This report documents the results of the second part of a research project jointly
funded by Noranda Inc. and the Mine Environment Neutral Drainage program. The first
part is covered in a separate report entitled Hydrogeochemistry of Oxidised Waste
Rock from Stratmat Site, N.B., published concurrently with this report (MEND 2.36.2a).
The overall objective of the project was to understand the geochemical and
hydrological interactions between partially oxidised waste rock and water, and to
improve our capabilities and techniques in the prediction of acidic drainage from waste
rock piles.

The main objective of this part of the research was to understand the hydrology and
solute transport within waste rock piles during infiltration and drainage events in
response to precipitation. Large column tests were conducted to achieve this objective.
Three columns measuring 1.2 m in diameter and 2 m in height and containing up to 3.7
t of partially oxidised Stratmat waste rock were subjected to ten rain simulations. The
bottom area of each column was divided into drainage partitions. During and after each
rain simulation, the volume and the chemistry of the drainage in each partition was
monitored independently over time.

Geochemically, the experimental results suggest that the concentrations of Ca, Pb,
and Al in the drainage are solubility-controlled by gypsum, anglesite, and jurbanite,
respectively. In contrast the concentrations of Zn, Fe, and sulphate in the drainage are
not subject to solubility controls. A dilution hypothesis is proposed to explain the
concentration variations of Zn and sulphate. The hypothesis states that the variations of
these concentrations and the pH are a result of successive dilutions and/or intermixing
of various-stage dilutions of the original pore water, subject to the regulation by redox
reactions and mineral precipitation. All soluble zinc seems to originate from the pore
water and not from dissolution of secondary minerals. The zinc loading in the drainage
is a function of the mass transfer that occurred during dilution and mixing processes.

Hydrologically, the experiments have demonstrated that channelling is a ubiquitous
phenomenon in the waste rock studied. Large channels representing < 5% of the total
drainage area conduct 20-30% of the total drainage flow. Intermediate-size channels
accounting for ~20% of the drainage area carry ~ 40% of the total flow. About 50% of
the drainage area has background or matrix flows that carry ~30-40% of the total flow.
Finally, ~30% of the column base area does not intercept any flow. Channelling is
more pronounced in earlier stages of drainage events and tends to attenuate as the
draining process continues. Channel stability is influenced by variables related to the
rock bed properties and simulated rain characteristics.

On solute transport, the Zn mass balance shows an efficiency of Zn removal from the
pore water that is comparable to the efficiency of a well-mixed system. Whereas the
mechanism giving rise to this observation is unclear, it is unlikely that the transport of
solutes takes place by pore water displacement. There is no simple relation between
solute concentrations and drainage flow rates. A conceptual dendritic-reticulate
channelling model is proposed on the basis of the experimental observations.

Statistical analysis suggests that the flow density and the zinc loading can be
appropriately described by the lognormal distribution, whereas the Zn concentrations
are distributed normally.

The main challenges in flow and solute transport modelling are channelling and
interactions between flows and geochemical processes. In this study, porous media
flow rock is differentiated from channelling flow rock based on hydraulic properties.
Furthermore, five basic component structures that make up a waste rock pile are
identified. Each structure has distinct characteristics and should be modelled with
different approaches. Factors influencing water flows and flow effects contributing to
flow heterogeneity are discussed. A mathematical representation of channelling
phenomena is developed, which can be coupled with statistical relationships between
solute concentrations and flow rates to model solute transport in waste rock piles.

The kinematic wave model, recommended by an earlier MEND study, was applied to
the experimental data. The model did not appear to adequately predict the channelling
flow characteristics of the waste rock. It over-predicted the extent of larger channel
flows at the price of smaller channel flows. The lack of applicability probably stems
from the fact that the kinematic wave model precludes merges and splits of flows within
waste rock. The model may be more appropriate to coarser waste rocks.

Further fundamental research and case studies are needed to advance our
understanding of flow and solute transport in waste rock piles to a point where
concentrations and loadings in the drainage can be reliably modelled.