Scale and Support of Observation

In practice, all these variabilities are never observed simultaneously, since this would require an enormous amount of data of quasi-point support covering the entire field of variability from $1 /mu m$ to $100 km$.

At the scale of observation, when counting elements on a thin section (where the information has a quasi-point support), it is not possible to distinguish between the first variability due to errors of measure (e.g., counting errors) and the second due to petrographical variability.

At the scale of evaluation of a mining face by channel samples, the support of the channel samples integrates the first two variabilities into one single undifferentiated variability that will be defined later as a nugget effect. However, it is possible at this scale to distinguish the third variability, due to lenticular beds in the same stratum.

At the scale of the evaluation of the entire deposit by drill cores spaced at 50 to 100$m$ intervals, the first three variabilities are indistinguishable. However, it is possible to distinguish the fourth variability due, for example, to the alternation of mineralized strata, or trends in grade values at the borders of the deposit or at lower depths.

Just as in a set of nested tables, each table covers all the smaller tables, so each scale of observation integrates the variabilities of all the smaller scales.