Who would have thought that when I sent out the first reLAKSation in the New Year of 2000 that I would be still writing and publishing this weekly commentary. Originally conceived as a comment on the salmon market, reLAKSation had evolved into a weekly view on the battle with the wild fish sector about the causes of the demise of wild salmon stocks. As I have mentioned more than once, the biggest surprise has been the total unwillingness of those who blame salmon farms to engage in any sort of discussion. I will continue to express a view regardless. I don’t know whether I will still be writing reLAKSation in another twenty five years’ time, I will certainly keep on trying.
Quote: Salama et al. (2018) (Journal of Fish Diseases) begins the discussion of their paper by expressing the following view.
Models can only be useful if they represent aspects of reality and features of the outputs have a potential application (Foreman et al 2015, Salama & Rabe 2013) However, models are not required to be completely parsimonious with reality as they would cease to have purpose due to the complexity, which may also lead to overcomplex models which would need to be highly specific and would be cumbersome (Murray 2008, 2009).
The use of the word ‘parsimonious’ is an odd one but it does have a place in science. Raykov & Marcoulides, (1999) state that
The principle of parsimony reflects the notion that researchers should strive for simple measurement models that use the minimum number of parameters needed to explain a given phenomenon.
Here, in a nutshell, is the root of the problem concerning the impacts of sea lice on wild fish. In attempting to keep the models as simple as possible, some basic but fundamental aspects of the science have been ignored.
I have written several times about the concept of ‘Model Land’ in which modellers can become trapped. The reality is that the key ‘sea lice models’ are far from reality because of the view that models are not required to be parsimonious with reality seemingly because of the complexity. However, in the case of sea lice, aiming for less complex models also ensures the models cease to have purpose. Its no wonder that Marine Directorate and SEPA scientists are so reluctant to talk to me because no-one else is challenging them on the relevance of their models, especially as a way of protecting wild salmon.
Traffic Lights: Late last year, the Sea Lice Expert Group put out a request for any information that might help them refine the Traffic Light assessment process. I therefore submitted details of an alternative narrative for sea lice based on science that does not seem to have been considered by the established research community. Just prior to Christmas I received a short communication from the head of the Sea Lice Steering Committee in which she wrote:
“The scientific content are known to the Expert group and has been taken into consideration when applicable.”
Given that the members of the Sea Lice Steering Committee and Expert Group are all part of the established research community, this reply did not come as any surprise. I have requested for references to any scientific studies that include this alternative science as known to the Expert Group, but I am not hopeful of receiving any response. The science of the Traffic Light System is the same science of the established research community and the Expert Group, and they are unlikely to accept that the established science may be wrong.
To be fair to the Expert Group and the Steering Committee, the alternative scientific content may well be known to them, but it is ignored because it doesn’t fit into the established narrative.
Over the festive holidays, I have been reanalysing the sea lice data collected by IMR as part of the assessment process for the Traffic Light System. As with the data from previous years, I struggle to understand how the Expert Group have reached the conclusions they have, other than by their use of the Taranger risk formula and the various models they have developed.
Five (of thirteen) production areas include data collected by trawling for wild salmon smolts. This is in addition to the data collected from sea trout caught either by traps or nets or a combination of both. Trawling for salmon smolts only too place in those production areas with the highest number of salmon farms (except PO5).
When the total lice infestation of the five areas (PO2-PO6) is plotted on a graph, the distribution is the typical aggregated distribution that is a defining feature of parasites. 65% of the salmon smolts collected were lice free and 90% carried 3 lice or less.
By comparison, the graph of sea lice infestation on sea trout caught by nets and traps shows that only 10% of the fish caught from these fish production areas were lice free.
The graph shows counts up to 50 lice but to compare the 90% coverage with that of trawled salmon smolts, sea trout carrying up to 90 sea lice must be included. Clearly, the sampled sea trout are much more infected with sea lice than the wild salmon smolts.
The question is why is there such a difference between the trawled and the trap/net data?
I would argue that part of the problem is the size of the sample. Small samples are not representative of the population and inevitably consist of the weaker more infested fish. The IMR data records how many fish were captured by what method on each day, but they do not record how many traps are used, whether it is just one for each location or several. In their report, IMR say that the traps are the preferred method of capture and nets are used to supplement the trapping where trapping is challenging such as if the currents are strong or if there is a large tidal difference. Seven of the 32 sites across all thirteen production areas, including one in PO2, do not appear to have deployed traps at all, relying just on nets. Eight of the 15 sites in the PO2-PO6 areas use both traps and nets but only one uses just nets, (8 out of the 32). What is unclear is whether one trap or more than one is used in each location because this will affect the sample size.
The total catch from netting and trapping for all thirteen production areas is shown in the following table. There is clearly a lot of variability in the days that the sampling occurs and how many fish are caught. The main focus appears to be on those production areas with the largest numbers of farms.
In their report, IMR state that initially, the sites selected for sampling were chosen on the results of their sea lice dispersal model but in later year, they say that the sites were selected independently of the dispersion model but assessed in relation to an infection model and the estimated migration time of the salmon. The earlier assessments sampled more sites with 66 sampled compared to 32 in 2024. It is unclear why IMR have chosen to ignore half the current number of sites.
In addition, the number of days sampling is variable ranging from just 2 in one production area to 42. The minimum days spent sampling at the sites in the areas from PO2 to PO6, where trawling also took place was 7 to 42.
It is worth noting that the trawling wasn’t that successful, which can be gauged from the data by the number of the trawl recorded for each fish. Typically, fish were only caught on 17% of the trawls, the maximum number in any area being 127.
Finally, as IMR do not publish protocol on sampling, I refer to the Scottish version used to catch fish by seine net. This states a minimum of 30 fish should be caught in each sample. If the Norwegian data is applied to this protocol, then an average of 92% of the samples were caught in numbers of less than 30 fish. 21% of the samples comprised of just one fish. These fish are hardly representative of the whole population.
Although the Norwegians do not seem to have a protocol, Taranger’s 2012 report recommended a minimum of 100 fish, which means that eight of the production areas fail to meet this minimum number in total.
IMR say that both traps and nets are passive in the way they catch fish because the fish must swim into them and potentially underestimate the true infection pressure of sea lice on sea trout. They provide a list of reasons:
Lice can affect the swimming ability of the infested fish by which IMR imply that infested fish do not actively swim into the traps and nets.
Infested fish tend to stay in brackish water such as the estuaries, and because no sampling takes place in such places, IMR believe that they are failing to assess the population.
Heavily infested fish may die and thus are not represented in the samples.
Lice infested fish may return to rivers seeking out freshwater to delouse thus masking the real infestation pressure.
On reading this list, seems that IMR are determined to establish that sea lice associated with salmon farms are damaging wild stocks. Yet, if they are so certain, then
- Don’t they use a more active method of sampling such as a seine net.
- Why don’t they sample in brackish water.
- Why don’t they put a counter or trap in selected rivers to catch returning sea trout.
It is interesting to note that 172 sea trout were caught by trapping and nets in 2024 with lice numbers ranging from 100 to 330, none of which were resident in brackish water, or had returned to freshwater or had died. Instead, this rather confirms the science that says a few hosts carry many parasites. Even with poor sampling, this was just 7% of the fish.
Unfortunately, the Sea Lice Steering Committee and Expert Group say that they know this science but seemingly don’t appear to know how to interpret it. Of course, the Traffic Light System is based on their science, so they are unlikely to want to admit that they have got it wrong.
Sentinel cages: Regular readers may remember that a Freedom of Information request to SEPA revealed that they were to spend £200,000 with the Norwegian Institute of Marine Research for help with sentinel cage work.
SEPA’s consultation document states:
‘Currently, the most effective way to obtain a measure of sea lice infestation pressure is using sentinel cage studies.’
However, in their latest assessment of sea lice infestation of wild fish along the Norwegian coast in 2024, IMR state:
‘The results from the sentinel cages are used primarily to compare infection pressure in time and space, and not as a direct estimate of the infestation of wild fish.’
If IMR do not use sentinel cages to measure infestation of wild fish, could it be that SEPA are going to pay out £200,000 and not learn how to achieve their aim. Unfortunately, SEPA have been misled into thinking sentinel cages will provide more than they can give. But as they are unwilling to talk to those who question their approach, SEPA appear intent on imposing a pointless regulation on one of Scotland’s most important sectors.
Salmon conservation: The Scottish Government published the outcome of the consultation on the proposed river gradings for the 2025 season on the 20th of December, just as many of us were packing up for the Christmas holiday. Perhaps, they thought that by burying it in the Christmas festivities, their proposals might go unnoticed. The consultation had closed on 6th September and received just 37 responses so it is unclear why it has taken so long to publish the outcome. In my experience, it doesn’t make much difference what views are expressed, the Scottish Government appears reluctant to implement any of the radical action needed to protect stocks of wild salmon. This is apparent from the main thrust of the consultation which included a proposal to allow catch and release fishing on the river Annan between 25th February and 31st March. The proposal intends to bring the river Annan in line with other districts in Scotland, which have early periods during the closed season where fishing is allowed. It doesn’t appear that the Scottish Government considered bringing other fishery districts into line with the restrictions on the Annan and put a stop to all early fishing across all of Scotland to protect wild salmon and especially the early spring salmon.
The Scottish Government point out that the change to the fishing on the Annan requires a minor amendment to the Conservation of Salmon (annual Close Time and Catch and Release) (Scotland) regulations 2014. However, anyone looking at this regulation should quickly realise that, at a time when salmon are considered endangered, that this regulation needs a major overhaul.
The regulation lists the Annual Close Time for all fishery districts, which one would assume is when fishing is not permitted. One would also assume that the Annual Close Time would apply equally across all Scotland. Sadly, this is not the case. Fifty-one districts have a close time from 27th August to 31st March. The Annual Close Time for another nineteen districts runs from 10th September to 31st March. Three others run from 21st August whilst two districts run from 1st September but to 30th April and one other from 14th September to the 31st March. Is it not time, that one Annual Close Time running from 27th August to 31st March should apply to all fishery districts.
As I have pointed out, one would assume that the Annual Close Time is when fishing is not permitted, however, when it comes to anything concerning Scottish angling, nothing could be so simple. The regulation allows for anglers to fish by rod and line within the Annual Close Time. For example, anglers fishing the Nith can continue to fish past the start of the Close Time on 10th September until 30th November, whilst others can also fish before the Close Time comes into force on the 31st March as long as no fish are killed. For examples, The Helmsdale opens for fishing on the 11th January. In effect, the Annual Close Time which would typically last for around 32 weeks is reduced to just 6 to 7 weeks.
The Annual Close Time was introduced for a purpose, presumably to protect the fish during the spawning period. The demands of the angler have eroded this over many years so that there is now only a handful of weeks during when angling does not take place, although in some rivers it is possible to fish for grayling when fishing for salmon and sea trout is prohibited.
The main outcome of the Scottish Government’s consultation on salmon conservation seems to be that anglers fishing the Annan, a river already considered to be of a poor conservation status, can now fish for longer than in previous years. How this will help conserve salmon stocks is a mystery.
However, I am guessing that the Scottish Government and anglers probably believe that the issues of salmon conservation will be solved through SEPA’s ever more Draconian sea lice risk framework. After all, once these salmon farms can no longer emit sea lice, they believe that the problems of wild salmon will be solved.