Conditional Simulation For Recoverable Reserve Estimation

Geostatistical literature has been dealing with recoverable reserve estimation techniques for the last 30 years. There have been well established support correction techniques (Le. affine correction, indirect log-normal correction, etc.) to deal with volume -variance relationship globally in determining what will be the reserves above a cutoff grade during mining. Kriging has become the accepted technique in the mining industry for ore reserve estimation. Many large companies are using this method for gold and copper deposit modeling. In general, the estimates coming from kriging are smooth. The variances of the block grades estimated by kriging are often lower than the material that will be mined selectively. As a result, kriging tends to under estimate reserves at high cutoff grades and overestimates them in low cutoff grades. It is necessary to perform additional steps after kriging if the goal is to determine recoverable reserves at some cutoff and not the in situ resources. In the mining industry, it is becoming more and more common for the low grade heap leach projects to consider whether the higher grade portion of a deposit may be processed by a mill at higher profits. Similarly, a concentrator -refinery operations are implementing heap leach -SX-EW option for the lower grade portion of their deposit. These options are making it very critical to estimate the low and high end of the grade distributions more accurately. Ore reserve estimation is normally done by using widely spaced exploration data which has very small size support (Le. core samples). Selective mining is carried out on ore control that is based on closely spaced blasthole data normally averaged into larger polygons based on some cutoff grade. Estimated grades from exploration data need to reflect the variability of the grades of the selectively mined units if they are used without further adjustments to predict recoverable reserves.