STK_MODEL generates a model with default covariance parameters

 CALL: MODEL = stk_model ()

   returns a structure MODEL (see below for a description of the fields in such
   a structure) corresponding to one-dimensional Gaussian process prior with a
   constant but unknown mean ("ordinary" kriging) and a stationary Matern
   covariance function.

 CALL: MODEL = stk_model (COVARIANCE_TYPE)

   uses the user-supplied COVARIANCE_TYPE instead of the default.

 CALL: MODEL = stk_model (COVARIANCE_TYPE, DIM)

   creates a DIM-dimensional model. Note that, for DIM > 1, anisotropic
   covariance functions are provided with default parameters that make them
   isotropic.

 In STK, a Gaussian process model is described by a 'model' structure,
 which has mandatory fields and optional fields.

   MANDATORY FIELDS: covariance_type, param, lm, lognoisevariance
   OPTIONAL FIELD: param_prior, noise_prior

 See also stk_materncov_iso, stk_materncov_aniso, ...