Biological Sampling on Research Vessels: The Weighing Workflow

The five weighing levels between haul and database

Biological sampling on a survey vessel produces weights at five nested levels: total catch, species or size category, subsample, individual fish, and organs or stomach contents. The catch, category and subsample levels are linked by ratios, so an error in one of those weights is carried into every raised value derived from it.

Level	What is weighed	Typical mass	Used for	DATRAS field (unit)
1. Total catch	Whole haul, when not fully sorted	Kilograms to tonnes	Raising partially sorted catches	Not a standard field; national systems
2. Catch category	All fish of one species, sex or size category	Grams to hundreds of kilograms	Catch rate by weight, raising factor numerator	CatCatchWgt (g)
3. Subsample	The fish actually measured for length	Grams to tens of kilograms	Raising factor denominator	SubWgt (g)
4. Individual fish	Fish sampled for age, sex and maturity	About 1 g to over 10 kg	Length–weight relationships, weight-at-age	IndWgt (g)
5. Organs and contents	Gonads, livers, stomach contents	Below 1 g to over 1 kg	Maturity indices, condition, diet	Project or stomach database formats

Units in the table follow the DATRAS units documentation published by the ICES Data Centre. The procedures below follow the North Sea IBTS manual (Revision 11); other ICES surveys use comparable steps with species- and gear-specific details.

Step 1: sorting the catch

The catch from every valid haul is sorted by species, and the IBTS manual requires full sorting wherever that is practically possible. Fish and shellfish are identified to the lowest taxonomic level possible, and only when that is not feasible may organisms be grouped by genus or family.

For large catches, very abundant species or size categories can be subsampled instead of sorted completely. Elasmobranchs and crustaceans are measured and weighed by sex, which means separate category weights for males and females.

From a weighing point of view, sorting creates a sequence of baskets or boxes that must each be weighed quickly while the next haul may already be on deck. Three practical points reduce errors at this stage:

• Use a consistent tare. Wet baskets hold water; drain them the same way each time and re-tare regularly rather than relying on a nominal basket weight.
• Keep capacity and resolution matched. A 60 kg basket scale with 10–20 g readability is fine for a 40 kg category, but not for a category of 150 g of a rare species.
• Weigh before fish lose water. Catch left draining on deck or in sun loses weight, which lowers category weights compared with individual weights taken later.

Step 2: subsampling and raising factors

A raising factor is the category weight divided by the subsample weight, and it scales the numbers-at-length counted in the subsample up to the whole category. Both weights must therefore be taken with the same care, on a suitable scale and with the same tare procedure.

Example. A haddock category weighs 48.6 kg. A random subsample of 12.15 kg is measured, giving 90 fish. The raising factor is 48.6 ÷ 12.15 = 4.0, so each length class count is multiplied by 4, giving an estimated 360 fish in the category. If the subsample had been recorded 0.5 kg too light, the factor would become 4.17 and every length class would be over-estimated by about 4 %.

The IBTS manual gives the following guidance for subsample size and composition:

1. For catches of more than 75 individuals of a category, measure a representative subsample of at least 75 fish; fewer may be enough for narrow length ranges, and more are needed for wide ones.
2. For catches above 1,000 individuals, increase the subsample to 150 fish.
3. When the catch contains distinct length groups with nothing in between, split it into size categories. The manual's example: in a catch of 999 fish of 18–26 cm plus a single 40 cm fish, a subsample of 100 would raise the large fish to either ten or zero. The large fish is measured separately as its own category and the remainder is subsampled.

Each size category then needs its own category weight and subsample weight, which increases the number of weighings per haul.

Step 3: length measurements

Length distributions are recorded for all fish species caught, and the measurement convention determines how weights can later be linked to lengths. In the IBTS, fish are measured as total length, from the tip of the snout to the tip of the caudal fin:

• to 0.5 cm below for herring, sprat and boarfish;
• to 1 cm below for all other fish species;
• to the millimetre below for the carapace of shellfish such as Nephrops and edible crab, and to the centimetre below for cephalopod mantle length.

Some deep-water taxa use other length types, such as pre-anal fin length for grenadiers. Whatever the convention, it must be stored with the data, because length–weight parameters are only valid for the length type used to fit them.

Step 4: individual fish – weight, sex, maturity and otoliths

Individual fish sampling links age to length and weight: selected fish are weighed individually, sexed, staged for maturity and have both otoliths removed. The IBTS manual requires sex, maturity and weight data for all target species for which age data are collected.

Species (IBTS examples)	Minimum otolith sampling per haul
Herring, sprat	1 otolith per 0.5 cm length class
Cod, mackerel, saithe, plaice	1 otolith per 1 cm length class
Haddock, whiting	2 per 5 cm class for 11–30 cm; 2 per 1 cm class above 30 cm
Norway pout	2 per 5 cm class for 5–15 cm; 2 per 1 cm class above 15 cm

Individual weights are usually taken on a separate, higher-resolution scale at the measuring station. In a Cefas study of survey data from 2009–2012, the smallest fish were weighed to 0.1 g, medium-sized fish to 1 g and large fish to 5 g, or 10 g for exceptionally large individuals (Silva et al., 2013). The same report describes an electronic data capture system that collects individual fish weights from linked marine scales and prompts the operator when an otolith is still needed to meet the sampling target.

These individual weights are the raw material for length–weight relationships and condition indices.

Step 5: organs and stomach contents

Organ and stomach weights are the lightest routine measurements on a survey and demand the most careful weighing. Gonad weights support maturity staging through the gonadosomatic index, and liver weights indicate energy reserves.

For diet studies, the ICES stomach sampling manual recommends 5 stomachs per 5 cm length group for most predators in the North Sea, and 15 per group for saithe, mackerel and horse mackerel, because many of their stomachs are empty. At the minimum sampling level, stomachs can be analysed on board; at the extended level they are frozen individually and analysed in the laboratory. In both cases total prey weight and the weight of each prey group are recorded.

Readability requirements, wind protection and a protocol for these samples are covered in Precision Weighing of Small Samples at Sea.

Where weighing errors enter survey data

Most weighing errors on survey vessels come from a handful of predictable sources, and each has a practical countermeasure.

Error source	Effect on the data	Mitigation
Vessel motion	Unstable or biased readings, especially for light samples	Motion-compensated scale, placement near midships and low in the vessel, wait for stable indication
Wet or variable tare	Systematic error in category and subsample weights	Drain containers consistently, re-tare often, record tare method
Coarse resolution	Poorer length–weight fits for small fish	Match readability to sample mass; use separate scales for small fish
Wind and draughts	Drifting readings on small platforms	Wind cover, sheltered position
Transcription	Digit swaps, wrong units, values linked to the wrong fish	Direct data capture from the scale into survey software (see choosing a scientific marine scale)
Zero drift and damage	Offset in all values	Zero before each series, check with a test weight every watch

The effect of resolution is measurable. Silva et al. (2013) found a length–weight fit for sprat of r² = 0.70 with data recorded to 0.5 cm and 1 g, against r² = 0.865 with data recorded to 0.1 cm and 0.1 g, and recommended that survey protocols review the resolution used for such species.

How WPL approaches this

WPL builds separate motion-compensated scales for the different levels of this workflow. At the sorting area, an M2 Series Marine Scale with a 30 kg or 60 kg range on a 400 × 300 mm or 500 × 400 mm platform handles baskets and categories, with USB, RS232 and Ethernet for data capture. At the measuring station and in the wet laboratory, the M3 Series Scientific Marine Scale covers individual fish and organs from 300 g (0.1–0.2 g readability) to 6,000 g. More context on survey weighing is collected in the Research Vessels hub.