The semivariogram provides a visual assessment of spatial dependence but without directly resulting in a single numerical value that expresses the measurement of such dependence. However, from the estimated parameters of the theoretical model adjusted to the experimental semivariogram, it is possible to construct a measure of this spatial dependence. Currently, there are two indices in literature that are widely used; however, these indices show some gaps. The aim of this study was to propose a new index to measure the spatial dependence of geostatistical data to overcome problems shown by existing indices. This new index uses the relationship between the semivariogram and the correlogram, thus including all aspects of spatial dependence. A comparison by simulation between the proposed index and existing indices was carried out, and the applicability of the proposed index was verified using actual published research, where adjustments of spherical, exponential, and Gaussian theoretical models were made. It was found that the proposed index was better than existing ones. Furthermore, it was observed that existing indices can lead to errors in the interpretation of the degree of spatial dependence, indicating that they should be avoided. As a result, the use of this new index to measure the degree of spatial dependence is recommended.
semivariogram; practical range; model factor; summary measures
