The Scottish Government

« Previous | Contents | Next »

Listen

8.0 SILICOFLAGELLATE BLOOMS AND FISH MORTALITY

8.1 Autecology of Silicoflagellates

Silicoflagellates are widely distributed in the ocean, but have been largely ignored because of their low abundance and species number. Silicoflagellates are better known to paleo-oceanographers for their biostratigraphic value than their ecology. Silicoflagellates are considered in this review because of recent reports that their blooms at fish farm sites caused mortality.

In common with diatoms, silicoflagellates require silicon for growth; their delicate siliceous skeleton, on which their taxonomy is based, is their distinctive morphological feature (Figure 10). The considerable variability and intergrades of their skeleton confound taxonomic identification (Boney, 1973), with one valid genus containing three species presently recognized: Dictyocha fibula, D. octonaria and D. speculum (Moestrup and Thomsen, 1990). Of these, D. speculum is best known. Its complex life cycle includes skeletal and naked cell stages, both of which are motile, capable of vegetative growth and harmful, and a non-motile multinucleate stage (Moestrup and Thomsen, 1990; Henriksen et al., 1993). The capacity of D. speculum to alternate between siliceous and naked cell stages, and to bloom in both modes (Table 6), enables it to exploit a wide range of niches. This ability, exhibited also by Phaeocystis spp. [Section 7.3], is unique among HAB taxa.

There is limited autecological data on D. speculum (Table 7). Naked stage cells are easy to maintain in culture, unlike siliceous cells which tend to die at relatively low population densities (Henriksen et al., 1993; Moestrup and Thomsen, 1990; van Valkenburg and Norris, 1970). The growth rates of naked and siliceous cells of D. speculum also differ. Naked cells grown in culture at optimal temperature (11°-15°C) and salinity (15-25 psu) achieved a maximal daily rate of µ = 1.0 (Henriksen et al., 1993); in field populations, µ ranged from 0.4 to 1.1 d ^-1 (Jochem and Babenerd, 1989). In contrast, the growth rate of cultured and natural populations of siliceous cells, µ = 0.5 d ^-1, is 50% lower than naked cell rates (van Valkenburg and Norris, 1970; Fanuko, 1989). Distephanus fibula did not grow in autoclaved media, and appears to be sensitive to turbulence since mechanical agitation of cultures and aeration stopped its growth (van Valkenburg and Norris, 1970).

The geographical distributions of D. fibula and D. speculum partially overlap, with 15°C considered the transitional temperature segregating D. speculum, the boreal species, from D. fibula, the warm water species, (Boney, 1973; Fanuko, 1989). Dictyocha speculum is the more common in European coastal waters, and the species implicated in farmed fish mortality. A five year monitoring program carried out at 31 coastal sites stretching from the English Channel to the Mediterranean Sea detected its frequent presence (Belin et al., 1995). In Scottish waters, autumn blooms of D. speculum have occurred in the Firth of Clyde (Marshall, 1924; Boney, 1973), where D. fibula is an even more minor component of the phytoplankton community. In contiguous waters, D. speculum has bloomed in Liverpool Bay (Voltolina and Oster, 1985) and was found in high abundance in coastal waters of western Ireland (Table 6; Gowen et al., 1990). The sites and magnitude of prominent D. speculum blooms elsewhere in northern European waters are given in Table 6 .

The first report of mariculture mortality attributed to a silicoflagellate bloom came from Danish waters, where farmed salmon ( Salmo salar) and trout ( Salmo gairdneri) died during blooms of the naked cell stage of D. speculum (Aertebjerg and Borum, 1984; Moestrup and Thomsen, 1990; Henriksen et al., 1993). In France and Spain, mortality of farmed salmon and, in Spain, of cultured turbot and octopus also occurred during D. speculum blooms; in those instances, the siliceous stage bloomed (Erard-LeDenn, 1991; Prego, et al., 1998). In the western Shetlands, a bloom of the siliceous stage caused mortality of farmed salmon (Bruno et al., 1989). Elsewhere in Scotland, there is evidence [Section 6.2] that fish farm mortalities in Loch Striven and upper Loch Fyne in 1979 and 1982 accompanied blooms of a naked stage silicoflagellate subsumed in designation of the bloom species as Flagellate X (see also Ayres et al., 1982; Gowen et al., 1982; Gowen, 1987; Tett, 1980). Blooms of D. speculum are not restricted to mariculture sites (Table 6), but mortality of the natural biota has not been reported during blooms at other sites. [Note: Erard-LeDenn and Oster (1985) state Aertebjerg and Borum (1984) [reference not seen] reported blooms of skeleton-bearing cells "damaged" fish in the Kattegat and Belt Sea area.]

The maximal population densities reported for D. speculum are given in Table 6. Three features are evident: the bloom sites of the two life cycle stages appear to differ, both stages attain high bloom populations, and bloom densities of naked cells are considerably higher. Blooms of siliceous cells tend to cluster in open, coastal waters; naked stage blooms predominate in more enclosed, onshore habitats. [The bloom behavior of D. speculum recorded from the Baltic Sea is unknown (Leppänen et al., 1995).] Blooms of naked cells usually attain population densities that reach millions of cells L ^-1, while bloom abundance of siliceous cells is usually an order of magnitude less [but note exceptional population densities recorded in French and Spanish waters (Table 6)]. When blooms of both stages co-occur, the ratio of naked stage to siliceous stage cells can range from about 3:1 to 100:1 based on events in German coastal waters (Jochem, 1989; Jochem and Babenerd, 1989). The blooms can be nearly monospecific (Moestrup and Thomsen, 1990; Jochem and Babenerd, 1989), with a sub-surface population maxima developing at the pycnocline (Jochem and Babenerd, 1989). The occurrence of sub-surface bloom maxima at the halocline is of interest given the prevalence of such features in Scottish sea-lochs (see Tett and Edwards, 2002). This sub-surface micro-habitat, into which nutrients can be advected from deeper layers, provides growth opportunities, access to which would be facilitated by cellular motility and migration. Jochem (1989) reported that naked cells undertake diel migrations, and Fanuko (1989) reported that siliceous cells of psychrophilic D. speculum migrated to below the thermocline, a behavior that she interpreted to be an avoidance reaction to the high temperatures found in the upper water column. However, the vertical depth-keeping attributed to D. speculum, particularly by the naked cell stage, is difficult to reconcile with its reported sluggish swimming velocity of only 20-25 µm s ^-1 (Moestrup and Thomsen, 1990). This rate corresponds to a vertical displacement of only 9 cm hr ^-1.

The finicky behavior of D. fibula and D. speculum reported from culture experiments (van Valkenburg and Norris, 1970; Henriksen et al., 1993) and the low growth rates of siliceous D. speculum present an enigma: the cause of the high bloom abundance achieved in situ (Table 6). This can not be attributed to protection against grazing. During a bloom in Irish coastal waters siliceous D. speculum cells were grazed by copepods, as evidenced by the Distephanus skeletons found in their fecal pellets (Gowen et al., 1999). Salps and echinoderm larvae also graze silicoflagellates (see Fanuko, 1989). There is no information on grazing of the naked cell stage, whose very high bloom population densities are notable, and whose blooms require prior metamorphosis from either the siliceous and/or large multinucleate life cycle stages, or possibly a still unrecorded resting stage (see Moestrup and Thomsen, 1990; Henriksen et al., 1993).

In summary, two distinct types of silicoflagellate blooms occur that are harmful to fish farms: blooms of naked cells - reported from Loch Striven, upper Loch Fyne, and Denmark - and blooms of siliceous stage cells, reported from the Shetlands, France and Spain. Collectively, the life cycle features and field population data suggest that three types of bloom-stimulation events characterize silicoflagellate ecology: bloom stimulation of each of the two types of life cycle stages, and stimulation of metamorphic life cycle transitions into the naked stage to seed its blooms. It is uncertain whether these must be concurrent stimulation or are independent events and presumably under different combinations of regulatory factors.

8.2 Silicoflagellate bloom regulation and ichthyotoxicity

Several lines of evidence suggest that nutrient enrichment from terrestrial runoff and anthropogenic input precedes silicoflagellate bloom events. Distephanus speculum blooms in the Shetlands (Bruno et al., 1989) and Danish coastal waters (Moestrup and Thomsen, 1990) were preceded by a rainfall-runoff event. Initiation of a 200 km bloom of D. speculum in New Zealand coastal waters coincided with the runoff delivery of sediment-rich plumes (Rhodes et al., 1993). The blooms of D. speculum reported for Nordaasvatn, Norway (Table 6) and Omura Bay (Iizuka, 1976) are in eutrophicated habitats. Novel silicoflagellate blooms in Kiel Fjord were attributed to its eutrophication (Jochem, 1989; Jochem and Babenerd, 1989). Elevation of the dissolved inorganic nitrogen pool, particularly NO ₃ concentrations, has also been considered to stimulate silicoflagellate blooms. Winter NO ₃ concentrations in Loch Striven (unpolluted) were twice those of other west coast sea lochs during a salmonid fish kill (see p. 38 in Tett, 1980) linked to a silicoflagellate bloom (= Flagellate X; Section 6.2). Waste nitrogen excreted by caged fish was not implicated in this bloom. In Spain, a silicoflagellate bloom coincided with a copious supply of NO ₃ (36 µM) delivered by the Rio Negro into a fish farm embayment, doubling NO ₃ concentrations above those in local coastal waters (Prego et al. (1999). Regional Si and PO ₄ concentrations were unchanged

The role of Si in D. speculum bloom dynamics has also been considered, given that naked stage cells have no apparent requirement for Si unlike skeletal stages. Riverine delivery of silica into the maricultural area at Merexo, Spain, is elevated because of drainage through a granitic geology. Prego et al. (1999) believed that this high Si input and high N concentrations provoked siliceous cells to bloom, which then became P-limited. For Liverpool Bay, Voltolina and Oster (1985) concluded that the seasonal occurrence of D. speculum was dependent upon Si concentrations, rather than temperature. Not unexpectedly, naked cell abundance in the Kiel Bight region was not correlated with Si concentrations (Jochem and Babenerd, 1989). The latter investigators concluded that the N:Si ratio selected for silicoflagellate blooms over those of diatoms, and that D. speculum becomes dominant after diatoms have become Si-depleted which increases the N:Si ratio in favor of D. speculum.

The association between chemically modified habitats and D. speculum blooms suggests that chemical conditioning of habitat waters is a bloom prerequisite, with nitrogen important in this. There is no evidence, however, that nutrients excreted from fish farms triggered the harmful blooms recorded there, nor that nutrients influence D. speculum toxicity . This differs from the well known stimulation of saxitoxin synthesis in P-limited dinoflagellates. In fact, the evidence suggests that D. speculum does not produce a toxin, and that the harmful mechanisms of naked and siliceous cells differ. Henriksen et al. (1993) tested directly for PST presence in naked cells of D. speculum and for whole cell toxic effects by intraperitoneal injection of these cells into test mice. Both tests were negative, which eliminated the presence of toxins. In experiments that directly exposed test-fish ( Salmo gairdneri) to naked cells, death resulted from hypoxia, an effect that could be reversed through oxygenation of the test chamber. Extrapolating from these results, Henriksen et al. projected that night-time depletion of oxygen caused by the respiration of the naked cells was the cause of mortality at Danish fish farms (Moestrup and Thomsen, 1990). This effect is most likely influenced by the bloom population density, which would help to explain the variable percentage of caged trout dieoffs reported to have occurred. Erard-LeDenn and Oster (1985) also reported the night-time mortality of farmed trout in French waters, the survivors showing behavioural signs of asphyxia. In that event, siliceous stage D. speculum bloomed. In addition to hypoxia induced mortality, histological analyses of dead fish revealed their gill lamellae were pierced, accompanied by edema and clogging by mucus. This histology was also reported during the farmed salmon dieoffs during blooms of siliceous stage D. speculum in Spain (Prego et al., 1998) and in the Shetlands (Bruno et al., 1989) . The Spanish investigators did not believe that respiration of D. speculum leading to hypoxia was the cause of mortality.

Thus, D. speculum can cause fish mortality by two different modes: through hypoxia, and piercing and abrasion of gills. The latter affect is similar to that reported to occur during spinous diatom blooms [Section 9.0]. Given these mortality modes, it seems likely that farmed fish are vulnerable to D. speculum and other silicoflagellate blooms because their caging prevents their avoidance responses, unlike the vagility retained by natural populations. A third inimical effect attributed to D. speculum has been reported by Fanuko (1989). A widespread anoxic event in the northern Adriatic Sea that caused significant benthic mortality was attributed to the decomposition of a D. speculum bloom that became P-limited, terminated and then sank to the bottom sediments.

Table 6. Maximum population abundance (as cells L ^-1) in European waters reported for the siliceous and naked cell stages of the silicoflagellate Dictyocha speculum.

(1). Erard-LeDenn & Rychaert, 1990; (2). Fanuko, 1989; (3). Gowen, 1987; (4). Gowen et al., 1999; (5). Jochem & Babenerd, 1989; (6). Moestrup & Thomsen, 1990; (7). Prego et al., 1998; (8). Voltolina & Oster, 1985.

« Previous | Contents | Next »