7.3. Background species

7.3.1. Introduction

Background species have first been introduced in Osmose by Fu Caihong (see [FOT+17]). They can be viewed as an intermediary between focal species (i.e. species of interest, whose full life cycle is simulated) and lower trophic levels (plankton for instance). They differ from lower trophic levels since they can predate on and be predated by focal species.

Background species are activated by setting the parameter simulation.nbackground to values greater than 0.

7.3.2. BackgroundSpecies class

To manage background species, a new class has been implemented (BackgroundSpecies.java). It contains the following attributes:

  • index of the background species (used in the configuration files)

  • ByClassTimeSeries object, which allows to extract biomass time series by time-step and by size class. It is defined by the "biomass.byDt.bySize.file.bkgX parameter.

Warning

For background species, the size classes are fixed!

  • The name of the background species (species.name.bkgX parameter)

  • Allometric factors for size/weight conversions (species.length2weight.condition.factor.bkgX and species.length2weight.allometric.power.bkgX parameters).

  • Distribution maps. They can be provided as CSV files, in the same way as focal species movements (see Section 4.8), with parameters starting with movement.bkgspecies. They can also be defined as NetCDF files (one per background species), which should have the following format:

netcdf movement_bkg0 {
dimensions:
        m = 3 ;
        y = 22 ;
        x = 45 ;
        t = 24 ;
variables:
        int yearmax(m) ;
        double map(m, y, x) ;
                map:_FillValue = NaN ;
        int class(m) ;
        int yearmin(m) ;
        byte season(m, t) ;
}
  • Trophic levels for each size class (species.trophiclevel.bkgX parameters).

In addition, the background species need to have the predation parameters that are also used by focal species ( predation.accessibility.stage.threshold.bkgX predation.efficiency.critical.bkgX predation.ingestion.rate.max.bkgX predation.predPrey.stage.threshold.bkgX predation.predPrey.sizeRatio.max.bkgX predation.predPrey.sizeRatio.min.bkgX, see Section 4.10)

7.3.3. Accessibility matrix

When using background species, the accessibility matrix must be changed accordingly. It must always have the following form:

Focal (pred)

Background (pred)

Focal (prey)

Background (prey)

LTL (prey)

7.3.4. Background biomass

Background species biomass is defined from a biomass time series (one per species and per size class) and by distribution maps. The distribution maps contain defines the distribution of the background species over space. They contain float values, which are normalize, so that the integral over space equals one:

\[\sum_{k=0}^{N_{cell}-1} D_k = 1\]

Before each time step, the background species biomass is reset by multiplying the biomass time series by the map distribution factor. Because of the normalisation, the spatially integrated biomass is equal to the biomass provided in the time-series.

7.3.5. Mortality

Background species come into play in the predation mortality process.

Warning

Background species have only be implemented in the stochastic mortality algorithm (StochasticMortalityProcess.java).

For each cell:

  • The list of background species schools is recovered. As for LTL, it is assumed that each cell contain one unique school for each background species and for each size class. This list is added to the list of potential preys (i.e. school of focal species, swarms of ltl that belong to the cell).

  • For all the background schools, the accessibility to all the preys is estimated.

  • The biomass of all the schools is initialized (see Section 7.3.4).

  • A random draft of predators (both focal and background species) is performed. And for each predator, the predation mortality toward each prey (focal, background and ltl species) is computed.

  • The number of dead individuals are then removed.