Predicting gold targets using cokriging in SURFER 17

Golden Software Inc. included the method of cokriging in the newest version of SURFER 17. This has opened a new tool for interpreting geochemical data. We can use cokriging in SURFER 17 to improve the quality of maps and to predict similar targets in nearby areas. We use cokriging when we want to process data from different datasets. One dataset is always smaller than the other. Here, I first tasted the method with a hypothetical geochemical model combining a smaller dataset of FA gold results with a larger dataset of ICP-MS multi-elements. Later, I applied this method to real data from a soil sampling project in Mozambique. I tested a known mineralized target and also an extended area to predict gold targets. I also had the gold results for the extended area. They allowed me to confirm the effectiveness of cokriging in predicting the new targets. There are many opportunities where we can apply cokriging as a prediction tool. One situation is when an initial sampling returned a group of interesting but isolated gold results. We can then use a cheaper method, like ICP-MS, to better understand the gold distribution in the area.

金软件公司（Golden Software Inc.）在其最新版SURFER 17中集成了协同克里金法（cokriging），该方法为地球化学数据解译提供了全新工具。借助SURFER 17中的协同克里金法，可提升成图质量并预测周边区域的同类找矿靶区。当需要处理多源数据集时，通常会采用协同克里金法，这类场景中往往存在一个规模更小的数据集。本次研究首先基于假想地球化学模型开展方法验证：模型将规模较小的FA金分析结果数据集与规模较大的电感耦合等离子体质谱法（ICP-MS）多元素数据集相结合。随后，将该方法应用于莫桑比克某土壤采样项目的实测数据。研究中先针对已知矿化靶区开展测试，再通过扩展区域预测金找矿靶区。同时获取了扩展区域的金分析结果，借此验证了协同克里金法在预测新型找矿靶区方面的有效性。协同克里金法作为预测工具的应用场景十分广泛，其中一类典型场景为：初始采样仅获得若干零散且具有研究价值的金分析结果，此时可借助成本更低的ICP-MS等手段，更精准地解析区域内金元素的分布特征。