epizootic
is an extension to poems
, a spatially-explicit,
process-explicit, pattern-oriented framework for modeling population
dynamics. This extension adds functionality for modeling disease
dynamics in wildlife. It also adds capability for seasonality and for
unique dispersal dynamics for each life cycle stage.
You can install the latest release on CRAN with:
install.packages("epizootic")
You can install the latest version of epizootic from GitHub with:
# install.packages("devtools")
install.packages("poems")
devtools::install_github("viralemergence/epizootic")
Because epizootic
is an extension to poems
, it is necessary to
install poems
first.
poems
and epizootic
run on
R6 classes. R is
primarily a functional programming language in which the primary units
of programming are expressions and functions. Here we use R6 to create
an object-oriented framework inside of R. R6 classes such as
DiseaseModel
and SimulationHandler
are used to store model
attributes, check them for consistency, pass them to parallel sessions
for simulation, and gather results and errors.
Here is the initial state of an idealized theoretical disease scenario, following a SIR disease model with three life cycle stages: juvenile, yearling, and adult.
library(poems)
library(purrr)
library(epizootic)
example_region <- Region$new(coordinates = data.frame(x = rep(seq(177.01, 177.05, 0.01), 5),
y = rep(seq(-18.01, -18.05, -0.01), each = 5)))
initial_abundance <- c(c(5000, 5000, 5000, 1, 0, 0, 0, 0, 0),
rep(c(5000, 5000, 5000, 0, 0, 0, 0, 0, 0), 24)) |>
matrix(nrow = 9)
example_region$raster_from_values(initial_abundance[3,]) |>
raster::plot(main = "Susceptible Adults")
example_region$raster_from_values(initial_abundance[4,]) |>
raster::plot(main = "Infected Juveniles")
Here I create a DiseaseModel
object, which stores inputs for disease
simulations and checks them for consistency and completeness.
model_inputs <- DiseaseModel$new(
time_steps = 10,
seasons = 2,
populations = 25,
stages = 3,
compartments = 3, # indicates disease compartments
region = example_region,
initial_abundance = initial_abundance,
# Dimensions of carrying_capacity are populations by timesteps
carrying_capacity = matrix(100000, nrow = 25, ncol = 10),
# Indicates length of breeding season in days for each population
breeding_season_length = rep(100, 25),
# One mortality value for each stage and compartment
mortality = c(0.4, 0.2, 0, 0.505, 0.25, 0.105, 0.4, 0.2, 0),
# Indicates that these are seasonal mortality values
mortality_unit = 1,
# No reproduction in this simple example
fecundity = 0,
fecundity_unit = 1,
fecundity_mask = rep(0, 9),
# Transmission rates from infected individuals, one for each stage
transmission = c(0.00002, 0.00001, 7.84e-06),
# Indicates that these are daily transmission rates
transmission_unit = 0,
# Indicates that all stages in the first compartment, S, can be infected
transmission_mask = c(1, 1, 1, 0, 0, 0, 0, 0, 0),
recovery = c(0.05714286, 0.06, 0.1),
recovery_unit = 0,
# Indicates that all stages in the second compartment, I, can recover
recovery_mask = c(0, 0, 0, 1, 1, 1, 0, 0, 0),
season_functions = list(sir_model_summer, NULL),
dispersal = list(disperser),
simulation_order = list(c("transition", "season_functions", "results"),
c("dispersal", "results")),
verbose = F
)
model_inputs$is_complete()
#> [1] TRUE
model_inputs$is_consistent()
#> [1] TRUE
The core simulation engine of epizootic
is the function
disease_simulator
, which simulates spatially explicit disease dynamics
in populations. Here I show the results from the non-breeding season in
the tenth year of the simulation.
results <- disease_simulator(model_inputs)
results$abundance_segments$stage_3_compartment_1[,10,2] |>
example_region$raster_from_values() |>
raster::plot(main = "Susceptible Adults")
results$abundance_segments$stage_3_compartment_2[,10,2] |>
example_region$raster_from_values() |>
raster::plot(main = "Infected Adults")