Skip to contents

Add fish traits (FishBase)

Source: R/add-fishbase.R
add_fishbase.Rd

Joins FishBase morphological and ecological traits to a taxify() result by looking up accepted_name.

Usage

add_fishbase(x, cols = NULL, verbose = TRUE)

Arguments

x

A data.frame returned by taxify().

cols

Which columns to attach: NULL (default) the curated set, "all" every column the source carries, or a character vector of names. See enrichment_cols.

verbose

Logical. Default TRUE.

Value

The same data.frame with additional columns:

fb_body_length_cm

Maximum body length in centimetres.

fb_body_mass_g

Maximum published weight in grams (FishBase SPECIES.Weight). This is a record maximum, not a typical or adult-mean mass.

fb_trophic_level

Trophic level.

fb_depth_min_m

Minimum depth in metres.

fb_depth_max_m

Maximum depth in metres.

fb_vulnerability

Vulnerability index (0–100).

fb_habitat

Habitat type (e.g. demersal, pelagic).

fb_importance

Commercial importance category.

fb_lw_a

Coefficient a of the length-weight relationship W = a * L^b (weight in g, length in cm of type fb_lw_type), from FishBase's POPLW table.

fb_lw_b

Exponent b of the length-weight relationship.

fb_lw_type

Length convention the coefficients were fitted against (TL total, SL standard, FL fork, WD width, ...). Chosen to match the species' maximum-length type where recorded, so it applies to fb_body_length_cm. Applying a coefficient to a different length type is a silent error.

fb_lw_method

How the fit was obtained (e.g. "type I linear regression", "single L-W pair with b=3").

fb_lw_sex

Sex the fit applies to (unsexed, mixed, female, male, juvenile).

fb_lw_n

Sample size the fit was based on.

fb_lw_r2

Coefficient of determination (r^2) of the fit.

Details

Source: FishBase via rfishbase (Froese & Pauly, CC BY-NC 4.0). Coverage: ~35k fish species. Fishes only.

The fb_lw_* columns give one representative length-weight fit per species from the POPLW table, so a length can be converted to a mass where fb_body_mass_g (a record maximum) is not what you want. FishBase's Bayesian congeneric estimates are not included; fb_lw_* are measured fits.

The build-from-source fallback requires the rfishbase package (available on CRAN). Pre-built .vtr files do not require rfishbase.

References

Froese R, Pauly D (eds.) (2024) FishBase. World Wide Web electronic publication, https://www.fishbase.org.

Examples

# Runs offline against the bundled example database.
old <- options(taxify.data_dir = taxify_example_data())

taxify("Gadus morhua", backbone = "gbif") |>
  add_fishbase()

options(old)