Application of Geostatistics to Mineral Resource Modelling and Estimation: Case Study of Gofolo Hill Iron Ore Deposit, Western Liberia

Abstract

The Gofolo Hill deposit, situated over a regional northwest trending shear zone (Todi Shear Zone) on the margin of the Archean West African craton, occurs in one of Liberia's historic iron ore mining districts notable for several high-grade deposits. The study area is in Western Liberia, Grand Cape Mount County and bounded by latitude 6°52ʹ46ʺ N to 6°53´11ʺ N and longitude 11°14ʹ19ʺ W to 11°13ʹ25ʺ W. Though the Gofolo Hill occurs with Koehnko Hill and Zaway Hill along the strike length of two historic but abandoned iron ore mines: Bomi Hills Mine and Bong Mine, the deposit is yet to undergo a full evaluation of mineral resources which will inform feasibility study and subsequently draw on conclusion about exploitation. The study aims to model the orebody of Gofolo Hill and evaluate the mineral resources by applying the geostatistical ordinary kriging method. Ordinary kriging and comparative basic estimation methods (inverse distance weighting and nearest neighbour polygon) are used to evaluate the Gofolo Hill iron ore deposit and test the precision and accuracy of the results. Results for all estimation methods were obtained by applying a 30% Fe cut-off grade and a global density of 3.00kg/m3. The ordinary kriging method estimates the project to contain 17.169Mt of ores at an average grade of 35.90% Fe. Comparable methods like IDW2, IDW3 and NNP reported 35.22%, 35.53% and 38.53% grade Fe respectively and 16.274Mt, 16.975Mt, and 14.757Mt, respectively. The comparative study through grade-volume-tonnage results, visualisation of grade block maps, statistics, and correlation coefficient shows that the IDW methods (power2 then power3) produced more comparable results to the OK method than the NNP method. Cross–validation by the slope of regression values calculated from original sample values and kriged estimates showed a strong correlation meaning good estimates, and provides a slope or mean of 0.792. Further validation was carried out on all methods by (1) comparing the 3D model volume result (estimated volume) to the block model result (calculated volume); (2) by comparing global mean grade difference; and (3) by comparing calculated standard error. These validation methods prove the precision and accuracy of estimates. The study recommends that infill drilling be done on the project to improve the estimates confidence from inferred resources to indicated or measured resources. It also suggests that other deposits within strike length and proximity, such as Gofolo north-east, Koehnko, and Zaway, which together make up the “Mofe Creek project”, be fully estimated, and a full feasibility study be conducted on the project for exploitation of the mineral resource.