MANAGING CORE SAMPLING BIAS IN RESOURCE MODELLING AT BATU HIJAU PORPHYRY COPPER GOLD DEPOSIT

PROCEEDINGS JOINT CONVENTION BALI 2007
The 32^nd HAGI, The 36^th IAGI, and The 29^th IATMI Annual Conference and Exhibition

MANAGING CORE SAMPLING BIAS IN RESOURCE MODELLING AT BATU HIJAU PORPHYRY COPPER GOLD DEPOSIT

Johan Arif, Nugraha Jakamartana, Wawan Hermawan & Koko Suhanto

Mine Geology Department,
PT Newmont Nusa Tenggara, NTB - Indonesia.

ABSTRACT

Batu Hijau is a copper-gold mining operated by PT Newmont Nusa Tenggara, a joint venture between Newmont Mining Corporation (45%), Sumitomo Corporation (35%) and PT Pukuafu Indah (20%). Project to date, the core generated exploration block model has underpredicted the blasthole generated ore controlled Cu metal by approximately 18% and Au metal by 15%. Monthly reconciliation shows a consistent underprediction suggesting that a bias exists in either the samples used or the model generation itself. Since start-up, various in-house sampling trials and statistical analysis were carried out to investigate the source of the bias. Although inconclusive, it became apparent that the core sampling was underestimating the grade while the blastholes were slightly overestimating the grade. It was also apparent that the degree of bias varied by geology, whereby the host lithologies showed a greater bias than the intrusives.

A consultant was engaged to review all data compiled to date and also review all sampling and resource modeling methodologies. The consultant’s findings supported the in-house findings that physical losses of sulphides during the drilling and sampling processes were the predominant cause for the bias. A mathematical correction function was derived to correct for the bias. A methodology to demonstrate the loss of sulphides during the drilling / sampling process was also proposed.

The consultant explored earlier in-house findings that the bias was spatially controlled solely by, or a combination of, lithology, alteration, depth and grade. The conclusion was that the majority, but not all of, the bias could be explained by grade. This followed earlier findings that there was less bias in the central high grades and a greater bias towards the lower grade peripheries. This relationship would also prove to be simpler to apply than relationships derived by individual geology types. It also followed the fines loss theory well – if the assumption is that a consistent amount of fines is lost across all grade ranges, then the percentage impact will be greater in the lower grade peripheries where a lower amount of sulphides exist. At the conclusion of the 2002 drilling, two models were built, one with the addition of the 2002 drilling and one with the addition of the 2002 drilling and the bias correction applied to the composites. A backward looking analysis was carried out for the period 2001 to October 2002. The model generated from grade bias corrected composites proved to be in line with actual production for the period. Proven to be robust, the end of 2002 Batu Hijau reserves were reported using the grade corrected model.