lidaRtRee (version 4.0.8)

aba_predict: Mapping of ABA prediction models

Description

Applies calibrated area-based prediction models output of aba_build_model to a raster of metrics to obtain a raster of predictions

Usage

aba_predict(
  model_aba,
  metrics_map,
  stratum = NULL,
  add_error = FALSE,
  pkg = "terra",
  ...
)

Value

a raster of predictions obtained by applying the model aba_build_model

to the observations in metrics_map

Arguments

model_aba

model returned by aba_build_model or aba_combine_strata

metrics_map

raster. metrics returned e.g by pixel_metrics

stratum

string. indicates which layer of metrics.map contains the stratum in case of a stratified aba.model. The layer should have a RAT including a column with the same name (see is.factor).

add_error

boolean. indicates whether errors sampled from a normal distribution N(0, sigma(residuals)) should be added to fitted values; implemented only for log transformation case

pkg

raster output format. Use pkg = "terra|raster|stars" to get an output in SpatRaster, RasterLayer or stars format.

...

other parameters to be passed to predict.lm, e.g. interval = "prediction",

See Also

aba_build_model for model fitting and aba_combine_strata for combining stratified models, clean_raster for applying spatial mask and value thresholds to a raster.

Examples

# load data
data(quatre_montagnes)
# build model
model_aba <- aba_build_model(quatre_montagnes$G_m2_ha, quatre_montagnes[, 9:76],
  transform = "boxcox"
)
# build example raster to apply model
quatre_montagnes$X <- rep(1:8, 12)
quatre_montagnes$Y <- rep(1:12, each = 8)
metrics_map <- terra::rast(quatre_montagnes[, c(2, 3, 9:76)], type = "xyz")
predict_map <- aba_predict(model_aba, metrics_map)

# plot map
terra::plot(predict_map, main = "predictions")