A Geostatistical Study of the Uranium Deposit at Kvanefjeld, The Ilimaussaq Intrusion, South Greenland

The uranium deposit at Kvanefjeld within the Ilimaussaq intrusion in South Greenland has been tested by 70 diamond drill holes. In total 5658 drill core samples were selected and analysed by various methods. A data base containing all the analytical data, geological information and characteristic drill hole parameters was constructed.

Based on different types of spatially distributed samples the uranium variation within the deposit was studied. The spatial variation, which comprises a large random component, was modelled, and the intrinsic function was used to establish gradetonnage curves by the best linear unbiased estimator of geostatistics (kriging). f roil Qu ta obtained by a ground surface gamma-spectrometrie survey it is shown that the uranium variation is possibly subject to a spatial anisotropy consistent with the geology.

It is also shown that, although anisotropy exists, the uranium variation has a second-order stationarity. A global estimation of the total reserves shows chat single block grade values are always estimated with high errors. This is mainly caused by the poor spatial structure and the very sparse sampling pattern.

The best way to solve this problem appears to be a selective type of kriging. The overall uranium reserves are estimated as approx. 23600 tons with a mean grade of 297 ppm using a cutoff grade of 250 ppm U. The effect of using different block sizes/bench heights is studied. Studies of data from the Kvanefjeld tunnel show that local geostatistical estimation can be done with acceptably small errors provided that a close sampling pattern is used.

Geostatistics is therefore regarded as a useful tool for the estimation of this deposit. A regression relationship is established to correct field gamma-spectrometrie measures of bulk grades towards truer values. Multivariate cluster and discriminant analyses were used to classify lujavrite samples based on their trace element content.

A number of mis-labelled samples were discovered. Misclassification is due to a possibly continuous transition between naujakasite lyjavrite and arfvedsonite lujavrite which was not recognized by the geologists. Some of the main mineralogical differences between the geological units are identified by the discriminating effect of the individual variable.