Introduction

As it has been emphasized in several previous studies (Tatsuhara et al. 2012; Tamoto et al. 2015; Tabelin et al. 2017a, b, 2018) that various hazardous trace elements (e.g. As and Se) in rocks are already subjected to progressive weathering processes resulting in contamination of geological environments. This risk increases even more with mining activities carried out in places where mineralization occurs. The rapidly growing economies of countries require new resources and raw materials and thus lead to the growth in the mining sector worldwide. Excavation of geological sequences for mining purposes brings simultaneously major environmental problems due to disturbing and changing of some naturally stabilized field conditions. Herrera et al. (2007) stated that many abandoned or closed mine sites ultimately resulted in the formation of acid mine drainage (AMD). This key mining-related ecological issue results in releasing contaminants (e.g. soluble metals, oxidation products) and decreasing the pH of surface and groundwater in the close vicinity of abandoned mined land. It is well known that this natural process is a product of the oxidation of sulphide minerals whenever oxygen dissolved in water comes into contact with these minerals found in rock and mine wastes. The remediation of pollution arising from AMD generation is too complicated and not so easy due to the cost and time required to solve such major environmental issue. According to EMCBC (1996), its treatment costs millions of US dollars and it can go on for many centuries to come with the current technology. The damage mitigation costs for the USA was estimated by Feasby and Tremblay (1995) to be between 2 and 5 billion dollars; its world costs were estimated by Weatherell et al. (1997) as over 10 billion dollars, and currently, in the USA, the mining industries spend more than one million dollars each day in the treatment of acid effluents (Grande et al. 2005). Park et al. (2019) emphasized that AMD is an important global problem for the environment as well as the public health of many countries. Considering the importance of AMD generation and given its negative impact on the environment, the understanding and simulation of weathering processes of any mine site containing sulphide minerals are very significant for the prediction of AMD hazard and developing convenient remediation methodologies before, during and after mining operations.

The determination of AMD potential of any site at laboratory condition is based on two main different testing techniques, which are known as static and kinetic tests. In comparison with a static test, the kinetic tests are better in representing field conditions due to its applicability to controlling the rate of acid generation. However, large volumes of the specimen are essential for kinetic tests and the chemical reaction of acid formation lasts longer time based on the oxidation rate of sulphide minerals. Due to relatively lower reliability of static tests completed offsite and do not reflect actual field measurements (Parbhakar-Fox and Lottermoser 2015), kinetic tests were preferred by many researchers (e.g. Strömberg and Banwart 1999; Demers et al. 2008; Sapsford et al. 2009; Parbhakar-Fox et al. 2011; Parbhakar-Fox et al. 2013). Up to now, different approaches considering kinetic principles have been developed and recommended for use in the prediction and/or assessment of AMD. Among these approaches, the column test is known as more representative for the simulation of actual weathering processes (Bradham and Caruccio 1991). However, Strömberg and Banwart (1999) verified that the behaviour of acid generation is still unclear because of inconsistencies between the laboratory and field-based AMD measurements in conjunction with variations in complexity and physical scale between these two different systems. Kalyoncu Erguler et al. (2014) conducted comprehensive research on the relationships among cumulative concentrations of Ca2++Mg2++Mn2+, SO42− and the column dimensions. They found statistically significant correlations between these parameters and emphasized the variations in amount of sulphate and Ca2++Mg2++Mn2+ can be clearly defined by the length, diameter and volume of the column, as well as, the particle size of samples. Despite extensively performing kinetic column test to predict AMD (e.g. Strömberg and Banwart 1999; Benzaazoua et al. 2004; García et al. 2005; Kalyoncu Erguler et al. 2014), the particle and column size-dependent mass release rate of oxidation products and other constituents have not been investigated in detail to upscale geochemical behaviour for understanding AMD generation for real field conditions. Therefore, considering the constantly increasing of AMD’s impact on the environment over time and deficiencies related to upscaling results of the kinetic column test, an ongoing research program has been conducted to achieve given objectives below:

  • To increase the reliability of kinetic column tests,

  • To determine effects of particle size and column dimensions on the rate and quantity (instant and cumulative values) of soluble metals, main oxidation (SO42−) and neutralization (Ca2+, Mg2+, Mn2+, etc.) products,

  • To assess the interaction principles between oxidation and neutralization products during kinetic column tests and to find a suitable way to incorporate effects of column dimensions and particle size within the system,

  • To find suitable practical approaches and/or understandings by using key parameters controlling the rate of acid generation for upscaling geochemical behaviour and kinetic principles of AMD generation.

To achieve objectives above, samples composed of sulphide minerals possessing a tendency to form AMD were taken from a mining site located in the northeastern part of Turkey. Numerous columns were manufactured in various sizes and the collected ore samples were subjected to the crushing process for assessing particle size-dependent AMD generation. The initial mineralogical compositions of samples were determined by X-ray diffraction (XRD), scanning electron microscope (SEM), mineral liberation analyser (MLA) and X-ray fluorescence (XRF). All columns were filled with crushed samples and tests were performed with an initial pH value of 8.1. The time-dependent changes in the concentration of many constituents and fluctuation in pH values of contaminated water recovered from columns were measured during experiments and then analysed. By performing statistical analyses, predictive models thought to be useful for upscaling kinetic principes-based geochemical behaviour of AMD processes were obtained by performing simple and multiple regression analyses among column dimensions and main parameters controlling the rate of acid generation.

Materials and methods

Physical properties of samples

The samples were collected from an open-pit quarry located in Murgul, Artvin City, in northeastern Turkey. The sampling site is given in Fig. 1. In order to determine all physical, mineralogical and AMD properties of unweathered specimens, fresh ore samples having a mass of 300 kg were collected from suitable places. In addition, taking into account the importance of joints on acid formation, the frequency, orientation, spacing, persistence, weathering and roughness of discontinuities found in studied rock masses were also measured by line surveying technique. The results reveal that discontinuities provide enough capacity to transmit freshwater for new AMD formation and to cause infiltration of contaminated water to the underground water resources in the studied site.

Fig. 1.
figure 1

Sampling site map (the longitude and latitude of sampling location are 41.589466° and 41.288814°, respectively)

Full size image

The porosity, water absorption, void ratio and unit weight (γd) of these samples were measured on ten randomly selected samples in accordance with ISRM (2007), and the results were 1.01–2.68 %, 0.27–0.92% and 0.01–0.03, 27.09–38.54 kN/m3, respectively. Samples were crushed to obtain relatively coarse (minus 3.35 mm) and fine (minus 0.625 mm) particles for investigating the effect of particle size distribution on the AMD generation. Particle size distribution curves of these samples are illustrated in Fig. 2. Based on the Brunauer-Emmett-Teller (BET) method, the precise specific surface areas were determined as 0.98 (m2/g) for coarse and 1.24 (m2/g) for fine samples. Samples consist of two different particle size distribution curves that were placed in columns for kinetic tests. Based on 12 different constant head permeability tests, it was determined that the hydraulic conductivity values of fine samples ranged between 0.0017 and 0.0231 cm/s, with a mean value of 0.0116 cm/s. Similarly, these values changed between 0.0028 and 0.0306 cm/s, with a mean value of 0.0126 cm/s for coarser particles. According to USBR (1974), the drainage capacities of all columns are enough to transmit water during the kinetic column test.

Fig. 2.
figure 2

Particle size distribution curves of crushed samples used in the kinetic column test

Full size image

Mineralogical and chemical composition of samples

X-ray diffraction (XRD) was utilized at the laboratories of the Mineral Research & Exploration General Directorate (MTA) situated in Ankara, Turkey, to identify initial mineralogical characterization of samples and then describing the oxidation of minerals during kinetic column tests. It was found in XRD measurements that, in addition to quartz, chalcopyrite and pyrite known main reason for acid formation, constitute a significant proportion of samples’ mineralogical composition. Generally, XRD provides a qualitative or semi-quantitative assessment for the mineral content of materials. Thus, in addition to XRD analyses, mineral liberation analyser (MLA) was also used to achieve the mineral percentage of fresh samples. The latter indicated that fresh ore samples were composed mainly of chalcopyrite (32–38%), quartz (32.5–37.5%), pyrite (19.5–23%), biotite (4.8–6.5%) and ankerite (2.3–3%) with less than 0.5 % of other minerals that unimportant for kinetic behaviour of tests. The results of MLA revealed that the total percentages of pyrite defined as the primary cause of acid mine drainage by Tabelin et al. (2017c) and chalcopyrite whose oxidation end up with acid generation reach to significant values (55% and 58%). Before performing column tests, chemical analyses were conducted to identify the amount of major and, especially, trace elements of initial samples by X-ray fluorescence (XRF) and ICP-OES at the laboratories of MTA. The results of the analyses are listed in Table 1.

Table 1 Initial chemical compositions of samples based on XRF and ICP-OES analyses
Full size table

Kinetic column tests

AMD generation is a time-dependent geochemical reaction and various factors such as availability moisture, oxygen, Thiobacillus ferrooxidans and other oxidation catalyzing bacteria, pH and temperature (Robertson et al. 1992) are key controlling parameters for this process. Many of these conditions cannot be provided by performing static tests utilized to obtain preliminary geochemical data. To overcome with limitations related to this tests, many different approaches have been developed and recommended by considering kinetic principles of AMD generation (Smart et al. 2002; Lapakko et al. 2006; Andrina et al. 2006; Sapsford et al. 2009; González-Sandoval et al. 2009; Smith et al. 2009; Lewis and Sjöstrom 2010; Tatsuhara et al. 2012; Parbhakar-Fox et al. 2013; Naka et al. 2016). Similarly, Tabelin et al. (2017a, b) performed laboratory-controlled column tests fulfilled with batch leaching to specify the leaching and transportation characteristics of several hazardous trace elements in rocks extracted during tunnel construction. In addition to laboratory-based tests, Tamoto et al. (2015) utilized the in situ column experimental setup to evaluate short- and long-term release mechanisms of arsenic, selenium and boron in sedimentary rocks collected from a tunnel project in Japan. Kinetic tests are usually preferred for evaluation time-consuming progressive AMD formation at both laboratory and field conditions. Although Lawrence and Marchant (1991) and Morin and Hutt (1997) emphasized the priority of the humidity cell and column techniques among other kinetic tests for understanding and describing AMD susceptibility of mining field, according to Bradham and Caruccio (1991), the column test can be considered as a more reliable approach to simulate acid generation at the field.

Previously, Hood and Oertel (1984), Murr et al. (1977), Strömberg and Banwart (1999) and Benzaazoua et al. (2004) designed leaching columns in different dimensions for investigation AMD generation in their research site. According to Hornberger and Brady (1998), the kinetic test apparatus should be simple and practical with respect to its size, shape and structure. However, the diameters of previously designed columns range from 3.30 cm (Hood and Oertel 1984) to 3.08 m (Murr et al. 1977). The inside diameter and height of columns used by Benzaazoua et al. (2004) were 0.106 m. and 0.7 m, respectively. Strömberg and Banwart (1999) carried out kinetic tests by using six large columns (0.8 × 2 m). The mass of samples in kinetic column tests change between 10 and 3000 kg (EPA 2003). Furthermore, Morin and Hutt (1997) investigated the AMD formation of samples having a mass of 20,000 kg.

The particle size of the sample is another controlling factor on the rate of AMD generation due to the positive relationship between the rate of acid generation and specific surface area of particles. Robertson and Broughton (1992) specified that samples having a particle size of − 2 mm should be used in humidity cells. According to Bradham and Caruccio (1990), column tests of well-sorted tailing materials with a particle size of > 5 mm accurately represent field test conditions. It is well known from the correlation between particle size and the rate of acid formation that an increase in the percentage of finer particle contents causes an acceleration in the rate of AMD in conjunction with time and cumulative mass release of contaminants (Kalyoncu Erguler et al. 2014). Thus, all particle sizes such as physical fragmentation of mining processes remained inside quarry and the particle size distribution of spoil and tailing materials should be noted for a representative kinetic column test.

By considering abovementioned statements, samples with relatively coarse and fine particle sizes were placed in twenty-four columns set up in different dimensions to investigate the effect of particle size and column dimensions on the rate of AMD generation (Fig. 3). The loss of fine-grain samples should be prevented during kinetic column tests, and an acrylic wool was placed at the bottom of each column to keep finer particles inside columns. The amount of distillated water required for initiation of kinetic column tests was defined based on 50-year meteorological records. It was found that the amount of water changes between 89 and 1090 ml depending on dimensions of columns. Two-litre capacity reservoirs were fixed to the top of platforms for supplying water to the columns without using a pumping system. Kinetic column tests were performed by dropping water inside columns. Depending on particle and column sizes, it was determined that the flow rates change between 1.0 and 25.2 ml/min without developing any channeling and preferential routes. Time-dependent settlements were measured for some columns particularly filled with fine-grained samples; however, the effect of settlements was not recorded for hydraulic conductivities of columns.

Fig. 3.
figure 3

Columns designed and used in this study

Full size image

After the water had flown through the column, contaminated solutions were collected in glass jars and then analysed for monitoring variations in pH and chemical composition at specific time intervals. The chemical analyses were performed by conducting ICP-OES and ion chromatography analysis techniques at the laboratories of MTA on leachate recovered periodically from columns to determine the geochemical behaviour of fine and coarse samples. For this purpose, the concentrations of the K+, Na+, Ca2+, Mg2+, NH4+, Li+, B3+, SiO2, Fe(2+,3+), Mn2+, Pb2+, Zn2+, Cd2+, Cu2+, Ni2+, Cr2+, Co2+, Mo2+, Sr2+, Ba2+, CO32−, HCO3−, Cl−, SO42−, F−, NO2−, NO3−, Br− and PO43− were recorded.

Results and discussion

The geochemical behaviour, the evolution of pH measurements and concentrations of several parameters during kinetic column tests and related simple regression analyses by utilizing the samples of this study were previously discussed by Kalyoncu Erguler et al. (2014). They stated that “lag time”, which is the time needed to monitor the first acid formation and generally predicted from the curve of pH versus time, can last for weeks, months or years depending on weathering rate of sulphide minerals and neutralization capacity of waste material. Based on F tests, Kalyoncu Erguler et al. (2014) concluded that lag time can be estimated from column dimensions by utilizing simple regression methods and they found statistically significant correlations with a good performance for samples having two different particle size distributions (Table 2). Thus, it can be said that increasing the column’s parameters decrease the rate of AMD formation. In this study, kinetic evaluations of column tests based on cumulative variations in pH, the concentration of Ca2++Mg2++Mn2+ and SO42− are mainly given in the below sections.

Table 2 Correlations between lag time and column dimensions
Full size table

Oxidation-neutralization–based approaches

Previous studies such as Scharer et al. (1991), Robertson et al. (1992), Cruz et al. (2001), Benzaazoua et al. (2004) and García et al. (2005) specified that while the total concentration of Ca2+, Mg2+ and Mn2+ represents the dissolution of carbonate minerals, and their presence indicated neutralization, the concentration of SO42− is considered as the main sulphide oxidation product. According to Benzaazoua et al. (2004), the correlation between the mass release rate of Ca+Mg+Mn and SO42− may be accepted a good indicator for evaluation AMD potential of any samples subjected to the column test. They presented a new approach in which the cumulative mass of sulphates was plotted versus the cumulative amount of Ca, Mg and Mn for interpretation of the AMD phenomenon of the Joutel mine tailings (Quebec, Canada). Benzaazoua et al. (2004) found linear relationships, which were called the oxidation-neutralization curves, between these two parameters, and used these curves to indicate the geochemically based assessment of the acid generation to neutralize element-produced ratios throughout the column tests. As seen in Fig. 4, two different linear correlations were obtained between cumulative concentrations of Ca2++Mg2++Mn2+ and SO42− for all columns. Rapid and accelerated decline in pH triggered an increase in the amount of neutralization products (Ca2++Mg2++Mn2+). However, such further increases in the cumulative concentrations of the Ca2++Mg2++Mn2+ were not enough to neutralize the acid already generated throughout column tests.

Fig. 4.
figure 4

Relationship between cumulative concentration of Ca2++Mg2++Mn2+ and SO42− for column 2 (a) and column 7 (b)

Full size image

In order to further understand the process and rate of AMD generation, the relationship between cumulative concentrations of Ca2++Mg2++Mn2+ and SO42− was investigated for the first part of these bilinear correlations. Strong correlations were found between these parameters, particularly for coarse samples with very high R2 values ranging between 0.998 and 0.999 (Table 3). All correlations given in Table 3 are linear function and can be represented by the following single empirical model with a changeable slope value depending on the column dimension.

Table 3 Correlations between cumulative Ca2++Mg2++Mn2+ (dependent variable, y) and cumulative SO42− (independent variable, x)
Full size table
$$ y={a}_1x $$
(1)

where y is the cumulative concentration of Ca2++Mg2++Mn2+, x is the cumulative concentration of SO42− and a1 is a dimensionless constant (slope). In this study, slope values range between 0.3398 and 0.4909 for coarse and 0.36 and 1.7128 for fine samples. The slope values found by Benzaazoua et al. (2004) vary between 0.33 and 0.41. As seen in Table 3, slope values change with the length, diameter and volume of column dimensions as well as the particle size of tested samples. Considering these results, charts were developed to estimate the constant of this general equation from the properties of a column filled by coarse samples subjected to kinetic column tests (Fig. 5).

Fig. 5.
figure 5

Correlations between column dimensions and slope of Eq. (1) (a1) for coarse samples

Full size image

Simple regression for correlations between time and the pH, Ca2++Mg2++Mn2+ and SO42− concentrations

Time is another important parameter for understanding the kinetic behaviour of AMD processes. In previous studies (e.g. Strömberg and Banwart 1999; Benzaazoua et al. 2004; García et al. 2005), investigators generally evaluated the weathering kinetics of samples by plotting changes in key indicator parameters (such as pH values, the concentration of Ca2+, Mg2+, Mn2+, Mn2+, SO42−) as a function of time, except for detailed research on time-dependent variations in cumulative oxidation and neutralization products. However, the evaluation of data should also include the changes in these parameters over time. Thus, in addition to the above-given relationships between the cumulative concentration of Ca2++Mg2++Mn2+ and SO42−, it is thought that evaluations including the time-dependent variations in these cumulative values would be useful for a better understanding of the process of acid generation throughout column tests. Furthermore, the change in cumulative pH values with time and its effect on the other concentration of parameters have not been evaluated so far.

When instant values are considered, the concentration of Ca2++Mg2++Mn2+, SO42− and pH measurements do not directly correlate with time (Fig. 6 a and c). Therefore, the time factor cannot be involved properly in the kinetic evaluation of the AMD process. However, the cumulative values of these parameters show significant correlations with time. One of the typical outcomes of these correlations is illustrated in Figs. 6 b and d. As seen from this figure, cumulative pH shows a rapid decrease at the beginning and end of the tests, whereas the cumulative values of Ca2++Mg2++Mn2+ and SO42− continuously increase with time.

Fig. 6.
figure 6

The variation of instant concentration of Ca2++Mg2++Mn2+, SO42− and pH measurements with time for column 1 (a) and column 7 (c), the cumulative variation of pH and concentration of Ca2++Mg2++Mn2+ and SO42− for column 1 (b) and column 7 (d)

Full size image

As stated by Robertson et al. (1992), the dominant weathering processes at the beginning stage and the end of the kinetic column test are totally different from each other. While the chemical weathering process is more effective at the beginning of the process (about pH > 5.0), with increasing the effect of chemical processes thus after AMD formation by the help of dropping pH values to the below of 4.0, biological oxidation becomes more significant for the last 1 or 2 weeks of kinetic column tests by rapidly accelerating the AMD formation process. To find the effect of time on cumulative values of Ca2++Mg2++Mn2+ and SO42−, and cumulative pH measurements during chemical oxidation processes, further analyses were performed by taking these parameters as dependent and time as independent variables. The results of these analyses are given in Table 4.

Table 4 Simple regression analyses results based on cumulative values of Ca2++Mg2++Mn2+, SO42− and pH measurements as dependent and time as independent variables
Full size table

As seen in Table 4, statistically significant correlations were obtained between time and cumulative values of Ca2++Mg2++Mn2+, SO42− and pH. Data obtained after the pH drop were not used in these analyses. While about 13–20 data pairs were used to develop empirical equations for coarse samples (Table 4), due to the short duration of the experiments, a lower number of measurements, between 4 and 9, were taken for fine samples. When empirical models given in this table are simplified, the general equations given below are obtained for prediction of time-dependent variations in cumulative values of Ca2++Mg2++Mn2+, SO42− and pH:

$$ \mathrm{Cumulative}\ {{\mathrm{SO}}_4}^{2-}={a}_2x $$
(2)
$$ \mathrm{Cumulative}\ {\mathrm{Ca}}^{2+}+{\mathrm{Mg}}^{2+}+{\mathrm{Mn}}^{2+}={a}_3x $$
(3)
$$ \mathrm{Cumulative}\ \mathrm{pH}={a}_4-b\ast \ln (x) $$
(4)

where x is time in week and a2, a3, a4 and b are dimensionless constants. The variations in the constants of the above-given equations are thought to depend on column dimensions. Therefore, to predict these constants for practical engineering application, statistically significant regression models were developed between the constants of the above equations and column dimensions based on simple regression analyses (Table 5). A typical graph of these regression models is illustrated in Fig. 7 as an example. These models might be important in the prediction of cumulative values of SO42− and Ca2++Mg2++Mn2+ for any desired length, diameter and volume of the column.

Table 5 Summary of the statistical analysis for prediction a2, a3, a4 and b from column dimensions
Full size table
Fig. 7.
figure 7

Prediction of the slope of Eq. (2) (a2) from length, diameter and volume of columns in which coarse samples used

Full size image

Multiple regression models for the kinetic process of acid generation

The concentration of dissolved metals and leachate quality is controlled by several parameters such as pH, time, oxygen and water availability, particle size distribution, the temperature at both natural field and laboratory conditions. Among these, particle size, temperature, pH and time are very important parameters for simulation acid generation models at laboratory conditions until decreasing of pH to about 4.0. It is known that biological oxidation processes start at such pH value and the rate of acid generation increase due to the dominant effect of oxidation catalysing bacteria. Therefore, it can be inferred that investigating the effect of only one parameter on changes in sulphate or main neutralization products (Ca2+, Mg2+, Mn2+, etc.) or the variation of other dissolved metals measured from leachate during kinetic column test over time would not be enough for understanding the entire kinetic processes of acid generation. In this study, kinetic column tests were performed by using two different particle sizes, the temperature measured during kinetic column tests changed between 18 and 27 °C, with a mean value of approximately 22 °C. Although time-dependent changes in cumulative values of Ca2++Mg2++Mn2+, SO42− and pH, the effect of column dimensions on these values was comprehensively investigated by simple regression models in the previous section; it is thought that simple regression models may not be enough for understanding the entire kinetic process of the column test because time is not the only factor controlling leachate quality but also the pH level is one of another important effective parameter on the amount and rate of oxidation processes throughout column tests.

Therefore, in the light of above statements, it is considered that the pH values, particularly cumulative pH, may be useful for the assessment of time-dependent changes in main oxidation (cumulative SO42−) and neutralization (cumulative Ca2++Mg2++Mn2+) products of kinetic column tests. For this purpose, to incorporate combine effect of time and pH, multiple regression analyses were performed by taking cumulative SO42− and cumulative Ca2++Mg2++Mn2+ as dependent variables (y) and cumulative pH and time, which are two important key parameters controlling the rate of acid generation at the laboratory condition, as independent variables (x). The statistically significant empirical approaches obtained as the results of these multiple regression analyses of all columns are given in Table 6. The R2 values for regression models of columns in which samples having coarse particle sizes were used range between 0.90 and 0.99. The R2 values reveal that these predictive empirical models explain 90–99% of the variability in cumulative SO42− and cumulative Ca2++Mg2++Mn2+ released from coarse samples. Except for models suggested for defining variation in cumulative Ca2++Mg2++Mn2+ of columns 22 and 24, again all empirical equations for fine samples have also high R2 values. In addition to these statistical evaluations, further analyses on the comparison between the predicted and experimentally measured cumulative SO42− and cumulative Ca2++Mg2++Mn2+ values were carried out. As obvious from Fig. 8, data are not scattered, and the margins of error, which are represented as the distance between data and 1:1 line, are generally very small. Hence, the empirical relationship given in Table 6 can be used to estimate cumulative SO42− and cumulative Ca2++Mg2++Mn2+ values by using time and pH parameters measured from water percolated through to the samples during the kinetic column test.

Table 6 Predictive empirical models based on multiple regression analyses by selected cumulative SO42− and Ca2++Mg2++Mn2+ as dependent variables (y) and cumulative pH and time as independent variables (x)
Full size table
Fig. 8.
figure 8

Typical examples for experimentally measured cumulative concentration of SO42− and Ca2++Mg2++Mn2+ versus their predicted values for columns 1 and 2

Full size image

Conclusions

In this study, a comprehensive research program was conducted on sulphide-bearing rocks for understanding and finding dominant factors controlling the kinetic behaviour of the acid generation during kinetic column tests. The main conclusions obtained from present study are summarized below:

After AMD generation, increases in the amount of cumulative main neutralization products as a result of dissolution of remaining calcium-, magnesium- and manganese-bearing minerals under very low pH values (< 4.0) are not enough to neutralize the generated acid during the kinetic column tests.

Unlike previous studies, it was found that the correlations between cumulative concentrations of Ca2++Mg2++Mn2+ and SO42−, for all designed columns were bilinear in the case of including data after the AMD formation. The first linear relationships may be represented by a single linear function (y = ax) with a changeable slope value depending on the column dimension.

Although the pH measurement is a key parameter to monitor the rate of the AMD generation, due to simultaneously processes of oxidation and neutralization, the variation in pH during kinetic column tests is the naturally unsteady and instant value of this parameter cannot be effectively used for evaluation of the kinetic of acid generation or this instant pH values do not correlate with other parameters such as time or the concentration of the dissolved element or compounds. Such kind of result indicates the importance of the incorporation of the cumulative pH for AMD assessments.

Contrary to the fluctuation in the instantaneous values of Ca2++Mg2++Mn2+, SO42− and pH measurements over time, the cumulative values of these parameters show significant correlations with time during the column tests. Statistically significant relationships were obtained between time and cumulative values of Ca2++Mg2++Mn2+, SO42− and pH, and furthermore, these empirical models were simplified to produce three general equations (Eqs. (2), (3) and (4)) to be used for prediction of time-dependent variations in cumulative values of Ca2++Mg2++Mn2+, SO42− and pH. In addition, to predict the constants (a2, a3, a4 and b) of these empirical models and determine cumulative values of SO42− and Ca2++Mg2++Mn2+ for any desired length, diameter and volume of the column, some practical approaches were suggested based on simple regression analyses.

Predictive empirical relationships were obtained based on multiple regression analyses by taking cumulative SO42− and Ca2++Mg2++Mn2+ values as dependent variables and cumulative pH and time as independent variables to incorporate combine effect of time and pH conditions, which are thought to be key parameters controlling the rate of acid generation.