Environ Earth Sci DOI 10.1007/s12665-014-3949-3
ORIGINAL ARTICLE
A model for prioritizing sites and reclamation methods at abandoned mines Owen E. Kubit • Christopher J. Pluhar Jerome V. De Graff
•
Received: 29 October 2013 / Accepted: 9 December 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Abandoned mines present numerous safety and environmental problems due to altered topography and poor management of mine waste. Few guidelines are available for selecting reclamation methods to address these problems, and many existing decision models lack transparency, leave out important parameters and reclamation methods, and/or lack model calibration. Consequently, a decision model was developed that includes: (1) a mine hazard index for prioritizing sites for reclamation, (2) a reclamation method screening table for narrowing viable reclamation methods, and (3) a reclamation method ranking matrix for ranking the applicability of reclamation methods at a site. These three processes form the abandoned mine decision model, optimized for topographic reconstruction and waste disposal at abandoned mines. The hazard index quantifies geologic and hydrologic hazards using measurable parameters, sub-parameters, and a range Electronic supplementary material The online version of this article (doi:10.1007/s12665-014-3949-3) contains supplementary material, which is available to authorized users. O. E. Kubit (&) Department of Earth and Environmental Sciences, California State University at Fresno, 2505 Alluvial Ave, Clovis, CA 93611, USA e-mail:
[email protected] C. J. Pluhar Department of Earth and Environmental Sciences, California State University at Fresno, 2576 East San Ramon Ave. M/S ST24, Fresno, CA 93740, USA e-mail:
[email protected] J. V. De Graff Department of Earth and Environmental Sciences, California State University at Fresno, 2576 East San Ramon Ave. M/S ST24, Fresno, CA 93740, USA e-mail:
[email protected]
of sub-parameter conditions. Main and sub-parameter weighting factors were determined using the analytic hierarchy process and Delphi method. Reconciling differences between initial parameter weighting factors for the two methods resulted in an improved decision-making technique. The decision model was calibrated with 25 abandoned metal mines in the western USA. Mine hazard index thresholds were determined for mines having low, medium, and high priority. The screening table and ranking matrix were effective at narrowing the number of viable reclamation methods. Further validation of the decision model was evident by the implemented reclamation method being among the four highest scoring alternatives 80 % of the time. This model provides a quantitative and transparent process that overcomes the deficiencies found in many existing mine reclamation decision models. Keywords Mine reclamation Abandoned mine Decision model Topographic reconstruction Analytic hierarchy process Delphi method
Introduction Abandoned mines constitute a legacy of safety and environmental problems throughout the USA. Estimates for the number of abandoned mines in the USA vary from 200,000 (USEPA 2000) to as high as 557,650 (Struhsacker and Todd 1998). The United States Government Accountability Office (2011) states that the 12 western states and Alaska contain over 160,000 abandoned hardrock mines. Many of these mines were developed before the advent of modern environmental laws in the 1970s and now present numerous safety and environmental hazards.
123
Environ Earth Sci
Common problems at abandoned mines include slope instability, dangerous highwalls (artificial cliffs), soil and water contamination, poor esthetics, erosion damaging the mine site and providing unwanted sediment to local watersheds, and hydrologic impacts such as modifications to natural waterways. All of these problems are related to topographic alteration of the land during the mining process and vary with the type of mining method. Surface mining can result in large open pits and vast quantities of waste material. Subsurface mining typically causes less surface disturbance, but often results in moderate to large waste rock piles that can cause many of the same problems found at surface mines. Abandoned mines were typically developed without any consideration for future reclamation, often resulting in scattered distribution of mine waste and more complex reclamation challenges. Topographic reconstruction commonly comprises 90 % of total reclamation costs at mines due to the high cost of earth moving (Black and Toy 2000; Harwood and Thames 1988). Therefore, the viability of a reclamation project is enhanced by using the most effective and economical topographic reconstruction methods. However, a review of hundreds of articles and books revealed few guidelines or recommendations for selecting a reclamation method after the assessment stage is completed. Another significant problem for geologists and other reclamation specialists is ranking and prioritizing mines to be reclaimed. Limited funding and personnel necessitate that mines will be reclaimed gradually over many decades, and studies have shown that many abandoned mines pose few or no problems (Struhsacker and Todd 1998). Reclamation geologists need methods to prioritize abandoned mines to ensure that they focus on those presenting the greatest hazards to the public and environment. To assist with these challenges, this study developed a decision model that prioritizes abandoned or inactive mines based on hazards and recommends the most appropriate reclamation alternatives. Existing decision models were found to have significant limitations such as only prioritizing mine sites or considering a limited numbers of reclamation methods. The abandoned mine reclamation model proposed here addresses methods to reclaim land disturbance and mine waste piles (e.g., tailings, overburden piles, processing waste). These materials are generally reclaimed through ‘topographic reconstruction’ methods that involve regrading and recontouring. Some specific topographic reconstruction methods include backfilling pits, reconstructing displaced soil or rock to the original contour, and modifying highwalls. The model uses an integrated approach to: (1) identify higher priority sites, (2) screen applicable reclamation methods, and (3) rank those reclamation methods for the particular site. Case studies from a
123
diverse range of abandoned mines were used to validate and calibrate the model. Decision-making tools Assessing the physical and environmental hazards at abandoned mines is difficult without substantial data, such as a detailed slope stability analysis, or a human health risk assessment. These analyses are typically not available during preliminary investigations, while many basic data such as contaminant concentrations or highwall height are readily measurable. Therefore, in early stages, assessments are often performed with subjective judgment of the relative importance of such basic parameters to the overall hazard. The Delphi method and analytic hierarchy process (AHP) are two common techniques that can help reduce subjectivity in decision making. These methods can be used to calibrate decision models and score and rank alternatives. This calibration is done by assigning weighting factors for important parameters, ensuring that the measured values of the most important parameters drive the decision making and insignificant parameters carry little weight. The Delphi method is a systematic decision-making technique relying on a panel of experts (Hsu and Sandford 2007). The method captures the expertise embodied in individuals and distills it into an algorithm used in decision making. Experts first answer a questionnaire about the relative importance of decision model parameters. They receive an anonymous summary of the responses and are encouraged to revise their answers after reviewing the range of responses. This process can occur for one or more rounds. During this process the range of answers typically decreases, and it is believed that the group converges toward the best answer (Hsu and Sandford 2007). The Delphi technique can help several decision makers with conflicting opinions or different backgrounds to arrive at consensus. The AHP is a common technique that uses pair-wise comparisons between parameters to assign parameter weighting factors in a decision model (Saaty and Vargas 1991). The user determines parameter weighting factors based on their relative importance: the ‘‘intensity of importance’’ during each pairwise comparison (Table 1) (Saaty and Vargas 1991). Table 1 also shows a revised AHP process developed by the authors, which is discussed later in ‘‘Mine hazard index’’. Online Resource 1 provides an example of how the AHP can be used to prioritize several reclamation-related parameters. The Delphi method and AHP were chosen among numerous other decision-making methods because they are simple, easy to use, easy to understand, and are widely accepted. They are also generally considered the classical
Environ Earth Sci Table 1 Scoring system for conventional and revised analytical hierarchy process
Intensity of importance
Intensity of importance
Weighting factors for pair of parameters (%)
Conventional
Conventional
Revised
Revised
Equal importance
1
1
50/50
50/50
Weak or slight importance
2
1.1
67/33
52/48
Minor importance
–
1.2
–
55/45
Moderate importance
3
1.5
75/25
60/40
Strong importance
5
3.0
83/17
75/25
Very strong importance
7
5.7
87/13
85/15
Extreme importance
9
9.0
90/10
90/10
approaches to subjective decision making and were considered ideal for this application. The authors also wanted to create a practical tool whose origins could be easily understood by field engineers and geologists. Miller’s ‘‘law’’, established through a series of psychological experiments, states that an individual cannot effectively compare more than 7 ± 2 attributes (Miller 1956). This ‘‘law’’ of cognitive science indicates that a system with more than seven attributes should be simplified, or separated into a hierarchy of parameters and subparameters, with no more than seven attributes in each level. Thus, before using the Delphi method or AHP, the number of attributes compared in each set needs to be reduced to about seven or less. Other reclamation decision models Several assessment indices and decision models have been previously developed for abandoned mines, some using the AHP or Delphi method. A literature review concluded that no existing model succeeded in satisfying all of the elements identified as important to guiding mine reclamation workers: (1) prioritizing sites for reclamation, (2) guiding decision making toward optimal reclamation methods for a given site, (3) completeness of input parameters, (4) transparency in how the decision model is constructed, and (5) model calibration/validation. This provides strong motivation for developing a new decision model. However, some exiting models did successfully satisfy a few of these elements and were useful examples in developing our model. Several studies incompletely address the creation of a comprehensive reclamation decision model. Bezuidenhout et al. (2009) developed a numerical index for assessing hazards at asbestos mines. This index includes a logical list of parameters, but the basis for selecting weighting factors is not discussed. Albert et al. (1991) developed a project prioritization model for abandoned mines. Some of the parameters used in this model are overly general (e.g., ‘‘degree of hazard’’) and many are non-technical (e.g., ‘‘local support’’). Assessment models have also been
developed for specific mine features, including highwalls (Klimstra et al. 1983) and waste piles (Dias et al. 2011). These models are useful when applied to specific hazards, but do not provide a comprehensive hazard assessment at a mine site. The Colorado Division of Minerals and Geology (2002), Brown and Sidle (1992), and Interstate Technology & Regulatory Council (2010) each developed decision trees for selecting mine reclamation methods. These decision trees are useful, but each is incomplete in the mine site hazards and number of reclamation methods considered. Robertson and Shaw (1998) describe the methodology for creating a mine reclamation hazard index, but do not provide an actual model. The United States Environmental Protection Agency (USEPA) (1992) developed a hazard ranking system (HRS) to prioritize Superfund sites, which can include abandoned mines where environmental contaminants are present. However, the HRS does not consider some features typically relevant to abandoned mines, such as highwalls or hydrologic impacts. The preliminary appraisal and ranking system (PAR), developed by the California Office of Mine Reclamation (2000), is one of the most useful hazard ranking systems developed specifically for abandoned mines. However, like all of the other assessment models described above, PAR does not include a component to help select a reclamation method. Several studies have approached mine reclamation decision making in a more quantitative fashion. Albert et al. (1991) developed a mine reclamation prioritization model using five different decision-making techniques, including the Delphi method. Hu and Linlin (2009) used the Delphi method to evaluate post-reclamation land use. Bascetin (2007), Soltanmohammadi et al. (2010), and Bandopadhyay and Chattopadhyay (1986) each used the AHP to develop post-reclamation land use models. These five papers illustrate the usefulness and popularity of the Delphi method and AHP in evaluating mine reclamation projects. Soltanmohammadi et al. (2008) and (2009) provide useful decision models with a transparent and rational methodology that considered a large number of relevant attributes (50) and reclamation options (14). However,
123
Environ Earth Sci
these models serve a different purpose than the abandoned mine decision model; they are used to select post-mining land use, as opposed to ranking mine hazards and selecting reclamation methods. While existing mine reclamation models demonstrate certain strengths, many also display deficiencies that limit their breadth, accuracy, and/or usefulness including: (1) lack of transparent and/or rational basis for selecting weighting factors, (2) use of overly general parameters or absence of some important parameters, (3) lack of model calibration using real data or case studies, and (4) limited number of reclamation methods considered in decision models. Geologists working to reclaim abandoned mines would benefit from improved tools for selecting reclamation methods. The decision model developed for this study addresses the aforementioned deficiencies and provides a single model for both prioritizing mines for reclamation by assessing a mine’s conditions and selecting appropriate reclamation alternatives.
Abandoned mine decision model The abandoned mine reclamation model focuses on topographic reconstruction, which is defined here as restoring, modifying, or stabilizing surface topography and/or removing mine waste materials to mitigate safety and environmental hazards. The general goals of topographic reconstruction include: 1. 2. 3. 4. 5.
Recontour topography to satisfy legal requirements, stabilize slopes, and prevent erosion. Eliminate hazards from highwalls. Isolate or remove contamination. Reconstruct or mimic natural hydrologic features. Provide a stable surface for revegetation and future land uses.
In many ways, optimization of topographic reconstruction and mine waste handling is critical to mine reclamation, because earthmoving is often the most costly portion of a project. Dutta et al. (2005), Struhsacker and Todd (1998) and the U.S. Bureau of Land Management (1992) provide the most comprehensive discussions on topographic reconstruction. However, no single source discusses all available methods. Figure 1 summarizes commonly used reclamation methods from numerous references, not including outdated methods (e.g., marine disposal) or methods rarely used to reclaim abandoned mines (e.g., new tailings dams). At present, this model does not address hazards due to shafts and adits, abandoned buildings, and features not specifically related to mines (trash, etc.)
123
The abandoned mine decision model (model) comprises three consecutive modeling tools, including: (1) mine hazard index, (2) reclamation method screening table, and (3) reclamation method ranking matrix. Several approaches were considered for developing an abandoned mine decision model, including decision trees and indexes. Decision trees were found to have practical limitations and could only evaluate a limited number of variables before becoming too physically large and complex to be practical. Numeric indexes, on the other hand, have been successfully used for several mine assessment models (Bezuidenhout et al. 2009; Albert et al. 1991). The selected approach involves a three-step process using a numeric hazard index, screening table and ranking matrix. This approach yields a model which: (1) determines a mine hazard score that prioritizes which mine is most in need of reclamation; (2) identifies applicable reclamation methods for the selected mine; and (3) ranks the viable reclamation methods based on geologic, hydrologic, and related factors. These three processes together form the abandoned mine decision model. The overall approach for developing the decision model is shown in Fig. 2. Mine hazard index The ‘mine hazard index’ quantifies the severity of environmental, human health, and public safety hazards at abandoned mines. The mine hazard index score is used to compare and rank different sites to prioritize mines for reclamation and is ultimately used as part of a ranking matrix to assist in reclamation method selection. The mine hazard index was developed by selecting parameters that captured the full range of mine hazards, identifying a range of potential conditions for each parameter and calibrating the model by determining weighting factors for each parameter using analytical methods. The mine hazard index is then computed by feeding measureable parameters into the model and receiving a score that indicates whether reclamation is needed or no action is acceptable. The score can then be used to rank sites relative to one another to assign priority. The list of relevant mine site parameters, and a range of conditions for those parameters, is based on guidance from: (1) mine reclamation literature; (2) laws and regulations; (3) other mine reclamation models; (4) 25 case studies; and (5) site visits to several abandoned mines. Initially, over 140 primary parameters were identified, but these were reduced to 5 main parameters and 18 sub-parameters, collectively called parameters. Next, a range of possible conditions was determined for each of the 18 sub-parameters. Figure 3 shows an example of a parameter with its sub-parameters and conditions.
Environ Earth Sci
Fig. 1 Reclamation methods
The main parameter and sub-parameter are equivalent to a category and sub-category of mine hazards, while the sub-parameter conditions span the range of conditions that could exist at a mine. The sub-parameter condition may be discrete data as shown in Fig. 3. In cases such as highwall height the sub-parameter condition is not a discrete value, but rather a numerical value equal to the height of the highwall. Consequently, the primary criteria for selecting parameters for the model included how well they describe or represent hazard severity and how easily they can be quantified during preliminary investigations. The authors identified a range of potential conditions for each mine hazard index sub-parameter. For example, surface waters typically exhibit pH values from 1 to 9, while site accessibility could be judged as poor, fair, good, or excellent. Important threshold values were determined for some parameters to help identify boundary values that would trigger action if exceeded. For example, highwalls and waste piles greater than 7.6 m (25 feet) require reclamation in all cases (States of Oregon and Washington 1997; U.S. Bureau of Land Management 1992). The authors determined ratios for each possible condition representing its relative impact on a specific mine hazard. A
ratio of 1.0 represented the worst possible condition, and a ratio of 0.0 represented no hazard. The experts consulted elsewhere in this study were not asked to assign ratios for each condition due to the large number of ratios listed for all 18 sub-parameters. The ratios were determined using one of three methods: (1) analytic hierarchy process; (2) an equation based on measurable parameters; or (3) directly assigning ratios when the answers are basic and descriptive, such as poor (1.0), fair (0.75), good (0.5), and excellent (0.0). To reduce multiple mine site parameter values into a single hazard index score, it is necessary to establish the relative importance of each parameter in contributing to the overall site hazard. This was accomplished by determining weighting factors (factors) for each parameter and subparameter, establishing the range of sub-parameter conditions, and establishing numerical values (a ‘‘ratio’’ 0.0–1.0) for each sub-parameter condition representing a condition’s severity. The weighting factors (factors) for the main and sub-parameters were determined using the Delphi method and the AHP. Use of these two independent methods enabled comparison and calibration. Fifteen geology and mine reclamation experts (experts) participated in the
123
Environ Earth Sci Fig. 2 Research methodology flowchart
Fig. 3 Main parameter, subparameter, and sub-parameter condition relationship
Delphi method parameter weighting, and the authors used the AHP. The experts were not asked to use the AHP. The group of experts was only asked to use the Delphi method because of its simplicity, and concern that some may not
123
participate if they were also asked to learn the AHP. Achieving a high rate of participation was considered important so that a sufficient number of responses were available to determine average weighting factors.
Environ Earth Sci Table 2 Parameter weighting factors
Sub-parametera
Delphi methodb Phase II
Conventional
Revised 0.14
00.16 – 0.06
0.17 – 0.06
0.09
0.40 ± 0.16
0.39 ± 0.15
0.25
0.40
Connection of erosion to water bodies
0.60 ± 0.16
0.61 ± 0.15
0.75
0.60
0.15 – 0.09
0.14 – 0.07
0.09
0.14
Slope stability Slope
0.23 ± 0.10
0.22 ± 0.08
0.18
0.20
Height
0.20 ± 0.07
0.20 ± 0.06
0.13
0.19
Condition of slope
0.28 ± 0.12
0.29 ± 0.12
0.40
0.23
Geologic composition
0.13 ± 0.08
0.13 ± 0.07
0.13
0.18
Accessibility Soil contamination COC screening criteria exceeded Mobility Volume Water contamination Water pH
The sum of the five main parameters equals 1.0, and the sum of each set of subparameters equals 1.0 b The Delphi Method values are shown with their standard deviation
Phase I
Erodibility
Erosion
a
Analytic hierarchy process
0.16 ± 0.09
0.16 ± 0.09
0.16
0.20
0.23 – 0.08
0.24 – 0.08
0.24
0.22
0.44 ± 0.18 0.32 ± 0.13
0.40 ± 0.12 0.36 ± 0.10
0.44 0.39
0.38 0.34
0.24 ± 0.08
0.24 ± 0.08
0.17
0.28
0.32 – 0.09
0.31 – 0.06
0.49
0.36
0.16 ± 0.07
0.16 ± 0.05
0.10
0.16
MCLs exceeded in water
0.25 ± 0.08
0.24 ± 0.09
0.33
0.26
Hydrologic connection to waterways
0.29 ± 0.10
0.29 ± 0.05
0.29
0.22
Water uses
0.20 ± 0.09
0.21 ± 0.09
0.20
0.22
Annual precipitation
0.10 ± 0.05
0.10 ± 0.05
0.08
0.14
0.14 – 0.06
0.14 – 0.05
0.09
0.14
Hydrologic Impacts Impacts to drainage patterns
0.30 ± 0.07
0.29 ± 0.06
0.18
0.21
Impacts to streambed/floodplain
0.48 ± 0.10
0.50 ± 0.09
0.71
0.60
Impacts to infiltration
0.22 ± 0.06
0.21 ± 0.06
0.11
0.19
Delphi method Expert participants were asked to assign weighting factors to the mine hazard index parameters and sub-parameters based on their experience in prior mine reclamation activities. An anonymous summary of the responses from all participants was then sent to each expert and they were encouraged to revise their results if their opinion had changed based on the other responses and further reflection. At the time of the surveys, the experts were employed in industry (3), academia (3), and state or federal agencies (9). Eleven participants were mine reclamation experts, and four were experts in related fields, such as environmental or engineering geology. Hsu and Sandford (2007) stated that a sample population of 10–15 is sufficient for the Delphi method if their backgrounds are homogenous. Analytic hierarchy process The authors determined weighting factors for the main parameters and sub-parameters using the analytical hierarchy process. The AHP factors were compared to the Delphi method factors to help calibrate the Delphi method results, and compare the answers obtained from two different methods. A revised AHP procedure was developed
to address some deficiencies revealed in the conventional AHP (Saaty and Vargas 1991) (see Table 1). Table 2 lists the main parameters and sub-parameters selected for the mine hazard index, and their weighting factors (factors) determined with two phases of the Delphi method and the conventional and a revised AHP. The weighting factors are used in the mine hazard index to simulate the judgment of a panel of experts in determining the relative hazard posed by a given mine site. The main parameters selected for the model represent the most common and significant surface hazards found at abandoned mines. The list of main parameters is similar to lists found in Struhsacker and Todd (1998) and Albert et al. (1991). The sub-parameters represent the most important variables for each main parameter that can also be easily quantified during preliminary studies. Subsurface mine hazards, such as portals, adits, and land subsidence, were not included in the main parameters because they require different types of analysis and reclamation methods, and probably need to be evaluated in a separate model. Other hazards, such as dilapidated buildings, were not included since the model focuses on geologic and hydrologic hazards. The number of parameters in each set was kept to less than seven to comply with Miller’s law.
123
Environ Earth Sci
Delphi method and analytic hierarchy process A questionnaire was sent to 18 mine reclamation experts (experts) asking them to assign weighting factors to the parameters and sub-parameters. Fifteen experts responded. The results of the first round were summarized and returned to the experts for possible revisions. Thirteen of the 15 experts responded in the second round. It was assumed that the two non-responsive experts had no changes. The changes to the weighting factors in the second round were relatively minor. Eight volunteers (53 %) modified their weighting factors. The standard deviation for the responses reduced from 0.09 (first round) to 0.08 (second round), and each parameter changed by an average of just 3 %. This shows that the experts generally agreed on the relative importance of the parameters during the first questionnaire. The authors calculated weighting factors for the main and sub-parameters using the conventional AHP. Table 2 shows that the factors from this initial round of AHP did not compare well with the Delphi method. In particular, when parameters had small to moderate differences in importance, the conventional AHP calculated a relatively large difference in factors. As a result, a revised scoring system including categories for small differences in importance was developed for the AHP (see Table 1), and the weighting factors were recalculated. The new weighting factors agreed much better with the Delphi method results (see Table 2). The average difference between the factors converged from 0.07 (Delphi method versus conventional AHP) to 0.02 (Delphi method versus revised AHP). The revised AHP results therefore validated the results from Phase II of the Delphi method. The Delphi method results were ultimately used in developing the mine hazard index. Figure 4 graphically compares the main parameter weighting factors determined with the Delphi method and AHP. The graph shows close similarity between the two phases of the Delphi method, a moderate difference between the Delphi method and conventional AHP, and a minor difference between the Delphi method and revised AHP. The error bars in Fig. 4 represent standard deviation of the range of expert responses. Mine hazard index template . The mine hazard index was developed using the main parameters, sub-parameters, range of sub-parameter conditions, and all of their associated weighting factors. The mine hazard index template is shown in Fig. 5. The hazard index is used by determining a score for each sub-parameter based either on the given formula, or the scores provided in the mine hazard index template for a
123
specific hazard condition. The summation of these values yields the hazard index score, ranging from 0 to 1,000. Figure 5 includes a range of conditions and scores for each sub-parameter. For example, under ‘Hydrologic impacts’, the sub-parameter ‘Impacts to infiltration’ has four possible conditions: none, minor, moderate, and severe, which are assigned respective scores of 0, 10, 20, and 29. These scores are a product of the main parameter weighting factor, sub-parameter weighting factor, and subparameter condition ratio. (The methodology that was used to determine these scores, including the multiple weighting factors used to determine scores for each condition, is described in Online Resource 2.) The user selects the appropriate condition for a mine site and enters the corresponding score in the mine hazard index table. Most of the parameters used in the model lend to subjective ranking, such as the relative risk of one hazard over another. Subjective methods (Delphi method and AHP) were used to capture input and opinions from a diverse range of experts. However, some limited objective criteria from mine reclamation literature were used (see Table 3). These criteria were utilized to develop threshold values or formulas that were then used in estimating weighting factors. In summary, the mine hazard index scores are a product of the three separate types of values, the parameter weighting factor, sub-parameter weighting factor, and condition ratio. Evaluation at these multiple levels was necessary to comply with Miller’s ‘‘law’’. All were determined by the Delphi method or AHP calculations as outlined above and are now embodied within the model. When using the mine hazard index template, one only needs to select from the ‘Conditions and scores’ column provided in Fig. 5 and does not need to directly use any of the weighting factors or ratios. It should be stressed that this formulation is easily transferable into a computer expert system or mobile application for use by semi-skilled individuals. Reclamation screening and ranking The reclamation method screening table (Screening Table—Table 4) was developed to help reduce the number of reclamation methods that should be considered by the model for a given site. The screening table first separates mines into four categories based on the type and extent of surface disturbances: pits/highwalls, waste piles, both or neither. The screening table lists viable reclamation methods for each of the four categories, and then eliminates methods from consideration based on site conditions (e.g., groundwater depth) or the presence of hazards (e.g., soil contamination). For instance, if local soils are contaminated, then terracing is removed as an option, since terracing only addresses slope stability and erosion issues.
Environ Earth Sci
Fig. 4 Main parameter weighting factors
The screening table can reduce the number of viable reclamation methods from 18 to as few as 3, depending on the site characteristics. The reclamation method ranking matrix (ranking matrix) uses the mine hazard index values to score and rank potential reclamation methods. Table 5 is a template of the ranking matrix. The ranking matrix includes a table of five main mine hazard parameters by 18 reclamation methods. Values of 0.0, 0.5, or 1.0 were specified for each matrix cell representing the ability of a reclamation method to address a certain hazard. These are called ‘reclamation potential values’. A value of 0.0 indicates that the reclamation method is not relevant to that hazard, while 1.0 indicates that the reclamation method can fully mitigate the hazard. The values were determined using general knowledge of each reclamation method and the following assumptions: 1.
2. 3.
All regrading/recontouring methods, except terracing, can fully mitigate erosion when accompanied with revegetation. All regrading/recontouring methods can fully mitigate waste pile stability when accompanied with revegetation. Regrading/recontouring methods can partially mitigate soil and water contamination by burying or relocating contaminated soils.
4.
5.
6.
Contaminated soils are consolidated and recontoured to stable slopes prior to being confined in a repository, cap, or cover. Hydrologic impacts can only be fully mitigated with construction to original contour/pit backfill. Other forms of recontouring can only partially mitigate hydrologic impacts. The slope stability hazard is separated into waste piles and highwalls, because some reclamation methods address these two features differently. For instance, off-site landfill disposal would remove a waste pile and thus eliminate waste pile stability concerns, but cannot address highwall stability.
The matrix is used by cross-multiplying the mine hazard index scores (which are entered into the table by the user) by the reclamation potential values (0.0. 0.5, or 1.0), resulting in a total score for each reclamation method. Reclamation methods with the highest scores are considered the most promising. Abandoned mine case studies The decision model was validated with a diverse group of 25 abandoned mine case studies that varied in location,
123
Environ Earth Sci Fig. 5 Abandoned mine hazard index template
123
Environ Earth Sci Table 3 Selected sub-parameter criteria Sub-parameter
Criteria
Source
Erodibility
Over 168 metric tons/hectare is a ‘high sediment risk’. More than 0.4 ha of disturbance requires attentiona.
California Stormwater Quality Association (2009)
Slope
Slopes 3:1 or flatter are generally stable. Slopes steeper than 2:1 are generally unstable.
CA Office of Mine Reclamation (2007), BLM (1992)
Slope height
Waste piles or highwalls should always be less than 7.5 ma
States of OR and WA (1997), BLM (1992)
Volume of contaminated soil
190 cubic meters is the threshold from minor to moderate waste pile volume.
US Environmental Protection Agency (1992)
pH
pH [ 9 or \5 is hazardous
States of WA and OR (1997)
Precipitation
Precipitation greater than 90 cm is significant at a mine site
Bezuidenhout et al. (2009)
Impacts to drainage patterns
Stormwater pollution prevention plans are needed in all cases where disturbance is greater than 2 ha1
California Stormwater Quality Association (2009)
a
Legal requirement
Table 4 Reclamation method screening table Category Topographic disturbance
1 Waste pile only
Pit/highwall Waste piles
\1.5 m [40 m3
2 Pit/highwall only [1.5 m \40 m3
3 Pit/highwall and waste pile [1.5 m [40 m3
4 No major disturbance \1.5 m –
Applicable reclamation methods
AEFGHIJKLMOPQR
BDIJLNO
DIJLNOPQ
ABCEFGHKMPQR
Water MCL exceeded, pH [ 9 or pH \ 5
-(I J L)
-(I J L)
-(I J L)
–
Soil COC screening criteria exceeded
-(G I J L M)
-(I J L)
-(I J L)
-(C)
Groundwater \1.5 meters deep
-(P Q R)
–
–(P Q)
-(C P Q R)
Groundwater [1.5 m deep
–
-(B)
–
-(B)
Slopes \H:1 V
-(I J L)
-(I J L)
-(I J L)
–
Off-site disposal/importation not feasible
-(E F)
-(N O)
–
-(E F)
Esthetics important
-(I L)
-(I L)
-(I L)
–
A diversion channel, B construct wetlands, C infiltration gallery, D modify pit into pond, E off-site landfill, F off-site soil treatment, G sell as bulk material, H general recontouring, I slope stabilization, J flatten slopes/highwall elimination, K valley fill, L terracing, M reuse on site as construction material, N partial pit backfill, O construct to original contour/backfill pit, P on-site repository, Q impermeable cap, R cover with soil or mine waste -() indicates that the reclamation methods are being removed from consideration
size, mine features, commodities, mining method, and post-mining topography. Approximately 200 mines were screened to identify the 25 best case studies with mixed characteristics and sufficient information to complete the mine hazard index and screening table. Case study data were obtained from mine reclamation reports and site visits to abandoned mines. Mine reclamation reports were obtained from the EPA, US Forest Service, Bureau of Land Management, mining companies, and the States of Colorado and Idaho. Site visits to abandoned mines in the Sierra Nevada and California Coast Range Mountains (Santa Lucia Range) provided supplemental information for seven of the case studies. The case studies included 20 subsurface mines, four surface mines, and one site including a surface and subsurface
mine. Most abandoned mines are subsurface, and only a few surface mines were found in the literature. This is likely because most abandoned mines are old operations developed before large excavation equipment was available. The case study mines are located in five western states: California, Oregon, Washington, Idaho, and Colorado. A large variety of commodities were recovered at the mines including gold, silver, copper, lead, antimony, zinc, tungsten, phosphate, chromite, arsenic, barite, and slate. The case studies described reclamation using 13 different methods out of the 18 methods being considered in this model. The case studies included sites that used a single reclamation method, some that used several reclamation methods in combination, and some that had no reclamation because the problems were considered minor or the sites were remote.
123
123 Modify pit into pond
0.5 0.5 0
Water contamination
Hydrologic impacts
Overall score
Flatten slope/ highwall eliminations
0.5
0
Overall score
0.5
Soil contamination
0.5
0.5
Slope stability—high walls
Hydrologic impacts
1
Slope Stability—Waste Piles
Water contamination
1
Erosion
0
0
0
0
1
1
1
Slope stabilizations
0
0
0
0
1
1
1
0
0.5
0.5
0.5
0
1
1
Valley fill
0
0.5
0.5
0.5
0
0.5
0
0
1
1
0.5
Terracing
0
0
0.5
0.5
0
Reuse on site
0
0.5
0.5
0.5
0
l
1
0
0.5
0.5
0.5
0.5
1
1
Partial backfill
0
0.5
0.5
0.5
0.5
0
1
0
0.5
1
1
0
1
1
0
1
0.5
0.5
I
I
I
Const. to original contour/pit backfill
0.5
Soil contamination
0
0
0.5
General reconstructing
0.5
Slope stability—high walls
0
0.5
Confinement
0.5
Topographic reconstruction
0.5
0
0.5
1
1
0
1
1
On-site repository
Off-site soil treatment
Infiltration gallery
Diversion channel
Construct wetlands
Removal
Hydrogic modifications
Slope Stability—Waste Piles
Hazard index score
Erosion
Mine site problem
Table 5 Reclamation method ranking matrix template
0
0.5
1
1
0
1
1
Impermeable cap
0
0.5
1
1
0
1
1
Off-site soil treatment
0
0.5
0.5
1
0
1
1
Cover with soil or mine waste
0
0.5
0
0
0
I
l
Sell as bulk material
Environ Earth Sci
Environ Earth Sci
Decision model validation The abandoned mine decision model was validated using the 25 case studies. The purpose of the model validation was to: 1. 2. 3.
4.
Determine whether the model returns reasonable results. Provide an opportunity to tune different components of the model to optimize for accuracy. Compare the reclamation methods selected by the model to those actually used or recommended in detailed reclamation studies. Adjust the model for ease of use.
It was assumed that reclamation methods used at case study sites were appropriate, and this was verified, when feasible, by reviewing post-reclamation monitoring reports. Several steps were taken to calibrate and fine-tune the model: 1.
2.
3.
4.
Several sample sites were initially evaluated to determine whether the model is easy to use and follows a logical process. Consequently, several improvements were made to the model layout, presentation, and descriptive text. The authors identified the sub-parameter conditions and determined ratios for them using the AHP. During initial trials, some of these were modified with new AHP calculations when index scores seemed too high or low based on the actual physical conditions at the case study sites. The AHP system was revised to address deficiencies in the conventional AHP model. This iteration led to the calculation of new parameter weighting factors that more closely agreed with the results of the Delphi method survey. The main parameters, sub-parameters, and their weighting factors were not modified after the case study evaluation began. These parameters and weighting factors are considered to have a firm foundation since they were developed by a panel of experts through an iterative process. The weighting factors were also confirmed with modified AHP calculations.
In conclusion, through a multi-step iteration process, geologic, hydrologic, and geochemical data and reclamation activity reports from the case studies were used to refine and improve the model accuracy. Model calibration runs using case studies Twenty-five case studies were selected for calibrating and iteratively revising the model. The case studies include 15 sites that were reclaimed and 10 that were recommended
for no action. Each case study was analyzed using the: (1) mine hazard index; (2) reclamation method screening table; and (3) reclamation method ranking matrix. Mine hazard index Figure 6 shows the mine hazard index scores for each case study. Data on this graph fall into three regions: 1.
2. 3.
Low priority: mines that do not need reclamation or need reclamation only for erosion, slope stability, and hydrologic impacts (i.e., no contamination). Medium priority: mines that represent moderate hazards and may or may not need reclamation. High priority: mines that constitute major hazards and require eventual reclamation.
The screening table was used to reduce the number of applicable reclamation methods for the 15 case study sites that were reclaimed. The screening table reduced the number of reclamation methods from 18 to an average of 6.6, with a minimum of 3 and a maximum of 11. The case studies included 11 projects that used one reclamation method, and 4 projects that use more than one method. The screening table results included the methods used at the 11 sites that had one reclamation method, and at least one of the methods used at the 4 sites using multiple methods. The remaining ten sites have not been reclaimed (thus far). Overall, the screening table results yielded 17 of the 20 methods used. Reclamation method ranking matrix The ranking matrix was used to score and rank the viable methods at each site. The method used was within the four highest scoring alternatives 80 % of the time. Although this part of the model is only intended to provide alternatives for feasibility studies, the reclamation method used at the case study sites matched the highest ranking method in the model 47 % of the time. Comparison to results from other models The authors also evaluated the 25 case studies using the California Department of Conservation’s Preliminary Appraisal and Ranking System (PAR), so that the results could be compared to the mine hazard index. The PAR system includes two separate components for physical and chemical hazards, each with a maximum score of five. The PAR model has a maximum score of 10 and the mine hazard index a maximum score of 1,000. The PAR system includes variables for subsurface features (e.g., adits and shafts), land subsidence, and old buildings. These hazards were not included in the mine hazard index, so the two
123
Environ Earth Sci Fig. 6 Hazard index scores for case studies
Fig. 7 Comparison of mine hazard index scores and Preliminary Appraisal and Ranking System values for 25 case studies
models were not directly comparable if they were considered. As a result, they were given scores of zero in the PAR system. The PAR score was also adjusted so that the hazards evaluated had a maximum score of 10, which was then multiplied by 100; so it had a maximum score of 1,000, the same as the mine hazard index. Figure 7 compares the index scores using both models. There is general correlation between the two methods, with values typically increasing together. However, if the
123
methods agreed, then the data should fall on the straight line with a slope of one. Instead, the two methods show fairly good agreement with the no-action sites, but PAR exhibits much less discrimination between sites that were deemed to require reclamation and those that did not. Averages for the normalized PAR values are 250 (no-action sites) and 347 (reclaimed sites) with a difference of 97. For the same sites, the mine hazard index yields average values of 285 (noaction) and 478 (reclaimed sites) with a difference of 193.
Environ Earth Sci
This shows that the mine hazard index represents an improvement in differentiating which sites that reclamation professionals deemed to require reclamation. The PAR system loses some precision, because the scores are whole numbers between 1 and 10, with no intermediate values. The PAR system also includes some subjective variables. For instance, highwalls are simply given a hazard ranking from 0 to 4, rather than being scored using relevant parameters, such as highwall height, slope, and geologic composition, as is used in the mine hazard index. Overall, these results suggest that the mine hazard index may provide a better method for prioritizing sites for topographic reconstruction and waste pile reclamation than PAR. The hazard index results could not be compared to other indexes in the literature because they use different parameters or have a different focus. For example, Bezuidenhout et al. (2009) developed an index that includes parameters unique to asbestos mines and therefore is not practical for assessing metal mines. The model results were also compared to the results determined by two published decision trees including Colorado Division of Minerals and Geology (2002) and the Interstate Technology and Regulatory Council (2010). However, these decision trees have limited utility because they only include a limited number of reclamation options and primarily ask simple yes/no questions. This is a reflection of the inherent limitations of decision trees.
Discussion Parameter weighting factors The first phase of the Delphi method survey showed fairly good agreement among the experts. Some experts said they made no changes in the second phase because their values were already similar to the mean values. However, others may have left their values the same because they were volunteers and had no time or desire to contribute further to the study. Nevertheless, since the first phase of the survey shows close agreement, the experts have general concurrence on the relative importance of mine site hazards. The survey identified contamination as the largest problem at abandoned mines. Soil and water contamination had the highest main parameter factors and combined for 55 % of the total. Water contamination had the highest average weighting factor, and 14 of 15 experts chose water contamination as the most important, or tied as the most important parameter. The five parameters would have weighting factors of 0.2 if weighted equally. However, the lowest weighting factors assigned were 0.14 for both slope stability and hydrologic impacts, indicating that none of the factors considered were minor.
Initially, the AHP results did not agree well with the Delphi method results. A close inspection revealed that the discrepancies were restricted to parameters that had smaller differences in importance than the AHP was designed to discriminate. For example, when two parameters have a ‘weak’ difference in importance, the conventional AHP provides weighting factors of 67 and 33 % for the two parameters. The authors consider these weighting factors more appropriate for two parameters that have a strong or significant difference in importance. The AHP scoring system was therefore modified to provide finer distinction between parameters. This revision provided results more consistent with the Delphi method. Here, a modified AHP is proposed with more levels of comparison between parameters than the traditional AHP (Table 1), and a revised scoring system that is more consistent with the verbal scale used in the AHP (Saaty and Vargas 1991). The revised AHP provides an improved method of comparing and ranking parameters that could be used by other disciplines and applications. In keeping with Miller’s law, the revised AHP still contains a tractable number of categories. Reclamation versus no action Mine reclamation literature and the case studies reviewed for this project revealed that many abandoned mines are not being reclaimed, because they represent minimal hazard, are located in remote areas distant from population centers, have naturally reclaimed over time, or public agencies must focus on higher priority mines. For instance, a review of 162 abandoned mine assessment reports by the State of Idaho showed that only 31 % were recommended for reclamation, 54 % were recommended for no action, and 22 % were recommended for additional studies. In addition, the Western Governor’s Association concluded that 80 % of abandoned mines have no major safety or environmental concerns (Struhsacker and Todd 1998). As a result, developing a threshold for deciding whether reclamation is needed became one focus of this study. To help determine this threshold, the authors purposely selected ten case study sites where no action was taken. Figure 7 shows that most case studies with mine hazard indexes below 300 were not reclaimed. Model uses and limitations The decision model was calibrated with 25 abandoned metal mines in the western USA. The selection of case study locations was not arbitrary. The western USA has been extensively mined in the past and has many remote areas that still harbor abandoned mines. However, the decision model should be applicable to other areas with
123
Environ Earth Sci
similar physiography, climate, and geology, such as Western Canada and Patagonia. The model purposely excluded coal mines and gravel pits, since they use different mining methods and exhibit different features than metal mines. For example, strip mines are commonly used for extracting coal, but are less commonly used for metal mining. Nevertheless, coal and aggregate mines share many similarities with metal mines. The model may offer some benefit for evaluating these, especially if it is modified and tested to be more universal through additional research. Furthermore, the model was developed for abandoned mines, but may also be useful at active mines. The model does not account for factors such as endangered species, public opinion, and esthetics. These issues are significant drivers of mine reclamation, but vary dramatically over time due to economic, political, and social factors. They were not included in the model primarily because they are difficult to quantify, and the model focused on geologic hazards rather than all issues pertinent to abandoned mines. Cost is not explicitly considered due to different operating expenses between geographic areas and changing economic conditions over time. Cost is a factor that is more useful in subsequent feasibility studies. The purpose of the model is to identify high priority sites and promising reclamation methods that should be evaluated in a feasibility study, at which time these other factors can also be considered.
Conclusions Abandoned mines present numerous safety and environmental problems from altered topography and the poor management of mine waste materials. These problems can be reduced or eliminated with proper reclamation. However, an extensive literature review concluded that few decision models were available for selecting mine reclamation methods. Most existing models also exhibit various deficiencies such as absence of important technical parameters, limited number of reclamation methods considered, lack of an analytical process for assigning weighting factors, and lack of model calibration. Lastly, no model was found that can both assess a mine and recommend reclamation methods. To address these problems, an abandoned mine decision model was developed that includes three consecutive decision guidance tools: (1) mine hazard index; (2) reclamation methods screening table (screening table); and (3) reclamation methods ranking matrix (ranking matrix). The mine hazard index was developed with input from a group of experts, thus embodying some of their expertise, and the model was calibrated with case studies for 25 abandoned metal mines in the Western USA.
123
The mine hazard index includes five main parameters, 18 sub-parameters, and several possible conditions or a continuum of values for each sub-parameter. The five main parameters and their weighting factors are: water contamination (31 %), soil contamination (24 %), erosion (17 %), slope stability (14 %), and hydrologic impacts (14 %). Contamination (soil and water) is considered the most important hazard and accounts for 55 % of the total importance. The mine hazard index is effective at ranking and prioritizing mine sites and provides information needed for subsequent parts of the decision model. Twenty-five case studies were scored with the mine hazard index, including 15 mines that were reclaimed and 10 that were recommended for no action. The results identified three general categories of abandoned mines: 1.
2. 3.
Low priority (score \300): mines that do not need reclamation or need reclamation only for erosion, slope stability, and hydrologic impacts (i.e., no contamination). Medium priority (score 300–600): mines that have moderate hazards and may or may not need reclamation. High priority (score [600): mines that have major hazards and require reclamation.
The screening table and ranking matrix were tested with the 15 case study mines that were reclaimed. These two components of the decision model were effective at narrowing and selecting a few of the most appropriate reclamation methods. The screening table reduced the number of reclamation methods from 18 to an average of 6.6 for the 15 case studies. Although it was not intended to always pick the best reclamation method, the ranking matrix correctly predicted the reclamation method used 47 % of the time. In addition, the method used was within the four highest scoring alternatives 80 % of the time. The abandoned mine decision model offers an improved method for prioritizing mine sites and narrowing the list of viable reclamation methods. The model allows for rapid assessment of large numbers of mines because it provides a systematic approach that can be easily repeated. As a result, the model can help save money by reducing the time to perform an analysis. In addition, the analyses conducted here indicate that the abandoned mine decision model represents an improvement for topographic reconstruction and mine waste piles over the PAR method (California Office of Mine Reclamation 2000), given that it more clearly discriminates sites that were judged by reclamation professionals as needing action. The model helps to convert a subjective decision process (e.g., best technical judgment) into a more objective process due to input from a panel of experts, model calibration, and a documented systematic procedure. The process is also transparent, rational, and consistent when repeated on multiple mine sites.
Environ Earth Sci Acknowledgments The authors would like to thank Dr. John Wakabayashi and Shay Overton for their review comments, which substantially improved the manuscript. We would also like to thank the following individuals who participated in the Delphi Method Survey: Barbara Brandl, David Norman, Fritz Wolff, Jeff Johnson, John Kirk, Maggie Baker, Pat Trainor, Peter Werner, Sam Hayashi, Dr. Stuart Jennings, Dr. Terrence Toy, and Tom Buchta.
References Albert EK, Ince E, Mutmansky JM (1991) Fuzzy multiple-criteria decision-making—application to AML project selection. Int J Surf Min Reclam 5(4):167–176 Bandopadhyay S, and Chattopadhyay A (1986) Selection of post mining uses of land via fuzzy algorithm. 19th Conference on Applications of Computers and Operations Research in the Mineral Industry, Society of Mining Engineers, Littleton, 321–332 Bascetin A (2007) A decision support system using analytical hierarchy process (AHP) for the optimal environmental reclamation of an open-pit mine. J Environ Geol 52:663–672 Bezuidenhout JJ, Claassens S, Jansen PJ, van Rensburg L (2009) Rehabilitation of asbestos mining waste—a rehabilitation prioritisation index (RPI) for South Africa. J Environ Geol 57:267–273 Black JP, Toy TJ (2000) Topographic reconstruction—the theory and practice. In: Bartels J et al (eds) Reclamation of drastically disturbed lands. ASA Publications, Madison, pp 41–75 Brown RW, Sidle RC (1992) Decision model for successful reclamation of disturbed lands. International Symposium on Land Reclamation—Advances in Research and Technology, Nashville, 14–23 California Office of Mine Reclamation (2007) Surface mining and reclamation act and associated regulations California Office of Mine Reclamation, Abandoned Mine Lands Unit (2000) California abandoned mines—a report on the magnitude and scope of the issues in the state, Volume II, pp 16-25 California Stormwater Quality Association (2009) Stormwater best management practice handbook portal—construction, pp 2-1 to 3-1 Colorado Division of Minerals and Geology (2002) Best practices in abandoned mine land reclamation—the remediation of past mining activities Dias MB, Rodriguez A, Rodriguez R, Vigil H (2011) Development of a user-friendly method to assess the present condition of old abandoned mining waste dumps in Asturias (Spain). Int J Min Reclam Environ 25(1):6–31 Dutta S, Rajaram R, Robinson B (2005) Mineland reclamation. In: Dutta S, Parameswaran K, Rajaram V (eds) Sustainable mining practices—a global perspective. A. A Blakema Publishers, Leiden, pp 179–191 Harwood GD, Thames JL (1988) Design and planning considerations in surface-mined land shaping. In: Hossner LR (ed) Reclamation
of surface-mined lands, vol I., CRC PressBoca Raton, FL, pp 138–155 Hsu C, Sandford BA (2007) The Delphi technique—making sense of consensus: practical Assessment. Res Eval 12(10):1–8 Hu Z, Linlin W (2009) Landscape restoration regionalization for resource-exhausted coal mine areas based on GIS. National Meeting of the American Society of Mining and Reclamation, Lexington, pp 1512–1519 Interstate Technology & Regulatory Council Website (2010) Mining waste treatment technology section. http://www.itrcweb.org/ miningwaste-guidance/decision_tree.htm. Accessed 2 March 2011 Klimstra WD, Nawrot JR, and Perkins MP (1983) Potential benefits and hazards of highwalls. Symposium on Surface Mining, Hydrology, Sedimentology and Reclamation, Lexington, 131–137 Miller GA (1956) The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychol Rev 63(2):81–97 Oregon Department of Geology and Mineral Industries and Washington Division of Geology and Earth Resources (1997) Best management practices for reclaiming surface mines in Washington and Oregon, Open File Reports O-96-2 and 96-2, pp 1.1–6.8 Robertson AM, and Shaw SC (1998) Alternatives analysis for mine development and reclamation. 22nd Annual British Columbia Mine Reclamation Symposium, Penticon, BC 51-62 Saaty TL, Vargas LG (1991) Prediction, projection and forecasting Applications of the analytic hierarchy process in economics, finance, politics, games and sports. Kluwer Academic Publishers, Norwell Soltanmohammadi H, Osanloo M, and Aghajani Bazzazi A (2008) Developing a fifty-attribute framework for mined land suitability analysis using AHP-TOPSIS approach. Proceedings of postmining symposium, Nancy, 1–12 Soltanmohammadi H, Osanloo M, Bazzazi AA (2009) Deriving preference order of post-mining land-uses through MLSA framework: application of an outranking technique. Environ Geol 58(4):877–888 Soltanmohammadi H, Osanloo M, Bazzazi AA (2010) An analytical approach with a reliable logic and a ranking policy for postmining land-use determination. Land Use Policy 27(2):364–372 Struhsacker DW, and Todd JW (1998) Reclaiming inactive and abandoned hardrock mine lands in the west. National Mining Association, pp 1–37 United States Government Accountability Office (2011) Abandoned mines—Information on the number of hardrock mines, cost of cleanup, and values of financial assurances, Statement of Anu K. Mittal, Director Natural Resources and Environment Team, GAO-11-834T US Bureau of Land Management (1992) Solid minerals reclamation handbook: BLM Handbook H-3042-1, (variously paged) US Environmental Protection Agency (1992) Hazard ranking system guidance manual—interim final, pp 21–40 US Environmental Protection Agency (2000) Abandoned mine site characterization and cleanup handbook, pp 1–1 to 11–15
123