Tailings Pond Water Balance

Water systems in modern mines are complex and an understanding of the tailings pond and plant water balance is necessary. Control of the water balance can avoid problems such as operating losses from water shortages or hazardous releases as a result of a water surplus. Here is a schematic of a typical water balance. By regularly monitoring the various components of the water balance, effective water management can be practiced.

For operators who depend on reclaim for process water, the water balance must be closely monitored. There are many factors that affect the balance and these should all be examined in deriving an equation.

Factors to consider include feed rate of tailing to pond, percentage of water released as supernatant and % retained in deposited tailing, evaporation losses, precipitation contribution both direct and run-off, seepage losses, amount of seepage collected and returned, amount of supernatant discharged from system, reclaim water removal rate and quantity of any other inflow such as pit dewatering. Most operators monitor the quantity of fresh water added to the process from outside sources. The operator should also equip himself with a calculation of pond level vs contained volume. Armed with these pieces of information, rates of rise and rates of advance can be calculated and long term adjustments are possible. Changes in consumption rate of fresh water and diversion of run-off either in our out of the pond may be used to bring the system into balance.

Conceptually, analysis of contaminant transport from mine tailings is not different from analysis of contaminant transport away from any concentrated source of elements that is chemically different from the surrounding media. Contaminant transport analysis includes the hydrologic flow component and the geochemical mobility factor. The relative significance of the geochemical and hydrologic phenomena depend upon

1. the chemical and hydrologic conditions of the tailings and environment,
2. the chemical constituents of interest, and
3. the distance along the flow path away from the tailings.

The large differences between geochemistry of tailings and the groundwater/soil system create a chemical disequilibrium within the environmental system. This implies that reactions will occur to readjust the perturbed system toward chemical equilibrium, which determines the mobility (i.e., retardation) of the contaminants in the groundwater. The significant reactions are a combination of precipitation/dissolution and adsorption reactions that change the chemical character of seepage from tailings to a chemical composition similar to the natural groundwater. The question for contaminant transport analysis is, what distance will the seepage travel before the chemical character is indistinguishable from the background.

This paper describes a model developed from the application to uranium mill tailings which illustrates the concepts of geochemistry to predict contaminant transport. Examples of the geochemical system from mill tailings in two environments are discussed. In one case the pH difference between the tailings and soil is the major control of geochemical mobility. Whereas, in the second case, geochemical mobility is dominated by redox potential (Eh) differences.

The model is not a continuous numerical function, but rather it defines spatial units that are chemically unique portions of the flow path and models each unit independently. The term “model,” as used herein, is the conceptual description of the chemical reactions and mechanisms important in the groundwater/soil system. The model is different from the computer code, i.e., PHREEQE, used to calculate the chemical equilibrium between the aqueous and solid phases. The computer code is merely an instrument for executing the model.

In an attempt to overcome the limitations of retardation coefficients and thermodynamic/flow models, engineers defined three units of the tailings/environmental system, each having distinct geochemical properties (Figure 1):

• Tailings as the source
• Mixing zone of seepage with groundwater/soils
• Background soil/groundwater systems

The mixing zone is distinguished from the local soil/groundwater system by the degree of chemical disequilibrium between the groundwater and soil/rock mineralogy. The system is far from equilibrium within the mixing zone: whereas, chemical equilibrium (or at least a steady state) exists in the background system.

Each unit is modeled independently and linked together by source terms. The concentration of an element exiting a unit is the source term for the next unit. Seepage from tailings into groundwater causes a change in the chemistry of the groundwater and the consequent disequilibrium between groundwater and minerals. This results in precipitation/dissolution reactions between the groundwater and minerals to re-establish background steady-state conditions.

Modeling the mobility of contaminants within a given unit must distinguish between major and trace elements. For purposes of modeling, major components are those that comprise the major mineral phases and play a role in controlling the pH and Eh (redox potential) of the system. Alternatively, trace components do not play a role in defining the chemical environment, but rather their behavior is determined by the major chemistry of the system. The aqueous concentrations of major elements are primarily determined by precipitation/dissolution reactions. Whereas, adsorption is often the dominant mechanism for controlling the concentrations of trace elements. Thermodynamic calculations of precipitation/ dissolution reactions address the major element chemistry directly. The major element chemistry influences the retardation of trace components indirectly by affecting the substrates for coprecipitation and adsorption and by determining the aqueous species of the trace components.

Freshly precipitated hydrous oxides of iron, manganese, aluminum, and silica are commonly observed substrates for adsorption. Different species are more likely to be adsorbed at different pH conditions. At a pH lower than the isoelectric point, hydroxyl groups on the substrate surface are positively charged and are anion adsorbers. At a pH greater than the isoelectric point, the surfaces become negatively charged and are predominantly cation adsorbers.

Based upon the thermodynamic approach to modeling the potential for impacts on the groundwater, the problem is reduced to;

• Characterization of the tailings mineralogy and the chemistry of the interstitial water.
• Characterization of the soil mineralogy and the chemistry of the groundwater.
• Characterization of the reactions that occur between the seepage from the tailings and the groundwater.
• Development of the relationship between the reactions of the major elements and the mobility of the trace elements.
• Expression of the geochemistry in a form suitable for use in contaminant transport models.

Mine Water and Mass Balance Models

“Rarely is there just the right amount of water, rather there is either too little or too much”. To effectively manage site-specific water challenges, mine operators need a comprehensive, calibrated, and continually updated water and mass balance model. This model accounts for all of the water and solutes entering and exiting a defined location. Water and mass balances arc essential for understanding the quantity and quality regarding water needs and water availability for the facility, and the environmental consequences of mine water use.

Todays mines typically have complex water management strategies as water is reused and recycled in an attempt to minimize environmental impacts. Water from one mine feature is often collected and pumped to another mine feature for reuse, resulting in a highly connected system. A water and mass balance should integrate all of the mine features and other processes that can affect the quantity or quality of water moving through the system. Without a comprehensive water balance, changes in the design or operation of one mine feature may result in unexpected and unwanted changes to water quantity or quality elsewhere in the system. A water and mass balance for the entire mine, not just the processing plant, is vital to the appropriate sizing of mine water infrastructure, the design of improvements to plant water systems, and the implementation of environmental controls.

Water balances can also assist in facility risk assessments where upset conditions, such as extreme weather events or plant shutdowns, can be evaluated and preparations made accordingly. A large rainstorm, a long drought, or other extreme weather events can be evaluated within a water balance, allowing for a better understanding of the impact the upset condition can have on the mine water systems, resulting in better preparedness for a real event.

A water and mass balance model is a decision support tool designed to assist mine operations with mine site water management and to help regulators assess potential environmental impacts on water resources. It can also be a tool to communicate the impacts of potential changes to internal stakeholders (e.g., mine management) and external stakeholders (e.g., agencies, communities) that may have differing perspectives on the issues that surround water management at the mine. This full range of stakeholder perspectives should guide the many choices that must be made when creating a water and mass balance model. The team setting up the model must consider the following factors:

• Model complexity
• Model inputs and data sources
• Model outputs required
• Modeling to manage uncertainty
• Modeling programs

Model Complexity

The complexity of the water and mass balance is driven by the questions being asked of the model. A water balance can be as simple or complex as needed for the application. Model complexity relates to the geographic area to be covered (model boundaries), level of detail necessary, time periods to be considered, and frequency at which model predictions arc needed (time-step). Complexity also depends on whether a deterministic or stochastic model is used, as discussed later in this chapter. A simple water balance might be centered around a solitary water body at a single point in time, while a complex water balance could include multiple water bodies and mine features as well as the water demand of the individual processes within the plant, evaluated continuously through time with a weekly time-step.

Water and mass balance models are not static and should be updated throughout the mine life cycle. The process of creating a water balance model should begin in the early planning phases of the mine, it possible, and the model should be continually updated and refined through construction, operations, and closure.

Over time, a mine water and mass balance generally grows in complexity and accuracy as the mine develops and data are continually collected and used in the model. This is driven by more detailed questions being asked of the mine water and mass balance, For instance, in the planning stage of the mine, the first question is typically “is there enough water?” with a general water balance with a long time-step (e.g., yearly). During operations, the question may change to “how can water be used more efficiently?”. As a project becomes more defined, the water balance can become more refined with shorter time-steps (e.g., daily or monthly), more inputs, including water quality information, and potentially fewer unknown parameters.

The model boundary for mining operations is most commonly the property boundary, but water and mass balances can be scaled to answer site-specific questions. For example, if there is a concern only with the water quantities within a certain area of the mine, such as in a single mine pit, then a water balance can be built around that specific feature. If, however, there are concerns about water quality beyond the mine property, about water availability for other stakeholders, or about water supply to culturally or environmentally important locations, then the boundaries should be expanded to answer these questions. For example, at Rio Tintos Australian iron ore operations, the mine water management strategy accounts for water supply and demand at regional, local, and mine-sire scales, and evaluates the potential impact mining operations can have on each of these scales. As a water balance is being created, it may become apparent that it is necessary to expand the area to comprise other water impacts that were not initially included or, conversely, to add detail within a smaller area to better represent a specific process relative to the question at hand.

Depending on the complexity of the model, it may become necessary to use separate speciality models to develop the inputs necessary for the water balance. Detailed models arc available that can evaluate individual aspects of a comprehensive water balance, such as groundwater flows, water quality predictions, or geochemical interactions. These separate speciality models may be necessary to expand the derail of the water and mass balance model, and the results can be used as inputs to an overall comprehensive water and mass balance.

Model Inputs and Data Sources

At its most basic, a mine water balance model consists of the known inflows, outflows, and associated production values that affect water usage, and solves for the unknown water flows. When combined with a mass balance model, the chemistry of the different water flows is included as well as processes, such as chemical weathering of waste rock, that can affect water quality. The Minerals Council of Australia Water Accounting Framework outlines the types of inputs needed as part of a mine water balance. Model inputs can generally be grouped into five categories:

1. Mine process and dewatering inputs, such as production rates, mine plan, and dewatering rates
2. Physical inputs, such as topography, land-use characteristics, and associated runoff coefficients
3. Climate inputs, such as precipitation, runoff, and evaporation
4. Hydrologic and hydrogeologic inputs, such as stream or groundwater inflows or losses
5. Geochemistry and water quality inputs, such as chemical weathering of waste rock or wall rock

Water and mass balance models rely on many data sources, including data that are measured, simulated, and estimated, The accuracy of model results depends on how many of the inputs must be estimated rather than measured. Accuracy generally improves over rime, because as the mine develops, more actual data can be measured and applied.

Some of the data necessary for the development of a water and mass balance may be readily available. For example, water flows appropriated for use or discharged from the mine may have been measured and reported to regulatory agencies, which can then be incorporated into the model. The data collection process typically includes reviewing engineering drawings and flow sheets of processing facilities, measuring water levels or flows, conducting a bathymetric survey of water bodies to evaluate storage, collecting water quality samples, delineating watersheds, and discussing operational practices and strategies with the people that control the water use at the mine. Some internal flows that are not known may be estimated based on pipe sizes and pump curves. Existing and former personnel can often be an invaluable source for data not otherwise documented based on their personal experiences in the operation.

An essential step in gathering data for the water balance model is to evaluate the accuracy and relevance of the information for the model that is being created. Data must fit the location and system being modeled. This may require adjusting conventional techniques to better suit local conditions. For example, in the Andean High Cordillera region, where evapotranspiration plays a critical role in the ground water recharge, mine operators realized that conventional methods for calculating potential evapotranspiration were not sufficiently accurate. A new method was created that calibrated the calculated values to observed data from the region, allowing the system to be more accurately represented.

After available data have been gathered, data gaps should be identified and analyzed to determine their impact on the water balance. Missing data can be measured, approximated, assumed, or solved in the model. The method of addressing data gaps is based on engineering judgment and needs to consider the significance of the missing data. Solving for the unknowns within the water balance is the final option for working with unknown data. There is typically a final unknown, typically an unmeasureable loss, which is determined by closing the water balance.

As with any model, the quality of the input data directly impacts the quality of the results. Deciding how to work with the accuracy, uncertainty, and variability in the inputs of the water and mass balance model can be a driving factor in determining whether to create a deterministic or a stochastic model, as discussed later in this chapter.

Model Outputs Required

When developing a mine water and mass balance model, it is important to understand what output from the model will be needed, which is driven by the questions that are being asked of the model. Model outputs typically include ranges in flows, water volumes, or water levels, and constituent concentrations at selected locations. In the case of water and mass balances, the outputs should provide the expected water quality in each location along the water balance, including at discharge points. Time series plots, tables, and charts are the common format for presenting and displaying model output.

Deterministic and Stochastic Modeling to Manage Uncertainty

In addition to answering the questions being asked of the model, a water and mass balance model should also provide an estimate on the range of potential answers, from average to extreme conditions, and identify the uncertainty around the results. This can take the shape of several different deterministic model runs or stochastic modeling, depending on the needs of the model. Deterministic models have single-value inputs creating single-value results, whereas stochastic models have probability distributions for the inputs producing probability distributions for the results. Depending on the type of water balance model and desired results, either type of model may work for a given situation. The model can transition from deterministic to stochastic as dictated by the project, similar to how the model complexity can increase as the model is updated.

Deterministic models essentially simplify the real world by using an average value or a worst-case value as the input. This produces a single-value result that is easy to use and apply. Deterministic models are best suited for well-defined data, such as historic records, and are often used with less-detailed models. Big picture models are generally looking at an overview of a system, and a deterministic model can provide a usable result that, although, not fully encompassing all of the unknowns of the system, will work for the situation. When a deterministic model is used with data that are not well defined, a single input value muse be selected to represent the range of data. The range of uncertainty can be addressed by running a deterministic model with average conditions, and then again to assess the model request with extreme conditions to evaluate a change in climatic variables, production rates, or other model inputs. However, selecting conservative values for multiple inputs of a model can produce a result that is too conservative, without providing an understanding of the range of uncertainty around the result.