Partial dependent plots for abundance-based distribution models

Create partial dependence plots to explore the marginal effect of predictors on modeled abundance

Usage

p_abund_pdp(
  model,
  predictors = NULL,
  resolution = 100,
  resid = FALSE,
  training_data = NULL,
  invert_transform = NULL,
  response_name = NULL,
  projection_data = NULL,
  rug = FALSE,
  colorl = c("#462777", "#6DCC57"),
  colorp = "black",
  alpha = 0.2,
  theme = ggplot2::theme_classic(),
  sample_size = NULL,
  training_raster = NULL,
  x_coord = NULL,
  y_coord = NULL
)

Arguments

model: A model object found in the first element of the list returned by any function from the fit_abund_ or tune_abund_ function families
predictors: character. Vector of predictor name(s) to calculate partial dependence plots. If NULL all predictors will be used. Default NULL
resolution: numeric. Number of equally spaced points at which to predict abundance values for continuous predictors. Default 50
resid: logical. Calculate residuals based on training data. Default FALSE
training_data: data.frame or tibble. Database with response and predictor values used to fit a model. Default NULL
invert_transform: logical. If TRUE, inverse transformation of response variable will be applied.
response_name: character. Name of the response variable. Default NULL
projection_data: SpatRaster. Raster layer with environmental variables used for model projection. When this argument is used, function will calculate partial dependence curves distinguishing conditions used in training and projection conditions (i.e., projection data present in projection area but not training). Default NULL
rug: logical. Add rug plot to partial dependence plot. Default FALSE
colorl: character. Vector with colors to plot partial dependence curves. Default c("#462777", "#6DCC57")
colorp: character. Color to plot residuals. Default "black"
alpha: numeric. Transparency of residuals. Default 0.2
theme: ggplot2 theme. Default ggplot2::theme_classic()
sample_size: vector. For CNN only. A vector containing the dimensions, in pixels, of raster samples. See cnn_make_samples beforehand. Default c(11,11)
training_raster: a terra SpatRaster object. For CNN only. A raster containing the predictor variables used in tune_abund_cnn or fit_abund_cnn.
x_coord: character. For CNN only. The name of the column containing longitude information for each observation.
y_coord: character. For CNN only. The name of the column containing latitude information for each observation.

Value

A ggplot object

Details

This function creates partial dependent plots to explore the marginal effect of predictors on modeled abundance. If projection_data is used, function will extract the minimum and maximum values found in a region or time period to which a model will be projected. If the range of projection data is greater than of the training data it will be plotted with a different color. Partial dependence plot could be used to interpret a model or to explore how a model may extrapolate outside the environmental conditions used to train the model.

Examples

if (FALSE) {
require(dplyr)
require(terra)

# Load data
envar <- system.file("external/envar.tif", package = "adm") %>%
  rast()

data("sppabund")
some_sp <- sppabund %>%
  filter(species == "Species one")

# Fit some models
mglm <- fit_abund_glm(
  data = some_sp,
  response = "ind_ha",
  predictors = c("bio12", "elevation", "sand"),
  predictors_f = c("eco"),
  partition = ".part",
  distribution = "ZAIG",
  poly = 3,
  inter_order = 0,
  predict_part = TRUE
)

# Partial Dependence Plots:

# In different resolutions
p_abund_pdp(
  model = mglm,
  resolution = 50,
  training_data = some_sp,
  response_name = "Abundance"
)

p_abund_pdp(
  model = mglm,
  resolution = 5,
  training_data = some_sp,
  response_name = "Abundance"
)

# Especific variables and different resulotions
p_abund_pdp(
  model = mglm,
  predictors = c("bio12", "sand"),
  training_data = some_sp,
  response_name = "Abundance"
)

# With residuals and rug plot
p_abund_pdp(
  model = mglm,
  training_data = some_sp,
  response_name = "Abundance",
  resid = TRUE
)

p_abund_pdp(
  model = mglm,
  training_data = some_sp,
  response_name = "Abundance",
  rug = TRUE
)

p_abund_pdp(
  model = mglm,
  training_data = some_sp,
  response_name = "Abundance",
  resid = TRUE,
  rug = TRUE
)

# Partial depence plot for training and projection condition found in a projection area
p_abund_pdp(
  model = mglm,
  training_data = some_sp,
  projection_data = envar,
  response_name = "Abundance",
  rug = TRUE
)

# Custumize colors and theme
p_abund_pdp(
  model = mglm,
  predictors = NULL,
  resolution = 100,
  resid = TRUE,
  training_data = some_sp,
  projection_data = envar,
  colorl = c("blue", "red"),
  colorp = "darkgray",
  alpha = 0.4,
  theme = ggplot2::theme_dark()
)
}