Determination of sampling configuration for near-shore diamondiferous gravel occurrence using geostatistical methods
J Jacob, C Prins, A Oelofsen

Diamondiferous linear beaches in Mining Area No. 1 have been the mainstay of Namdeb’s diamond production for more than 80 years. Most of the onshore beaches have been mined out, but in recent years mining has been extended offshore into the surf zone through a process of beach accretion. A total of 61.6 million carats has been produced from the gravel beaches since their discovery in 1928, and Mining Area No. 1 is considered to have great remaining potential from areas currently submerged under water.
To date, the surf zone has remained largely unexplored due to the consistently strong winds, currents, and large waves which make access to this area extremely difficult. This paper describes the processes used in developing a practical methodology for exploration of the surf zone in a domain extending approximately 22 km long in a northwest–southeast direction and 0.3 km wide in a perpendicular direction adjacent to the current Mining Area No 1.
The vigorous surf zone poses multiple technical challenges in terms of obtaining geological and diamond information. In order to access the area for sampling, a jack-up walking probe drill platform (PDP) with a 5-inch diameter reverse-circulation drill has been developed to carry out exploration drilling in the dynamic surf zone. The hydraulically driven platform has eight legs, each of which is 18 m long. Four of the legs are in fixed positions on the fore and aft sides of the platform. The other four moveable legs are fitted to sliding frames attached to the port and starboard sides of the platform. The sets of fixed and moveable legs can be raised and lowered by hydraulically powered jacking stations. By alternately lifting and lowering the sets of fixed and moveable legs in conjunction with the frames moving back and forth horizontally, a walking action is performed by the PDP. The platform can walk at a speed of 10–15 m/h, depending on weather and footwall conditions. Optimization of sampling for diamondiferous gravel occurrence in Mining Area No.1 was undertaken through creation of a virtual orebody on which different sample configurations were tested. The input data for the construction of a virtual orebody comprises a set of drilling data, collected from recently accreted areas directly adjacent to the 22 km × 0.3 km target domain. The input drilling data covers only 34% of the domain, and for the purpose of this study, a single realization is deemed to be sufficient.
The texture obtained from using only the drilling data to construct a simulation did not make geological sense; hence it was necessary to make use of analog data in order to improve the simulation. The first analog data used consists of the gully patterns found in the metasedimentary Precambrian bedrock footwall. Gullies are, in general, gravel filled and it is fair to assume that gully patterns form a subset of the total gravel occurrence. Total gravel occurrence includes marine terraces (governed by sealevel stands) above bedrock peaks, together with gravel within gullies below bedrock peaks. The second analog ‘data’ introduced is through the use of ‘expert interpretation’. The indicator drilling data is interpreted by expert opinion and the 2-dimensional result is hand-sketched, digitized, and then pixelated. The pixelated data set is then used as input for variogram calculation.
This study cannot provide a single definite optimization result as the nature of the data does not permit this. The use of different validation approaches (conditioning data, where available; expert interpretation; and gully pattern data), however, can give a very good indication of how to balance sampling effort with de-risking aspects related to geological continuity. Work undertaken by sampling the simulation will give an assessment of the relative probabilities of determining gravel occurrence in the study area. This study showed that a 50 m × 400 m cross-configuration will be a good initial sample spacing for highlighting areas where gravel may be absent, and further infill drilling may be required. The relative efficiency of a 50 m × 400 m cross off-set sampling configuration has been demonstrated using a trumpet curve versus sampling effort when using kriging as the estimation method.
Keywords: sampling configuration, marine diamonds, orebody simulation.