« Previous | Contents | Next »
Listen
10.0 FISH FARMS, SHELLFISH CULTIVATION AND HARMFUL BLOOMS
10.1 Aquacultural wastes: composition and habitat impacts
Fish farming is an intensive aquacultural activity - the fish are fed a processed diet rich in nitrogen, whereas shellfish aquaculture is extensive: it relies on growth of the cultured stocks through feeding on the natural phytoplankton flora and in competition with the natural herbivore populations (Gowen et al., 1990; Handy and Poxton, 1993). Fish and shellfish cultivation are open culture systems: their wastes are discharged into the habitat and enrich the water column and bottom sediments with particulate and dissolved organic and inorganic material. There are multiple sources of enrichment: the excretion of soluble wastes and faecal production by the cultured stocks, remineralization of uneaten food, desquamation, and the release of mucus, vitamins and therapeutic agents (Arzul et al., 1996). The pellets provided as fish feed are composed of two basic types of chemicals: food components and medicinals (Haya et al., 2001). Food components include fish oils, meals, essential minerals, dyes and antioxidants. Medicinals are used as therapeutics, and include pesticides, disinfectants, antibiotics and chemotherapeutants (Haya et al., 2001). This rich mixture of food pellet chemical additives, including a high protein content, is metabolized for growth, the efficiency of which influences the amount and chemical composition of the soluble and particulate wastes. Undigested food pellets leach their chemical additives; food wastage is an important source of the high solid waste production characterizing aquaculture. Copious production of faecal matter occurs: solid waste production by farmed fish exceeds cultured shellfish production by one to two orders of magnitude (Handy and Poxton, 1993). Molluscs sort particles during filter feeding using their labial palps to reject undesirable particles and prey, and void these as mucoid covered pseudofaeces and faeces (Bougrier et al., 1997; Cognie and Barillé, 1999). Even though less than fish farm production rates, cultured shellfish produce significant amounts of solid wastes [see Section 10.5].
Dissolved nutrients and growth factors are excreted along with the production of particulate wastes. Although fish utilize dietary protein very efficiently, large quantities of nitrogenous waste are produced, with ammonia (NH 4+) the main end product in teleost fish (Dosdat et al., 1996). The main excretory organs for soluble nitrogenous excretion in fish are the gills and, secondarily, the skin and faeces (Figure 12; Handy and Poxton, 1993; Dosdat et al., 1993). Fish excrete a variety of inorganic and organic N compounds, with NH 3 excreted as un-ionized NH 3 or ammonium ion (NH 4+). Dosdat et al. (1993) state that NH 4+ represents 75-90% of the total nitrogenous waste, with a significant proportion of the N excreted as urea (5-10%). Both compounds are assimilated by phytoplankton. Un-ionized NH 3 is acutely toxic to fish at concentrations between 0.09 and 3.35 mg L -1 dependent on temperature, salinity and pH (Handy and Poxton, 1993). Gowen et al. (1990) point out that NH 4+ is the main excretory product in marine fish and crustacean farming, dissolving directly into the water. Up to 70% of the N fed to farmed fish may be excreted as NH 4+, and only 1% of the total NH 4+ is released from solids sinking to bottom sediments. Ammonium ions can be toxic to phytoplankton depending on concentration and the species, as discussed in Sections 10.2 and 10.3.
Fish and shellfish cultivation enrich the water column with dissolved inorganic and organic nitrogen (and phosphorus) in multiple ways, to which the phytoplankton respond. This response is not limited to the classical dose - yield relationship, i.e. the amount (= carrying capacity) of biomass produced is determined by the amount of nitrogen available. The aquacultural waste products are a complex chemical mixture expected to have diverse potential impacts on the local flora. Excreted wastes may provide essential nutrients required for growth; have allelochemical and allelopathic power stimulating, or inhibiting local phytoplankton growth; and excreted wastes can also influence species composition through changes in the inorganic N:P ratio and in "water quality". The issue raised is whether the cultivation of finfish and shellfish can cause changes in aquatic ecosystems: in the case of finfish farming, by chemical enrichment of the receiving water and associated changes in water quality; in the case of shellfish cultivation, both through nutrient release and grazing induced changes in phytoplankton community structure (Gowen et al., 1990). Within this, the issue of interest is whether aquacultural excretions can stimulate harmful algal blooms detrimental to local aquaculture, or generally. Several other aspects of aquacultural waste production need to be considered before addressing this harmful bloom issue.
A direct relationship is not to be expected between aquacultural waste discharge and phytoplankton response, whether as species selections or bloom behavior. Aquacultural wastes have other habitat and biotic impacts that influence the ultimate phytoplankton response to the altered nutrient conditions. With respect to impacts on the natural biota, shellfish cultivation affects both pelagic and benthic communities; fish farming primarily affects benthic communities. The effects on benthic communities lie in their biotic compositional changes induced by the manuring of bottom sediments through deposition of solid wastes which alter the chemical and redox properties of the sediments underlying cultivation pens and rafts. Video assessments of the benthic environment around/under penned salmon organically enriched by undigested food particles and faeces reveals changes occur in sediment colour, bacterial mats ( Beggiatoa) and in fauna and flora (Crawford et al., 2001). Lumb's (1989) survey of 57 salmon farms in Scotland revealed acute organic enrichment occurred under the culture cages, usually accompanied by outgassing. Lumb termed this effect self-pollution. Assessment of organic waste delivery from salmon culture in loch Spelve led Brown et al. (1987) to conclude that " waste generated from salmon farming has similar effects on the benthos to other forms of organic enrichment". The most important impacts of this enrichment include development of anoxic sediments, outgassing of methane, H 2S formation, long-term residence of bacteria pathogenic to fish, altered bacterial community, and impoverished infauna or azoic sediments (Ervik et al., 1997; Brown et al., 1987). Gowen et al. (1990) give a daily deposition rate of organic wastes (consisting mostly of carbon and nitrogen) beneath floating fish cages ranging from 17-361 g m -2 d -1, with H 2S production the most important consequence. They report that H 2S "souring" at 1 ppm can stress fish, with levels in outgassing bubbles reaching 17,000 ppm.
Similar manuring under mussel culture rafts changed the sedimentary chemistry and benthic bacterial and faunal communities at a Swedish growth site (Dahlbäck and Gunnarsson, 1981). The sedimentary sinks of fish farm wastes that develop result from the gravitational sinking of particles whose velocity of descent influences the amount of nutrient solubilized and released into the water column during settling to the bottom sediments. In the case of undigested feed pellets, Frid and Mercer (1989) determined that it would take only 2-3 minutes for pelleted food sinking at the estimated rate of 7.2 m minute -1 to fall to the sea bed at a Welsh fish farm site. Beyond the influence of the amount and particle sizes of the wastes produced on the partitioning of excreted nutrient wastes delivered into the water column and to the sea floor, this partitioning is further influenced by tidal flow, mixing and flushing characteristics at the aquacultural site. The speed with which benthic community changes can occur is impressive: the original benthic community beneath long-line mussel culture rafts almost completely disappeared three months after cultivation began (Mattson and Linden, 1983).
How do these benthic habitat changes influence the response of the phytoplankton to excreted aquacultural nutrients? Fish farm sites are often located in relatively shallow waters that exhibit tight benthic-pelagic coupling (see Tett and Edwards, 2002; Rydberg et al., 2002). Evidence for the quick shift in benthic community structure in response to aquacultural waste deposition was presented above. As organic enrichment proceeds, species and functional groups are progressively eliminated resulting in the elimination of all but a few annelid species, as shown for Loch Linnhe (see Pearson, 1975). The induced shift away from molluscan filter feeders reduces grazing pressure on the phytoplankton; i.e. grazing of species previously filtered by shellfish is relaxed. Benthic remineralization and release of nutrients into the overlying water column are an important source of pelagial nutrient. The change in sedimentary redox potential accompanying organic enrichment disrupts the natural flux of nutrients, further loosening benthic-pelagic coupling. Despite the significant changes in benthic properties and biota that occur under fish and shellfish cultivation rigs, the effects are confined to relatively small areas of the sea bed directly beneath and adjacent to the fish cages and shellfish rafts (Brown et al., 1987; Gowen and Bradbury, 1987). Gowen and Bradbury (1987) state that the proportion of the sea bed receiving fish waste is probably <1% in the Scottish lochs used for fish farming. This feature diminishes the scope of any bloom stimulation by aquacultural nutrients. However, the chemical spatial heterogeneity of elevated nutrient at the culture sites is a potential stimulant of intense, local blooms that can provide seed populations delivered through advective spreading leading to blooms elswhere. Notwithstanding the latter, aquacultural wastes modify benthic grazing impacts on phytoplankton whether stimulated, or not, by aquacultural waste nutrients.
There is also competition for excreted N among bacteria, phytoplankton and macroalgae. Not all of the excreted N becomes available for phytoplankton assimilation; the microbial and macroalgal communities also assimilate nitrogen. And of the particulate N deposited onto sediments, the amount that is remineralized and fluxes into the water column, where it becomes available for assimilation by phytoplankton and macroalgae, is a function of three major bacterial processes that take place in the sediments (Gilbert et al., 1997; Hargreaves et al., 1998). Some of the NH 4+ released will be oxidized to NO 3- during nitrification, a two-step process mediated predominantly by two bacterial genera carrying out the following reactions:
NH 4+ + 1.5 O 2 ? NO 2- + 2H + + H 2O
mediated by Nitrosomonas, and the oxidation of nitrite to nitrate mediated by Nitrobacter:
NO 2- + 0.5 O 2 ? NO 3-
Following its production by nitrification, NO 3- may follow several biochemical pathways, including assimilation by phytoplankton and macroalgae. Denitrification also occurs:
NO 3- ? NO 2- ? NO ? N 2O ? N 2
At least 14 genera of bacteria are capable of denitrification. From 15 to 70% of the organic N deposited onto sediments can be lost by denitrification (Gilbert et al., 1997). Kaspar (1985) showed that denitrification rates of sediments at a mussel farm were 7-fold greater than at a control site. Denitrification is an important N purging mechanism that removes aquacultural waste N which, if allowed to build up in excess of the assimilatory capacity of the aquacultural site, could lead to deterioration of water quality through the accumulation of NH 4+ and NO 2- which have` toxic effects on the farmed fish and phytoplankton (Hargreaves, 1998).
Nitrate can also be reduced to ammonia through dissimilatory ammonium production:
NO 3- ? NH 4+
The type of aquaculture carried out influences the nitrogen cycle. For example, cultivation of Crassostrea gigas at a French site lowered the nitrification rate and stimulated the dissimilatory reduction of NO 3- to NH 4+ (Gilbert et al., 1997).
Thus, excreted aquacultural nitrogenous wastes are subject to a series of biochemical processes and physical chemistry reactions associated with the N cycle involving several trophic levels, the water column and bottom sediments. The biochemical processes and chemical reactions transform, sequester and purge waste N, with the ultimate result that a variable but decreased amount of N from the loading level becomes available for phytoplankton use. This complexity precludes reaching a general, definitive conclusion in response to the question of whether the open system features of fish and shellfish cultivation can (and do) enrich the growth habitats with nutrients that stimulate microalgal blooms detrimental to cultured stocks and the natural biota. In fact, there is an ongoing debate over whether mariculture has resulted in increased HAB events or stimulates the toxicity of potentially toxic species (Rhodes et al., 2001). Gowen et al. (1990), who have concluded that there is little evidence that mariculture has caused algal blooms, rightly point out that the waste nutrient-phytoplankton bloom issue relates to whether HAB species respond to hypereutrophicated conditions and whether those conditions occur at aquacultural sites. Gowen and Bradbury (1987) concluded that " in general widespread hypereutrophication and eutrophication are unlikely" to result from salmonid fish farms. If there is an effect, its spatial scale is expected to be localized and restricted to the immediate vicinity of fish farms. Tett and Edwards (2001) provide a thoughtful review of nutrient-phytoplankton dynamics in considering the potential impact of aquacultural wastes on blooms. Their comparison between Loch Creran and Loch Striven suggested to them that increased nutrient inputs increase the size of algal blooms and the " probability of harmful effects from mildly toxic algae".
The evidence whether increased harmful bloom events have accompanied the surge in fish farming in Scottish coastal waters, and the potential for such blooms are evaluated in the following sections. This is followed by a similar analysis of the observed and potential effects of shellfish cultivation on blooms in Scottish coastal waters [Secton 10.5].
10.2 Fish farms, nutrient wastes and blooms in Scotland
Fish farm production in Scotland is dominated by Atlantic salmon ( Salmo salar) reared in salt-water cages, with a small production of rainbow trout ( Onchorhynchus mykiss). In 2000, salmon production was 129,000 t and rainbow trout production 800 t (see Rydberg et al., 2002). The growth of fish farming has been extraordinary, having increased from an annual production of 598 t in 1980 (Tett and Edwards, 2001) to 159,000 t in 2001 (Rydberg et al., 2002). The 265-fold increase in production over two decades was punctuated by surges in production over five year intervals, ca. a 12-fold increase occurred between 1985/1990; a 5-fold increase between 1990/1995, and a 1.7-fold increase between 1995/2000 (see Table on p. 13 in Tett and Edwards, 2002 and Table 6 in Rydberg et al., 2002).
The significant quantities of nutrient wastes discharged at fish farms become even more evident when compared to muncipal loadings. For example, 8,700 t of the 10,000 t of salmon produced in Scotland in 1986 were farmed on the West and North coasts and the Hebridean Islands (= WNH region). The amount of nutrients discharged from fish farms in the WNH region that year was estimated to have exceeded that in the waste from the human population in that region (Tett and Kennedy, 2002). Post-1985, the WNH salmon production increased to 25,000 t in 1990 and 81,000 t in 2000. All else equal, the regional WNH nutrient loading from fish farms in the 15-year period from 1985-2000 increased from parity with domestic waste delivery to exceeding the latter by 9-fold. [The year 1985 will be used as a branch point in which fish farm wastes discharged into Scottish waters began to shift from acute to chronic delivery.] Elsewhere, the amount of organic enrichment released at fish farm sites in eastern Canada equalled the untreated municipal sewage of a city of 300,000 (Smith et al., 2001). Holby and Hall (1991) calculated that the phosphorus excreted at a fish farm in Gullmarfjord, Sweden, with an annual production of 50 t of rainbow trout ( Onchorhynchus mykiss), would correspond to the discharge from a sewage treatment plant in a town of 7,000 inhabitants, assuming 90% removal of phosphorus. These comparisons put into perspective the high fertilization potential of fish farm wastes delivered into coastal waters. Rydberg et al. (Table 11 in 2002) estimated the total annual nitrogen and phosphorus loadings into Scottish waters from fish farming in 2001 were 7,900 and 1,580 t, respectively, which compare with rates of 345 and 69 t, respectively, for 1985. [The 1985 amounts were calculated based on the reported production of 6,921 t that year (Tett and Edwards (2002) and excretion rates of 50 kg N t -1 and 10 kg P t -1 fish (Rydberg et al., 2002)].
When considered regionally, the 23-fold increase in annual N and P loadings from fish farms since 1985 has not been accompanied by an increase in phytoplankton blooms, either from benign, or harmful species, based on the available regional data described in Sections 2 to 9 of this report. Using 1985 as the branch point when fish farming accelerated, the regional bloom patterns and occurrences characterizing the pre- and post-1985 bloom events in Scottish coastal waters are not distinctive. Similarly, the patterns and trends in harmful and benign species of Alexandrium, Dinophysis, Pseudo-nitzschia, harmful phytoflagellates, diatoms and ichthyotoxic dinoflagellates do not show detectable associations with increasing delivery of fish farm nutrients. Tett and Edwards (2002) and Rydberg et al. (2002) came to a similar conclusion applying different, but related analytical approaches in analyzing the relationship between fish farm intensity and phytoplankton blooms.
There are significant limitations to the regional approach used to evaluate the relationship between blooms and fish farm nutrient wastes, which require comment. The regional approach approximates a mass balance exercise that masks the within region- and local bloom-nutrient waste relationships. An even greater limitation is the lack of adequate time series to allow detection of any long-term patterns and trends that might have occurred, and symptomatic of a positive relationship between bloom and fish farm nutrients. There is also a presumptive bias: the assumption that the fish farm habitats will degrade with fish farm nutrient discharge. Ross et al. (1993) modelled nutrient enrichment effects in Loch Creran assuming a hypothetical fish farm (500,000 salmon of 2 kg average weight). They concluded that because of low incident irradiance and the consequent early onset of self-shading, phytoplankton in west coast lochs tend to be limited by light and not nutrients. They extrapolated this modelled result to reach the general conclusion that nutrient (= N) enrichment alone has little or no effect in Scottish lochs. They also evaluated various combinations of irradiance, nutrient enrichment and grazing pressure and concluded that in very small lochs of relatively high irradiance or low attenuation nutrient enrichment may indeed stimulate blooms. In larger systems, they concluded, the effects of nutrients are linked to grazer structure, and if nutrient enrichment is accompanied by a large concomitant reduction in zooplankton grazing, this may stimulate blooms leading to anoxia. The value of the modelled projections of Ross and co-workers lies in their emphasis that nutrient assimilation and the consequences of nutrient enrichment are under multifactorial control [in their instance by irradiance and grazing], and the impact of this interaction is site-specific. This leads to the question of whether there is indeed evidence that there is a correlation between fish farm nutrient release and increased bloom events at specific growth sites.
The amount of N consumed and excreted by farmed fish is a function of their size, species and food composition. Figure 12 is a general schematic of the routes and rates of nitrogen excretion, wastage and the efficiency of N assimilation and retention of farmed fish under conditions of high food wastage and poor N assimilation and retention (i.e. high excretion). In this "worse case scenario" developed by Handy and Poxton (1993), the fish are fed a ration of 1-4% of body weight N d -1 (= 1-4 g N kg -1 fish, equal to 365-1,460 kg N t -1 production). Of this, 40.5% of the food provided is wasted and deposited onto the sea floor; 59.5% is ingested. And, of the N added as food, about 95% is wasted to the environment as excreta or unused nitrogen. Gowen and Bradbury (1987), in their review of salmonid fish farming, calculated that 78% of the N consumed is lost as faecal and soluble N, of which NH4 and urea accounted for 68 to 86%. This corresponded to 32 kg NH4 t -1 of food provided, with the amount of excreted NH 4 t -1 of fish produced estimated to range from 45 to 55.5 kg. The amount of waste production reported by investigators is quite variable, but tends to converge towards 50 kg N and 10 kg P t -1 fish produced (see Rydberg et al., 2002). Davies (2002) arrived at even lower N waste production rate: 34-45 kg N t -1 fish based on a mass balance model for 1997-1999 using data from Scottish fish farms. These more recent estimates are much lower than the rate of 75-120 kg N t -1 fish used by GESAMP (1996).
The progressive reduction in waste N loading that has occurred partly reflects the development of high energy food pellets that allow more efficient utilization of N. The ratio (by weight) of feed provided to fish produced ( FCR) has progresively improved from 2.0 to 1.3 (Davies, 2002). While the reduction in N loading t -1 fish produced helps to diminish the threat of phytoplankton blooms, this potential impact is influenced by the amount of farmed fish produced per unit area (kg m -2), and which I term fish farm intensity ( FFI). The FFI values for 30 lochs and voes ranged from 3.25 to 0.05 kg m -2 (see Tables 8a and 8b in Rydberg et al., 2002). The areas with the highest biomass produced per unit area (>2.0 kg m -2) tended to be small voes and embayments <2 km 2 in area, and where the yield of farmed fish ranged from 266 to 6,610 t. There are lochs which sustain a higher yield, such as Loch Fyne and Loch Sunart with an annual fish production of ca. 9,000 t and 6,000 t, respectively, but whose FFI is very low, 0.05 and 0.11 kg m -2, respectively (Rydberg et al., 2002). Based on first principles, it would be expected that blooms occurring in response to fish farm nutrient wastes would be much more likely to occur in lochs and voes having very high FFI values than in low FFI systems. This can not be examined rigorously because of the lack of suitable time series at representative fish farm and control sites during pre- and post-fish farming initiation. However, the limited evidence available summarized in Sections 2 to 9 and the review of Tett and Edwards (2001, 2002) do not suggest that fish farming has stimulated increased, or novel occurrences of harmful blooms, either within the region or when new sites have been opened up to fish farming.
Ideally, analyses seeking to nullify (or validate) the waste nutrient - bloom stimulation hypothesis should focus on the period within the annual cycle when harmful blooms are likely to develop. For Scottish waters, this would appear to be the summer period for two reasons. During this post-spring bloom period, nutrient levels, particularly N, are diminished and limit the growth of phytoplankton. Loch Creran is an example of this condition (Solórzano and Ehrlich, 1979). Extrapolating the nutrient limitation theory classically applied to marine coastal waters, the voes, lochs and open coastal waters of Scotland are N sensitive systems into which addition of N is expected to stimulate phytoplankton growth. Fish farms are a source of this nitrogen. Davies (2000) showed that the maximal excretion at fish farms occurs during the second summer (June-August) of the salmon grow-out cycle. Inspection of Figure 3 in his report shows that the amount of dissolved N excreted monthly during June-August ranges from 3,000 to 3,500 t per 1,000 t fish production, considerably above pro-rated monthly rates during the remainder of the year. This coincidence between high N loadings and the increased N sensitivity of the N-limited phytoplankton during the summer months, a period of elevated temperatures, is a well known recipe for bloom formation (Smayda, 2000). The lack of a long-term time series for nutrients and phytoplankton at representative fish farm and control sites precludes rigorous evaluation of whether summer (seasonal) phytoplankton blooms have increased along with the development and expansion of fish farming in Scotland. The limited data available suggest that this has not been the case. A caveat to this impression: in areas of sustained, high yield fish farming, or at sites of both low and high FFI enrichment, stimulation of the indigenous flora by fish farm N excretion without development of novel and harmful blooms can not be excluded. This favorable and beneficial (to the local food web) response might continue up to a threshold level, above which the changing water quality and species composition might then select for harmful species blooms. That is, the relationship between fish farm waste nutrients and blooms develops as a process, over time, rather than transitions suddenly, similar to a square wave, into a condition where such blooms become common occurrences at fish farm sites.
The apparent absence of a stimulating effect of fish farming on harmful blooms in Scottish coastal waters should not be interpreted to indicate that salmonid farm wastes lack nutritional value and allelochemical and allelopathic properties to influence blooms. The following section documents the growth promoting properties of fish farm wastes.
10.3 Fish farm wastes and phytoplankton growth
The effects of aquaculture wastes on phytoplankton growth and blooms are complex. The responses vary among receptor species and with the chemical structure and animal source of the exudates. The effect is not uniform: some inorganic and organic compounds in the excretion bolus may be stimulatory, other compounds may suppress, or even inhibit phytoplankton growth. And, the responses of the phytoplankton species within the community can also differ: some species may be stimulated, some repressed, with others indifferent to the excretions. This section documents such effects reported for fish farm exudates; the effects of cultured shellfish wastes are considered in Section 10.5.
Aquacultural wastes have three primary modes of impact on phytoplankton: as nutrients, as growth factors, and as "chemical conditioners" influencing water quality. The copious excretion of NH 4+ during fish farming (Figure 12) supplies the major inorganic N source assimilated by phytoplankton. NH 4+ is well known to be preferentially utilized over NO 3- whose uptake is repressed at NH 4+ concentrations >1-2 µM. Most of the post-spring bloom N in support of phytoplankton growth comes from NH 4+ which has a higher turnover rate via animal excretions than NO 3-, which is termed "new nitrogen" and is usually supplied through physical oceanographic processes which flux NO 3- from its deep water reservoir into the euphotic zone. Both NH 4+ and NO 3- are also delivered through riverine discharge. But NH 4+ has dual consequences: while it is the major dissolved inorganic N source for growth, at high concentrations it becomes inhibitory. This has been shown for the dinoflagellates Lingulodinium polyedrum and Akashiwo sanguineum (Thomas et al., 1980). Inhibition by NH 4+ can be heightened by NO 3- presence, as shown for domoic acid producing Pseudo-nitzschia multiseries (Hillebrand and Sommer, 1996). The influence of elevated NH 4+ concentrations inhibitory to Pseudo-nitzschia spp., and the role of fish farms as bloom triggers of this toxic genus were considered in Section 2.11. Notwithstanding the inhibitory effects of NH 4+, given the dilution gradient at fish farm sites its excretion is probably more beneficial than harmful to phytoplankton growth, subject to the FFI conditions discussed in Section 10.2.
Urea is excreted at fish farms and is an important source of organic N (Figure 12) assimilated by phytoplankton (Antia et al., 1990), although urea can also inhibit growth (Arzul et al., 1996). Urea added to sea water at low concentrations (1 to 2.5 µM) inhibited the growth of toxic Alexandrium minutum, but at <1 µM stimulated growth of K. mikimotoi (Arzul et al., 1996) . Dinoflagellate blooms commonly occurred in hybrid striped bass ( Morone saxatilis) aquacultural ponds when urea concentrations were in excess of 1.5 µM N (Glibert and Terlizzi, 1999). Summer bloom species included Akashiwo sanguineum, Karloodinium galatheanum and Prorocentrum minimum, and winter blooms of a Katodinium sp. When urea levels were <1.5 µM, dinoflagellate blooms were not observed, and diatoms then dominated.
There is increasing evidence that the capacity of phytoplankton to assimilate amino acids is greater than expected (see Antia et al., 1991; Flynn, 1990; Bonin and Maestrini, 1981). While the concentrations of amino acids excreted at fish farms are unknown, the spiking of food pellets with amino acids makes these nitrogenous compounds available for phytoplankton growth following their release from undigested pellets and through direct excretion. [Handy and Poxton (1993) state that Atlantic cod prefer diets containing glycine or alanine.] Admiraal et al. (1986) report that diatoms take up and excrete amino acids, while Flynn (1990) showed that this ability varies among phytoplankton species. Flynn recognized three distinct physiological groups: species which assimilate a wide range of amino acids, but secrete little; those with limited uptake ability, but are copious secretors, and those with limited uptake and release capacity. The work of Granéli et al. (1995) on DSP species showed that dark uptake of aspartic acid enhanced the growth of Dinophysis acuta. While the levels and impact on phytoplankton of dissolved organic nitrogen wastes from fish farms are poorly known, the limited data available suggest that their excretion would lead to higher growth.
Bioassays of fish farm conditioned water and elutriates of food pellets and faecal wastes provide more direct evidence that undigested food and wastes of farmed fish can impact phytoplankton growth. Arzul et al. (1996) found that elutriates of food pellets supplied to farmed turbot, Psetta maxima, markedly stimulated growth of ichthyotoxic K. mikimotoi (Section 5.1) and, to a lesser extent, the PSP-species A. minutum (Section 4.2.2). Alexandrium minutum was strongly inhibited when grown in water conditioned by turbot growth and diluted to 10% with coastal water; it was also inhibited when exposed to elutriates of diluted faeces. These treatments were without effect on K. mikimotoi, while the spinous diatom Chaetoceros gracile, closely related to chaetocerids shown to be harmful to farmed fish [Section 9.0], was stimulated. Arzul et al. (1996) also tested the responses to the antibiotic, oxytetracycline, to evaluate the potential effects of antibiotics contained in fish feed. The responses of the three test-species differed: growth of the diatom C. gracile was depressed, but stimulated in A. minutum, and either stimulated or inhibited in K. mikimotoi depending on the concentration of oxytetracycline. Research on the impact on phytoplankton growth by medicinals (pesticides) used to combat sea-lice infections and disinfectants applied to prevent the spread of viral infections, such as salmon anaemia (Krossøy et al., 2001), appears to be rare. In a related study, Haya et al. (2001) tested the effect of various pesticides, anti-parasitic drugs and medicated feed used to protect against infestation by the parasitic copepds Lepeophtheirus salmonis and Caligus elongatus. They found that the chemicals used in the treatment of sea-lice infestations were lethal to lobster and sand shrimp ( Crangon septemspinosa). Given these results, similar studies using phytoplankton test species are desirable and are recommended.
Another study (Arzul et al., 2001) examined the response of the harmful species Alexandrium catenella ( PSP), Alexandrium minutum ( PSP), Heterosigma akashiwo (ichthyotoxic), K. mikimotoi (ichthyotoxic) and the diatom Chaetoceros gracile to dilutions of excreta from farmed Atlantic salmon and sea bass, Dicenthrarchus labrax, compared to that produced by the bivalves Mytilus chilensis and Crassostrea gigas [see also Section 10.5]. Notable differences in response were found among species. The PSP producers A. catenella and A. minutum were unaffected by Atlantic salmon and sea bass excretions, including NH 4+ and urea exudation. In contrast, K. mikimotoi was inhibited by sea bass excretion, being more sensitive to the repressive effects of this excretion than to stimulation by NH 4+ or urea. Heterosigma akashiwo was inhibited by Atlantic salmon and sea bass excretions, an unexpected response given the evidence that its blooms are stimulated at fish farm sites [Section 6.4.1b].
In the Seto Inland Sea, Japan, sea water samples collected from a yellowtail ( Seriola quinqueradiata) fish farm supported good growth of K. mikimotoi which was also stimulated when exposed to extracts of mackerel flesh and yellowtail faeces (Nishimura, 1982). The ichthyotoxic raphidophyte species Chattonella antiqua was not stimulated in these bioassays; its growth improved only when provided chelated Fe. Nishimura concluded that dissolved organic matter released from fish farms influences blooming of K. mikimotoi. Turner et al. (1987) reported that high concentrations of the vitamin biotin during natural blooms of K. mikimotoi, and in experiments, induced formation of gill lesions in farmed Atlantic salmon. It is unknown whether biotin stimulated the growth of K. mikimotoi leading to the observed histopathology.
The addition of salmon food pellets and faeces to microcosms increased bacterial abundance and led to a bloom of the heterotrophic dinoflagellate Oxyrrhis marina (Parsons et al., 1990). In evaluating the potential influence of fish farm [and shellfish] wastes on phytoplankton, the associated involvement (stimulation) of bacteria can not be ignored. It is now well established that algicidic bacteria are active against a wide range of phytoplankton species which they kill either through direct attack or by their toxins (see Imai, 1997). The complex relationship between waste nutrients - phytoplankton - bacteria - animal mortality is revealed by the study of Connell et al. (1997). Larval Crassostrea gigas died within 48 hrs of exposure to Heterosigma akashiwo, suggestive of an allelopathy against this oyster. However, addition of the antibiotic streptomycin sulfate to the medium prevented mortality, which suggested that bacterial infection and not a phycotoxin was the actual cause of larval oyster death. This was confirmed when a Vibrio sp. was isolated into culture and its growth enhanced by the addition of medium conditioned by H. akashiwo. It may be difficult to distinguish between fish farm mortality due to harmful phytoplankton and bacterial infection. Bruno et al. (1998) have shown that bacterial infections of Vibrio spp. can cause significant mortality among farmed fish during a condition sometimes referred to as "winter ulcers". Navarro (2000) reported that higher concentrations of NH 4+ and organic phosphorus, and higher abundances of bacteria and their protozoan grazers occurred near fish farms sites in Loch Fyne. He concluded that nutrient wastes from fish farms enhances growth of the bacterial and microbial loop components that feed on bacteria. No information on phytoplankton species or blooms was given.
The tight benthic-pelagic coupling that usually characterizes fish farm sites, and the significant enrichment of the underlying sediments by wastes were pointed out earlier. Several investigators have reported that elutriates or extracts of these manured sediments can stimulate flagellate growth, including H. akashiwo, K. mikimotoi and Prorocentrum minimum (Hirayama et al., 1972; Irie, 1973; Kondo et al., 1990). Some investigators have used field evidence to conclude that fish farm wastes stimulate the growth of selected phytoplankton species, such as toxic Prymnesium parvum (Kaartvedt et al., 1991) and the important spring bloom diatom, Thalassiosira nordenskioeldii (Smith et al., 2001). The use of field data in such efforts has many pitfalls, since it requires a time series data set and an analytical approach which applies ecological and oceanographic principles and processes, but usually not available, nor practised. The controversy over the 1988 bloom stimulation of toxic C. polylepis in southern Scandinavian waters considered in Section 7.1.2 is relevant to the misconnects that can result from analysis of incomplete field data and reliance on statistical correlations in seeking to detect and account for changes in HAB events.
Compounds, termed polyamines, associated with the death and decomposition of organisms have begun to attract research attention. Although their actual sources and dynamics are obscure, polyamines have some special features relevant to the fish farm wastes - phytoplankton growth issue, and are evaluated in the following section.
10.4 Fish farm mortality, polyamines and blooms
The excretion of fish farm wastes builds up a chemically diverse, dissolved organic nitrogen pool. The chemical composition of this pool is diverse because of the different catabolic pathways active in the excretion of waste products. Excretion of an array of biologically active compounds is expected to occur, including purines, pyrimidines, polyamines and vitamins, many of which can be assimilated by the phytoplankton. Purines, for example, form the organic base found in nucleotides and nucleic acids. Two major purines - adenine and guanine - are found in DNA and RNA, while hypoxanthine and xanthine are important in the synthesis and degradation of nucleic acids. Consider the excretion of urea, whose assimilation by phytoplankton was dealt with in Section 10.3. Urea formation is near the end point of the purine catabolism pathway:
adenine, guanine ? hypoxanthine ? xanthine ? uric acid ? allantoin ?
allantoic acid ? urea ? NH 4+ + CO 2
Experiments have shown that all of the purines and their break down products serve as good, sole N sources for phytoplankton whose uptake, as shown for guanine, increases with N-deprivation (Allison and Syrett, 1987; Antia et al., 1975; Iwasaki, 1973; Oliveira and Huynh, 1990; Shah 1984a,b; Shah and Syrett, 1982; Ueno et al., 1977). However, the capacity to assimilate purines varies among species, and both stimulation and inhibition of growth has been reported in response (Oliveira and Huynh, 1990). The ichthyotoxic flagellates Heterosigma akashiwo and Prymnesium parvum, whose blooms at salmonid fish farm have been discussed in Sections 6.4.1.b and 7.2, differ in their purine uptake capacity (Oliveira and Huynh, 1990). Prymnesium parvum can use both hypoxanthine and xanthine as a sole N source, but not allantoic acid. In contrast, H. akashiwo assimilates none of these compounds and takes up urea only if nickel (Ni 2+) is added to the medium. Prymnesium parvum does not have a strict Ni 2+ requirement for urea uptake.
The intent of this brief summary of purine catabolism is to point out that it is a source of biologically active organic nitrogenous compounds expected to be found at fish farm sites and available for phytoplankton assimilation. It is unknown whether purine assimilation and the concentrations of purines at fish farm sites influence phytoplankton species selections and blooms.
Polyamines are short-chain, aliphatic amines found in procaryotic and eucaryotic cells where they regulate cellular growth, differentiation and division, and control senescence [see Lu and Hwang, 2002; Nishibori and Nishi, 1997). The polyamine composition of PSP toxic Alexandrium minutum and Alexandrium tamarense has been reported (Hwang and Lu, 2000; Lu and Hwang, 2002; Nishibori and Nishio, 1997). Polyamines also regulate diatom silicification (Kröger et al., 2000), respond to physiological stress (Lu and Hwang, 2002) and are precursors of alkaloids which defend against herbivory - more details may be found at: http://www.hort.purdue.edu/rhodcv/hort640c/polyam/polyam.htm.
Three major polyamines are formed either by decarboxylation of ornithine, with putrescine the first step in polyamine biosynthesis:
ornithine ? putrescine ? spermidine ? spermine
or by the transamination of arginine:
arginine ? ornithine ? putrescine ? spermidine ? spermine
Iwasaki (1984), based on experiments with red tide species, suggested that exogenous polyamines stimulate and regulate blooms. More recently, a possible link between decay of the spring diatom bloom, its release of putrescine and blooming of ichthyotoxic K. mikimotoi in French coastal waters has been suggested (see Gentien, 1998). The connection made between the decay of the diatom bloom and polyamine release reflects its catabolic release during decarboxylation/transamination of amino acids, as in the pathways shown above. It is also well known that cadaverine, also a polyamine, and putrescine are common bacterial degradation products. Cadaverine is produced by decarboxylation of the amino acid lysine, and putrescine is formed from ornithine and arginine. High putrescine and cadaverine levels are found in spoiled fish, and are thought to be potentiators of histamine poisoning resulting from fish consumption (Todd and Holmes, 1993). Polyamine concentrations in sea water occur at nM levels. Summer concentrations in the Seto Inland Sea, an area of intense aquaculture, showed putrescine to be most abundant (32 nM) followed by sperimidine and spermine (14 nM) and lesser quantities of cadaverine (Nishibori et al., 2001). The investigators concluded that putrescine and spermidine contributed significantly to the DON pool. Gentien (1998) reported putrescine levels reached 100 nM at the pycnocline of a frontal system where diatoms were in decay; significant ornithine levels were also found.
The limited experimental data are consistent with the ideas of Iwasaki (1984) and Gentien (1998) that polyamines can be phytostimulants. Putrescine at concentrations ranging from 0.1 to 5.0 µM stimulated growth of ichthyotoxic K. mikimotoi (Gentien, 1998). Putrescine and cadaverine at concentrations of <110 µM increased the growth rate of the haptophyte Chrysochromulina leadbeateri [Section 7.1.3] in cultures, but high polyamine concentrations (1,100 µM) depressed growth (Legrand et al., 2001). Putrescine, spermidine and cadaverine have been shown to stimulate the growth, cell division and photosynthesis of the flagellate Dunaliella tertiolecta (see Legrand et al., 2001). Growth of the cyanobacterial bloom species Microcystis aeruginosa increased by 45 to 75% over that on NO 3 when supplied with putrescine and spermine (Maestrini et al., 1999). In addition to serving as growth factors, polyamines appear to be co-factors enhancing the ichthyotoxicity of toxins, as shown for spermine and prymnesin produced by Prymnesium parvum (Shilo, 1967). The addition of polyamines increased the haemolytic activity of extracts of C. leadbeateri cells and of sea water containing lysed cells and bacteria (Legrand et al., 2001).
[Note: the high experimental concentrations of polyamines used, at µM levels, contrast with the nM levels reported from in situ. The possibility that the experimental results exaggerate actual in situ responses to polyamines requires evaluation.]
The presumed relevance of polyamine synthesis, release and their potency as growth factors leading to their potential stimulation of harmful blooms by fish farm wastes stems from the following: dead fish release growth promoting polyamines supplementing their excretion during the catabolism of amino acids by the living stock. The view that a surge in polyamine release from dead fish can stimulate harmful blooms has been particularly espoused by Johnsen et al. (1999) and Legrand et al. (2001), who have evaluated the toxic 1991 bloom of C. leadbeateri in Norwegian waters near the Arctic Circle killing farmed salmon. The details of this bloom and the role of dead fish and polyamine release in its ichthyotoxicity have been considered in Section 7.1.3. Dead fish would appear to provide two nutritional supplements to a bloom population: stimulatory polyamines and nutrients released during remineralizaton of the carcasses. With regard to Scottish fish farms, Davies (2000) has stated that there is a ca. 10% mortality of salmon smolts in the first three months following their transfer to cages, and a mortality of 1.6% during the subsequent growth cycle. Should the expected release of polyamines initially stimulate bloom species, nutrients released thereafter during decomposition of the dead fish would help to prolong the bloom. This autotrophic benefit could be further enhanced by the ability of the bloom species to ingest (=phagocytosis) colloidal and larger particles, a nutritional capacity increasingly being demonstrated for dinoflagellates (see Stolte et al., 2002).
The intent of this brief summary, as that given for purine catabolism, is to point out that polyamines are potentially an important class of phytoplankton growth factors found at fish farm sites. It is unknown, however, whether the polyamine composition and concentrations at fish farm sites in Scotland influence species selections and blooms. Both the occurrence of polyamines and bioassays of their potential impact on the indigenous flora at fish farm sites in Scottish coastal waters should be investigated.
10.5 Shellfish cultivation and blooms
10.5.1 Shellfish cultivation in Scotland
Five species of shellfish are cultured in Scottish coastal waters: mussel ( Mytilus edulis), Pacific oyster ( Crassostrea gigas), king scallop ( Pecten maximus), queen scallop ( Chlamys opercularis) and flat or European oyster ( Ostrea edulis) (Pendrey and Fraser, 2002; Anonymous 2002). Shellfish cultivation in Scotland is primarily a mussel culture industry: of the total shellfish yield (3,350 t) in 2001 at 261 growth sites valued at 4 million £, mussels accounted for 89% and Pacific oysters 8%. The primary mussel growth areas are in the Strathclyde region (45% of the 2001 production) and Shetlands (28% production). There is considerable regional variation in mussel yield, the causes of which are uncertain. This has prompted extension of the traditional mussel growth sites in the south west and Highlands regions to the Shetland Islands where the greatest increase in mussel production has occurred (Pendrey and Fraser, 2002; Anonymous, 2002). Mussel cultivation is expanding; between 1993-2001 production increased 324% to 2,988 tonnes. Pacific oyster production (as individuals) during this period increased 36%, while king and queen scallop yields decreased approximately 50%, and flat oyster production was about 25% lower. Over 80% of the Pacific oyster yield is produced in the Strathclyde region.
Shellfish cultivation is projected to continue to increase over the short-term, particularly for mussel and Pacific oyster (Pendrey and Fraser, 2002). The outlook for king and queen scallop is less certain since the main cultivation method depends on natural spat settlement which is then on-grown in pearl and lantern nets suspended from long-lines [scallops increasingly are being on-grown on the sea bed (Anonymous, 2002). Queen scallop production is subject to considerable variation in natural settlement, and is further subject to restricted harvest because of exceedance of DA thresholds and potential ASP illness [see Section 2.11]. Mussel cultivation also depends on natural spat settlement, with the attached spat continuing to grow on the ropes suspended from long-lines or rafts until marketable.
Some companies have begun bi-culture of species: king + queen scallop; Pacific + flat oyster, and mussel + Pacific oyster cultivation. The cultivation of Pacific oyster in combination with the other species is cause for concern over potential harmful algal consequences. Crassostrea gigas is not native in Scottish waters, where it is unable to reproduce due to the prevailing low water temperatures (Anonymous, 2002). Growers depend on spat imported from hatcheries and nursery beds elsewhere in Europe. This importation has the danger of introducing harmful bloom species novel to Scottish waters in the transplanted seed stock. In an importation of Pacific oyster, Crassostrea gigas, to Ireland from France, 67 phytoplankton species were recorded from the gut contents and sediments of the consignment, including 15 dinoflagellate cyst producing species, three of which were harmful (O'Mahoney, 1993). The diversity and magnitude of phytoplankton species introductions in shellfish transplantations can be staggering. Dijkema (1992) calculated that 2.5 million viable dinoflagellate cysts can be transferred per tonne of mussels imported into The Netherlands from "red tide" areas, yielding an annual introduction of ca. 10 10 dinoflagellates into Dutch coastal waters. The sudden appearance in western Japanese waters in 1988 of Heterocapsa circularisquama, which produces a molluscicide, followed by rapid regional expansion has caused widespread devastation of edible oyster and pearl oyster ( Pinctada fucata) culture and mass mortality of short necked clam, Ruditapes philippinarum, and Mytilus galloprovincialis edulis (Horiguchi, 1995; Matsuyama et al., 2003). Within Japanese waters, regional spreading has accompanied transfer of shellfish seed stocks (Honjo et al., 1998). It has been confirmed recently that H. circularisquama is found in Hong Kong waters, from where it was probably introduced into Japanese waters through consignment of bivalve stocks (Iwataki et al., 2002). Blooms of Heterocapsa circularisquama in Hong Kong waters have been accompanied by dieoffs of natural bivalve stocks, but have not attracted much attention there since shellfish cultivation in those waters is limited (Fukuyo, pers. comm.).
It is recommended that Scottish authorities monitor Pacific oyster importations periodically for potential harmful species contamination and release at growth sites, and to develop precautionary shellfish importation and transplantation criteria. Particular attention should be placed on imported shellfish stock from French and Spanish growth sites where the assemblage of toxic species differs somewhat from that in Scottish waters [see Section 11].
10.5.2 Extensive nature of shellfish cultivation: impacts on natural phytoplankton community
Shellfish cultivation relies upon the availability of natural food, whereas farmed fish are fed processed diets. This feature, which Gowen et al. (1990) term extensive and intensive aquaculture, respectively, results in fundamental differences in the impact that these two aquacultural modes have on the indigenous phytoplankton flora, and in the vulnerability of the cultured stocks to harmful species. These differences primarily reflect the herbivorous (filter feeders) vs. carnivorous grazing modes of the cultured shellfish and fish, respectively, and override their maricultural similarities, i.e. both are open culture systems and their wastes are discharged directly into the natural habitat [see also Section 10.1; Handy and Poxton (1993)].
The intrinsic capacity of coastal waters to support bivalve culture and the impacts that can quickly develop at growth sites are demonstrated by introduction of the bay scallop Argopecten irradians into Chinese coastal waters (Yan et al., 2002). Introduced into China in 1983, an annual production of 50,000 t was reached in five years, the industry having developed from the introduction of <50 animals from the U.S. The rapid development of shellfish and shrimp aquaculture the past two decades in China has elevated this country to the world's largest producer, accounting for 63% of the total shellfish global production in 1995 (Yan et al., 2002).
Shellfish culture is a form of animal husbandry within a husbandry: the cultured stock consumes phytoplankton, excretes nutrients which stimulate new phytoplankton growth, upon which the shellfish filter feed. Filter feeding on the phytoplankton produced by this "internal cycling" mechanism is supplemented by filtration of phytoplankton advected into the growth sites from offshore waters [see Sections 11.5, 11.6]. Huge volumes of water are processed and "chemically conditioned" by shellfish during their filter feeding. Mussels cultured in a 7,000 ha lagoon were estimated to filter daily a water volume equivalent to 3.6-times that of the lagoon (Caroppo, 2000). The amount of phytoplankton biomass that can be consumed and the accompanying pseudofaeces production and release of nutrients are equally impressive [see Section 10.5.3]. A large mussel bank located in the Öresund (10 6 t wet wt) removed 75% of the chlorophyll in water being transported through the Sound at a rate of ca. 36,000 m 3 s -1, accompanied by a release of NH 4 from the sediments equalling 0.7 tonne hr -1 (Haamer et al., 1999). In Killary Harbour, Ireland, 47% (mean) of the chlorophyll was filtered from water flowing past mussel culture rafts (Rodhouse et al., 1985). The great filtration capacity of bivalves allows their continous exploitation of the indigenous phytoplankton, on which they selectively feed. This can cause significant changes in phytoplankton composition and community structure. Norén et al. (1999) concluded that intense mussel grazing favors the selection of fast growing nanophytoplankton over larger sizes, and this competitive advantage removes slower growing HAB species as they pass over a mussel bank.
Fish farms and shellfish cultivation impact the benthos and water column through their production of faecal and pseudo-faecal (by bivalves) wastes, deposition of unused food (finfish) and excretion of soluble wastes [Section 10.1]. However, shellfish cultivation in its reliance on the natural phytoplankton flora, and in competition with other planktonic herbivores, can consume large quantities of the natural phytoplankton stocks and change the local ecosystem structure. In Rîa de Arosa, the large scale cultivation of mussels replaced copepods as the dominant herbivore (see Gowen et al., 1990). Rodhouse et al. (1985, 1987) evaluated what the natural carrying capacity for mussel farming in Killary Harbour, Ireland, would be without major impact on foodweb structure. Based on the phytoplankton carbon production rate, an annual mussel production level of 340 tonnes could be supported without impact on the natural ecosystem. Above this threshold, the cultured mussels would begin to compete with natural zooplankton stocks for available phytoplankton, and alter the ecosystem. Fish farming, in contrast, does not lead to ecosystem changes because of competitive grazing with indigenous stocks; the use of processed food precludes such competition.
10.5.3 Vulnerability of cultured bivalves to indigenous, harmful phytoplankton species
A major difference between fish and shellfish aquaculture is the heightened vulnerability of shellfish to natually occurring blooms of indigenous, harmful species and phycotoxin accumulation. This is a result of their filter feeding and the large filtration volumes processed. Shellfish, unlike farmed fish, are directly exposed to, and accumulate phycotoxins that cause ASP, DSP and PSP. While these shellfish borne toxicities do not cause stock mortality, marketing closures lead to financial loss. Harmful blooms leading to mortality and great financial loss are commonplace at shellfish cultivation sites (Shumway, 1990). In China, for example, rapid development of shellfish and shrimp aquaculture the past two decades has elevated this country to the world's largest producer, accounting for 63% of the total shellfish global production in 1995 (Yan et al., 2002). This productivity has been accompanied by major harmful algal blooms, resulting in a loss of US $104 million in the period from 1981-1998. Increased frequency of PSP and DSP contamination and related sub-standard shellfish quality issues led the European Union to curtail export of marine bivalves from China from 1997-1999.
Where fish suffer toxic effects, this occurs through toxin vectoring and not from the direct uptake of phytoplankton [excluding clupeoids which graze Pseudo-nitzschia; see Section 2.1]. Ingestion of toxic cells by first-feeding stages of fish can lead to mortality (see Smayda, 1992), but this is not a factor in fish farming. Farmed fish mortality is usually limited to ichthyotoxic blooms and anoxia. Anoxic dieoffs of cultured shellfish also occur, as in Korea where an oyster dieoff was valued at >$60 million (Cho, 1979). A bloom of Ceratium tripos in the offshore waters of the New York Bight caused an anoxia induced dieoff of 143,000 metric tonnes of the surf clam Spisula solidissima (Mahoney and Steimle, 1979).
There are numerous examples of the vulnerability of cultured bivalves to blooms of harmful algae. The remarkable appearance and dispersion of the HAB dinoflagellate K. mikimotoi (= Gyrodinium aureolum) in northern European waters, including western Scotland, has been described in Section 5.1. This species is a serious threat to fish farms and shellfish cultivation: it has been frequently implicated in fish kills (Jones et al., 1982; Roberts et al., 1983; Turner et al., 1987) and invertebrate mortalities (Helm et al., 1974; Ottway et al., 1979; Partensky and Sournia, 1986), including commerically important shellfish species such as clams, mussels, oysters and scallops (Heinig and Campbell, 1992; Shumway, 1990; Smolowitz and Shumway, 1997; Tangen, 1977; Widdows et al., 1979). The impacts of K. mikimotoi on shellfish are species-specific (see Smolowitz and Shumway, 1997; Lesser and Shumway, 1993), ranging from cessation of feeding and mortality of post-larvae and juveniles ( Pecten maximus), reduction in clearance and growth rates ( Mytilus edulis), inhibition of feeding ( Argopecten irradians), and reduced larval survival ( Crassostrea gigas). Blooms of Cochlodinium polykrikoides caused mortality of Crassostrea virginica larvae and also disrupted calcium metabolism leading to deformed umbo stages in the presence of this toxic species (Ho and Zubkoff, 1979). Massive, regionally widespread mortality of bivalves in western Japan accompanied the appearance of the novel species Heterocapsa circularisquama which has become endemic in those waters (Horiguchi, 1995; Oda et al., 2001). Exposure of bivalves to toxic algae can also affect valve closure and cardiac ability, alter feeding and respiration rates and the production of mucus (Lesser and Shumway, 1997). These harmful species assaults on bivalves periodically compromise reliance on natural spat settlement as the maricultural basis for mussels and scallop, and may be a factor in the variable, interannual scallop yields experienced in Scotland [see Section 10.5.1].
K. mikimotoi blooms caused mass mortality of Pecten maximus post-larval stages, stopped the growth of juvenile stages, and reduced the growth and reproduction of adults (Erard-Le Denn et al., 1990). The toxins responsible for these harmful effects have not been identified. Two toxin types are suspected, including a fat-soluble, cytotoxin (Gentien and Arzul, 1990; Partensky et al., 1989), and compounds isolated from K. mikimotoi known to be haemolytic and ichthyotoxic (Yasumoto et al., 1990). Given the threat to fish farming and shellfish cultivation posed by K. mikimotoi, does aquaculture favor its blooms over naturally occurring blooms? Offshore frontal regions appear to be the preferred niche of K. mikimotoi in northern European waters (Gentien, 1998), a habitat preference with physical and chemical conditions that differ from those in nearshore habitats where aquaculture is usually carried out, and where K. mikimotoi is less frequently encountered. Since K. mikimotoi is expected to be relatively rare at nearshore shellfish culture (and fish farm) sites, nutrients released during cultivation will probably not stimulate blooms of local populations. More likely, the well known advection of K. mikimotoi populations from offshore into aquacultural sites, with or without local stimulation by excreted nutrients, is a greater potential problem [see Section 5.1]. For example, an offshore bloom (<10 6 cells L -1) of K. mikimotoi advected into Donegal Bay resulted in 80% mortality of cultured clams ( Tapes semidecussata) (O'Boyle et al., 2001). Advection of Dinophysis spp. from offshore into aquacultural sites, i.e. species that grow well without nutrient enrichment and result in DSP contamination, poses a particularly grave threat to shellfish culture, as discussed in Sections 3.3 and 11.6. Cultured bivalves are also vulnerable to qualitative changes in their appearance following ingestion of nuisance species which affect their marketability. For example, in New Zealand coastal waters, ingestion of Mesodinium rubrum and Noctiluca scintillans by cultured, greenshell mussel ( Perna canaliculus) discoloured their gut contents blue-red (Rhodes et al., 2001).
Shellfish appear to have several defense strategies to protect against physiological impairment from phycotoxins and their accumulation. They can tolerate high toxin body burdens; the content of PSP toxins in field populations of bivalve molluscs ranged from 2,000 to 10,000 µg STXeq 100 g -1 wet tissue weight (Table 1 in Bricelj and Shumway, 1998). At high phytoplankton population levels, such as characteristic of high density, low biomass blooms ( HDLB), filtration ceases. For example, a toxic bloom of the pelagophyte Aureococcus anophageffens in Narragansett Bay killed about 99% of the mussel population ( Mytilus edulis), but the clam Mercenaria mercenaria largely survived (Smayda and Villareal, 1989). The survival of Mercenaria mercenaria appears to be partly related to having stopped filtering at population levels >250 million cells L -1. [ Aureococcus reached a maximum population of 10 9 L -1] Bivalves also exhibit selective feeding rejecting undesirable species in their pseudofaeces [see Section 10.5.3; Shumway, 1900].
In summary, shellfish culture is extensive in its reliance on the natural phytoplankton flora, which is subjected to intense grazing pressure during filter feeding by bivalves. Cultured bivalves compete directly with natural stocks; the natural ecosystem is then subject to alteration depending on the intensity of cultivation. Shellfish cultivation is a dual husbandry process, and the question that arises is whether the waste products stimulate blooms of undesirable species increasing the high vulnerability of cultured shellfish to phycotoxins. This is considered in the following section.
10.5.4 Shellfish cultivation, wastes and blooms
Molluscan filter feeders exposed to a mixture of phytoplankton and seston particles exhibit preingestive selection and selectively graze available particles (Bougrier et al., 1997; Shumway et al., 1985). Molluscs sort particles during filter feeding using their labial palps. This allows preferential rejection of undesirable particles and phytoplankton species, and their voidance as pseudofaeces. The differential selection and rejection of dietary matter are influenced by the size, shape and composition of the particles. Thus, filter feeding bivalves exhibit some control over their ingestion of phytoplankton species and phycotoxins.
The capacity for, and the results of selective feeding vary among bivalves. Ostrea edulis fed a mixture of three species of diatom, dinoflagellate and cryptomonad preferentially consumed the dinoflagellate ( Prorocentrum minimum) and rejected the diatom as pseudofaeces. In contrast, Crassostrea virginica ingested the cryptomonad ( Chroomonas salina) and enriched its pseudofaeces with P. minimum (Shumway et al., 1985). Crassostrea gigas, when fed a mixture of seven phytoplankton species, preferentially filtered and rejected (as pseudofaeces) the four proffered diatoms relative to the flagellates, whereas Mytilus edulis preferentially filtered the flagellate Tetraselmis suecica and significantly rejected more diatom and other flagellate species than did C. gigas (Bougrier et al., 1997). It is unknown to what extent, and whether bivalves reject toxic species, a capacity found among zooplankton (Smayda, 1992).
Organic waste deposition rates in shellfish cultivation are typically one- to two-orders of magnitude lower than in fish farming (Gowen et al., 1990), but are still considerable and can change the sedimentary chemical and benthic community structure (Brown et al., 1987; Dahlbäck and Gunnarsson, 1981; Mattson and Linden, 1983). The accumulation of organic material under mussel culture platforms at daily deposition rates of up to 3 g C m -2 d -1 in Swedish coastal waters altered the sediment chemistry and the bacterial acommunity (Dahlbäck and Gunnarsson, 1981). There was almost complete disappearance of the original benthic community three months after initiation of mussel culture (Mattson and Linden, 1983). The general impact of aquaculturally derived changes on benthic habitat chemistry and the biota and the impact of these changes on blooms were discussed in Section 10.1.
Shellfish farming significantly influences the nitrogen cycle at the water-sediment interface, as shown by field studies carried out at scallop ( Crassostrea gigas) and mussel farm sites in France and New Zealand (Gilbert et al., 1997; Kaspar et al., 1985). This disruption is of ecological interest since benthic remineralization is an important pathway by which recycled N released from the sediments fluxes into the water column where it becomes available for phytoplankton growth. The effect of aquacultural wastes on the N cycle was discussed in Section 10.1. Harvesting of cultured bivalves which filter-feed on the natural phytoplankton communites removes large quantities of N from the habitat. Nitrogen loss due to denitrification can also become significant. Kaspar et al. (1985) showed that the denitrification rate of sediments at a mussel farm were 7-fold greater than at a control site. In the case of C. gigas, farming lowered the nitrification rate and stimulated dissimilatory reduction of NO 3 to NH 4. Enhancement of the latter process reduces the denitrification rate and loss of N available for phytoplankton growth. It is unknown to what extent the altered N dynamics associated with shellfish cultivation influences the response of the indigenous flora.
The impact of filter-feeding bivalves on phytoplankton composition and abundance is the outcome of two different effects: an antagonistic effect that results from grazing (filter feeding), and stimulation by excreted nutrients and those released from decomposition of pseudofaeces and faeces. Gowen et al. (1990) cited a Japanese report that a typical oyster rack holding 420,000 tonnes of oysters will generate 16 tonnes of faecal and pseudo-faecal material over a 9-month growing season. The amounts of pseudofaeces and faeces produced can be huge. Cognie and Barillé (1999) estimated that the 110,000 tonnes of Crassostrea gigas farmed in France produce annually about 1.1 million tonnes of pseudofaeces and 390,000 tonnes of faeces, corresponding to a ratio of 9.6 and 3.5 tonnes per tonne C. gigas, repectively. Copious mucus production associated with faecal waste production also occurs: 1.14 t:t. Cognie and Barillé showed that the mucus coating of pseudofaeces and faeces stimulated the growth of eight diatoms and a flagellate, including the diatom Haslea ostrearia which causes "greening" in oyster. [Unfortunately, Cognie and Barillé did not examine the responses of Pseudo-nitzschia spp. that produce domoic acid leading to Amnesic Shellfish Poisoning.] Thus, similar to marine gastropods, mucus production by bivalves provides nutrients for µ-algae on which they feed. Limpet mucus has also been reported to be stimulatory (Conner and Quinn, 1984).
Arzul et al. (1996) biossayed the response of Alexandrium catenella ( PSP), Alexandrium minutum ( PSP), Heterosigma akashiwo (ichthyotoxic), K. mikimotoi (ichthyotoxic) and the diatom Chaetoceros gracile to dilutions of excreta produced by Pacific oyster ( Crassostrea gigas) , the mussel ( Mytilus chilensis), farmed Atlantic salmon, and sea bass ( Dicenthrarchus labrax). Notable differences in response were found among species and to the source of the excreta. Oyster and mussel excreta generally stimulated growth rate, while the Atlantic salmon and sea bass excretions were inhibitory. Within this, the bivalve excretions were more potent stimulators of growth in H. akashiwo and C. gracile than NH 4+; K. mikimotoi was unaffected by oyster waste, but inhibited by sea bass excretion; while A. catenella and A. minutum were unaffected by the extracts of the bivalve and fish excreta. In other studies, dilutions of decomposed oyster faeces stimulated flagellate growth, including H. akashiwo and dinoflagellates (Hirayama and Numaguchi, 1972; Iwasaki, 1969,1973).
The limited experimental data available suggest that shellfish wastes can stimulate phytoplankton growth. However, the overall impact of bivalve biofiltration on phytoplankton species composition and on selective stimulation of bloom species, whether harmful or benign, is unclear. There is stronger evidence that fish farm wastes influence phytoplankton growth [see Sections 10.3, 10.4], a conclusion that may be related to the greater number of studies that have been carried out. However, there may be a real difference between the extensive and intensive aquaculture of shellfish and fish, respectively, in altering phytoplankton behavior. While shellfish cultivation is more vulnerable to harmful species than fish farming, the latter may be potentially more stimulating to harmful species because of the chemical diversity and composition of its wastes. The corollary of extensive vs. intensive aquaculture is that shellfish cultivation removes nutrients, fish farming adds nutrients.
10.6 Summary and recommendations
Although shellfish cultivation is still limited in Scotland, based on the analyses presented here, it is safe to conclude that irrespective of the annual production level and independent of the species cultured this industry will be compromised periodically by unpredictable occurrences of harmful algal blooms and phycotoxins. The current ASP problem compromising native scallop harvesting in Scottish waters reflects this probability [see Section 2.2]. This prediction is made not because of an anticipated stimulation of harmful blooms by fish farm and shellfish cultivation wastes, such an effect is unlikely for the reasons given previously in this review. Rather, ASP, DSP and PSP toxin accumulation at shellfish cultivation sites is commonly observed and the result of bivalve filtration of local or advected populations, and whose presence and abundance are not directly related to waste nutrients released at the cultivation site [see also Sections 3.4 and 3.5].
The aperiodic and unpredictable nature of HAB events at cultivation sites will require increased vigilance and monitoring by Scottish agencies and aquaculturists to ensure seafood safety and to protect public health. The successful cultivation of the greenshell mussel ( Perna canaliculus) can serve as a guide for this required, and recommended effort (Rhodes et al., 2001). The New Zealand greenshell mussel industry is valued at US $62 million, with export to 55 countries. To protect the industry, weekly phytoplankton monitoring is carried out at 30 commercial sites and biotoxin monitoring at ca. 80 sites nationwide. Failure to implement an adequate monitoring program not only compromises stock protection, but also marketing. In South Africa, because of toxic bloom problems an embargo was placed on the export of shellfish to EU countries until appropriate monitoring protocols were developed and implemented (Pitcher, personal communication). Such restrictions and monitoring for quality control are not only needed because of harmful bloom impacts. The extent to which shellfish aquaculture increases harmful blooms and/or stimulates the toxicity of species, or whether the increased production of shellfish through cultivation merely magnifies an effect that is normally found in natural stocks remains unresolved. This uncertainty is complicated by novel and unusual phytoplankton behavior, such as the sudden appearance and persistence of Heterocapsa circularisquama in the Seto Inland Sea where its blooms are causing mass mortality of bivalves and financial loss (Horiguchi, 1995; Oda et al., 2001). It is suggested that the experience of Irish investigators in coping with the problems of DSP with respect to the bloom ecology and advection of the causative Dinophysis spp., the diverse toxins encountered, and the impacts of DSP occurrences on shellfish cultivation [see Section 3.0] be carefully considered by Scottish agencies and aquaculturists in designing monitoring protocols to protect the burgeoning shellfish cultivation industry and human health. This will require increased oceanographic investigation of coastal waters, including the occurrences and behaviour of frontal systems in response to physical forcing (winds, etc.), and the quantitative assessment of the distribution, abundance and species composition of phytoplankton communities. The use of remote sensing surveillence in such investigation is encouraged.
« Previous | Contents | Next »