The Norwegian salmon lice expert group does not evaluate the lice models that form the knowledge base for the traffic light regulation system.
Frank Nilsen, professor of fish health at the University of Bergen and deputy chair of the expert group, is the first expert witness in the ongoing trial between 25 Western Norwegian salmon farmers and the government. Nilsen was questioned on Friday morning by two of the lawyers, Trond Hatland and Hilde Lund.
The plaintiff’s lawyer Trond Hatland wondersed if the expert group has evaluated the models that form the assessment basis for the production areas.
“We have not evaluated the models beyond the knowledge of the group’s members. We are assessing whether data points are in the same direction or not, and how much uncertainty in predictions these give,” Nilsen replied.
Uncertainty
“If the conclusions of the data sources we have point in the same direction, we have increased security, if the data sources point in different directions there is increased uncertainty,” he added.
“Can you go in and test the models yourself?” asked Hatland.
“It is not in the mandate we have to test these models. There is a representative in the expert group from each of these modelers in the institutions, the Institute of Marine Research (IMR), the Veterinary Institute and Sintef, and we ask them if there is uncertainty. We as an expert group have no responsibility for the models working. We will not generate any research in the export group,” said Nilsen.
“We have some cases where the models have shown different results, so we discuss it,” said Nilsen, referring to the following example: “In 2018, the models showed that there were more lice than field data showed. That year, there was long and late snowmelt that resulted in lower salinity in the fjords. None of the models took that into account,” he added.
Infection pressure
“In practice, it is impossible to continuously measure the level of lice on wild salmon. Practically all lice are from salmon in fish farms, so we use this to measure infection pressure,” said Nilsen.
Hatland then asked about the models’ migration profile for the wild salmon smolt, including the time of migration and how this is distributed over 40 days.
“Is it even, is it in groups, or as distributed as a Gaussian curve?” asked Hatland.
“I can’t remember what this profile is like. But it’s important to see migration time out of the fjord, and there’s a lot of data,” replied Nilsen.
“In terms of the representativeness of observation data, what does it take for observations by trawl to be good enough?,” asked Hatland.
The IMR is using small shrimp trawlers to find samples of wild smolt, to find sea lice on them.
“We assess all findings that have been made. If you have zero fish, you can’t emphasise it. Few fish have less value than more fish, of course,” said Nilsen.
“Have you used statisticians to create a table for representativeness in trawl finds? Number of fish per trawl?”
“No, we haven’t made it. We use all the data that is there. If you have a few observations, it is less emphasised than if one has more fish.”
Scientific basis
“The lawyers of the counterpart believe that the scientific basis is “really bad”, were the words used, do you agree with that?” asked lawyer Hilde Lund.
“No, we do not believe that the scientific foundation is bad. What common people think is a different matter,” Nilsen replied, visibly annoyed. “We don’t do research here. The research we use is from others, and our mandate is to use this.”
“The mandate you have to measure lice on wild salmon. Sea trout are not part of it. Why do you still use sea trout as a measure of lice?” asked Lund.
“The number of lice on sea trout says something about the pressure of infection. When looking at the stage of lice on sea trout, you can look at the amount and when it becomes infected,” he replied.
“From what you have heard from your counterpart so far, is there any indication that you have been wrong?” asked Lund.
“No, I have not seen anything so far that allows us to change our conclusions. Sea lice are actually capable of killing salmon,” said Nilsen. “It has been proven in the laboratory several times. Those who have more than 0.3 lice per fish will die. It will happen to wild fish, for it will not be deloused as farmed salmon will be.”
“Smolts are very vulnerable when they go out [from the rivers]. It’s small and it won’t take much more lice before it dies. When there are 300-400 million farmed salmon around, it does not take much until there are large amounts of lice,” Nilsen concluded.
