Archives

November 18, 2015

How many turkeys do we kill for Thanksgiving?

March 16, 2015

The fish we kill to feed the fish we eat

June 23, 2014

Do you know someone who buys meat only from a small local farm?

April 30, 2014

The forgotten mothers of the chickens we eat

September 10, 2013

A child raised to weigh 500 pounds by age 10?

July 23, 2013

Factory farming and the price of meat

March 27, 2013

Animals we use and abuse for food we do not eat

February 13, 2013

More cows harmed for cheese than for milk

November 20, 2012

Meat consumption and demand both in decline

October 9, 2012

Electoral politics and the meat industry

August 23, 2012

Meat consumption patterns by race and gender

July 12, 2012

Look who is talking animal welfare!

May 7, 2012

Meat industry advertising

February 6, 2012

How many animals does a vegetarian save?

January 9, 2012

The paradoxical increase in the use of primates in research

October 24, 2011

Is Vegan Outreach right about how many animals suffer to death?

September 1, 2011

The pain of animals used in research

August 5, 2011

Do pathogens in animal products cause deadlier illnesses?

July 19, 2011

Where is the cow whose milk we drink?

July 12, 2011

Vegan titles poised to overtake vegetarian?

July 11, 2011

Animal rights as contentious as the Israeli-Palestinian conflict?

“Not everything that can be counted counts, and not everything that counts can be counted.”

— William Bruce Cameron, 1963

The fish we kill to feed the fish we eat

by Harish

Posted on March 16, 2015

The majority of the fish we eat are not ones caught in the wild. They are farmed—confined and raised in overcrowded indoor tanks or sea cages—under the more decorous name of aquaculture. Yet, we kill billions of wild fish. Some of them end up on our dinner plates. But, most of the wild fish we kill die at our hands because we want to turn them into feed for the caged fish we eat.

Just how many wild fish do we Americans kill to feed the farmed fish and shellfish we eat? For several reasons, obtaining an answer to this question is not straightforward. Firstly, government agencies report on the fish and shellfish we capture in the heartless metric of tonnages and not in the numbers of individual animals. Secondly, the complexity and the variety of processing methodologies at fisheries renders elusive any hope for precision. Finally, the lack of international trade data that distinguishes between aquacultured and wild-caught animals forces assumptions which, even when they appear reasonable, cannot be confirmed as being true.

But, with the caveat that this will always be a work in progress, let’s try anyway, beginning with a look at our per capita consumption of sea animals in 2013. In this post, the animals who become what is commonly called “seafood” will be called sea animals.

Each year, the National Fisheries Institute, using raw data from the Fisheries Statistics Division of the National Marine Fisheries Service, releases data on per capita consumption of sea animals by United States residents. This data is based on the total US supply of fish and shellfish calculated as total domestic landings (caught in the seas and brought to domestic shores) plus our imports minus our exports. However, since it does not include US-produced aquaculture, it slightly underestimates the total US supply and the per capita consumption.

US-produced aquaculture data are not yet released for 2013, but we can reach a reasonable approximation using the 2012 data from the Fisheries of the United States (2013) report released by the National Marine Fisheries Service. With this adjustment to the data from the National Fisheries Institute, in the bar graph below, I plot the estimated per capita consumption in the U.S. in 2013 of the edible weight of fish and shellfish.

Shrimp, salmon, and tuna, in that order, are the most consumed sea animals in the United States. Most of the sea animals we eat besides those in the bar graph above are sardines, oysters, scallops and lobsters. Of all these animals we eat, cod, pollock, tuna, and lobsters are rarely aquacultured—they are mostly caught in the wild in the oceans. Mollusks such as clams, oysters, and scallops are often products of aquaculture, but they do not feed on wild fish; they are fed planktonic organisms such as algae and microscopic bacteria. That leaves only six major categories of sea animals we eat which are aquacultured using wild fish products in their feed: shrimp, salmon, tilapia, pangasius, catfish and crabs.

The live weight of aquacultured animals we eat

Almost all the fisheries data is reported in tonnages of live weight of animals captured or raised—that is, the weight of the whole animal while alive. The edible weight of an animal, in the above bar graph, is just the weight of what we actually eat. It does not include the parts of an animal we do not eat—such as the shell, the head or the legs of a shrimp; or the head, the skin, the guts or the bones of many fish.

In counting the fish we kill to feed the fish we eat, therefore, we have to first convert the edible weight of the animals we eat into their equivalent live weight. Admittedly, there is a significant element of approximation involved in the process of deriving these conversion factors, primarily because of variations by region, year and processing methodologies in addition to the variations by type of animal and type of preparation. In the following, the conversion factors I use are gleaned from four different organizations which have done the exhaustive work of compiling this data—the Food and Agriculture Organization of the United Nations (FAO), the European Market Observatory for Fisheries and Aquaculture (EUMOFA), the National Marine Fisheries Service (NMFS) and the EU Fish Processors and Traders Association.

Since the conversion factors depend on both the type of animal and the type of preparation, one has to incorporate into these numbers a wide array of data such as the percentage of edible products sold in canned form, the percentage sold as fresh, frozen or chilled fillets and steaks, the percentage eaten as sticks and portions, and the percentage sold or eaten cured by smoking, salting or drying. Considering these percentages and using data on conversion factors from the organizations mentioned earlier, Table 1 below lists the equivalent live weight of the aquacultured sea animals we consumed in 2013.

The first column in the table is the edible weight consumed (in 1,000 lbs.) of each category of animals by the U.S. civilian resident population. The population size assumed is 314,886,749 based on the US census estimate for July 1, 2013. For example, the per capita consumption of 2.772 lbs. of edible weight of salmon translates to 2.772 × 314,886,749 ≈ 872,861,000 lbs. of edible weight of salmon consumed in the U.S. (Because of rounding, the results of additions and multiplications reported in this post may not be exact.)

Of course, some fish we eat are indeed wild-caught and not aquacultured. While we know that 87% of the sea animals we eat is imported and that over half of it is aquacultured, the reported trade data do not distinguish between farmed and wild-caught animals; so, we cannot know for sure exactly how much of each type of sea animal we eat is aquacultured. However, given that such a vast majority of the sea animals we eat is imported, the percentage of worldwide production that is aquacultured offers a reasonable estimate of the percentage of aquacultured animals in our diet of sea animals. Based on this reasoning, Table 1 above includes the percentage of total consumption that is a product of aquaculture using data on the total global fisheries production from worldwide catches and from aquaculture as reported in the FAO Yearbook on Fishery and Aquaculture Statistics (2012).

From the estimates in Table 1 above, about 5.5 billion pounds (in live weight) of aquacultured sea animals were consumed by Americans in 2013.

The live weight of wild fish killed for aquaculture

Wild fish caught at sea are processed into fishmeal (ground up dried fish) and fish oil, both of which are ingredients in the feed fed to aquacultured animals. Diets for different types of aquacultured animals require different percentages of fishmeal and fish oil in them. Collecting this information from fisheries across the world, a comprehensive technical paper by the FAO on feed ingredients used in aquaculture reports on both historical and projected amounts of fishmeal and fish oil required for various aquacultured species. Using the 2010 numbers reported in the paper, which are less speculative than its 2015 numbers, Table 2 below computes and lists the amounts of fishmeal and fish oil required to feed the aquacultured animals eaten by US residents in 2013.

Let’s use salmon as an example to explain the entries in Table 2 above. US residents consumed about 1,295 million pounds of aquacultured salmon (in live weight) in 2013. Since the feed conversion ratio for salmon is 1.3, the salmon were fed 1,295 × 1.3 ≈ 1,683 million pounds of feed. From the table, since fishmeal comprises 22% of the feed in the diet of salmon, it required a total of 1,683 × 0.22 ≈ 370 million pounds of fishmeal to feed the salmon. Similarly, since fish oil comprises 12% of the feed in the diet of salmon, it required a total of 1,683 × 0.12 ≈ 202 million pounds of fish oil to feed the salmon.

The sea animals killed to feed the pigs and chickens we eat

It may surprise many that we actually also feed wild-caught fish to the pigs and chickens we eat! In fact, according to the International Fishmeal and Fish Oil Organisation, 20% of the world’s production of fishmeal in 2010 was used to feed weaning pigs, and an additional 5% was used to feed day-old chicks in the poultry industry. According to the Food and Agriculture Organization of the United Nations (FAO), the worldwide production of fishmeal in 2012 was 13.5 billion pounds. We can, therefore, estimate that 2.691 billion pounds (≈ 20% of 13.5 billion) of fishmeal was fed to pigs worldwide, and another 673 million pounds (≈ 5% of 13.5 billion) was fed to chickens worldwide.

Combining USDA data on pigs and chickens in my post on the number of animals saved by a vegetarian with FAOSTAT data from the Statistics Division of the FAO, we know that 8.06% of the worldwide production of pig meat enters the US supply for consumption by US residents. Similarly, 14.61% of the worldwide production of chicken meat enters the US supply. Yes, we Americans eat more than our share of pigs and chickens!

Given the above, we can estimate that a total of 217 million pounds (≈8.06% of 2.691 billion pounds) of fishmeal is used each year to feed the pigs eaten by Americans. Also, a total of 98 million pounds (≈14.61% of 673 million pounds) of fishmeal is used each year to feed the chickens eaten by Americans. Table 3 below lists these along with the fishmeal and fish oil we feed to the aquacultured animals we eat.

According to a detailed analysis of worldwide fishmeal and fish oil yields from wild fish, 100 pounds of wild fish yields between 20 and 25 pounds of fishmeal (we will assume an average yield of 22.5%). According to the same analysis, 100 pounds of wild fish yields between 4 and 5 pounds of fish oil (we will assume an average yield of 4.5%). Based on these yield percentages, Table 3 also computes the demand placed on wild fish by US consumption of each category of animals.

Let me briefly explain the last row in Table 3. We know from the table that US consumption of pigs, chickens, and aquacultured animals demands a total of about 1,269 million pounds of fishmeal and 252 million pounds of fish oil. At a 22.5% yield of fishmeal from wild fish, the need for fishmeal corresponds to a demand of about 1,269 × 100 / 22.5 ≈ 5,641 million pounds of wild fish. At a 4.5% yield of fish oil from wild fish, the need for fish oil corresponds to a demand of about 252 × 100 / 4.5 ≈ 5,602 million pounds of wild fish. So, the derived demand (from American consumption of all animals) for the capture and death of wild fish for fishmeal slightly exceeds that for fish oil. Since the same wild fish can be used to produce both fishmeal and fish oil, the total demand is the larger of these two demands—American consumption of farmed animal meat, therefore, demands the capture and death of 5,641 million pounds of wild-caught fish!

Note that the entries in the last column in Table 3 do not add up to the total. This is because the same fish can be used to produce fishmeal for pigs and fish oil for salmon—we cannot simply add up the demands. The total demand for wild fish has to be derived from the total demand for fishmeal and fish oil.

Who are we grinding down into fishmeal and fish oil?

A disturbingly indiscriminate variety of sea animals—small and large, managed and unmanaged—are described as “feed fish” and caught to be turned into fishmeal and fish oil. They include, among others, anchovies, capelins, hake, herrings, mackerels, krill, menhadens, sardines, and sand eels. In Table 4 below, I list the tonnages of the majority of these species of sea animals captured in 2012 from the FAO Yearbook on Fishery and Aquaculture Statistics (2012) (in this post, I consider only the species for which the capture was more than 150 million metric tonnes). Some of these species are more popular for direct human consumption than others. But, in the following, for lack of better data, we will assume that the percentage of the catch of a species used for direct human consumption is approximately the same for all “feed fish” species. This allows an estimation of the mean weight of the individual animals used as “feed fish”.

The estimated mean weights of individuals of these species in this table primarily come from this report on the welfare of wild-caught fish. In some cases, they come from other sources (such as this paper for Antarctic krill and this paper for jumbo flying squid).

Fishmeal and fish oil produced from fish residues

Recently, a steadily rising percentage of fishmeal is being produced from trimmings, heads, tails, bones and other offal from fish processed for human consumption—the by-products of fisheries—instead of from whole fish caught at sea. According to the FAO’s annual report, State of World Fisheries and Aquaculture (2014), about 35% of fishmeal production in 2012 was from such by-products. According to a recent issue of the monthly newsletter, Update, from the International Fishmeal and Fish Oil Organization, a similar percentage (30%) of fish oil is also being produced from residue at fisheries which process fish for human consumption.

In weighing the total impact of American consumers on wild fish, there is an important question we have to answer: Should the use in aquaculture of such by-products count as contributing to the pressure on wild fish populations? Or, instead, should we ignore their use altogether because these by-products would be thrown away anyway if not fed to the farmed fish we eat?

The use of any by-product of wild-caught fish—whether caught for human consumption or not and whether caught for American consumption or not—supports and sustains the capture of wild fish. For example, veal is often described as a by-product of the dairy industry. Even though the veal industry may not exist if the dairy industry did not, we rightly hold both the dairy consumer and the veal consumer partially responsible for the harm done to cows and calves.

One may argue that when a by-product is of low economic value, our ethical responsibility in consuming it is lower than it would be if the economic value was high. For example, it is likely true that those who eat the tenderloins of a pig create a larger demand for the slaughter of pigs than those who eat the lower-valued head of a pig. But, these demands are linked from the animal’s point of view—the pig would protest just as hard whether you ask for her head or her tenderloins; after all, they both require her slaughter. Economic prices may track demand or value but they do so only from a human-centric point of view, if at all. In general, divvying up ethical responsibility based on economic arguments alone is a fraught affair.

In this post, I will assume a simpler ground rule: if I consume, directly or indirectly, x percent of the live weight of an animal, I am x percent responsible for the death of that animal. Therefore, whether whole or as by-products, American consumers of pigs, chickens, and aquacultured animals have had a causal role in the capture and death of over 5,641 million pounds of wild fish in 2013.

Considering the impact of aquaculture alone, American consumers of aquacultured animals have had a causal role in the capture and death of 5,602 million pounds of wild fish in 2013. A key number in this context is the fish-in-fish-out (FIFO) ratio, defined as the number of pounds of wild fish (in live weight) needed to produce the feed required to grow one pound (in live weight) of aquacultured animals. Since we require about 5,602 million pounds of wild fish to feed and grow the 5,522 million pounds of aquacultured animals eaten by us, the FIFO ratio that applies to American consumption of aquacultured animals with fish input is 5,602 / 5,522 ≈ 1.01.

We cannot absolve ourselves of all ethical liability toward wild fish when we eat farmed fish simply because the farmed fish were fed by-products of wild fish eaten by others and not whole wild fish. But, the aquaculture industry, especially the International Fishmeal and Fish Oil Organization, routinely quotes FIFO numbers that ignore these wild fish entirely in a self-serving attempt to wipe away all ethical responsibility for their capture and death. That brings us to the common mistakes—sometimes likely deliberate—made by the industry and activists alike in calculating or using the FIFO ratio.

Common mistakes in computing or using the FIFO ratio

The impact of aquaculture on wild fish populations is not easily reduced to a single number such as the fish-in-fish-out (FIFO) ratio. Computing this number requires many uncomfortable approximations and, not surprisingly, different people with different interests promote different numbers for the FIFO ratios. An approximation is not a mistake—but, in the debate on the FIFO ratio, there have also been many fallacious computations of the FIFO ratio and some trumpeting of convenient but incorrect implications. I will address these common mistakes made in the computation of the FIFO ratios before proceeding to compute the number of wild fish killed for aquaculture feed.

Common mistake #1: The first and most common mistake made in computing the FIFO ratio is to apply the same global aquaculture production numbers to all regions of the world, including the United States, without recognizing that American consumption patterns are far more oppressive on wild fish than global consumption patterns. For example, according to the FAO Yearbook on Fishery and Aquaculture Statistics (2012), global aquaculture production of fish, mollusks, and crustaceans in 2012 was 66.6 million metric tonnes, while about 21.7 million metric tonnes of fishery production was used for the manufacture of fishmeal and fish oil. This suggests a global FIFO ratio of 21.7 / 66.6 ≈ 0.33 for all of aquaculture. But, it is incorrect to assume this number for the aquacultured animals that American consumers eat. American consumption tends to be higher in salmon and shrimp (the two most popular sea animals eaten in America) which are more oppressive on wild animals than most other sea animals eaten across the world.

Common mistake #2: The aquaculture industry prefers to quote small FIFO numbers. So, they tend to include in their FIFO calculations even the production that does not generally require wild fish products in their feed. In 2012, as much as 23% of global aquaculture production was mollusks which feed on microscopic planktonic organisms and do not need fishmeal or fish oil in their diets. Including them in the FIFO calculations leaves the false impression that the FIFO numbers for aquacultured fish with wild fish input are smaller than they actually are. The fish-in-fish-out numbers are meaningful only when quoted specifically for the production of only those animals which require fish input in their feed.

Common mistake #3: A mistake made frequently by opponents of aquaculture is to not recognize that FIFO ratios vary significantly by species. While salmon consumption is the most oppressive on wild fish populations, it is not correct to use the FIFO ratio for salmon for all of aquaculture. This is the flip side of the common mistake #2.

Common mistake #4: It was frequently the case that FIFO numbers were calculated and quoted separately for each individual species of the aquacultured animals. There is nothing wrong with doing so—the error is in using these numbers to add up the total wild fish used by a collection of aquacultured species. This requires a deeper explanation and so, let’s take a step back and examine, using salmon as an example, how FIFO ratios have traditionally been computed.

As mentioned earlier in this post, the yield of fishmeal (FM) from wild fish is about 22.5% (let’s call this number FM-Yield) and the yield of fish oil (FO) from wild fish is about 4.5% (let’s call this number FO-Yield). The feed conversion ratio (FCR) for salmon is 1.3—this means that producing 100 pounds of salmon requires 130 pounds of feed. If salmon diets contain 22% fishmeal (let’s call this number FM-in-Diet) and 12% fish oil (let’s call this number FO-in-Diet), it means that producing 100 pounds of salmon requires 130 × 22% = 28.6 pounds of fishmeal and 130 × 12% = 15.6 pounds of fish oil. Now, 28.6 pounds of fishmeal can be produced from 28.6 × 100 / 22.5 ≈ 127 pounds of wild fish. Similarly, 15.6 pounds of fish oil can be produced from 15.6 × 100 / 4.5 ≈ 347 pounds of wild fish. So, at a minimum, producing 100 pounds of salmon would require 347 pounds of wild fish, giving us a FIFO ratio of 3.47. This was the method used in a widely cited paper, published in 2008, by Tacon and Metian. Basically, the formula for obtaining the FIFO ratio is:

FIFO = FCR × max {	FM-in-Diet	,	FO-in-Diet	}
	FM-Yield		FO-Yield

The above method of computing the FIFO ratio is disputed by some industry groups such as the International Fishmeal and Fish Oil Organisation (IFFO). They argue, I think legitimately, that the 347 pounds of wild fish in the example above would produce about 347 × 22.5% ≈ 78 pounds of fishmeal of which only 28.6 pounds will be used for feeding the aquacultured salmon. The remaining 49.4 pounds of fishmeal are unused and assumed wasted in the above computation when, in reality, they do get used up by other units of the aquaculture industry. So, they argue that the above computation exaggerates the FIFO ratios for salmon.

While there is merit in the above argument made by the industry, the real error is not in the computation of the FIFO ratios by the above formula but in how the computed FIFO numbers are used. For example, if salmon has a FIFO ratio of 3 and shrimp has a FIFO ratio of 2, we cannot simply add up the wild fish killed for each to say that producing 1 pound of salmon and 1 pound of shrimp will require a total of 5 pounds of wild fish (because the unused fishmeal from the 3 pounds of wild fish used to make fish oil for salmon could be used for feeding the shrimp). This is why, in this post, I have avoided relying on FIFO numbers for individual species.

Common mistake #5: To suppress the FIFO numbers for aquaculture in general and salmon in particular, the industry has sometimes promoted flawed formulas even when the driving concern behind it is legitimate. For example, in this paper widely and enthusiastically cited by the industry, the author promotes a new formula to compute the FIFO ratio for individual species. This manufactured formula tries to address the common mistake #4 above by artificially suppressing the FIFO ratios for species with the highest FIFO ratios. This makes for better publicity for aquaculture—no one species (think salmon) looks scarily oppressive on wild fish and makes less likely that people will misquote a large FIFO number and apply it across all aquacultured species. While this formula promoted by the industry does indeed give the correct answer for the FIFO ratio corresponding to a collection of species, it is a flawed solution for a legitimate problem. It misrepresents the actual derived demand for wild fish generated by the consumption of species such as salmon. The traditional formula for calculating the FIFO ratio serves well as an indicator of the derived demand for the capture and death of wild fish—there is no need to change the formula; we just have to make sure to use the result of the formula correctly without making the common mistake #4 above.

Common mistake #6: As explained earlier in this post, quoting FIFO ratios based exclusively on the use of whole fish for fishmeal and fish oil ignores the fact that consumption of by-products also places a certain demand on wild fish. They distort and under-represent the actual impact of aquaculture on wild fish populations.

In this post, I believe I have avoided the above common mistakes. If I made a new different mistake, I hope my few readers who have gotten this far will enlighten me.

The number of wild fish killed

As computed in Table 3, American consumption of farmed animals requires about 1,269 million pounds of fishmeal and 252 million pounds of fish oil each year. Of these amounts, the portion that is produced from whole wild fish has a different impact on wild fish than the portion that is produced from by-products because of the different sets of wild sea animals used in the production.

When whole sea animals are used to produce fishmeal and fish oil, it is from animals we usually describe as feed fish. To convert the tonnage of these animals into the number of animals killed for feed, we need the mean weight of the individuals of these species. From the last line in Table 4, this mean weight lies between 0.0474 and 0.0926 pounds.

Fishmeal and fish oil produced from fishery residues come from wild sea animals who are processed for human consumption. It used to be difficult to process small pelagic fish—used traditionally only for fishmeal and fish oil—for direct human consumption. But, as made clear in a technical paper by the FAO on feed inputs for aquaculture, several factors have led to an increase in the use of traditional “feed-fish” species for direct human consumption. In fact, except for the very smallest of them such as the Antarctic krill, about 10% of the “feed fish” are used for direct human consumption. As a result, fishmeal and fish oil produced from residues come from most of the species of sea animals we catch—the sea animals we eat directly and 10% of the “feed fish.” Combining weight estimates with data from the FAO Yearbook on sea animals captured in 2012, we find that the mean weight of these individuals lies between 0.17 and 0.51 pounds.

In considering the sea animals whose by-products end up as aquaculture feed to serve American consumption, we do not wish to double count them if we are already counting them as being eaten directly by American consumers. To this end, we should subtract the portion of the catch of these sea animals who are consumed directly by the US resident civilian population. We know, from the calculations detailed here and from the FAO Yearbook, that about 4% of the worldwide capture of sea animals for direct human consumption is consumed by Americans. In counting the animals who are turned into fishmeal and fish oil for the American diet, we should only count the 96% used for non-American consumption since the remaining 4% is already counted as having been directly consumed by Americans. Given all these facts, the following table computes the total impact of American consumption on wild sea animals. (These numbers do not include the sea animals we unintentionally catch and throw away dead or dying—they are addressed in my post on the number of animals saved by a vegetarian.)

As seen from the last two rows of Table 5 above, American consumption of farmed animals places a slightly higher demand on wild sea animals for its derived demand for fish oil than for fishmeal. From Table 3, all of the demand for fish oil comes from our consumption of aquacultured animals and not from our consumption of pigs and chickens. American consumption of aquacultured fish and shrimp, therefore, demands the capture and death of 45 to 92 billion wild-caught sea animals each year!

Of the sea animals typically processed to produce aquaculture feed and which are captured in large numbers worldwide (more than 150,000 metric tonnes per year), the following pie chart depicts the proportions of their numbers. The majority of them are small animals such as Peruvian anchovies and the Antarctic krill.

Assuming a US resident civilian population of 314,886,749 in 2013, the number of wild sea animals captured and killed to feed the aquacultured animals eaten by the average US consumer is between 144 and 293 per year. Parts of some of these animals are also used to feed pigs and chickens eaten by the average American.

When we eat animals, we do so in the knowledge that the animals we eat were killed so we could eat them. We meet them on our dinner plates or between slices of bread—dead, but we meet them nevertheless. These 144 to 293 sea animals, however, die in service of someone’s appetite who rarely, if ever, learns of their existence. They die unseen, unheard, and unknown by those of us who cause their deaths.

py3.9.4 (default, Apr 5 2021, 09:56:39) [GCC 7.5.0]Django(3, 2, 1, 'final', 0)