Welcome to Model Land: Earlier this year, physicist and former editor of Nature and New Scientist, Mark Buchanan wrote in Nature Physics about the lure of models. He says that scientists cannot now help using models and such models underlie some of sciences’ most impressive achievement ranging from Einstein’s theory of relativity to today’s numerical weather predictions. Yet at the same time, he points out that:
“Even so, models should come with safety warnings. Most people realise that model predictions can be wrong, and often are wrong. Yet, misplaced confidence can draw individuals, groups, or entire governments into making painfully poor decisions. Worse, as Erica Thompson of the London School of Economics examines in her incisive new book ‘Escape from Model Land’, models have an almost magical capacity to lure their users into mistaking the sharp, tidy, and analytically accessible world of the model with actual reality with unfortunate consequences.
Over the last twenty years or so, models have increasingly pervaded into the aquaculture sector and nowhere are they used as much than to predict the infestation of wild salmon by sea lice emanating from salmon farms. This has proved to be a highly contentious subject and so the publication of this book makes extremely interesting and relevant reading. Of course, the book does not include any reference to sea lice but focusses on major issues such as modelling Covid 19, climate change and the economy. However, the principles behind using models are all the same.
In her book, Erica Thompson creates the idea that the world of models can be very distinct from the real world. She writes that:
“Though Model Land is easy to enter, it is not so easy to leave. Having constructed a beautiful internally consistent model and a set of analysis methods that describe the model in detail, it can be emotionally difficult to acknowledge that the initial assumptions on which the whole thing is built are not literally true. This is why so many reports and academic papers about models either make and forget their assumptions, or test them only in a perfunctory way. Placing a chart of model output next to a picture of real observations and stating that they look very similar in all important respects is a statement of faith, not an evaluation of model performance and any inferences based on that model are still in Model Land.” One example of this of which everyone will be familiar occurred during the pandemic when graphs of actual deaths were compared with the number predicted by the models.
The above paragraph rings so many alarm bells for me in relation to the application of modelling to sea lice. Erica Thompson says that comparing outputs from the model with a real observation can be more an act of faith, yet for sea lice, we seem to be only presented with the model output and no real observations at all. If such real observations are absent, the model is surely then well beyond any act of faith. In the last issue of reLAKSation, I presented an image taken from the time series of sea lice dispersal generated by the model for the SPILLS project. This showed where the larval lice should be on any given day and at what concentration. Yet when the research team undertook collection of larval lice from the area, only 4.8% of a total of 372 samples contained any lice at all. The Scottish Government’s SPILLS report stated that failure to capture sea lice larvae does not equate to their absence.
Perhaps if this had been the first attempt to catch dispersed sea lice larvae, then there might be some belief in this view, but papers published by researchers trying to identify sea lice larvae in the sea stretch back to 1995. Clearly, if no-one has yet managed to find significant quantities of larval lice by now, then is it time to ask whether this failure to capture larval lice does mean that they are in fact absent.
There are actually two possible alternative explanations as to why sea lice larvae have not been identified in the numbers claimed. The first is that the workings of the model are wrong, and the dispersed lice do not follow the patterns that the model predicts. Thus, sampling for lice in the area where the model predicts that they will be is therefore pointless as the lice are simply not there. However, this assumes that the lice larvae may be somewhere else, yet to be identified.
The second explanation is that as Erica Thompson states the initial assumptions on which the models are built are literally not true. This would suggest that the lice are not captured because the model wrongly assumes that lice are dispersed by wind and currents when they are not.
However, with sea lice there is a second initial assumption that also could be wrong and that is sea lice are a significant threat to wild salmon and sea trout. As yet, there are no graphs showing predictions of modelled wild salmon mortality against actual fish that have died.
There is possibly a good reason why larval sea lice have not been identified in the water column, but unfortunately, no-one has been willing to provide me with a better explanation. This brings me to an additional point made by Erica Thompson and that is the issue of accountability.
In her book, Erica Thompson writes: “Reliance on models for information tends to lead to a kind of accountability gap. Who is responsible if a model makes harmful predictions? The notion of ‘following the science’ becomes a screen behind which both decision makers and scientists can hide saying ‘the science says we do must X’ in some situations and ‘its only a model’ in others. The public are right to be suspicious of the political and social motives behind this kind of dissimulation.” She continues “Scientists and other authorities must do better at developing and being worthy of trust by making the role of expert judgement much clear, being transparent about their own backgrounds and interests and encouraging wider representation of different backgrounds and interests.”
The problem when it comes to sea lice in relation to what happens in Scotland is that the main source of limited research, the development of models, the development of and provision of advice to Government and any assistance to impose regulation policy are all under the control of Marine Scotland Science. Yet, at the same time, Marine Scotland Science are still unable to locate actual real sea lice larvae in the seas that form the heart of their own Model Land.
Escapes: The Scottish Government has recently written to one of the long-term complainers about salmon farming to say that work is ongoing to strengthen the escape regulatory regime. and they are exploring how penalties for fish farm escapes events could be introduced with a view to redistributing this money to support environmental measures related to wild salmon conservation and research.
When will the Scottish Government start to understand that salmon farming has minimal impact on wild salmon and sea trout whether from sea lice or escapes? The reason why wild salmon are in trouble has more to do with the fact that the wild fish sector has spent more time and effort catching and killing wild fish than trying to understand the long-term changes happening to wild fish stocks. If they want to penalise salmon farmers why not penalise anglers too and make them pay for every fish that they kill and want to keep. Fisheries managers claim wild salmon are worth several hundred pounds so why shouldn’t that amount be paid to the Scottish Government when a fish is killed. Had this policy been introduced in 1952, just think what revenue for wild salmon conservation could have been raised from the 5.9 million fish killed since then. However, anglers don’t think that the current crisis is their fault so will do everything possible to make sure that they don’t have to pay any more than they already do for their fishing. They already balk at the thought of paying a small amount for a fishing licence.
A recent study has shown that river managers think that sea lice and escapes are damaging to wild fish stocks, but this is just what they think. There is very little hard evidence to support these claims. The link to Norwegian stock is primarily made through genetic markers such as those found used in human genetic detection to find links to Viking, Irish or European stock. In the case of salmon, many fish of Norwegian origin were bought from salmon farms by river managers to release in some Scottish rivers, so it is no surprise to find fish with Norwegian markers. At the same time, there is no knowledge as to how many wild salmon stray from Norwegian rivers to Scottish and vice versa. It is so much easier to blame salmon farms than to ask such questions.
The problem is that there is so much nonsense spoken about escapes that they are perceived to have an impact even if they don’t. The scale of the nonsense is easily apparent from a report that has just been published by Intrafish. This relates to farmed cod.
According to the report, a crew of a fishing boat noticed something abnormal about one of the cod they caught. It was deformed. In the wild such fish would have succumbed to predation very early on but if reared in a farm, the disadvantage of any deformity might be overcome. As the fish was deformed the fishing boat crew concluded it must be an escaped farmed cod.
The fish was collected by the Norwegian Directorate of Fisheries which sent samples to the Institute of Marine Research for analysis, who then concluded that the fish had been raised on a farm. Intrafish point out that there have been some escapes of fish from the resurging cod farming sector. This has prompted scientists and fishermen to become alarmed at the prospect that these escaped fish could breed with wild stocks.
What do they think is going to happen?
This is a progression of the anti-aquaculture view initially generated by the angling fraternity who blame salmon farming for the decline of wild salmon and the risk that the genetic integrity of salmon stocks in individual rivers have been compromised. How quickly the wild fish sector has forgotten the £1 million they spent on the FASMOP project trying to prove that salmon in different Scottish rivers were genetically different. At the time, they blamed the technique for the failure to get the results they wanted, even though one might have thought that they would have tested the technique ahead of time. Equally, if it was just the technique, then why ten years on, has there been no attempt to repeat the project, albeit on a smaller scale.
Farmed cod are just wild cod held on a farm. They are the same fish. What do Norwegian scientists believe will happen if a handful of farmed cod should breed with fish not held in a pen? I suspect they see the prospect of future research grants to investigate such interactions and as for the fishermen – undoubtedly, they see competition for their markets!