Using zooMizer as zooplankton component in mizer #6
Description
I have thought a bit about how to use a zooMizer model as a zooplankton component in a Mizer fish model. It turns out that the new mizer extension mechanism is perfect for this.
ZooMizer needs to supply three functions. First of all, a function for setting up a MizerParams object describing the zooplankton model. Here is the skeleton:
#' Set up MizerParams object for zooplankton model
#'
#' @param group TODO Patrick: document this
#' @param input
#'
#' @return A MizerParams object describing the zooplankton model
newZooMizerParams <- function(groups, input) {
# This is essentially the current function fZooMizer_run() but without
# actually running the model
# TODO Patrick: put the code here
}
The idea is then to use this zooMizer model as a component of a mizer fish model. This would be set up as follows:
zoo_params <- newZooMizerParams()
fish_params <- NS_params # or any other mizer params object
fish_params <- fish_params %>%
setComponent(component = "zoo",
initial_value = zoo_params@initial_n,
dynamics_fun = "zoo_dynamics",
component_params = zoo_params) %>%
setResource(resource_dynamics = "resource_zooMizer")
This registers two other functions that zooMizer needs to supply, one to run the zooplankton dynamics and one to aggregate all the plankton spectra into a single resource spectrum for the fish model.
#' Project zooplankton forward by one fish time step
#'
#' This is called by the fish model at each fish time step to update its
#' zooplankton component.
#'
#' @param params The MizerParams object describing the fish model
#' @param n_other A list containing the current values of all other ecosystem
#' components. Only `n_other@zoo` will be needed by this function. It holds
#' the array (type x size) with the current zooplankton abundances.
#' @param rates A list of rate functions calculated by the fish model, see
#' [mizer::getRates()] for details. Probably only `` will
#' be needed.
#' @param t The current time
#' @param dt The amount of time to project forward. This is the fish time step
#' which will be much larger than the plankton time step.
#' @param ... Unused other parameters
#'
#' @return An array (type * size) with the updated zooplankton abundances at
#' time `t + dt`.
zoo_dynamics(params, n_other, rates, t, dt, ...) {
# get the MizerParams object for the zooplankton
zoo_params <- other_params(params)$zoo
# get predation mortality imposed by the fish
mort_from_fish <- rates$resource_mort
# TODO Patrick: use this mortality to set the mortality in zoo_params
# TODO Patrick: choose the plankton step size as a divisor of fish dt
steps <- 30
zoo_dt <- dt / steps
# get array (type x size) with the current zooplankton abundances
n <- n_other@zoo
l <- new_project_simple(zoo_params, n = n,
n_pp = zoo_params@initial_n_pp,
n_other = zoo_params@initial_n_pp,
t = t, dt = zoo_dt, steps = steps,
effort = list(),
resource_dynamics_fn = resource_constant,
other_dynamics_fns = list(),
rates_fns = lapply(zoo_params@rates_funcs, get))
return(l$n)
}
#' Return total plankton spectrum
#'
#' This function will be used as the resource dynamics function by the fish
#' model. It is used at each fish time step to update the total plankton
#' abundance used by the fish model. However this function does not have to
#' calculate any dynamics because that will already have been done by
#' [zoo_dynamics()]
#'
#' @param params The MizerParams object describing the fish model.
#' @param n_other A list containing the current values of all other ecosystem
#' components. Only `n_other$zoo` will be needed by this function. It holds
#' the array (type x size) with the current zooplankton abundances.
#' @param ... Unused other parameters
#'
#' @return A vector with the total plankton abundances, aggregated over all
#' plankton types, with one entry for each weight in `params@w_full`
resource_zooMizer(params, n_other, ...) {
# get the MizerParams object for the zooplankton
zoo_params <- other_params(params)$zoo
# phytoplankton abundances stay constant
n_pp <- zoo_params@initial_n_pp
# get array (type x size) with the current zooplankton abundances
# and aggregate over all types
n <- colSums(n_other$zoo)
# TODO Patrick:
# Now add n and n_pp and put it into a vector of the right length
}