The "nugget effect" refers situations where coincidental sampling in a location with an abnormally high grade may skew interpretation and estimation of mineral reserves. The term gets its name from the irregularly high precious metal assays resulting from the analysis of samples that include "nuggets" of the metal.

The nugget effect is often discussed in geostatistics (or kriging). A semi-variogram is used to illustrate variability of sampling data with regard to spatial distribution (i.e., how similar are two samples depending on the distance between them.) When looking at a semi-variogram, the nugget value (i.e., y value) where x=0. In practical terms, it accounts for the fact that two samples could be taken side by side (i.e., there is practically no distance between them) but have significantly different values.

Figure: example semi-variogram showing the nugget value.