- Title
- Optimizing geochemical sampling sizes and quantifying uncertainties for environmental risk assessment using Anglogold-Ashanti Gold Mines as a case study
- Creator
- Chihobvu, Elizabeth
- Subject
- Environmental risk assessment
- Subject
- Geochemical prospecting
- Date Issued
- 2010-04
- Date
- 2010-04
- Type
- Master's theses
- Type
- text
- Identifier
- http://hdl.handle.net/10353/24443
- Identifier
- vital:62796
- Description
- Generally, and particularly in South Africa, limited work done on the development of methodologies for sample sizing and quantifying uncertainties in geochemical sampling and analyses. As a result, little trust is placed on the long-term predictions of geochemical modelling for Environmental Risk Assessment (E.R.A). In addition, this leads to the slow approval of mining authorisations, water use licenses and mine closure plans. This dissertation addresses this deficiency in geochemical sampling and analyses specifically for ERA and proposes two methodologies (i) for quantifying uncertainties in geochemical sampling and analysis as a function of sample size and analyses and (ii) for determining the optimum sample size to ensure data quality. The statistical analysis approach was adopted as the best method for sample size determination. The approach is based on the premise that the size of the study sample is critical to producing meaningful results. The size of the required samples depends on a number of factors including purpose of the study, available budget, variability of the population being sampled, acceptable errors and confidence level. The methodology for estimating uncertainty is a fusion of existing methodologies for quantifying measurement uncertainty. The methodology takes a holistic view of the measurement process to include all processes involved in obtaining measurement results as possible uncertainty components. Like the statistical analysis approach, the methodology employs basic statistical principles in estimating the size of uncertainty, associated with a given measurement result. The approach identifies each component of uncertainty; estimates the size of each component and sums the contribution of each component in order to approximate the overall uncertainty value, associated with a given measurement result. The two methods were applied to Acid-Base Accounting (ABA) data derived from geochemical assessment for ERA of the West Wits and Vaal River (Ashanti Gold mines) tailings dams undertaken by Pulles and Howard de Lange Inc. on behalf of AngloGold Ltd. The study was aimed at assessing and evaluating the potential of tailings dams in the two mining areas to impact on water quality and implications of this in terms of mine closure and rehabilitation. Findings from this study show that the number of samples needed is influenced by the purpose of the study, size of the target area, nature and type of material, budget, acceptable error and the confidence level required, among other factors. Acceptable error has an exponential relationship with sample size hence one can minimize error by increasing sample size. While a low value of acceptable error value and high confidence are always desirable, a tradeoff among these competing factors must be found, given the usually limited funds and time. The findings also demonstrated that uncertainties in geochemical sampling and analysis are unavoidable. They arise from the fact that only a small portion of the population rather than a census is used to derive conclusions about certain characteristics of the target population. This is further augmented by other influential quantities that affect the accuracy of the estimates. Effects such as poor sampling design, inadequate sample size, sample heterogeneity and other factors highly affect data quality and representivity hence measurement uncertainty. Among these factors, those associated with sampling, mainly heterogeneity was found to be the strongest contributing factor toward overall uncertainty. This implies an increased proportion of expenditure should be channelled toward sampling to minimise uncertainty. Uncertainties can be reduced by adopting good sampling practices and increasing sample size, among other methods. It is recommended that more information be made available for proper uncertainty analysis.
- Description
- Thesis (MSc) -- Faculty of Science and Agriculture, 2010
- Format
- computer
- Format
- online resource
- Format
- application/pdf
- Format
- 1 online resource (180 leaves)
- Format
- Publisher
- University of Fort Hare
- Publisher
- Faculty of Science and Agriculture
- Language
- English
- Rights
- University of Fort Hare
- Rights
- All Rights Reserved
- Rights
- Open Access
- Hits: 220
- Visitors: 236
- Downloads: 21
Thumbnail | File | Description | Size | Format | |||
---|---|---|---|---|---|---|---|
View Details Download | SOURCE1 | Chihobvu (M Sc) Geology.pdf | 2 MB | Adobe Acrobat PDF | View Details Download |