simulation.experiment: Runs simulation experiments with VirtualCom
In VirtualCom: Assembling virtual plant communities

Description Usage Arguments Details Value See Also Examples

This function creates (1) a species pool (see create.pool), (2) assembles communities (see tamaure) and (3) computes functional and phylogenetic diversity indices. After natural communities are assembled invasion can occur.

1	simulation.experiment(parameters)

parameters

The only argument is a named vector of input parameter values, the entries are given in details

n.species.pool: number of species in the species pool (including native and invasive species)
evol.model: choice of the trait evolution model (see also create.pool)
evol.model.param: parameterization of the trait evolution model (see also create.pool)
species.pool.abundance: vector of abundances in the species pool; if NA then all species are equally abundant
min.phyl.signal: minimum value of phylogenetic signal in species pool (see also create.pool)
years: number of simulated timesteps; If plots=TRUE equilibrium conditions can be checked by eye
n.communities: number of simulated communities
env: value of the environment in the simulated community (e.g. temperature)
K: value of carrying capacity, i.e. number of individuals in the community
niche.breadth: value of standard deviation of the Gaussian distribution that describes the niche traits (identical for all species)
beta.env: value of the strength of the environmental filter (see details)
beta.comp: value of the strength of the competition filter (see details)
beta.abun: value of the strength of the recruitment filter, i.e. the advantage of already being present in the community (see also tamaure)
invasion.time: number of time-steps to simulate invasion (after native community has established); if 0 than there is no invasion
n.invader.pool: number of invaders in species pool
InvDistrib: The invader's distribution in the phylogney can be either clusterd ("1"), random ("2") or overdispersed ("3)
n.rep.null.model: number of repetitions for null models to test the diversity indices describing community composition
null.model: different null models can be choosen, see details

The following null models can be choosen to test for the structure of the native community: "1" = taxa.labels,"2" = richness, "3" = frequency, "4" = independentswap, "5" = trialswap. These models are explained in more detail in the function ses.mpd in the package picante. The output lists for natives and invaders consist of 5 matrices describing the diversity structure. They contain information on (1) observed diversity, and based on the chosen null models: (2) z-scores, (3) ranks, (4) mean of the null distribution and (5) standard deviation of the null distribution. Each matrix gives this information for all sites and for a set of diversity indices (mpd, mntd, ...).

Information on input parameters, species pool, native assembly structure and invader assembly structure are returned as a list:

parameter: a vector of the input parameters
pool: a list of different objects that define the species pool. "func" is a dataframe of trait values and invader identity. "phylo" is a phylo object. "invader" is a list of the invader's ID, their niche optima and their performances given the environment in the community
natives: a list of the native commuity structure and diversity measures. "all.abundances" is a site-by-species matrix of abundances. "indices" is a list of matrices containing functional and phylohgenetic diversity information (see details)
invaders: a list of community structure and diversity measures after the invasion processes, see "natives" for the structure

tamaure for the community assembly function

# load pre-prepared parameter table
  data(simple_param)
  # run the first line of the parameter table
  single.run <- simulation.experiment(simple_param[1,]) # run the first line
  str(single.run)

  # setting up a full experiment (this may take a few minutes)
  wrapper <- function(a){
  library(VirtualCom)
  data(simple_param)
  print(a)
  return(try(simulation.experiment(simple_param[a,])))
  }

  # running the experiment either with lapply
  output <- lapply(1:nrow(simple_param), wrapper)

  # # or running the experiment with sfLapply using snowfall to allow for parallel computing
  # require(snowfall)
  # sfInit(parallel=TRUE, cpus=11)
  # output <- sfLapply(1:nrow(simple_param), wrapper)
  # sfStop()

  # extract results from list
  result.table <- get.results(output=output, myvar="obs", invader="FALSE")
  result.table$process <- ifelse(result.table$beta.env==0 & result.table$beta.comp==0, "Random", ifelse(result.table$beta.env!=0 & result.table$beta.comp==0, "Env", ifelse(result.table$beta.env==0 & result.table$beta.comp!=0, "Comp", "Both") ))

  # plot the funtional diversity (mpd) of communities in dependence on assembly processes
  require(ggplot2)
  ggplot(data=result.table, aes(x=process, y=FD_ab_mpd)) + geom_boxplot(width=0.8) +  xlab("Assembly rules") +  ylab("Functional diversity")