Journal of Cleaner Production 19 (2011) 580e587

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

A simplified matrix of environmental impacts to support an intervention

program in a small-scale mining site
Rodolfo N. Sousa a, *, Marcello M. Veiga a, John Meech a, Janis Jokinen a, Aguinaldo J. Sousa b
a
University of British Columbia, NBK Institute of Mining Engineering, 517, 6350 Stores Road, Vancouver, BC V6T 1Z4, Canada
b
Suzano Pulp and Paper, Av. Brigadeiro Faria Lima, 1355, 5th to 8th floor, São Paulo 01452-919, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: This paper describes a tool developed to evaluate environmental risks at a Brazilian artisanal small-scale
Received 11 August 2009 gold mining site. The tool consists of a matrix used to determine the relevance of such risks through the
Received in revised form classification of mining activities according to the likelihood that they may cause environmental impacts.
26 November 2010
The language and method are very simple, and thus appropriate for local stakeholders. It ensures their
Accepted 27 November 2010
Available online 4 December 2010
participation when developing a plan of action for intervention and training. When applied to a pilot site
in the Brazilian Amazon, the matrix revealed the variables that pose higher environmental risks. These
activities were subsequently addressed in a training program to promote better practices. As a result,
Keywords:
Artisanal mining
the mining activities were improved by 28.8% when comparing compliance to environmental requisites
Matrix of risks before and after the training program. Specifically, the training program resulted in improved gold
Environmental indicators recovery, and reduced environmental and health impacts.
Environmental impacts Ó 2010 Elsevier Ltd. All rights reserved.
Heuristic system
Training
Self-assessment

1. Introduction and background worldwide (Brundtland, 1987). Education is a critical determinant

of public health (WHO, 2009). To further discuss the link between
Artisanal and small-scale gold mining (ASGM) generally poverty and education, Krishnamurti (1956) defines the latter as
involves rudimentary mining and processing methods, and occurs something broader than merely acquiring knowledge, or gathering
in conditions where few safety precautions are exercised. Often, and correlating facts; it also involves recognizing the significance of
artisanal gold miners are poorly trained, have low levels of formal life as a whole, and is deeply connected with health and the envi-
education, produce very little gold, and have minimal consideration ronment. In the ASGM context, miners struggle to survive and have
for the environment. Moreover, ASGM is typically undercapitalized little access to education, demonstrating the cycle of poverty and
and illegal, carried out by individuals who do not have concession depressed safety, health and environmental conditions and stan-
rights (Labonne and Gilman, 1999). With few exceptions, ASGM is dards that persist in ASGM locations.
strongly associated with poverty, and employs at least 30 million The need to address poverty as a means to reduce environ-
people directly and indirectly worldwide (UNIDO, 2009). Losses of mental and health impacts and other social related problems in
mercury e used to amalgamate gold e to the atmosphere, soils and ASGM has been recognized. Hilson (2010a), studying artisanal
water contamination are the industry’s most significant environ- mining in Africa concluded that the phenomenon of child labour in
mental impacts, although these activities cause a range of other ASGM is a poverty-driven problem and increased educational
problems, including deforestation, water siltation, cyanide support is capable of discouraging children from engaging in mine-
contamination and soil degradation (Sousa and Veiga, 2009). related work. Also in Africa, attempts by governments and policy
While there are many issues facing the ASGM sector, this paper makers to encourage mining communities to move into farming
focuses specifically on poverty and (the lack of) education. Poverty (“re-agrarianize”) has failed and artisanal mining has become the
has been identified as a main driver of environmental degradation primary indispensable economic activity for poor communities
(Banchirigah and Hilson, 2010). Other African studies demonstrate
that despite the social and environmental problems related with
* Corresponding author. Tel.: þ1 604 8226717; fax: þ1 604 822 5599. artisanal mining, evidence suggests that people working in ASGM
E-mail address: [email protected] (R.N. Sousa). are less likely to be impoverished than those with other

0959-6526/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jclepro.2010.11.017
 R.N. Sousa et al. / Journal of Cleaner Production 19 (2011) 580e587 581

occupations (Fisher et al., 2009; Hilson, 2010b). In many aspects, This paper examines some of the differences between conven-
the situation in Africa resembles the ASGM scenario in the Amazon tional EIA/ERAs and the developed tool, a simplified matrix of
region, where people also see mining as a better alternative to environmental impacts. This paper also describes how the matrix
traditional agriculture and farming. This scenario demands atten- can be applied as a practical and useful tool for local stakeholders
tion to the better organization and regulation of ASGM, making the and implementing agencies to classify the significance of impacts
application of simple tools to help control environmental and resulting from ASGM activities. The matrix tool is not as compre-
health impacts invaluable. hensive as conventional EIA/ERAs, however it is the authors
Worldwide, the formal mining sector is frequently the target of opinion that it is sufficient enough to support interventions
stringent environmental regulations and legislation, a result of grounded in the belief that the most relevant activities are identi-
pressures exerted by environmentalists and NGOs, and society as fied, and can be addressed in interventions to improve environ-
a whole. However, informal mining sectors such as ASGM in most mental conditions and standards.
cases are not subject to proper regulations, or if they are, govern- Concepts related to a matrix of environmental impacts have
ments lack the capacity to provide adequate technical assistance or been evolving for decades. The NEPA (USA’s National Environ-
to enforce compliance. Overall, environmental impacts caused by mental Policy Act), developed in 1969, is an original tool used for
the world’s poorest population are less subject to monitoring and the evaluation of environmental aspects and impacts related to
control than impacts caused by formal sectors. As the poorest development projects (NEPA, 1969). NEPA established the need to
global population is concentrated in developing countries, most accomplish evaluations of environmental impacts, including
governments lack efficient mechanisms to implement programs to accounting for the beneficial and adverse impacts resulting from
mitigate and control environmental risks, or to enforce their own actions capable of significantly affecting environmental quality.
legislation. Within developing countries, the misuse of mercury is The evaluation of environmental impacts can be subdivided into
a prominent focus for governments, researchers, NGOs and inter- three distinct phases: “identification”, “estimation” and
national organizations. Despite the existence of alternatives solu- “comparison”. In the identification phase, environmental elements
tions based on gravity concentration and cyanidation of are defined traditionally as physical, biological and human
concentrates, for instance, mercury amalgamation remains the components of a specified ecosystem. The identification of envi-
preferred method of application at most ASGM sites (Vieira, 2006; ronmental impacts establishes the possible consequences
Sousa et al., 2010). (impacts) of a quantified number of activities (aspects). The
In Brazil, regulations have been established to attempt control concept of an identification matrix, broadly used to support
the ASGM activities. For instance, the Brazilian government has management systems, is based on ISO 14001(Barrow, 1999), and is
adopted bureaucratic procedures that require environmental a useful tool according to Leopold (1971). The matrix of impacts
impact or risk assessment studies for ASGM activities. The defini- correlates activities to the affected elements. Leopold’s system
tions of these processes (Environmental Impact Assessment/EIA (1971) was used as a guide and is realized through a bi-dimen-
and Environmental Risk Assessment/ERA) may have distinct sional matrix. In one dimension, the existing characteristics and
meanings depending on the authors (Covello and Merkhofer, 1993; conditions of the environment are characterized and include the
Fairman et al., 1999; Murray and Claassen, 1999; Glasson et al., following components: physical and chemical characteristics,
2005). For instance, Glasson et al. (2005) define an EIA as “the biological conditions, cultural factors, and ecological relationships.
process of identifying, predicting, evaluating and mitigating the The other dimension entails the proposed actions which may
biophysical, social, and other relevant effects of development cause environmental impacts and includes commentary related
proposals prior to major decisions being taken and commitments to: modification of regime, land transformation and construction,
made”. The USEPA (2010) defines ERA as a process that evaluates resource extraction, processing, resource renewal, chemical
the likelihood that adverse ecological effects are occurring or may treatment, and accidents. After the identification phase, it is
occur as a result of exposure to one or more stressors. Suter II et al. necessary to estimate the evolution of all environmental impacts.
(2000) offer a similar explanation, defining ERA as the process of Models are developed to estimate the magnitude of the current
collecting, organizing and analysing information to estimate the impacts and there are at least three employed methods for this
likelihood of undesired effects on non human organisms, pop- representation: simple mathematical functions, dynamic models
ulations, or ecosystems. and verbal declarations (Veiga and Meech, 1995). Simple mathe-
While conventional EIA/ERAs are appropriate tools for large matical functions of the models can be heuristic equations rep-
organizations, they are inappropriate for regulating artisanal resenting the knowledge of an expert. These functions can be
miners, who are not organized at the desired scale or sufficiently defined mathematically, or simply by employing a ranking system.
empowered to develop and use these elaborate tools. Most ASGM The linguistic expression allows for the treatment of variables in
activities take place in poor communities in isolated locations, ranks, in which a group of logical statements can be algebraically
where people lack formal education, interest and/or resources to represented. These logical affirmatives are later transferred to
invest in any comprehensive monitoring systems. In such condi- a flowchart tree, defined for a group of experts. The final phase,
tions, it is imperative that technological assistance and education called the comparison phase, uses the information generated in
are simplified and made user-friendly (Hilson, 2005). To ensure the previous phases to identify a “best fit” solution that takes into
successful implementation of EIA/ERAs in such an environment, consideration the magnitude and the significance of the envi-
risk assessment tools must be simple, inexpensive and easy to ronmental impact.
apply. In addition, local stakeholders (local government, local The method adopted for the identification, evaluation and
environmental agency, miners’ associations, community leaders) control of the environmental aspects with significant impact is
must be able to participate, discuss, understand and use the risk based on information gathered from the local interested parties,
assessment tool. Respecting different contexts and realities, it is identification of pertinent legislation, numeric evaluation of the
hypothesized that the ASGM sector requires a far more simplified significance of potential impacts, and monitoring and control of
approach, and the development of simple alternatives to evaluate significant environmental parameters. The term “interested party”
the environmental risks posed by activities carried on by informal is defined by ISO 14001 (1999) as an individual or group inter-
sectors such as ASGM is of great usefulness to support intervention ested or affected by the environmental performance of an
programs. organization.
582 R.N. Sousa et al. / Journal of Cleaner Production 19 (2011) 580e587

2. The study The proposed matrix of environmental risks presented in this

study represents a collaborative effort between researchers and
2.1. Artisanal and small-scale gold mining in the Brazilian Amazon: local stakeholders that preceded the implementation of a field
an Overview training regime in mining sites as promoted by the United Nations
Global Mercury Project (GMP) (2008) that ran from 2006 to 2008
The Brazilian Amazon, considering the number of miners, gold (Sousa and Veiga, 2007b). The matrix considered both directly
production, mercury consumption and the overall environmental measurable impacts, and the stakeholders’ perception of such
impacts derived from these activities, is one of the most important impacts. The collaborative process that occurred between
ASGM sites in the world. It is locally known as “garimpo”, (Veiga and researchers and stakeholders was a fundamental step in imple-
Baker, 2004; Veiga et al., 2006). The number of individuals esti- menting GMP goals, and partners represented individuals who
mated to be involved in ASGM activities in garimpo is believed to be could assure the sustainability of actions after the GMP project was
around 200,000 (DNPM, 2008). Despite the large number of concluded. The matrix was developed specifically for mines located
informal miners operating in the Amazon for at least the past three in the Tapajos River Basin region, a tributary of the Amazon River.
decades, the Brazilian government has provided little assistance to However, the concepts associated with the matrix can be adapted
formalize and improve their working conditions. to any other ASGM site. The implemented matrix was useful in
In Brazil, in line with the international trend, environmental determining the relevance of risks based on the classification of
regulations and legislation have been strengthened over time. activities according to the likelihood (or Degree of Belief) they
Regulations at the Federal, State and Municipal levels address might cause environmental impacts, and has proved useful in
environmental criteria for companies and individuals to operate supporting actions that promote good practices and training to
their businesses. In 1986, the Federal Environmental Council mitigate environmental impacts.
(CONAMA) established the compulsory criteria for the evaluation of
environmental impacts as part of the licensing process for mining
operations. As such, new mines are required to provide a plan to 2.2. Study site and methods
mitigate environmental impacts. In the Tapajos region, where
miners extract alluvial and colluvial gold, and in some cases, process The selected site for this study was the municipality of Itaituba,
primary quartz-vein ore, it is estimated that 99% of miners do not in the Tapajos River Basin, in the Central Brazilian Amazon.
attain the legal criteria to obtain the environmental license, and According to SEMMA (Itaituba Environmental Agency) (2006), it is
thus, are considered illegal miners (Sousa and Veiga, 2007a). estimated that 40,000 miners work in this location, extracting
Regardless, they have been mining for decades, despite existing legal colluvial and primary gold. Specifically, the artisanal mining area
requirements. known as “Garimpo Canaan” was the first area to be evaluated
Legislation alone does not assure that artisanal miners comply through the use of the matrix of environmental risks.
with environmental, safety and health requirements. For example, At least three scenarios should be analyzed when considering
the Brazilian decree 97507-89 does not allow the use of mercury in the use of environmental risk assessment (ERA) tools for ASGM. The
“garimpos” without an environmental license (Ministry of classical ERA and the developed matrix tool are highlighted and
Environment, 1989), but amalgamation is still the preferred and compared in Table 1.
simplest method to extract gold in this location. Hinton et al. (2003) At one extreme, there is the classic ERA model. This process
and Hilson (2006) highlight that, worldwide, the amount of money manages the uncertainty and probability of adverse impacts and
spent on educating miners is considerably lower than those of potential risks of multiple stressors on complex ecological
other approaches, such as monitoring and enforcement. communities (Suter II, 1993). All information and data are subject to
To add to the complexness of the ASGM context, artisanal gold interpretation, and professional judgment is required to make
miners live and work in remote areas where government (Munic- decisions and find solutions in the management of potential risks.
ipal, State, Federal) is absent and do not participate in the devel- Classical ERA processes involve problem formulation, exposure
opment of economic options that could bring enhanced benefits to assessments, effect assessments, risk characterizations (including
miners. In the present scenario, there is no simple solution to extent, pathway and form of contamination), risk assessments, risk
improve conditions; however, for administrative processes and management and risk communication (Morris and Therivel, 1996;
regulations to be effective, they need to be simplified to reflect Suter II et al., 2000). The determination and analysis of contami-
more accurately the realities of ASGM. This includes consideration nants of potential concern (COPC), pathways, and the receptors of
of simple criteria to identify environmental risks associated with concern (ROC) during classical ERA studies, do not fit within the
artisanal miners’ activities, and enabling the implementation of ASGM reality. This classical process is costly, time consuming,
environmental programs that address such impacts. demands a higher level of expertise, and the conversion of the risk

Table 1
Comparison between classical ERA, simplified matrix of risks and the current practices in ASGM.

Classical environmental risk assessment (ERA), Simplified matrix of environmental risks, The current scenario of ASGM, in terms of using
if applied to ASGM if applied to ASGM environmental risk assessment
Stochastic methods: requires data, sampling, calculations Heuristic methods: based on the existing knowledgea No formal control in place
A probabilistic process based on data and statistics A probabilistic process based on experience
Requires expert people Requires basic understanding
High costs involved for ASGM Low costs involved
Less flexible More flexible
Not every stakeholder can understand Better chances to be understood by all the stakeholders
Generates information to support a plan of action Generates information to support a plan of action A plan of action is possible, but may
not address the key variables
Reduced level of subjectivity Intermediate level of subjectivity High level of subjectivity
a
Note: knowledge, in this case, is obtained from experts or traditional miners.
 R.N. Sousa et al. / Journal of Cleaner Production 19 (2011) 580e587 583

outcomes into practical implementations of preventive actions or 2.3. The matrix of environmental impacts in the pilot site
remediation procedures would be difficult.
Classical ERA or EIA are appropriate technical tools for the Programs that address the reduction and control of environ-
formal sector, but the language and methods exceed the artisanal mental impacts in ASGM generally rely on the experience and
miners’ comprehension and capacity. However, a total lack of any expertise of a project manager in order to set objectives and
tool or environmental assessment also does not help the ASGM identify the most important variables to be controlled. This method
situation. In this context, when environmental concerns are of selecting objectives and goals, and giving relevance for different
brought into consideration, generally, conjectures and isolated environmental impacts, can be improved by reducing the level of
opinions have prevailed. A third option, as proposed in this study, is subjectivity, and through the promotion of local stakeholder
the heuristic matrix for determining environmental risks. This participation.
approach is situated between the aforementioned extremes and The development of the matrix involved local stakeholders
represents an alternative tool that is simple, not expensive and is (with representatives of the GMP management team, the local
easily understood by the main local stakeholders. When simplifi- miners’ association and the local government) in order to support
cation is a requirement, heuristic systems are generally more the risk assessment and to enhance understanding of environ-
appropriate than the stochastic ones (Zadeh, 1965; Veiga and mental impacts. The tool is a measurable way to determine the
Meech, 1995). most relevant environmental variables in small-scale mining sites.
The proposed matrix of environmental risks was developed by This heuristic method classifies the variables according to the
the GMP team and designed in collaboration with local stake- likelihood (or Degree of Belief) they may be associated to envi-
holders, considering their perception and judgment on such ronmental impacts. This process allowed for the establishment of
impacts. While the current scenario in these garimpos is clearly the action plans focused on the control of the most critical variables.
total absence of controls, a heuristic and intuitive method based on Through this matrix of evaluation, every activity was classified
fuzzy logic seems to fulfill this gap adequately. according to parameters with the summation indicating its signif-
The Fuzzy Logic technique was created by Zadeh (1965) and icance for the environment. Variables with higher likelihood to
employs human analysis to provide an approximate but effective cause environmental impact thus were given a higher priority
means to describe the behaviour of situations which are too when controls, training, awareness and any other action were
complex to allow for precise mathematical analysis. As mentioned implemented. The matrix uses similar concepts to the HgEX System
by Veiga and Meech (1995), heuristic equations and fuzzy sets are developed by Veiga and Meech (1995), and aims to assess the
suitable techniques to mimic the reasoning of an expert, conferring significance of all main environmental impacts, going beyond
elasticity to the Degree of Belief on the conclusions. These linguistic mercury related issues.
terms have the same effect as complex mathematical models that There are many activities involved in artisanal mining, and some
usually demand high costs, large amount of data and complex hold a higher potential to cause environmental damage than
software to quantify the relationship among variables. The heuristic others. ASGM activities vary from simple panning to dismantling
system developed by these authors, called HgEX, aims to evaluate colluvial deposits with hydraulic monitors (a process that removes
mercury bioaccumulation in artisanal mining sites. large amount of soil and vegetation). Every variable (mining
It is likely that different experts on ASGM issues classify the activity), therefore, carries a potential environmental impact. In
relevance of environmental impacts by using the simplified matrix order to compose the matrix, a flowchart of the mining process in
will be led to equivalent conclusions, as the method classifies the the area was evaluated step by step, and these activities were dis-
impacts based on a few relevant factors. If there is only one factor, cussed with the participants. A summary of the variables is given in
as observed by Veiga and Meech (1995), what is considered a low Table 2.
impact for some people may not be considered as low for others. ASGM processes in the Tapajos River Basin vary depending on
There is a higher uncertainty at using a single factor, however, the type of geological deposit. This study considered the most
when the criteria involve the distribution and classification into typical process, involving the excavation of pits with hydraulic
more factors (seven in the present case), chances are that different monitors, the concentration of gold by passing the ore slurry
people with reasonable knowledge on the matter would converge through sluice boxes layered with carpets, and the recovery of
to equivalent classification, expressing the degree of belief in which gold through amalgamation, as described by Veiga and Hinton
an activity may cause a significant environmental impact. (2002).

Table 2
List of mining activities (or variables) with potential environmental impacts.

# Activities # Activities (continuation)

1 Surveying 17 Backfilling pool amalgamation
2 Vegetation removal 18 Backfilling old pits
3 Equipment transportation 19 Backfilled pit accommodation
4 Generator and motor installation 20 Ploughing refilled pit for revegetation
5 Excavation with hydraulic monitors 21 Revegetation
6 Sluice box installation 22 Fertilization of revegetation
7 Water catchment for using in sluice boxes 23 Fighting pests and diseases
8 Carpet washing 24 Infra structure (roads, bridges)
9 Mercury use 25 Camping construction
10 Manual amalgamation 26 Well excavation
11 Concentrate gathering 27 Garbage disposal
12 Panning pool construction 28 Garbage burning
13 Panning pool operation 29 Used oil disposal
14 Amalgam burning without retort 30 Latrine construction
15 Use of retort / mercury recycling 31 Motors and pumps maintenance
16 Mercury reactivation
584 R.N. Sousa et al. / Journal of Cleaner Production 19 (2011) 580e587

The term environmental aspect is defined by ISO 14001 (1999) Table 4

as any element of the activities, products or services of an organi- Potential environmental impacts associated with the environmental aspects.

zation that can interact with the environment. It must be noted that # Potential Environmental Impacts # Potential Environmental
a significant environmental aspect is the one that has or can have Impacts (cont.)
a significant environmental impact. ISO 14001 (1999) also defines 1 Soil physical alteration 8 Flora alteration
environmental impact as any modification of the environment, 2 Soil chemical alteration 9 Fauna alteration
3 Soil microbiological alteration 10 Depletion of renewable
adverse or beneficial, that results, in the whole or partly, of the
resources
activities, products and services of an organization. For the purpose 4 Water contamination 11 Noise emission
of this study, the terms environmental aspect and environmental (chemical / biological)
impact were differentiated, although in some other studies terms 5 Water siltation 12 Human health alteration
like effect, aspect and impacts are sometimes used interchangeably. 6 Water, soil and fauna alteration 13 Green house effect
7 Air quality alteration
For instance, the activity involving removal of vegetation is asso-
ciated with the environmental aspect “deforestation”, and is asso-
ciated with environmental impacts such as “soil exposure”, For example, for the variable “use of fire to remove vegetation”
“erosion” and “soil microbiological alteration”. Environmental a hypothetical interpretation according to the selected criteria
impacts refer to the physical, chemical or biological alterations in could be: Public perception is high, since forest fires are commonly
the environment, and cannot always be observed. perceived by society as a serious issue. For the purpose of this study,
In general, the public understands environmental impact as technical issues related to each variable are not fully explored, as
something necessarily adverse. For the purpose of this study, the objective is to show the practical side and usefulness of this tool
however, environmental impacts are defined as any anthropogenic for stakeholders and implementers to classify the relevance of
alteration to the natural environment. In this case, they can be impacts resulting from ASGM activities. For this example, the extent
adverse or beneficial. Adverse impacts are characterized as those varies depending on the history of fires registered in the region,
that negatively affect the environment. For example, the indis- based on comparison between a current extent and historical
criminate use of fire to control pastures, pests, generating soil occurrences in the region. Frequency also depends on the current
exposure and alteration of soil microbiological activity are exam- frequency of fires when compared to historical numbers in the
ples of adverse impacts. Beneficial impacts are those that positively region. For example, how often are fires expected to occur in a year/
affect the environment and could include a prescribed burn of month/week? Intensity is determined by the percentage of trees
surrounding vegetation to avoid uncontrolled fires (natural or not) individually affected, varying from superficial to total loss (when
that may produce more damage is considered a beneficial impact. only ashes remain). A possible final conclusion: “use of fire to
The variables listed in Table 2 are associated with potential remove vegetation” has high public perception (weight 3), inter-
environmental aspects (Table 3) and impacts (Table 4). The matrix of mediate extent (weight 2, depending on the area), low frequency
environmental impacts provides linkages between the parameters (weight 1, assuming the frequency is low in the area) and high
indicated in the tables, and estimates the significance of each vari- intensity (weight 3, assuming hypothetically the damage is high).
able, based on the degree of belief that they may cause significant This information provides the basis for further classification of the
environmental impact. This evaluation relies on relevance factors to significance of each variable (activity), in terms of its potential to
measure the importance of each variable, as shown in Tables 3 and 4. cause environmental impact.
In order to give the sense of relevance, each variable (activity) In addition to the previous factors, three other “non-quantita-
was classified according to different factors, chosen heuristically by tive” factors were considered: legislation, interested parties and
the key stakeholders. These factors included public perception, detectability. As they are not ordinarily quantitative, no weight was
extent, frequency and intensity. Aiming to simplify the system, considered, only their pertinence. If they applied to an activity, one
the classification criteria considered that each factor has the same additional point was considered; otherwise no point was added.
importance when compared one another. Four levels of relevance Legislation: this factor adds one more point to the variable
for each factor were considered, with different weights (0, 1, 2 or 3 related to legal requisites. The analysis is a simple yes (1 point) or
points). The weights, according to the relevance of the factors, no (0 point). Interested Parties: this factor adds another point in case
were: no relevance, representing 0 point (or does not apply for the variable may affect directly any interested parties. For this
a particular activity); low relevance, representing 1 point; interme- study, the main interested parties were represented by the local
diate, equal to 2 points, and high relevance, representing 3 points. government, local garimpos and neighbouring communities.
Public perception refers to the likelihood that the general public Detectability: this factor adds another extra point for variables that
may perceive or associate a variable with an environmental impact. generate detectable and traceable impacts.
It varies from low to intermediate or high. Extent refers to the As resources to implement actions are limited, the method
geographical area affected by the activity. Frequency refers to the prioritizes variables that are legal requirements, affect other parties
frequency that certain activity is expected to occur, generating and are detectable and traceable. Example: the “use of fire to
impact. Intensity refers to the intensity in which a variable is remove vegetation” is regulated by law, affects directly neigh-
believed to impact the environment. bouring areas and is easily detectable. As a result, considering

Table 3
Potential environmental aspects associated with the mining activities.

# Potential Environmental Aspects # Potential Environmental Aspects (cont.)

1 Soil compression 8 Mercury release
2 Soil exposure 9 Particulate material emission
3 Soil removal (organic material) 10 Sediment release
4 Water consumption 11 Garbage production and disposal
5 Material consumption 12 Used oil disposal
6 CO2 emission 13 Use of pesticide
7 Mercury vapour emission
 R.N. Sousa et al. / Journal of Cleaner Production 19 (2011) 580e587 585

one point for each factor, 3 extra points are considered for this 3. Results and discussion
variable.
Ultimately, the matrix evaluates the significance of each variable Table 5 presents an organization of variables, correlates the
in the whole environmental impact. Thus, if a hypothetical variable associated environmental aspects and impacts, and classifies such
X was highly significant, this means that it may contribute to aspects and impacts for decision on inclusion of activities for
produce a highly significant environmental impact. further consideration of the significance of risk. The original table is
As variables are ranked from 0 to 3 based on their relevance, not limited to 20 variables, but a summary was selected to
a Degree of Belief in the significance of these variables is derived by demonstrate the use of this tool. In the final stage, the DoB (Degree
summation of the relevance factors divided by the maximum of belief that a variable is associated with significant environmental
number of points. According to this, the significance of a variable is impact) was calculated considering the factors defined by the group
classified as negligible, acceptable, moderate, substantial or extreme. of stakeholders.
This classification allows the variable ranking based on significance. Table 6 shows the classification of activities, was generated
The maximum relevance factors summation is 15 points (Public based on DoB the might cause significant environmental impact,
perception (3); Extent (3); Frequency (3); Intensity (3); Legislation employing the weights assigned for the method. The final score was
(1); Interested party (1) and Detectability (1)). The DoB of the converted into a linguistic term of the level of significance of the
significance is shown in %. environmental impact. In this study pilot site, variables with levels
The general DoB (Degree of Belief) of the significance of each of significance “substantial” or “extreme” were selected by the GMP
variable was given by: team and local partners to be addressed in an action plan involving
P the training of miners.
Relevance factors As demonstrated in Table 6, the following activities were iden-
DoB significance ¼ P  100
of Max Relevance Factors tified to have a “substantial” or “extreme” adverse environmental
impact: vegetation removal, excavation with hydraulic monitors,
The classification of the DoBs is interpreted in intervals of 20%, water catchment for using in sluice boxes, mercury use, manual
producing the following scale for levels of significance of the amalgamation, amalgam burning without retorts. The following
results: Negligible: 0e20%; Acceptable: 21e40%; Moderate: activities were identified to have a beneficial “substantial” or
41e60%; Substantial: 61e80%; Extreme: 81e100%. This is called “extreme” environmental impact: the use of retorts, mercury
linguistic defuzzification, as mentioned by Veiga and Meech (1995). reactivation, backfilling pools for amalgamation, backfilling old pits.
Therefore, in order to evaluate the significance of a potential Once the most significant activities were identified, a training
environmental impact of a hypothetical activity, the user of this program addressing the critical variables was prepared, and aimed
heuristic method takes into consideration the following questions: to mitigate adverse impacts and to promote beneficial impacts. The
Is the potential environmental impact of this activity perceived by GMP training implemented in this area involved approximately
community as high? Is the frequency of this activity high 4200 artisanal miners between 2006 and 2008 and this matrix was
(comparatively with other activities in ASGM in the same area)? the basis for the scope of the training program. In addition to the
Are the extent (affected area) and intensity high (comparatively environmental variables, training also addressed steps necessary
with other activities in the same area)? Does it affect directly for the formalization of “garimpos”, the improvement of gold
interested parties, such as neighbours? Is this a legal requirement recovery, the optimization of costs, and the improvement of health
(are there laws regulating the variable)? Is such impact easily and sanitation. The most critical variables were made mandatory
detectable (for example, visually)? These considerations lead to the topics in the training program and performance indicators were
level of significance of the potential environmental impact, varying developed to evaluate these variables. Mining sites were pre-
from negligible to extreme, indicating whether the variable evaluated prior to the intervention and 120 days post training.
requires monitoring and controls in order to mitigate its potential A general improvement in environmental performance by 28.8%
impacts. was observed in the 141 sites evaluated, as described in details in

Table 5
ASGM activities and associated environmental aspects and impacts.

Variable (activity) Environmental aspects Environmental impacts

Direct Indirect Adverse Beneficial

1 Surveying Soil removal x Soil physical alteration x
2 Vegetation removal Soil exposure x Depletion of renewable resources x
3 Equipment transportation Soil compression x Soil physical alteration x
4 Generator and motor installation CO2 emission x Air quality alteration x
5 Excavation hydraulic monitors Soil removal x Water siltation x
6 Sluice box installation Soil compression x Soil physical alteration x
7 Water supply Use of water x Depletion of non-renewable resources x
8 Carpet washing Use of water x Depletion of non-renewable resources x
9 Mercury use Mercury use x Water/soil/air contamination x
10 Manual amalgamation Mercury release x Human health alteration x
11 Concentrate gathering Use of water x Depletion of non-renewable resources x
12 Panning pool construction Soil removal x Soil physical alteration x
13 Panning pool operation Mercury release x Water contamination x
14 Amalgam burning without retort Mercury vapour x Human health alteration x
15 Use of retort/mercury recycling Mercury vapour x Human health alteration x
16 Mercury reactivation Mercury release x Human health alteration x
17 Backfilling amalgamation pool Soil removal x Soil physical alteration x
18 Backfilling old pit Soil removal x Soil physical alteration x
19 Backfilled pit accommodation Soil removal x Soil physical alteration x
20 Ploughing refilled pit for revegetation Soil removal x Soil physical alteration x
586 R.N. Sousa et al. / Journal of Cleaner Production 19 (2011) 580e587

Table 6
Classification of activities based on the significance of potential environmental impacts.

# Variable (activity) Relevance factors DoB relevance (%) Significance

Public perception Frequency Extent Intensity Legislation Interested parties Detectability

1 Surveying 2 1 3 2 1 60.0 Moderate
2 Vegetation removal 3 1 2 3 1 1 1 80.0 Substantial
3 Equipment transportation 2 1 1 2 1 46.7 Moderate
4 Generator and motor installation 2 1 2 3 1 60.0 Moderate
5 Excavation with hydraulic monitors 3 3 3 3 1 1 1 100.0 Extreme
6 Sluice box installation 2 1 2 3 1 60.0 Moderate
7 Water supply 3 3 3 3 1 1 1 100.0 Extreme
8 Carpet washing 1 1 1 2 1 40.0 Acceptable
9 Mercury use 3 2 2 3 1 1 1 86.7 Extreme
10 Manual amalgamation 2 2 2 3 1 1 1 86.7 Extreme
11 Concentrate gathering 2 2 2 2 1 60.0 Moderate
12 Panning pool construction 3 2 1 3 1 1 73.3 Substantial
13 Panning pool operation 3 3 1 3 1 1 80.0 Substantial
14 Amalgam burning without retort 3 3 3 3 1 1 1 100.0 Extreme
15 Use of retort/mercury recycling 3 3 3 3 1 1 1 100.0 Extreme
16 Mercury reactivation 3 2 2 3 1 1 80.0 Substantial
17 Backfilling amalgamation pool 3 3 3 3 1 1 93.3 Extreme
18 Backfilling old pit 3 3 3 3 1 1 93.3 Extreme
19 Backfilled pit accommodation 2 3 3 2 1 1 80.0 Substantial
20 Ploughing refilled pit revegetation 2 1 2 2 1 53.3 Moderate

Sousa and Veiga (2007b, 2009). The most successful performance a transparent, participatory, and accurate manner. As this partici-
indicators (such as the use of pools for amalgamation and the patory process involved local stakeholders (GMP team, SEMMA,
confinement of mercury, the use of retorts, the reactivation and miners’ representatives), it significantly reduced the subjectivity
reuse of mercury, the use of barriers to retain tailings preventing and bias of the criteria selected. In addition, it established
the discharge of sediments to local rivers, and proper garbage commitment amongst participant’s through the discussion and
disposal) demonstrated that miners responded well to training, and classification of the importance of mining activities based on the
have adopted cleaner mining practices. DoB of their environmental significance and has lead to the
SEMMA (Secretary of Environment of Itaituba), the local envi- sustainability of cleaner practices through training programs.
ronmental authority in charge of issuing environmental licences to Variables that generated “substantial” or “extreme” DoBs on
the artisanal miners, was one of the partners in this program. environmental impacts (such as vegetation removal, excavation,
SEMMA’s representatives acknowledge their limited resources to mercury use, and manual amalgamation), as demonstrated in this
enforce the existing legal requirements. A local survey demon- study, were addressed in training, monitoring and control; aiming
strated that less than 1% of miners attain the current requisites for to mitigate their adverse impact. Variables with “substantial” or
legalization, but they have been operating without proper licensing “extreme” beneficial environmental impacts (such as mercury
for decades (Sousa and Veiga, 2007a). This situation is the direct reactivation, use of retorts, backfilling pool for amalgamation, and
result of the inappropriateness of the laws, the lack of technical backfilling of pits,) were promoted in the program for 4200 arti-
assistance to miners, lack of training and lack of commitment or sanal miners and their community members.
empowerment of stakeholders towards these issues. One of the In the opinion of an environmental expert, this classification
rare cases of successful environmental licensing in the region, may seem rather obvious. However, it is important to reiterate that
known as “Garimpo Canaan”, used this heuristic approach to justify this was intended for or meant to take into consideration the
its environmental controls in 2007. This area has set a standard for spectrum of the stakeholders involved in discussing the environ-
others in the region to follow. mental aspects and impacts associated with ASGM. The matrix
A scenario where 99% of miners operate without any formal represents the introduction of some environmental concepts by the
controls reinforces the idea that only simple tools have real chances GMP team to people who may not have had any such prior expe-
to be implemented. The matrix has helped the GMP project rience with such concepts.
management team, the local authorities and miner’s representa- Variables with level of significance “acceptable” or “moderate”
tives to discuss environmental issues. This has lead to an identifi- were also considered but the prioritization for actions depended on
cation of priorities requiring address in intervention training. the level of difficulty and costs to implement prevention and
Importantly, this study has highlighted the need for Brazilian mitigation procedures.
regulators to reconsider the current environmental assessment In comparison to stochastic processes, environmental experts
process for issuing environmental licences to garimpos. tend to view heuristic approaches with some disbelief. However, in
the ASGM context, simplification not only is a pre-condition for
4. Conclusion understanding environmental processes and impacts by local
stakeholders, it also represents a real chance for the practical
As in most ASGM sites worldwide, the pilot site of this study application of cleaner mining practices. This matrix has proved to
lacked an environmental management system. The proposed be simple and efficient for ASGM, and it can be easily adapted for
matrix provided an intermediate alternative between the existing use in other sites elsewhere.
scenario (a complete lack of management) and a comprehensive
environmental assessment that is beyond the regional government References
and miners’ capacities. The simplified matrix of environmental
Banchirigah, S.M., Hilson, G., 2010. De-agrarianization, re-agrarianization and local
impacts selected the most significant variables to be monitored economic development: re-orientating livelihoods in African artisanal mining
and controlled in the environmental program for ASGM in communities. Policy Sciences 43 (2), 157e180.
 R.N. Sousa et al. / Journal of Cleaner Production 19 (2011) 580e587 587

Barrow, C.J., 1999. Environmental Management, Principles and Practice. In: Envi- Murray, K., Claassen, M., 1999. Interpretation and evaluation of the US environ-
ronmental Management Series. Routledge, p. 336. mental protection agency ecological risk assessment guidelines. Water South
Brundtland, G., 1987. Our common future: The World Commission on Environment Africa 25 (4), 513e518.
and Development. Oxford University Press, Oxford. NEPA e National Environmental Policy Act, 1969 Available at. http://www.epa.gov/
Covello, V.T., Merkhofer, M.W., 1993. Risk Assessment Methods Approaches for regulations/laws/nepa.html Accessed on Set/110/2009.
Assessing Health and Environmental Risks. Plenum, New York, p. 319. SEMMA e Mining and Environment Secretary of Itaituba Municipality e Brazil.
DNPM (“Departamento Nacional da Producao Mineral”), 2008. Brazilian Mineral 2006. Internal Report.
Yearbook (In Portuguese). http://www.dnpm.gov.br consulted on Apr/20/2009. Sousa, R.N., Veiga, M.M., 2007a. Implementing programs to improve gold recovery
Fairman, R., Mead, C.D., Williams, W.P., 1999. Environmental Risk Assessment e and reduce environmental impacts by artisanal gold mining in Brazil. In: Proc.
Approaches, Experiences and Information Sources. Monitoring and Assessment 31st Annual British Columbia Mine Reclamation Symposium. Squamish, BC,
Research centre, King’s College, London. European Environment Agency e EEA. Sep/2007, pp. 17e20.
Environmental issue report No 4. http://reports.eea.eu.int/GH-07-97-595-EN- Sousa, R.N., Veiga, M.M., 2007b. Final Report on the Global Mercury Project in Brazil.
C2/en/riskindex.html. UNIDO e United Nations Industrial Development Organization, p. 82. Available
Fisher, E., Mwaipopo, R., Mutagwaba, W., Nyange, D., Yaron, G., 2009. "The ladder at. http://www.globalmercuryproject.org/ (accessed 10.07.09.).
that sends us to wealth": artisanal mining and poverty reduction in Tanzania. Sousa, R.N., Veiga, M.M., 2009. Using performance indicators to evaluate an envi-
Resources Policy 34 (1e2), 32e38. ronmental education program in artisanal gold mining communities in the
Glasson, J., Therivel, R., Chadwick, A., 2005. Introduction to Environmental Impact brazilian amazon. AMBIO: A Journal of the Human Environment 38 (1), 40e46.
Assessment. Routledge, London, p. 345. Sousa, R.N., Veiga, M.M., Klein, B., Telmer, K., Gunson, A.J., Bernaudat, L., 2010. Strat-
Global Mercury Project, 2008. http://www.globalmercuryproject.org/ (accessed egies for reducing the environmental impact of reprocessing mercury-contami-
02.02.09.). nated tailings in the artisanal and small-scale gold mining sector: insights from
Hilson, G., 2005. Strengthening artisanal mining research and policy through Tapajos River Basin, Brazil. Journal of Cleaner Production 18, 1757e1766.
baseline census activities. Natural Resources Forum 29 (2), 144e153. Suter II, G.W., 1993. Ecological Risk Assessment. Lewis Publishers. Chelsea, MI. US
Hilson, G., 2006. Abatement of mercury pollution in the small-scale gold mining EPA (1998) Guidelines for Ecological Risk Assessment.
industry: restructuring the policy and research agendas. The Science of the Suter II, G.W., Efroymson, R.A., Sample, B.E., Jones, D.S., 2000. Ecological Risk
Total Environment 362 (1e3), 1e14. Assessment for Contaminated Sites. Lewis Publishers, Boca Raton, FL, p. 438.
Hilson, G., 2010a. Child labour in African artisanal mining communities: experi- UNIDO e United Nations Industrial Development Organization, 2009 Global
ences from Northern Ghana. Development and Change 41 (3), 445e473. Mercury Project Website. http://www.globalmercuryproject.org/ (accessed
Hilson, G., 2010b. “Once a miner, always a miner”: poverty and livelihood diversi- 05.08.09.).
fication in Akwatia, Ghana. Journal of Rural Studies 26 (3), 296e307. USEPA e United States Environmental Protection Agency, 2010. http://www.epa.
Hinton, J., Veiga, M., Veiga, T.C., 2003. Clean artisanal gold mining: a utopian gov/emap2/html/pubs/docs/resdocs/mglossary.html#rr (accessed 20.01.10.).
approach? Journal of Cleaner Production 11, 99e115. Veiga, M.M., Baker, R., 2004. Protocols for Environmental and Health Assessment of
ISO 14001/1996, 1999. International Organization for Standardization. Environ- Mercury Released by Artisanal and Small-scale Gold Miners. GEF/UNDP/UNIDO
mental Management e Environmental Performance Evaluation e Guidelines Global Mercury Project, Vienna, ISBN 92-1-106429-5, p. 289.
Geneve. Veiga, M.M., Hinton, J.J., 2002. Abandoned artisanal gold mines in the Amazon:
Krishnamurti, J., 1956. Education and the Significance of Life London: Victor. a legacy of mercury pollution. Natural Resources Forum 26, 15e26.
Labonne, B., Gilman, J., 1999. Towards Building Sustainable Livelihoods in the Veiga, M.M., Meech, J.A., 1995. HgEx e A heuristic system on mercury pollution in
Artisanal Mining Communities. Paper Presented at the Tripartite Meeting on the Amazon. Water, Air and Soil Pollution 80, 123e132.
Social and Labour Issues in Small-scale Mines. ILO, Geneva. Veiga, M.M., Metcalf, S.M., Baker, R.F., Klein, B., Davis, G., Bamber, A., Siegel, S.,
Leopold, L.B., 1971. A Procedure for Evaluating Environmental Impact. U.S. Singo, P., 2006. Manual for Training Artisanal and Small-Scale Gold Miners. GEF/
Geological Survey, Washington, D.C., p. 29. UNDP/UNIDO, Vienna, Austria, p. 144.
Ministry of Environment, Brazil. 1989. Decree # 97.507/89. Regulation of mercury Vieira, R., 2006. Mercury-free gold mining technologies: possibilities for adoption in
use for gold extraction. Brazilian Official Gazette February,14th, 1989. the guianas. Journal of Cleaner Production 14 (3e4), 448e454.
(In Portuguese). WHO e World Health Organization, 2009. http://www.who.int/en/ (accessed
Morris, Therivel, R., 1996. Methods of Environmental Impact Assessment. UCL Press, 13.09.09.).
London. NationalAcademy Press. Washington, D.C. Zadeh, L.A., 1965. Fuzzy sets. Information and Control 8, 338e353.