In Jpomz/IGPtoy: Simulate Intraguild Predation Dynamics in Three Species Community

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)
library(IGPtoy)

IGPtoy: functions to simulate trophic dynamics in 3-species communities in spatially explicit patches.

Justin Pomeranz r format(Sys.time(), '%d %B, %Y')

• Overview
• Installation
• Instruction
• igp_sim()
  • Basic usage
  • Quick start
  • Custom setting: visualization
  • Custom setting: trophic parameter
  • Custom setting: disturbance
  • Custom setting: productivity
• Using with mcbrnet
  • Custom setting: combine brnet() and igp_sim()
  • Model description

Overview

This is a developmental package for simulation models for Intra-Guild Predation (IGP). It will be rolled into the mcbrnet package upon completion.

• igp_sim: Function igp_sim() simulates trophic dynamics within 3-species communities. By default, it produces a square landscape with patches randomly distributed by sampling x- and y- coordinates from a uniform distribution. Habitat structure can be controlled by supplying a distance matrix. This function follows the framework of Zhou et al. (2013), with the addition of a 3rd species and intraguild predation. See model description below for further details.

installation

You can install the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("Jpomz/IGPtoy")
library(IGPtoy)

Instruction

igp_sim()

basic-usage

The key argument is the number of patches (n_patch). Trophic dynamics are simulated through 1) local dynamics, and 2) dispersal (emigration and immigration).

Example:

igp <- igp_sim(n_patch = 10)

The function returns:

• sp_dynamics a data frame containing simulated trophic dynamics.
• patch: patch ID
• time: time-step
• basal_k: carrying capacity, K, for the basal species, B
• disturbance: Binary indicator variable for if a disturbance occured (1) or did not occur (0) in that patch at that time-step
• fcl: food chain length in that patch at that time-step. Values (0, 1, 2, 3)
• obs_P_pref: The observed predator preference for resource B in that patch at that time-step
• species: Species ID (B, C, P)
• value: Population abundance for a species
• df_patch a data frame containing patch attributes
• patch_id: patch ID
• disturb_value: disturbance level (i.e., proportional mortality), given a disturbance occurs, in each patch.
• basal_k: carrying capacity, K, for the basal species, B
• dens_dep_reg: density dependence regulation variable = (r_max - 1) / K
• fcl_state_prop a data frame containing the proportion of patches that are in each FCL "state" at each time-step.
• time: time-step
• fcl_state: character string indicating FCL state p0 = No species, p1 = B, p2 = B + C, p2.5 = B + P, p3 = B + C + P
• value: numeric value from 0 to 1, indicating the proportion of patches in a given fcl state.
• sim_params A data frame containing all of the simulation parameters

quick-start

The following script simulates trophic dynamics with n_patch = 5. By default, igp_sim() uses 200 time-steps for burn-in and records trophic dynamics for 1000 time steps.

igp <- igp_sim(n_patch = 10)

A named list of return values:

igp

custom-setting-visualization

Users can visualize the simulated dynamics in 5 random patches using plot_patch_dynamics = TRUE. In addition, the proportion of FCL state in all patches across time can be visualized using plot_fcl = TRUE

igp <- igp_sim(n_patch = 10, plot_patch_dynamics = TRUE, plot_fcl = TRUE)

custom-setting-trophic-parameters

Users may use the following arguments to customize trophic dynamic simulations regarding 1) predation / search effort, 2) conversion efficiency, 3) productivity, and 4) dispersal ability.

Predation

Arguments alphabc, betabc and alphap and betap control the search effort and predation rate for species C and P, respectively. Argument P_pref controls the preference of predator P for resource B.

The beta variables control how fast the predation curve reaches the asymptote, and alpha / beta determines the location of the asymptote. See Model Description for more details. Default values for alpha and beta parameters are 4 and 20, respectively.

P_pref: default is NULL, and preference is calculated for each patch at every time step based on prey resource availability (see model description for more details). If a value is supplied, it is assumed constant across space and time and ignores resource availability.

Default values are set such as to make coexistence between all three species likely. For example, coexistence is more likely if C outcompetes P for resource B. We can make P a superior competitor by setting alphap = 10:

igp_alpha10 <- igp_sim(alphap = 10, plot_patch_dynamics = TRUE)

Here, C generally goes locally extinct, and B and P are prone to extreme oscillations

Conversion efficiency

Arguments ebc, ebp, and ecp, control the efficiency of converting consumed resources into new consumer species. Parameters have the form e_ij where i is the resource species and j is the consumer species. The default conversion efficiency values are set at 2.

Productivity of basal species, B

Arguments k_function, k_base, and r_max control the productivity level for the basal species B.

k_function can be one of (NULL, "environment", "patches-upstream"). If k_function = NULL, carrying capacity does not vary across patches, and is fixed as k_base. k_function = "environment" : Carrying capacity varies across patches according to an environmental value. The observed carrying capacity, K, for a patch is scaled from 50% to 150% of k_base value, based on the environmental value. This is designed to work in conjunction with the brnet() function from the mcbrnet package (See Custom setting combine brnet and igp_sim below). For 2-dimensional networks (i.e., when a distance matrix is not supplied), an environmental value for each patch is sampled randomly from a normal distribution with mean = 0, and sd = 1. E.g., there is no spatial correlation in environmental values in 2D networks.

r_max: controls the maximum reproductive value of the basal species, B.

igp1 <- igp_sim(n_patch = 100, k_function = NULL)
igp2 <- igp_sim(n_patch = 100, k_function = "environment")

c(mean(igp1$df_patch$basal_k), sd(igp1$df_patch$basal_k))
c(mean(igp2$df_patch$basal_k), sd(igp2$df_patch$basal_k))
#plot(density(igp1$df_patch$basal_k))
#plot(density(igp2$df_patch$basal_k))

Dispersal ability

Arguments p_dispersal and theta.

p_dispersal controls the probability of an individual emigrating from a patch. This can be thought of as the proportion of the population emigrating at the beginning of each time step. If one value is supplied, it is assumed the same for all three species. If a vector of length = 3 is supplied, this sets the probability for B, C and P, respectively. p_dispersal does not vary across patches or through time.

theta: controls the rate of exponential decay for distance travelled for emigrants according to $theta$^-1}. i.e., theta = 0.1 emigrants travel far, theta = 10 emigrants essentially do not travel.

WIP, below not finished

custom-setting-disturbance

custom-setting-productivity

mcbrnet

custom-setting-combine-brnet-and-igp_sim

model-description

The number of prey consumed follows a Type II response according to $$W_{ij} = \frac{\alpha_{ij} N_{i} N_{j}} {\beta_{ij} N_{j} + N_{i}}$$

Jpomz/IGPtoy documentation built on Aug. 2, 2021, 5:28 a.m.