Click
on Figure 1 to enlarge itThis case study is copyright (©) 1998 by Kevin A. Morin and Nora M. Hutt, and is adapted from Chapter 4 of Environmental Geochemistry of Minesite Drainage: Practical Theory and Case Studies.
The comparison of minesite-drainage chemistry to rain at first seems puzzling, but there are many reasons why they are similar. First and foremost, there would be no drainage chemistry if there is no drainage. Since drainage is ultimately derived from precipitation, "no rain" means "no drainage" which in turn means "no drainage chemistry".
Some mines in the Atacama Desert of Chile know this well - there is no rainfall during some years. However, as open pits and underground workings at these minesites move below the deep water table, there is eventually drainage and its chemistry to deal with.
There are less obvious similarities between drainage chemistry and rainfall. This includes the primary focus of this Internet case study - their cyclical nature.
Over the period of many decades, an mean annual amount of rainfall can be calculated. In reality, each year's rainfall will be somewhat above or below the mean, with some occasional extreme amounts. However, most years are close to the mean. Also, over a one-year period, there can be seasonal fluctuations as wet and dry seasons.
Drainage chemistry is like this too. Compilations of long-term monitoring programs spanning almost 30 years at several minesites (e.g., see Chapter 4 of our book and excerpts below) show that mean annual aqueous concentrations of a particular metal are similar year after year as long as pH remains generally constant. Also, the range of concentrations over each annual period is about the same year after year, even if pH varies (Figure 1).
Click
on Figure 1 to enlarge it
With rain, there are wetter days and drier days. With drainage-chemistry, there are days with higher concentrations and days with lower ones. No one can predict months in advance the exact rainfall on a particular day, but one can often estimate when the wetter months will be. No one can predict months in advance what the aqueous concentration will be on a particular day, but one can often estimate in what months the higher concentrations will occur. There can be a fierce storm over the period of a few hours, releasing torrents of rain. There can be an extremely high aqueous concentration over the period of a few hours, but this may be rarely detected by grab samples of drainage collected once a month.
Further to the last point of drainage-chemistry "storms" over the period of a few hours, we have presented talks on the fictitious "Lucky Mine" and the "Unlucky Mine". The Lucky Mine collects grab samples of drainage once a month, but by coincidence has never collected one during the drainage-chemistry storm. The Unlucky Mine happened to collect one monthly sample during the drainage-chemistry storm and there is panic. The analytical laboratory is asked to repeat the analysis, and the laboratory confirms a high concentration. The mining company suspects contamination during collection or analysis, and collects another sample days or weeks after the drainage-chemistry storm. Of course, the concentration in this new sample is much lower, and the former analysis is dismissed as anomalous and erroneous.
We have been curious for a long time on the cyclical nature of drainage chemistry, and several sources of information have allowed us to compile examples. We have compiled and interpreted drainage-chemistry data from minesites that have operated for almost 30 years, which show the annual repetition of annual mean concentrations and of ranges as annual standard deviations. We have also compiled and interpreted high-frequency analyses, particularly samples from one minesite collected every four hours at eight monitoring stations for up to four years. This information is in Chapter 4 of our book and is summarized below.
However, we recognize that not all minesites show this regular cyclical behavior in drainage chemistry . In our experience, about 15% of minesites do not have it.
The main reason for the cyclical pattern often lies in secondary-mineral equilibrium (a topic for another monthly Internet case study), which establishes the mean annual concentration (Table 1). Then the seasonal variation (standard deviation, Table 1) is caused by all other natural physical, chemical, and biological processes (like dilution and temperature) and artificial factors (like analytical error). Because an open-environment system like a minesite is so complex, it is basically impossible to delineate the contribution of each factor to the variation.
TABLE 1 Annual Statistics for Drainage Chemistry at a Monitoring Station
Receiving Drainages from Several Waste-Rock Dumps and an Open Pit |
||||||||
Log10 Mean1 |
Log10 Standard Deviation1 |
|||||||
90-91 |
91-92 |
92-93 | 93-94 | 90-91 |
91-92 |
92-93 | 93-94 | |
| pH | 6.27 | 6.93 | 7.08 | 6.86 | 0.71 | 0.85 | 0.67 | 0.69 |
| Conductivity | 3.06 | 3.09 | 3.1 | 3.16 | 0.12 | 0.09 | 0.11 | 0.06 |
| Alkal. (mg/L) | 1.74 | 1.71 | 1.7 | 1.62 | 0.07 | 0.45 | 0.36 | 0.35 |
| Acidity (mg/L) | 1.32 | 1.33 | 1.51 | 1.6 | 0 | 0.33 | 0.24 | 0.05 |
| Cu (mg/L) | -0.94 | -1.23 | -1.49 | -1.43 | 0.41 | 0.46 | 0.42 | 0.46 |
| Zn (mg/L) | 0.41 | 0.22 | 0.25 | 0.41 | 0.38 | 0.44 | 0.4 | 0.34 |
| Cd (mg/L) | -1.74 | -1.82 | -1.79 | -1.72 | 0.17 | 0.22 | 0.23 | 0.21 |
| Sulfate (mg/L) | 2.89 | 2.9 | 2.91 | 2.93 | 0.06 | 0.09 | 0.1 | 0.08 |
| Ca (mg/L) | 2.37 | 2.39 | 2.43 | 2.44 | 0.04 | 0.09 | 0.1 | 0.07 |
| Mg (mg/L) | 1.46 | 1.51 | 1.59 | 1.6 | 0.04 | 0.16 | 0.1 | 0.07 |
| Al (mg/L) | -0.78 | -0.71 | -0.56 | -0.57 | 0.68 | 0.33 | 0.35 | 0.37 |
1 All values are logarithms except pH |
||||||||
Another way of showing the data is through scatterplots of one aqueous parameter against another. Scatterplots of pH against aqueous metal concentrations, showing up to 30 years of monitoring data, demonstrate the remarkable regularity of concentrations that can arise at minesites (Figures 2 and 3, below).
Click
on Figure 2 or 3 to enlarge it
These scatterplots (Figures 2 and 3) can be summarized using (1) the "best-fit" equations on the scatterplots to define the mean annual concentration and (2) the standard deviation above and below the line to define the seasonal variation. This information can then be compiled for each drainage-chemistry parameter at a minesite to form an "empirical drainage-chemistry model" (EDCM, Table 2).
TABLE 2 Example of an Empirical Drainage-Chemistry Model (EDCM) Including an
Open Pit, Several Waste-Rock Dumps, and a Tailings Impoundment |
|||
Parameter |
pH Range | Best-Fit Equation | Log(Std Dev) |
| Acidity | pH < 3.5 | log(Acid) = -0.932pH +5.864 | 0.345 |
| pH > 3.5 | log(Acid) = -0.360pH + 3.862 | ||
| Alkalinity | pH > 4.5 | log(Alk) = +0.698pH - 3.141 | 0.654 |
| Dissolved Aluminum | pH < 6.0 | log(Al) = -0.925pH + 4.851 | 0.429 |
| pH > 6.0 | Al = 0.2 mg/L | ||
| Dissolved Arsenic | < 0.2 mg/L | 0 | |
| Dissolved Cadmium | pH < 3.0 | Cd = 0.07 mg/L | 0 |
| pH > 3.0 | Cd = 0.015 mg/L | ||
| Dissolved Calcium | log(Ca) = +0.619log(SO4) + 0.524 | 0.375 | |
| Dissolved Copper | pH < 3.4 | log(Cu) = -1.485pH + 6.605 | 0.692 |
| 3.4<pH<5.4 | log(Cu) = -0.327pH + 2.666 | ||
| pH > 5.4 | log(Cu) = -1.001pH + 6.307 | ||
| Total Copper | log(CuT) = +0.962log(CuD) + 0.180 | 0.23 | |
| Dissolved Iron | pH < 4.4 | log(Fe) = -1.429pH + 6.286 | 0.807 |
| pH > 4.4 | log(Fe) = -0.455pH +2.000 | ||
| Total Iron | If diss Fe>1.0, total Fe=diss Fe | 0 | |
| Dissolved Lead | Pb = 0.05 mg/L | 0 | |
| Dissolved Nickel | log(Ni) = -0.317pH + 0.853 | 0.607 | |
| Total Nickel | total Ni = diss Ni | 0.613 | |
| Dissolved Selenium | Se = 0.2 mg/L | ||
| Dissolved Silver | Ag = 0.015 mg/L | ||
| Dissolved Zinc | log(Zn) = -0.441pH + 1.838 | 0.667 | |
| Total Zinc | total Zn = diss Zn | 0.144 | |
Turning now to high-frequency sampling, this field study involved thousands of chemical analyses of drainage samples collected every four hours from eight surface-drainage-monitoring stations for up to four years (drainage only flowed for six months each year). The mean annual concentration and standard deviation at each station, based on the entire year's database, was compared to calculated means and standard deviations based on only a portion of the database. The intent of this was to identify whether less frequent sampling (like quarterly or monthly) would provide similar values to those from the entire database. If they did not, then high-frequency sampling would always be needed to properly define annual mean concentrations and standard deviations, including drainage-chemistry "storms" (see beginning of this Internet case study).
For example, the mean concentration and standard deviation were calculated for simulated monthly sampling by randomly selecting by computer an analysis from each month, for a total of six months when the drainage was flowing. This random selection was done 25 times for each simulated period of sampling (monthly, weekly, etc.).
The results of this (Figure 4, to the left, click to enlarge) showed that, as the simulated frequency of sampling increased, the range of the 25 means and standard deviation began to narrow and converge on the values from the full database. This demonstrated that, to calculate the annual mean concentration and standard deviation within general analytical accuracy, the frequency of sampling should be monthly at some stations and weekly at others.
Click on Figure 4 to enlarge it
Once reasonable estimates of the mean annual concentration and standard deviation are available, peak concentrations that may be encountered during drainage-chemistry "storms"can be calculated. This is based on standard statistics and probability values, where the highest mean monthly concentration, for example, will be +1.73 standard deviations above the mean concentration.
© 1998 Kevin A. Morin and Nora M. Hutt
References:
Morin, K.A., N.M. Hutt, and I.A. Horne. 1995a. Prediction of future water chemistry from Island Copper Mine's On-Land Dumps. 19th Annual British Columbia Mine Reclamation Symposium, Dawson Creek, B.C., June 19-23, p. 224-233.
Morin, K.A., N.M. Hutt, and R. McArthur. 1995b. Statistical assessment of past water chemistry to predict future chemistry at Noranda Minerals' Bell Mine. Proceedings of the Conference on Mining and the Environment, Sudbury, Ontario, May 28 - June 1, Volume 3, p.925-934
Morin, K.A., I.A. Horne, and D. Riehm. 1994a. High-frequency geochemical monitoring of toe seepage from mine-rock dumps, BHP Minerals' Island Copper Mine, British Columbia. IN: Proceedings of the Third International Conference on the Abatement of Acidic Drainage, Pittsburgh, Pennsylvania, USA, April 24-29, Volume 1, p.346-354.
Morin, K.A., I.A. Horne, and D. Flather. 1993. The appropriate geochemical monitoring of toe seepage from a mine-rock dump. Proceedings of the 17th Annual Mine Reclamation Symposium, Port Hardy, British Columbia, May 4-7, p.119-129. Mining Association of British Columbia.
For more details and case studies, see Environmental Geochemistry of Minesite Drainage: Practical Theory and Case Studies.
Created by K.Morin