Fits a smooth surface to x, y, z data using generalised additive models

This is based on mgcv::gam() and derived from openair::polarPlot()

Usage

fit_gam_surface(
  data,
  x,
  y,
  z,
  weights = NULL,
  k = 100,
  extrapolate = FALSE,
  force_positive = TRUE,
  dist = 0.05
)

data: a data.frame or tibble containing the data (wide format)
x: string giving the u (wind) component or x coordinate, respectively
y: string giving the v (wind) component or y coordinate, respectively
z: string giving the response variable
weights: vector of weights for fitting x, y value pair; can be NULL
k: numeric, smoothing degree in gam model mgcv::gam(z ~ s(x, y, k = k)
extrapolate: TRUE/FALSE, result of fit extends over NA values in z, thus providing a way of extrapolation. If FALSE, only u, v pairs with !is.na(z) are returned, if TRUE, also fitted z values within a certain distance (dist) from x, y are returned based on mgcv::exclude.too.far()
force_positive: TRUE/FALSE, shall fitted values forced to be positive?
dist: input for mgcv::exclude.too.far(): how far away counts as too far. Grid and data are first scaled so that the grid lies exactly in the unit square, and dist is a distance within this unit square

tibble with variables u, v, z