Skip to contents

Compute MESS (Multivariate Environmental Similarity Surface)

Source: R/model_diagnostics.R
maxent_mess.Rd

Measures how similar each cell is to the training environment using the full distribution of reference values. Negative MESS values indicate novel (non-analog) environments.

Usage

maxent_mess(env_grids, reference_values, feature_names)

Arguments

env_grids

List of external pointers to Grid<float> objects.

reference_values

List of numeric vectors with reference values for each variable (e.g. values at training sites).

feature_names

Character vector of variable names.

Value

A named list with:

mess_grid

External pointer to Grid<float> with MESS values

mod_grid

External pointer to Grid<float> with Most Dissimilar Variable index (1-based)

Examples

# \donttest{
g1 <- maxent_grid_from_matrix(matrix(runif(50), 5, 10),
        -120, 35, 1, name = "temp")
g2 <- maxent_grid_from_matrix(matrix(runif(50, 50, 200), 5, 10),
        -120, 35, 1, name = "precip")
temp_train_vals <- runif(20)
precip_train_vals <- runif(20, 50, 200)
result <- maxent_mess(list(g1, g2),
            list(temp_train_vals, precip_train_vals),
            c("temp", "precip"))
mess_mat <- maxent_grid_to_matrix(result$mess_grid)
mod_mat  <- maxent_grid_to_matrix(result$mod_grid)
# }