Myths and legends in resource modelling have merit in some circumstances but are not always the most appropriate approach to problem solving. In this article, we consider legends, myths or ‘rules of thumb’ associated with top capping, block size and grade interpolation.
Myths and Legends: Case examples
A) High-grade capping above the 97.5th percentile. While this potentially has limited significance in well-sampled deposits with homogenous grade distributions, high-grade ‘outliers’ may also occur at values less than the 97.5th percentile in other deposits where grade variability is significant, for example nuggety gold, and particularly those at a relatively early stage of drilling. These values, if left uncapped and otherwise not restricted or sub-domained, may result in highgrade ‘blow outs’. This could lead to upward bias in the resulting grade estimate and risk of overstating the metal content. This is illustrated in Figure 1, an example from SRK’s review of a third-party estimate for an underground gold prospect.
B) Block size at half the average drill spacing. In general, this is a good rule of thumb to help avoid the estimation of overly small blocks, which can result in grade distributions that are poorly supported by the input sample data. However, block size selection should also consider the interpreted mineralisation style. For example, where there is a predictable grade trend observed in drilling from top to bottom contact within a mineralised orebody, use of a smaller block size (less than half drill spacing) may be justified to appropriately reflect the interpreted distributions of high and low grades. This was the case at a sediment-hosted borate project where SRK recently estimated the mineral resource (Figure 2).
C) Search neighbourhood ellipse dimension set to two-thirds of the variogram range. In general, this is a reasonable rule of thumb where variograms are well informed by sample data and their range exceeds multiple drill fences. However, in early-stage exploration projects where the variogram range is based on limited data and is similar to the drill spacing, a search dimension at two-thirds of this distance could result in poorly informed block estimates based on single drill holes. A larger ellipse dimension (less than two thirds of the variogram range) is likely to be more appropriate in this scenario.
Conclusions
The resource modelling myths and legends considered in this article often form a reasonable starting point or rule of thumb for analysis; however, as highlighted in the cases above, every dataset and geological model should be assessed individually, with project specific parameters to avoid sub-optimal mineral resource estimates.
