Ecological Effects of Sea Lice Treatments

Ecological Effects of Sea Lice Treatments
Scottish Association for Marine Science
The Marine Laboratory, Aberdeen
Plymouth Marine Laboratory
This report presents the results of a five year project investigating the ecological
effects of medicines used for controlling sea lice on farmed salmon. The objective of
this report is to explain the background to the project, describe the research
methodology, and provide a synopsis of the project results and conclusions.
Measurement and sampling programmes were undertaken at four active salmon farm
sites on the west coast of Scotland in Lochs Diabaig, Craignish, Kishorn and Sunart.
Each of these sites initially had access to the bath-treatment Excis as the main
medicine for sea lice control but in 2001 the in-feed medicine Slice became available
for use and was used at three of the sites.
Aspects studied include: examination of littoral and sub-littoral settlement panels to
assess whether settlements of flora and fauna are affected by chemical usage;
sampling of sediments around and away from the farms for meiofauna, macrofauna
and the presence of Slice; zooplankton sampling before during and after medications
and as a time-series to assess natural variability; time-series measurements of
phytoplankton populations and nutrient concentrations; acoustic ground
discrimination to determine substrate type around each farm; experiments to measure
hydrographic parameters using Differential Global Positioning System (DGPS)
drifters and current meter arrays to allow modelling of the dominant water movement
processes and simulate dispersion of the treatments used at each site.
The project's primary aim was not to determine local effects of sea lice medicines: all
discharges to the marine environment have some effect on the receiving environment
and it is a general principle of pollution control that such perturbations should be
confined to the immediate environment, the mixing zone. We attempted to look
beyond ephemeral, local effects and concentrated on trying to detect long-term
changes beyond the immediate mixing zone. This was a difficult task and, taking all
the results together, we did not detect any clear effect of medicine usage, or indeed
other farm activities, beyond the local scale. The processes of species succession and
population dynamics that we observed were well within the range of what might be
expected or predicted for fjordic sea loch systems.
The project has achieved much by helping to improve our understanding of natural
variability in relatively unstudied systems and, most especially, by demonstrating that
wide-scale ecosystem-level effects from medicine use, if they exist at all, are likely to
be of the same order of magnitude as natural variability and, therefore, inherently
difficult to detect.
Sea lice are a major problem for salmon farmers in terms of fish health costing the
industry in the region of £15-30 million per annum. Anglers and conservationists are
also concerned as excessive background lice larval numbers may be one factor in the
decline in wild sea trout and salmon on the west coast of Scotland. Several strategies
are pursued to reduce sea lice on farmed fish but the availability of an effective range
of sea lice treatment medicines is crucial. Although these products have passed a
variety of regulatory tests, only when they are used on a commercial scale over long
periods can the large-scale, long-term effects of these products be accurately assessed.
A number of new treatments have recently been granted Marketing Authorisations.
Before they can be used, these treatments also require discharge Consents and much
effort has gone into ecotoxicity testing to establish Environmental Quality Standards
(EQS). Discharge consents are granted so that after discharge the concentration in the
receiving water at the edge of the allowable zone of mixing does not exceed the EQS,
thus avoiding acute or chronic harm to the environment. Ecotoxicity tests are
generally performed on indicator or sentinel species that give a good indication of the
likely environmental risk. However, any wider effects can only be ascertained once
the new treatments are in use at production scale. Therefore, this project was designed
specifically to look for any medium to long-term (1-4+ years) ecosystem responses
attributable to the use of some of these veterinary medicines.
The objective of the study from the outset was to detect changes in the natural species
assemblages in Scottish sea lochs and determine whether such changes might result
from the use of sea lice treatments. To this end, the study was designed to assess
changes in the plant and animal assemblages from a range of habitats in the vicinity of
the farm locations selected for study by measuring a wide range of ecosystem
Sea Lice
Figure 1. Adult Lepeophtheirus salmonis. Photo courtesy of Alan Pike.
Salmon farms in Scotland are affected by sea lice, Lepeophtheirus salmonis (Figure 1)
and to a lesser extent Caligus elongatus (Figure 2, Figure 3), ectoparasitic copepods
that feed on the skin of the salmon. Sea lice, in common with other copepods, have a
free-living life stage called the nauplius. After developing into the copepodid (at
which stage the lice are actively searching for a fish host), the sea lice metamorphose
into the chalimus stage on contact with a fish. Lepeophtheirus salmonis will only
infect salmonids, while C. elongatus have been found on a range of marine fish. The
sea lice undergo four chalimus stages while they develop (Figure 4) attached to the
host fish before they metamorphose into the pre-adult and adult stages, which are
unattached stages, allowing the sea lice access to the whole of the fish’s skin surface
(Figure 5).
Figure 2. Caligus elongatus. Photo
courtesy of Paul Tatner.
Figure 3. Adult Caligus females. Photo
courtesy of Gordon Rae.
Caligus may be found all over the body of the fish, while the larger Lepeophtheirus
show some preference for the head region. The open wounds caused by untreated sea
lice weaken the fish and allow the proliferation of secondary bacterial and viral
infections that can lead to mortality.
Figure 4. Life cycle of sea lice. Drawing courtesy of Gordon Rae.
Figure 5. Damage to fish from sea lice. Photo courtesy of Jim Treasurer.
Veterinary Medicines for Sea Lice
Recently eleven compounds representing five modes of action were identified as
being used worldwide on commercial salmon farms in the period 1997 to 1998,
although the substances used in individual countries varies. These included two
organophosphates (dichlorvos, azamethiphos); three natural pyrethrin/pyrethroid
compounds (pyrethrum, cypermethrin, deltamethrin); one oxidizing agent (hydrogen
peroxide); three avermectins (ivermectin, emamectin benzoate, doramectin) and two
benzoylphenyl ureas (teflubenzuron, diflubenzuron). In the EU, a substance for which
a Maximum Residue Limit has been established can only be used as a veterinary
medicine following the granting of a Marketing Authorisation. In the period 1999 to
2004 (the duration of this project), only Salmosan, Excis, Salartect, Slice and Calicide
were licensed for use in Scotland (Table 1).
Table 1. Sea lice medicines used on commercial salmon farms in Scotland.
Active Ingredient
Emamectin benzoate
Hydrogen peroxide
Trade Name
Medicine Type
Benzoylphenyl urea
Oxidizing agent
Calicide (teflubenzuron) has not been widely used in Scotland and Salartect
(hydrogen peroxide), which degrades rapidly to water and oxygen, is not considered
to be a hazard to marine life. Excis and Slice were the two most commonly used sea
lice medicines in Scotland during this project:
Excis (Cypermethrin)
Cypermethrin is the active ingredient in Excis. Cypermethrin is a synthetic pyrethroid
which acts on the nervous system by increasing sodium permeability of the nerve
membrane during excitation, resulting in prolonged nerve stimulation. It has low
water solubility, degrades rapidly and is applied as a bath treatment. Cypermethrin
released following a bath treatment is rapidly diluted in the sea and the majority is
eventually adsorbed onto particulate material, which will settle to the sea bed. The
main degradation products of cypermethrin are approximately 1000 times less toxic
than the parent compound. It does not accumulate in fish as they readily metabolise
pyrethroids. Bioconcentration factors are greater in shellfish, which can be exposed
through feeding on particulates or though direct uptake, apparently due to slower rates
of metabolism and depuration.
Slice (Emamectin benzoate)
Although all the appropriate ecotoxicological testing has been carried out to allow this
product to be licensed for use, there is relatively little information available in the
published literature on the fate and ecological effects of Slice in the marine
environment. Emamectin benzoate is the active ingredient of Slice. Emamectin
benzoate is a semi-synthetic avermectin that acts by increasing membrane
permeability to chloride ions and disrupting physiological processes. Emamectin
benzoate is administered as a component of salmon feed and enters the marine
environment associated with uneaten feed and fish faeces. Salmon readily assimilate
emamectin benzoate and excretion continues for an extended time period after
administration. Given its association with particulate material and low water solubility
most emamectin benzoate will settle to the sea bed. The organisms most likely to be
exposed to this chemical are, therefore, sediment dwelling animals in the vicinity of
fish farm cages. Slice use is increasing in Scotland and, in many loch systems,
strategic treatments are being undertaken simultaneously at several farm sites.
At the outset of the project Excis became widely available to fish farmers including
the farms selected for study. Discharge consents for Slice were delayed for a variety
of regulatory reasons and it was not until the summer of 2001 that we were able to
identify a suitable site with a Slice discharge consent.
The Study Sites
During the first three years of the project four fish farms on the Scottish west coast
(one each in Lochs Craignish, Sunart, Diabaig and Kishorn) were identified as being
appropriate for the project. One problem encountered at the start of the project was
that much of the industry around the Oban area (where the lead institute SAMS is
based) was closed owing to an outbreak of ISA (Infectious Salmon Anaemia). This
meant that farms had to be selected from further away than was logistically optimal.
The longest time-series of data was collected from a fish farm at the head of Loch
Sunart, with samples collected from late 1999 through to mid-2004. This site initially
used Excis with occasional use of Salmosan, but in 2002 gained access to and used
Slice. Sampling also began at Loch Diabaig in 1999, but most sampling ceased there
in 2001. The Loch Diabaig farm used Excis, Salmosan and Slice but proved generally
unsatisfactory as a result of its high winter exposure that resulted in damage to
moored equipment. The third site in Loch Craignish was intensively sampled in 2000,
but most sampling was discontinued in 2001 as only one lice treatment was reported
during the 12 month period and the site had not applied for a Slice consent. In 2001
we transferred our attention from Loch Craignish to a site in Loch Kishorn that had
consent for Slice and planned to use it. The work in Kishorn continued until the end
of 2002, when the site became fallow.
Study Components and Results
In order to determine the broad range of basin-scale effects from sea lice treatment
medicines over the medium to long term, we focussed on the main components of the
ecosystem that were likely to be affected:
Benthic meiofauna - animals < 1.0 mm in size, living on or in the sea bed
Benthic macrofauna - animals > 1.0 mm in size, living on or in the sea bed
Inter-tidal organisms - plants and animals settling on the sea shore
Sublittoral organisms - plants and animals settling on hard substrates deployed
in the water column
Zooplankton - small animals drifting in the water column
Phytoplankton - microscopic plants drifting in the water column
Each sampling programme involved multiple stations at varying distances from the
fish farm in order to examine whether there were any differences that might be
attributed to sea lice treatments. Fish farms have some well known local ecological
effects as a consequence of organic enrichment so particular attention was paid to
Crustacea as the component of the community most susceptible to effects from sea
lice medicines.
In addition to the ecological work, the study was complemented by measurements of
sediment medicine concentrations and the development and use of computer models
to simulate the dispersion of medicines. Each of these components is discussed below:
Meiofauna in Sediments
Meiofauna is a collective name given to animals that are small enough to pass through
the sieve meshes (1 or 0.5 mm) which are used to separate larger animals
(macrofauna) from sediments. Specialised techniques are used to separate the
meiofauna from sediments, but essentially they are collected on a fine mesh with
apertures 63 µm across. In sediments such as occur in sea lochs the most abundant
meiofaunal groups would be expected to fall into two major categories, namely freeliving nematodes (roundworms) and tiny crustaceans in a group called harpacticoid
copepods. Although generally too small to be seen with the naked eye, meiofaunal
animals are useful indicators of environmental change for several reasons. Their small
size and high densities mean that smaller volumes of sediment are required, and the
majority of the sediment processing may be done under the controlled conditions of
the laboratory. Apart from such practical advantages, there are also sound scientific
reasons for expecting meiofaunal assemblages to be sensitive to environmental
change. Unlike the macrofauna, many species of which have a planktonic larval stage
and lifespans in excess of a year, meiofauna have no specific dispersal phase and
spend their whole lives in one place, living, breeding and feeding. Meiofaunal
reproduction tends to be rapid, with generation times of a few months, and
continuous, with juveniles occurring throughout the year. This means that they are far
more likely to react rapidly to changes in environmental conditions.
We initially looked at nematode and copepod assemblages in the vicinity of salmon
farms where bath treatments for sea lice using Excis, were being used. In Loch Sunart,
meiofauna samples were collected from the same sites, and at the same time, as
macrofauna samples in October 1999 and March 2001. Excis was used on several
occasions between these two surveys. Samples were also collected in Loch Diabaig in
November 2000 and August 2001, again before and after Excis treatments. Slice was
also used at this site during the period, but as the spatial dispersal and potential
impacts of in-feed treatments are expected to be rather different to those of the bath
treatment Excis, a separate intensive study on the effects of Slice was undertaken at a
farm in Loch Kishorn in 2001.
Nemotode community composition close to the fish farm cages was similar in Lochs
Kishorn, Diabaig and Craignish, and differed from assemblages at the reference
stations (Figure 6). Nemotode assemblages at Loch Sunart were generally different
from those observed in the other lochs. Copepod community composition displayed a
similar, if less clear, pattern (Figure 7).
At Loch Kishorn, where samples were collected from very close to the farm, strong
gradients in meiofaunal assemblage structure were apparent, and there was evidence
of greater differences in copepod community structure between stations in posttreatment surveys. Meiofaunal community structure was most related to station
distances from the cages and it is likely that deposition from the cages impacted
nematode and copepod assemblages, with a combination of changes in sediment
composition and the presence of in-feed treatment chemicals causing the observed
changes in copepod community structure. At Loch Diabaig, there was no evidence
that sea lice treatments were a major driver of patterns in meiofaunal community
structure, instead most of the differences between stations can be explained by
variation in sediment type, and to a lesser extent other factors (distance from the
cages, depth, % total organic carbon or nitrogen) which may or may not reflect
organic enrichment effects from the cages.
Figure 6. MDS ordination of similarities between abundances of nematode genera
averaged within stations from different surveys. Contours group samples at a level of
40 % similarity. The stations are denoted as follows: K11A = Kishorn, Survey 1 (June
2000), transect 1, station A; D22 = Diabaig, Survey 2 (August 2001), Station 2; C13
= Craignish, Survey 1 (October 2000), Station 3; S11 = Sunart, Survey 1 (October
1999), Station 1; K3R = Kishorn, Survey 3 (January 2001), Reference station; and so
forth. All stations represented in the plot were sampled with a van Veen grab except
the reference stations at Kishorn.
At Lochs Sunart and Craignish, changes in meiofaunal assemblages were also
unrelated to sea lice treatments. Only small differences in meiofaunal community
structure were observed at Loch Sunart, which were related primarily to site location,
with differences in depth and sediment structure rather than fish-farming activity
driving the differences seen. Seasonality and different sampling methods may also
have caused the shift in pattern. At Loch Craignish, species known to be indicators of
disturbance and organic enrichment contributed to differences between stations,
leading to the conclusion that the observed differences between stations reflect the
normal organic enrichment gradient associated with fish-farming activity.
Figure 7. MDS ordination of similarities between abundances of copepod taxa
averaged within stations from different surveys. Contours group samples at a level of
25 % similarity. The stations are denoted as follows: K11A = Kishorn, Survey 1 (June
2000), transect 1, station A; D22 = Diabaig, Survey 2 (August 2001), Station 2; C13
= Craignish, Survey 1 (October 2000), Station 3; S11 = Sunart, Survey 1 (October
1999), Station 1; K3R = Kishorn, Survey 3 (January 2001), Reference station; and so
forth. All stations represented in the plot were sampled with a van Veen grab except
the reference stations at Kishorn.
The comparative studies between sites have proved to be useful, with the results
indicating that, at some level, the effects of fish-farming on meiofaunal communities
are similar in different sea lochs. Therefore, the results of in-depth studies in one sea
loch may provide useful information for the purposes of regulation at another. The
results also suggest that it may be possible to select a subset of species which could
act as indicators of aquaculture impacts.
Macrofauna in Sediments
The identification and enumeration of bottom-living macro-invertebrates
(macrobenthos) and subsequent analysis of the population data is a widely-recognised
means of assessing impacts of discharges of materials on the marine environment, and
is usually an important component of monitoring requirements stipulated by
regulators when granting consents to discharge.
The macrofauna studies were designed to detect any medium term (up to 3-4 year)
effects on benthic populations (meiofaunal and zooplankton studies detect more rapid
changes). Macrobenthic invertebrate populations (retained on a 1 mm mesh after
sieving samples of sea-bed sediments) were sampled along transects in the vicinity of
each fish farm and at reference sites anticipated to be well beyond the zone of impact
of the fish farms in each sea loch. Where possible, samples were collected before and
after sea lice medicine treatments. Additional samples were collected at annual
intervals in an attempt to quantify the influence of natural seasonal effects on the
populations being compared.
The effects of aquaculture on the macrobenthic communities in this study were
broadly comparable between the different sea lochs, although macrofaunal
assemblages at Lochs Kishorn and Diabaig were more similar than those at Loch
Sunart (Figure 8).
Figure 8. MDS plot of similarities between macrofaunal abundances averaged within
survey/station combinations from Lochs Sunart, Diabaig and Kishorn. The first letter
indicates the site (S, Sunart; D, Diabaig; K, Kishorn). The first numeral indicates
surveys (For Sunart 1, October 1999; 2, March 2000; 3, November 2000; 4, March
2001; 5, November 2003. For Diabaig 1, July/August 2000; 2, November 2000; 3,
August 2001. For Kishorn 1, June 2001; 2, August 2001) and the second numeral
indicates stations (1 to 4 at Sunart, 1 to 3 at Diabaig, 1 to 3 and R at Kishorn).
Differences in macrofaunal community composition between sample stations in each
loch indicated a gradient of organic enrichment originating at the fish farm cages,
with no evidence to indicate that sea lice treatments adversely affected the
macrobenthic communities in any of the lochs studied.
At Loch Sunart, although macrobenthic assemblages close to the fish cages differed
from the reference station, they did not change over time indicating that sea lice
treatments between surveys did not cause species abundance or composition to
change. The gradient of organic enrichment was considerably stronger at Loch
Kishorn and differences in community composition were observed between surveys
undertaken before and after a sea lice treatment with Slice. In particular, reductions in
the number of crustacean and echinoderm species and their abundance were observed
at one station close to the cages after the Slice treatment. These groups are sensitive
to both organic enrichment and sea lice treatment agents, thus either (or both) could
have been responsible for the differences seen, and overall there was no evidence that
Slice adversely affected the macrobenthic community at Loch Kishorn. Similarly, at
Loch Diabaig, differences in macrofaunal assemblages between stations and surveys
were more likely the result of seasonal variability and organic enrichment. Species
diversity was lowest closest to the fish cages and dominated by nematodes and small
polychaetes, while species typical of normal shell/muddy sediments dominated at the
reference station. Following a Slice treatment, dominance increased at all stations and
crustaceans declined closest to the cages.
Shore Flora and Fauna
At all four fish farm sites studied the inter-tidal zone consisted mainly of bedrock
and/or boulders. Following initial ecological surveys, three or four shore stations were
established in each area, two stations at locations where information on current flows
suggests that any effects of the treatment chemicals would be most likely to be
detected, and one or two reference stations, situated beyond the anticipated zone of
influence of activities associated with the fish farms.
All of these shores were characterised by abundant growths of brown algae mixed
with abundant invertebrate populations of attached barnacles and mussels, limpets and
mobile snails: mainly Littorina species and the dog whelk Nucella lapillus, which
feeds on barnacles.
Barnacles (Semibalanus balanoides) are the only sedentary crustaceans that settle
predictably on these shores in significant amounts. Barnacles belong to the same
group as sea lice, and are therefore appropriate ‘target’ organisms for studies aimed at
detection of any effects of the sea lice treatments. In each area, sets of four slate
panels were bolted to the rock at approximately Mid Tide Level to receive any
settlements (Figure 9). The barnacles and other biota on the surrounding rocky shores
were also studied and photographed at each visit (approximately three times per year
at each location). Settlement on the slates revealed that barnacle densities and
subsequent growth varied considerably between sea lochs, and between study sites in
each sea loch.
Figure 9. Slates showing heavy barnacle settlement.
The highest barnacle settlement densities occurred at Loch Diabaig where exposure to
wave action was the highest of the four lochs. There was also evidence of a
relationship between reduced barnacle egg production and predicted impact at this
At Loch Diabaig, barnacles at the predicted high impact station (Station C) produced
fewer egg masses than barnacles at the predicted intermediate impact (Stations B) and
reference stations (Station A; Figure 10). The difference was greatest in 2000 and
2001 with 60 - 70 % of barnacles without eggs at the high impact station. In 2002 and
2003, this had fallen to 30 - 50% at the high impact station, but had increased at the
other two stations.
Thus, the overall trend was of increasing frequency of egg mass occurrence at the
predicted high impact station and decreasing frequency at the intermediate and
reference stations (Figure 10).
Figure 10. Reproductive failure of littoral barnacles (Semibalanus balanoides) at
Loch Diabaig. Error bars are 95% confidence limits.
The reproductive failure of barnacles observed in this study could have been caused
by a variety of natural factors such as degree of exposure, and not just fish farm
activities such as sea lice treatments. The trend diminished with time over a period
when fish farm activity and the use of sea lice treatments also declined at the site, but
we do not know whether this is correlated or not. However, the results indicate that
the influence of environmental factors on egg production in littoral barnacles merits
further investigation, including study areas remote from fish farm activities.
Sublittoral Fauna Settlement
The study of organism settlement on the sublittoral arrays of slate panels suspended at
three different depths and at three or four stations in Lochs Sunart, Diabaig and
Craignish yielded a clear picture of natural seasonal and annual successions, and
fauna and flora abundances.
Overall, species diversity of fauna settling on the sublittoral slates was limited, but
similar at all sites studied. At all sites and stations the most significant ecological
event each year was the spring settlement of the dominant sublittoral barnacle
Balanus crenatus. At all stations the slate panels were heavily colonised, with the
barnacles growing quickly to attain 100% cover, especially at the optimal upper and
mid-water depths. The results clearly show that barnacle settlement was similar at all
stations, regardless of distance from the fish farm cages. It follows that sea lice
treatments at the salmon farms did not affect barnacle settlement at any of the sites
Other crustacean species settled on the panels only rarely, such as Balanus balanus at
Lochs Sunart and Diabaig and the small crab Porcellana longicornis at Loch Sunart.
In December 2003, the non-native (alien) caprellid amphipod Caprella mutica was
observed for the first time at two of the sample stations in Loch Sunart.
Macroscopic algae (seaweeds) scarcely featured on the sublittoral slates, although
they often colonised the near-surface ropes and buoys; and mussel spat (Mytilus
edulis) settled readily on ropes, hydroids and any fine-branched attached seaweeds.
The main settlement events observed are summarised in the following figures (Figure
11 to Figure 14):
In April to June heavy settlements of barnacles (Balanus crenatus) usually occur
followed, through the summer months, by mussels (Mytilus) and ascidians (seasquirts) and lesser numbers of a variety of other species (a total of around 40).
Meantime, the main predators of barnacles and mussels, namely the sea-slug
Onchidoris bilamellata, followed by the even more voracious starfish, Asterias
rubens, settle from the plankton, grow rapidly and kill all the barnacles. The strong
empty shells of the barnacles then often remained for many months, greatly altering
the character of the surface for all later settling animals through the late summer and
autumn. These included increasing numbers of ascidians and often polychaete worms
including Pomatoceros and Hydroides in calcareous tubes and those in soft muddy
tubes such as Sabella pavonina (peacock worm) (Figure 11).
Figure 11. Loch Sunart Station 1, middle array. 12 Dec 2001, showing abundant
polychaete worms Pomatoceros, Hydroides, and Sabella pavonina.
Figure 12. 6 July 2000. Massive settlement of barnacle Balanus crenatus
(superabundant; 100% cover), with young ascidians Ascidiella aspersa and Ciona
intestinalis and hydroid Obelia longissima all common. The barnacles were being
eaten by nudibranch Onchidoris fusca (frequent) settled directly from the plankton.
Figure 13. 7 Nov 2000. The barnacles are now mostly empty shells having been eaten
by the predators Onchidoris fusca and the starfish Asterias rubens (common), which
also settles (later) from the plankton. The ascidians Ciona intestinalis (common) are
now much larger. Slate #2 (deployed 6/7/00) has extensive cover (~30%) of the
gelatinous prostrate ascidian Diplosoma listerianum, also being browsed by Asterias;
also present occasional young tube worms Pomatoceros triqueter. Note mussels
Mytilus edulis abundant on the upper bar and Asterias feeding on them.
Figure 14. 3 March 2001. The empty barnacle shells are mostly still present; some
loss of the mature ascidians; loss of Mytilus from top bar and predator Asterias.
Figure 15. Massive growth of ascidians overloading settlement arrays.
In several lochs, the settlement of ascidians was massive and completely obscured
everything else present on the slates, the mooring ropes and the buoys. At one
location, the over-loaded arrays had to be removed at the end of one season (Figure
Copepods are the main constituent of the zooplankton and are the major herbivores in
sea loch ecosystems, playing a fundamental role both as consumers of phytoplankton
and as a food resource for larger animals. Because zooplankton have limited mobility
they are largely dependent on currents and tides for their location, consequently they
are unable to avoid unfavourable water conditions. The response of zooplankton to
chemical exposure is generally considered to be informative of the relative impacts on
the whole ecosystem. Significant declines in copepod abundance and consequent
reductions in their grazing rates could potentially result in serious environmental
problems such as algal blooms. Planktonic copepods have a similar life cycle to
ectoparasitic copepods (sea lice) and are likely to be adversely affected by sea lice
medicines, which are highly toxic to crustaceans.
Zooplankton are most likely to be exposed to cypermethrin in soluble form given its
mode of administration as a bath treatment. In the immediate post-treatment period,
zooplankton entrained within a cypermethrin treatment plume may be exposed to
potentially lethal concentrations for several hours. Cypermethrin is then predicted to
adhere to particulate material in the water column providing a further route of
exposure to zooplankton through ingestion.
Zooplankton have a lower risk of exposure to emamectin benzoate as it is
administered as a component of the salmon feed. The most plausible exposure route is
associated with ingestion of feed and faecal particles. Excretion of emamectin
benzoate by salmon continues for an extended period post-treatment, increasing the
duration of any exposure to particulate associated emamectin benzoate. Resuspension
of freshly deposited material will also reintroduce emamectin benzoate to the water
column for a considerable period post-treatment thus making it potentially available.
To determine the effects of sea lice medicines on the zooplankton community,
sampling campaigns were undertaken at fish farms using the bath treatments Excis
and Salmosan, and the in-feed treatment Slice. Initially, sampling campaigns were of
relatively short duration with plankton samples collected before, during and after sea
lice treatments. However, because of difficulties distinguishing treatment-related
changes in zooplankton community composition from naturally occurring changes
observed during the short-term surveys, a long-term fortnightly sampling campaign
was subsequently undertaken in Loch Sunart, during which Excis and Slice were the
sea lice medicines predominantly used.
At each site, zooplankton samples were collected from four or five sample stations
positioned around the farm cages. Three or four sample stations were located within
100 m of the salmon cages, and a reference station was located at a distance from the
cages. At each sample station, five samples were collected using a small boat using a
plankton net (120 µm mesh) hauled from a maximum depth of 24 m.
For all sea lice treatments undertaken during this study no adverse affects on
zooplankton were detected at either the species or community level. Changes
observed were naturally occurring, with patchiness in distribution, life history
characteristics, and water currents being the most influential factors affecting
zooplankton distribution and community composition. The long-term sampling
programme undertaken in Loch Sunart from November 2001 to April 2004 allowed us
to identify the natural seasonal cycles in abundance and species composition and
confirmed that sea lice treatment events did not significantly alter the seasonal trends.
The seasonal cycles of species and abundance were similar in 2002 and 2003, with
peaks in abundance during the summer months between May and September in both
years, although total numbers were higher in late summer 2003 (Figure 16). Declines
in abundance in late summer (September/October) could have mistakenly been
attributed to sea lice treatments with Slice (2002) and Salmosan (2003) had a two year
data set not been collected to identify the seasonal cycles of abundance.
Figure 16. Total zooplankton abundance (number m-3) at Loch Sunart during the
long-term monitoring program (November 2001 to May 2004). Vertical bars are the
stations (A, B, D) positioned close to the fish cages (0 and 100 m). Closed circles are
the reference station. Timing of sea lice treatments is indicated by the arrows; Excis
(C), Salartect (HP), Salmosan (A), Slice (EB1, EB2, EB3).
The small copepods which dominate zooplankton communities in Scottish sea lochs
are the group most likely to be affected by sea lice treatment agents because of their
toxicity to crustaceans. Nonetheless, it is perhaps unsurprising that we did not detect
treatment related affects on the zooplankton communities in this study.
Notwithstanding the patchy nature of zooplankton distribution, which makes detection
of treatment related effects difficult, it is unlikely that water column concentrations of
sea lice treatment agents were maintained at sufficiently high levels to cause
discernable or widespread changes in the zooplankton communities.
In the case of Excis, measured and predicted cypermethrin water column
concentrations following a single cage treatment at Loch Sunart were considerably
lower than concentrations shown to cause acute toxicity to copepods in laboratory
studies. The results also indicated that the 3-hour EQS of 16 ng l-1 was unlikely to
have been breached following single cage treatments with Excis. It is possible that
there may have been localised affects on zooplankton within dispersing treatment
plumes. However, the wider zooplankton community within the loch was not
adversely impacted at a detectable level following any sea lice treatments in this
Phytoplankton play an important role in marine ecosystems. Their growth is directly
influenced by physical factors such as temperature, salinity, light and nutrients, and
they are first link in the marine food chain. Information on seasonal changes in
phytoplankton species and abundance is vital as it allows identification of responses
that may be caused by the use of sea lice treatment agents, as opposed to those caused
by physical factors such as changing salinity, temperature, light and nutrients, or
biological variables such as grazing pressure.
Phytoplankton and supporting salinity and nutrient data were collected in high
frequency sampling programmes so that effect(s) of sea lice treatments on
phytoplankton communities could be separated from the effects of continually
changing environmental conditions.
Despite differing physical conditions and sea lice treatment histories, comparison of
the four sea loch phytoplankton communities revealed significant similarity between
them. This reflects the widespread distribution of many of the phytoplankton species
in Scottish coastal waters. In all lochs, the phytoplankton community was typically
dominated by a relatively small number of species with many other species present in
low numbers at different times of the year. Some of the more common (and bloomforming) species were observed year-round at all sites (e.g. small Chaetoceros sp.,
Skeletonema costatum, Gymnodinium sp., and unidentified cryptophytes).
Phytoplankton blooms occurred with normal frequency and duration for Scottish
coastal waters and were caused by species commonly observed at all sites.
Summer populations were more species-rich than winter populations, and some
typical species observed include the diatoms Eucampia zodiacus, Leptocylindrus
minimus, Pseudo-nitzschia sp., Asterionellopsis glacialis, a variety of pennate
diatoms, and Cylindrotheca closterium. Some representative dinoflagellates observed
during summer months include Dinophysis (e.g. D. acuta, D. acuminata),
Protoperidinium sp. and Heterocapsa sp.
Cell abundance during the winter months (November to February) was lower than
during summer and dominated by microflagellates, with low numbers of diatoms and
dinoflagellates (Figure 17). Species typical of winter populations included the
dinoflagellates Gymnodinium sp. and Gyrodinium sp., and the diatoms Cylindrotheca
closterium, unidentified naviculoids, Skeletonema costatum and Pseudo-nitzschia sp.
The presence or absence of any phytoplankton species could not be attributed to the
use of sea lice treatments in any of the four lochs. Instead, changes in phytoplankton
community composition and abundance were more closely related to season,
temperature, salinity and nutrient concentrations, and to a lesser extent zooplankton
Population percentage
Others *
Figure 17. Gross phytoplankton community composition at Loch Sunart between June
2000 and April 2004. A Skeletonema costatum bloom (17.2 x 106 cells l-1) in March
2004 is included to illustrate its dominance at this time. “Others*” includes other
flagellates, silicoflagellates, cryptophytes, ciliates, tintinnids and resting stages/cysts.
Modelling Chemical Dispersion
In order to fully understand the ecological fate of sea lice medicines it is important to
be able to predict where these substances might go when released into the
environment and at what concentrations. It was by modelling that we hoped to
improve our understanding of physical, chemical and biological processes and to
allow predictions of impact at other sites.
Sea lice medicines in bath form disperse in the water current after being released from
salmon cages. Designing sampling protocols for medicines and their impacts is made
much easier by the use of computer models that can predict how water moves around
in the vicinity of the cages. Information on current movements was collected using
current meters deployed at each site, and acoustic surveys provided information on the
sea bed topography and sediment types in each loch basin. Drifting buoys were also
used to track water movements during sea lice treatments to determine the horizontal
dispersion characteristics at each site and thus give an idea of the rate of dispersion of
the treatment chemicals. The drifters have sophisticated electronics for receiving
satellite Global Positioning System (GPS) signals and radioed GPS corrections from a
land base. The differentially corrected positions (DGPS) are consistently accurate to
±2 m with a maximum error of ±4 m. During a deployment, the position of the drifters
was recorded every 30 seconds at the shore base. The drifters were attached to socks
hanging in the water column at a depth of 6 m thus allowing water movement at that
depth to be tracked at the surface. This novel technology is a vast improvement on
more traditional methods as more frequent and accurate position fixing can be
The data from the current meters and drifters were then used in mathematical models
to simulate dispersion of sea lice medicines following a treatment. The models
predicted chemical concentrations in the dispersing treatment plume and final
sediment concentrations following settlement. The results from the model simulations
were then compared with chemical concentrations measured in water and sediment
At Loch Sunart, cypermethrin water column concentrations were measured in an
effluent plume following a single cage treatment with Excis in January 2002. The
treatment plume was tracked for a period of three hours post release using drifter
buoys so that water samples could be collected from the plume.
Figure 18. Observed and predicted cypermethrin concentrations in the effluent plume
following a single cage treatment at Loch Sunart in January 2002. The drifter buoy
track is shown in northings and eastings. Sample collection times post-release are
shown in brackets; bars show predicted and observed (means of duplicate samples)
concentrations (ng l-1).
Cypermethrin concentrations measured in the dispersing treatment plume in the first
43 minutes post release ranged from 2.2 ng l-1 to 21 ng l-1 (Figure 18). Cypermethrin
was not detected in the majority of samples collected after this. Only four samples
taken after 43 minutes contained cypermethrin concentrations higher than the
detection limit (1 ng l-1). Predicted cypermethrin water column concentrations were
within the same order of magnitude as measured concentrations for the same times
and locations (Figure 18). Given that concentrations decreased 1000 times from the
treatment concentration of 5000 ng l-1 over a short period of time (within 30 minutes),
this is considered an acceptable level of accuracy for the model.
The model was used to predict sediment cypermethrin concentrations following a
farm wide treatment (Figure 19). The centre of deposition was positioned to the west
of the cage group, with peak concentrations of about 20 ng g-1 dry weight. The area
of sediment predicted to contain concentrations above the detection limit of 2.25 ng g1
was about 108,000 m2, with a length and breadth of roughly 530 and 260 m
Concentration (ng g
-1 dry weight)
Figure 19. Predicted sediment cypermethrin concentrations (ng g-1) following a full
farm treatment at Loch Sunart, 30 January to 6 February 2002. The hatched
rectangle represents the cage group.
The results suggest that multiple cage treatments may result in the deposition of
detectable quantities of cypermethrin on the seabed in the vicinity of fish farms.
However, because cypermethrin may remain in the water column for several hours,
whether in soluble or adsorbed form, the location of peak deposition will be
determined by the prevailing wind-driven currents during and after treatment, in
addition to the tidal current direction. The dependence on wind-driven currents makes
prediction of cypermethrin dispersion, and selection of sample sites, more difficult.
This is in contrast to in-feed medicines, whose settlement is primarily determined by
tidal currents, allowing more accurate prediction of deposition patterns.
The sediment concentrations depicted in Figure 19 are due to direct deposition of
cypermethrin; chemical decay was not included in the simulations. Decay of the
compound in high organic marine sediments is thought to halve concentrations after
35 days. Processes of sediment resuspension and redistribution, which were not
included in the model, are also likely to reduce the predicted concentrations.
Emamectin benzoate deposition was modelled using the particle tracking model
DEPOMOD, which is used by SEPA in consenting Slice applications. Prediction of
sediment emamectin benozoate deposition footprints below the fish farm cages
assisted positioning of sample stations for benthic fauna and sediment chemistry. The
predicted post-treatment concentrations were also compared with measured
concentrations from sediment samples and have helped with further validation of the
Figure 20. Predicted emamectin benzoate sediment concentrations at Loch Sunart on
days 9, 114 and 269 post-treatment.
At Loch Sunart, the predicted footprint of sediment emamectin benzoate deposition
was below and slightly to the west of the cage group as a result of the residual current
(Figure 20). Following sea lice treatments with Slice in 2002, predicted sediment
emamectin benzoate concentrations were considerably higher than measured
concentrations of the chemical in sediment samples collected from the stations along
transects S1 and S2 (indicated in Figure 20). In most of the sediment samples
collected after Slice treatments at Loch Sunart, emamectin benzoate concentrations
were either trace or not detected. Stations where emamectin benzoate was detected
were directly below the cages and concentrations were generally two orders of
magnitude lower than predicted concentrations. The chemical was also not detected in
samples collected from the macrofauna stations following two Slice treatments at the
Given that the DEPOMOD model has been validated for different environments and
farmed fish species, possible reasons for the large differences between predicted and
measured sediment emamectin benzoate concentrations during this project are: 1)
inaccurate or insufficient representation of emamectin benzoate behaviour or
properties in the model; 2) inaccurate model input data, particularly use of default
data; 3) unreliable or insufficient measured concentrations to test the model across a
range of temporal and spatial scales.
Discussion and Conclusions
This project was conducted in the real world of regulated commercial salmon farming.
In many cases this created difficulties in planning the field sampling campaigns. For
example, because of the lengthy discharge consenting process for Slice, work was
unable to start on this medicine until the summer of 2001 nearly two years after the
start of the project. Our philosophy had been to establish long term sample collection
programmes at three sites and to continue them for the duration of the project but this
strategy had to be abandoned when the risk that we might not get adequate time to
sample a site using Slice became too great. At that point we transferred our attention
from Loch Craignish to Loch Kishorn.
It is important to emphasise the dynamic nature of the marine ecosystem, which is
driven by changes in season, weather and biological factors (e.g. in variations in the
timing of larval supply), leading to natural differences between years and even
decades. These natural changes can obscure more subtle changes caused by the
discharge of medicines and other substances into the sea.
Long-term zooplankton and phytoplankton sampling campaigns confirmed that sea
lice treatments did not alter natural seasonal trends as the processes of species
succession and population dynamics were well within the range of what might be
expected or predicted for fjordic sea loch systems. Similarly, the study of organism
settlement on sublittoral arrays yielded a clear picture of natural seasonal and annual
species successions and abundances, beginning with the spring settlement of barnacles
each year.
Gradients in meiofaunal and macrofaunal community structure were observed at most
of the sites, however, the differences observed between stations generally reflected the
normal organic enrichment gradient associated with fish farming activity. At Loch
Kishorn, it is possible that a combination of changes in sediment composition and the
presence of in-feed treatment chemicals may have impacted meiofaunal assemblages,
however, these variables are co-linear with the strong organic enrichment gradient at
this site. In contrast to the other sites, stations at Kishorn were deliberately positioned
very close to the cages to try to detect some effect.
Barnacles are the only sedentary organisms that settle predictably on Scottish rocky
shores in high numbers. The highest barnacle densities occurred at Loch Diabaig, the
most exposed site. It is possible that the low frequency of egg masses observed at this
site may have been related to sea lice treatments at the fish farm and this aspect of the
study merits further investigation to determine the effects of natural and
anthropogenic processes on barnacle reproduction.
The broad objective of the project was to determine the ecological effects of sea lice
treatments in Scottish sea lochs, and in those terms that objective has been met, with
no gross effects of medicines on the receiving environment distinguished. The project
has achieved much by helping to improve our understanding of natural variability in
relatively unstudied systems and, most especially, by demonstrating that wide-scale
ecosystem-level effects from medicine use, if they exist at all, are likely to be of the
same order of magnitude as natural variability and, therefore, inherently difficult to
The scientific consortium responsible for this work are pleased to acknowledge the
assistance we have received from the farm managers and their staff at Lochs Sunart,
Diabaig, Kishorn and Craignish. Without their logistical support this project would
not have been possible. Thanks also to Jim Treasurer, the late Gordon Rae, and Alan
Pike for photos. Paul Tatner supplied the photo of Caligus elongatus from the
Biomedia Museum Project - a joint funded project originally compiled by Dr Andrea
Fidgett and directed by Dr Douglas Neil, University of Glasgow in association with
the Universities of Paisley and Strathclyde.
This project was sponsored by:
The Department for Environment, Food, and Rural Affairs Veterinary Medicines
Directorate and Chemicals and Genetic Modification Policy Directorate;
Scottish Executive Environment and Rural Affairs Department;
Scottish Natural Heritage;
The Scotland and Northern Ireland Forum for Environmental Research and Scottish
Quality Salmon.
Questions or Comments
Any questions or comments on this work should be emailed to the scientific coordinator, Dr Kenny Black: [email protected] and/or to the sponsors co-ordinator Dr
Nick Renn: [email protected]
Glossary of Technical Terms
Ectoparasitic: External parasites
Transect: A line along which samples are taken.
Reference station: Control sample station positioned beyond the area where impacts
are expected.
Environmental Quality Standard (EQS): A value, generally defined by regulation,
which specifies the maximum allowable concentration of a potentially hazardous
chemical in an environmental sample such as water or air.
ng l-1: 1/1,000,000,000 gram per litre = 1 part per trillion.
Crustaceans: A group of animals that have an external skeleton, jointed legs, and a
segmented body. Most crustaceans live in water.
Copepods: Small shrimp-like crustaceans.
Echinoderms: Spiny-skinned invertebrates that live on the sea floor such as sea
urchins and starfish.
Polychaetes: Marine segmented worms, usually with bristles on each segment.
Hydroids: Small animals with stinging cells that build colonies of polyps. Related to