Any (wave) model approximates reality. This motivates the researcher to improve the model, whereas the user should realize the limitations of the model.

 

How to trust a model? Consider four levels of validation:

  • conceptual level; which processes are included in the equations?
  • algorithmic level; which methods are used to solve the equations?
  • software level; is the program free of errors?
  • user level; is human error involved?