Critical evaluation of acid mine drainage modelling, 1986-1991 for Curragh, down valley tailings, Faro, Yukon

Abstract

A critical evaluation of the test plot program and tailings modelling by SRK of Curragh's Down Valley tailings system has been performed. Two decommissioning options described by SRK were considered. The first option, Alternative #4, includes a combination of flooding (water cover) over the intermediate tailings with a composite cover of fine tailings (slimes) and rock over the original and secondary tailings. The second, Alternative #5, involves a water cover after removing original and some secondary tailings for reprocessing. Predictions of zinc loadings to, and concentrations in the downstream creek were the focus of the modelling. All zinc values are considered incremental from the effects of the Down Valley tailings only. The cover test plot program represents a good and unique database for comparing various options for tailings disposal. It is agreed that a composite (slimes) cover provides the best protection (excluding the water cover) from oxidation among the cover options tested. The data, however, are inadequate and therefore should not yet be used for estimating oxidation and metal release rates in actual tailings impoundments. The very shallow water table levels in the test plots make comparison to drained tailings inappropriate. Careful measurements of oxygen and moisture content could improve the interpretation of the oxidation rates and complement the currently used mass balance approach. The tailings modelling approach used by SRK was reasonable and represented state-of-the-art methods. However, critical uncertainties of some variables were not considered by SRK and the modelling of oxidation rates in tailings below covers (Alternative #4) has not yet been verified by actual field measurements; this remains a research topic. One major uncertainty associated 1 with this modelling is the prediction of the quantity of water (degree of saturation) retained in the cover material or tailings above the water table. Small errors in the estimates of the degree of saturation can translate to large variations in oxidation rates. Because the oxidation rates in tailings are very sensitive to the saturation values above the water table, any modelling estimates are prone to large errors based on small variations in the degree of saturation of the cover. In turn, the saturation values are very sensitive to the estimated evaporation that is very difficult to predict precisely. It is concluded that the modelling predictions used by SRK for Alternative #4 were highly optimistic with respect to cover performance and estimates of saturation were very high under field conditions. More conservative calculations with reasonable uncertainty for moisture content in the composite covers provide average annual zinc concentrations in Rose Creek that are a factor of 10 to 100 times higher than those predicted by SRK. This large discrepancy suggests that cover performance will require close monitoring and evaluation to verify the predicted values. The water cover (Alternative #5) will also allow zinc release by oxidation induced by movement of dissolved oxygen through the tailings with flowing water. This mode of oxidation was ignored by SRK in the modelling estimates. Although zinc levels will be sensitive to release rates, even the water cover could produce incremental concentrations as high as 0.06 mg-L"1 in Rose Creek. These estimates are based on the correlation between sulphate and zinc in humidity cells and more experimental evidence for underwater covers will improve the confidence and may reduce predicted concentrations in these conservative estimates. The alternate concentrations of zinc in Rose Creek predicted in this review should not be construed to contradict the values suggested by SRK. The higher values suggested here, however, represent deviations from the overly optimistic SRK values and are based on reasonable uncertainties in critical variables that have not yet been "field tested". The most significant uncertainties that impact the predictions are: (1) the long-term degree of saturation of the covers and (2) the zinc release rates. The latter are correlated with sulphate production in humidity cell tests and field data from tailings water samples and these are factors of 100 and 10, respectively, greater than the value used by SRK. Therefore, increasing confidence in the predictions of zinc loadings and concentrations in Rose Creek could be attained through: (1) verification of moisture content behaviour in the covers (measurements in modified test plot monitoring); and (2) further verification of zinc release rates for covered and underwater conditions. The zinc release rates may be possible to verify with laboratory experiments. There are MEND projects in progress that may provide useful information for this purpose. The long-term performance of covers may require a few years of monitoring in the field to verify. In addition to further measurements for prediction, careful monitoring and data interpretation of the active oxidation zones in the tailings will be required to verify the zinc production rates in the tailings and loading rates to Rose Creek after decommissioning.