Sampling errors and their influence on accuracy of zinc and lead content evaluation in ore from the Trzebionka mine (Silesian–Cracow Zn–Pb ore district, Poland)

Abstract

The Trzebionka ore field is located in the SE part of the Silesian–Cracow ore district. The mine extracts zinc and lead ore from a Mississippi Valley-type deposit, hosted in Triassic dolomites. These are stratabound tabular orebodies with very complex internal structure. Main ore minerals are sphalerite (ZnS) and galena (PbS). Ore grade may be defined as low: 4.2% Zn and 1.7% Pb (on average). The accuracy of the mean metal estimation within typical reserve calculation blocks of 100×100 m horizontal dimensions is low with the relative kriging standard errors: 20–25%. The study aimed at examination to what extent the sampling and analytical errors affect the accuracy of the mean metal estimation in mine sub-blocks (area about 1 ha) and in broken ore. The results of routine and experimental sampling of orebodies in mine workings, and automatic sampling of the broken ore transported on a conveyor belt, were used in this research. Variability of metal contents in the deposit “on site” and in the broken ore was characterized geostatistically by means of relative semivariograms. Variances of sampling errors were estimated using P.Gy formula while variances of analytical errors using the experimental data. It was found that total variability of metal contents in the deposit “on site” is only marginally affected by sampling and analytical errors. In contrast, the contribution of these errors to the total variability of metal contents in the broken ore is significantly high. However, the latter has no practical impact on the accuracy of metal content estimation due to low variation of metal tenors caused by ore homogenization and large volume of samples taken from conveyor belt.