 100 ml−1), confined to a narrow
depth range between 4 and 5m. The A. catenella layer was in a region
of minimal current shear, and thus presumably in the depth interval
associated with the lowest turbulence intensity in the vertical profile. The
highest values of shear in the profile were in the layers above and below
it. In this profile, there were marked changes in density, with the A.
catenella layer in about the middle of the largest density gradient
between approximately 3.5 and 6m depth. The next summer, 1998, A.
catenella populations were again found in high cell concentrations
(approximately 40 ml−1) in a narrow depth range between approximately
8 and 10m (Fig. 8). Turbulence intensity in the profile was maximal near
the surface (ε approx. 10−5 m2 s−3), but decreased to its minimum values
(less than ε approx. 10−7 m2 s−3) in the A. catenella layer, before rising
again toward the bottom” (Sullivan et al., 2003).
Blooms of A. catenella can be broken up by turbulence, due to excessive
wind. Turbulence increases in the fall and winter and can be sustained for days
to weeks on end. This may help to stimulate cells of A. catenella to begin
forming gametes and zygotes which then turn into cysts and fall to the ocean
floor as inoculums waiting for next year’s bloom.
16
2.3
Precipitation and Runoff
Precipitation and runoff also have an important part in the bloom dynamics
of A. catenella. Precipitation and runoff supply important nutrients that are the
energy that fuels the cell division of A. catenella. This can be especially
important during summer months when nutrients can become depleted in the
surface mixed layer. Not only does precipitation and river runoff supply nutrients,
but they also affect the physical environment. It has long been thought that the
physical changes of a water column that are associated with heavy spring rains
and the spring freshet (peak snowmelt) is an important stimulate for germination
of toxic dinoflagellates cysts, such as those of A. catenella . A study done in
Japan’s Hiroshima Bay, using Alexandrium tamarense a dinoflagellate similar to
A. catenella, demonstrated that runoff and precipitation also affect the salinity
gradient of the water column.
“In 1996, freshwater input had continued from 18 March to 7 May,
followed by a short break 13-20 May, and then increased again between
20 and 28 of June. Although the river runoff affected the density
distribution, it did not likely affect the temperature profile, particularly in
the first half of the period. In 1997, the situation was basically the same
as in 1996, showing that the density stratification was fundamentally
formed by the salinity decrease in the surface water due to river runoff”
(Yamamoto et al., 2002).
However Yamamoto et al. also found that if the rains are too heavy and
runoff too great in terms of volume of flow blooms of A. tamarense can be
dispersed (Yamamoto et al., 2002). To the contrary Weise et al. found that after
17
analyzing a ten year data set on environmental conditions and populations of A.
tamarense in the Gulf of St. Lawrence, Quebec, Canada, higher volumes of river
runoff, such as during the spring freshet, were required for blooms to develop.
“Nonetheless, our results indicate that other factors, such as higher then average
river runoff and low wind speeds, that further intensify vertical stratification, are
required for A. tamarense bloom to fully develop at Sept-Iles” (Weise et al.,
2002). But what defines higher or too great in terms of rain and runoff. Each
species of phytoplankton may have different thresholds, so these types of factors
may need to be assessed on a species by species basis.
2.4
Salinity
Salinity directly influences the density of the water column and contributes
to stratifying the water column. As precipitation and runoff increase the surface
mixed layer becomes less saline. As the surface layer becomes less saline the
thickness of the mixed layer will increase due to the separation of the more saline
deeper waters. This change in water column stratification from salinity is
important as it affects the biological processes of A. catenella both negatively
and positively. However, not much is known about how salinity affects the
dormancy and germination periods of A. catenella. Nor has any data been
compiled on how salinity affects the cell division rate or growth rate of A.
catenella.
18
Salinity is generally used as an indicator of the water column stability; by
looking at changes in salinity over depth and time it has been suggested that
salinity can be used as indicator of the initiation time of blooms of dinoflagellates.
Yamamoto et al. found that in Japan’s Hiroshima Bay salinity induced
stratification may be a good indicator of the timing of blooms of A. tamarense.
“At the beginning of the observations in 1996, the volume of river
discharge may have been too high to sustain the population, while the
bloom coincided with the less stratified period after the flushing, the
bloom also coincided with the less stratified period developing between
the two marked flushing episodes on 31st of March to the 14th of April and
the 6th thru the 19th of May. These results imply that river water runoff
does not support A. tamarense blooms despite the formation of salinity
induced stratification. Water column stability is considered an important
parameter determining species dominancy: dinoflagellates usually
dominate in stratified conditions, and diatoms in mixed conditions.
However, these general criteria do not seem to be the case for spring
blooms of A. tamarense in Hiroshima Bay” (Yamamoto et al., 2002).
Weise et al. found that blooms of A. tamarense in the Gulf St. Lawrence
had a strong negative correlation with surface salinity. “Salinity can be
considered an indication of freshwater input and water column stability and the
importance of both these factors is inferred by the strongly significant negative
correlation between surface salinity and the occurrence of A. tamarense cells at
Sept-Iles in the Gulf of St. Lawrence, during the study period” (Weise et al.,
2002). However this may be region specific, and needs further investigation.
2.5
Water Temperature
19
Water Temperature also affects water column stability and stratification.
This in turn has direct affects on the biological process of A. catenella. The time
it takes a cyst of Alexandrium species to mature during dormancy may be
regulated by water temperatures. “For a single species, this dormancy can vary
with soil storage temperature and the duration of this process can have
significant effect on the timing of recurrent blooms, as cysts with a long
maturation requirement may only seed on or two blooms per year, whereas those
that can germinate in less time may cycle repeatedly between the plankton and
the benthos and contribute to multiple blooms in a single season” (Anderson,
1998). Field studies of A. tamarense have shown that cysts that are subject to
cold water temperatures remain dormant until water temperatures increase
above a certain level, the exact value may be geographically linked and is
undetermined at this time. However, the same pattern was found to be true for
cysts in high water temperatures, which stay dormant until water temperatures
decrease (Anderson, 1998). This indicates that a specific temperature window
may be required for dinoflagellates cysts of the species Alexandrium to
germination. This may help to explain why we see such a seasonal pattern in the
initiation and timing of blooms of A. catenella.
“Various authors have discussed the importance of water temperature in
the dynamics of HAB’s, with general agreement that this parameter plays a
fundamental role both in the germination of cysts, and in vegetative growth of the
cells. Temperature can affect division rates, photosynthesis and respiration, cell
size, and other factors in species participating in blooms” (Navarro et al., 2006).
20
Temperature was one of the main environmental factors which affected the
development of dinoflagellate blooms, with optimal values of 5-8 degrees Celsius
observed at high latitudes” (Navarro et al., 2006).
Cell growth and division can be heavily influenced by water temperature
and the subsequent layering of the water column that is associated with it. It has
long been thought that water temperature may be one of the key factors driving
fluctuations in blooms of toxic dinoflagellates. Navarro et al. (2006) found
optimal cell concentrations of the Chilean species A. catenella at experimental
water temperatures of 12 degrees Celsius, and significantly lower cell
concentrations and growth rates at16 degrees Celsius. This is indicative of the
fact that optimal water temperature requirements may be necessary for extensive
blooms of A. catenella to perpetuate. Not only does water temperature influence
the initiation of blooms as discussed above, but it may also have significant
affects on the toxin content of individual cells. Navarro et al. also found, during
his experiment, an inverse relationship between toxicity and water temperature in
a laboratory experiment, which was reflected in observations of A. catenella in
the field. This suggests that during the spring when water temperatures are
warm but not too hot yet, growth is optimal and as summer progresses toward
fall and water temperatures start to cool down at this time cell growth slows down
and toxin production goes up. Water temperature most likely is also an important
trigger in the termination of blooms of A. catenella. As water temperatures In the
Puget Sound decrease in the fall A. catenellas cell growth and division may go
21
up at first, but eventually slow down and this may trigger the formation of zygotes
and gametes.
In order to understand how environmental drivers and environmental
conditions such as turbulence, precipitation/runoff, salinity and water temperature
set the stage for the initiation of blooms of A. catenella, it is pertinent to
understand how these processes affect different aspects of the life cycle of A.
catenella. By understanding how the physical properties of the aquatic
environment affect A. catenella s life cycle processes we can extrapolate how
large scale oceanic circulation patterns such as ENSO affect the different
physical and biological process associated with the life cycle stages of this
unique toxic dinoflagellate.
22
Chapter 3
Large Scale Oceanic Circulation Patterns Such as
ENSO Affect Regional Weather, and Drive Small
Scale Physical Processes
The term El Niño is Spanish for “the boy Christ-Child” and was originally
used by fisherman to refer to the Pacific Ocean warm currents near the coast of
Peru and Ecuador that appeared periodically around Christmas time and lasted
for a few months. Only later, when it was linked to its atmospheric component
the “Southern Oscillation”, was the phenomenon give the name “El Niño
Southern Oscillation” (ENSO). Hanley et al. described the ENSO as a natural,
coupled atmospheric-oceanic cycle that occurs in the tropical Pacific Ocean on
an approximate timescale of 2-7 years. Which has three phases: warm tropical
Pacific Ocean sea surface temperatures (El Niño), cold tropical Pacific Ocean
sea surface temperatures (La Niña), and near neutral conditions (ordinary
periods). ENSO is a complex ocean atmospheric circulation phenomenon and
many aspects of it are still not well understood.
3.1
The El Niño Southern Oscillation (ENSO)
23
The Southern Oscillation is a seesaw of atmospheric pressure between
the eastern equatorial Pacific and Indo-Australian areas, and is closely linked
with El Niño (Glantz et al, 1991). The Southern Oscillation is the result of winds
blowing over the equatorial Pacific, commonly referred to as the trade winds,
which blow from east to west and are driven by an area of average high pressure
in the eastern part of the Pacific Ocean and a low pressure area over Indonesia.
The southern Oscillation consists of irregular intervals of strong and weak trade
winds, which are related to changing sea surface pressures.
“A physical explanation for the existence of the Southern Oscillation is
provided at least in part by the “Walker Circulation”, a large-scale
atmospheric circulation consisting of sinking air in the eastern Pacific and
rising air in the western Pacific and caused feedback between trade
winds and ocean temperatures” (Katz, 2002).
In General, the tropical Pacific Ocean is characterized by warm surface
water (29-30ºC) in the west and much cooler sea surface water temperatures in
the east (22-24ºC) (Webster et al., 1997). This is due to the fact that during
ordinary (non El Niño/La Niña) periods, eastern trade winds produce cool surface
water in the eastern Pacific Ocean and at the same time they pool, or corral,
warm surface waters in the far western Pacific Ocean. The cooler sea surface
temperatures in the eastern Pacific Ocean are the result of evaporation and the
upwelling of colder sea water from below the surface by the trade winds
converging with the warm pool. The warmer waters in the west create what is
known as the Pacific warm pool. The warm pool is relatively deep, with the
temperature decreasing slowly with depth to the thermocline (see chapter 2)
before dropping off more rapidly.
24
Approximately every two to seven years the state of equilibrium that exists
during ordinary (non El Niño/La Niña) periods breaks down.
“The onset of an El Niño is often marked by a series of prolonged
westerly wind bursts in the western Pacific, which persist for one to three
weeks and replace the normally weak easterly winds over the Pacific
warm pool. At which time during an El Niño event, warming of sea
surface temperatures occurs across the entire Pacific Ocean basin,
which can last for a year or occasionally longer” (Webster et al., 1997).
Figure 4: Mean and anomalies of sea surface temperature (SST) from 1986 to present,
showing El Niño events left 1986-1987, 1991-1992, 1993, 1994 and 1997 and La Niña
events right in 1985 and 1995 (www.pmel.noaa.gov/tao/elNiño/la-Niña-story.html).
During periods of El Niño the easterly trade winds weaken, causing the
upwelling of deep water to cease or slow down in the eastern Pacific. The
consequent increase in sea surface temperatures further weakens winds and
strengthens the El Niño affect. As the trade winds weaken, so does the
containment of the warm water in the west and the maintenance of the cooler
waters in the east. As a result the relatively warmer water of the Pacific pool
25
becomes ubiquitous all across the Pacific Ocean basin (Figure 4, left). This is
commonly known as an El Niño event; typically the Pacific warm pool is
displaced to the east, causing a shift in the major precipitation regions of the
tropics and the disruption of normal climate patterns at higher latitudes (Webster
et al., 1997).
3.2
La Niña
Typically an opposite and cooler state of the tropical Pacific follows a year
or so after El Niño. This phenomenon is El Niño’s twin sister and is known as La
Niña (“the little girl” in Spanish). La Niña is characterized by unusually cold
ocean temperatures in the eastern equatorial Pacific (Figure 4, right), as
compared to El Niño, which is characterized by unusually warm ocean
temperatures in the eastern Equatorial Pacific (Figure 4, left). This cooling of the
Pacific Ocean is brought about partly by increases in the intensity of the eastern
trade winds. The result of the eastern trade winds strengthening is more intense
periods of upwelling in the eastern Pacific Ocean. The upwelling brings more
nutrient rich deep cold water to the surface, and this colder water further cools
the eastern Pacific. At the same time the Pacific warm pool is pushed farther
west due to the stronger easterly winds. These warming (El Niño) and cooling
phases (La Niña) are interspersed with the more common, normal, or quasiequilibrium state, which for the sake of this paper we will call ordinary events or
years.
26
3.3
ENSO Influence on Pacific Northwest Climate and Weather
The ENSO has global impacts on regional weather patterns, however, for
the purpose of this study; we will concentrate on the impacts to Pacific Northwest
weather and climate. The Climate Impacts Group examined monthly average
temperature and precipitation values for El Niño versus La Niña years (19311999). They found that that in the Pacific Northwest El Niño winters tend to be
warmer and drier then average, versus La Niña winters which tend to be cooler
and wetter then average (C.I.G., 2008).
Figure 5: a (left)-b (right): Box-and-whisker plots showing the influence of ENSO on OctoberMarch (a) temperature and (b) precipitation (1899-2000). For each plot, years are categorized as
cool (La Niña), neutral (ENSO neutral), or warm (El Niño). For each climate category, the
distribution of the variable is indicated as follows: range of values (whiskers); median value for the
phase category (solid horizontal line); regional mean for all categories combined (dashed
horizontal line); 75th and 25th percentiles (top and bottom of box). Area-averaged Climate
Division data are used for temperature and precipitation (C.I.G., 2008).
More specifically they found that October through March temperature is
approximately .7 to 1.3 degrees Fahrenheit higher on average, during El Niño vs.
La Niña years (Figure 5a)(C.I.G., 2008). The also found the October through
27
March precipitation is 14% less, on average, during El Niño vs. La Niña years
(Figure 5b)( C.I.G., 2008). Reductions in winter time precipitation amounts
during El Niño years reduce the amount of annual Snowpack in the mountains.
Less snow pack in the mountains means less than average stream flow
during the spring, which when combined with lower precipitation amounts from El
Niño can drastically reduce river flows in the Pacific Northwest. Ropelewski et al
found that areas of Alaska and Canada experienced positive temperature
anomalies in 17 out of 21 (81%) ENSO episodes from 1890 to 1984, during the
“season” defined by December of the ENSO year through the following March.
During El Niño, the northern part of the country experiences a positive shift in
winter mean temperature anomalies central tendency, agreeing with the common
understanding of the El Niño signal (Gershunov et al., 1998). Gershunov et al.
also found the opposite during La Niña years, where the patterns for cold
extreme frequency anomalies are roughly inverses of those for warm extremes
with more cold outbreaks in the north western & western part of the country.
Not only does El Niño affect temperature and precipitation patterns in the
Pacific Northwest, it also has dramatic effects on wind patterns. The strongest
most, persistent wind gust patterns occur during the fall and winter for both El
Niño and La Niña. This is not uncommon as both the warm and cold phases of
ENSO reach maturity during the fall. Enloe et al. found that after examining peak
wind gust from 1948 to 31st of August 1998, that during La Niña events for the
months of November to March, the Pacific Northwest experiences an overall
increase in the gustiness of the winds. While La Niña is associated largely with
28
positive differences (greater than 20%) in monthly means of the peak wind gust
and the frequency of gale force wind gusts El Niño is often associated with
reduced peak wind gust in the Pacific Northwest (Enloe et al., 2004).
“Though strong differences in means are not observed, there is a
general reduction in peak wind gusts over the entire Contiguous United
States during El Niño, as well as a reduction in occurrences of gale force
wind gusts. One persistent warm phase signal, though weak, is
observed in the Northwest. From the beginning of the El Niño year
(October) through February, a general weak reduction n the monthly
mean peak wind gust (0 to -5%) spans most of the country. However,
percent changes of -5% to -10% are more common in the Northwest,
with some magnitudes as great as -15% to -20% occurring during these
months. In a warm phase November, Seattle averages a daily peak
wind gut of 10.3% (19.2kt (knots) or 9.9 meters per second) smaller then
the average neutral (ordinal event) phase wind gust of 21.4kt or 11
meters per second” (Enloe et al., 2004).
Now that we know how different the affects that El Niño and La Niña have on
weather patterns in the Pacific Northwest we can begin to see how they would
have different effects on the different life cycle stages of A. catenella.
29
Chapter 4
Climate Changes Impacts to
Pacific Northwest Weather and ENSO
Twentieth century climate shows evidence of change. We see headlines
in the newspapers on an almost daily basis, but what does it mean for P.S.P.
levels in Washington? According to the projections of a recent study, the Pacific
Northwest will experience warmer, wetter winters and hotter, drier summers (UW
CIG, 2004). While natural climate variability has caused (and will continue to
cause) fluctuations in Pacific Northwest climate on seasonal and decadal scales,
analysis of observed twentieth century conditions shows evidence of longer term
trends that are consistent with modeled projections of twenty-first century climate
change (UW CIG, 2004). These trends include region-wide warming, increased
precipitation, declining snow pack, earlier spring runoff, and declining trends in
summer stream flow.
“To understand the implications of rising temperatures and potentially
increased winter precipitation on the PNW water cycle it may be helpful
to consider 1998 and 1999 water years. In 1998 one of the strongest El
Niño's on record created unusually warm winter temperatures in the
PNW. While snow-pack was only a little below normal for the winter due
to near normal precipitation, the snow began melting approximately one
month earlier then normal and very rapidly due to unusually warm spring
conditions. A warm, dry summer followed and the lengthened summer
season caused by the early melt created water supply problems in some
areas in the late summer. These effects are similar to what we think
would be experienced under climate change on average in the summer.
30
Some years would be dryer, some wetter then the average (Hamlet,
1997)”.
Climate change affects Pacific Northwest weather in two ways, first by
changing the temperature, and second by changing precipitation patterns. With
relation to harmful algal blooms the first would have the most impact, though the
second does play an important role as well. Climate models simulate average
winter temperature increases ranging from about 3.2-3.9 degrees F by the 2020s
and 4.8-6.1 degrees F by the 2050s (Hamlet, 1997). In snowmelt-dominated
systems, these higher winter temperatures would cause more of the dominate
precipitation to fall as rain in the winter, leaving less water stored as snow to
supply summer stream flow. Higher spring and summer temperatures melt the
snow earlier, increasing the length of the growing season, and increasing
summer evapotranspiration, which results in less spring, summer, and fall stream
flow. The earlier melt also effectively lengthens the period between the end of
snowmelt and the onset of fall rains. In hydrologic terms this is like making
summer several months longer then it is now.
This may have drastic affects on the duration and initiation of harmful algal
blooms. As we know A. catenella thrives in a stratified environment (chapter 2).
In the Puget Sound Basin during the summer when stream flow is lowest (low
flow period), this is also the period when the basin becomes heavily stratified.
With warmer climate trends this period of low flow will be arriving sooner and
lasting longer, according to a recent analysis done on streamflow data, spring
streamflow during the last five decades has shifted so that the major April-July
streamflow peak now arrives one or more weeks earlier, resulting in declining
31
fractions of spring and early summer river discharge (Stewart, 2005). This could
have implications for the monitoring and prediction of when blooms of A.
catenella may occur and how long (duration) they may last. If low flow events
are occurring earlier then it follows suit that blooms of A. catenella may occur
earlier in response to environmental conditions. If low flow events will occur over
a longer temporal period than it seems likely that blooms of A. catenella may
occur over a longer temporal period (duration). With this in mind it is important
that we understand how physical processes influence the life history of A.
catenella, before changing hydrological and weather patterns make it even
harder to predict and monitor blooms of this species. This in turn may make it
easier in the future to separate out the anthropogenic influences (nutrient inputs)
from the influences of natural climate drivers such as ENSO.
The International Panel on Climate Change recently published a new
report about the impacts of climate change on ENSO. They found that the 19971998 El Niño event was the largest on record in terms of sea surface
temperatures anomalies and they also found that the global mean temperature in
1998 was the highest on record, at least until 2005 (IPCC 4th assessment, 2007).
It was estimated that the global mean surface air temperatures were 0.17º
Celsius higher for the year centered on March 1998 due to the influence of an El
Niño phase of ENSO. Since the El Niño phase of ENSO is associated with
warming sea surface temperatures it makes sense that if sea surface
temperatures continue to rise, then there will likely be more frequent and more
persistent El Niño phases of ENSO if global warming persist. However, the exact
32
influence of global warming, whether it is anthropogenic or natural, on the phases
of ENSO is not understood at this moment so this is merely speculation. At the
same time whether it is natural or anthropogenic, understanding how changing
climate patterns will affect P.S.P. levels may be crucial to understanding the year
to year fluctuations in P.S.P. levels.
33
Chapter 5
Data Sources
This study involves the analysis of ENSO influence on Paralytic Shellfish
Poisoning (P.S.P) levels present in shellfish throughout Washington States
marine waters from January of 1977 to December of 2007. Statistical analysis
was performed to determine if there is a relationship between P.S.P. levels and
the phases of ENSO.
5.1
ENSO Index
Until recently there has been no consensus within the scientific community
as to which index best defines ENSO years, including the strength, timing, and
duration of events. However, during 2005 the National Weather Service, the
Meteorological Service of Canada, and the National Meteorological Service of
Mexico have reached a consensus on an index (Anonymous, 2005). The index
is defined as a 3-month average of sea surface temperature departures from the
normal average temperatures as defined by the period of time 1971-2000, for a
critical region of the equatorial Pacific (Niño 3.4 region; 5ºN-5ºS, 170º-120ºW).
“Departures from average sea surface temperatures in this region are critically
important in determining major shifts in the pattern of tropical rainfall, which
34
influence jet streams and patterns of temperature and precipitation around the
world” (Anonymous, 2005). For this study the Niño 3.4 region was used to
calculate shifts from ordinary weather events to El Niño and La Niña phases.
The Niño 3.4 region overlaps portions of the Niño-3 and Niño-4 regions covering
an area between 5ºN-5ºS latitudes and 170º-120ºW longitudes. Hanley et al
found the Niño-3.4 and the Niño-4 indices to be the most sensitive for predicting
ENSO events. The Climate Impacts Group found the Niño-3.4 index to be the
best indicator of weather pattern anomalies in the Pacific Northwest associated
with the ENSO (C.I.G., 2008).
The data for this study was acquired from the National Weather Service
Climate Prediction Center (http://www.cpc.noaa.gov). Each phase of ENSO,
whether it be El Niño(+) or La Niña (-), is determined based on a threshold of +/0.5ºC, which is calculated from a 3-month running mean of sea surface
temperature anomalies, in the Niño 3.4 region, departures from normal. An El
Niño or La Niña event occurs when the threshold of +/- 0.5ºC is met or exceeded
for a minimum of 5 consecutive months.
5.2
WDOH Database
Paralytic Shellfish Poisoning (P.S.P.) data were provided by the
Washington State Department of Health (WDOH) Office of Shellfish and Water
Protection. The WDOH has been routinely monitoring P.S.P. throughout
Washington State in both commercial and recreation shellfish since the 1930’s
35
(Cox, 2001). Though A. catenella produces an entire family of toxins, the best
known and the one thought to be the cause of P.S.P. is Saxitoxin, and thus is the
toxin the WDOH test for (Cox, 2001). The minimum Lethal Dose of Saxitoxin
(P.S.P.) for humans is 9µg/Kg of (human) body weight (Halstead, 1965). Figure
6 shows the chemical structure of saxitoxin.
Testing in Washington State began in 1930 after a large outbreak of
P.S.P. related illnesses and deaths that occurred in California, and was
expanded in 1957 to include the northern inland waters of the state after a severe
outbreak of P.S.P. in nearby British Columbia (Cox, 2001). Monitoring was
expanded again in 1978 to include most of Puget Sound after a widespread
outbreak, with P.S.P. levels as high as 30,000 µg per 100g of mouse, caused 10
serious illnesses in the Whidbey Island area. The monitoring program was
expanded again in 1988 when oysters in Carr Inlet had detectable levels of
P.S.P. as high as 2,200 µg each, causing the first shellfish area closure, due to
P.S.P. levels, south of the Tacoma Narrows Bridge (Cox, 2001). Monitoring is
now conducted uniformly throughout all marine waters in Washington State,
regardless of past P.S.P. history.
Figure 6: Chemical structure of Saxitoxin.
36
The WDOH maintains two biotoxin monitoring programs; one for P.S.P,
and one for domiac acid poisoning. The larger of the two is designed to monitor
P.S.P. biotoxin in numerous species of shellfish, including clams geoducks and
oysters. These samples are collected from hundreds of locations throughout
Puget Sound by volunteers, WDOH employees, other state employees, and
industry representatives and then delivered to the Washington State Department
of Health’s Food Safety lab for testing. A 100 grams of shellfish tissue are
collected (needed) for the toxin analysis, this is about a 100 average size (1-2
inches in length) mussels, and then placed in one-gallon size baggies, packed in
Styrofoam containers with frozen gel packs and shipped to the Washington State
Department of Health’s Food Safety lab for testing. Once an area has tested
positive for P.S.P., it will be tested until it has had no positive results for three
consecutive years. Up until 1990 this was generally how samples were collected
and disseminated for analysis. In 1990 the Sentinel Monitoring Program was
started to assess if the WDOH can become more predictive then reactive to
P.S.P. levels.
The Sentinel Monitoring Program is designed to act as an early warning
system for the onset of P.S.P. activity. The Sentinel Monitoring Program helps
guide regional monitoring under the larger general sampling program. A single
species of shellfish is sampled, typically the blue mussel Mytilus edulis however
M. galloprovincialis (a possible subspecies of m. edulis) and the oyster M.
californianus are used at a few Puget Sound sites, from about 40 fixed locations
throughout Washington’s marine waters (Determan, 2003). At most sites, wire
37
mesh cages are periodically stocked with the blue mussel and suspended about
one to two meters deep. Sampling occurs on the frequency of every 2 weeks
during the year, except following an event of high detection levels when samples
are found to have detectable levels of P.S.P. (>38 µg STX/100g). “In other
words, there was increased sampling during a toxic event, to characterize the
extent and severity of the event, resulting in a greater proportion of tests that are
positive for toxin” (Trainer et al., 2003).
The shellfish samples are analyzed using the mouse bioassay procedure.
The mouse bioassay is the most commonly used method for routine analysis of
P.S.P. in shellfish throughout the world and is the accepted method for regulatory
purposes in the United States. During its initial use the mouse bioassay results
were typically expressed as mouse units. However, the mouse bioassay has
been modified since it was first used in the 1920’s. Since then, the US Food and
Drug Administration (FDA) has added a saxitoxin standard, the results are now
expressed in saxitoxin equivalents, STXeq (Trainer et al., 2003). Now, results
are given as micrograms of saxitoxin equivalents per 100 grams of shellfish
meats (µg STX/100g).
The results are then reported to the WDOH office of Shellfish and Water
Protection for entry in their database. Test results are coded and entered into
the database using the following classification system: A -1 indicates that no
toxins were detected; A -2 indicates that no test was preformed, which can occur
for several reasons ranging from not enough meat submitted to the shellfish
spoiling before it reaches the lab; A -3 indicates that only trace amounts were
38
detected, but not enough to determine the amount; A -4 indicates that the level
of P.S.P. present in the shellfish was less than 38µg STX/100g of shellfish; A -5
indicates that the test was unsatisfactory. Results that are greater than 38µg
STX/100g of shellfish are simply reported as the amount of saxitoxin present in
the shellfish (#µg STX/100g of shellfish). Molluscan shellfish with a P.S.P.
content of less than 80 µg/100g meats are permitted to be harvested, processed
and sold.
39
Chapter 6
Analysis and Results
Using the WDOH data set on P.S.P. levels in Washington States marine
waters and the ENSO index from NOAA, statistical analysis can be performed to
determine if a relationship exists between the two. In order to see if there is any
relationship between the P.S.P. levels and ENSO index, the available data were
examined over time. By visually comparing the patterns of distribution between
the ENSO 3-month moving averages departures from normal over time with
P.S.P. levels over time we can see if any similar patterns exist. Time plots of the
ENSO 3.4 region index and the P.S.P. levels over the period January 1 st of 1977
through December 31st of 2007 are shown in Figures 7 & 8.
As we can see from Figure 7 there is a cyclical pattern to the phases of
ENSO but what drives the length of each phase is still being determined by
climatologists (chapter 3). Similar descriptive statistics were used in an attempt
to examine potential seasonal patterns of P.S.P. levels over time. And as we can
see from Figure 8, there is a seasonal pattern of fluctuation to the P.S.P. levels in
Washington’s marine waters; however due to some extreme events of unusually
high P.S.P. levels this seasonal pattern is difficult to discern.
40
2.5
2
1.5
1
0.5
0
-0.5
-1
-1.5
-2
2007
2002
1997
1992
1987
1982
-2.5
1977
3 month running average departures
from normal
3
Time (1977-2007)
ENSO region 3.4 3 month running average departures
normal of +/- 0.5ºC which is calculated from a 3
Figure 7: ENSO over time based on afrom
threshold
month running mean of sea surface temperature anomalies or “departures from normal”
in the Niño 3.4 region which is based on the 1971-2000 base period. An El Niño or La
Niña event occurs when the threshold of +/- 0.5ºC is met or exceeded for a minimum of
5 consecutive months.
35000
PSP results µg STX/100g)
30000
25000
20000
15000
10000
5000
0
Jun-04
Oct-00
Jul-97
Apr-92
Apr-87
Jun-82
Jan-77
Time (1977-2007)
PSP (µg STX/100g) results over time
Figure 8: P.S.P. levels for the period 1/1/1977-12/31/2007, over time.
41
No obvious visual similarities immediately stand out between the ENSO
3.4 region index and the P.S.P. data when visually comparing Figures 7 & 8.
However, as we can see from Figure 8, there are a few outliers (>5,000µg
STX/100g) that are dominating the data and thus masking any regular seasonal
or cyclical pattern. These outliers are likely the result of a localized phenomenon
and not indicative of the overall pattern of P.S.P. seen in Washington’s marine
waters. Removing these extreme events may reveal patterns in the P.S.P data
over time and allow us to see whether they are similar to the patterns of the
ENSO index over time. Further investigation of these outliers should be done to
determine the conditions associated with them, but this is not within the scope of
this thesis.
6000
PSP µg STX/100g results
5000
4000
3000
2000
1000
Jul-04
Oct-00
Time (1977-2007)
Aug-97
May-92
May-87
Jun-82
Jan-77
0
PSP (µg STX/100g) results over time
Figure 9: P.S.P. levels equal to or less than 5,000µg STX/100g over time.
From the distribution of the P.S.P. levels in Figure 9 we can see the
seasonal fluctuations more clearly, but strong variation in the pattern exist from
42
year to year. What causes this variation is still unknown at this time; it could be
weather related, which may become evident through this analysis, or it could be
due to other parameters of A. catenella’s environment, such as trophic
structures.
In order to more appropriately assess whether there is a cyclical pattern to
P.S.P. levels in Washington State marine waters, the monthly P.S.P. values were
converted to a three-month moving average. The three-month moving average of
the P.S.P. data will smooth out the daily variations and show the overall trend of
the P.S.P. data. The three-month moving average transformation was chosen to
match the averaging time of the ENSO index, so that the P.S.P. levels and
ENSO index can be compared, and to provide a better visual graphic of the
distributions of the phases of ENSO and P.S.P. levels over time. From Figure 10
we can see that there is still a great deal of variation but now we can see that
during certain years there are larger peaks in P.S.P. levels.
6000
PSP results µg STX/100g)
5000
4000
3000
2000
1000
0
Oct-04
Sep-01
Sep-99
Oct-96
Sep-92
May-89
Aug-84
Jun-81
Jan-77
Time (1977-2007)
3-Month moving average o f P.S.P.
levels over time
Figure 10: 3 month moving average of P.S.P. µg STX/100g levels over time.
43
However, the ENSO index is expressed as moving average departures
from a baseline and reveals both increases and decreases from it. In order to
determine whether the three-month moving average of P.S.P. levels over time
coincide with the phases of ENSO the P.S.P levels were converted to moving
average departures from the all time average (1977-2007) and the two
parameters were graphically displayed superimposed on the same graph. As we
can see from Figure 11, qualitatively it appears that there are similarities in the
phases of ENSO and P.S.P. levels in Washington States marine waters. This
graphic display suggests that there may be a relationship between the phases of
ENSO and P.S.P. levels.
3
2
1
0
-1
-2
-3
2007
2002
Time (1977-2007)
1997
1992
1987
1982
1977
PSP & ENSO 3 month moving
average departures from normal
4
PSP over time
ENSO over time
Figure 11: Area plot of 3 month running means departures from normal for both ENSO
and P.S.P. levels in Washington States marine waters.
44
Despite being suggestive, this evidence is not adequate to show that there
is a significant relationship between the two variables. In order to determine if
ENSO has a direct affect on P.S.P. levels in Washington States marine waters,
further analysis is needed to determine if there is a relationship between the two.
A simple linear regression was chosen to test the significance of the
relationship between the two parameters. However, because the distribution of
the P.S.P. data and the 3-month moving average showed significant departures
from a normal distribution, the 3-month moving average P.S.P. data were logtransformed to be converted to a normal distribution. Therefore, a simple linear
regression analysis was performed using the ENSO 3.4 region index as the
explanatory variable and the log of the 3-month moving average of P.S.P. levels
as the response variable to see if a linear relationship exists between the two
variables.
The regression showed that there was a significant linear relationship
between these parameters with an F=0.0000 and p =4.75E-10, although the slope
of the line was very low, b1=0.0250 (95% CI=0.0171, 0.0328). Indeed, a very low
R2 value of 0.0000 indicates that the variability of the log of the 3-month moving
average of P.S.P. levels is not adequately explained by the variability of the
ENSO index, and suggests that ENSO alone is not enough to explain the pattern
of P.S.P. levels in Washington States marine waters (Figure 12).
45
4
Log of P.S.P. 3 Month
Moving Average
3
2
1
0
-2
-1
0
1
2
3
ENSO Index
Linear
(=)
Figure 12: Linear Regression of Log of 3 month moving average vs. ENSO Index.
To see if there is any relationship between extreme weather events of
both the cold and warm phases of ENSO and higher P.S.P. levels, linear
regression was performed between the 3-month moving average of P.S.P. levels
and the absolute ENSO index values to indicate only the degree of departure
from baseline. This is the variation of P.S.P. levels related to extreme ENSO
events regardless of the direction away from the baseline, in other words,
extreme departures from normal regardless of whether they are El Nino or La
Nina. The regression showed that there was a significant linear relationship
between these parameters with an F=0.0084 and p =0.0084, although the slope
of the line was very low, b1=0.0163 (95% CI=0.0042, 0.0284). Indeed, a very low
R2 value of .0009 indicates that the variability of the log of the 3-month moving
average of P.S.P. levels is not explained by the variability of the absolute ENSO
46
index, and suggests that absolute ENSO index alone is not enough to explain the
pattern of P.S.P. levels in Washington States marine waters.
As discussed in chapter 3, the phases of the ENSO have different affects
on local weather patterns, and from chapter 2 we know that different weather
parameters have different effects on the life cycle of A. catenella. As such we
would expect to see some influence on the levels of P.S.P. by localized weather
patterns. Ten years (1997-2006) worth of weather data was collected from a
weather station at Seattle Tacoma international airport and entered into Excel
(http://www.wunderground. com/history/airport/KSEA.html). The parameters that
more directly influence the life cycle of A. catenella include salinity, water
temperature, Turbulence, and precipitation. Unfortunately, despite the close
relationship between the A. catenella biotic cycle and these parameters, no
related data was available for this study. These are not measured on a routine
basis. Therefore, other local weather parameters were chosen as proxy
indicators for them. Data available for this study are mean daily air temperature
(Cº), max daily wind speed (mph), average daily wind speed (mph), and daily
precipitation (inches). Initially a simple linear regression was performed for each
of the available parameters.
6.1 Air temperature
Water Temperature affects water column stability and stratification. This
in turn has direct affects on the biological process of A. catenella. Cell growth
47
and division can be heavily influenced by water temperature and the subsequent
layering of the water column that is associated with it. Air temperature was
chosen as a good indicator of water temperature. A linear regression analysis of
the log of the 3-month moving average of P.S.P. levels and mean daily air
temperature was performed to see if linear relationships exist between the two
variables.
The regression showed that there was a significant linear relationship
between these parameters with an F=0.0000 and p =7.87E-55, although the slope
of the line was very low, b1=0.0107 (95% CI=0.0093, 0.0120). Indeed, a very low
R2 value of 0.0300 indicates that the variability of the log of the 3-month moving
average of P.S.P. levels is not explained by the variability of mean daily air
temperature, and suggests that mean daily air temperature alone is not enough
to explain the pattern of P.S.P. levels in Washington States marine waters
(Figure 13).
Log of Moving Average
4
3
2
1
0
0
5
10
15
20
25
30
Mean Daily Air Temperature (C)
Linear (Predicted Log of Moving Average )
48
Figure 13: Linear Regression of Log of 3 month moving average vs. mean daily air
temperature (C).
6.2 Wind
According to Smayda, (1997) turbulence can negatively influence
dinoflagellate blooms by three mechanisms: physical damage; physiological
impairment; and behavioral modification. It has long been noted that blooms of
A. catenella typically occur during periods of weak winds, which creates less
turbulence in the water column thus inducing stratification.
Log of Moving Average
4
3
2
1
0
5
10
15
20
25
Max Wind speed (mph)
30
35
Linear (Predicted Log of Moving Average)
Figure 14: Linear Regression of Log of 3 month moving average vs. max daily wind
speed.
Two indicators of wind were used for this study, max daily wind speed and
average daily wind speed. Max daily wind speed is representative of peak wind
gust for the day while average daily wind speed is representative of the overall
49
wind speed for the day. A linear regression analysis of the log of the three-month
moving average of P.S.P. levels and max daily wind speed (mph) was performed
to see if linear relationships exist between the two variables.
The regression showed that there was a significant linear relationship
between these parameters with an F=0.0000 and p =5.21E-07, although the slope
of the line was very low, b1= -0.0043 (95% CI= -0.0059, -0.0026). Indeed, a very
low R2 value of 0.0031 indicates that the variability of the log of the 3-month
moving average of P.S.P. levels is not explained by the variability of the max
daily wind speed, and suggests that max daily wind speed alone is not enough to
explain the pattern of P.S.P. levels in Washington States marine waters (Figure
14).
A linear regression analysis of the log of the 3-month moving average of
P.S.P. levels and the average daily wind speed (mph) was performed to see if
linear relationships exist between the two variables. The regression showed that
there was a significant linear relationship between these parameters with an
F=0.0000 and p =3.77E-07, although the slope of the line was very low, b1= 0.0054 (95% CI= -0.0074, -0.0033). Indeed, a very low R2 value of 0.0032
indicates that the variability of the log of the 3-month moving average of P.S.P.
levels is not explained by the variability of the average daily wind speed, and
suggests that average daily wind speed alone is also not enough to explain the
pattern of P.S.P. levels in Washington States marine waters (Figure 15).
However it should be noted that the slope is negative for both wind variables
indicating a decrease in P.S.P. with higher wind speeds and by extension
50
supporting the reasoning about higher winds speeds breaking up or dissipating
blooms of A. catenella.
Log of Moving Average
4
3
2
1
0
0
2
4
6
8
10
12
14
16
18
20
22
Daily Wind Speed (mph)
Linear (Predicted Log of Moving Average)
Figure 15: Linear Regression of Log of 3 month moving average vs. average daily wind
speed.
6.3 Precipitation
Precipitation supplies important nutrients that are the energy that fuels the
cell division of A. catenella. This can be especially important during summer
months when nutrients can become depleted in the surface mixed layer. A linear
regression analysis of the log of the 3-month moving average of P.S.P. levels
and the daily precipitation amount (inches) was performed to see if linear
relationships exist between the two variables.
51
The regression showed that there was a significant linear relationship
between these parameters with an F=0.0000 and p=7.64E-06, although the slope
of the line was very low, b1= -0.0622 (95% CI= -0.0895, -0.035). Indeed, a very
low R2 value of 0.0025 indicates that the variability of the log of the 3-month
moving average of P.S.P. levels is not explained by the variability of the daily
precipitation amount, and suggests that daily precipitation alone is not enough to
explain the pattern of P.S.P. levels in Washington States marine waters (Figure
16).
Log of Moving Average
4
3
2
1
0
0
0.5
1
1.5
2
Daily precipitation amount (in)
Linear (Predicted Log of Moving Average)
Figure 16: Linear Regression of Log of 3 month moving average vs. daily precipitation
amount (inches).
6.4 Multiple liner regression analysis
52
Though each of the above variables alone do not explain the variation in
P.S.P. levels seen in Washington States marine waters, perhaps their
combination does allow a more complex relationship to become evident. A
multiple linear regression analysis was performed using the log of the 3-month
moving average of P.S.P. levels as the response variable and a number of
explanatory variables including the ENSO index, daily ambient air temperature
(Cº), daily max wind speed (mph), average daily wind speed (mph), and daily
precipitation amount (inches).
Table 1: Multiple regression of the log of the 3 month moving average vs. the ENSO
index, daily air temperature, max daily wind speed, daily wind speed, and daily
precipitation.
Adjusted R Square
F
Significance F
ENSO Index
Mean Daily Air
Temperature (C)
Max Wind speed (mph)
Daily Wind Speed (mph)
Daily precipitation
amount (in)
0.0383
64.6784
0.0000
slope
0.0314
Confidence
Interval Lower
95%, Upper 95%
0.0236, 0.0393
t Stat
7.8599
P-value
4.35E-15
0.0106
-0.0009
-0.0026
0.0092, 0.0119
-0.0031, 0.0013
-0.0052, 0.0000
15.1930
-0.7939
-1.9441
2.05E-51
0.4273
0.0519
-0.0236
-0.0516, 0.0044
-1.6509
0.0988
As we can see from the significance of F (0.0000) in Table 1 the model is
significant. Also from Table 1 we see that, despite low slope values, the first two
variables, ENSO index and Mean Daily Air Temperature, show a significant linear
relationship with P.S.P. levels. However the last three variables t-stat and pvalues indicate that these variables are not significant. The adjusted R2 indicates
53
that only 3.8% of the variability of P.S.P. levels is explained by the variability of
all these five variables, indicating the model is incomplete. However, the model
including all five explanatory variables has helped increase R2 (from <0.3% to
3.8%) compared to the simple linear models with each variable alone.
However some of the independent variables may be correlated to each
other and that may affect the outcome of the regression. A correlation analysis
shows that only two variables are correlated (Table 2), Max wind speed (mph) is
correlated with average daily wind speed (mph). This is not surprising, since a
higher average daily wind speed is likely associated with a higher max wind
speed for the same day. It also happens to be that the least significant variable
in the first multiple regression output (Table 1) is max wind speed. Thus
removing max daily wind speed from the regression may improve the model.
Table 2: Correlation output among the Independent variables.
Mean Daily Air
Temperature
(C)
ENSO
Index
ENSO Index
Mean Daily Air
Temperature (C)
Max Wind speed
(mph)
Max Wind
speed
(mph)
Daily
precipitation
amount (in)
1
-0.0824
1
0.1513
-0.2053
1
Daily
precipitation
amount (in)
0.0234
-0.1468
0.2804
1
Daily Wind
Speed (mph)
0.0577
-0.1394
0.6264
0.1709
54
The results of the updated multiple regression model (Table 3) without
max daily wind speed (mph) are described in Table 3. The significance of F
(0.0000) in table 3 indicates that this model is significant. In addition, three of the
four explanatory variables, ENSO index, daily ambient air temperature (Cº), and
average daily wind speed (mph), have a significant linear relationship with P.S.P.
levels as indicated by their t-stats and p-values. However, daily precipitation
(inches) has a low t-stat, and the p-value of 0.0627 indicates that this variable is
not significant. The adjusted R2 has not been impacted by the removal of max
wind speed from the model, indicating this second model is improved compared
to the previous model. However, only 3.8% of the variability of P.S.P. levels is
explained by these four variables, indicating the model is still poor.
Table 3: Multiple regression of the log of the 3 month moving average vs. the ENSO
index, daily air temperature, daily wind speed, and daily precipitation amounts.
Adjusted R Square
F
Significance F
ENSO Index
Mean Daily Air
Temperature (C)
Daily precipitation
amount (in)
Daily Wind Speed
(mph)
0.0384
80.6941
0.0000
Confidence
Interval Lower
slope
95%, Upper 95%
0.0310 0.0232, 0.0387
t Stat
7.8262
P-value
5.67E-15
0.0106 0.0093, 0.0120
15.3982
9.60E-53
-0.0260 -0.0534, 0.0014
-1.8616
0.0627
-0.0032 -0.0053, -0.0011
-3.0441
0.0023
Daily precipitation is not significant in model 2 and a new multiple
regression (Table 4) was performed to see whether removing this variable has
55
any impact on the overall fit of the model. The results of model 3 are described in
Table 4. The significance of F (0.00) in Table 4 indicates that this model is
significant, and all the remaining variables are significant as indicated by each
variables t-stat and p-values. However, this final model with only three of the five
explanatory variables available is simpler without compromising the overall fit:
the adjusted R2 remains at the same level. Overall, given the low values of the
coefficients and the low R2, only 3.8% of the variability of P.S.P. levels is
explained by the variability of these three variables, indicating this model is also
poor. In conclusion, all these explanatory variables combined explain little of the
P.S.P. variability in Washington States marine waters and they are probably poor
predictors of high P.S.P. events.
Table 4: Multiple regression of the log of the 3 month moving average vs. the ENSO
index, daily air temperature, and daily wind speed.
Adjusted R Square
F
Significance F
ENSO Index
Mean Daily Air
Temperature (C)
Daily Wind Speed
(mph)
0.0381
106.4042
0.0000
Confidence Interval
slope
Lower 95%, Upper 95%
0.0309 0.0232, 0.0387
t Stat P-value
7.8171 6.10E-15
0.0108 0.0095, 0.0121
15.7536
4.36E-55
-0.0035 -0.0056, -0.0015
-3.3687
0.0008
56
Chapter 7
Implications of This Study
Before we can speculate how this study may be used we need to discuss
who may want to use it. There are four groups of individuals who would benefit
the most from this study. The first group would be the regulators and would
consist of groups or entities who are involved in the regulation of shellfish to
protect public health, such as the United States Food and Drug Administration
and the WDOH who provided the P.S.P. monitoring data as is mentioned earlier
in chapter 5. These are only a few local examples of what a regulator is;
however there are many different types of regulatory agencies worldwide who
would benefit from this information and who may be able to apply a similar study
to other toxic algal species.
The second group would consist of those who harvest shellfish to sell.
This would include shellfish harvesters, growers, and distributors (not equal to
processors). Washington’s shellfish industry is multi-billion dollar industry, and is
economically a very important part of Washington State’s economy. Washington
is the leading producer of farmed bivalve shellfish in the United States,
generating an estimated $77 million in sales (Figure 17) and accounting for 86
percent of the West Coast's production in the year 2000 (http://www
.psat.wa.gov/Programs/shellfish/fact_sheets/economy_web1.pdf). The cost of
57
P.S.P. to the commercial fishery, recreational harvesters, and the aquaculture
industry is believed to exceed $10 million annually (http://www.
nwfsc.noaa.gov/hab/outreach /pdffiles /RedTides2000.pdf).
The third group would be comprised of consumers and individuals who
harvest shellfish for recreation, as well as Native American tribes in Washington
State who harvest shellfish for subsistence, ceremonial harvesting, or for market.
The fourth group would be fellow scientists who may find this study useful in their
own endeavors to better understand why toxic species of phytoplankton exhibit
such variability in between years as well as within year.
Figure 17: Estimated West Coast production of farmed oysters, clams, mussels and
geoduck, 2000 http://www.psat.wa.gov/Programs/shellfish/fact_sheets/economy_
web1.pdf.
The remainder of this chapter is divided into three sections to discuss the
implications this study may have for those that are impacted by P.S.P. in
Washington’s marine waters. The first section will cover the regulatory
implications this study may have. The second section will look at the benefits
58
this study may have for shellfish harvesters, growers, and distributors, whom for
the purpose of this study will be called the shellfish industry, this includes tribal
harvest for commercial sale. The third section will cover how this information
may be important to consumers of shellfish including recreational harvesters, and
the Native American tribes around Puget Sound and on Washington States
Coastal areas.
7.1
Regulatory Agencies and/or Groups
“The Washington State Department of Health (WDOH) monitors biotoxin
levels in shellfish from sites throughout Western Washington to protect
consumers from poisoning by naturally occurring biotoxin’s that if present can
accumulate in shellfish tissue” (Determan, 2003). Better understanding how and
why levels of P.S.P. in Washington State’s marine waters fluctuate the way they
do would be a great advantage in trying to predict when a bloom of P.S.P. may
occur and prevent illness outbreaks before they happen.
The WDOH has been very successful to date with their P.S.P. monitoring
program in preventing illness. However, sometimes there are questions about
how to treat rising levels of P.S.P. that have not gone above the 80µg STX/100g
of shellfish tissue needed to close a beach. When shellfish areas are open for
harvesting the WDOH samples biweekly unless a sample has detectable levels
of P.S.P. which is equal to or greater than 38µg STX/100g, at which time they
59
start sampling weekly. Samples are also submitted for testing prior to any tribal
harvest for commercial, ceremonial and subsistence harvesting (Cox, 2001).
One final recommendation would be to collect more local environmental
data such as water temperature, salinity, air temperature, wind speed and
direction, and rain data and look for relationships that may help to predict blooms
of P.S.P. A complete data set of environmental parameters taken in conjunction
with P.S.P. samples would greatly facilitate the effort to find the most reliable
environmental indicator and predictors of rising P.S.P. levels.
7.2
Shellfish Industry
The Shellfish industry primarily grows, harvests, and distributes the many
diverse and abundant species of commercial shellfish in Washington State.
However, many of the industry are involved in one way or another with the
regulatory aspects and decision making processes of the government. The
shellfish industry is also involved in projects to enhance shellfish beds through
environmental enhancement of the surrounding areas. Since there are many
species of shellfish harvested for commercial purposes in Washington State and
just as many ways to harvest them, this study may need to be applied on a case
by case basis as needed. However, this study may prove useful to the shellfish
industry in a variety of ways such as the ones listed below.
If a relationship can be found between the phases of ENSO and local
P.S.P. levels, then the shellfish industry may want to plan to stop or reduce their
60
harvesting of shellfish during times and phases of ENSO that may be more likely
to have higher levels of P.S.P. present. In order to limit the amount of product
recalled or destroyed, because it gets harvested but does not get sold. This may
be especially beneficial to those companies that harvest species of commercial
shellfish that retain P.S.P. longer in their systems then other shellfish thus
prolonging the unsafe harvest period. For example “Blue mussels (Mytilus
edulis) are quick to pick up P.S.P. toxin and also quick to purge the toxin, once
the mussels stop feeding on the toxic algae. At the other extreme are shellfish
such as Butter clams (Saxidomus nuttalli), that are slow to pick up P.S.P. toxin
and are also slow to purge the toxin” (Cox, 2001).
The shellfish Industry may want to plan for periods when ENSO has the
possibility to impact their business negatively such as mentioned above. They
can potentially avoid the costs of P.S.P. related closures to their businesses by
choosing to not harvest or test more rigorously/frequently during times when
P.S.P. may have the potential to be higher, thus ensuring that the product will not
get recalled and result in immediate loss of money as well as keeping the
shellfish for future sales.
An outbreak of P.S.P. during November and December in 1997 (this
apparently is not the rule: it just happened to be an El Nino phase: it doesn’t
mean that the same should be expected in future El Nino phases, there is not a
good correlation or relationship between the two, and therefore it has no
predictive value) is a good example of the economic hardships P.S.P. has had on
the Shellfish industry that can potentially be mitigated, if a relationship between
61
ENSO and P.S.P. levels exist and can be used to predict when levels of P.S.P.
will be higher. The 1997 P.S.P. blooms severely impacted the oyster harvest in
Puget Sound and in the coastal estuaries of Willapa Bay and Grays Harbor.
Many of the small farms in these areas were closed and suffered great financial
losses.
“A Puget Sound-area farmer of clams and oysters said that he was forced to
close for eight weeks, causing him to miss out on Thanksgiving, Christmas, and
New Year’s sales, and estimated his losses to be $5,000 per week. The P.S.P.
bloom in Willapa Bay and Grays Harbor was felt just before Thanksgiving Day,
which in the oyster industries is the busiest time of the year, accounting for 40%
of the year’s business. Although the coastal bays were reopened by midDecember, sales during the Christmas season were also lost because out-ofstate competitors had moved into the market. About 34 coastal shellfish farms
lost approximately 50% of their sales, reducing average sales by about $8
million. Over 100 workers were laid off and many more had hours reduced”
(http://www.nwfsc.noaa.gov/hab/outreach/pdffiles/RedTides 2000.pdf). Not only
did the small oyster farms suffer in 1997, the entire shellfish industry felt the
economic impacts of the P.S.P. bloom. Large companies had to scale back
shellfish farming in the coastal estuaries as well as in Puget Sound resulting in
the loss of, oyster diggers and shuckers, jobs throughout the state.
One of the processes of shellfish enhancement is to seed the actual
shellfish beds with oysters, clams, and geoduck seeds or inoculums. Different
environmental parameters work better for seeding shellfish beds. Though it is
62
beyond the scope of this paper, enhancing shellfish beds during one phase of
ENSO may prove to have more productive results then the other phases.
7.3
Recreational Harvest and Tribal Ceremonial/Subsistence
Harvesting
Thousands of recreational shellfish harvesters participate in the extremely
popular razor clam fishery in Washington State each year. When recreational
shellfish harvest is closed due to the presence of toxic algae, recreational
harvesters are deprived from their favorite activity and food, but it is not only
them that bear the impact. Recreational shellfish harvesting provides a
significant economic support to local coastal communities especially for service
businesses. Therefore, it is also the hundreds of business owners, and the
thousands employed by them, who greatly suffer from the loss of money spent by
clam diggers that stay overnight or pass through Washington’s many small
coastal communities. The annual value of the coastal razor clam fishery is
estimated at $12 million recreationally and another $7 million commercially
(http:// www.psat.wa.gov/ Programs/shellfish/factsheets/economyweb1.pdf).
Washington’s Puget Sound basin and coastal regions are a unique area
where treaty Indian tribes reside. For ceremonies, subsistence, and commercial
sales, these tribes depend on the harvest of marine species such as oysters,
razor clams, California mussels, littleneck clams, horse clams, butter clams,
gooseneck barnacles and Dungeness crabs. Unfortunately, all of these species
63
can accumulate toxins by filtering seawater. When toxins reach levels too
dangerous for human consumption, tribes can face tremendous economic and
quality-of-life losses. Because of declining fish stocks in the Pacific Northwest,
including rockfish and salmon, tribes are relying more heavily on shellfish then
ever before. Shellfish and crustaceans are a primary source of income to many
tribal members.
The commercial harvest of shellfish by Washington’s Indian tribes would
benefit from this study in the same way as indicated above for the rest of the
shellfish industry. The current WDOH monitoring program will ensure the
continued safety and health of the tribe’s subsistence and ceremonial harvests
by continuing to have samples submitted and tested prior to their consumption or
use. However if one phase of ENSO tends to have higher average P.S.P. levels,
perhaps the tribes may want to be prepared to offset the harvest of shellfish with
alternative food sources and income, though this may not be a viable solution for
all tribes. It is beyond the scope of this paper to provide alternatives, but this is
an important subject to look into in the future so that the Tribes can continue to
sustain their way of life and cultural integrity during those times when P.S.P.
makes the harvest of shellfish unsafe.
64
Chapter 8
Conclusion
Knowing how A. catenella and other harmful algal booms are affected by
the physical processes of their environment may lead us to an answer about why
they are becoming more widespread globally. It may also provide some insight
into why we are seeing the annual patterns of harmful algal blooms, in
Washington State’s marine waters and other coastal areas, that we are and why
they are becoming more widespread globally. From a policy and health
perspective understanding how and why A. catenella is affected by its
environment, and the physical processes associated with it, may lead policy
makers and local health officials to a better understanding of how to predict and
monitor outbreaks of harmful algal species such as A. catenella.
Understanding why P.S.P. levels in Washington State’s marine waters
have such variability inter-annually and seasonally will help in mitigating the
impacts to Washington’s economy and health. This understanding will also
enable regulatory officials to predict when levels of P.S.P. are more likely going
to be higher then average. This will also help shellfish growers and harvesters to
attempt to mitigate the impacts of P.S.P. to their businesses and in turn
Washington’s economy and health as stated in chapter 7. However,
understanding how the physical properties of A. catenellas environment
influences its production of P.S.P. is still something that needs to be determined
65
more accurately. One direction to take would be to attempt to understand how
levels of P.S.P. in Washington change with regional and global weather patterns
such as ENSO.
Even though no relationship was found between ENSO and P.S.P. levels
in this study it does not mean that one does not exist. It may be a more complex
relationship involving other phenomena and combined with variations in local
parameters. The limitations of the data availability at this time limit this study’s
ability to determine what is the relationship between large scale climate
phenomena’s such as ENSO and localized P.S.P. levels, if one does exist.
Given that the global scale ENSO index is not adequate by itself in explaining
variations of P.S.P. levels, and neither ENSO nor three local weather parameters
combined explain more then 3.8% of P.S.P. variability, more detailed information
about local parameters is needed. It would be important to continue to monitor A.
catenellas response to localized weather patterns and relate these to large scale
climate indicators such as ENSO for future studies.
Due to the long-term variability (centennial) of the phases of ENSO, this
study will be more complete as more data are gathered, thus as time goes on we
will have a longer timescale beyond just a few decades to compare P.S.P. levels
to the phases of ENSO. At the same time the monitoring program at the WDOH,
will hopefully continue to collect more P.S.P. samples and add a environmental
data collection aspect to their monitoring program, though more funding may be
required for this. This will also help to ensure a more robust analysis in the
future. If more toxic algae monitoring programs continue to gather and maintain
66
long term databases, in other states and countries, then perhaps others groups
will find that the phases of ENSO, or similar large scale climate phenomena’s, do
have an influence on the levels of toxins present in their areas. The more we
know about toxic species of algae as a whole the more we can begin to
understand the patterns of toxicity (present in space and time) associated with
them.
Determining how regional climate and weather affect P.S.P. levels will not
help us to completely stop the impact of P.S.P. to Washington’s economy and
health; however, if used properly they should help to reduce the impact. And if
more local environmental data are collected in conjunction with P.S.P.
samples over time, perhaps we will be able to determine a relationship
between a physical property of A. catenellas environment and its P.S.P.
production. If a relationship can be determined then we can try to correlate the
physical property responsible with the different phases of ENSO in an attempt to
become more precise in predicting P.S.P. outbreaks. The more we understand
the better we can mitigate the impacts of P.S.P.
67
References:
1. Anderson M. Donald. 1998. Physiological Ecology of Harmful Algal
Blooms: Physiology and Bloom Dynamics of Toxic Alexandrium Species,
with Emphasis on Life Cycle Transitions. Ecological Sciences, vol. 41: pg
29-48
2. Anderson M. Donald. 2004. The Growing Problem of Harmful Algae.
Oceanus, vol. 43, No. 1: pg. 34-39
3. Anderson M. Donald, Pitcher C. Grant, Estrada Marta. 2005. The
comparative “Systems” Approach to HAB Research. Oceanography, vol.
18, No. 2: pg. 148-157
4. Anonymous. 2005. North American Countries Reach Consensus on El
Niño, La Niña Definitions. Bulletin of the American Meteorological
Society, vol. 86, Iss. 4: pg. 473-474
5. Climate Impacts Group. 2008. Impacts of Natural Climate Variability on
Pacific Northwest Climate. University of Washington, Seattle, WA.
http://cses.washington.edu/cig/pnwc/ci/shtml
6. Frank Cox. 2001. Washington’s Marine Biotoxin Monitoring Program.
Washington State Department of Health, Olympia, WA.
7. Determan Tim. 2003. Paralytic Shellfish Poisoning (P.S.P.) Patterns in
Puget Sound Shellfish in 2001. Washington State Department of Health,
vol. 1: pg. 1-14
8. Donaghay L. Percy, Osborn R. Thomas. 1997. Towards a Theory of
Biological-Physical Control of Harmful Algal Bloom Dynamics and
Impacts. Limnology and Oceanography, vol. 42, No. 5, part 2: pg. 12831296
9. Enloe Jesse, O’Brien James, Smith R. Shawn. 2004. Notes and
Correspondence: ENSO Impacts on Peak Wind Gusts in the United
States. American Meteorological society, vol. 17: pg. 1728-1737
10. Estrada Marta, Berdalet Elisa. 1998. Physiological Ecology of Harmful
Algal Blooms: Effects of Turbulence on Phytoplankton. Ecological
Sciences, vol. 41: pg 601-618
11. Fourth Assessment Report of the IPCC. 2007. Observation: Surface and
Atmospheric Climate Change. IPCC, vol. 4 chapter # 3: pg. 287-292
68
12. Gentian Patrick, Donaghay Percy, Yamazaki, Raine Robin, Reguera
Beatriz, Osborn Thomas. 2005. Harmful Algal Blooms in Stratified
Environments. Oceanography, vol. 18, No. 2: pg. 172-183
13. Gershunov Alexander, Barnett P. Tim. 1998. ENSO influence on
Intraseasonal Extreme Rainfall and Temperature Frequencies in the
Contiguous United States: Observations and Model Results. American
Meteorological society, vol. 11: pg. 1575-1586
14. Glantz M., Katz R., Nicholls N. 1991. Teleconnections Linking Worldwide
Climate Anomalies. Cambridge University Press, Cambridge.
15. Glibert M Patricia, Seitzinger Sybil, Heil A. Cynthia, Burkholder M. Joann,
Parrow W. Mattew, Codispoti A. Louis, Kelly Vince. 2005. The Role of
Eutrophication in the Global Proliferation of Harmful Algal Blooms New
Perspectives and New Approaches. Oceanography, vol. 18, No. 2: pg.
198-209
16. Halstead B.W. 1965. Poisonous and Venomous Marine Animals of the
World, Volume 1, Invertebrates. U.S. Government Printing Office, D.C.
Washington: pg. 25
17. Hamlet F. Alan. 1997. Climate Change in the Columbia River Basin.
JISAO Climate Impacts Group-University of Washington, pg.1-6
18. Hanley E. Deborah, Bourassa A. Mark, O’Brien J. James, Smith R.
Shawn, Spade R. Elizabeth. 2003. Notes and Correspondence:
quantitative Evaluation of ENSO Indices. American Meteorological
society, vol. 16: pg. 1249-1258
19. Katz W. Richard. 2002. Sir Gilbert Walker and a Connection between El
Niño and Statistics. Statistical Science, vol. 17, No. 1: pg. 97-112
20. Navarroa J. M., Munoz M.G., Contreras A.M. 2006. Temperature as a
factor regulating growth and toxin content in the dinoflagellate
Alexandrium catanella. Harmful Algae, vol. 5: pg. 762-769
21. Ropelewski C.F., Halpert M.S. 1986. North American Precipitation and
Temperature Patterns Associated with the El Niño/Southern Oscillation
(ENSO). Monthly Weather Review, vol. 114: pg. 2352-2362
22. Sharples Jonathan, Moore C. Mark, Rippeth P. Tom, Holligan M. Patrick,
Hydes J. David, Fisher R. Neil, Simpson H. John. 2001. Phytoplankton
Distribution and Survival in the Thermocline. Limnology and
Oceanography, vol. 46 No. 3: pg. 486-496
69
23. Smayda J. Theodore. 1997. Harmful algal Blooms: Their Ecophysiology
and General Relevance to Phytoplankton Blooms in the Sea. Limnology
and Oceanography, vol. 42, No. 5, Part 2:pg. 1137-1153
24. Stewart T. Iris, Cayan R. Daniel, Dettinger D. Michael. 2005. Changes
toward Earlier Streamflow Timing across Western North America. Journal
of Climate, vol. 18: pg. 1136-1155
25. Sullivan M. James, Swift Elijah, Donaghay L. Percy, Rines E.B. Jan. 2003.
Small-scale turbulence affects the division rate and morphology of two
red-tide dinoflagellates. Harmful Algae, vol. 2: pg. 183-199
26. Trainer L. Vera, Eberhart L. Bish-Thuy, Wekell C. John, Adams G.
Nicholaus, Hanson Linda, Cox Frank, Dowell Judy. 2003. ParalyticShellfish Toxins in Puget Sound, Washington State. Journal of Shellfish
Research, vol. 22, No. 1: pg. 213-223
27. Webster J. Peter, Palmer N. Timothy. 1997. The Past and the Future of El
Niño. Nature, vol. 390: pg. 562-564
28. Weise M. Andrea, Levasseur Maruice, Sucier J. Fancois, Senneville
Simon. 2002. The link between precipitation, river runoff, and blooms of
the toxic dinoflagellate Alexandrium tamarense in the St. Lawrence.
Canadian Journal of Fisheries and Aquatic Sciences, Vol. 59, No. 3: pg.
464-474
29. UW Climate Impacts Group (CIG). 2004. Overview of Climate Change
Impacts in the U.S. Pacific Northwest. University of Washington Climate
Impacts Group, vol. 1:pg. 1-13
30. yamamoto Tamiji, Hahimoto Toshiya, Tarutani Kenji, Kotani Yuichi. 2002.
Effects of winds, tides, and river water runoff on the formation and
disappearance of the alexandrium tamarense bloom in Hiroshima Bay,
Japan. Harmful Algae, vol. 2: pg. 301-312
70