Mineralisation in alluvial diamond or vein gold deposits is often hosted in trapping mechanisms. For instance, diamonds are nested in bedrock depressions such as gullies and potholes which were carved out by the sea long before the diamond deposition took place. Similarly, gold mineralisation can precipitate in the necks of a boudinage structure._x000D_
In such deposits, the spatial distribution of the mineralisation is governed by the mutual arrangement of the trapsites and by the grade variations within each trapsite. Reliable estimates can be obtained only if they rest on a proper knowledge of both degrees of variability. This creates a challenging sampling problem, as during the exploration phase it is important to have a good global mean estimate to enable decision making towards the future. A similar problem occurs for local estimates at a later stage of development._x000D_
To understand and manage the sampling for global means, specific sampling procedures are required based on an appropriate choice of sample support size and spacing. Estimates from such sampling campaigns should be unbiased, and the less variability they display, the better the sampling parameters. The performance and impact of sampling parameters are tested and compared on different simulated deposits.