Dynamic stability analysis for community time series

compute_dynamic_stability() runs the full dynamic stability analysis. The analysis has multiple steps:

1. run simplex projection on each time series to identify the optimal embedding dimension

2. generate surrogate time series, assumming that the data have just a seasonal pattern

3. run ccm on each pairwise interaction, including the surrogate data

4. identify the significant interactions by comparing the CCM for the real time series against the calculations for the surrogate data

5. run S-map models for each time series, using the appropriate number of lags, and including the important interacting variables

6. extract out the s-map coefficients from the models and assemble matrices for the system

7. perform eigen-decomposition on the s-map coefficient matrices

8. perform singular-value-decomposition on the s-map coefficient matrices

9. compute volume contraction from the s-map coefficient matrices

10. compute total variance from the s-map coefficient matrices

build_dynamic_stability_plan() creates a drake plan for the dynamic stability analysis.

compute_dynamic_stability(block, results_file = NULL, id_var = NULL,
  max_E = 16, E_list = seq(max_E),
  surrogate_method = "annual_spline", num_surr = 200,
  surr_params = list(), lib_sizes = seq(10, 100, by = 10),
  random_libs = TRUE, num_samples = 100, replace = TRUE,
  RNGseed = 42, silent = TRUE, rescale = TRUE,
  rolling_forecast = FALSE)

build_dynamic_stability_plan(id_var = NULL, max_E = 16,
  E_list = seq(max_E), surrogate_method = "annual_spline",
  num_surr = 200, surr_params = list(), lib_sizes = seq(10, 100, by =
  10), random_libs = TRUE, num_samples = 100, replace = TRUE,
  RNGseed = 42, silent = TRUE, rescale = TRUE,
  rolling_forecast = FALSE)

Arguments

block	a data.frame containing time series for the community. Each column is a time series of abundances, and a censusdate column is used as the column containing the time value.
results_file	the location of the results to be stored on disk.
id_var	the name of the index variable of block that is useful for labeling the time-varying output, if it exists
max_E	largest E to examine using simplex projection; this sets the default range for E_list, but any setting for E_list will override the value for max_E
E_list	the embedding dimension or range of embedding dimensions to search over.
surrogate_method	which surrogate method to use: options are "annual_spline" or methods available in rEDM::make_surrogate_data()
num_surr	number of surrogates to compute
surr_params	a list of named optional arguments to be passed into the surrogate data function
lib_sizes	the vector of library sizes to try
random_libs	indicates whether to use randomly sampled libs
num_samples	is the number of random samples at each lib size (this parameter is ignored if random_libs is FALSE)
replace	indicates whether to sample vectors with replacement
RNGseed	will set a seed for the random number generator, enabling reproducible runs of ccm with randomly generated libraries
silent	prevents warning messages from being printed to the R console
rescale	A logical, indicating whether to rescale each time series
rolling_forecast	A logical, indicating whether to make individual rolling forecasts for the second half of the time series.

Value

a list with named components for the individual output objects XXX