Sampling Density Variance

This methodology enables you to measure the quality of estimation regardless of the block size we use the Spatial density variance which is independent of the block size:

where is the kriging variance of the volume V.

The unit of the Spatial density variance being %².m³ it makes it hard to manipulate. For that reason, the Spatial density variance is normalized by the average of the domain squared. This quantity is homogenous with a volume, it is called the Specific Volume:

This quantity is not linked to the block size nor the average value. This can help comparing estimations of different domains or deposit by solely focusing on the variogram and the sampling layout.

In order to help defining the quality of the estimation, it possible to calculate the coefficient of variation of the estimation on a given production volume:

Thresholds can be applied on this quantity to classify blocks. For instance C. Dohm defined in 2004 (“A logical approach”) thresholds of 2.5% and 5%:

• Measured: CV < 2.5%
• Indicated: 2.5% < CV < 5%
• Inferred: 5% < CV

To obtain the Spatial density variance, we need the Kriging variance which is obtained by a Super Kriging: A super block V is defined and centered around each block v, data inside the super block is used to obtain the kriging variance of the super block . The Spatial density variance is then the same for the super block and the block: