Abstract
The distinction between artisanal versus small-scale mining operations has caused much confusion within the sector over the years. The objective of this research was to utilize multi-criteria decision-making and expert opinion tools to develop a definition of small-scale mining. The model developed would need to derive an “acceptable” and “optimized” definition of small-scale mining that uses pre-defined criteria as inputs, and the output would be the correct classification of the mining operation.
The foundation of the research undertaken was the utilization of decision, probability, and systems theory as the framework. The Analytic Hierarchy Process (AHP) is a tool developed to assist decision-makers in problems with multiple attributes. AHP was used to develop a multi-criteria decision model with multiple perspectives, i.e., political, economic, social, technological, legal, and environmental, based on selected literature reviews.
The Delphi technique, which dates to the 1950s regarding its use in research, was selected as the primary data collection tool. It is a “powerful” technique used in research methodology to bring about consensus or group decision-making amongst expert panels to achieve a specific objective. In this case, it was to determine the final criteria and their respective weights for the decision model. The data collected was both quantitative and qualitative.
The outcome of the Delphi process was that experts deemed Volume of Production, Level of Technology, Annual Turnover, Capital Cost, and Staff Educational level to be the most important criteria to be used in the development of a definition. The next step of the process was determining the rankings and weights for the above criteria for the five most mined commodities in Africa: gold, diamonds, semi-precious gemstones, ornamental & dimension stones, and construction & industrial minerals. The resulting multi-criteria decision model has average weights of 8.4% on the Social Perspective, 46.7% on the Technological Perspective, and 44.9% on the Economic Perspective.
A definition was formulated using a linear combination based on multiplying each variable weight by a rating and adding the results to achieve an output between 0 and 1. The movement of mining operations was modeled as a continuum that uses a combination of business and
10
mining classifications. The result included five business categories corresponding to five classes mapped to four mining categories: artisanal, small-scale, medium-scale (junior), and large-scale.
Due to a lack of plant data and COVID-19 travel restrictions, one could not obtain data for model testing. As a result, the “stratified sampling-based scenario generation” method of the Monte Carlo technique was used to generate a random baseline dataset for modeling purposes. It was found that there is a moderate to high correlation between the commodity type weights, which indicates that in a multiple linear regression model, there will not be a significant change in the output (classification of an operation) due to a variation in commodity type for the same input values.
Using the 1008-point baseline dataset developed, one tested the movement of operations across categories by changing commodity type. The number of operations that moved up a category or dropped down a category is +- 3.4%. The result corroborates the finding that the different commodity weightings do not affect the final rating or classification substantially but can result in operations on the border of a category interval moving to another classification.
A neural network was developed based on the linear model. The error using prediction in the neural network is significantly higher using four criteria instead of five, but one can still predict the classification (rating) of the mining operation with some level of accuracy.
This definition will provide guidance and direction in terms of the available opportunities and support structures needed for the sector. It will clarify country-specific laws, policies, and regulations as there tends to be confusion and ambiguity in practice. According to experts, it will assist in bringing about both transparency and accountability within the sector.
All research questions posed during the study were answered, and all the objectives were met. The hypothesis that small-scale mining could be differentiated from artisanal, medium, and large-scale mining based on the measurement and linear combination of pre-defined criteria and the relative ranking and weighting thereof was proved.