Category Archives: Mathematics

Calling shenanigans — again — on Morton’s junk graph

Alexandra Morton has decided to make 2014 the year of the Salmon Food Scare, and she’s trying as hard as she can to stir up the masses using half-truths and emotion.

She’s no doubt trying to recreate the panic from 2004, when one highly-publicized study was used to suggest farmed salmon was dangerous to human health. Ah the good old days, eh Alex? When the gullible masses would believe anything you said without checking your claims? Must be tough these days, judging by the “DONATE NOW” button on every one of your websites. 

For a good overview of how that 2004 panic all shook out, including an enlightening look at the millions of dollars invested in slagging farmed salmon, we recommend you read this opinion piece by Vivian Krause.  

Oh, and by the way, there’s no reason to be concerned about dioxins or PCBs in any food sold in North America. None.

Eat a balanced diet, get some exercise and be excellent to each other and you’re shiny.

But that’s not stopping Morton, who knows how to play on people’s fears like a cheap fiddle.

A while ago we blogged about a ridiculous graph she is using to try and claim that farmed salmon contains dangerous levels of dioxins.

Apparently, ashamed to have been caught in the act of deliberately misrepresenting the facts, she has revised this graph with an explanation in the fine print on her new website. She says:

Please note the source data for salmon is provided in pg/g wet weight, while the other values were provided in pg/g fat. NIFES reports farmed salmon is 15.6% fat and so the conversion to pg/g fat = 41.6

Yeah, no. This is just stupid, Alex.

Wet weight is a common distinction made in weighing fish, because so much fish sold is smoked, cured, salted or dried. Wet weight is merely a measurement of fish with the water content in, i.e. before being processed. The non-fish products on this list are always weighed with their water content included: their default measurement is wet weight.

What you’re doing here is like assuming that if your car can go 160 km/h before the governor kicks in, and the speed limit is 80 km/h, it should only take you half an hour to get where you’re going.

But since you aren’t inclined to provide people with facts, we’ll do it for you.

Here’s a comparison of the limits on dioxins in food set by the EU, versus the amounts that are actually in said food.

dioxin_limits_vs_actual

Facts hurt, don’t they Alex. There’s no reason to avoid any of these foods because of dioxins.

Want to check our math? Feel free. Here’s the data table we used to make this graph, complete with comprehensive citations.. Which is more than activists like Morton will give you.

Oh, and by the way, if the anonymity on our blog bothers you? Feel free to check all the links in the document above and retrace our steps. Thinking for yourself: does a body good.

Activist math, contaminants and the art of fear

We’ve been sitting on this one for a while because we wanted to see how far it would go. Apparently, our favourite scaremongerer is pretty serious about it so we figured it was time to expose this.

In a poor attempt to scare people about food contamination, Alexandra Morton has created a graph that people might assume shows that farmed salmon is contaminated with PCBs.

Mortoncantdomathgraph

Oooh, scary, right? Two things though.

1) These numbers, as shown in the EU regulation cited by Morton as the source of information for this graph, DO NOT represent ACTUAL amounts of dioxins and PCBs in ANYTHING. They are limits set by the EU on what is a safe amount in those foods. Morton neglects to point this out, giving a false impression.

2) This “41.6” number on Morton’s graph comes out of thin air. The actual limit for salmon and all other fish, as set in the EU document source for the other numbers on the graph, is 6.5.

Check it for yourself, it’s on page 4.

Because 6.5 doesn’t have the sort of shock value Morton was hoping for, she did some of her own math and came up with a greater number, conveniently forgetting to show where in the world this number comes from, and also neglecting to explain that these numbers represent the limits set by the EU, not actual test results.

But that’s all boring, right? Who cares, Morton raises a valid point, a scary story about how farmed salmon are more contaminated, right? Who cares that she just made up a number and misrepresented what her source actually says, she’s just getting the truth out there right? We’re sure someone will comment here saying something like that.

That’s the art of fear in action, and Morton is damn good at it. She starts with a scary story, and then manipulates data to make it look like science is on her side. By the time people like us come along and pull back the curtain, it doesn’t matter because people really want to believe in Oz.

People believe in stories, not facts.

But we’ll keep bringing our readers the facts, in the hopes that they will learn how to pull apart these anti-salmon farming stories and see that they are something much more vulgar: manipulative lies.

The true salmon feedlots

Conventional salmon farms versus land-based salmon farms in density

There’s one more point we wanted to make about on-land salmon farms as a follow-up to our last post.

Critics of salmon farms love to use the term “feedlot” to describe ocean farms. In their minds this brings a negative connotation to salmon farms (even though science is showing that feedlots may actually be more “green” than finishing beef on grass).

They try and paint word pictures of cramped quarters where poor farmed fish can barely move.

Yet these same people turn around and promote land-based salmon farms as the best possible way to raise salmon.

Either they’ve never bothered to research the numbers, or they are lying when they say they care about fish welfare. Because the latest land-based salmon farm, currently under construction near Port McNeill, will be operating with fish in their tanks at a density of 90 kilograms per cubic metre.

Based on our conversations with the different farming companies, a conventional ocean salmon farm operates with fish in the nets at a density of about 18 kilograms per cubic metre.

This land-based farm will have fish in a density equivalent to 2.6 salmon in a bathtub.

A conventional ocean farm has fish in a density equivalent to 0.5 salmon in a bathtub.

Which one is the true feedlot?

A true salmon feedlot.
A true salmon feedlot. And the linked article makes a lot of assumptions to suggest that this is profitable! See DFO’s feasibility study for a thorough economic comparison.

One last thought: if animal welfare is truly your concern, take a look at this report from 2007, titled “Closed Waters: The Welfare of Farmed Atlantic Salmon, Rainbow Trout, Atlantic Cod and Atlantic Halibut.” The report suggests, based on an analysis of several different studies, that the top density for farmed Atlantic salmon in ocean farms is 22 kilograms per cubic metre. Salmon farmers are well below this threshold. The report also points out that salmonids in higher densities can be more aggressive, grow smaller, nip and bite at each other, are more susceptible to illness and disease, and get worn fins.

Again, which form of salmon farming is the true feedlot?

Growing awareness of aquaculture is prompting people to start calling for welfare rules for farmed fish. We agree this is necessary.

And we are confident that salmon farmers in the ocean are already above reproach when it comes to looking after the health of their livestock.

Sea lice science ain’t done yet

I don’t work on sea lice any more because I figured it out. Where there’s fish farms, there’s sea lice. It’s an extremely easy thing to study, way easier than whales.

— Alexandra Morton, author and co-author of numerous sea lice studies, speaking at the Cohen Commission, Sept. 7, 2011

The science is never done, even if some people think they’ve “figured it out.” And it’s a good thing researchers haven’t given up on researching sea lice in B.C. because we apparently still have an awful lot still to learn.

Six UBC scientists have just published an interesting paper titled “Physiological consequences of the salmon louse (Lepeophtheirus salmonis) on juvenile pink salmon (Oncorhynchus gorbuscha): implications for wild salmon ecology and management, and for salmon aquaculture.” It was published by the Philosophical Proceedings of the Royal Society B and is available to read online (we will be adding the PDF version to our library at a later date).

There is a lot of interesting information in this study, which looked at the real-world physiological effects of sea lice infestation on wild salmon.

For years, critics of salmon farms have claimed that sea lice from salmon farms are killing wild salmon. However, there has never been any evidence to support this claim, only sketchy predictions from mathematical modelling studies.

No one denies that salmon farms can increase the number of sea lice in an area; when wild fish pass salmon farms, the lice, a natural parasite found in the Pacific Ocean, can infest farmed salmon, which can act as a breeding ground for lice if not managed properly.

But what is up for debate is this: IF lice from farms infest tiny pink salmon passing by in the spring months, when they head out to sea, do they have any effect on wild salmon survival?

This new study basically says, “maybe, maybe not.”

But while this is inconclusive, this study makes several interesting new discoveries, and puts old information in a new light.

Shedding

It’s well-known that juvenile pink salmon can shed sea lice, dependent on a wide variety of factors in the ocean as well as their own physiology.  But this new study makes the interesting point that previous mathematical modelling studies, particularly the well-known and well-publicized 2006 sea lice study by Morton, Krkosek et al, are seriously flawed when they do not consider shedding rates.

A high rate of louse shedding is also of importance in mathematical models, where the incorporation of realistic shedding rates [67] reduces the predicted mortality of pink salmon owing to salmon louse infection by 95% relative to an earlier model where shedding was not considered [10]. Collectively, these findings of high rates of shedding of attached lice suggest that the majority of the salmon lice that successfully infect a fish will be shed before they reach motile and reproductive stages. Clearly, the interactions between salmon lice, juvenile pink salmon and their environment are extremely complex and change temporally as both salmon louse and fish develop and the implications of this need to be considered in relation to conservation efforts.

Ionoregulation

What is ionoregulation? It’s a self-regulatory process in a salmon’s body which, if it is out of balance, can kill a salmon within days or leave survivors severely crippled. Researchers in this new study decided to see if sea lice can affect this process when pink salmon are at their most vulnerable, right after they enter saltwater. The research found that sea lice can cause an imbalance, but only until the fish are about 0.5 grams, and the implications to their survival is unclear. As well, once the salmon reach 0.7 grams, they have sufficiently increased their ability to resist and shed sea lice.

Conclusions

Sea lice from salmon farms may harm wild salmon, but it’s not certain. The effects of sea lice on tiny wild salmon will certainly have some effect on them at some point, so to reduce any risks, the study recommends that salmon farmers and their managers continue doing what they are doing: voluntarily fallowing farms during outmigration periods to keep sea lice levels low. This is sound advice, and a good use of the precautionary principle.

In a recent study to investigate the benefit of fallowing fish farms in the Broughton Archipelago, it was determined that fallowing reduces salmon louse levels around the farm and reduces juvenile pink infection levels to background levels [70], indicating that this recommendation and voluntary compliance may mitigate salmon louse effects on the most sensitive stages of pink salmon.

The study closes with a very interesting footnote, which leaves us with the suggestion that in the grand scheme of things, sea lice really aren’t that big of a deal for salmon. The study’s closing paragraph is a perfect way to close this blog post:

Perhaps the only study that has truly considered the impact of salmon louse infection on juvenile salmon is the 10-year study of the return of Atlantic salmon smolts that had been treated with SLICE (emamectin benzoate) to protect them from salmon louse infection for the first 90 days of their outward migration, a period which easily extended beyond the contact with net-pen aquaculture and associated salmon lice. Remarkably, a comparison with non-treated fish revealed that protection of juveniles from salmon louse infection represented a minor component to overall marine survival. Indeed, during the 10-year study, adult Atlantic salmon returns fell a similar 10-fold in both treated and non-treated fish [71]. A similar study on juvenile pink salmon treated with SLICE in the Broughton Archipelago could be very revealing in assessing the true impact of sea lice on pink salmon fitness.

We agree. This study should be done, because the science is never truly finished.

Context counts (and sexy graphs help)

We’ve had a lot of fun with Ms. Alexandra Morton’s ridiculous use of one particular graph to claim that salmon farms in BC brought about a decline in wild salmon productivity.

Without any context, her argument sounds reasonable.

But in context, it all falls apart as we showed in an earlier post and again in this post.

Recently, DFO released its 2012 projections for Fraser River sockeye, estimating a 90 per cent probability of a maximum run size of 6.6 million.

Their predictions are summarized in a graph, including an estimate of the productivity of this year’s run, placing it just above the average.

This is, incidentally, the same graph Ms. Morton uses, but with all the context included.

Projected 2012 Fraser sockeye returns and productivity
Projected 2012 Fraser sockeye returns and productivity

See that? 2012 will likely be an average year.

What does that mean?

It means Ms. Morton’s predictions of doom and gloom are, once again, false prophecies.

Salmon runs fluctuate, and have done so ever since we started recording these numbers half a century ago.

In fact, according to recent research in Alaska, salmon runs have fluctuated for more than 2,000 years.

Evidence from Alaska lake shows natural fluctuations in sockeye population over 2000 years
Evidence from Alaska lake shows natural fluctuations in sockeye population over 2000 years

Humans have been impacting Fraser sockeye stocks and the stocks of every other kind of salmon in B.C. for thousands of years, and our impacts have increased ever since we started catching salmon in massive amounts and damaging their habitat more than a century ago.

But it seems apparent that ocean conditions have far more long-term impacts on salmon.

But Ms. Morton and anti-salmon activists don’t care about those facts. They focus on a small window in time and on one river system, asking, why did Fraser sockeye productivity decline for roughly 20 years starting in the 1990s?

It was salmon farms, they say, answering their own question and fingering the “new kid on the block” before anyone can raise any other possibilities or use science.

In contrast to their easy answers to hard questions, a scientific approach requires we look at as many possible factors as we can find.

What did cause the decline of Fraser sockeye productivity in the 1990s and 2000s?

Was it an explosion of Alaskan ranched fish entering the North Pacific feeding grounds in the early 1990s? Take a look at this graph.

Comparison of Alaska salmon releases with Fraser sockeye productivity
Comparison of Alaska salmon releases with Fraser sockeye productivity

Was it ocean temperatures? The ocean has gotten hotter since the 1950s. Interestingly, the decline in temperatures in the 1960s could be correlated with the dip in Fraser sockeye productivity at the same time. But the rise above average (0) since about 1990 can also be correlated with the decline in productivity in the 1990s and 2000s.

Ocean heat content 1955-2007
Ocean heat content 1955-2007

Was it The Pacific Decadal Oscillation? The natural fluctuation of ocean temperatures, which changes every 15-30 years? Interestingly, the brief change in the 1960s can be correlated with the decline in Fraser sockeye productivity at the same time, and the major shift from a long-term trend in the 1970s and 1980s to quick fluctuations in the 1990s and 2000s can be correlated to the long-term stability of Fraser sockeye productivity in the 1970s and 1980s, and the wild fluctuations and decline in the 1990s and 2000s.

Pacific Decadal Oscillation observed monthly values 1900-2011
Pacific Decadal Oscillation observed monthly values 1900-2011

Was it any of these things? All of these things? None of these things? We don’t know. We haven’t even included catch data in this post, that has to be considered as well. But it sure looks like ocean conditions have way more impacts than a few salmon farms ever could.

Plus, there is actually evidence to suggest that ocean conditions do have impacts (in contrast to the speculations and simplistic correlations anti-salmon farming activists use to support their claims).

But we can’t say for sure.

One thing we know for sure, however, is that, given all these factors which could affect Fraser sockeye productivity, suggesting that “salmonfarmsdidit” is a facile conclusion to the mystery, especially when we don’t even have a body.

The “corpse” of wild salmon is still swimming strong, despite all the predictions of doom and gloom, and with an average run predicted for the Fraser, exceptional runs predicted for the Alberni-Clayoquot region and good returns expected elsewhere in BC, we are confident in saying that all the evidence shows that wild and farmed salmon can coexist together in the same ocean with negligible risks to either species.

Farmed versus wild is a false choice.

Choose both.

Heavy on the graphs

We have three graphs we would like to show here.

Together, they show how important aquaculture has become, and will be, for the entire world in the near future.

The debate over salmon farming is a tempest in a teapot. Whatever happens to salmon farming in particular, aquaculture in oceans, lakes, ponds and on land is going to happen, whether anti-aquaculture and anti-salmon farming activists want it or not.

Fisheries have peaked. The world’s population keeps growing, especially in the developing world. Not surprisingly, aquaculture is experiencing massive growth in the developing world.

This is the way of the future.

Global aquaculture and capture fisheries

Global population 1984-2004

On sea lice, UFOs and faulty logic: analyzing a new sea lice study

One of our Twitter followers (thanks @farmsalmon4ever) called our attention to a new sea lice study published Jan. 4, 2012 about sea lice epidemics.

The study, titled “Critical thresholds in sea lice epidemics: evidence, sensitivity and subcritical estimation” was published in the Proceedings of the Royal Society’s Biological Sciences journal and was written by Neil Frazer, Alexandra Morton and Martin Krkosek.

We decided to take a read through and see what it says.

Morton and Krkosek have published many studies together on sea lice; their work is the basis for much of the criticism of salmon farms when it comes to sea lice. In particular, they have tried to show that salmon farms amplify levels of sea lice, which are naturally present in the ocean.

The problem, they have tried to show, is that farmed salmon provide a place where sea lice can survive and increase their numbers at times of the year when lice levels would naturally drop.  This, they then conclude, threatens passing juvenile wild salmon which have to contend with unnaturally high levels of sea lice around salmon farms.

But, as we asked in our last post, have sea lice from salmon farms harmed wild salmon?

After more than a decade of extensive research, scientists are divided, with the only real consensus being “maybe, but we’re not seeing it,” a conclusion which was recently reiterated by scientists at the Cohen Commission (see project 5).

While farms may increase the number of sea lice in a small area, scientists see no meaningful connection between farm lice levels and actual productivity of wild salmon stocks.

And doomsday prophecies that sea lice would wipe out wild salmon have been proven wrong.

So where does that leave us? Again, in the world of “maybe, but we’re not seeing anything definitive.”

Reverse-engineering epidemic data

This new paper by Frazer, Morton and Krkosek is different than many of their previous sea lice studies together because it appears to be more of an attempt to develop a tool to better understand and predict sea lice levels than an attempt to prove anything.

The paper presents a mathematical model for predicting how many farmed fish should be allowed in an area before they reach a threshold and pose a risk of sparking a sea lice epidemic.

“Two observations motivate this paper,” they write. “One is that sea lice are seldom a problem in areas with low production even when lice are present on local wild hosts. The other is that lice are seldom a problem when sea-cage aquaculture is new to an area.”

It takes several farm production cycles and increasing density of fish to spark any notable increases in sea lice, they continue.

To prove this, the authors use the examples of a sea lice epidemic in New Brunswick in 1994 and an epidemic in the Broughton Archipelago in 2001, taking the numbers from the epidemic, reverse-engineering them and calculating the threshold beyond which stocking densities in salmon farms can provoke a sea lice epidemic.

Bad logic

There are problems with this. The study doesn’t have very much data for the Broughton Archipelago. “Lice data during subcritical stocking in the area are not available because monitoring of lice began after the epidemic emerged. However, assuming that the critical stocking threshold was exceeded in 2000-2002, the critical band is estimable.”

Wait, what?

That is a BIG assumption to make. The authors are assuming that because there was an epidemic of sea lice in 2001, farms must have exceeded the critical stocking threshold for the region.

Frazer, Morton and Krkosek are assuming their theory is right based on completely circular reasoning.

  • They believe there is a “critical threshold” for the amount of farmed salmon which can be in an area.
  • They believe that beyond that threshold there will be sea lice epidemics.
  • There was a sea lice epidemic in 2001.
  • Therefore salmon farms crossed that “critical threshold.”

It’s fallacious reasoning. Using that logic, we could argue that:

  • We believe UFOs are real.
  • We believe UFOs leave crop circles.
  • There was a crop circle in the farmer’s field this morning.
  • Therefore UFOs are real.
A crop circle
This showed up in the field yesterday. Therefore UFOs are real.

Just like the UFO believer fails to consider that crop circles may have come from something other than a UFO, this paper fails to consider that a sea lice epidemic may have been prompted by something other than salmon farms.

The study also seems designed to reach the conclusions the authors were looking for.

“For the Broughton Archipelago, the Pacific Salmon Forum (PSF) recommends limiting Broughton Archipelago farmed salmon production to less than 18.5 kilotons per year,” it says.

How convenient that is just about exactly where the authors estimate the “critical threshold” production levels for the region.

Production level critical threshold

A useful tool

Before anyone thinks we are trying to dismiss this study out-of-hand, we want to point out that it actually provides quite a useful tool. It considers water temperature and salinity, which can have a huge impact on sea lice abundance, and the model could be adjusted to help farmers forecast and plan production cycles.

“If changes in temperature and salinity could be forecasted, farmers could pre-emptively harvest or treat,” the authors write. “Also, locating farms in low-saline conditions may raise threshold values and prevent epidemics.

“Thus, if good records are kept while an aquaculture industry expands in a particular region, the critical stocking level can be estimated without ever experiencing an epidemic.”

This could be very useful for salmon farmers, who now have more than a decade’s worth of sea lice data as well as salinity and temperature data and would love to be able to keep sea lice levels as low as possible.

However the authors err by assuming their model is right.

Perhaps someone could take the decade of real-world data that is out there, plug it into this mathematical model and see if it actually lines up with any observed increases in sea lice levels on wild salmon. Then we would know if it’s right or not.

Perhaps the authors should have done that themselves.