The Scottish Government

« Previous | Contents | Next »

Listen

6.0 PHYTOFLAGELLATE SLOOMS IN SCOTTISH COASTAL WATERS

6.1 Introduction

Phytoflagellates, also referred to as µ-flagellates, micro-flagellates and naked flagellates, are primarily nanophytoplanktonic (<10 µm) cells whose fragile cellular structure and integrity make them difficult to investigate. They are usually damaged during sample preservation and by overheating during light microscopy on living material. It is often difficult to distinguish phytoflagellates from detritus. Their taxonomy, also difficult, requires use of transmission and scanning electron microscopy for proper species identification. Numerous phylogenetic groups are represented among the phytoflagellates, including chlorophytes, chrysophytes, cryptomonads, haptophytes, prasinophytes, prymnesiophytes, and the motile life cycle stages of silicoflagellates. Each group, with the possible exception of the cryptomonads, has species that are harmful or a nuisance. Dinoflagellates of similar size and problems of detection and species identification also occur. They are usually treated as separate group among the phytoflagellates, as in this review.

Phytoflagellates in Scottish waters have not received much attention from investigators, nor, excluding enigmatic Flagellate X, have their harmful blooms been reported. This contrasts with Scandinavian waters where novel phytoflagellate blooms the past two decades increasingly have caused mortality of farmed fish and the natural biota. It is therefore of interest to establish whether investigator disinterest and the singular harmful bloom reported for Flagellate X reflects the occurrence of a more or less depauperate phytoflagellate flora in Scottish coastal waters, unlike in Scandinavian waters. If not the case, is there an intrinsic barrier in Scottish coastal waters that represses phytoflagellate blooms? The potential for harmful phytoflagellate blooms impacting aquaculture in Scottish waters is also examined in the following sections.

6.2 Flagellate X - what is it?

An unidentified flagellate designated as Flagellate X bloomed in Loch Striven and Loch Fyne between 1972 and 1982, accompanied by three major kills of farmed salmon (Ayres et al., 1982; Gowen, 1984, 1987; Tett, 1980). Bloom populations ranged from 8 to 34 million cells L ^-1. Flagellate X also bloomed during 1982 at Ullapool (35,000 cells L ^-1), accompanied by a modest salmon kill (Ayres et al., 1982) and, without mortality, in Loch Kanaird and Loch Broom (Gowen, 1987). Ayres et al. characterized Flagellate X as a " west coast phenomenon", occurring from 56° to 58°N, from the Clyde Sea to Ullapool. Gowen (1987) characterized it as " a normal component in the Clyde Sea area", but found throughout the inshore waters of western Scotland. The taxonomic identity of Flagellate X has challenged local investigators who have referred it to different taxa. Droop et al. (1980) noted a morphological similarity to the raphidophyte genus Olisthodiscus (Figure 9). Gowen et al. (1982) and Ayres et al. (1982) considered Flagellate X to be either Olisthodiscus luteus or a Chattonella sp. (Figure 9), another raphidophyte. [ Heterosigma akashiwo is now the accepted name for pelagic cells of an organism that is a look-alike to Olisthodiscus luteus, an epibenthic species, and used here.] Gowen (1987) later applied Tangen's suggestion that Flagellate X might be the naked form of the silicoflagellate Dictyocha (= Distephanus) speculum (Figure 10). All three taxa are ichthyotoxic, with Heterosigma and Chattonella infamous for their lethal blooms at fish farm sites. Hence, the taxonomic identity of the Scottish populations of Flagellate X is of great interest, and is attempted here based on the drawings and descriptions of Flagellate X presented in the reports of its blooms. Archived material was not available for microscopic examination. Irish investigators have also referred to a Flagellate X as the cause of farmed salmon mortality in Ireland (Doyle et al., 1984; Dunne, 1984; Silke and Jackson, 1993).

Drawings of Flagellate X are presented by Gowen et al. (Figures 3a-c in 1982) and Droop et al. (Figure 5 in 1980), with morphologically similar flagellates depicted in Tett (Figure 2 in 1980). Gowen et al. (1982) state that during the Flagellate X bloom uni- and bi-flagellate cells containing numerous chloroplasts were found. That feature is clearly evident in the three depictions presented in their Figure 3, with two of the drawings (3a,c) showing the organism to be uniflagellate. A bi-flagellate cell is not shown; the organism in Figure 3b is without a flagellum. The presence of numerous chloroplasts is a group trait of the ichthyotoxic raphidophyte genera, Chattonella, Fibrocapsa and Heterosigma (Figure 9).

The naked stage of the silicoflagellate D. speculum, also implicated in kills of farmed fish [Section 8.0], has numerous chloroplasts and a prominent anteriorly directed flagellum that overshadows the presence of its second, rudimentary flagellum (Figure 10; Moestrup and Thomsen, 1990). These features are distorted in preserved samples and during examination of live material under light microscopy, making it difficult to capture these features in the drawings rendered. Based on the drawings in Gowen et al. (1982), Figure 3c has features consistent with those described for the naked stage of D. speculum (Figure 10), and is also consistent with Gowen's (1987) identification of Flagellate X as D. speculum.

In contrast, Figure 3a in Gowen et al. (1982) is notable for the numerous peduncular protuberances emergent around the cell perimeter, making its appearance reminiscent of a " naval mine". This habitus is strikingly similar to that of Chattonella verruculosa (see Figure 9-9), the raphidophyte that has recently exhibited novel blooms in Scandinavian waters (Backe-Hansen et al., 2001). A similar morphotype is depicted by Droop et al. (see cell # 3 in Figure 5 in 1980), identified by them as being "possibly Olisthodiscus (= Heterosigma) or Chattonella". Of the three cells included in their Figure 5, the bi-flagellated, potsherd (or potato)-shaped flagellate containing numerous chloroplasts (depicted as #1 in Figure 5) is strongly suggestive of Heterosigma akashiwo. These drawings are consistent with Ayres et al. (1982) decision that Flagellate X is " tentatively identified as Olisthodiscus (= Heterosigma) or Chattonella".

This taxonomic assessment suggests that the blooms of Flagellate X that caused farmed salmon mortality in Loch Striven and upper Loch Fyne in 1979 and 1982 were not monospecific; more likely, three known ichthyotoxic species bloomed concurrently: Chattonella cf. verruculosa, Dictyocha speculum and Heterosigma akashiwo. In addition to their cellular similarities, the timing of the Flagellate X bloom (in May) and the bloom habitats are consistent with the behavior of these three organisms elsewhere. Blooms of naked stage D. speculum in Danish and German (Kiel Bight) coastal waters are often May events (Moestrup and Thomsen, 1990; Jochem, 1989), as are blooms of H. akashiwo (see Smayda, 1998).

Blooms in Loch Kanaird and Loch Broom in June 1982 ascribed to Flagellate X were not accompanied by mortality (Gowen, 1987). Should this not have been because of low population densities, it can not be excluded that phytoflagellate species other than the three suggested to be conspecific with Flagellate X have been subsumed into the latter. Consider the June 1982 bloom (16 million cells L ^-1) in Loch Striven where a farmed salmon dieoff occurred. Observations on preserved and live material indicate that a different flagellate community then prevailed. Tett (1980) records the presence of Chrysochromulina sp., Pseudopedinella sp. and a Pyramimonas sp., with the latter two species established into culture. The Chrysochromulina sp. was identified to genus based on its body shape, coiled haptonema and two flagella. Of these three taxa, Pseudopedinella only has not been reported to be ichthyotoxic, a capacity that varies among species of Chrysochromulina and Pyramimonas (see Section 7.2). Excellent drawings of a Chrysochromulina sp., collected in the vicinity of Loch Striven fish cages, are given in Figure 2 of Tett (p. 107 in 1980), along with Prymnesium parvum, a well known ichthyotoxic species maintained in culture. The source of this culture is not stated, but P. parvum occurs in Scottish coastal waters (Hannah and Boney, 1983).

This reconstruction of the Flagellate X events and the candidate species that may have been subsumed into this appelation suggest that a diverse phytoflagellate flora consisting of at least six genera having species of proven ichthyotoxic capability have bloomed in Scottish sea lochs.

6.3 Ubiquity and abundance of phytoflagellates

Review of the literature indicates that the phytoflagellate flora in Scottish coastal waters is taxonomically diverse, abundant and biogeochemically important, whether expressed as contributions to biomass, to primary production, numerical abundance, or species bloom events. Phytoflagellates are important in the ecology of sounds, firths and sea lochs irrespective of flushing rates and degree of nutrient enrichment. The conclusion of Hannah and Boney (1983) based on their Firth of Clyde investigation would appear to apply generally to Scottish waters: nanophytoplankton (= phytoflagellates) are the stable, background component of the phytoplankton community, upon which blooms of species from within this group and of diatoms and dinoflagellates occur. The evidence for these extrapolations includes the following.

For Loch Etive, where µ-flagellates of chlorophyte, haptophyte, chrysophyte and euglenoid phylogenies are common, Wood et al. (p. 574 in 1973) report that winter standing crop levels " are not particularly low" because of phytoflagellate abundance, and that along with diatoms they are " responsible for high standing crops throughout the summer". Jones and Gowen (1985) calculate the flushing time of Loch Etive to be several weeks. In rapidly flushed (1-2 days) Loch Ardbhair, the site of a fish farm, small flagellates were important during June/July, as in Loch Etive (Gowen et al., 1983). Blooms of Chrysochromulina sp., Pseudopedinella sp. and a Pyramimonas sp. have been reported from Loch Striven (Tett, 1980). During a mesocosm study of the spring bloom in Loch Ewe, small naked flagellates (<5 µm), mostly prymnesiophytes, almost completedly dominated the population initially (Morris et al., 1985). For Easdale Quarry, a sea-flooded slate quarry on Easdale Island in the Firth of Lorn, Tett et al. (1988) found that small phytoflagellates were dominant in May; in July (when cryptomonads also appeared), phytoflagellate abundance reached ca. 10 ⁷ cells L ^-1, with species of Pedinomonas (ca. 10 ⁵ cells L ^-1) and Pyramimonas (ca. 10 ⁶ cells L ^-1) achieving great abundance (see Figure 6 in Tett et al., 1988). Loch Craiglin, a small loch (18 acres) opening into Loch Sween, is exceptional in its shallowness, eutrophication and lack of free flushing (Gowen et al., 1983). Marshall's (1947) fertilization experiment in Loch Craiglin revealed an " enormous [baseline] population of µ-flagellates" that responded rapidly to enrichment with NO ₃ and PO ₄, reaching up to 15 million cells L ^-1. The phytoflagellate species composition was not reported. Marshall and Orr (1948) subsequently described the unique behavior, great abundance and year-round occurrence of a euglenoid, Euglena proxima, at the oxic/anoxic boundary layer in Loch Craiglin. The maximum population abundance reported was 54 million cells L ^-1, with the bloom virtually monospecific. Euglenoid blooms (ca. 10 ⁶ cells L ^-1) have also been reported for Loch Fyne (Tett et al., 1986). Euglenids as a group generally appear to prefer nutrient rich habitats.

In the Sound of Jura, abundant summer phytoflagellate populations developed in stratified regions (Gowen et al., 1983). The investigators speculated that this predominance reflected the ability of some species to assimilate dissolved organic nitrogen ( DON) to achieve such abundance, given the slow diffusion of dissolved inorganic nitrogen ( DIN) fluxed to surface waters during stratification. Elsewhere in related habitats, an unusual spring bloom dominated by phytoflagellates =10 µm attained 1.6 to 2.4 million cells L ^-1 in offshore waters along the western Irish coast (Gowen et al., 1999).

The most detailed study of phytoflagellates in Scottish coastal waters was carried out by Hannah and Boney (1983) in the outer and inner reaches of the Firth of Clyde. During their 2.5-year investigation (1976-1978), the nanophytoplankton, which consisted primarily of phytoflagellates, accounted for 60 to 70% of the total annual community biomass (as chlorophyll) and 50 to 60% of the annual primary production. Analyses of preserved samples and nutrient enrichment experiments revealed a diverse phytoflagellate flora persisted year-around in the Firth of Clyde and, by extrapolation, in western Scottish coastal waters generally. Winter populations were dominated by species of Micromonas, Pyramimonas and cryptomonads, and in summer blooms of Apedinella spinifera, Chrysochromulina spp., and Heterosigma akashiwo occurred. The detection of ichthyotoxic H. akashiwo, considered to be included in the Flagellate X cluster discussed earlier, is particularly notable: it provides direct evidence that it is found in Scottish coastal waters. Table 5 lists the 28 principal flagellate species (excluding dinoflagellates) recorded by Hannah and Boney. Of these, 18 are among the 79 species of phyto- and zoo-flagellates found in Norwegian coastal waters recorded by Throndsen (1969) in enrichment cultures. This apparent regional difference in phytoflagellate diversity probably reflects differences in investigative intensity and techniques, rather than is a real floristic difference.

It is concluded that a common and diverse phytoflagellate community occurs in both Scottish and Scandinavian coastal waters, and that phytoflagellates are not depauperate in Scottish waters accounting for their rarely reported blooms and relative neglect by investigators. Since phytoflagellates are among the most harmful species reported to date, their potential for harmful blooms leading to aquacultural loss in Scottish coastal waters, based on events in European (primarily Scandinavian) coastal waters, is assessed in the following sections.

6.4 Raphidophyte blooms

Raphidophycean flagellates are well known for their allelopathy and ichthyotoxic blooms that pose a serious global threat to fish farming (Figure 11). Raphidophytes have a large number of chloroplasts (Figure 9), relatively high growth rates, are closely regulated by temperature, attain very high population levels, and have at least two fish-killing mechanisms (Imai et al., 1998; Smayda, 1998). There are four raphidophyte genera: Chattonella, Fibrocapsa, Heterosigma and Olisthodiscus, with six species of Chattonella recognized; the three other genera are monospecific (Figure 9; Hallegraeff and Hara, 1995).

Table 5. Principal flagellate species other than dinoflagellates recorded in the Firth of Clyde during 1976-1978 (from Table VII in Hannah and Boney (1983), with the taxonomic updates given in Throndsen (1996).

Raphidophyte occurrences in Scottish coastal waters are poorly known, with detection of Heterosigma in the Firth of Clyde the only known direct observation (Hannah and Boney, 1983). The indirect evidence that Heterosigma and Chattonella occur in Scottish waters was presented in Section 6.2 dealing with Flagellate X. The apparent range expansion and heightened bloom activity of raphidophytes occurring in continental European waters (Peperzak, 2002), and their ichthyotoxic blooms generally at aquacultural sites (Figure 11) prompt this review of their bloom behavior. Heterosigma akashiwo is the focus, with some relevant material on Chattonella spp. and Fibrocapsa japonica provided. The treatment of H. akashiwo is detailed because this species, above all others, is potentially the major threat to fish farming in Scottish coastal waters. While Pseudo-nitzschia blooms and ASP related illness are the major threat to commercial shellfish harvesting in Scottish waters; DSP inspired harvesting closures of cultured shellfish during Dinophysis blooms are anticipated to become a problem as that industry grows, and Alexandrium tamarense blooms undoubtedly will continue to cause regional, aperiodic PSP outbreaks, none of these harmful bloom types impact fish farming in the lethal manner of H. akashiwo blooms. While it shares this potential with K. mikimotoi, mortality of farmed fish during blooms of the latter is not as widespread, nor devastating as those reported for H. akashiwo.

6.4.1 Heterosigma akashiwo

6.4.1a Taxonomic status

The taxonomic status of H. akashiwo has been confused by its close resemblance to Olisthodiscus luteus, first described from a brackish tide pool on the Isle of Wight (Carter, 1937), and now considered a separate species and member of the benthic psammon community. Most pelagic blooms attributed to O. luteus are almost certainly those of H. akashiwo, which has also been referred to as Heterosigma carterae. Throndsen's (1996a) conclusion that the epithet akashiwo has priority over carterae is generally accepted, and is applied here.

6.4.1b Blooms and fish farm mortality

Heterosigma akashiwo is probably the most versatile harmful algal species known: it is antagonistic to organisms ranging in size from bacteria to invertebrates to fish. Its multiple modes of antagonism range from nutritional inadequacy, to suppression of feeding and growth, and to toxicity causing mortality (see Black, 1990; Tomas, 1980a; Tomas and Deason, 1981; Verity and Stoecker, 1982, Wilson, 1981). Its allelopathy against the diatom Skeletonema costatum, against which it competes throughout its distributional range, was demonstrated in Pratt's (1966) classical study. More recently, H. akashiwo has become infamous for its fish killing blooms at fish farm sites. In the Seto Inland Sea, Japan, its harmful blooms over a 16-year period [1972-1987] caused mortality of cultured yellowtail ( Seriola quinqueradiata) and red sea bream ( Pagrus major) carrying a financial loss of ca. 2 billion ¥ (Honjo, 1993). Mortality of farmed Atlantic salmon ( Salmo salar) has occurred during H. akashiwo blooms in Canada (Taylor, 1993; Taylor and Haigh, 1993; Whyte, 1999), Chile (Clement and Lembeye, 1993) and the United States (Horner et al., 1991). A regional, 4-month bloom in British Columbia in 1997 resulted in mortality of pen-reared salmon valued at $20 million, a loss equivalent to the combined financial loss caused by H. akashiwo blooms during the previous decade (Whyte, 1999). In Big Glory Bay, New Zealand, a novel H. akashiwo bloom resulted in a 600 tonne dieoff of farmed chinook salmon ( Oncorhyncus tshawytscha) valued at $ NZ 17 million (Chang et al., 1990, 1993). Blooms of H. akashiwo compromise fish farming in China (Tseng et al., 1993) and Korea (Park et al., 1989), with kills of farmed salmon reported from Tasmania (see Figure 1 in Smayda, 1998). In Spain, a 1989 H. akashiwo bloom led to mortality of cultured sea bass ( Dicentratus labrax) (Fraga, 1988).

Heterosigma akashiwo blooms can also disrupt shellfish aquaculture. Mortality of oyster spat in British Columbia (Whyte et al., 1999), and bivalve mortality in Portugese waters (see Peperzak, 2002) (reported as Heterosigma inlandica) are some examples. Rhodes et al. (1993) reported that H. akashiwo produces an irritant to mussels, resulting in a 'peppery taste' upon ingestion and a tingling sensation to the tongue lasting several minutes. Fisherman working in the bloom area reported skin irritations.

A notable feature of the H. akashiwo blooms at these globally dispersed fish farm sites (Figure 11) is that it was either not reported previously from those regions, or it was locally insignificant prior to its fish-killing blooms. For Chile, Clement and Lembeye comment (p. 223 in 1993): " since the beginnings of fish farming, salmonid mortality during phytoplankton blooms [was] observed", with H. akashiwo appearing soon after initiation of fish farming. In South Korea, H. akashiwo blooms began in 1981 after the initiation of large-scale shellfish aquaculture (Park et al., 1989). In Big Glory Bay, New Zealand, H. akashiwo blooms were first observed in 1989, about five years after fish farming began (Chang et al., 1990, 1993). And, H. akashiwo was not a significant bloom species in northwest Pacific waters (U.S., British Columbia) prior to fish farming. It is notable that all of these reports are from Pacific sites; Atlantic occurrences are seemingly rarer, as will be discussed in a later section.

This common and provocative association between novel H. akashiwo blooms and fish culture suggests there is a causal relationship, i.e. that fish farming, once initiated, will be followed at some point by novel blooms of H. akashiwo at the fish farm site. Since farmed fish are pen-fed, H. akashiwo blooms do not develop because the farmed fish selectively graze on the local food web. [Selective grazing may be a factor at shellfish culture sites where the cultured stocks filter-feed on the natural flora.] The excreted wastes of the penned fish and remineralization of unused food potentially contribute to the apparent stimulation of Heterosigma blooms at fish farms. A nitrogen mass balance budget for Big Glory Bay, New Zealand, revealed that excretion from salmon farms increased nitrogen by 30% above pre-fish farm levels (Chang et al., 1993). In experimental cultures, growth of H. akashiwo was stimulated by several organic growth factors that may be present in fish excreta (Iwasaki, 1979). Decomposing oyster faeces added to experimental cultures also stimulated growth of H. akashiwo (Iwasaki, 1973), as did extracts of bottom sediments collected from fish farm sites in Japan (Uyeno and Nagai, 1973). Section 10 of this report considers, in greater detail, the issue of whether aquaculture generally may stimulate harmful blooms.

6.4.1c Toxic mechanisms

Within minutes of exposure to toxic raphidophytes, fish become stressed, exhibit enhanced activity, constant arching of the dorsal backbone (with protracted exposure), and suffer elevated arterial and cardiac pressure, with death following after 90 minutes of exposure (Endo et al., 1988, 1992; Black, 2000). Older, brood fish appear to be most susceptible (Black, 2000). Penned salmon are vulnerable to raphidophyte blooms because their bloom avoidance options are removed, but mortality of wild stocks can also occur. In Puget Sound, a modest dieoff of wild coho salmon ( Onchorhyncus kisutch), chum salmon ( Onchorhyncus keta), chinook salmon ( Onchorhyncus tschawytscha) and several species of flatfish and sculpin occurred during a H. akashiwo bloom (Hershberger et al., 1997).

Raphidophytes have at least two, and possibly three toxic mechanisms, a remarkable versatility compared to other HAB species. However, the actual toxic principles and mechanisms are still unclear. The copious mucoid layering of gill tissue usually found in fish killed during raphidophyte blooms has been taken as evidence for asphyxiation (Hishida et al., 1997; Ishimatsu et al., 1996). This is problemmatic. Biopsy and behavioral results suggest that mucus buildup is neither the toxicological agent, nor adventitiously derived from H. akashiwo cells; rather, mucus appears to be produced in response to another mechanism (toxin) (Endo et al., 1985). Extracellular, labile reactive oxygen species ( ROS) are suspected ichthyotoxic agents that can alter gill structure and function, and cause asphyxiation (Twiner and Trick, 2000). Raphidophytes secrete significant amounts of ROS, including superoxide, hydroxyl radicals and hydrogen peroxide (H ₂O ₅) (Oda et al., 1992a,b; Twiner and Trick, 2000; Yang et al., 1995). However, recent studies suggest that the amount of H ₂O ₅ produced is 10-fold lower than required to induce mortality of vertebrate and invertebrate cell lines (Twiner and Trick, 2001). Twiner and Trick (2000) have suggested that extracellular ROS are not the ichthyotoxins, but are compounds that have a non-toxic function in serving as a "chemical defense" system against pathogens and/or used to regulate iron availability to Fe-limited cells. They suggest, instead, that H. akashiwo produces four compounds that correspond to the neurotoxin termed brevetoxin, exposure to which caused paralysis and death of juvenile red sea bream ( Pagrus major) (Khan et al., 1997). These brevetoxin-like compounds induce depolarization of the vagal nerve lowering heart rate and blood pressure. This impedes oxygen transfer in the gill lamellae, with the hypoxic condition that is created leading to death (see Tyrrell et al., 2001). It has been speculated that H. akashiwo produces ROS to control brevetoxin auto-toxicity (Twiner and Trick, 2001). Raphidophytes also produce haemolytic and/or haemoagglutinating compounds that affect gill lamellae leading to hypoxia and death (Tyrell et al., 2001). Whatever the specific toxic mechanism(s) of H. akashiwo, the resultant fish mortality rates are also influenced by its population density and habitat temperature (Black et al., 1991).

6.4.2 Heterosigma akashiwo blooms and fish kills in European waters

This section summarizes the known dynamics of H. akashiwo in continental European waters; its behavior in Scottish and Irish coastal waters is considered in Section 6.4.4. Heterosigma akashiwo has been recorded in continental European waters from Norway (60°N) to Portugal ( ca. 40°N), with blooms reported in Oslofjord (Braarud, 1969), Skagerrak (Naustvoll et al., 2002), the coastal waters of The Netherlands (Rademacher et al., 1995, 1997), France (Erard-LeDenn and Crassous, 1998), Portugal (Sampayo, 1989), and in Spanish rías (Wyatt and Reguera, 1989). The distributional and bloom features of H. akashiwo in these waters conform to its global pattern. Three habitat -distinct regional populations (metapopulations) are recognizable, with each habitat characterized by a different primary environmental feature: a physically, chemically or biologically dominated habitat (Smayda, 1998). These ecotypes contrast with " normal" Heterosigma habitats, where significant blooms also occur (Smayda, 1998). " Normal" habitats are defined as relatively pristine sites in terms of nutrient levels, and where upwelling and aquaculture are not factors, e.g., Narragansett Bay (Li and Smayda, 2000) and Sechelt Inlet, British Columbia (Taylor et al., 1994).

Physically dominated systems in European coastal waters where Heterosigma blooms occur are regions of intermittent upwelling: the Spanish Galician rías (Pazos et al., 1995; Tilstone et al., 1994). This metapopulation is the European counterpart of those occurring in the Peruvian (Rojas de Mendiola, 1979) and Chilean (Clement and Lembeye, 1993) upwelling systems. Field and experimental evidence reveals that H. akashiwo has a greater tolerance to turbulence than dinoflagellates (Berdalet and Estrada 1993; Pazos et al., 1995).

Chemically dominated habitats also support significant H. akashiwo populations. Precipitous blooms in nutrient-enriched habitats are often the first recording of its local presence. Blooms in eutrophicated inner Oslofjord (Braarud, 1969), Kastela Bay [Adriatic Sea] (Marasovic and Pucher-Petkovic, 1985) and Cascais Bay, Portugal (Sampayo and Moita, 1984) are the European counterparts of blooms that occur in nutrient-enriched New York Bight, Tokyo Bay, and in sub-regions of the Seto Inland Sea and Hakata Bay (see Smayda, 1998). Exposure to high nutrient conditions favors growth of H. akashiwo since it is relatively inefficient at nutrient uptake. It has high half-saturation constants (K _s) for uptake of inorganic nitrogen and phosphate which range from about 2.0 to 2.5 µM (Smayda, 1997).

Biologically dominated habitats classified as stimulatory to H. akashiwo blooms are aquacultural sites, particularly salmonid fish-farms. Bloom stimulants are speculated to be present in the excreted fish wastes and/or derived during remineralization of undigested food that "chemically condition" the habitat. The presence of H. akashiwo has often been revealed only after initiation of fish farming, as discussed earlier . Blooms of H. akashiwo have caused farmed fish mortality along the southern Norwegian coast (Naustvoll et al., 2002), in Spanish rías and along the Iberian Peninsula (Fraga, 1988; Peperzak, 2002), dieoffs that are the European counterparts of the mortalities reported at Pacific fish farm sites. Fish mortality and H. akashiwo blooms in European waters are further considered in Section 6.4.5.

In summary, the distributional features and metapopulation dynamics of H. akashiwo reveal that it has a very wide niche; it blooms: 1). both in nutrient-enriched and at moderate nutrient levels; 2). in highly stratified and moderately mixed watermasses; 3). in habitats heavily influenced by runoff; 4). in regions of relatively constant high salinity; and 5). in habitats in which natural grazing and associated nutrient turnover have become unbalanced by fish farming and shellfish aquacultural activities. These features are considered in greater detail by Smayda (1998).

6.4.3 Is Heterosigma akashiwo indigenous in European coastal waters?

Heterosigma akashiwo was first recorded in Spanish rías in 1982, where it has been designated a "newcomer" (Wyatt and Reguera, 1989). In France and The Netherlands, its blooms were first recorded in the early 1990s. These relatively recent detections led Nehring (1998b) to classify H. akashiwo as an exotic species in European coastal waters. Elbrächter (1999) dismissed this conclusion, stating that H. akashiwo was previously overlooked because of taxonomic misidentification, a view in agreement with Rademacher et al. (1997) who state: " due to uncertainties in identifying H. akashiwo in preserved samples it is often overlooked ". [This difficulty of identification parallels the taxonomic problem encountered by investigators in Scotland and Ireland in seeking to establish whether Flagellate X was H. akashiwo.] In 1964, an enormous bloom of H. akashiwo occurred in the Oslofjord, attaining 53 million cells L ^-1 (Braarud, 1969). This early report of H. akashiwo in European coastal waters also counters Nehring's bioinvasion theory. Peperzak (Figure 23 in 2002) has mapped the European occurrences of H. akashiwo between 1979 and 1994, which shows that it is distributed in the region extending from Iceland to Portugal.

Despite the historical evidence that H. akashiwo is indigenous in continental European waters, Connell (2000) concluded that it has indeed has spread into these waters. She based this on the remarkable genetic similarity found among 19 strains of H. akashiwo isolated from Atlantic and Pacific habitats, a molecular kinship that is unusual based on genetic screenings of other geographically dispersed HAB species. Connell also contrasted the absence of fish mortality during H. akashiwo blooms in western Atlantic coastal waters, where it is distributed from New England to Florida, with its impact on Pacific fish farms. This contrast and her molecular data led her to conclude that H. akashiwo spread from the Pacific into Atlantic habitats in geological recent times, " possibly by human means". Notwithstanding the molecular similarity, I consider Connell's attempt to characterize the regional toxicity patterns of H. akashiwo as a function of human assisted bioinvasions to be forced.

6.4.4 Heterosigma akashiwo in Scottish and Irish coastal waters

The presence of H. akashiwo in the coastal waters of Ireland and Scotland, where fish farms have largely escaped ichthyotoxic blooms, is enigmatic. In both regions, fish kills in the 1980s were attributed to Flagellate X, which may have been a cluster of species, including raphidophytes, as discussed in Section 6.2. Despite this uncertainty, investigators generally agree that a Heterosigma-like species was among the bloom species during farmed salmon kills in Loch Fyne, Loch Striven, and in western Ireland lochs and harbours (Ayres et al.,1982; Doyle et al., 1984; Droop et al., 1980; Dunne, 1984; Gowen, 1987); Gowen et al., 1982; Silke and Jackson, 1993; Tett, 1980). For both Scottish and Irish coastal waters, Silke and Jackson (p. 1 in 1993) state that Flagellate X " has now been tentatively identified as Heterosigma akashiwo". There is also the direct observation of H. akashiwo blooms in the Clyde Sea (Hannah and Boney (1983). Peperzak (Figure 23 in 2002) mapped H. akashiwo as being present in the Shetland Islands from 1988 to 1990, citing the ICES- IOCHAEDAT (= Harmful Event Database) as the authority for that occurrence. Those sitings require confirmation and more information as to the investigators having identified H. akashiwo, etc. At best, the evidence that H. akashiwo occurs at fish farm sites in Scotland and Ireland, and its attributed bloom-induced mortality of farmed salmon there is circumstantial, but consistent its behaviour elsewhere.

Accepting that H. akashiwo is present in Scottish coastal waters, the issue of interest is: why, unlike at Pacific fish farms, the rare occurrence (if any) of its ichthyotoxic blooms at fish farms in Scotland [and in Ireland and Norway]? One reason may be that the environmental conditions in these waters marginally support H. akashiwo blooms. Another possibility is that these marginal conditions are forcing a longer lag-period between the initiation of fish farming and emergence of H. akashiwo as a local bloom species than occurred at Pacific fish farm sites. Should this be the case, future blooms may be probable once the requisite environmental conditions (= "biological conditioning") and local availability of bloom-seed stock are reached. These prospects are partially assessed in the following summary of H. akashiwo bloom ecology based on a comparative regional analysis of the association between its blooms and fish kills.

6.4.5 Rarity of Heterosigma akashiwo blooms at fish farm sites in European waters: an enigma

Blooms of H. akashiwo and other raphidophytes are the primary HAB threat to salmon aquaculture, and pre-eminent among known ichthyotoxic HAB species. The fish-killing H. akashiwo blooms that occur at Pacific fish farm sites have been described. Atlantic populations of H. akashiwo, in contrast, are rarely implicated in fish farm mortality despite their well documented allelopathy. And when implicated, the evidence for this is circumstantial or superficial rather than definitive, such as the modest mortality of farmed fish in Scotland and Ireland circumstantially associated with H. akashiwo (= Flagellate X) blooms in the 1980s. An 1100 tonne mortality of farmed Atlantic salmon in southern Norway occurred in 2001 during a massive bloom (maximum = 13.5 million cells L ^-1) of H. akashiwo intermixed with Chattonella marina (Naustvoll et al., 2002). It was not established whether H. akashiwo or C. marina, or both, was the toxic species, however. Along the Iberian Peninsula, Peperzak (see Figure 23 in 2002) depicts multiple (n = 5) fish kills as having occurred during H. akashiwo blooms between 1987-1996. He gives no other information about these mortalities. Peperzak also depicts a fish kill in 1994 off the French coast, citing Nézan et al. (1995) as the authorities; their report was not available for my review.

Whatever the real situation, mortality of farmed fish during H. akashiwo bloom events in Scotland and Norway are very rare (unless under-reported) events, if not absent. This rarity is conspicuous since these two nations are major farmed salmon producers globally, and in whose coastal waters H. akashiwo has bloomed. The intensity of fish farming, the ecophysiology and bloom behavior of H. akashiwo, and the Pacific fish farm experience lead to the expectation that the fish farms in Scotland and Norway would be compromised by H. akashiwo blooms. Since this expectation is unrealized, the potential role of temperature and iron ecology as bloom barriers to H. akashiwo at those sites is considered in the following sections. These two variables are thought to be important regulators of H. akashiwo blooms (Smayda, 1998; Twiner and Trick, 2000, 2001).

6.4.5a Influence of temperature

Temperature exerts a major influence on the growth and seasonal bloom cycle of H. akashiwo throughout its distributional range. Its resting stage requires exposure to a minimal temperature during its "over-wintering" quiescent period to complete its life cycle. This minimum appears to be <10°C (Yamochi and Joh, 1986; Itakura et al., 1996). In culture, H. akashiwo tolerates temperatures from <5° to >30°C (Tomas, 1978; Yamochi, 1989), with mortality occuring at 33°C (Yamochi, 1984). Thermal strains occur (Watanabe et al., 1982) and it is capable of very rapid growth, reported to reach µ = 2 to 3 d ^-1 (Honjo, 1993). The growth rate of H. akashiwo is temperature-dependent: for a Narragansett Bay clone, it increased ca. 3-fold between 10° and 20°C (Tomas 1978). Consistent with experimental results, natural populations of H. akashiwo occur over a wide temperature range. At the extremes, it has been found at ca. 3°-4°C in Narragansett Bay and Norwegian coastal waters (Tomas 1980; Throndsen 1969) and at 1° to 3°C during a Skagerrak bloom (Naustvoll et al., 2002) and in Kamchatka embayments (Konovalova, 1995). At the upper extreme, blooms have occurred at 30°C in Hakata Bay (Honjo, 1974).

The temperature-bloom relationship for H. akashiwo is remarkably persistent throughout its distributional range. At most bloom sites, temperature must be minimally 15°C for bloom initiation, with bloom maxima clustering within a narrow range between 15°-20°C. This relatively fixed and narrow "bloom temperature" window has been observed in Oslofjord, the coastal waters of British Columbia, Chile, Peru, Masan Bay (Korea), New Zealand, in Iberian Peninsula upwelling habitats and coastal lagoons, and in Narragansett Bay, U.S. (see Smayda, 1998). For Narragansett Bay, analysis of a 38-yr time series revealed an 86% probability that the initial annual appearance of H. akashiwo will be delayed until water temperatures reach 10°-11°C, with an 80% probability that the bloom maximum will then occur sometime between weeks 23 to 26 (Li and Smayda, 2000). This maximum is achieved soon after water temperatures reach 15° to ca. 20°C. This bloom is followed by a lull at 22°C, with a second maximum sometimes occurring when temperature drops to between 20° and 15°C (Tomas, 1980).

The 15°C "bloom threshold window" is most commonly observed, but two other, less frequent temperature-bloom relationships have been reported: warm and cold water blooms. In some Japanese coastal waters, H. akashiwo blooms begin at >18°C and then peak between ca. 22°-<31°C (Honjo, 1974; Yamochi, 1983, 1989). A similar warm water bloom, at >25°C, has occurred in Kastela Bay, Adriatic Sea (Marasovic and Pucher-Petkovic, 1985). Two cold-water blooms of H. akashiwo have been recorded: one in British Columbia at 9°-10°C (Whyte et al., 1999), and an extraordinary bloom at 1° to 3°C in the Skagerrak (Naustvoll et al., 2002). Not only is H. akashiwo eurythermal, it is capable of significant bloom development at very low, intermediate (mostly), and very high temperatures, a capacity consistent with the experimental demonstration that thermal strains occur (Watanabe et al., 1982).

The annual temperature range in Scottish coastal waters is ca. 5° to 16°-17°C (Figure 4 in Craig, 1959; Milne, 1972). Summer temperatures exceeding this upper range are unusual, with Loch Sween, classified as a fjordic-type sea loch by Milne (1972) clearly anomalous. Milne (p. 35 in 1972) reports a July temperature there that reached 21.6°C. The seasonal temperature in Scottish inshore waters, at its lower end, would accommodate the <10°C that resting stages require for successful dormancy ("overwintering") and germination into motile cells (Yamochi and Joh, 1986; Itakura et al., 1996). "Cold water" blooms, such as reported for British Columbia (Whyte et al., 1999) and the Skagerrak (Naustvoll et al., 2002), also could be accommodated. At the upper end, the 15°C "bloom threshold window" is reached during summer. This suggests that sub-optimal temperatures probably are not the barrier responsible for the fewer than expected blooms of H. akashiwo at fish farm sites in Scotland. This extrapolation should be tempered for several reasons, however. The seasonal increase to the maximal temperature level and its duration are relatively narrow: the maximum temperatures reached are usually only 1° to 2°C above the 15°C threshold, and the duration of the temperature maximum is brief. The resultant exposure of H. akashiwo cells, should they then be present, to the optimal temperature range (15° to 20°C) for growth (Tomas, 1978) is very limited. That is, the period and intensity of favorable growth opportunity are very reduced. The role of temperature in regulating H. akashiwo blooms may also be site-specific. The maximal seasonal temperatures of the estuarine and fjordic sea lochs, following Milne's (1972) classification, differ between and among these two types of sea lochs. Since the hydrography and circulation of sea lochs vary with topographical features and meteorological conditions, and to which maximal summer temperatures are sensitive, the duration of the 15°C "bloom threshold window" and its "width" within the optimum temperature range (15° to 20°C) attained is expected to vary. All things considered, it seems likely that at least a partially limiting temperature effect may be a barrier thwarting H. akashiwo blooms in Scottish waters.

Temperature may be a more important barrier to H. akashiwo blooms in Norwegian coastal waters than in Scotland. The most intense bloom of H. akashiwo recorded to date in European coastal waters (53 million cells L ^-1) occurred in polluted inner Oslofjord (Braarud, 1969) where summer temperatures can reach 24°C (Braarud, 1945). In Throndsen's (1969) survey of the flagellates in Norwegian coastal waters, from Oslofjord (60°N) to Varangerfjord at the Arctic Circle (70°N), H. akashiwo is listed as present only in Oslofjord. Seasonal phytoplankton studies in Trondheimsfjord (ca. 64°N) and Balsfjord (ca. 70°N), where the annual temperature ranged from about 1.5° to 16.5°C (Sakshaug, 1972) and 1° to 7°C, repectively, did not record H. akashiwo. This latitudinal temperature gradient along the Norwegian coast suggests that blooms of H. akashiwo, if present, would have to be primarily of the "cold water" type, as in the 2001 Skagerrak bloom event (Naustvoll et al., 2002. Such blooms are relatively uncommon, as pointed out above.

6.4.5b Role of nutrients

The correlation between nutrient enrichment and H. akashiwo blooms described in Section 6.4.2 has two components: a response to macronutrients (N, P) and to micro-nutrients, including Fe. While there is not much data on the phosphorus metabolism of raphidophytes, a notable feature is their inability to synthesize the enzyme alkaline phosphatase induced by inorganic phosphate limitation (Smayda, 1998). The inability of H. akashiwo to synthesize alkaline phosphate prevents its nutritional use of dissolved organic phosphorus, unlike most HAB species. It is not known whether this influences its occurrence in Scottish coastal waters.

Several investigators have demonstrated that Fe and/or Mn enrichment is required to trigger H. akashiwo blooms and to persist (Honjo, 1974; Takahashi and Fukazawa, 1982; Yamochi 1983, 1989). This requirement is embedded within the apparent dependency of H. akashiwo blooms, in some regions, on river runoff, as reported from Canada, Japan and New Zealand (Honjo, 1974; MacKenzie, 1991; Taylor and Haigh, 1993). The stimulatory effect of runoff in these cases is not induction of watermass stratification, but the delivery of Fe, a riverine supply mechanism considered to be "crucial" to blooms in Osaka Bay (Yamochi, 1989). Similar dependency on Fe has been reported for the ichthyotoxic Chattonella spp., a raphidophyte genus which shares with H. akashiwo a relatively high Fe requirement (Okaichi et al., 1989). Heterosigma akashiwo grew vigorously when provided extracts of bottom sediments in culture experiments, a response cited as evidence that micro-nutrients and complexation ligands are significant bloom-factors (Honjo, 1974). Twiner and Trick (2000, 2001) have linked Fe metabolism in H. akashiwo to secretion of hydrogen peroxide (H ₂O ₅), its release inversely related to Fe availability. In their view, extracellular secretion of H ₂O ₅ provides reducing power which facilitates the reduction of free ferric iron, or organically bound Fe, into a ferrous state available for uptake. The production of H ₂O ₅ that progressively increases with advancing stages in the population growth of H. akashiwo, and presumably stimulated by Fe limitation (Twiner and Trick, 2001), may reflect a strategy in which ROS are released to overcome Fe limitation, in addition to its ichthyotoxic effects.

It is unlikely that Fe is the factor (barrier) limiting H. akashiwo blooms at fish farm sites in Scottish coastal waters given the runoff that occurs into sea lochs (Tett and Edwards, 2002; Rydberg et al., 2002). This is also expected to be the case in Norwegian coastal waters; runoff accompanying snow melt presumably delivers Fe in copious quantities. This subjective conclusion should be tested following the recommendation of Twiner and Trick (p. 1404 in 2001) that " much knowledge could be obtained by measuring ambient ROS and Fe concentrations in situ during a H. akashiwo bloom".

6.4.5c Summary

The analyses provide no definitive explanation for the unexpectedly low incidence of fish-killing blooms of H. akashiwo at farmed fish sites in Scotland, unlike in Pacific localities, despite the apparent presence of H. akashiwo in Scottish waters, its capacity for rapid growth µ = 2 to 3 d ^-1), its eurythermal and euryhaline nature, and that it can bloom in waters ranging from being chemically pristine to eutrophicated. The exudation of phyto-stimulatory substances from farmed fish and the release of growth promoting substances from the chemically enriched sediments underlying the caged fish are also expected to favor blooms, as demonstrated at other localities. Fe availability does not appear to be the limiting parameter, while the role of temperature is ambiguous. There is evidence that the limited levels and duration of temperatures in the optimal growth range from 15° to 20°C characterizing Scottish sea lochs may be a factor. The sub-optimal growth temperatures prevailing along the western coast of Norway would appear to be an even greater deterrent responsible for the failure of H. akashiwo to bloom at Norwegian fish farm fish sites. It is unlikely that temperature acts alone in its influence on H. akashiwo growth in Scottish, Irish and Norwegian waters, but the limited data available prevent evaluation of the combinations of temperature and other growth parameters that may be at play. Nonetheless, the conclusion can not be avoided that Scottish fish farms are highly vulnerable to potential ichthyotoxic blooms of H. akashiwo. Where such blooms have occurred at other fish sites, they have been sporadic and unpredictable rather than an annual feature. The toxic blooms that occurred in the first half of the 1980s in Scotland and Ireland, and attributed to H. akashiwo, are consistent with that experience. It also suggests that such bloom events may be linked to regional features, possibly climatologically driven, in addition to the local habitat conditions modified by fish farm operations.

6.5 Chattonella spp. and Fibrocapsa japonica blooms

The raphidophyte genera Chattonella and Fibrocapsa are also well known for their ichthyotoxicity (Imai et al, 1998; Kahn et al., 1996). The ecophysiology of Chattonella spp. has been reviewed by (Imai et al.,1998). Chattonella antiqua first appeared in the Seto Inland Sea, Japan, in 1964 where it caused severe mass mortalities of cultured yellowtail ( Seriola quinqueradiata) and financial loss of about 19 billion ¥ during 10 major blooms between 1970 and 1987 (Okaichi, 1989). A 1972 bloom in Hiroshima Bay killed 14 million cultured yellowtail (Imai et al., 1998). In South Australia, a Chattonella marina bloom killed 1,700 tonnes of bluefin tuna ( Thunnus maccoyii) kept in holding tanks, and valued at $40 million (Hallegraeff et al., 1998). Unidentified Chattonella and Heterosigma spp. have bloomed in shrimp culture ponds in Thailand, the first record of raphidophyte occurrence in Thai coastal waters (Lirdwitayaprasit et al., 1996). Blooms of Ch. marina in the S. China Sea (Qi et al., 1993), first observed in 1991, were followed by novel appearances of Ch. antiqua, Ch. globosa, Ch. ovata (Lu and Hodgkiss, 2001). Fibrocapsa japonica appeared in 1991 in New Zealand (Rhodes et al., 1993). In Florida waters, Ch. marina, Ch. subsalsa, unidentified Chattonella spp., F. japonica and H. akashiwo have progressively appeared and bloomed since 1986 (Tomas, 1998). The progressive eutrophication of Alexandria Harbor, Egypt, has been accompanied by novel blooms of previously undetected Ch. antiqua and an unidentified Chattonella sp. causing limited wild fish and faunal mortality (Ismael and Halim, 2001; Mikhail, 2001).

European coastal waters are also exhibiting this apparent global increase in novel blooms of the Chattonella-Fibrocapsa-Heterosigma triad. Chattonella blooms were first recorded in 1991: Ch. aff. minima bloomed within the thermal plume from a French nuclear power plant (Erard LeDenn and Crassous, 1998), and Ch. antiqua, Ch. marina and F. japonica appeared in Dutch coastal waters (Vrieling et al., 1995). Chattonella blooms are now frequent and intense, particularly in the Skagerrak region, where three major blooms occurred in the four-year period from 1998-2001. In 1998, a massive spring bloom of Chattonella aff. verruculosa (up to 24 million cells L ^-1) (Figure 9-9) in the Skagerrak and off western Denmark was accompanied by a 350 t dieoff of farmed salmon in southern Norway (Backe-Hansen et al., 2001). Mortality of wild fish stocks also occurred, including herring ( Clupeus harengus), eels ( Ammodytes spp.) and mackerel ( Scomber scombrus). This was the first reported occurrence of Ch. aff. verruculosa occurs in European waters and its ichthyotoxicity. Anthropogenic nutrients delivered by the Jutland Coastal Current flowing from the German and Dutch Wadden seas into the Skagerrak were the suggested bloom trigger, with an unusual weather pattern influencing subsequent dynamics (Aure et al., 2001).

The spring bloom of Ch. aff. verruculosa in 2000, which extended from the German Bight to the Skagerrak, was even more intense: up to 30 million cells L ^-1 and 80 mg m ^-3 of chlorophyll were recorded (Lu and Gobel, 2000; Naustvoll et al., 2002). The investigators do not mention whether mortality occurred. The Chattonella bloom that developed in early March, 2001 in the NE Skagerrak, with a loss of 1,100 t of farmed salmon in southern Norway, differed significantly from previous blooms (Naustvoll et al., 2002). Abundance of the dominant species, Chattonella marina and H. akashiwo, ranged from about 12 to 14 million cells L ^-1, and accompanied by a diverse assemblage of less abundant phytoflagellate species: Apedinella, Chrysochromulina, Pseudopedinella spp., and a species similar to Ch. verruculosa (Naustvoll et al., 2002). This bloom was remarkable for its occurrence at 1° to 3°C and the species diversity and succession of the phytoflagellate community. The extremely low bloom temperature contrasts with the bloom temperature of 15°C "normally" reported during Chattonella and H. akashiwo blooms in Japanese coastal waters (Imai et al., 1998). The causes and regulation of the year 2000 and 2001 blooms are unknown.

Fibrocapsa japonica, described from Japanese waters in 1973 (see Elbrächter, 1999), was first recorded in European coastal waters in 1991, both off France (Billard, 1992) and The Netherlands (Vrieling et al., 1992), and in the German Bight in 1993 (Elbrächter, 1999). Fibrocapsa has since increased in abundance and become established in the North Sea, reaching a population maximum of 2.4 million cells L ^-1 during 1997 in the German Bight (Rademacher et al., 1998). It has not, however, achieved the same impact and bloom importance exhibited by the Chattonella community. Fibrocapsa has been implicated in the death of seals, but the neurotoxic evidence for this is equivocal (Rademacher et al., 1998).

The conspicuous emergence and bloom dynamics of the Chattonella-Fibrocapsa-Heterosigma triad in continental European coastal waters during the past decade has provoked interest as to their origins. Elbrächter's (1999) analysis of the various claims of origin led him to conclude that F. japonica is an introduced species, while the Chattonella spp. and H. akashiwo are indigenous species often misidentified because of their difficult taxonomy. Should this be the case, the Chattonella spp. would then have been components of the hidden flora. This gives rise to a different question: what has stimulated their growth during the past decade making them more evident? There is reason to believe that this query has to be addressed at the raphidophyte group level, rather than at the species level. In European waters and other regions, where new occurrences of Chattonella spp. and F. japonica are increasingly being reported, H. akashiwo is usually present. These three raphidophyte genera appear to co-occur, detection of one of the three genera is symptomatic of the presence of the other two genera. That is, a raphidophyte niche may be opening up globally, particularly in chemically modified habitats and at fish-farming sites, and possibly in "competition" with the dinoflagellate life-form niche. Peperzak (2002) has mapped the regional expansion of raphidophyte blooms and locations of associated fish kills in European coastal waters. Most of the known, major raphidophyte species are now reported to be present in European coastal waters: Ch. antiqua, Ch. marina, Ch. aff. minima, Ch. subsalsa, Ch. aff. verruculosa, F. japonica and H. akashiwo (Figure 9).

The raphidophyte expansion and species assemblages in European coastal waters are relevant to fish farm activities in Scotland. While the presence of Chattonella in Scottish coastal waters has not been directly established, the evidence that Ch. verruculosa (Figure 9-9) may be one of the species in the ichthyotoxic Flagellate X cluster reported from Loch Striven and Loch Fyne has been presented in Section 6.2. And, as discussed earlier, there is more definitive evidence that H. akashiwo occurs in these waters and contiguous areas. Given this and the evidence that there is a raphidophyte niche in which the presence of one member of the Chattonella - Fibrocapsa - Heterosigma triad signals the likelihood that all three ichthyotoxic genera are present, it is suggested that this niche is also present in the sea lochs of Scotland open to raphidophyte exploitation under appropriate growth conditions. It may be relevant that the three subordinate flagellate genera present during the 2001 fish-killing bloom of Ch. marina in the Skagerrak (Naustvoll et al., 2002), the non-toxic Apedinella and Pseudopedinella and the toxic genus Chrysochromulina, have been reported from Scottish coastal waters (Table 5; Hannah and Boney, 1983). This may indicate niche overlap occurs among raphidophytes and other phytoflagellate groups (species). I conclude that fish farms in Scottish coastal waters are probably open to the raphidophyte expansion and their ichthyotoxic blooms being experienced in continental European waters.

« Previous | Contents | Next »