2. MONITORING PROGRAMME

Factors taken into account in the selection of the sites included:

• the potential for large-scale drinking water use,

• the identification of significant point-source discharges from mines,

• the need to establish, as far as possible, natural background levels.

2.4 Sampling Frequency and Duration

Since, for chronic radiation exposures, it is the cumulative radiation dose that is important, doses to the public are normally integrated over a full year of exposure for the purposes of assessment. The exact yearly dose from environmental radioactivity, which varies over time, particularly in water sources, can only be determined with high frequency monitoring, ideally on a continuous basis. This was, however, not possible in practice due both to analytical capacity constraints and to budgetary constraints. A compromise had to be reached to ensure reasonable accuracy of the estimation of the integrated annual dose. Thus to achieve a reasonable estimate of integrated annual radiation dose, a weekly sampling frequency and a 25 week sampling duration was adopted for the first phase of the study (7 January to 25 June 1997). Preliminary analysis of the data from the first phase of the study showed that significant autocorrelation existed for the radioactivity data gathered at intervals of less than one month (see Appendix 3). This implied that the sampling frequency could be reduced to once a month without a significant loss in the ability to estimate the annual dose with a reasonable degree of accuracy. Thus, during the second phase of the study (July to December 1997), data was gathered on a monthly rather than on a weekly basis.

2.5 Geological and Radiological Characteristics of the Catchment

Because gold mining was established in the Mooi River catchment long before radioactivity measurements were made, it was not possible to establish unequivocally the true natural background level, especially as the natural ground water recharge constitutes a significant proportion of the base flow of the river. Recent gamma ray spectrometric surveys and a large body of radioactivity measurements on geologically similar areas for airborne radiometric mapping of the environmental impact of gold and uranium mining in Gauteng Province, South Africa, were also reported by Coetzee, H, (1995) [10]. The pertinent geological factors are as follows:

• The dolomitic areas (most of the Mooi River catchment is underlain by dolomite) have very low (~10% of crustal average) radio-element contents. These dolomites also constitute the major groundwater source in the area.

• The quartzites and shales in the area tend to be enriched in potassium, uranium and thorium and consequently, the daughter nuclides of uranium and thorium reach levels generally at 1.5-3 times the crustal average.

• The granites tend to contain slightly elevated uranium concentrations and elevated potassium and thorium concentrations.

The highest naturally occurring uranium series activities in the area are found in the gold reefs of the Witwatersrand Supergroup. These, however, are extremely limited in outcrop, generally sub-outcropping below hundreds or thousands of metres of younger cover rocks.

2.6 Variables Measured and Data Collected

2.6.1 Radionuclides

The three natural radioactive decay series of relevance are those headed by the radionuclides uranium-238, uranium-235 and thorium-232. Details of these decay series and an explanation of terms are given in Appendix 2. The radiological variables originally requested from the AEC for analysis were gross alpha activity and the individual activities of uranium-238, radium-226 and thorium-232. The AEC contributed significantly to the study by determining, in addition, gross beta activity and the individual activities of radium-223, radium-224 and uranium-235. During the second phase of the study the number of radiological nuclides measured was increased to include polonium-210, lead-210, thorium-230, thorium-227, uranium-234, and radium-228. This was done in order to clarify uncertainties in the dose calculated, relating to the non-equilibrium of nuclides with the parent nuclides in the water phase. It was also decided that the protactinium-231 and actinium-227 in the water samples had to be determined on a limited set of samples.

Additional analyses on the last batch of samples were also performed. These analysis included radiological variables on the suspended solids that were left on the filter in the samples.

The use of gross beta measurements for estimating the contributions of beta emitters to the total radiation dose could not be considered, because the measurements were deemed to be unreliable owing to analysis problems caused by the effects of water chemistry. The AEC concurred that the well-established gross beta measurement techniques used by them could not be regarded as suitable for the determination of very low beta activity concentrations in waters characteristic of those sampled in this study. It was accordingly decided not to accept the gross beta data set, but rather to measure those beta emitters likely to contribute significantly to the total ingestion dose, in phase two of the study. Beta emitters measured included lead-210, radium-228, and actininium-227

The methods used for radiological analysis of the samples are given in Appendix 2.

2.6.2 Chemical Variables

Chemical variables, both major inorganic and trace metal constituents, were measured by the IWQS laboratories. The primary reason for collecting chemical variables was to establish whether a relationship could be found between dose and the chemical variables, so as to answer the question as to whether any of the chemical variables could be used as surrogate parameters.

The chemical variables measured were:

1. The following metals (dissolved fraction): aluminium, barium, bismuth, iron, manganese, lead, yttrium and germanium.

2. The following major inorganic determinands: pH, electrical conductivity, total alkalinity, sodium, potassium, calcium, magnesium, ammonium, chloride, fluoride, sulphate, nitrate + nitrite (as N), phosphate as P, and silicate as Si.

The most significant of the chemical variables measured was possibly sulphate, which is formed by the oxidation of pyrite in the mine residue deposits, leading to acidic conditions conducive to the mobilization of some radionuclides into water.

2.6.3 Other Data

Although the radiological data gathered in this study related primarily to radioactivity in the dissolved constituents of the water, limited data were gathered also on radioactivity in the suspended solids. No data on environmental levels of radioactivity in sediments, river banks, vegetation or other possible elements of the human food chain were gathered. Instead, potential radiological impacts from exposure pathways other than drinking water were estimated on an order-of-magnitude basis through the use of screening models.

Other data collected were flow and rainfall data where available. From the very limited river flow and rainfall data that was available for the catchment, no correlation could be established with the radiological data. Unfortunately very few radiation monitoring sites corresponded with flow gauging sites. In the few sites that did correspond, the flow was heavily influenced by man made structures such as dams, weirs, canals and treatment works. This resulted in a highly modified pattern of flow which displayed little or no correlation with radioactivity.

2.6.4 Access to Analytical Data

Analytical results collected during the study can be obtained from the Hydrological Information System (HIS) of the Department of Water and Sanitation. Requests for data from the HIS can be sent directly to:

Directorate: Hydrology

Department of Water and Sanitation

Paterson 536

Private Bag X313

Pretoria 0001

Tel: (012) 338 7500, ask for the Data Supply Section in Directorate: Hydrology

Fax: (012) 326 1488

The official departmental station numbers, provided elsewhere in the report (example C2H073) should be provided with all data requests. Data can be provided in an ASCII format and files can be provided via e-mail.

2.7 Quality Control

A number of actions were taken to address quality control. As a quality control measure, split samples were analyzed by three laboratories, as part of phase two (Appendix 5). These confirmed the accuracy of the radiological analyses.

The AEC conducted the radiometric analyses of the water samples for the study. As a CNS-recognised laboratory, the AEC adopts approved methods and procedures for analysis, and incorporates specific quality control methods. The quality control and validation done by the AEC’s Radioanalytical Laboratory is shown in Appendix 6.

Measurements of uranium by both radiochemical and ICP-MS techniques, during the second phase of the study, allowed comparisons to be made as an additional quality control check. The following good correlation for uranium concentration in mg/l was obtained by linear regression from the 98 samples analysed:

[U]_ICP-MS = 0,993 x [U]_{Radiochemical} - 0,563 (r² = 0,906)

Thorium-232 was also measured by both techniques in the second phase, but a correlation between the two techniques could not be established because the ICP-MS measurements were frequently at the lower limit of detection and therefore inapplicable.

In natural uranium, the activity ratio between uranium-238 and uranium-235 is 21,719. The following good correlations, between the data for the two isotopes, were obtained by linear regression:

Radiochemical (phase 2), 98 data: ²³⁸U / ²³⁵U = 21,341 ± 0,115 (r² = 0,996)

ICP-MS (phase 1), 570 data: ²³⁸U / ²³⁵U = 20,785 ± 0,030 (r² = 0,999)

ICP-MS (phase 2), 63 data: ²³⁸U / ²³⁵U = 22,171 ± 0,571 (r² = 0,860)