Create Structural Equation Model (SEM) Equations and Time Series Data

This function prepares the inputs for a Dynamic Structural Equation Model (DSEM). It generates the specific path equations (drivers to targets) and compiles the corresponding time-series data based on environmental links and diet composition thresholds.

Usage

create_dsem_inputs(
  data,
  focal_functional_groups,
  diet_composition_threshold = 0.1
)

Arguments

data: A tibble/list containing data_om, data_environment, and data_diet_composition.
focal_functional_groups: Character vector specifying one or more focal groups of interest (e.g., c("menhaden 0", "meanhaden 1")).
diet_composition_threshold: Numeric. The minimum proportion of diet required to include a predator-prey link in the SEM. Default is 0.1, which represent 10% of the diet.

Value

A tibble with four columns:

data_time_series_all: The complete matrix of available time series.
data_time_series_sem: The subset of time series used specifically in the SEM.
sem_tibble: A detailed tibble of drivers, targets, lags, and link types.
sem_lines: A character string formatted for the dsem package.

Details

The function constructs three types of linkages:

Environmental: Links environmental indices to impacted functional groups.
Bottom-up: Links prey biomass to the focal functional group.
Top-down: Links predator biomass to the focal functional group.

Examples

# Create mock data for demonstration
data_om <- tibble::tibble(
  year = rep(1984, 2),
  month = rep(1, 2),
  functional_group = c("Phytoplankton", "Menhaden (0yr)"),
  functional_group_snake_case = c("phytoplankton", "menhaden_0yr"),
  type = "biomass",
  value = c(100, 50)
)

data_environment <- tibble::tibble(
  index = "amo",
  year = 1984,
  month = 1,
  value = 0.679,
  unit = "NA",
  lag_months = 1,
  impacted_group = "Menhaden (0yr)",
  impacted_group_snake_case = "menhaden_0yr"
)

data_diet_composition <- tibble::tibble(
  prey = "Phytoplankton",
  predator = "Menhaden (0yr)",
  proportion = 0.5,
  prey_snake_case = "phytoplankton",
  predator_snake_case = "menhaden_0yr"
)

data <- list(data_om = list(data_om), data_environment = list(data_environment),
             data_diet_composition = list(data_diet_composition))

# Generate DSEM inputs
dsem_inputs <- create_dsem_inputs(data, "Menhaden (0yr)", 0.1)

# Extract the SEM equations for use with the dsem package
cat(dsem_inputs$sem_lines)
#> amo -> menhaden_0yr, 1, amo_menhaden_0yr
#> phytoplankton -> menhaden_0yr, 0, phytoplankton_menhaden_0yr

# Extract the time series data required for the SEM
print(dsem_inputs$data_time_series_sem[[1]])
#> Time Series:
#> Start = 1 
#> End = 1 
#> Frequency = 1 
#>   year month   amo phytoplankton menhaden_0yr
#> 1 1984     1 0.679           100           50