click on the '*' to jump to a section) Abstract Because of its dependence on computers, geostatistics has been criticised as being a black-box approach to grade estimation. Though geostatistical resource estimations are becoming more common, there is a widespread lack of understanding of the method's mechanics. The objectives of this investigation were to carry out a geostatistical resource estimate for the Poura deposit and to increase the industry's awareness and understanding of

more »

... l methods. The Poura Mine in Burkina Faso, West Africa consists of steeply-dipping, gold-silver mineralized quartz veins. The structure has a strike length of two kilometres and an average vein thickness of about two metres. This resource estimate concentrated on the remnant ore blocks and pillars in the underground workings. Using the supplied information, coordinates were assigned to 3,232 samples. Semi-variograms were calculated for grade and vein thickness. The raw data followed clear trends to which models could readily be fit. Ordinary block kriging was chosen as the appropriate kriging method. The author wrote a block kriging computer program called Krige2D to carry out the kriging calculations. For each block, the estimated grade and variance, as well the estimated vein thickness and variance, were calculated using Krige2D. The undiluted, mean block grade ranged between 4.7 g/tonne and 36.3 g/tonne with variance ranging between 9.4 (g/tonne) 2 and 143 (g/tonne) 2 . Estimated thickness ranged from 0.77 metres to 3.81 metres. Thickness variance ranged between 0.0014 m 2 and 0.37 m 2 . The mean grade was used to calculate the remnant block resources. When no cut-off grade is considered, undiluted resources consist of 468,000 tonnes with an average grade of 14.8 g/tonne. The resource contains a total of 6.95 tonnes (220,000 troy ounces) of gold and has an average vein thickness of 1.86 metres. Using a cut-off grade of 5.0 g/tonne, the resource consists of 464,000 tonnes with an average grade of 14.9 g/tonne. A previous resource estimate had been carried out using arithmetic methods. Block kriging estimated the average grade to by 2.5 g/tonne higher and the total amount of gold to be 1.2 tonnes higher. When confidence in the estimate and a 5.0 g/tonne cut-off grade were considered, the undiluted resource drops from 464,000 tonnes to 367,000 tonnes (a 21 % decrease) when the confidence in the estimate is increased from 50 % (mean value) to 70 %. The average grade decreases from 14.9 g/tonne to 13.5 g/tonne (a 9 % decrease). Additional work is warranted. An economic analysis should be carried out. Block kriging should be used with that economic analysis to optimise the sample spacing of future exploration programs. Further exploration is recommended along strike and down dip the current workings.