Heuristic rules for a reliable variogram parameter tuning

The variogram incorporates the basic information for the geostatistical analysis of spatial data, making it one of the most popular and widespread approaches for the evaluation and assessment of environmental data. Fitting an experimental variogram with a proper model, capable of catching the main characteristics of the sampled data, is a critical stage of spatial analysis. This stage is generally carried out using a semi-Automatic approach, with different levels of user's contribution. Once a model and the related starting parameter values have been defined, they can be refined using a trial-And-error strategy and the impact of such changes can be evaluated through several calibration indices. In this paper, we present an overview of the main indices used for the refinement of the variogram model parameters, highlighting the specific information provided by each of them together with a few heuristic strategies for driving the trial-And-error refinement process.