Feed Conversion Efficiency in Aquaculture: Do We Measure It Correctly?

Supplementary Information for

Feed conversion efficiency in aquaculture: Do we measure it correctly?

Jillian P. Fry1,2,3*, Nicholas A. Mailloux1, David C. Love1,2, Michael C. Milli1, Ling Cao4,5

1 Johns Hopkins Center for a Livable Future, Johns Hopkins University, 615 N. Wolfe Street, Baltimore, Maryland, USA

2 Department of Environmental Health and Engineering, Bloomberg School of Public Health, Johns Hopkins University, 615 N. Wolfe Street, Baltimore, Maryland, USA

3 Department of Health, Behavior and Society, Bloomberg School of Public Health, Johns Hopkins University, 624 N. Broadway, Baltimore, Maryland, USA

4 Center on Food Security and the Environment, Stanford University, 616 Serra St, Stanford, California, USA

5Institute of Oceanography, Shanghai Jiao Tong University, Shanghai, China

* Correspondence to: Jillian P. Fry, PhD, MPH, 615 N. Wolfe Street, W7010, Baltimore, MD 21205, , 410-502-5069

Contents:

1. Species selection criteria

2. Data extraction for aquatic and terrestrial animal inputs and outputs

3. Protein and calorie retention calculations

4. Supplementary tables S1-S5

1

1. Species selection criteria

We extracted global aquaculture production data for 2009 through 2013 using United Nations Food and Agriculture Organization (FAO)FishStatJ data software [1]. We identified the top five aquaculture species groups produced globally, which were carps, tilapias, shrimps, catfishes, and salmonids. Four of the five species groups (catfish, salmonids, shrimp, tilapia) contain fish or crustacean species that are among the most commonly consumed seafood in the U.S. [2], and carp is the most widely farmed fish species group in the world [1]. The total production of these five species groups averaged 37.8 million metric tons (MMT), accounting for 60 percent of global aquaculture production per year from 2009 through 2013 (including fed and unfed aquaculture, excluding cultivation of aquatic plants) (Table S2). Within each species group, we selected the two top-produced species per group for analyses (Table S3), excludingspecies that were largely non-fed (e.g., filterfeeders such as silver carp, bighead carp, etc.). Species group information was used for tilapias because species-specific information was not available. Nile tilapia comprise over 71 percent of global tilapia aquaculture production, andmost other tilapias are classified by FAO as “nei,” or “not elsewhere included.” The final list of aquatic species included in the study was: common carp, grass carp, channel catfish, pangas catfish, Atlantic salmon, rainbow trout, giant tiger prawn, whiteleg shrimp, and tilapia. The top three terrestrial animal species produced for meat in the U.S. and globally are cattle, chicken, and pigs[3, 4]; these species were included in our analysis for comparison.

2. Data extraction for aquatic and terrestrial animal inputs and outputs

We collected informationfor each species including feed conversion ratio (FCR), edible portion, feed composition, and nutrition content of edible flesh.When species-specific information was unavailable for aquatic animals, we collected data at the species group level.Below we describe each data element used in the retention equations.

Feed conversion ratio

Aquatic animal FCRs used in our analysesare values reported for species groups and/or species inTacon and Metian’s 2008 paper, Global overview on the use of fish meal and fish oil in industrially compounded aquafeeds: Trends and future propects[5]. LivestockFCRs are from Table 4.1 (pg. 140) of Vaclav Smil’s 2013 book: Should We Eat Meat? Evolution and Consequences of Modern Carnivory[6] (the values are based on U.S. Department of Agriculture (USDA) data), and additional sources for each type of livestock: a 2013 report on beef cattle feed efficiency [7], a pig industry report on feed efficiency from 2015[8], and a 2014 paper in Poultry Science[9].

Edible portion

We collected edible portion values from a variety of sources for each species, including the FAO and other government reports, peer-reviewed literature, industry data, and individual aquaculture experts (Table S4). For seafood, edible portions are reported in different ways, including drawn (gutted, with or without a head), dressed (fins and tail removed, head-on or head-off), and fillets or steaks (edible flesh with or without skin). For our analyses, we excluded values that included heads because fish heads are not typically eaten in the U.S., and including heads would be inconsistent with nutrition values for edible flesh.Some of the values include inedible parts of fish (e.g., skin, tail), which may slightly overestimate yield.Edible portion values for terrestrial animals were extracted from FAO reports and other sources (Table S4).

Feed composition

For each of the animal species (aquatic and terrestrial), we collected information on commercial feed nutrient composition including energy/calories, protein, lipid/oil/fat, carbohydrate, ash, and/or moisture (Table S5). When limited or no feed nutrient composition information was available for aquatic species, we used nutritional requirements of the fish species or species group. Some commercial feeds are only recommended for use with a single species while others are sufficient for the nutritional needs of several farmed aquatic animal species; feed nutrient composition datawere only used for a given species if that species or its species group was listedas a recommended species for the feed. Nutritional requirements were used for grass carp, channel catfish, rainbow trout, giant tiger prawn, and tilapia. Nutritional requirements were combined with actual feed composition information for all of these species except grass carp, where it was the only available feed information. We also communicated with aquaculture experts to fill gaps in the dataset.

The information we collected on protein levels in feed is consistent with protein requirements reported elsewhere. Typical protein requirements for food producing animals (including multiple life-stages) are: beef cattle (5-18%), pigs (13-26%), chickens (18-23%), and farmed fish (26-55%)[10, 11].

The amount of metabolizable energy, or calories (kcal), in feed was generally not reported for aquatic animal species. We estimated caloric content using protein, fat, and carbohydrate content (Equation S1). Approximate calorie values per g of macronutrient are as follows: 4 kcal per g of protein, 9 kcal per g of fat, and 4 kcal per g of carbohydrate [12]. There is some variation in calories per g of macronutrient based on food type, but feed ingredients were not available for most species so we used these approximate values. If carbohydrate content was not reported, we filled in themissing value by subtractingfrom 100 the percent of protein, fat, moisture, and ash. In cases where carbohydrate content was not reported and moisture and/or ash content was unknown, we used relevant literature, expert opinion, and/or other feed formulations for the same species to estimate those values in order to determine carbohydrate levels in feed.

Equation S1. Energy content in feed

Nutrition

For most species, nutrition data was compiled fromRelease 28 of the USDA Nutrient Database for Standard Reference (NDSR)[13]; this resource did not have information on pangas catfish. We found a nutrition label for pangas catfish on a seafood organization website[14]. We used information from an expert in China (personal communication, Shauhua Zahn, Nanyang Technological University)for species-specific information on carps and a publication from Mississippi State University for channel catfish [15]. All values represent the nutritional components of the raw, edible portions of the animal (Table 1). Values for protein and calorie content are reported as g and kcal per 100 g of edible meat, respectively.

Where available, information for aquatic animals was reported to the species level; otherwise, information was collected at the species group level. Some seafood nutrition content was reported for farmed species; otherwise, nutrition data was listed without distinction between wild and farmed.

3. Protein and calorie retention calculations

Many analytical approaches and terms exist for assessinganimal production, efficiency, yield, and resource use (See Table S1 for terms used in aquaculture). We developed protein and calorie retentionequations to calculate the proportion of protein and calories (kcal) that are fed to ananimalthatultimately enter the human food supply. The equation is based on weight of animalsproduced per unit of feed used (FCR), the portion of the whole animal that is edible, the protein/calorie content of feed, and the protein/calorie content of the edible part of the animal.Equations S2 and S3 are the precursors to Equations 1 and 2 in the manuscript.

Equation S2. Protein retention

Equation S3. Calorie retention

1

4. Supplementary tables

Table S1. Analytical approaches for assessing aquaculture production, efficiency, yield, and resource use

Name / Numerator (units) / Denominator (units) / Utility of measure / Citation example
Calorie conversion efficiency / FCR * calorie content in feed (%) / culture species calorie content (%) / - / -
Dry matter ratio (DMR) / FCR * percent dry matter in feed / percent dry matter in culture species / "The dry matter ratio (DMR) is an indicator of the efficiency with which nutrients in feed are converted to animal biomass" / Boyd et al. (2007)[16]
Feed conversion ratio / total feed used in production (kg/tonnes) / net production of cultured species (kg/tonnes) / “The lower the FCR, the lower the production cost, the greater the efficiency of feed use and the smaller the waste load” (Boyd et al. 2015)
“Excellent measure of feed use and economic efficiency” (Boyd 2005) / Boyd (2005)[17]; Boyd et al. (2015)[18]
Fish in / fish out ratio / wild fisheries inputs (kg/tonnes) / farmed fish outputs (kg/tonnes) / - / Naylor et al. (2009)[19]
Fish in / fish out ratio / [fishmeal or oil in diet (g/kg)]/ [fishmeal or oil reduction efficiency from foraged fish (g/kg)] * FCR / - / Two conversion ratios: 1. "conversion ratio of forage fish into fish meal and fish oil", "condensation efficiency is a more appropriate term"; 2. FCR / Ytrestøyl et al. (2014)[20]
Forage fish dependency ratio / percent fishmeal/oil in feed from forage fisheries * FCR / meal yield / "quantity of forage fish required to produce the amount of fishmeal and oil used to produce a unit of farmed fish" / Ytrestøyl et al. (2014)[20]
Marine protein dependency ratio / weight of marine-derived protein in feed (kg/tonnes) / fish protein weight gain (kg/tonnes) / “Reflect resources used by aquaculture because feed manufacturers use proteins and lipids, not whole fish”
“Allow for comparison of MNDRs (marine nutrient dependency ratios) between farmed species, despite differences in the body composition of these species” / Crampton et al. (2010)[21]
Marine oil dependency ratio / weight of marine-derived oil in feed (kg/tonnes) / fish oil weight gain (kg/tonnes) / “Reflect resources used by aquaculture because feed manufacturers use proteins and lipids, not whole fish”
“Allow for comparison of MNDRs [marine nutrient dependency ratios] between farmed species, despite differences in the body composition of these species” / Crampton et al. (2010)[21]
Nutrient retention / Amount of nutrient or energy incorporated in animal (whole body or edible part) / Amount of nutrient or energy used in feed / “[Nutrient-to-nutrient ratios] are a
measure of the proportion of the dietary nutrients and energy that isretained in the animal product.” / Ytrestøyl et al. (2014)[20]
Protein conversion efficiency / FCR * crude protein content in feed (%) / culture species crude protein content (%) / "Protein efficiency is an estimate of the ratio of feed protein applied to protein contained in the net harvest biomass of the culture species" / Boyd (2005)[17]
Protein conversion ratio / [FCR * feed crude protein content (%)]/100 / - / "The protein conversion ratio (PCR) is the ratio of feed protein to net harvest biomass" / Boyd (2005)[17]
Protein efficiency ratio / body weight or biomass produced (kg/tonnes) / protein fed (kg/tonnes) / Describes protein utilization as a measure of weight increase per amount of protein fed / Ytrestøyl et al. (2014)[20]
Protein production value / fish crude protein weight gain * 100 / crude protein weight in feed / - / Pucher et al. (2014) [22]
Waste production ratio / [[DMR - 1] * percent dry matter in culture species]/100 / - / "The waste production ratio (WPR) is the ratio of waste to live weight production of the culture species" / Boyd et al. (2007)[16]

1

Table S2. Global aquaculture production (in million metric tons) for selected species groups

Species group / Average / 2009 / 2010 / 2011 / 2012 / 2013
Carpsa / 23.4 / 21.6 / 22.5 / 23.1 / 24.3 / 25.7
Tilapiasb / 4.0 / 3.1 / 3.5 / 3.9 / 4.5 / 4.8
Shrimpsc / 4.1 / 3.5 / 3.8 / 4.2 / 4.4 / 4.5
Catfishesd / 3.5 / 2.8 / 3.2 / 3.4 / 3.9 / 4.3
Salmonidse / 2.8 / 2.4 / 2.4 / 2.7 / 3.2 / 3.1
Subtotal / 37.8 / 33.5 / 35.3 / 37.3 / 40.3 / 42.4
Total global productionf / 62.7 / 55.7 / 59.0 / 62.0 / 66.6 / 70.2

Data source: FAO FishStatJ; FAO World Fisheries and Aquaculture 2014[1, 23]

a Grass carp, silver carp, common carp, bighead carp, catla, crucian carp, etc.

b Nile tilapia, tilapias nei, blue-Nile tilapia hybrid, Mozambique tilapia, blue tilapia, etc.

cWhiteleg shrimp, giant tiger prawn, Penaeus shrimps nei, Kuruma prawn, fleshy prawn, etc.

dPangas catfish nei, torpedo-shaped catfish nei, channel catfish, Amur catfish, yellow catfish, etc.

e Atlantic salmon, rainbow trout, coho salmon, trout nei, etc.

fIncludes fed and unfed aquatic animal production, excludes aquatic plants.

Table S3. Production (in million metric tons) and top countries for selectedaquaculture species

Species Group / Species / Average MMT
(2009-2013) / Top producing countries
Carps / Common carp / 3.8 / China, Indonesia, Myanmar, Viet Nam
Grass carp / 4.7 / China
Catfish / Channel catfish / 0.4 / China, United States
Pangas catfish / 1.4 / Viet Nam, Indonesia
Salmonid / Atlantic salmon / 1.8 / Norway, Chile, United Kingdom, Canada
Rainbow trout / 0.8 / Chile, Iran, Turkey, Norway
Shrimp / Giant tiger prawn / 0.8 / Viet Nam, Indonesia, India
Whiteleg shrimp / 3.0 / China, Thailand, Ecuador
Tilapia / Nile tilapia / 2.9 / China, Egypt, Indonesia, Thailand
Tilapias neia / 0.7 / Brazil, Viet Nam, Bangladesh, Philippines

Data source: FAO FishStatJ; FAO World Fisheries and Aquaculture 2014 [1, 23]

anei: not elsewhere included

1

Table S4. Edible portion data and sources

Species / Edible Portion / No. of values used / Source
Carps
Common carp / 0.36-0.54 / 2 / FAO (1989)[24]
Grass carp / 0.36-0.54 / 2 / FAO (1989)[24]
Catfish
Channel catfish / 0.35-0.63 / 7 / Gregory Whitis (personal comm.); Silva and Dean (2001)[25]
Pangas catfishes / 0.35-0.63 / 7 / Lynch (2007)[26]; Gregory Whitis (personal comm.)
Salmonids
Atlantic salmon / 0.58-0.88 / 13 / Crapo et al. (1993)[27]; FAO (2001)[28]; Lynch (2007)[26]; Seafish[29]
Rainbow trout / 0.40-0.82 / 15 / Crapo et al. (1993)[27]; Ron Hardy (personal comm.); Lynch (2007)[26]; Seafish[29]
Shrimp
Giant tiger prawn / 0.40 / 1 / FAO (2001)[28]
Whiteleg shrimp / 0.62-0.65 / 2 / Monterey Bay Aquarium Seafood Watch Report [30]
Tilapias / 0.37-0.45 / 4 / FAO (1989)[24]; Lynch (2007)[26]; Seafish[29]
Cattle / 0.52-0.64 / 4 / MSU Extension (adapted from Principles of Meat Science, 4th Ed.)(2011)[31]; FAO (n.d.)[32]
Chicken / 0.70-0.78 / 4 / MSU Extension (adapted from Principles of Meat Science, 4th Ed.)(2011)[31]; FAO (n.d.)[32]
Pigs / 0.68-0.76 / 4 / MSU Extension (adapted from Principles of Meat Science, 4th Ed.)(2011)[31]; FAO (n.d.)[32]

1

Table S5. Feed nutrient composition data and sources

Feed Content (g per 100 g of feed) / Data Sources
Species / Protein (g) / Fat (g) / Carbohydrate (g) / Calories (kcal) / No. of feed profiles used
Carps
Common carp a / 17-45 / 2.2-15 / 22-59.8 / 175.8-554.2 / 12 / Aller Aqua (Poland)*
HaltapKft. (Hungary)*
Coppens International (Holland)*
FAO Feed information
Grass carpb / 25 / 4-5.5 / 47.5-49 / 326-345.5 / 2 / FAO Feed information
Catfish
Channel catfishc / 28-32 / 5-6 / 47-52 / 345-390 / 7 / USDA Southern Regional Aqu. Center (2013)[33]
Greg Whitis (personal comm.)
Robinson and Li (2005)[15]
Chapman (2015)[34]
Pangas catfishd / 26-32 / 5-6 / 47.5-51.5 / 339-388 / 3 / Greg Whitis (personal comm.) (USA)
Nhut (person comm.) (Vietnam)
Salmonids
Atlantic salmone / 35.5-44 / 20-32.5 / 12.5-21.5 / 372-554.5 / 3 / EWOS (multiple countries)*
Ytrestoyl, Aas, and Asgard (2015)[35]
Ye, Anderson, and Lall (2016)[36]
Rainbow trout / 40-47 / 21-24 / 8.5-12.5 / 383-454 / 3 / Bio-Oregon (USA)*
FAO Feed information
Ron Hardy, personal communication
Shrimps
Giant tiger prawne / 25-45 / 5-9 / 20-43 / 225-433 / 8 / Cargill*
Zeigler*
Presidents Feeds (Taiwan)*
FAO Feed information
Whiteleg shrimpe / 25-45 / 9 / 24-39 / 277-417 / 4 / Cargill*
Zeigler (multiple feeds)*
Tilapias / 20-32 / 4-8 / 25-51.1 / 216-404.4 / 18 / Charoen Pokphand (Thailand)*
San Miguel Foods, Inc. (Philippines)*
Santeh Feeds Corp. (Philippines)*
Vitarich Corp. (Philippines)*
Aquanutro (S. Africa)*
Cargill Viet Nam (Vietnam)*
FAO Feed information
Cattle / 7-15.4 / - / - / 188-339 / - / Protein: USDA NRCS[37]
Calories: Galyean et al. (2016) review article[38]; feed profiles included in thisstudywerepublishedin 2005 or later.
Chicken / 18-23 / - / - / 320 / - / Protein: National Research Council (1994) [39]Table 2-6
Calories: National Research Council (1994) Table 2-6; FAO document
Pigs / 13.2-20.9 / - / - / 326.5-335.1 / - / Protein: National Research Council (1998) [40]Table 10-1
Calories: National Research Council (1998) Tables 10-1, 10-3; US Pork Center

a Common carp: Assumes 10% moisture content based on:

and 11% ash content based on:

bGrass carp: Assumes 10% moisture based on:

c Channel catfish: Assumes 10% moisture content based on:

and 5% ash based on personal communication with Greg Whitis

d Pangas catfish: Assumes 5% ash and 10% moisture based on Channel catfish feed and

e Atlantic salmon, Giant tiger prawn, and Whiteleg shrimp: Some moisture and/or ash values were filled in based on other feed profiles for that species.

* Information source is a commercial feed company.

1

References

[1] FishStatJ (2016).

[2] NOAA Fisheries. The Surprising Sources of Your Favorite Seafoods (accessed 27 July 2016).

[3] U.S. Department of Agriculture Economic Research Service. Livestock & Meat Domestic Data meat statistics (2016, accessed 25 January 2017).

[4] Alexandratos N, Bruinsma J. World agriculture towards 2030/2050: The 2012 revision. 12–3, Rome (2012, accessed 29 June 2016).

[5] Tacon AGJ, Metian M. Global overview on the use of fish meal and fish oil in industrially compounded aquafeeds: Trends and future prospects. Aquaculture 2008; 285: 146–158.

[6] Smil V. Should We Eat Meat? : Evolution and Consequences of Modern Carnivory. West Sussex, UK: Wiley-Blackwell, 2013.

[7] Shike DW. Beef Cattle Feed Efficiency (2013, accessed 2 November 2017).

[8] Rabobank Food & Agribusiness Research. Pigs Might Fly: Peak Pork Production Potential (2015, accessed 2 November 2017).

[9] Zuidhof MJ, Schneider BL, Carney VL, et al. Growth, efficiency, and yield of commercial broilers from 1957, 1978, and 2005. Poult Sci 2014; 93: 2970–2982.

[10] Natural Resources Conservation Service (NRCS). Nutrient Management Documents and Technical References (2003, accessed 24 January 2017).

[11] Council NR. Nutrient Requirements of Fish and Shrimp. Washington, D.C.: National Academies Press. Epub ahead of print 25 May 2011. DOI: 10.17226/13039.