Global change and the future of harmful algal blooms in the ocean

27 Pages·2012·0.82 MB·English

Vol. 470: 207–233, 2012 MARINE ECOLOGY PROGRESS SERIES
Published December 6
doi: 10.3354/meps10047 Mar Ecol Prog Ser
Contribution to the Theme Section ‘Biological responses in an anthropogenically modified ocean’ OPPEENN
ACCCCEESSSS
Global change and the future of harmful algal
blooms in the ocean
Fei Xue Fu*, Avery O. Tatters, David A. Hutchins
The University of Southern California, Department of Biological Sciences, 3616 Trousdale Parkway, Los Angeles,
California 90089, USA
ABSTRACT: The frequency and intensity of harmful algal blooms (HABs) and phytoplankton
community shifts toward toxic species have increased worldwide. Although most research has
focused on eutrophication as the cause of this trend, many other global- and regional-scale
anthropogenic influences may also play a role. Ocean acidification (high pCO /low pH), green-
2
house warming, shifts in nutrient availability, ratios, and speciation, changing exposure to solar
irradiance, and altered salinity all have the potential to profoundly affect the growth and toxicity
of these phytoplankton. Except for ocean acidification, the effects of these individual factors on
harmful algae have been studied extensively. In this review, we summarize our understanding of
the influence of each of these single factors on the physiological properties of important marine
HAB groups. We then examine the much more limited literature on how rising CO together with
2
these other concurrent environmental changes may affect these organisms, including what is pos-
sibly the most critical property of many species: toxin production. New work with several diatom
and dinoflagellate species suggests that ocean acidification combined with nutrient limitation or
temperature changes may dramatically increase the toxicity of some harmful groups. This obser-
vation underscores the need for more in-depth consideration of poorly understood interactions
between multiple global change variables on HAB physiology and ecology. A key limitation of
global change experiments is that they typically span only a few algal generations, making it
difficult to predict whether they reflect likely future decadal- or century-scale trends. We con-
clude by calling for thoughtfully designed experiments and observations that include adequate
consideration of complex multivariate interactive effects on the long-term responses of HABs to a
rapidly changing future marine environment.
KEY WORDS: Climate change · CO · Ocean acidification · Temperature · Stratification · Nutrient
2
limitation · HAB· Algal toxins · Phycotoxins
Resale or republication not permitted without written consent of the publisher
INTRODUCTION In addition to disturbance of natural nutrient
cycles, humans are also causing a massive perturba-
Anthropogenic eutrophication has been linked to tion of the global carbon cycle. The atmospheric par-
worldwide increases in harmful algal bloom (HAB) tial pressure of CO (pCO ) has risen by >30% due
2 2
frequency and intensity in recent decades (Honjo to the burning of fossil fuels, deforestation, indus -
1993, Anderson et al. 2002, Glibert et al. 2005, Halle- trialization, and cement production (IPCC 2007).
graeff 2010). In response to these observations, a These already elevated current CO levels will ap -
2
great deal of research has focused on the influence of proximately double from ~385 to 750−800 ppm by
changing nutrient availability on algal bloom estab- 2100, and ocean pH will consequently decrease by as
lishment and growth. Eutrophication, however, is much as 0.77 units over the next several hundred
only one of multiple global anthropogenic biogeo- years (Caldeira & Wickett 2003), with unknown con-
chemical impacts. sequences for many pH-sensitive marine organisms
*Email: [email protected] © Inter-Research 2012 · www.int-res.com
208 Mar Ecol Prog Ser 470: 207–233, 2012
(Royal Society 2005, Orr et al. 2005, Hoegh-Guldberg investigate how they will respond to changing CO
2
& Bruno 2010). We know that ocean acidification can and/or pH both alone and in combination with other
have potentially far-reaching consequences for the variables. Accurately predicting the responses of HABs
physiology of many algal groups, including altered to these many interacting anthropogenic changes is
growth and carbon fixation rates, shifts in nutrient a top priority for everyone who must deal with the
uptake, changes in elemental ratios, and increased negative impacts of toxic algal blooms, including
sensitivity to ultraviolet radiation (Riebesell 2004, marine resource managers, policy makers, govern-
Fuet al. 2007, 2008a,b, 2010, Feng et al. 2008, 2009, mental management agencies, and marine resource
2010, Hutchins et al. 2007, 2009, Riebesell et al. 2007, users such as the seafood harvesting and aquaculture
2008, Rost et al. 2008, Beardall et al. 2009a, Gao et al. industries. The goal of this review is to summarize
2012a, this Theme Section). These physiological re- the effects of rising pCO in concert with other global
2
sponses may be reflected at the ecosystem level change factors on the physiological and ecological
through changes in algal competitive interactions, responses these organisms.
ecological dominance, and overall community struc- Our review is intended to expand on the excellent
ture (Tortell et al. 2002, Riebesell 2004, Hare et al. recent review of HABs and global change by Halle-
2007, Feng et al. 2009, 2010). graeff (2010), by focusing on how we may apply the
Primary producers including HAB species must results of a large body of prior work on environmen-
adjust not only to altered seawater carbonate chem- tal perturbation effects on HABs to understand their
istry, but also to numerous other concurrent environ- responses to a rapidly changing ocean. Equally im -
mental changes. Over the next 50 to 100 yr, green- portantly, we also emphasize the results of new ex -
house warming will increase average sea surface periments specifically targeting global change effects
temperatures by as much as 5°C, and increased pre- on HABs in this rapidly expanding field. First, we
cipitation, runoff, and ice melting will lower surface briefly review the relevant literature on the effects
salinities in many parts of the ocean (Bopp et al. 2001, of individual global change-relevant variables on
Sarmiento et al. 2002). The combined influence of growth and toxicity, including nutrients, tempera-
warming and freshening on the density of seawater ture, solar radiation, and salinity. In the final sections
will cause much of the surface ocean to become of our review, we cover the limited but particularly
morestratified, driving fundamental shifts in key bio - important body of HAB-related research examining
logical variables such as nutrient supplies and light multivariate interactions between these environmen-
exposure regimes (Boyd & Doney 2003, Boyd et al. tal factors and CO -driven ocean acidification.
2
2008, 2010, Cermeño et al. 2008, Hutchins et al.
2009). As a consequence of these environmental
changes, marine ecosystems all over the world are EFFECTS OF INDIVIDUAL GLOBAL CHANGE
currently changing at an alarming rate. Long-term FACTORS ON HARMFUL ALGAE
data sets from around the world suggest that ongoing
changes in coastal and estuarine phytoplankton com- pCO /pH
2
munities are likely due to the combination of climate
shifts and other anthropogenic influences (Edwards Despite the extensive recent research effort that
et al. 2006, Smetacek & Cloern 2008). has been directed toward understanding ocean acid-
Only a few studies to date have directly addressed ification effects on diverse marine organisms, only a
the implications of ocean global change for HAB spe- handful of studies have so far addressed how CO
2
cies, and most of these have considered the effects or pH changes affect HAB physiology and toxicity.
ofwarming (Peperzak 2003, 2005, Cloern et al. 2005, Some bloom-forming dinoflagellates may especially
Moore et al. 2008, 2009, Paerl & Huisman 2008). benefit from higher pCO , due in part to their CO -
2 2
However, it is virtually certain that harmful blooms fixing enzyme, a type II Rubisco (ribulose-1,5-bis-
of the future will also be simultaneously affected phosphate carbo xylase-oxygenase; Tortell 2000, Rost
by interactions with the complex network of other et al. 2003). Type II Rubisco has a low affinity for car-
changing environmental variables discussed above. boxylation and is therefore extremely inefficient at
Recent evidence demonstrates that some coastal eco- processing CO at present-day atmospheric concen-
2
systems and estuaries are already experiencing sig- trations, comp ared to the type I Rubisco found in
nificant levels of anthropogenic acidification (Feely most other algae. Dinoflagellates overcome this limi-
et al. 2008, Cai et al. 2011). Since HABs often occur in tation, in part, by compartmentalizing Rubisco within
these types of ecosystems, there is an urgent need to the chloroplast to avoid photorespiration (Jenks &
Fu et al.: Effects of global change on harmful algae 209
Gibbs 2000, Nassoury et al. 2001). This low-affinity bubbling the seawater with CO -enriched air) with
2
CO -fixing system may also be compensated for by nutrient limitation. The authors speculated that pCO -
2 2
efficient carbon-concentrating mechanisms (CCMs) induced domoic acid production is perhaps a conse-
such as various forms of carbonic anhydrase (CA), quence of an excess in carbon supply when elevated
which allow algae to access the much more abundant CO occurs together with nutrient-limited growth
2
pool of HCO − in seawater. These adaptations allow conditions. Interestingly though, 2 previous studies
3
some dinoflagellates to grow rapidly at present day found results that differ from these 2 recent studies,
pCO levels, and it is unknown for most species in that domoic acid levels increased instead at higher
2
whether elevated CO will enhance growth further pHs (e.g. lower pCO ; Lundholm et al. 2004, Trim-
2 2
by offsetting the physiological constraints of their born et al. 2008). It is worth noting that unlike the 2
type II Rubisco. more recent studies, in these earlier experiments pH
Dason et al. (2004) showed that the marine dino - was adjusted by HCl and NaOH addition rather than
flagellates Amphidinium carterae and Heterocapsa by CO bubbling. The 2 recent multivariate studies,
2
oceanica do not possess an external CA, and thus and possible reasons for these apparently contradic-
their photosynthesis is dependent on free CO alone. tory results, are considered further in the ‘Interactive
2
Consequently, the growth of these 2 species is sug- effects of CO and nutrients’ section below. Despite
2
gested to be CO -limited (Colman et al. 2002, Dason the differences in their findings, all 4 of these studies
2
et al. 2004), and their growth could potentially be support the suggestion that pCO /pH can strongly
2
stimulated by increasing CO concentrations in the influence the production of domoic acid by this glob-
2
future ocean. In contrast, the growth rates of 3other ally distributed diatom genus.
marine dinoflagellates (Prorocentrum minimum, H. The impact of elevated CO on the growth of CCM-
2
triquetra, and Ceratium lineatum) are most likely utilizing diatoms versus algal species without a CCM
notlimited by dissolved inorganic carbon, since they (Riebesell 2004) suggests that those species which
preferentially take up HCO − instead of CO to sup- lack CA will likely benefit most from rising CO levels.
3 2 2
port photosynthesis (Rost et al. 2006). This obser - Notably, the raphidophyte Heterosigma akashiwo
vation is supported by the finding that increasing does not appear to use CA (Nimer et al. 1997), sug-
CO does not significantly affect the growth rate of gesting that it may be especially favored by rising
2
another isolate of P. minimum(Fu et al. 2008a). CO levels. In fact, the growth of H. akashiwo is
2
Because phycotoxin biosynthesis is directly linked significantly stimulated by increasing CO , again
2
to the autotrophic metabolism of most HAB species, it achieved by bubbling the seawater with air/CO
2
is perhaps not entirely surprising to find that chang- mixtures (Fu et al. 2008a). However, this finding may
ing CO availability can also affect cellular toxicity. not apply to all raphidophytes. For instance, the
2
Photosynthesis is not only the essential process in pri- growth of Chattonella marina is not affected by pH
mary metabolism, but is also required for toxin pro- over a range from 7.5 to 8.5 (adjusted by acid and
duction (Pan et al. 1996a). For example, the yield of base additions), although growth greatly decreases
saxitoxin per cell in the dinoflagellate Alexandrium at pH values over 9.0 (Liu et al. 2007). Coupled with
catenella is proportional to hours of daylight (Proc- these reduced growth rates, rates of ichthyotoxic
tor et al. 1975). Also, A. minutum is not capable of reactive oxygen species (ROS) production by C.
producing saxitoxin after a 22 d incubation period in marinaalso increase at this elevated pH, but remain
the dark, while parallel light-grown cultures pro- stable within the pH range of 7.5 to 8.5. Liu et al.
duced 1.17 µg per 10000 algal cells (Maas & Brooks (2007) suggested that high pH may enhance the
2010). activities of enzymes that regulate ROS production,
pCO and/or pH changes affect toxicity of the and/or that high pH may reduce iron bioavailability
2
diatom genus Pseudo-nitzschia, which causes no - to the algae.
toriously damaging blooms along the Pacific coast The prymnesiophyte Phaeocystis globosacan form
of North America and elsewhere around the world massive harmful blooms in temperate areas such as
(Scholin et al. 2000, Trainer et al. 2000, 2009, the North Sea. Recent evidence suggests that its
Schnetzer et al. 2007). Two recent studies have physiological responses to changing pCO may be
2
examined the influence of seawater pH on the toxi- dependent on its polymorphic life history, which
city of cultures of Pseudo-nitzschia spp.: Sun et al. alternates between solitary flagellated cells and
(2011) and Tatters et al. (2012) found that domoic colonies composed of numerous cells embedded in a
acid concentrations increase dramatically in treat- gelatinous matrix. Wang et al. (2010) demonstrated
ments combining high pCO /low pH (adjusted by that bubbling P. globosa cultures with elevated CO
2 2
210 Mar Ecol Prog Ser 470: 207–233, 2012
stimulates the formation and growth rates of nying stratification seem to promote the proliferation
colonies, but the growth rates of solitary cells are of many microalgae, including several harmful spe-
unchanged. Based on the observed in creases in cies (Paerl & Scott 2010). Cellular toxicity can also be
colony formation, these authors suggest that future sensitive to rising temperature. For instance, cultures
rising CO may affect carbon and sulfur cycles as and field samples of Karlodinium veneficum exhibit
2
well as marine trophic structure both locally and increased cellular toxicity at temperatures >25°C
regionally. However, a natural assemblage of the (Kempton et al. 2002, Adolf et al. 2009).
closely related polar species P. antarcticais relatively The relationship between HABs and warming is
unaffected by extended incubation at elevated pCO not always straightforward. As toxic diatoms of the
2
(Feng et al. 2010). genus Pseudo-nitzschiatypically respond to seasonal
This short list summarizes the published studies on patterns, temperature is likely a critical driver in their
HAB species responses to ocean acidification in iso- bloom development. Depending on geographical
lation; these experiments and a few others examining region, seasonal blooms have been correlated with
rising pCO in combination with other variables are pulses of cool, nutrient-rich upwelled water (Horner
2
summarized in Table 1and are reviewed below. This et. al 1997, Trainer et al. 2002, Kudela et al. 2010),
surprising paucity of information on high pCO /low and also with warmer, stratified conditions (Bird &
2
pH effects needs to be remedied by further studies Wright 1989, Buck et al. 1992, Horner et al. 1997,
using a variety of environmentally relevant species, Scholin et al. 2000). In the laboratory, growth rates
thus there is likely to be new information available of a temperate isolate of P. pseudodelicatissima in -
on this subject within the next few years. crease up to 25°C (Lundholm et al. 1997). Tempera-
ture could also play a role in regulation of enzymatic
pathways involved in domoic acid biosynthesis by
Temperature Pseudo-nitzschia. Although there have been labora-
tory studies of Pseudo-nitzschiaspp. growth rates as
Temperature is probably the most widely recog- a function of temperature (Lundholm et al. 1997,
nized component of climate change and also plays a Thessen et al. 2009), effects on domoic acid pro -
crucial role in determining potential algal growth duction have been examined surprisingly seldom.
rates. Consequently, temperature can influence com- In 2 culture studies, warmer temperatures were not
munity dynamics of harmful bloom species relative demonstrated to accelerate domoic acid production
to their competitors and grazers. In diatoms, for ex - (Lundholm et al. 1994, Bates et al. 1998). Similarly,
ample, nitrate uptake and reduction decline rapidly little or no correlation was observed between cellular
at elevated temperatures (Lomas & Glibert 1999), domoic acid and temperature during field observa-
potentially favoring competing algae. Likewise, tem- tions in Chesapeake Bay and the northern Gulf of
perature can differentially impact the growth rate, Mexico (Thessen & Stoecker 2008, MacIntyre et al.
pigment content, light-harvesting capacity, and pho to - 2011).
synthetic carbon fixation of many microalgae (Sosik Temperature shifts may affect the spread of
& Mitchell 1994, Coles & Jones 2000, Anning et al. Pseudo-nitzschia to new habitats. The persistent
2001, Stramski et al. 2002). seasonal nature of these blooms once seed popula-
Increasing sea surface temperatures are already tions become established in supportive areas can be
leading to prolonged and more intense temperatures quite remarkable. For instance, recent data from Bar-
during bloom seasons (Peperzak 2003, 2005, Edwards ron et al. (2010) suggest that cooling waters of the
et al. 2006, Hallegraeff 2010, Paerl & Scott 2010), and North Pacific influenced by the negative Pacific
this trend is likely to continue with the potential for Decadal Oscillation have coincided with the sudden
establishment of temporally and spatially expanded 1999 appearance of P. australis and P. multiseries in
bloom windows (Fig. 1; Moore et al. 2008). Many the sedimentary record of California’s Santa Barbara
HABs have a window of temperature that is reached basin. These HAB diatoms remained significantly
and often exceeded within a given year (Gobler et more abundant relative to other diatoms such as
al. 2005, Moore et al. 2008). Therefore, in some Chaetoceros and Rhizosolenia spp. as of 2003. Al -
cases increasing temperature may not intensify though blooms of toxic Pseudo-nitzschia spp. are
HABs throughout the growing season, but perhaps now a common annual feature of this region, this
instead change the timing of their initiation and study could also be taken to suggest that long-term
termination during the annual seasonal cycle. Warm warming trends might contract the ranges of these
water temperatures, calm conditions, and accompa- organisms.
Fu et al.: Effects of global change on harmful algae 211
et;
ure of CO; UV: ultraviol2 Source Fu et al. (2007) Present work(Fig. 4) Liu et al. (2007) Fu et al. (2010) Present work(Fig. 5) Hwang & Lu (2000) Flores-Moya et al. (2012) Sun et al. (2011) Tatters et al. (2012) Lundholm et al. (2004), Trimborn et al. (2008) Chen & Gao (2011) Hoogstraten et al. (2012) Wang et al. (2010)
s
s
e
pr a-
acidification in isolation. pCO: atmospheric partial 2R: photosynthetically active radiation Response Increased growth at high pCOin Heterosigma2but not ; increased growth with Prorocentrumwarming in both species Increased growth at high pCOand temperature2 Growth reduced and reactive oxygen species production at pH >9.0 Increased growth and toxicity at high pCO; 2karlotoxins induced under P limitation Increased saxitoxin at high pCOand low 2temperature Growth and cellular gonyautoxin-1 and -4 quota highest at pH 7.5 Growth increase under combined increasing temperture and low pH, but toxicity trends unpredictable Increased domoic acid under P limitation and under high pCO2 Increased domoic acid under Si limitation and under high pCO2 Domoic acid levels increased at high pH in -spp. but unchanged in PseudonitzschiaNitzschia Under high PAR, high COinhibited growth 2without UVA and UVB, but either type of UVR further inhibited growth Under high light, growth increased with decreasing CObut was unaffected by CO22under low light Elevated COstimulated the formation and 2growth of colonies, but growth rates of solitary cells were unchanged
n A
aP
(HAB) responses to oceVR: ultraviolet radiation; Variable ranges 380−750 ppm20−24°C 380−750 ppm20−24°C 7.5−9.5 190−750 ppm0.5 and 20 µM P 380−800 ppm15−19°C pH 5.5, 7.5 and 8.5 pH 7.5 and 8.0;20 and 25°C 190−750 ppm;0.5−20 µM P 200−765 ppm;10.6−106.1 µM Si 7.9−8.9 393 ppm and 1013 ppm;varying PAR, UVA, and UVB 190−750 ppm80 and 240 µmol −2−1photons ms 380 and 750 ppm
m U
o
mful algal blo Variables and CO2temperature COand 2temperature pH COand 2phosphate COand 2temperature pH COand 2temperature COand 2phosphate COand SiO24 pH COand UV2 COand light2 CO2
r
a
h
able 1. Published studies on HAB taxon and Heterosigma akashiwoProrocentrum minimum(rhaphidophyte and dinoflagellate) Chattonella subsalsa(rhaphidophyte) Chattonella marina(rhaphidophyte) Karlodinium veneficum(dinoflagellate) Alexandrium catenella(dinoflagellate) Alexandrium minutum(dinoflagellate) Alexandrium minutum(dinoflagellate) -Pseudonitzschia multiseries(diatom) -Pseudonitzschia fraudulenta(diatom) -, Pseudonitzschia seriata, P. multiseriesNitzschia -(diatoms)navisvaringica Phaeocystis globosa(prymnesiophyte) Phaeocystis globosa(prymnesiophyte) Phaeocystis globosa(prymnesiophyte)
T
212 Mar Ecol Prog Ser 470: 207–233, 2012
are commonly found in tropical/sub-tropical waters
(Usup et al. 2012). The western Atlantic form of P.
erature (°C)1250 13°C threshold + + +462°°C°CC 11 629971 mmmooorrreee dddaaayyysss btafeishictas ecahhau alipnm.nm ue 2uttue0halnra0etosit6 vteeIo)ned.x dw sTiinsaahaasxnsex i stiaRotfesooix svxrrioemniecnr si essuLao rtlalhuteygsard ocv steohehw.n oo iA buwtihnegs ehFtwPnhtlS oi atfrPnthoi ,od Puotab.n t hubd(tLoeta arihr nnpdeadrcmiponsedubeonufneftfllcrseaygeer-,
p
m gellates, laboratory studies have dem onstrated that
e
T they have broad temperature windows, e.g. 22 to
10 Historical window
of increased opportunity 35°C (Usup et al. 1995) and 23 to 37°C (Gedaria
68 days et al. 2007). The distribution of P. bahamense var.
bahamensein coastal waters of Florida suggests that
J F M A M J J A S O N D the minimum temperature that limits its occurrence
Month is 20°C (Phlips et al. 2006). These temperature toler-
ances support po tential climate-related range ex -
Fig. 1. Alexandrium catenella.Expansion of the annual tem-
poral window for blooms of the saxitoxin-producing dinofla- pansion (Usup et al. 2012). Although the majority
gellate in Puget Sound, Washington, USA, at projected sea of environmental variables examined influence the
surface warming levels of +2, +4, and +6°C. The species is PSP toxin profile, not total toxin content, one study
limited to blooming at temperatures over 13°C; with a 6°C
demonstrated increased toxicity at low temperature
average increase, the annual bloom window could expand
(Usup et al. 1994).
from the historical value of 68 d to as long as 259 d (from
Moore et al. 2008) The increased abundance, geographical range ex -
pansion, and growing severity of ciguatera fish poi-
Proliferation of paralytic shellfish poisoning (PSP)- soning occurrences are likely indicators that several
producing dinoflagellates of the genus Alexandrium members of the benthic/epiphytic dinoflagellate genus
also tends to be seasonally and regionally specific. As Gambierdiscus are responding to warming sea sur-
described by Anderson et al. (2012, p. 29) face temperatures and habitat transformation by con-
current spreading of the marine macroalgae with
Overall, the Alexandrium species that have been
which they are associated (Morton et al. 1992, Hales
studied in detail have proven to be remarkably resili-
ent and capable of colonizing a wide spectrum of et al. 1999, Chateau-Degat et al. 2005, Parsons et al.
habitats and hydrographic regimes. It is thus of no 2012). One culture study examining effects of tem-
surprise that the biogeographic range of these species perature on ciguatoxins of G. toxicusdemonstrated a
has expanded in recent times and that associated PSP
positive correlation (Bomber et al. 1988). The range
outbreaks remain a significant global problem.
of Gambierdiscus is rapidly expanding along with
Once cyst beds become established in a given another toxic dinoflagellate genus, Ostreopsis,which
locality, temperature may determine periods of is not closely related to Gambierdiscusand produces
excystment and vegetative growth (Anderson et al. quite different toxins, but also shares a benthic/
2005). Annual variability in PSP-contaminated shell- epiphytic lifestyle (Tindall & Morton 1998, Rhodes
fish could result from either changing seasonal inci- 2011,Parsons et al. 2012). The majority of laboratory
dence of toxic Alexandriumblooms, variations in tox- experiments examining temperature suggest that
icity by resident dinoflagellates, or a combination of Os treopsis grow more efficiently at high tempera-
both (Siu etal. 1997). Arguably, each of these scenar- tures, but are more toxic at lower temperatures (Shears
ios could be temperature-related. Correlations & Ross 2009, Granéli et al. 2011, Rhodes 2011).
between cooler temperature and enhanced Alexan- Growth and toxicity of other HAB dinoflagellates
driumtoxicity have been reported by numerous cul- can also be positively or negatively related to sea -
ture and field investigations (e.g. Hall et al. 1982, water warming. Temperature affects toxicity in some
Ogata et al. 1987, Cembella et al. 1988, Anderson et diarrhetic shellfish poisoning (DSP)-producing Pro-
al. 1990). In contrast, enhanced toxicity at median or rocentrum spp. (Morton et al. 1994) and Dino physis
increased temperature is less common but has also spp. (Kamiyama et al. 2010, Tong et al. 2011). In a
been documented (Siu et al. 1997, Etheridge & study by Peperzak (2003), P. micansand P. minimum
Roesler 2005, Lim et al. 2006). doubled their growth rates in simulated warm strati-
Pyrodinium bahamense var. bahamense and var. fied conditions. The majority of studies on yesso-
compressum, 2 other PSP-producing dinoflagellates, toxinand analogues produced by Protoceratium reti -
Fu et al.: Effects of global change on harmful algae 213
culatumsuggest that toxicity in creases with temper- nities due to altered ocean temperature and mixing,
ature (Guerrini et al. 2007, Paz et al. 2007). The whether climate change will similarly affect the
brevetoxin-producing dinoflagellate Karenia brevis, abundance or distribution of nearshore and estuarine
which causes mass mortality of marine life in the Gulf HABs is far from clear.
of Mexico, has been observed in the field between
7 and 34°C (Brand et al. 2012). However, optimal
growth in laboratory cultures is between 22 and 29°C Nutrients
(Magana & Villareal 2006, Vargo 2009). The closely
related K. mikimotoi has also been found over a Future climate variations such as changing storm
widerange of temperatures (4to 32°C; Gentien 1998, frequencies and wind patterns will affect coastal
Brand et al. 2012). Toxin production in K. brevis water column dynamics, including frequency and
demonstrates a trend of slightly higher toxicity at intensity of upwelling events, tidal mixing, and
lowtemperatures that impair growth (Lamberto et al. mixed layer depths (Doney et al. 2009, Hallegraeff
2004), suggesting the possibility of reduced breve- 2010). Both warming and freshening of the surface
toxin impacts in a future warming ocean. ocean from increased precipitation will promote
A recent 50 yr time series study in the northeast increased seasonal water column stratification in
Atlantic and North Sea shows that phytoplankton coastal waters (Hallegraeff 2010, Paerl & Scott 2010),
community structure has shifted away from dino - as well as increases in permanent stratification in the
flagellates, including harmful species such as some open ocean gyres (Gentien et al. 2005, Polovina et al.
Prorocentrum spp. and non-harmful taxa such as 2008). The implications of this increased stratification
Cera tixum fuca, and towards diatoms such as for HABs are likely profound, since many coastal and
the potentially toxic Pseudo-nitzschia spp. and non- offshore blooms depend on vertical mixing to supply
HABs such as Thalassio siraspp. (Hinder et al. 2012). nutrients from below (Cermeño et al. 2008, Boyd et
The combined effects of increasing sea surface tem- al. 2010). More rapid depletion of surface nutrients
perature and increasingly windy conditions in sum- and concurrent decreases in replenishment from
mer were suggested to be the main reasons for this deeper water will likely favor pico- and nano-size
observation. However, Hinder et al.’s (2012) results species (Hallegraeff 2010). Reductions in bioavail-
do not neces sarily apply to many HAB species, since able silicate (Goffart et al. 2002) may also lead to
the survey focused on an open ocean phytoplankton decreased diatom abundance (Hallegraeff 2010),
community, and most HABs occur in estuaries or which could inhibit harmful blooms in some cases (e.g.
coastal waters. Local physical dynamics in these 2 those of the toxic diatom genus Pseudo-nitzschia),
regions are completely different. Nutrients are gen- and promote them in others (e.g. when non-toxic
erally much more enriched in estuaries than in the diatom species are important competitors with toxic
open ocean, and estuaries and bays are usually less dinoflagellates).
affected by wind-driven physics. Some harmful taxa In a classic aquatic ecology paper, Margalef (1978)
are warm-water species and hence slightly increas- suggested that diatoms are best adapted to nutrient-
ing temperature may favor their growth, in particular enriched, well-mixed water columns, while dinofla-
many dinoflagellates. Calm winds and warmer tem- gellates dominate in stratified, more oligotrophic
peratures will stratify the water column and suppress environments. More recently, it has been repeatedly
mixing long enough for motile dinoflagellates to suggested that motile species such as many harmful
grow and accumulate in surface waters, and hence dinoflagellates and raphidophytes have a distinct
allow them to bloom. advantage in obtaining nutrients by vertical migra-
Recent data link harmful dinoflagellate blooms to tion (Smayda 1997, Handy et al. 2005, Hallegraeff
warmer temperatures. For instance, increasing tem- 2010, Paerl & Scott 2010); thus enhanced stratifica-
perature stimulates blooms of the toxic dinoflagellate tion could offer these groups a competitive advan-
Alexandrium in Puget Sound in Washington state tage. For instance, Alexandrium tamarense cells liv-
(Moore et al. 2009). A large unprecedented dinofla- ing in N-limited waters are likely able to sustain
gellate bloom was observed in San Francisco Bay in growth and moderate toxicity if they are able to per-
September 2004, and one of the conditions that was form diel vertical migration to N-rich depths (Mac -
thought to have caused this bloom was high air Intyre et al. 1997). Along with many other genera
temperatures (Cloern et al. 2005). Although the study (Gentien et al. 2005), blooms of Alexandrium are
by Hinder et al. (2012) convincingly dem onstrated usually found subsurface under stratified conditions
multi-decadal changes in oceanic plankton commu- (Cembella & Therriault 1989).
214 Mar Ecol Prog Ser 470: 207–233, 2012
The mixotrophic capabilities of many dinoflagel- and karlotoxins, are less dependent on the availa-
lates (Stoecker 1999, Glibert & Burkholder2011) may bility of this nutrient (Adolf et al. 2009).
afford these organisms even more flexibility under In addition to various limiting nutrient scenarios,
future stratified, low-nutrient conditions (Caron & the chemical form or speciation of nutrients can also
Hutchins in press). Growth rates of some faculta- affect algal toxicity. The bloom-forming dinoflagel-
tively mixotrophic harmful species increase when late Karenia brevis shows little response of breve-
they are supplemented with prey (Adolf et al. 2006, toxin production to nutrient limitation (Lekan &
Glibert et al. 2009). This alternative metabolic strat- Tomas 2010), but is enhanced when grown on urea
egy would offer a potential for broader niches and versus nitrate (Shimizu et al. 1995). N spec iation can
alternative resource exploitation under both nutri- also have implications for toxicity in Alexandrium
ent-poor and eutrophic conditions. spp., since saxitoxin production is enhanced when
Of course, in many coastal and estuarine regimes, cultures are grown on ammonium as opposed to
cultural eutrophication may be more important than either nitrate or urea (Levasseur et al. 1995, John
increased stratification in determining future nutri- & Flynn 2000, Hamasaki et al. 2001). In natural
ent availability. In particular, in future climate re - blooms of the diatom Pseudo-nitzschia, domoic acid
gimes, some regions of North America are predicted levels increase with N source in the order urea >
to experience either more precipitation, or the same nitrate > ammonium (Armstrong Howard et al.
amount of precipitation delivered in fewer and thus 2007). However, a recent Pseudo-nitzschia spp. lab-
larger pulses (IPCC 2007). This could result in heav- oratory study showed that N sources may affect cel-
ier and more intense nutrient loading to coastal and lular domoic acid content in species- or strain-spe-
estuarine ecosystems, and this perhaps stimulation of cific ways (Thessen et al. 2009). That work (op.cit.)
HAB events. suggests that there is no general trend regarding
In addition to generally increased nutrient load- effects of N source on cellular domoic acid levels,
ing, coastal ecosystems can also experience unbal- and that it is consequently important to consider
anced nutrient ratios from anthropogenic inputs, intra- and interspecies variability in ecophysiology
potentially leading to limitation by either phospho- and toxicity.
rus (P) (high N:P ratios) or nitrogen (N) (low N:P Ocean acidification may be relevant to this de -
ratios) (Smayda 1997). These skewed nutrient ratios pen dency of HAB toxicity on specific N sources,
can have significant but genera-specific effects on since low pH has been shown to inhibit nitrification
physiological characteristics, in particular their cel- and so could ultimately shift the speciation of the
lular toxicity. For instance, Guerrini et al. (2007) overall ocean N inventory away from nitrate and
found that yessotoxin production by the dinoflagel- towards reduced species such as ammonium and
late Protoceratium re ticulatum is stimulated by P organic nitrogen (Hutchins et al. 2009, Beman et
limitation, but not by N limitation. Likewise, saxi- al. 2011). A model of the North Sea at 1000 ppm
toxin-producing members of the dinoflagellate genus CO suggests that ammonia oxidation rates could
2
Alexandrium demonstrate increased toxicity only be in hibited by as much as 20%, resulting in a de -
under P limitation (Boyer et al. 1987, Anderson et al. crease of the nitrate to total dissolved inorganic
1990, Siu et al. 1997). In contrast, for the dinoflagel- ratio by up to 10% (Blackford & Gilbert 2007).
late Karlodinium veneficum, karlotoxin concentra- Such a substantial shift in the chemical form of
tions increase significantly under conditions of either N supplied to phytoplankton communities under
N or P limitation (Adolf et al. 2009). In the dinofla- acidified conditions could potentially favor smaller
gellate Prorocentrum lima, N and P limitation both organisms that are more competitive for ammo-
inc rease cellular concentrations of the toxin okadaic nium, such as pico eukaryotes and cyanobacteria,
acid (Vanucci et al. 2010), but in Dinophysis acumi- as well as some HAB species such as dinoflagel-
nata, okadaic acid levels increased only under N lates and raphidophytes.
limitation (Johansson et al. 1996). Intracellular con-
centrations of domoic acid in some toxic Pseudo-
nitzschia species are enhanced by Si and P limita- Solar irradiance
tion, but not by N limitation (Bates et al. 1991, Pan
et al. 1996b,c). Often, the synthesis of N-rich toxins Light is obviously a key factor affecting the physio-
such as domoic acid and PSPs is reduced with N logical responses of all photoautotrophs, including
limitation (Boyer et al. 1987, Bates et al. 1991), while HAB species. Irradiance regimes will change for pri-
that of t oxins containing no N, such as yessotoxins mary producers in many areas of the future ocean due
Fu et al.: Effects of global change on harmful algae 215
to the increased stratification and mixed layer shoaling In contrast to these HAB species, there is evidence
(discussed in the ‘Nutrients’ section above). Phyto- that the dinoflagellate Karenia brevis appears to be
plankton circulating in a shallower mixed layer will relatively low-light adapted. This species has a low
necessarily be exposed to higher mean daily doses of light saturation point of around 65 µmol m−2 s−1
photosynthetically active radiation (PAR; Boyd et al. (Shanley & Vargo 1993, Magana & Villareal 2006),
2010), as well as to more potentially deleterious ultra- and its light compensation point is around 20 to
violet (UV) radiation (Gao et al. 2012a; UV is consid- 30µmol m−2s−1(Wilson & Collier 1955, Aldrich 1962,
ered further in the ‘Interactive effects’ section below). Eng-Wilmot et al. 1977). Brown tides (Aureococcus
In general, it should not be surprising that in the and Aureoumbra) are another group of HABs that
absence of other limiting factors, HAB growth in - benefit from low light, as both genera commonly
creases with PAR, within physiologically tolerable bloom in severely light-attenuated environments
limits. For instance, light-dependent growth kinetics (Gobler & Sunda 2012). These 2 genera can attain
occur in many species of Alexandrium (Anderson et nearly maximum growth rates under a light intensity
al. 1984, Maranda et al. 1985, Ogata et al. 1987, 1989, of 50 µmol m−2 s−1 at 20°C (MacIntyre et al. 2004).
Parkhill & Cembella 1999, Lim et al. 2006). Laabir et Consistent with their low light adaption, genetic evi-
al. (2011) found a positive relationship between light dence for adaptation to low light was obtained from
intensity and growth rates and biomass of a Mediter- the Aureococcusgenome (Gobler et al. 2011). These
ranean Alexandrium caten ella isolate up to 90 µmol results suggest that both K. brevis and brown tides
photons m−2 s−1; photoinhibition was not observed have an advantage when growing at depth, and also
until a light intensity of 260µmol photons m−2s−1was may have a competitive advantage during dense
reached. Another A. catenella culture showed no self-shaded blooms. This trait, however, means that
sign of photoinhibition up to 800 µmol photons m−2 they may not benefit from future increases in mean
s−1 (Carignan et al. 2002). These results support the light exposures as much as many other non-HAB
suggestion that the genus Alexandrium is adapted taxa.
to high light (Smayda 2008), which could provide it Light is required for production of many algal tox-
with a competitive advantage in future shallower ins, including PSPs, domoic acid, and DSP toxins
mixed layers. (Proctor et al. 1975, Bates et al. 1991, Pan et al. 1996a,
Baek et al. (2008)showed that optimal growth rates Carneiro et al. 2009, Tong et al. 2011). Parkhill &
of the bloom-forming dinoflagellates Ceratium furca Cembella (1999) and Etheridge & Roesler (2005)
and C. fususoccur at irradiances ranging from 216 to revealed that the highest cellular toxin levels in
796 µmol photons m−2 s−1. Like Alexandrium, these Alexandrium tamarense and A. fundyense were ob -
results indicate that Ceratium is well-adapted to in - served at light intensities between 100 and 150 µmol
tense light levels and hence has an advantage in photons m−2 s−1. Analysis of PSP composition in
highly transparent or shallow mixed layers (Baek et Alexandrium demonstrated that toxin composition
al. 2008). did not vary with increasing light (Boyer et al. 1987,
Similarly, increasing light intensity stimulates the Cembella et al. 1987, Ogata et al. 1987, Oshima et al.
growth of the estuarine raphidophytes Heterosigma 1990, Cembella & Destombe 1996, Parkhill & Cem-
akashiwo and Chattonella subsalsa, which exhibit bella 1999, Lim et al. 2006), suggesting that light
maximum growth rates over a light range of 100 to regulated the total toxin concentrations rather than
600 µmol photons m−2 s−1; no sign of photoinhibition changing the toxin profiles. In contrast to these
was observed for either species even at the highest observations of toxin stimulation by increasing light,
light intensity tested, >600 µmol m−2 s−1 (Zhang et an inverse relationship between cellular toxicity and
al. 2006). These results support the suggestion that light-dependent growth was also documented (Ogata
raphid ophycean flagellates generally can tolerate et al. 1987, Hamasaki et al. 2001, Cembella 1998).
and even prefer very high light intensities (Kahn et There appears to be no general trend that applies to
al. 1998). Another recent study demonstrated that the effect of light on the production of PSPs in all
light effects on the growth of H. akashiwo are tem- Alexandrium species and strains, and in general the
perature-dependent (Martinez et al. 2010). They also effect of light variation on toxicity is less remarkable
found differences in growth responses to light be - compared to other factors such as temperature, salin-
tween H. akashiwo strains, suggesting that light ity, and nutrients (Ogata et al. 1987, Lim et al. 2006).
could play a role in intraspecific dominance shifts Thus, the responses of these dinoflagellates to any
and that generalizations for the whole genus may future increases in irradiance doses due to mixed
need to be made cautiously. layer shoaling are difficult to predict.
216 Mar Ecol Prog Ser 470: 207–233, 2012
The responses of toxicity to light are also complex
in HAB raphidophytes. In Heterosigma, there is an
inverse relationship between light-limited growth rates
and toxicity (Ono et al. 2000). Conversely, hemolytic
activity of Fibrocapsa japonica is positively affected
by light (de Boer et al. 2009). Other studies with
Chattonella marinahave related light intensity posi-
tively to its ichthyotoxicity (Ishimatsu et al. 1996,
Marshall et al. 2001). Similar results were also ob -
served by Oda et al. (1997)and Marshall et al. (2001,
2005), who reported that light is also involved in
the production of ROS by raphidophytes, including
F.japonica.
The influence of light on the growth and toxin pro-
duction of dinoflagellates has been more extensively
investigated compared to toxic diatoms. Several stud-
ies have documented physiological responses of
Pseudo-nitzschia spp. to light intensity (Bates et al.
1991, Whyte et al. 1995, Pan et al. 1996a, Fehling et
al. 2005, Thessen et al. 2009), but only one of these
described light as a basic requirement for domoic
acid production (Bates et al. 1991). Domoic acid pro-
duction by Pseudo-nitzschiais inhibited in darkness,
but resumes soon after cultures are shifted into the
light (Bates et al. 1991). Although cultures of P. seri-
ata exposed to a long photoperiod (18 h light: 6 h
dark) compared to a short photoperiod (9 h light: 15 h
dark) have higher growth rates, biomass, and total
domoic acid production, their cellular domoic acid
content is reduced (Fehling et al. 2005).
The majority of published studies showing light-
regulated growth and toxin production in HAB
species have been done with laboratory cultures.
Recently, however, field observations assessing the
environmental factors regulating Pseudo-nitzschia
blooms in the northern Gulf of Mexico have found Fig 2. Pseudo-nitzschia spp. Environmental correlations
that the mean cell toxin quotas and abundance of with abundance of domoic acid-producing diatoms in sam-
Pseudo-nitzschiaspecies were strongly correlated with ples in which domoic acid was detected across an estuarine
salinity gradient in the Gulf of Mexico. Plotted against salin-
several factors, including high irradiance (Fig.2C,D;
ity are (A) total dissolved inorganic carbon (DIC), (B) silicate
MacIntyre et al. 2011).
concentrations, (C) mean down-welling photosynthetically
active radiation (PAR) in the water column, and (D) Pseudo-
nitzschia spp. abundance. ND: counts that were below the
Salinity limit of detection; R-values represent correlation coefficients
for each entire data set (from MacIntyre et al. 2011, used
by permission)
Altered future rainfall and climate patterns could
significantly increase salinity variability in coastal
areas, and especially in estuaries (Hallegraeff 2010, lima, growth rate and toxicity were inversely corre-
Paerl & Scott 2010). Such salinity fluctuations may lated with salinity (Morton et al. 1994). In the dinofla-
favor halotolerant and euryhaline organisms such as gellate Karlodinium veneficum, red uced growth
many HAB dinoflagellates and raphidophytes. For rates due to low salinity significantly enhance cellu-
instance, many species of the dinoflagellate Proro- lar toxin quotas (Adolf et al. 2009). However, the
centrum are euryhaline in culture and in nature diatom Pseudo-nitzschia multiseries demonstrates
(Grzebyk & Berland 1996). In a clonal culture of P. reduced growth rates and cell-normalized toxicity at