In MortenVinther/FishStomachs: Compilation of stomach contents data
knitr::opts_chunk$set(
collapse = TRUE,
comment = "#>"
)
Compilation of stomach content data for estimation of population diet and food ration.
Compilation of data from individual stomachs to population level is not
trivial and the method will depend highly on the question asked. Stomach
data are often collected to get information on the average diet or food
ration of a given species within a given area to inform multispecies
assessment models (e.g. SMS or Gadget). Such models require data on diet
and biomass eaten (food ration) by predator species and size classes.
The average "population" diet or food ration should basically be
calculated from a stratified mean of the individual stomach content
samples, weighted by the strata density of the predator and the area of
the strata. This seems simple, but incomplete and patchy sampling makes
it often necessary to use a series of ad hoc solutions.
In general, a series of data compilation methods could be applied:
- Read and check data from the agreed exchange format;
- Bias correct to take into account variable evacuation rate;
- Assign size classes for predators and preys;
- Bias correct to take into account regurgitated stomachs within
sample units;
- Aggregate stomach contents within sample_id and size classes.
- Allocate unidentified or partly identified prey items;
- Calculate population diet and food ration from a weighted average.
The sequence of these steps, of which some of the steps might be
repeated, will depend on the individual sampling design and the quality
of the analysed data.
The FishStomachs package defines data structures suitable for stomach
data, and provides the necessary methods to compile observed stomach
data into population diet and biomass eaten, used for multispecies
models. The methods applied for a set of observations are stored within
the data output to document the compilation steps taken.
With my experience from compiling stomach data for the North Sea for use
in the MSVPA/SMS models, this document describes the methods above and
how these can be applied.
1) Data exchange format
Data must be recorded, stored and exchanged in an agreed and documented
format suitable for the purpose of sampling. This R-package uses a user
defined format (see Table 1) with the possibility to rename variable
names such that it is possible to include existing databases (e.g. ICES
stomach contents database). The new format includes a number of
mandatory (system defined) and optional (user defined) variable names,
which uses "aliases" (See Table 2) as a link to "internal" variable
names used for processing by FishStomachs.
Table 1. Database format. "field" is the variable name used in
FishStomachs; "types" is the type of input(c=character, i=integer,
d=double, l=logical); "mandatory" is the field required as minimum
(T=TRUE, F=FALSE); "PRED" is the variable part of predator information;
"PREY" is the information part of prey information; "units" is the unit
or values applied. The data base format is specified in the file
stomach_format.csv.
a<-readr::read_csv(file.path(system.file( package = "FishStomachs"),"stomach_format.csv"),col_types = readr::cols())
a<- dplyr::mutate_if(a,is.logical,as.character())
a<- dplyr::mutate(a,mandatory=dplyr::if_else(mandatory=='FALSE','F','T'),
PRED=dplyr::if_else(PRED=='FALSE','F','T'),PREY=dplyr::if_else(PREY=='FALSE','F','T'))
a<- dplyr::mutate(a,units=dplyr::if_else(is.na(units),' ',units))
knitr::kable(a[,4:10])
Table 2. Alias names accepted as default. Aliases can be changed and
added by using a local version of the database format. The data base
format and aliases are specified in the file stomach_format.csv.
a<-dplyr::mutate(a,alias_1=dplyr::if_else(is.na(alias_1),' ',alias_1),alias_2=dplyr::if_else(is.na(alias_2),' ',alias_2),alias_3=dplyr::if_else(is.na(alias_3),' ',alias_3))
knitr::kable(a[,1:4])
The presently used ICES data exchange format for stomachs (see Appendix
1 in this document) is based on the "Year of the stomach" sampling for
North Sea stocks in the period 1981-1991 (see Appendix 2). The new
version includes some extensions to better include stomachs sampled
outside the "Year of the stomach", e.g. from national sampling programs
with less standardized recording. Another change is that the new
exchange format uses the WoRMS AphiaID
(http://www.marinespecies.org/index.php) species taxonomic ode, where
the old format used the NODC codes
(https://www.nodc.noaa.gov/General/CDR-detdesc/taxonomic-v8.html).
In this document, the examples and variable names used are based on a
data set on the ICES exchange format (Appendix 1).
2) Bias correction to take into account variable evacuation rate
It is often assumed that diet (eaten food) is the same as the observed
stomach content. This is however not the case as the stomach evacuation
rate of an individual stomach depends on e.g. the average meal size,
average energy density of the meal, prey armament and temperature. Large
energy rich meals will for example have a lower evacuation rate than
small energy poor meals. Therefore, large energy rich meals will be
present in a stomach for a longer time and thereby detectable for a
longer time, than stomach contents from small energy poor meals.
In the two sub-sections below, it is described how to bias correct the
effect of variable stomach evacuation rate and thereby how to bias
correct the estimated average diet (or consumption rate) estimated from
the observed stomach contents data. Ideally, this bias correction should
take place at the level of sampling, i.e. by the individual stomach or
by the pool of stomachs by sample site.
Estimation of food ration from stomach contents data
Food rations (evacuation rate of stomach contents) can be estimated from
the observed stomach contents using the methods suggested by Andersen
and Beyer (2005 a, b). This model takes into account the differences in
evacuation rates between prey types due to their energy density and
their resistance to digestion (armament).
Ration (R) (per hour) of prey group (i) for pooled stomachs is
calculated from:
$$R = \sum_{i}^{}{(\rhoM_{i}b_{i}e^{\delta T}L^{\lambda}}E^{\xi})K\left(\frac{N_{A}}{N_{F}} \right)^{\alpha - 1}S^{\alpha}$$
M= armament of individual prey (group) i
b=proportion of prey (group) i
T= temperature (^o^C)
L= length (cm) of the predator
E= average energy density (kJ/g wet weight) of the stomach (or of the
pooled stomachs)
N= Number of stomachs in the sample, total (A) and with food (F)
S = average stomach contents in grams
$\rho$, $\delta$, $\lambda$, $\xi$ and K are parameters to the model
(Table 3)
Values for armament (M) and average energy density of individual prey
(groups) as used in the compilation of the North Sea stomachs are
presented in Table 5.
When data exist by the individual stomachs, the ration eaten per
(feeding) fish can be calculated as:
$$R = \sum_{i}^{}{\rho * M_i b_i e^{\delta T} * L^{\lambda} * E^{\xi} * S^{\alpha}}$$
Table 3. Parameter values of the generic cylinder model of gastric
evacuation used for the compilation of stomachs from predators in the
North Sea.
| Predator | $\rho$ | $\lambda$ | $\delta$ | $\xi$ | $\alpha$ | K |
|:---------|:--------|-----------|----------|:-----:|----------|:----:|
| Cod | 0.00224 | 1.30 | 0.083 | -0.85 | 0.5 | 0.85 |
| Haddock | 0.00191 | 1.30 | 0.083 | -0.85 | 0.5 | 0.85 |
| Saithe | 0.00171 | 1.35 | 0.081 | -0.85 | 0.5 | 0.85 |
| Whiting | 0.00171 | 1.35 | 0.081 | -0.85 | 0.5 | 0.85 |
| Mackerel | 0.00174 | 1.30 | 0.080 | -0.85 | 0.5 | 0.85 |
Estimation of diet from stomach contents data
The method for bias correction in the calculation of food ration as
outlined above can also be applied to calculate the ration of the
individual prey item and thereby the average diet. For example, for
individual stomachs the Ration (R) of prey item i becomes:
$$R_{i} = \rho\ M_i *b_i * e^{\delta T} * L^\lambda * E^{\xi}$$
Based on the set of $R_{i}$ for the individual predators from e.g. a
sampling site (trawl haul) the average Ration of the of the individual
prey item can be calculated by taking the number of feeding, empty and
regurgitated stomachs into account. This average Ration by prey group is
in this case expressed as g wet weight per hour (with $\rho$ included in
the equation) and can easily be aggregated with average rations of preys
from other strata. These rations can finally be transformed into diet
data (weight proportions by prey item).
Table 4. Average energy density and armament of prey groups as applied
in the compilation of stomach for use in the North Sea SMS.
a<-readr::read_csv(file.path(system.file( package = "FishStomachs"),"extdata","energy_density_and_armament.csv"),col_types = readr::cols())
knitr::kable(a,col.names=c("Species","Quarter","Lower length (mm)","Upper length (mm)","Energy density (kJ/g wet weight)", "Armament"))
3) Assign predator and prey size classes.
Stomach data can be sampled, analysed and recorded by the individual
fish or by size group ("pooled stomach data") for the individual sample
unit (sample_id), e.g. trawl haul. To save time during sampling and
analysis of the stomach contents, the "Year of the stomach" data were
sampled and analysed by pre-determined size classes (Table 4). Pooled
stomachs do not provide information on the individual fish and on the
variation in stomach contents, which is preferable in e.g. estimation of
daily food ration and diet. Grouping of stomachs by size classes is
however often necessary in the later compilation of data, as data are
too scarce to be compiled by e.g. each cm group separately.
In general, I will suggest using size classes, e.g. the size classes
used in 1991 by "Year of the Stomach" sampling (Table 3). That means
that the variables are in the data exchange format.
-
pred_ll (predator lower length bound)
-
pred_lu (predator upper length bound)
-
prey_ll (prey lower length bound)
-
prey_lu (prey upper length bound)
are set from the length information on the individual predators and
preys, if not already included in the data and defined by the sampling
design.
Table 5. Size classes (for items larger than 5 cm) used for sampling
and compilation of stomach data in the North Sea "Year of the stomach"
1991 sampling.
| Length class, 1991 | Length (cm) |
|:-------------------|:------------|
| 50 | 5-5.9 |
| 60 | 6-6.9 |
| 70 | 7-7.9 |
| 80 | 8-9.9 |
| 100 | 10-11.9 |
| 120 | 12-14.9 |
| 150 | 15-19.9 |
| 200 | 20-24.9 |
| 250 | 25-29.9 |
| 300 | 30-34.9 |
| 350 | 35-39.9 |
| 400 | 40-49.9 |
| 500 | 50-59.9 |
| 600 | 60-69.9 |
| 700 | 70-79.9 |
| 800 | 80-99.9 |
| 1000 | 100-119.9 |
| 1200 | >120 |
4) Bias correction: Taking number of empty and regurgitated stomachs into account
Ideally each stomach should be classified as: with food, with food but
regurgitated, with skeletal remains only or empty. This information is
included in the presently used exchange format.
| Database variable | Comments |
|:------------------|:-----------------------------------------------------------------------------------------------|
| n_food | Stomachs with (recently ingested) food |
| n_regur | Stomach with food, but evidence of regurgitation of parts or all the stomach contents. |
| n_skel | Stomachs with practically indigestible remains and the fish is not considered feeding recently |
| n_empty | No stomach contents |
Stomach contents from regurgitated (or everted) stomachs should not be
included in the estimation of diet as the proportion regurgitated is
unknown. The number of fish with regurgitated stomachs should however be
recorded ("n_regur"). With the assumption that the regurgitated stomachs
sampled within a "sample_id" have had the same stomach contents as the
feeding (non-regurgitated) fish each observed prey item weight and
number should be corrected by a factor to calculate the mean stomach
contents of a predator within "sample_id".
$factor = \frac{(N_{F} + N_{R})}{N_{F}*(N_{F} + N_{SR} + N_{R} + N_{E})}$
Where
$N_{F}$ = Number of feeding fish ("n_food")
$N_{R}$ = Number of feeding fish, but regurgitated ("n_regur")
$N_{\text{SR}}$ = Number of non-feeding fish, stomachs containing only
skeletal remains ("n_skel")
$N_{E}$ = Number of non-feeding fish with empty stomachs ("n_empty")
In practice, it is difficult to detect fish with regurgitated stomachs,
and later data analysis may often show that there is a clear difference
in the proportions of regurgitated stomachs between vessels surveying in
the same area at the same time.
Another issue is how to distinguish "n_skel" from "n_food". In the
original compilation of the 1991 "Year of the Stomach" (ICES, 1997),
$N_{SR}$ was added to the numerator in the calculation of the factor
above. However, I think "n_skel" should be considered as stomachs from
not (recently) feeding fish and should be treated as empty stomachs,
i.e. that the stomach contents should be disregarded in the compilation
of data for feeding fish.
5) Aggregate stomach contents within sample_id and size classes
6) Allocate unidentified and partly identified prey items
Parts of the stomach contents are often not fully identified to species
(group) or size class. With the assumption that the unidentified items
follow the same species and size distribution as the identified prey
items, it is possible to allocate the unidentified items proportionally
to the identified species and size class. However, with the same
assumptions it can be claimed that the unidentified items can be ignored
for diet studies as the weight proportions of the individual food items
would be the same with and without the unidentified items. I am not
aware of published studies that quantify bias for the two approaches.
The allocation of unidentified items to identified items is not simple
in practice. A stomach may for example contain remains of an
unidentified fish and identified invertebrates. In such case, it would
not be sensible to allocate the remains to the identified preys. However
other stomachs from the same predator and size class from the same haul
might include identified fish preys which could be used as a
"distribution key" for unidentified items. If a given haul does not
include identified preys for a "distribution key" an alternative key
could be obtained from the neighbouring hauls or from another size
classes within the haul.
It is not possible to give specific rules on how to allocate
unidentified prey items, other than the distribution key should be
derived from "local" data as far as possible. By "local" is meant from
the same haul, time period, predator size class, ICES rectangle,
roundfish area etc.
Assigning prey group
Stomach contents data may include prey identified to species species
level, even for preys that are considered as "other prey" in the later
data set produced for multi species assessments.
With 250.000 stomachs sampled in the North Sea as the basis for
multispecies models a pragmatic approach was needed for allocation of
unidentified items. The former used NODC codes for preys are not
maintained anymore by NOAA, however NODC codes are still very useful for
the compilation of stomach data as they are based on a 10-digit
"intelligent" code. By "intelligent" code numbers it is meant that
information about taxonomy was built into the codes through the use of
2-digit couplets to represent one or more levels of the taxonomic
hierarchy. For example, a species assigned a 10-digit code would belong
to the genus represented by the first 8 digits of the code. In the
compilation of stomach data, this hierarchy makes it easier to assign
(by an algorithm) partly identified food items, to the identified items
within the same genus, family or order. Lookup tables exist in the ICES
system to assign a NODC code to the WoRMS AphiaID code system presently
used by ICES. The section "Compilation of stomach data for use in the
North Sea SMS model" provides examples of how allocation of unidentified
items could be done.
To be more specific. For the North Sea data compilation, the model prey
species (e.g. cod, whiting, sprat) were maintained as named preys. The
remaining prey items were grouped according to the table below.
Invertebrates (NODC \< 8500000000) are treated as "other food" in the
model and aggregated into one group. Two groups of prey items, Clupeidae
and Gadoidae were maintained as they were used to name e.g. preys that
most likely were sprat or herring (Clupeidae) but could not be
determined to species level. "other fish" includes fish preys that are
not either a Clupeidae or Gadoidae. The "unknown" group (NODC >
900000000) was used for unidentified prey remains (either fish or
invertebrates).
| Species group | First NODC | Last NODC |
|---------------|------------|------------|
| other food | 400000000 | 8499999999 |
| other fish | 8500000000 | 8747009999 |
| Clupeidae | 8747010000 | 8747019999 |
| other fish | 8747020000 | 8791029999 |
| Gadidae | 8791030000 | 8791031999 |
| Other fish | 8791032000 | 8999999999 |
| unknown | 9000000000 | 9999999999 |
to allocate the partly identified preys, preys within the Clupeidae and
Gadidae were first allocated to the identified species within the group
(if such exists)
7) Aggregation of stomachs contents into population metrics
The average diet should basically be calculated as a stratified mean of
the individual stomach content samples, weighted by strata abundances
(indices) of the predator and strata areas. For the initial compilation
of the "Year of the stomach" the average stomach contents were first
calculated by ICES rectangles and finally aggregated to the population
level (ICES, 1997). When there was more than one sample within a
rectangle the rectangle value was calculated as the weighted mean of the
sample (haul) average values, the weighting factors being the number of
stomachs from each sample. Four methods were tried to calculate the
"population" average stomach:
i. The rectangle values were averaged disregarding the number of
stomachs sampled within each rectangle.
ii. The rectangle values were weighted by the sample size (total number
of stomachs sampled within the rectangle).
iii. The rectangle values were weighted by the survey catch rates within
the rectangle.
iv. The rectangle values were weighted by the square root of survey
catch rates within the rectangle; this reduces the influence of
occasional very large catch rates
The compilation (ICES, 1997) ended up using option iv for the
predators cod, haddock, whiting and mackerel, while option ii was used
for saithe because the samples (survey catches) were patchily
distributed.
The compilation of data for the SMS model deviates from the original
method used. Average stomach contents by ICES rectangle are derived in
the same way as done initially, but the SMS method calculates then the
weighted mean of stomach contents by ICES roundfish area using the
square root of the mean rectangle survey cpue (option iv) as weighting
factor (but option ii are used for saithe).
The calculation of the population diet for SMS was done as a weighted
average of the strata (roundfish area) average diet weighted by the
product of the average strata survey cpue and area of the strata.
References
Andersen N.G., Beyer J.E. 2005a. Mechanistic modelling of gastric
evacuation applying the square root model to describe surface-dependent
evacuation in predatory gadoids. J Fish Biol 67:1392--1412.
Andersen N.G., Beyer J.E. 2005b. Gastric evacuation of mixed stomach
contents in predatory gadoids -- an expanded application of the square
root model to estimate food rations. Journal of Fish Biology
67:1413--1433.
Berg, C. W., A. Nielson, and K. Kristensen. 2014. Evaluation of
alternative age-based methods for estimating relative abundance from
survey data in relation to assessment models. Fisheries Research
151:91--99.
Daan, N., Bromley, P.J., Hislop, J.R.G and Nielsen, N.A. 1990. Ecology
of North Sea fish. Netherlands Journal of Sea Research,
26(2--4):343--386.
ICES. 1989. Database report of the Stomach Sampling Project 1981. Coop.
Res. Rep. 164: 1--145.
ICES. 1997. Database report of the Stomach Sampling Project 1991. Coop.
Res. Rep. 219: 1--442.
ICES. 1997. Report of the Multispecies Assessment Working Group, ICES
headquarters, 11--19 August 1997. ICES CM 1997/Assess:16, 235 pp.
Pinnegar, J.K. (2014) DAPSTOM - An Integrated Database & Portal for Fish
Stomach Records. Version 4.7. Centre for Environment, Fisheries &
Aquaculture Science, Lowestoft, UK. February 2014, 39 pp.
\
Appendix 1. ICES data exchange format for stomach data
(http://ices.dk/marine-data/data-portals/Pages/Fish-stomach.aspx )
+-----------------------------------+-----------------------------------+
| Field | Description |
+-----------------------------------+-----------------------------------+
| Dataset | Dataset name |
+-----------------------------------+-----------------------------------+
| RecordType | SS for single stomach |
+-----------------------------------+-----------------------------------+
| Country | Country that collected the data |
+-----------------------------------+-----------------------------------+
| Ship | Vessel that collected the data |
+-----------------------------------+-----------------------------------+
| Latitude | Data sampling position -- |
| | latitude |
+-----------------------------------+-----------------------------------+
| Longitude | Data sampling position -- |
| | longitude |
+-----------------------------------+-----------------------------------+
| Estimated_Lat_Long | Flag whether the sampling |
| | position based on the reported |
| | area |
+-----------------------------------+-----------------------------------+
| ICES_StatRec | ICES statistical rectangle |
+-----------------------------------+-----------------------------------+
| ICES_AreaCode | ICES area code |
+-----------------------------------+-----------------------------------+
| Year | YYYY |
+-----------------------------------+-----------------------------------+
| Month | MM |
+-----------------------------------+-----------------------------------+
| Day | DD |
+-----------------------------------+-----------------------------------+
| Time | Sampling time like HHMM |
+-----------------------------------+-----------------------------------+
| Station | Station reference |
+-----------------------------------+-----------------------------------+
| Haul | Haul number |
+-----------------------------------+-----------------------------------+
| Sampling_Method | Predator sampling method |
+-----------------------------------+-----------------------------------+
| Depth | Sampling depth |
+-----------------------------------+-----------------------------------+
| Temperature | ^o^ C |
+-----------------------------------+-----------------------------------+
| SampleNo(FishID) | Predator reference code |
| | |
| | Fish ID unique for Country, year, |
| | quarter and ship |
+-----------------------------------+-----------------------------------+
| ICES_SampleID | ICES predator reference |
+-----------------------------------+-----------------------------------+
| Predator_AphiaID | Predator WoRMS AphiaID |
+-----------------------------------+-----------------------------------+
| Predator_LatinName | Predator taxon Latin Name |
+-----------------------------------+-----------------------------------+
| Predator_Weight(mean) | (Mean) predator weight |
+-----------------------------------+-----------------------------------+
| Predator_Age(mean) | (Mean) predator age |
+-----------------------------------+-----------------------------------+
| Predator_Lengh(mean) | (Mean) predator length |
+-----------------------------------+-----------------------------------+
| Predator_LowerLengthBound | Predator's length lower bound |
+-----------------------------------+-----------------------------------+
| Predator_UpperLengthBound | Predator's length upper bound |
+-----------------------------------+-----------------------------------+
| Predator_CPUE | Predator catch per hour |
+-----------------------------------+-----------------------------------+
| GallBladder_stage(class) | Gall bladder stage |
+-----------------------------------+-----------------------------------+
| Stomach_METFP | Method of stomach preservation |
+-----------------------------------+-----------------------------------+
| Stomach_TotalNo | Total number of stomachs in the |
| | pool -- for single stomachs |
| | always 1. |
+-----------------------------------+-----------------------------------+
| Stomach_WithFood | Number of stomachs with food |
+-----------------------------------+-----------------------------------+
| Stomach_Regurgitated | Number of stomachs regurgitated |
+-----------------------------------+-----------------------------------+
| Stomach_WithSkeletalRemains | Number of stomachs with skeletal |
| | remains |
+-----------------------------------+-----------------------------------+
| Stomach_Empty | Number of empty stomachs |
+-----------------------------------+-----------------------------------+
| Stomach_ContentWgt | Stomach content weight |
+-----------------------------------+-----------------------------------+
| Stomach_EmptyWgt | Stomach empty weight |
+-----------------------------------+-----------------------------------+
| Stomach fullness | Stomach fullness |
+-----------------------------------+-----------------------------------+
| Stomach_Item | Stomach item name (see Appendix |
| | B) |
+-----------------------------------+-----------------------------------+
| ICES_ItemID | ICES stomach item ID |
+-----------------------------------+-----------------------------------+
| Prey_AphiaID | Prey WoRMS AphiaID |
+-----------------------------------+-----------------------------------+
| Prey_LatinName | Prey taxon Latin Name |
+-----------------------------------+-----------------------------------+
| Prey_IdentMet | Prey identification method |
+-----------------------------------+-----------------------------------+
| Prey_DigestionStage | Prey digestion stage |
| | |
| | 0= Intact prey (skin, fins, legs |
| | and flesh is complete), 1= |
| | partially digested prey (prey in |
| | more advanced stages of |
| | digestion), 2= partially digested |
| | prey (prey in more advanced |
| | stages of digestion), 3= skeletal |
| | material (no flesh, only bones, |
| | shells, otoliths) |
+-----------------------------------+-----------------------------------+
| Prey_TotalNo | Total number of preys |
+-----------------------------------+-----------------------------------+
| Prey_Weight | Prey weight in grams |
+-----------------------------------+-----------------------------------+
| Prey_LengthIdentifier | Prey length identifier |
+-----------------------------------+-----------------------------------+
| Prey_Length | Prey length in cm |
+-----------------------------------+-----------------------------------+
| Prey_LowerLengthBound | Prey length lower bound |
+-----------------------------------+-----------------------------------+
| Prey_UpperLengthBound | Prey length upper bound |
+-----------------------------------+-----------------------------------+
| Prey_MinNo | Minimum number of preys |
+-----------------------------------+-----------------------------------+
| Remarks | Any relevant comments |
+-----------------------------------+-----------------------------------+
\
Appendix b. Data exchange format used by the North Sea stomach
sampling program 1981-1991,
Source ICES Coop. Res. Rep. No. 219 , pp 421-422
| Position | Name | Type^a^ | Range of values | Comment |
|:-------------|:-----------------------------|:------------|:--------------------|:-------------------------------------------------|
| 1-2 | Ecosystem name | 2N | 1-8 | Footnote^b^ for coding scheme |
| 3-4 | Year | 2N | 01-99 | Year - 1900 |
| 5 | Quarter | 1N | 1-4 | X |
| 6-9 | Square | 4AN | | ICES rectangle |
| 10-19 | Predator code | 10N | | NODC 10-digit |
| 20-24 | Sample number | 5N | 1-99999 | Unique fish ID |
| 25-27 | Country | 3A | | ICES alpha codes, No |
| 28-31 | Ship | 4A | | ICES alpha, if available, No data: XXXX |
| 32-34 | Sampling method | 3A | | Footnote^d^ for coding scheme No data: XXX |
| 35-40 | Station/haul | 6AN | | Use national system,No data: XXXXXX |
| 41-42 | Month | 2N | 01-12 | Not known: 99 |
| 43-44 | Day | 2N | 01-31 | Not known: 99 |
| 45-48 | Time of day | 4N | 0-2359, | Local time, hh/mm, start of tow. Not known: 9999 |
| 49 | Quadrant | lN | 1-4, 9 | See footnote^d^ Not known: 9 |
| 50-53 | Latitude | 4N | 0-9000, | dd/mm. Not known: 9999 |
| 54-58 | Longitude | 5N | 0-18000, | ddd/mm. Not known: 99999 |
| 59-61 | Depth (meters) | 3N | 1-999 | Mean depth of tow. Not know: 999 |
| 62-64 | Temperature (bottom) | 3N | -9.9 to 99.8 | XX.X one implied decimal. Not known: 999 |
| 65-68 | Predator (mean) length | 4N | 1-9999 | mm XXXX |
| 69-73 | Predator (mean) weight | 5N | 1-99999 | grams XXXXX. Not known: 99999 |
| 74-75 | Predator (mean) age | 2N | 0-99 | Not known: 99 |
| 76-79 | Predator lower length bound | 4N | 1-9999 | mm XXXX |
| 80-83 | Predator upper length bound | 4N | 1-9999 | mm XXXX |
| 84-90 | CPUE | 7N | 1-9999999 | Weighting coefficient for sample.Not known: 1 |
| 91-93 | Number with food | 3N | 0-999 | 0, 1 for individual samples |
| 94-96 | Number regurgitated | 3N | 0-999 | 0, 1 for individual samples |
| 97-99 | Number with skeletal remains | 3N | 0-999 | 0, 1 for individual samples |
| 100-102 | Number empty | 3N | 0-999 | 0, 1 for individual samples |
| 103-112 | Prey species code | 10N | | NODC 10-digits |
| 113-116 | Prey lower length bound | 4N | 1-9999 | mm XXXX. Not known: 9999 |
| 117-120 | Prey upper length bound | 4N | 1-9999 | mm XXXX. Not known: 9999 |
| 121-128 | Prey weight | 8N | 1-99999999 | Total weight mg XXXXXXXX |
| 129-134 | Prey number | 6N | 1-999999 | Total number. Not known: 999999 |
| 135 | Digestion stage | 1N | 0-2, 9 | See footnote^e^ |
a) All numeric fields (N) right justified, zero filled; all apha (A)
and mixed alpha/numeric fields (AN) left justified, space filled
b) North Sea=l, Baltic Sea=2, Barents Sea=3, Iceland=4, Northeastern
Newfoundland=5, Northeastern USA=6, Southern Shelf=7, Faroes=8
c) DEM=demersally caught by trawling or seining gears PEL=pelagically
caught by trawling or seining gears DHL=demersal hook and line
PHL=pelagic hook and line DGN=demersal gill net PGN=pelagic gill net
d) Quadrants for position identification are: l=NE, 2=NW, 3=SW, 4=SE
(The axes of the quadrants are the equator and the Greenwich
meridian)
e) Digestion stages are: 0 =Pristine, 1 =Affected by digestion,
2=Skeletal remains, 9=Unknown
MortenVinther/FishStomachs documentation built on Feb. 14, 2025, 7:33 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
knitr::opts_chunk$set( collapse = TRUE, comment = "#>" )
Compilation of stomach content data for estimation of population diet and food ration.
Compilation of data from individual stomachs to population level is not trivial and the method will depend highly on the question asked. Stomach data are often collected to get information on the average diet or food ration of a given species within a given area to inform multispecies assessment models (e.g. SMS or Gadget). Such models require data on diet and biomass eaten (food ration) by predator species and size classes. The average "population" diet or food ration should basically be calculated from a stratified mean of the individual stomach content samples, weighted by the strata density of the predator and the area of the strata. This seems simple, but incomplete and patchy sampling makes it often necessary to use a series of ad hoc solutions.
In general, a series of data compilation methods could be applied:
- Read and check data from the agreed exchange format;
- Bias correct to take into account variable evacuation rate;
- Assign size classes for predators and preys;
- Bias correct to take into account regurgitated stomachs within sample units;
- Aggregate stomach contents within sample_id and size classes.
- Allocate unidentified or partly identified prey items;
- Calculate population diet and food ration from a weighted average.
The sequence of these steps, of which some of the steps might be repeated, will depend on the individual sampling design and the quality of the analysed data.
The FishStomachs package defines data structures suitable for stomach data, and provides the necessary methods to compile observed stomach data into population diet and biomass eaten, used for multispecies models. The methods applied for a set of observations are stored within the data output to document the compilation steps taken.
With my experience from compiling stomach data for the North Sea for use in the MSVPA/SMS models, this document describes the methods above and how these can be applied.
1) Data exchange format
Data must be recorded, stored and exchanged in an agreed and documented format suitable for the purpose of sampling. This R-package uses a user defined format (see Table 1) with the possibility to rename variable names such that it is possible to include existing databases (e.g. ICES stomach contents database). The new format includes a number of mandatory (system defined) and optional (user defined) variable names, which uses "aliases" (See Table 2) as a link to "internal" variable names used for processing by FishStomachs.
Table 1. Database format. "field" is the variable name used in FishStomachs; "types" is the type of input(c=character, i=integer, d=double, l=logical); "mandatory" is the field required as minimum (T=TRUE, F=FALSE); "PRED" is the variable part of predator information; "PREY" is the information part of prey information; "units" is the unit or values applied. The data base format is specified in the file stomach_format.csv.
a<-readr::read_csv(file.path(system.file( package = "FishStomachs"),"stomach_format.csv"),col_types = readr::cols()) a<- dplyr::mutate_if(a,is.logical,as.character()) a<- dplyr::mutate(a,mandatory=dplyr::if_else(mandatory=='FALSE','F','T'), PRED=dplyr::if_else(PRED=='FALSE','F','T'),PREY=dplyr::if_else(PREY=='FALSE','F','T')) a<- dplyr::mutate(a,units=dplyr::if_else(is.na(units),' ',units)) knitr::kable(a[,4:10])
Table 2. Alias names accepted as default. Aliases can be changed and added by using a local version of the database format. The data base format and aliases are specified in the file stomach_format.csv.
a<-dplyr::mutate(a,alias_1=dplyr::if_else(is.na(alias_1),' ',alias_1),alias_2=dplyr::if_else(is.na(alias_2),' ',alias_2),alias_3=dplyr::if_else(is.na(alias_3),' ',alias_3)) knitr::kable(a[,1:4])
The presently used ICES data exchange format for stomachs (see Appendix 1 in this document) is based on the "Year of the stomach" sampling for North Sea stocks in the period 1981-1991 (see Appendix 2). The new version includes some extensions to better include stomachs sampled outside the "Year of the stomach", e.g. from national sampling programs with less standardized recording. Another change is that the new exchange format uses the WoRMS AphiaID (http://www.marinespecies.org/index.php) species taxonomic ode, where the old format used the NODC codes (https://www.nodc.noaa.gov/General/CDR-detdesc/taxonomic-v8.html).
In this document, the examples and variable names used are based on a data set on the ICES exchange format (Appendix 1).
2) Bias correction to take into account variable evacuation rate
It is often assumed that diet (eaten food) is the same as the observed stomach content. This is however not the case as the stomach evacuation rate of an individual stomach depends on e.g. the average meal size, average energy density of the meal, prey armament and temperature. Large energy rich meals will for example have a lower evacuation rate than small energy poor meals. Therefore, large energy rich meals will be present in a stomach for a longer time and thereby detectable for a longer time, than stomach contents from small energy poor meals.
In the two sub-sections below, it is described how to bias correct the effect of variable stomach evacuation rate and thereby how to bias correct the estimated average diet (or consumption rate) estimated from the observed stomach contents data. Ideally, this bias correction should take place at the level of sampling, i.e. by the individual stomach or by the pool of stomachs by sample site.
Estimation of food ration from stomach contents data
Food rations (evacuation rate of stomach contents) can be estimated from the observed stomach contents using the methods suggested by Andersen and Beyer (2005 a, b). This model takes into account the differences in evacuation rates between prey types due to their energy density and their resistance to digestion (armament).
Ration (R) (per hour) of prey group (i) for pooled stomachs is calculated from:
$$R = \sum_{i}^{}{(\rhoM_{i}b_{i}e^{\delta T}L^{\lambda}}E^{\xi})K\left(\frac{N_{A}}{N_{F}} \right)^{\alpha - 1}S^{\alpha}$$
M= armament of individual prey (group) i
b=proportion of prey (group) i
T= temperature (^o^C)
L= length (cm) of the predator
E= average energy density (kJ/g wet weight) of the stomach (or of the pooled stomachs)
N= Number of stomachs in the sample, total (A) and with food (F)
S = average stomach contents in grams
$\rho$, $\delta$, $\lambda$, $\xi$ and K are parameters to the model (Table 3)
Values for armament (M) and average energy density of individual prey (groups) as used in the compilation of the North Sea stomachs are presented in Table 5.
When data exist by the individual stomachs, the ration eaten per (feeding) fish can be calculated as:
$$R = \sum_{i}^{}{\rho * M_i b_i e^{\delta T} * L^{\lambda} * E^{\xi} * S^{\alpha}}$$
Table 3. Parameter values of the generic cylinder model of gastric evacuation used for the compilation of stomachs from predators in the North Sea.
| Predator | $\rho$ | $\lambda$ | $\delta$ | $\xi$ | $\alpha$ | K | |:---------|:--------|-----------|----------|:-----:|----------|:----:| | Cod | 0.00224 | 1.30 | 0.083 | -0.85 | 0.5 | 0.85 | | Haddock | 0.00191 | 1.30 | 0.083 | -0.85 | 0.5 | 0.85 | | Saithe | 0.00171 | 1.35 | 0.081 | -0.85 | 0.5 | 0.85 | | Whiting | 0.00171 | 1.35 | 0.081 | -0.85 | 0.5 | 0.85 | | Mackerel | 0.00174 | 1.30 | 0.080 | -0.85 | 0.5 | 0.85 |
Estimation of diet from stomach contents data
The method for bias correction in the calculation of food ration as outlined above can also be applied to calculate the ration of the individual prey item and thereby the average diet. For example, for individual stomachs the Ration (R) of prey item i becomes:
$$R_{i} = \rho\ M_i *b_i * e^{\delta T} * L^\lambda * E^{\xi}$$
Based on the set of $R_{i}$ for the individual predators from e.g. a sampling site (trawl haul) the average Ration of the of the individual prey item can be calculated by taking the number of feeding, empty and regurgitated stomachs into account. This average Ration by prey group is in this case expressed as g wet weight per hour (with $\rho$ included in the equation) and can easily be aggregated with average rations of preys from other strata. These rations can finally be transformed into diet data (weight proportions by prey item).
Table 4. Average energy density and armament of prey groups as applied in the compilation of stomach for use in the North Sea SMS.
a<-readr::read_csv(file.path(system.file( package = "FishStomachs"),"extdata","energy_density_and_armament.csv"),col_types = readr::cols()) knitr::kable(a,col.names=c("Species","Quarter","Lower length (mm)","Upper length (mm)","Energy density (kJ/g wet weight)", "Armament"))
3) Assign predator and prey size classes.
Stomach data can be sampled, analysed and recorded by the individual fish or by size group ("pooled stomach data") for the individual sample unit (sample_id), e.g. trawl haul. To save time during sampling and analysis of the stomach contents, the "Year of the stomach" data were sampled and analysed by pre-determined size classes (Table 4). Pooled stomachs do not provide information on the individual fish and on the variation in stomach contents, which is preferable in e.g. estimation of daily food ration and diet. Grouping of stomachs by size classes is however often necessary in the later compilation of data, as data are too scarce to be compiled by e.g. each cm group separately.
In general, I will suggest using size classes, e.g. the size classes used in 1991 by "Year of the Stomach" sampling (Table 3). That means that the variables are in the data exchange format.
-
pred_ll (predator lower length bound)
-
pred_lu (predator upper length bound)
-
prey_ll (prey lower length bound)
-
prey_lu (prey upper length bound)
are set from the length information on the individual predators and preys, if not already included in the data and defined by the sampling design.
Table 5. Size classes (for items larger than 5 cm) used for sampling and compilation of stomach data in the North Sea "Year of the stomach" 1991 sampling.
| Length class, 1991 | Length (cm) | |:-------------------|:------------| | 50 | 5-5.9 | | 60 | 6-6.9 | | 70 | 7-7.9 | | 80 | 8-9.9 | | 100 | 10-11.9 | | 120 | 12-14.9 | | 150 | 15-19.9 | | 200 | 20-24.9 | | 250 | 25-29.9 | | 300 | 30-34.9 | | 350 | 35-39.9 | | 400 | 40-49.9 | | 500 | 50-59.9 | | 600 | 60-69.9 | | 700 | 70-79.9 | | 800 | 80-99.9 | | 1000 | 100-119.9 | | 1200 | >120 |
4) Bias correction: Taking number of empty and regurgitated stomachs into account
Ideally each stomach should be classified as: with food, with food but regurgitated, with skeletal remains only or empty. This information is included in the presently used exchange format.
| Database variable | Comments | |:------------------|:-----------------------------------------------------------------------------------------------| | n_food | Stomachs with (recently ingested) food | | n_regur | Stomach with food, but evidence of regurgitation of parts or all the stomach contents. | | n_skel | Stomachs with practically indigestible remains and the fish is not considered feeding recently | | n_empty | No stomach contents |
Stomach contents from regurgitated (or everted) stomachs should not be included in the estimation of diet as the proportion regurgitated is unknown. The number of fish with regurgitated stomachs should however be recorded ("n_regur"). With the assumption that the regurgitated stomachs sampled within a "sample_id" have had the same stomach contents as the feeding (non-regurgitated) fish each observed prey item weight and number should be corrected by a factor to calculate the mean stomach contents of a predator within "sample_id".
$factor = \frac{(N_{F} + N_{R})}{N_{F}*(N_{F} + N_{SR} + N_{R} + N_{E})}$
Where
$N_{F}$ = Number of feeding fish ("n_food")
$N_{R}$ = Number of feeding fish, but regurgitated ("n_regur")
$N_{\text{SR}}$ = Number of non-feeding fish, stomachs containing only skeletal remains ("n_skel")
$N_{E}$ = Number of non-feeding fish with empty stomachs ("n_empty")
In practice, it is difficult to detect fish with regurgitated stomachs, and later data analysis may often show that there is a clear difference in the proportions of regurgitated stomachs between vessels surveying in the same area at the same time.
Another issue is how to distinguish "n_skel" from "n_food". In the original compilation of the 1991 "Year of the Stomach" (ICES, 1997), $N_{SR}$ was added to the numerator in the calculation of the factor above. However, I think "n_skel" should be considered as stomachs from not (recently) feeding fish and should be treated as empty stomachs, i.e. that the stomach contents should be disregarded in the compilation of data for feeding fish.
5) Aggregate stomach contents within sample_id and size classes
6) Allocate unidentified and partly identified prey items
Parts of the stomach contents are often not fully identified to species (group) or size class. With the assumption that the unidentified items follow the same species and size distribution as the identified prey items, it is possible to allocate the unidentified items proportionally to the identified species and size class. However, with the same assumptions it can be claimed that the unidentified items can be ignored for diet studies as the weight proportions of the individual food items would be the same with and without the unidentified items. I am not aware of published studies that quantify bias for the two approaches.
The allocation of unidentified items to identified items is not simple in practice. A stomach may for example contain remains of an unidentified fish and identified invertebrates. In such case, it would not be sensible to allocate the remains to the identified preys. However other stomachs from the same predator and size class from the same haul might include identified fish preys which could be used as a "distribution key" for unidentified items. If a given haul does not include identified preys for a "distribution key" an alternative key could be obtained from the neighbouring hauls or from another size classes within the haul.
It is not possible to give specific rules on how to allocate unidentified prey items, other than the distribution key should be derived from "local" data as far as possible. By "local" is meant from the same haul, time period, predator size class, ICES rectangle, roundfish area etc.
Assigning prey group
Stomach contents data may include prey identified to species species level, even for preys that are considered as "other prey" in the later data set produced for multi species assessments.
With 250.000 stomachs sampled in the North Sea as the basis for multispecies models a pragmatic approach was needed for allocation of unidentified items. The former used NODC codes for preys are not maintained anymore by NOAA, however NODC codes are still very useful for the compilation of stomach data as they are based on a 10-digit "intelligent" code. By "intelligent" code numbers it is meant that information about taxonomy was built into the codes through the use of 2-digit couplets to represent one or more levels of the taxonomic hierarchy. For example, a species assigned a 10-digit code would belong to the genus represented by the first 8 digits of the code. In the compilation of stomach data, this hierarchy makes it easier to assign (by an algorithm) partly identified food items, to the identified items within the same genus, family or order. Lookup tables exist in the ICES system to assign a NODC code to the WoRMS AphiaID code system presently used by ICES. The section "Compilation of stomach data for use in the North Sea SMS model" provides examples of how allocation of unidentified items could be done.
To be more specific. For the North Sea data compilation, the model prey species (e.g. cod, whiting, sprat) were maintained as named preys. The remaining prey items were grouped according to the table below. Invertebrates (NODC \< 8500000000) are treated as "other food" in the model and aggregated into one group. Two groups of prey items, Clupeidae and Gadoidae were maintained as they were used to name e.g. preys that most likely were sprat or herring (Clupeidae) but could not be determined to species level. "other fish" includes fish preys that are not either a Clupeidae or Gadoidae. The "unknown" group (NODC > 900000000) was used for unidentified prey remains (either fish or invertebrates).
| Species group | First NODC | Last NODC | |---------------|------------|------------| | other food | 400000000 | 8499999999 | | other fish | 8500000000 | 8747009999 | | Clupeidae | 8747010000 | 8747019999 | | other fish | 8747020000 | 8791029999 | | Gadidae | 8791030000 | 8791031999 | | Other fish | 8791032000 | 8999999999 | | unknown | 9000000000 | 9999999999 |
to allocate the partly identified preys, preys within the Clupeidae and Gadidae were first allocated to the identified species within the group (if such exists)
7) Aggregation of stomachs contents into population metrics
The average diet should basically be calculated as a stratified mean of the individual stomach content samples, weighted by strata abundances (indices) of the predator and strata areas. For the initial compilation of the "Year of the stomach" the average stomach contents were first calculated by ICES rectangles and finally aggregated to the population level (ICES, 1997). When there was more than one sample within a rectangle the rectangle value was calculated as the weighted mean of the sample (haul) average values, the weighting factors being the number of stomachs from each sample. Four methods were tried to calculate the "population" average stomach:
i. The rectangle values were averaged disregarding the number of stomachs sampled within each rectangle.
ii. The rectangle values were weighted by the sample size (total number of stomachs sampled within the rectangle).
iii. The rectangle values were weighted by the survey catch rates within the rectangle.
iv. The rectangle values were weighted by the square root of survey catch rates within the rectangle; this reduces the influence of occasional very large catch rates
The compilation (ICES, 1997) ended up using option iv for the predators cod, haddock, whiting and mackerel, while option ii was used for saithe because the samples (survey catches) were patchily distributed.
The compilation of data for the SMS model deviates from the original method used. Average stomach contents by ICES rectangle are derived in the same way as done initially, but the SMS method calculates then the weighted mean of stomach contents by ICES roundfish area using the square root of the mean rectangle survey cpue (option iv) as weighting factor (but option ii are used for saithe).
The calculation of the population diet for SMS was done as a weighted average of the strata (roundfish area) average diet weighted by the product of the average strata survey cpue and area of the strata.
References
Andersen N.G., Beyer J.E. 2005a. Mechanistic modelling of gastric evacuation applying the square root model to describe surface-dependent evacuation in predatory gadoids. J Fish Biol 67:1392--1412.
Andersen N.G., Beyer J.E. 2005b. Gastric evacuation of mixed stomach contents in predatory gadoids -- an expanded application of the square root model to estimate food rations. Journal of Fish Biology 67:1413--1433.
Berg, C. W., A. Nielson, and K. Kristensen. 2014. Evaluation of alternative age-based methods for estimating relative abundance from survey data in relation to assessment models. Fisheries Research 151:91--99.
Daan, N., Bromley, P.J., Hislop, J.R.G and Nielsen, N.A. 1990. Ecology of North Sea fish. Netherlands Journal of Sea Research, 26(2--4):343--386.
ICES. 1989. Database report of the Stomach Sampling Project 1981. Coop. Res. Rep. 164: 1--145.
ICES. 1997. Database report of the Stomach Sampling Project 1991. Coop. Res. Rep. 219: 1--442.
ICES. 1997. Report of the Multispecies Assessment Working Group, ICES headquarters, 11--19 August 1997. ICES CM 1997/Assess:16, 235 pp.
Pinnegar, J.K. (2014) DAPSTOM - An Integrated Database & Portal for Fish Stomach Records. Version 4.7. Centre for Environment, Fisheries & Aquaculture Science, Lowestoft, UK. February 2014, 39 pp.
\
Appendix 1. ICES data exchange format for stomach data (http://ices.dk/marine-data/data-portals/Pages/Fish-stomach.aspx )
+-----------------------------------+-----------------------------------+ | Field | Description | +-----------------------------------+-----------------------------------+ | Dataset | Dataset name | +-----------------------------------+-----------------------------------+ | RecordType | SS for single stomach | +-----------------------------------+-----------------------------------+ | Country | Country that collected the data | +-----------------------------------+-----------------------------------+ | Ship | Vessel that collected the data | +-----------------------------------+-----------------------------------+ | Latitude | Data sampling position -- | | | latitude | +-----------------------------------+-----------------------------------+ | Longitude | Data sampling position -- | | | longitude | +-----------------------------------+-----------------------------------+ | Estimated_Lat_Long | Flag whether the sampling | | | position based on the reported | | | area | +-----------------------------------+-----------------------------------+ | ICES_StatRec | ICES statistical rectangle | +-----------------------------------+-----------------------------------+ | ICES_AreaCode | ICES area code | +-----------------------------------+-----------------------------------+ | Year | YYYY | +-----------------------------------+-----------------------------------+ | Month | MM | +-----------------------------------+-----------------------------------+ | Day | DD | +-----------------------------------+-----------------------------------+ | Time | Sampling time like HHMM | +-----------------------------------+-----------------------------------+ | Station | Station reference | +-----------------------------------+-----------------------------------+ | Haul | Haul number | +-----------------------------------+-----------------------------------+ | Sampling_Method | Predator sampling method | +-----------------------------------+-----------------------------------+ | Depth | Sampling depth | +-----------------------------------+-----------------------------------+ | Temperature | ^o^ C | +-----------------------------------+-----------------------------------+ | SampleNo(FishID) | Predator reference code | | | | | | Fish ID unique for Country, year, | | | quarter and ship | +-----------------------------------+-----------------------------------+ | ICES_SampleID | ICES predator reference | +-----------------------------------+-----------------------------------+ | Predator_AphiaID | Predator WoRMS AphiaID | +-----------------------------------+-----------------------------------+ | Predator_LatinName | Predator taxon Latin Name | +-----------------------------------+-----------------------------------+ | Predator_Weight(mean) | (Mean) predator weight | +-----------------------------------+-----------------------------------+ | Predator_Age(mean) | (Mean) predator age | +-----------------------------------+-----------------------------------+ | Predator_Lengh(mean) | (Mean) predator length | +-----------------------------------+-----------------------------------+ | Predator_LowerLengthBound | Predator's length lower bound | +-----------------------------------+-----------------------------------+ | Predator_UpperLengthBound | Predator's length upper bound | +-----------------------------------+-----------------------------------+ | Predator_CPUE | Predator catch per hour | +-----------------------------------+-----------------------------------+ | GallBladder_stage(class) | Gall bladder stage | +-----------------------------------+-----------------------------------+ | Stomach_METFP | Method of stomach preservation | +-----------------------------------+-----------------------------------+ | Stomach_TotalNo | Total number of stomachs in the | | | pool -- for single stomachs | | | always 1. | +-----------------------------------+-----------------------------------+ | Stomach_WithFood | Number of stomachs with food | +-----------------------------------+-----------------------------------+ | Stomach_Regurgitated | Number of stomachs regurgitated | +-----------------------------------+-----------------------------------+ | Stomach_WithSkeletalRemains | Number of stomachs with skeletal | | | remains | +-----------------------------------+-----------------------------------+ | Stomach_Empty | Number of empty stomachs | +-----------------------------------+-----------------------------------+ | Stomach_ContentWgt | Stomach content weight | +-----------------------------------+-----------------------------------+ | Stomach_EmptyWgt | Stomach empty weight | +-----------------------------------+-----------------------------------+ | Stomach fullness | Stomach fullness | +-----------------------------------+-----------------------------------+ | Stomach_Item | Stomach item name (see Appendix | | | B) | +-----------------------------------+-----------------------------------+ | ICES_ItemID | ICES stomach item ID | +-----------------------------------+-----------------------------------+ | Prey_AphiaID | Prey WoRMS AphiaID | +-----------------------------------+-----------------------------------+ | Prey_LatinName | Prey taxon Latin Name | +-----------------------------------+-----------------------------------+ | Prey_IdentMet | Prey identification method | +-----------------------------------+-----------------------------------+ | Prey_DigestionStage | Prey digestion stage | | | | | | 0= Intact prey (skin, fins, legs | | | and flesh is complete), 1= | | | partially digested prey (prey in | | | more advanced stages of | | | digestion), 2= partially digested | | | prey (prey in more advanced | | | stages of digestion), 3= skeletal | | | material (no flesh, only bones, | | | shells, otoliths) | +-----------------------------------+-----------------------------------+ | Prey_TotalNo | Total number of preys | +-----------------------------------+-----------------------------------+ | Prey_Weight | Prey weight in grams | +-----------------------------------+-----------------------------------+ | Prey_LengthIdentifier | Prey length identifier | +-----------------------------------+-----------------------------------+ | Prey_Length | Prey length in cm | +-----------------------------------+-----------------------------------+ | Prey_LowerLengthBound | Prey length lower bound | +-----------------------------------+-----------------------------------+ | Prey_UpperLengthBound | Prey length upper bound | +-----------------------------------+-----------------------------------+ | Prey_MinNo | Minimum number of preys | +-----------------------------------+-----------------------------------+ | Remarks | Any relevant comments | +-----------------------------------+-----------------------------------+
\
Appendix b. Data exchange format used by the North Sea stomach sampling program 1981-1991,
Source ICES Coop. Res. Rep. No. 219 , pp 421-422
| Position | Name | Type^a^ | Range of values | Comment |
|:-------------|:-----------------------------|:------------|:--------------------|:-------------------------------------------------|
| 1-2 | Ecosystem name | 2N | 1-8 | Footnote^b^ for coding scheme |
| 3-4 | Year | 2N | 01-99 | Year - 1900 |
| 5 | Quarter | 1N | 1-4 | X |
| 6-9 | Square | 4AN | | ICES rectangle |
| 10-19 | Predator code | 10N | | NODC 10-digit |
| 20-24 | Sample number | 5N | 1-99999 | Unique fish ID |
| 25-27 | Country | 3A | | ICES alpha codes, No
a) All numeric fields (N) right justified, zero filled; all apha (A) and mixed alpha/numeric fields (AN) left justified, space filled
b) North Sea=l, Baltic Sea=2, Barents Sea=3, Iceland=4, Northeastern Newfoundland=5, Northeastern USA=6, Southern Shelf=7, Faroes=8
c) DEM=demersally caught by trawling or seining gears PEL=pelagically caught by trawling or seining gears DHL=demersal hook and line
PHL=pelagic hook and line DGN=demersal gill net PGN=pelagic gill net
d) Quadrants for position identification are: l=NE, 2=NW, 3=SW, 4=SE (The axes of the quadrants are the equator and the Greenwich meridian)
e) Digestion stages are: 0 =Pristine, 1 =Affected by digestion, 2=Skeletal remains, 9=Unknown
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.