R/lsm_c_te.R

In landscapeecology/landscapemetrics: Landscape Metrics for Categorical Map Patterns

#' TE (class level)
#'
#' @description Total (class) edge (Area and Edge metric)
#'
#' @param landscape A categorical raster object: SpatRaster; Raster* Layer, Stack, Brick; stars or a list of SpatRasters.
#' @param count_boundary Include landscape boundary in edge length
#' @param directions The number of directions in which patches should be
#' connected: 4 (rook's case) or 8 (queen's case).
#'
#' @details
#' \deqn{TE = \sum \limits_{k = 1}^{m} e_{ik}}
#' where \eqn{e_{ik}} is the edge lengths in meters.

#' TE is an 'Area and edge metric'. Total (class) edge includes all edges between class i and
#' all other classes k. It measures the configuration of the landscape because a highly
#' fragmented landscape will have many edges. However, total edge is an absolute measure,
#' making comparisons among landscapes with different total areas difficult. If
#' \code{count_boundary = TRUE} also edges to the landscape boundary are included.
#'
#' \subsection{Units}{Meters}
#' \subsection{Range}{TE >= 0}
#' \subsection{Behaviour}{Equals TE = 0 if all cells are edge cells. Increases, without limit,
#' as landscape becomes more fragmented}
#'
#' @seealso
#' \code{\link{lsm_p_perim}}
#' \code{\link{lsm_l_te}}
#'
#' @return tibble
#'
#' @examples
#' landscape <- terra::rast(landscapemetrics::landscape)
#' lsm_c_te(landscape)
#'
#' @references
#' McGarigal K., SA Cushman, and E Ene. 2023. FRAGSTATS v4: Spatial Pattern Analysis
#' Program for Categorical Maps. Computer software program produced by the authors;
#' available at the following web site: https://www.fragstats.org
#'
#' @export
lsm_c_te <- function(landscape,
                          count_boundary = FALSE, directions = 8) {
    landscape <- landscape_as_list(landscape)

    result <- lapply(X = landscape,
                     FUN = lsm_c_te_calc,
                     count_boundary = count_boundary,
                     directions = directions)

    layer <- rep(seq_along(result),
                 vapply(result, nrow, FUN.VALUE = integer(1)))

    result <- do.call(rbind, result)

    tibble::add_column(result, layer, .before = TRUE)
}

lsm_c_te_calc <- function(landscape, count_boundary, directions, resolution, extras = NULL) {

    if (missing(resolution)) resolution <- terra::res(landscape)

    if (is.null(extras)){
        metrics <- "lsm_c_te"
        landscape <- terra::as.matrix(landscape, wide = TRUE)
        extras <- prepare_extras(metrics, landscape_mat = landscape,
                                            directions = directions, resolution = resolution)
    }

    # all values NA
    if (all(is.na(landscape))) {
        return(tibble::new_tibble(list(level = "class",
                              class = as.integer(NA),
                              id = as.integer(NA),
                              metric = "te",
                              value = as.double(NA))))
    }

    # get class id
    classes <- extras$classes
    class_patches <- extras$class_patches
    resolution_x <- resolution[[1]]
    resolution_y <- resolution[[2]]

    if (length(classes) == 1 && !count_boundary) {

        tibble::new_tibble(list(
            level = "class",
            class = as.integer(classes),
            id = as.integer(NA),
            metric = "te",
            value = as.double(0)))

    } else {

        # resolution not identical in x and y direction
        if (resolution_x != resolution_y) {

            top_bottom_matrix <- matrix(c(NA, NA, NA,
                                          1,  0, 1,
                                          NA, NA, NA), 3, 3, byrow = TRUE)

            left_right_matrix <- matrix(c(NA, 1, NA,
                                          NA, 0, NA,
                                          NA, 1, NA), 3, 3, byrow = TRUE)
        }

        te_class <- lapply(X = classes, function(patches_class) {

            # get connected patches
            landscape_labeled <- class_patches[[as.character(patches_class)]]

            # set all non-class patches, but not NAs, to -999
            edge_cells <- which(!is.na(landscape) & landscape != patches_class)

            landscape_labeled[edge_cells] <- -999

            # add one row/column to count landscape boundary
            if (count_boundary) {
                landscape_labeled <- pad_raster_internal(landscape = landscape_labeled,
                                                         pad_raster_value = -999,
                                                         pad_raster_cells = 1,
                                                         global = FALSE)

                # set NA to -999
                landscape_labeled[is.na(landscape_labeled)] <- -999
            }

            # resolution identical in x and y direction
            if (resolution_x == resolution_y) {

                # get adjacencies
                neighbor_matrix <- rcpp_get_coocurrence_matrix_single(landscape_labeled,
                                                                      directions = as.matrix(4),
                                                                      single_class = -999)

                # sum of all adjacencies between patch id and non-class patches (-999) converted to edge length
                edge_ik <- (sum(neighbor_matrix[2:nrow(neighbor_matrix), 1])) * resolution_x
            }

            else {

                # get adjacencies
                left_right_neighbours <- rcpp_get_coocurrence_matrix_single(landscape_labeled,
                                                                            directions = as.matrix(left_right_matrix),
                                                                            single_class = -999)

                # sum of all adjacencies between patch id and non-class patches (-999) converted to edge length
                edge_ik_left_right <- sum(left_right_neighbours[2:nrow(left_right_neighbours), 1]) * resolution_x

                # get adjacencies
                top_bottom_neighbours <- rcpp_get_coocurrence_matrix_single(landscape_labeled,
                                                                            directions = as.matrix(top_bottom_matrix),
                                                                            single_class = -999)

                # sum of all adjacencies between patch id and non-class patches (-999) converted to edge length
                edge_ik_top_bottom <- sum(top_bottom_neighbours[2:nrow(top_bottom_neighbours), 1]) * resolution_y

                # add edge length in x- and y-direction
                edge_ik <- edge_ik_left_right + edge_ik_top_bottom
            }

            tibble::new_tibble(list(
                level = rep("class", length(edge_ik)),
                class = rep(as.integer(patches_class), length(edge_ik)),
                id = rep(as.integer(NA), length(edge_ik)),
                metric = rep("te", length(edge_ik)),
                value = as.double(edge_ik)))
        })

        do.call("rbind", te_class)
    }
}