Science of the Total Environment 951 (2024) 175761

Contents lists available at ScienceDirect

Science of the Total Environment

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

Remote sensing of artisanal and small-scale mining: A review of scalable

mapping approaches
Ilyas Nursamsi a,*, Stuart R. Phinn a,b , Noam Levin a,b,c , Matthew Scott Luskin a,d,
Laura Jane Sonter a,d
a
School of the Environment, The University of Queensland, St Lucia, Brisbane, QLD, Australia
b
Remote Sensing Research Centre, The University of Queensland, St Lucia, Brisbane, QLD, Australia
c
Department of Geography, Hebrew University of Jerusalem, Mount Scopus Campus, Jerusalem 91905, Israel
d
Centre for Biodiversity and Conservation Science, University of Queensland, Australia

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• It is imperative to upscale Artisanal

Small-scale Mining (ASM) monitoring.
• Disproportionate regional focus limits
global ASM spatial data.
• Data fusion is a promising approach for
accurate large-scale ASM mapping.
• Local community involvement is vital
for on-ground spatial data.
• Integrate cloud processing and partici­
pation approach for large-scale
monitoring.

A R T I C L E I N F O A B S T R A C T

Editor: Martin Drews Artisanal and small-scale mining (ASM) significantly influences the socio-economic development of many low-to-
middle-income countries, albeit sometimes at the expense of environmental and human health. Characterized by
Keywords: its labor-intensive extraction from confined (<5 ha) or peripheral mineral reserves, congregated ASM practices
Artisanal small-scale mining can rival the spatial footprint of industrial mines. The unregulated and informal nature of many ASM activities
Remote sensing
presents monitoring challenges that remote sensing (RS) methods aim to address. While local-scale ASM mapping
Earth observation
has seen success, scaling these methods to regional or global levels remains unclear. We review literature on
Land management
Land cover change monitoring mapping ASM to determine: (1) if studies represent the global distribution and diversity of ASM activities, (2)
how ASM's unique characteristics influence the choice of RS methods, and (3) which RS approaches are the most
accurate and cost-effective. We found current studies disproportionately focused on ASM regions in Africa, which
highlights the need to extend the research to other regions with unique ASM characteristics, such as coal and
sand mining in India and China. The selection of RS approaches is heavily influenced by local ASM contexts, the

* Corresponding author.
E-mail address: [email protected] (I. Nursamsi).

https://doi.org/10.1016/j.scitotenv.2024.175761
Received 13 October 2023; Received in revised form 16 July 2024; Accepted 22 August 2024
Available online 23 August 2024
0048-9697/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

scale of analysis, and resource constraints such as funding for high-resolution imagery and validation data
availability. We argue that accurate regional-scale ASM mapping (>100,000 km2) requires innovative combi­
nations of data and methods to overcome data management and storage challenges. Local community partici­
pation, including miners, is vital for on-ground mapping and monitoring capacity. We outline a research agenda
needed to develop a range of approaches for mapping and monitoring ASM in under-studied regions. By syn­
thesizing effective methods, we provide a foundation for generating accurate and comprehensive spatial data,
addressing the issues of inaccurate and incomplete data that global ASM platforms aim to resolve. This spatial
data can guide policymakers, NGOs, and businesses in making informed decisions and targeted interventions to
improve ASM sector safety, sustainability, and efficiency. Leveraging cloud-based geoprocessing platforms, with
regularly updated global satellite image archives, combined with crowd-sourced on-ground information offers a
potential solution for sustained regional-scale monitoring.

1. Introduction et al. (2022), to detect ASM areas as small as 5 ha, facilitating meth­
odological trials for ASM mapping.
Artisanal and small-scale mining (ASM) represents a labour-intensive The mapping of rudimentary mineshafts associated with lode de­
extraction process from small (<5 ha) or marginal mineral deposits posits, such as gemstones, demands the precision of pricier high spatial-
using rudimentary tools, predominantly occurring in lower to middle- resolution imagery (<5 m pixels), due to their limited surface footprints
income countries (Cooke et al., 2019; Zvarivadza, 2018). While ASM (< 1 ha). While such imagery provides greater accuracy and detail, cost
significantly contributes to socio-economic benefits such as employment and data volume restrictions often confine its use to localized studies
generation, poverty alleviation, and foreign exchange, it is also associ­ (Ahmed et al., 2021; Chirico and DeWitt, 2017). Unoccupied aerial
ated with unregulated environmental degradation, child labor, and systems (UAS) provide even finer resolution, but their limited coverage
localized disruptions with potential regional impacts (Fritz et al., 2018; limits their optimal suitability to project-specific applications or tar­
Rustad et al., 2016; Schwartz et al., 2021; Zvarivadza, 2018). geted sampling strategies. (Carreiras et al., 2012).
One of the central debates surrounding ASM revolves around its There is limited published work successfully mapping ASM beyond
formalization, which involves integrating ASM into the formal economy, the local scales, e.g., for whole countries, continents, or globally. Studies
subjecting it to legal and regulatory frameworks, and providing miners such as Maus et al. (2022) and Tang and Werner (2023) have endeav­
with access to support services and markets (Hilson et al., 2022). Pro­ oured to spatially capture the global extent of various mining activities,
ponents argue that formalization can improve working conditions, including both industrial mining and ASM. However, these datasets do
enhance environmental stewardship, and increase government revenue not comprehensively represent ASM activities and thus hold limited
(Jiménez et al., 2024). Additionally, ASM often serves as a coping utility in many regions.
strategy for poverty alleviation and livelihood diversification (Hilson Scaling-up ASM mapping would enrich our understanding of ASM's
and Maconachie, 2020). In rural areas, ASM and farming are inter­ immediate and cumulative impacts (De Haan et al., 2020; Hilson, 2020;
twined, with ASM providing supplementary income during agricultural Hilson and Maconachie, 2020). In practice, this would require con­
off-seasons and farming acting as a fallback during periods of low verting a local-scale workflow to regional, national, and global scales
mineral prices or resource depletion (Bryceson and Jønsson, 2010). consistently. However, while the RS literature and the case studies to
Policymakers and donors face the challenge of balancing the eco­ date demonstrate that a range of data and techniques can be used to
nomic benefits of ASM with addressing its negative impacts. Effective detect and monitor ASM, the applicability of those specific methods
policies should promote responsible ASM practices, provide alternative tested locally to other locations or their scalability to regional or global
livelihoods, and tackle the root causes of poverty and marginalization levels remains uncertain.
(Hilson, 2024). Conflicts between large-scale mining (LSM) and ASM This paper reviews RS approaches employed for mapping ASM to
often emerge, as LSM operations encroach on ASM areas causing date to provide the basis for then developing a regional (100 − 500 k
displacement and environmental degradation, while ASM activities can km2) to continental (> 500 k km2) scale mapping and monitoring pro­
result in illegal mining within LSM concessions, leading to environ­ grams. We sought to identify areas where additional research could
mental damage and security concerns. (Hilson et al., 2020; Kemp and provide the most valuable insights by filling knowledge gaps, building
Owen, 2019). Therefore, it is important to understand the spatial di­ and understanding the role of local factors in choosing appropriate RS
mensions and geographical distribution of ASM activities. Analyzing approaches enabling the transferability of methods developed in one
spatial data can inform targeted interventions, identify overlaps with study to other contexts and identifying cost-effective methods for further
biodiversity hotspots, and support sustainable initiatives. It also aids in application, and where RS could address current limitations of field-
legalizing and supporting artisanal miners through targeted monetary based approaches to monitoring ASM in some contexts. Specifically,
interventions. Comprehensive spatial data also helps resolve conflicts we were motivated by three research questions:
between LSM and ASM by delineating designated areas, preventing
encroachment, and identifying feasible coexistence zones. This infor­ (1) To what extent do current RS studies represent the geographical
mation supports sustainable and equitable resource-sharing strategies distribution and diversity of ASM activities globally?
and policies, ensuring both sectors thrive without conflicts. (2) How do the unique characteristics of ASM operations and the
To date, a broad range of RS approaches have been applied to detect specific local contexts influence the selection of RS approaches
and monitor ASM in diverse local contexts. Artisanal and small-scale for mapping ASM?
gold mining (ASGM), for instance, typically occurring near river (3) Based on an evaluation of RS approaches employed in ASM
streams, creates distinct surface footprints such as irregular lunar- studies thus far, which methods exhibit the highest levels of ac­
shaped pits, tailings, and pit-lakes spanning areas of <5 ha but some­ curacy and cost-efficiency?
times extending to over 100 ha due to feature merging. These features
are usually detectable by medium-resolution (5–30 m pixels) airborne or 2. Methodology
satellite images, provided the pixel size is less than a quarter of the
mining pit (Jensen, 2009). The global coverages and regular acquisitions 2.1. Collating peer-reviewed research papers
(every 5–16 days) of the Landsat and Sentinels programs offered op­
portunity for researchers, such as Barenblitt et al. (2021) and DeWitt We identified 150 papers using Google Scholar with the following

2
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

key phrases: “artisanal small-scale mining remote sensing”, “artisanal location and the scope and methodology of the studies. This includes
small-scale mapping”, and “small-scale mining identification and map­ ASM commodity, mining or extraction method, the average size of the
ping”, as of October 2022. We identified an additional eight papers in mining site, features used for identification, sensors (type of RS imagery
Scopus and the Web of Science (WoS) using the following query string: or data), and the recognition method (Table 1).
“TI/AB = (‘artisanal’ AND ‘small-scale mining’ AND (‘remote sensing’ We documented the ‘commodities’ mined to identify the predomi­
OR ‘satellite’ OR ‘earth observation’ OR ‘identification’ OR ‘monitoring’ nant resources extracted by ASM. We classified studies that reported the
OR ‘spatial analysis’))”. extraction of more than two commodities into two distinct categories:
Our search string focused on studies using RS as the primary tool for “Metalliferous” for studies reporting the mining of various metals and
detecting and mapping ASM, as our review aims to synthesize literature “Crystalline” for those associated with the mining of crystal resources
specifically on remote sensing techniques. Broader terms like ‘identifi­ (McKenna et al., 2020). Furthermore, we recorded the type of extraction
cation’ or ‘monitoring’ would have included studies not prioritizing method or “mining methods”. This included a broad spectrum of tech­
remote sensing. We excluded papers using GIS modeling, spatial anal­ niques such as open-pit mining, dredging, sluicing, shaft mining, un­
ysis, or other non-RS methods, as they do not meet our objective of derground mining, and a category termed “mixed method.” The mixed
evaluating remote sensing's effectiveness in ASM contexts. method category encompasses studies that report a combination of
The lead author screened titles, abstracts, and full papers, removing extraction techniques for different commodities within a single
duplicates manually. This narrative review focuses on synthesizing investigation.
literature specific to RS applications in ASM detection and mapping. Following the categorization by McKenna et al. (2020), we classified
Given that ASM is loosely defined in the academic literature, for the RS sensors used by reviewed studies into five groups. For studies
initial search, we accepted the definition provided by the authors and employing optical satellite imagery, we classified the sensors as follows:
then followed a three-step process to collate papers for our final review. “EO High” represents Earth observation with high spatial resolution (<5
First, we examined the abstracts of all papers (n = 158) to assess their m); “EO Medium” denotes Earth observation with medium spatial res­
relevance to any ASM detection, mapping, or monitoring attempts using olution (5–30 m); and “EO Low” signifies Earth observation with low
various RS approaches. Only those addressing ASM activities distinct spatial resolution (>30 m). The remaining two sensor categories
from industrial large-scale mining (LSM) and applying RS as a main or encompass “SAR” for satellite synthetic aperture radar and “UAS” for
auxiliary tool were deemed relevant. Second, we thoroughly read all high-resolution imagery obtained via unoccupied aerial systems.
pertinent papers (n = 85) to further select those that (1) employed an RS We classified the average size of ASM sites as either smaller or larger
approach to identify and map ASM sites or proxy indicators of ASM, and than 5 ha. This distinction was based on an estimation of the minimal
(2) defined ASM and outlined a method for mapping ASM distinct from mining site size that can easily be identified, with minimal errors, by
the identification of any LSM activity. Lastly, we applied snowball optical satellite imagery with medium to low spatial resolution using a
literature sampling on the identified papers (n = 55) to incorporate any pixel-based classification technique (Smith et al., 2019). Furthermore,
additional relevant articles meeting the criteria, resulting in 57 papers the “identification” of mining areas was categorized based on whether
which hereafter we refer to as ‘study’. This review may have biases: they were directly discerned from the surface footprint (e.g., mining pit,
language bias from only English papers, selection bias from specific sluicing area, and tailings) or inferred from proxy indicators (e.g.,
criteria, snowballing bias from key studies, and database bias excluding nightlights and water siltation).
niche journals. We also excluded grey literature like technical reports, The “Recognition Method” encompasses both the identification of
white papers, theses, and conference proceedings. surface features (detection) and the categorization of these features
(classification), which we classified into seven categories. “Pixel-based”
methods assess the similarity of pixels for specific land cover types,
2.2. Data compilation and categorization while “Subpixel-based” techniques analyze the blend of land cover
present within each pixel. In contrast, “Object-based” and
We compiled a database to summarize key variables related to the

Table 1
Variables used to categorize the scope and methodology of each study in this review.
Variables

Commodity Mining method Identification feature RS imagery Recognition method

Metalliferous Dredging Mining site EO Low (> 30 m) Advanced classification

EO Medium (5–30 m) techniques
– Nickel – Underwater excavation of placer deposit. – Mining pit EO High (< 5 m) Digitization
– Tin Sluicing – Mining sluicing area SAR (Synthetic Aperture Interferometry
– Iron – Tailing facilities Radar) Pixel-based
– Gold – Gold-bearing gravels are washed down by sprayed jets of – Mining dredging area UAS (Unoccupied Aerial Subpixel-based
– Copper water. Proxy indicator System) Segmentation-based
– Coltan Shaft Object-based (GEOBIA)
– Manganese – Nightlights
– Bauxite – Ores and minerals were excavated using vertical shafts. – Greenness and moisture
Underground index
Crystalline – Water siltation
– Extraction of ores and minerals from underground deposit – Backscatter changes
– Amethyst veins. – Surface soil displacement
– Diamond Open-pit
– Cassiterite
– Monazite – Extract ores and minerals from an open pit after top-soil
– Tantalite removal.
– Tourmaline Mixed mining methods
Crocidolite
Cobalt
Emerald
Limestone

3
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

“Segmentation-based” approaches focus on detecting objects and seg­ the accuracy of the methods were not considered (Lyons et al., 2018;
ments rather than individual pixels. “Interferometry” is applied to SAR Morales-Barquero et al., 2019).
images. Studies that manually mapped and digitized ASM from RS im­
agery without utilizing the aforementioned methods were categorized as 3. Results and discussion
“Digitization.” The term “Advanced classification techniques” refers to
detection and classification methods that employ more complex algo­ 3.1. Geographical distribution and commodity focus of RS-based ASM
rithms, such as deep learning models (e.g., Convolutional Neural mapping
Networks).
Our analysis shows a significant research focus on RS-based ASM
mapping in central-western African countries, specifically Congo and
2.3. Data analysis
Ghana, with gold, cassiterite, and diamonds as the main mined com­
modities (e.g., Barenblitt et al. (2021), Kranz et al. (2017), and Snapir
To assess the global distribution of ASM mapping studies (Question
et al. (2017)). This observation is substantiated by the 15 and 19 pub­
1), we extracted the country of origin for each paper based on the re­
lications generated in these countries on the subject, underscoring the
ported study location. We subsequently mapped the distribution of
growing interest in ASM monitoring within the region (Fig. 1). However,
studies using QGIS (ver. 3.20.0) and the ASM-dependent population
of the 72 countries with a large ASM-dependent population (Dorner
map derived from Dorner et al. (2012). We quantified the count of each
et al., 2012), only 15 countries have been the subject of active, peer-
categorical variable and compiled the reported accuracy results. The
reviewed RS-based studies aimed at monitoring ASM.
average size of ASM target studies was calculated from the average
Consequently, the current RS efforts inadequately reflect the global
extent of mining sites reported in the papers to elucidate the impact of
geographical distribution of ASM activities, given the considerable
satellite image selection on the extent of detectable ASM. We used the
research gap for countries with a large ASM-dependent population, such
“ggalluvial” package (Brunson and Read, 2020) to create alluvial charts,
as the Central African Republic (CAR), Niger, Sierra Leone, and
illustrating the proportions and interconnectedness of data within and
Zimbabwe (Fig. 2). The call to extend RS to these lesser-studied regions,
between variables (Question 2).
rich in minerals such as nickel, copper, and diamond, is clear (Hilson
Furthermore, we compiled the accuracy levels reported in the
et al., 2019; Jaillon and De Brier, 2019; Maconachie, 2009). Such
studies, sorted them into the “recognition method” categories and the RS
diversification would not only aim in an effort to manage this sector for
sensors used, and then calculated minimum, maximum, and average
environmental and social benefits, but it would also offer a much more
values (Question 3). It is essential to note that we collected accuracy
holistic global view of ASM. Enhanced datasets are instrumental in
levels reported by authors from the collated studies, each employing
devising effective policies (McQuilken and Hilson, 2016).
different validation approaches, and then averaged these values within
Several factors, ranging from site accessibility and data availability
their respective ‘recognition method’ groups. However, data training
to policy priorities, shape this research bias. Remote regions with ASM
and test distribution, specific environmental, local biophysical factors,
activities pose significant research challenges, due to infrastructure
particular ASM types, and other site-specific factors that might influence

Fig. 1. Geographic distribution of the studies, percentage population that depends on ASM (World-Bank-Group, 2019), and the example of ASM surface footprint
(Google Earth 2023 imageries). A. Small-scale gold mining in Amazon Forest Brazil; B. Crocidolite mining in Prieska town, South Africa; C. Small-scale gold mining in
West-Kalimantan, Indonesia.

4
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

Fig. 2. A schematic representation depicts the relationship between selected countries' population dependency on ASM and the intensity of ASM RS studies, ranging
from nonexistence to the highest recorded number of peer-reviewed studies. The axis represents a spectrum with non-quantitative units. The countries featured in this
figure address both legal and illicit ASM concerns. Symbols denote regional categorization: ( ) represents African nations, (− ) corresponds to Asian nations, and (+)
signifies American nations.

deficits, security threats, and logistical hurdles, hampering field imagery data also varies regionally, with certain areas having better
research and vital ground data acquisition (Wagner and Hunter, 2020). coverage. For example, high cloud coverage and frequency in wet-
Furthermore, the availability of publicly accessible, open-access satellite tropical regions limit usable optical satellite imagery availability

Fig. 3. Stacked bar charts of the publication numbers for each year, according to the mined commodity.

5
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

compared to drier environments (Laborde et al., 2017). g., Kumi-Boateng and Stemn (2020) and Elmes et al. (2014)). These
Policy priorities and funding allocation also influence research focus activities often occur near rivers, producing distinct surface footprints,
in ASM mapping. Certain countries or regions may receive more atten­ such as interconnected irregular pits, tailings, and pit-lakes.
tion due to political, economic, or environmental considerations Contrary to clustered ASM, some mining sites are more dispersed,
(Hilson, 2002; Hilson and McQuilken, 2014). Furthermore, the dearth of covering areas of <5 ha and utilizing less discernible techniques like
knowledge regarding the complexity of ASM-related impacts, lack of panning and dredging. These methods may not be readily identifiable
clarity in the policy framework, and resource limitations contribute to through lower-resolution imagery. Such types of ASM can be mapped
the monitoring approaches deployed by the countries hosting ASM using high-resolution imagery, employing object-based, segmentation-
problems (Hentschel et al., 2002; Miserendino et al., 2013). These fac­ based, and digitization approaches (e.g., Aggrey et al. (2021), Bona et al.
tors collectively contribute to the observed biases in RS-based ASM (2018), and Janse van Rensburg and Kemp (2022)).
mapping, highlighting the need for a more comprehensive and inclusive We also found that studies were often conducted in regions with
approach to address the global distribution of ASM activities and their mixed ASM extraction methods, which encompass a range of extraction
associated environmental impacts. techniques for different commodities within a single investigation (e.g.,
Published RS-based mapping studies to date focused on gold as the open-pit mining, sluicing, and panning). These studies exhibit a similar
most extensively investigated commodity, followed by diamond, inclination toward sensor type selection based on the average ASM sizes
cassiterite, and a variety of metalliferous materials (Fig. 3: e.g., Ibrahim within the studied area (e.g., Isidro et al. (2017) and Nyamekye et al.
et al. (2020), Ngom et al. (2020), and Obodai et al. (2019)). However, a (2021)), Sentinel-2 for clustered ASM sites with an extent of >10km2.
comprehensive global picture of ASM commodities, including coal and They focus on identifying easily recognizable features such as lunar-
sand, remains scarce. This bias may arise from the global prevalence of shaped pits, pit-lakes, sluicing machines, or clusters of miner tents in
ASGM, attributed to its simple extraction process and easy market access alluvial areas.
(Schwartz et al., 2021; Seccatore et al., 2014). Broadening RS research, Research on mapping underground ASM, such as cobalt tunnel mines
especially in areas like China and India where coal and sand ASM are in the DRC, is limited to 5 papers. Detecting underground mining fea­
common (Deb et al., 2008; Shen and Gunson, 2006), is crucial. Such tures, such as tunnel openings, excavated materials, semi-permanent
expansion would contribute to determining the suitable RS mapping buildings, and pathways associated with underground mining, re­
strategies for these distinct environmental settings. quires costly EO High imagery with a resolution of <1 m, such as
Worldview-2 and UAS (DeWitt et al., 2022). More cost-effective alter­
natives such as open-access SAR data can be used for surface deforma­
3.2. Context-dependent selection of RS data and recognition methods for tion and subsidence analysis cover (e.g., Ammirati et al. (2020) and
ASM mapping Brown et al. (2020)), overcoming the EO Medium limitations (Moreira
et al., 2013).
The alluvial chart (Fig. 4) illustrates the distribution and inter-
relationships among variables and categories in RS studies of ASM.
Each category, represented by a bar (nodes), represents the proportion 3.3. Accuracy and cost-effectiveness of RS approaches in ASM mapping
of related publications. Our findings show a focus on open-pit ASM,
being the most studied extraction method (27 studies). This type of ASM, Evaluating the effectiveness of RS mapping methods revealed vary­
typically clustered and easily identifiable, is mostly mapped using EO ing accuracy levels, influenced by the mining operation's context and
Medium and well-established pixel-based or sub-pixel-based methods (e. scale. We do note that accuracy estimates for one study may not be

Fig. 4. Alluvial diagram demonstrating the proportions within variables and the correlation between variables of ASM mapping efforts using RS technologies
and methods.

6
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

suitable for comparison with those of another – as the studies used review, employing EO Medium imagery stands out as the most cost-
widely varying types of error and accuracy assessment approaches, with effective option, striking a balance between cost and accuracy
large variations in critical factors such as the number and distribution of (Table 3). This is due to most EO medium images being freely accessible,
validation or reference data (Lyons et al., 2018; Morales-Barquero et al., and expenses may primarily arise from fieldwork or software
2019). The manual digitization method, predominantly utilizing EO procurement.
High, reported the highest average accuracy (97 %; Table 2) and While EO High images provide greater detail and superior accuracy,
appeared most effective for monitoring ASM sites smaller than 5 ha, they come at higher costs, making them more applicable for site-specific
informally regulated mining sites, and even underground mining, or local investigations. Similarly, UAS offers higher resolution than
enabling detailed assessments and precise spatial records (e.g., Ahmed satellite imagery, yet is only optimal for site-specific studies (e.g.,
et al. (2021) and DeWitt et al. (2022)). Object-based and segmentation- Chirico and DeWitt (2017)). High costs notwithstanding, detailed high-
based approaches, also mainly applied to EO High, achieved relatively spatial-resolution data is crucial for handling individual, informally
high average accuracies of 87.6 % and 87.1 %, respectively. These regulated, or conflict-associated ASM sites (Kranz et al., 2016; Schoepfer
methods were capable of mapping small ASM sites (<5 ha) albeit limited et al., 2010).Finally, SAR imagery offers a unique advantage in detecting
to site-specific or local-scale investigations (e.g., Schoepfer et al. (2010) changes in surface topography related to mining activities and remains
and Stoll et al. (2022)). relatively unaffected by cloud coverage (Moreira et al., 2013). However,
The pixel-based approach using EO medium was the most common the cost-effectiveness of acquiring commercially high-resolution SAR
and reliable method for congregated ASM sites (Fig. 5), with an average data should be considered, despite the current availability of cost-free,
reported accuracy of 89.6 % (e.g., Hausermann et al. (2018)). To map in medium-resolution SAR.
forested areas where there is mixed spectral information per pixel, the
sub-pixel-based classification utilizing low- to medium-spatial resolu­ 3.4. Future research directions and RS advances needed to map ASM
tion imagery can produce high accuracy (average of 93.3 % in Asner beyond the local scale
et al. (2013) and Elmes et al. (2014). More advanced classification
techniques, such as deep learning algorithms, have achieved 95.8 % Scaling up ASM mapping requires rigorous methodologies, robust
accuracy for mapping open-pit ASM (e.g., Camalan et al. (2022) and data resources, and advanced processing capabilities. Mapping ASM in
Kimijima et al. (2022)). lesser-explored regions offers a unique opportunity to assess various
The recent use of interferometry from SAR imagery shows promise in methodologies in diverse contexts. Integrating open-source, multi-type
mapping both open-pit and underground ASM through surface defor­ RS data with cloud-based platforms like Google Earth Engine (GEE) is
mation and subsidence analysis (Monserrat et al., 2014). However, crucial due to GEE's exemplary storage and computational capabilities.
interferometry is sensitive to vegetation coverage, temporal decorrela­ This approach provides access to global, consistently collected datasets.
tion, and temporal phase aliasing, affecting the accuracy of this Enhancing these datasets requires building local capacity, particularly in
approach and restricting its applicability to specific environmental collecting ground-validated ASM spatial data and interpreting proxy
contexts such as sparsely vegetated areas and not forests (Manconi, data as foundational or supplementary resources.
2019; Rocca et al., 2000). The effectiveness of ASM mapping beyond the local scale is con­
Taken together, the selection of an appropriate recognition method strained by the quality of satellite imagery. Common challenges with
for ASM mapping depends on the specific characteristics of ASM activ­ multispectral products include cloud cover and forest canopy obstruc­
ities (size, above or below ground), available data sources, context tions (Isidro et al., 2017; Lobo et al., 2016). While SAR images are less
(forests versus sparse vegetation), and the spatial and temporal detail affected by these issues, speckle issues can compromise classification
required. There is a diverse range of approaches with strengths and accuracies (Maghsoudi et al., 2012). The integration of multispectral
limitations, and decisions also depend on skills capacity and computa­ and SAR datasets, known as data fusion, can offset the limitations of
tional resources. each type individually. While multispectral-SAR data fusion offers the
In the context of mapping ASM using RS, cost-effectiveness analysis potential for enhanced ASM mapping accuracy, challenges related to
illuminates the trade-offs between the detail of maps, spatial resolution, data storage and processing persist (Moomen et al., 2022). Nevertheless,
and financial requirements. Among the methods examined in this leveraging cloud-based geoprocessing platforms may mitigate these
challenges (Mutanga and Kumar, 2019). This combined approach pro­
vides a promising, cost-effective solution for efficient and sustained ASM
Table 2 monitoring at a regional scale, especially vital for developing nations
Summary of overall map accuracies taken from studies in the review that addressing ASM-related issues.
quantified overall map accuracies. Effective ASM monitoring also requires ample training and test data,
Recognition Sensor (Avg[min-Max]: N) acknowledging that data collection is often labour-intensive and
method resource-demanding. When ground-checked data is scarce, alternative
EO Low EO EO High SAR UAS avenues to source reliable spatial data are needed. These options may
Medium include (1) manual interpretation of free EO High imagery like Google
Advanced Earth, (2) utilization of global digitized mining datasets, such as Allan
Classification 95.8 79 et al. (2023), Maus et al. (2022), or Tang and Werner (2023), or (3)
Techniques [92–99.6]:2 [76–82.1]:2 reports from governmental and non-governmental organizations doc­
94.6 umenting confirmed ASM sites. Additionally, diving deeper into proxy
data sources—from RS-analyzed social media feeds to citizen science
Digitization 93:1 [83.3–99]:8
85.1
Interferometry [83–89.6]:4 insights and environmental markers like tainted vegetation, soil anom­
87.2 alies, and nightlight patterns—can shed light on the more informal ASM
Object-based 88:1 [84–93]:8 96.5:1 operations although with some limitations (e.g., Kyba et al. (2019);
92 89.6 95
Kimijima et al. (2021); Levin et al. (2020); Kimijima et al. (2022); and
Pixel-based [85.1–99]:2 [72.6–98]:2 [92–98]:2 90.9:1
Segmentation- 87 (Allan et al., 2023)).
based [85–89]:2 Furthermore, to enhance training and validation efforts, we propose
93.3 establishing localized training programs in collaboration with local
Subpixel-based 86:1 [87–99]:2 universities and research institutions. These programs will build ca­
pacity in remote sensing techniques specific to ASM detection and

7
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

Fig. 5. Illustration of surface features in a congregated artisanal small-scale gold mining area in Kalimantan, Indonesia, depicted using multispectral Google and
Sentinel-2 imagery (5–4-3 bands composite; 10 m pixel resolution), as well as Sentinel-1 SAR with VV (vertical-vertical) and VH (vertical-horizontal) polarizations
(10 m pixel resolution).

monitoring adapted to local conditions and needs. In regions with sig­ spatial data emphasizes collaboration and knowledge-sharing platforms
nificant security and logistical challenges, such as Central African like DELVE (https://delvedatabase.org/) for disseminating spatial data
Repulic and Niger, remote training modules and virtual collaboration and best practices. These platforms facilitate information exchange
tools can provide ongoing support and capacity building, including among stakeholders, supporting effective policy development and
online courses, webinars, and workshops. Partnering with local NGOs interventions.
already working in these regions can support and facilitate on-the-
ground implementation. 4. Conclusion
Transitioning from research to application, local governments
should proactively incorporate RS approaches into ASM mapping and ASM mapping encounters challenges due to the site-specific nature of
monitoring strategies rather than relying solely on external research. We ASM issues, including the diverse nature of this activity, and limitations
encourage personnel from mining regulatory agencies or related public to map and monitor it with RS. However, opportunities for rapid studies
sectors to invest strategically in enhancing their RS expertise. Collabo­ persist in regions heavily dependent on ASM and where extensive
ration with the scientific community and local universities to foster extraction of ASM commodities occurs. Adapting RS techniques to the
capacity-building and consultation efforts may also yield benefits unique features of ASM, whether they are congregated or dispersed
(Moomen et al., 2019). International organizations and funding agencies informal mining sites, using appropriate imagery and analysis methods,
should actively engage in discussions and dissemination of RS oppor­ can significantly enhance detection and monitoring capabilities.
tunities, especially for monitoring ASM activities in remote regions Nevertheless, it is essential to emphasize that further research is still
(Moomen et al., 2022). Strengthening local stakeholder capacity, needed to address existing knowledge gaps in mapping underground
including researchers, officials, community representatives, and miners, ASM and to explore more cost-effective mapping approaches.
is crucial for a participatory approach in ASM monitoring. Organizing Another pivotal aspect in scaling up ASM mapping is the intricate
regional workshops and conferences in safer neighboring countries, like balance between the level of detail, spatial resolution, and budgetary
Ghana or Senegal, and creating regional hubs with existing centers of constraints. Cloud-based geoprocessing platforms with vast satellite
excellence in remote sensing and mining research will foster cross-sector archives emerge as promising tools for scalable ASM mapping. Addi­
partnerships. Formal collaborations with international research orga­ tionally, the fusion of multispectral and imaging radar data can leverage
nizations and leveraging China's technological infrastructure through the strengths of both RS data types while mitigating their individual
partnerships with institutions like the Chinese Academy of Sciences can limitations. Crucially, ensuring a thorough gathering of training and
pilot advanced remote sensing techniques for ASM detection and validation data is critical for dependable analysis. Collaborations with
monitoring, ensuring integration of cutting-edge methodologies. local communities and a deeper grasp of proxy data can bolster data
When extended beyond local scales, spatial data offers new tools to acquisition, especially in data-scarce regions. By crafting bespoke
understand the socioeconomic and environmental complexities of ASM mapping systems that engage local stakeholders and incorporate addi­
activities, providing insights into formalization, cultural shifts, liveli­ tional data such as social media insights, a more inclusive and
hoods, infrastructure, and community welfare (Ang et al., 2023; Hilson partnership-driven ASM mapping approach emerges. Harnessing cloud
et al., 2022). Unlike traditional non-spatial or field-based methodolo­ platforms to amalgamate these elements fosters a cooperative frame­
gies, this approach visualizes connections between causative factors and work, simplifying RS processing available online and enabling regular
mining consequences, as demonstrated by case studies on land-use updates—paving the way for a cooperative regional or even global ASM
conflicts and social acceptance (Malone et al., 2023; Rustad et al., mapping and monitoring initiative.
2016). Integrating various socio-economic and environmental metrics, Accurate mapping of ASM can reveal its true contribution to

8
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

Table 3 development and counter negative narratives about the sector. We

Key characteristics and appropriate mapping applications for selected examples emphasize the importance of standardized remote sensing methods
of active and passive RS imagery in ASM studies. facilitating better data sharing and aggregation among stakeholders for
EO Cost No. of Features ASM mapping more effective collaboration and knowledge exchange. Global platforms
programs Studies application for ASM data, which aim to address issues of inaccurate and incomplete
(sensors) data, can benefit significantly from these improved remote sensing
UAS Very high 3 These products Appropriate only for methods.
(Drone) (Hired services achieve a very high mapping local scale
CRediT authorship contribution statement
start from spatial resolution of mining sites,
USD1,400/ up to centimetres including small sites
100 ha) and provide varied (<5 ha) and
multispectral data underground Ilyas Nursamsi: Conceptualization, Data curation, Methodology,
types depending on mining sites, as Writing – original draft. Stuart R. Phinn: Writing – review & editing.
the attached sensor. well as the Noam Levin: Writing – review & editing. Matthew Scott Luskin:
equipment used
by miners.
Writing – review & editing. Laura Jane Sonter: Supervision, Writing –
Geoeye-1 Very high 2 These products Suitable for local- review & editing.
(USD10.00/ achieve a very high scale mapping to a
Km2) spatial resolution of larger extent than
Declaration of competing interest
40 cm for drones and
panchromatic and potentially lower
1.65 m for cost per square The authors declare that they have no known competing financial
multispectral kilometre. It is interests or personal relationships that could have appeared to influence
imagery, which capable of mapping the work reported in this paper.
includes RGB and very small mining
Near InfraRed-NIR sites (<5 ha),
bands. including Data availability
Additionally, it has underground
a high temporal mining sites. Data will be made available on request.
resolution of 2–9
days and a
sufficient footprint Acknowledgments
width of 15.2 km.
Sentinel-2 Freely available 14 These products have Suitable for regional We thank Jenny Mahuika for the comments that improved the pre­
vious draft.
a medium to coarse to global-scale
spatial resolution mapping, this
ranging from 10 m program is capable
to 60 m and 12 of mapping Funding
bands, including individual mining
the SWIR band, sites with an extent
with a radiometric of >5 ha. This research did not receive any specific grant from funding
resolution of 16 However, it is agencies in the public, commercial, or not-for-profit sectors.
bits. It also offers a incapable of
decent temporal mapping smaller
Consent to participate
resolution of 5 mining sites,
days and can cover including
an area of 170 × underground All authors consented to participate in the research and the elabo­
185 km per scene. mining. The ration of the manuscript.
program has a
higher temporal
resolution than Consent for publication
the Landsat
program, which All authors have approved the contents of the manuscript and agreed
increases the
with its submission.
accuracy of
temporal analysis.
Sentinel-1 Freely available 5 Sentinel-1 is a Sentinel-1 is Ethical approval
Synthetic Aperture capable of regional
Radar (SAR) to global-scale
satellite that mapping and can
Not applicable.
provides high- effectively map
resolution (up to 5 mining sites with an References
m) images with a extent of less than
wide swath (up to 5 ha, including Aggrey, J.J., Ros-Tonen, M.A., Asubonteng, K.O., 2021. Using participatory spatial tools
400 km) and all- underground to unravel community perceptions of land-use dynamics in a mine-expanding
weather, day-and- mining, with a landscape in Ghana. Environ. Manag. 68 (5), 720–737.
night imaging higher temporal Ahmed, A.I., Bryant, R.G., Edwards, D.P., 2021. Where are mines located in sub Saharan
capabilities. resolution than Africa and how have they expanded overtime? Land Degrad. Dev. 32 (1), 112–122.
the Landsat Allan, H., Levin, N., Kark, S., 2023. Quantifying and mapping the human footprint across
program, thereby Earth's coastal areas. Ocean & Coastal Management 236, 106476.
increasing the Ammirati, L., Mondillo, N., Rodas, R.A., Sellers, C., Di Martire, D., 2020. Monitoring land
accuracy of surface deformation associated with gold artisanal mining in the Zaruma City
(Ecuador). Remote Sens. (Basel) 12 (13), 2135.
temporal analysis.
Ang, M.L.E., Owen, J.R., Gibbins, C.N., Lèbre, É., Kemp, D., Saputra, M.R.U.,
Note: The current estimated prices are based on the rates for services found on Everingham, J.-A., Lechner, A.M., 2023. Systematic Review of GIS and Remote
Sensing Applications for Assessing the Socioeconomic Impacts of Mining. J. Environ.
the internet as of 16/05/2023.
Dev. 10704965231190126.
Asner, G.P., Llactayo, W., Tupayachi, R., Luna, E.R., 2013. Elevated rates of gold mining
in the Amazon revealed through high-resolution monitoring. Proc. Natl. Acad. Sci.
110 (46), 18454–18459.

9
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

Barenblitt, A., Payton, A., Lagomasino, D., Fatoyinbo, L., Asare, K., Aidoo, K., Pigott, H., Janse van Rensburg, G., Kemp, J., 2022. The use of C-band and X-band SAR with
Som, C.K., Smeets, L., Seidu, O., 2021. The large footprint of small-scale artisanal machine learning for detecting small-scale mining. Remote Sens. (Basel) 14 (4), 977.
gold mining in Ghana. Sci. Total Environ. 781, 146644. Jensen, J.R., 2009. Remote Sensing of the Environment: An Earth Resource Perspective
Bona, D.S., Arymurthy, A.M., Mursanto, P., 2018. Classification of Limestone Mining Site 2/e. Pearson Education India.
Using Multi-Sensor Remote Sensing Data and OBIA Approach a Case Study: Biak Jiménez, A.D., Smith, N.M., Holley, E.A., 2024. Towards sustainable ASM-based
Island, Papua. 2018 International Conference on Advanced Computer Science and livelihoods: the role of institutional arrangements in the formalization of artisanal
Information Systems (ICACSIS). and small-scale mining. Resources Policy 94, 105129.
Brown, C., Daniels, A., Boyd, D.S., Sowter, A., Foody, G., Kara, S., 2020. Investigating the Kemp, D., Owen, J.R., 2019. Characterising the interface between large and small-scale
potential of radar interferometry for monitoring rural artisanal cobalt mines in the mining. The Extractive Industries and Society 6 (4), 1091–1100.
Democratic Republic of the Congo. Sustainability 12 (23), 9834. Kimijima, S., Sakakibara, M., Nagai, M., 2021. Detection of artisanal and small-scale gold
Brunson, J.C., Read, Q.D., 2020. Alluvial Plots in ggplot2. In: Cran R-Project. mining activities and their transformation using earth observation, nighttime light,
Bryceson, D.F., Jønsson, J.B., 2010. Gold digging careers in rural East Africa: small-scale and precipitation data. Int. J. Environ. Res. Public Health 18 (20), 10954.
miners’ livelihood choices. World Dev. 38 (3), 379–392. Kimijima, S., Sakakibara, M., Nagai, M., 2022. Investigation of long-term roving artisanal
Camalan, S., Cui, K., Pauca, V.P., Alqahtani, S., Silman, M., Chan, R., Plemmons, R.J., and small-scale gold mining activities using time-series Sentinel-1 and global surface
Dethier, E.N., Fernandez, L.E., Lutz, D.A., 2022. Change detection of amazonian water datasets. Int. J. Environ. Res. Public Health 19 (9), 5530.
alluvial gold mining using deep learning and sentinel-2 imagery. Remote Sens. Kranz, O., Schoepfer, E., Spröhnle, K., Lang, S., 2016. Object-based image analysis for the
(Basel) 14 (7), 1746. assessment of mineral extraction in conflict regions–a case study in the Democratic
Carreiras, J.M., Vasconcelos, M.J., Lucas, R.M., 2012. Understanding the relationship Republic of the Congo. Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.
between aboveground biomass and ALOS PALSAR data in the forests of Guinea- 891–896.
Bissau (West Africa). Remote Sens. Environ. 121, 426–442. Kranz, O., Lang, S., Schoepfer, E., 2017. 2.5 D change detection from satellite imagery to
Chirico, P.G., DeWitt, J.D., 2017. Mapping informal small-scale mining features in a monitor small-scale mining activities in the Democratic Republic of the Congo. Int. J.
data-sparse tropical environment with a small UAS. Journal of Unmanned Vehicle Appl. Earth Obs. Geoinf. 61, 81–91.
Systems 5 (3), 69–91. Kumi-Boateng, B., Stemn, E., 2020. Spatial analysis of artisanal and small-scale Mining in
Cooke, A.J., Panayides, T., Tarras-Wahlberg, H., Levin-Nally, E., Gomez, B.R., De the Tarkwa-Nsuaem Municipality of Ghana. Ghana Mining Journal 20 (1), 66–74.
Vries, P., Stacey, J.R., Ocker, C., Gronwald, V., 2019. Forest-smart mining: Kyba, C.C., Giuliani, G., Franziskakis, F., Tockner, K., Lacroix, P., 2019. Artisanal and
identifying good and bad practices and policy responses for artisanal and small-scale small-scale mining sites in the Democratic Republic of the Congo are not associated
mining in forest landscapes, 1, 22–30. with nighttime light emissions. J 2 (2), 152–161.
De Haan, J., Dales, K., McQuilken, J., 2020. Mapping artisanal and small-scale mining to Laborde, H., Douzal, V., Ruiz Piña, H.A., Morand, S., Cornu, J.-F., 2017. Landsat-8 cloud-
the Sustainable Development Goals. Newark DE: University of Delaware (Minerals, free observations in wet tropical areas: a case study in South East Asia. Remote Sens.
Materials and Society program in partnership with PACT). Available online via Lett. 8 (6), 537–546.
http://www.pactworld.org and https://sites.udel.edu/ceoe-mms/.(Accessed 10th Levin, N., Kyba, C.C., Zhang, Q., de Miguel, A.S., Román, M.O., Li, X., Portnov, B.A.,
January 2021). Molthan, A.L., Jechow, A., Miller, S.D., 2020. Remote sensing of night lights: a
Deb, M., Tiwari, G., Lahiri-Dutt, K., 2008. Artisanal and small scale mining in India: review and an outlook for the future. Remote Sens. Environ. 237, 111443.
selected studies and an overview of the issues. Int. J. Min. Reclam. Environ. 22 (3), Lobo, F., Costa, M., Novo, E., Telmer, K., 2016. Distribution of artisanal and small-scale
194–209. gold mining in the Tapajós River Basin (Brazilian Amazon) over the past 40 years
DeWitt, J.D., Chirico, P.G., O’Pry, K.E., Bergstresser, S.E., 2022. Mapping the extent and and relationship with water siltation. Remote Sens. (Basel) 8 (7). https://doi.org/
methods of small-scale emerald mining in the Panjshir Valley, Afghanistan. Geocarto 10.3390/rs8070579.
International 37 (1), 246–267. Lyons, M.B., Keith, D.A., Phinn, S.R., Mason, T.J., Elith, J., 2018. A comparison of
Dorner, U., Franken, G., Liedtke, M., Sievers, H., 2012. Artisanal and small-scale mining resampling methods for remote sensing classification and accuracy assessment.
(ASM). URL: http://www.polinares.eu/docs/d2-1/polinares_wp2_chapter7.pdf (дата Remote Sens. Environ. 208, 145–153.
обращения 01.10. 2012). Maconachie, R., 2009. Diamonds, governance and ‘local’ development in post-conflict
Elmes, A., Yarlequé Ipanaqué, J.G., Rogan, J., Cuba, N., Bebbington, A., 2014. Mapping Sierra Leone: lessons for artisanal and small-scale mining in sub-Saharan Africa?
licit and illicit mining activity in the Madre de Dios region of Peru. Remote Sensing Resources Policy 34 (1–2), 71–79.
Letters 5 (10), 882–891. Maghsoudi, Y., Collins, M.J., Leckie, D., 2012. Speckle reduction for the forest mapping
Fritz, M., McQuilken, J., Collins, N., Weldegiorgis, F., 2018. Global Trends in Artisanal analysis of multi-temporal Radarsat-1 images. Int. J. Remote Sens. 33 (5),
and Small-Scale Mining (ASM): A Review of Key Numbers and Issues. International 1349–1359.
Institute for Sustainable Development, Winnipeg, MB, Canada. Malone, A., Smith, N.M., Zeballos, E.Z., Aquino, R.Q., Huamaní, U.T., Soncco, J.M.G.,
Hausermann, H., Ferring, D., Atosona, B., Mentz, G., Amankwah, R., Chang, A., Salas, G., Coaquira, Z.M., Bedoya, J.H., 2023. Analyzing a deadly confrontation to
Hartfield, K., Effah, E., Asuamah, G.Y., Mansell, C., 2018. Land-grabbing, land-use understand the roots of conflict in artisanal and small-scale mining: a case study
transformation and social differentiation: deconstructing “small-scale” in Ghana’s from Arequipa, Peru. The Extractive Industries and Society 15, 101274.
recent gold rush. World Dev. 108, 103–114. Manconi, A., 2019. Limitations on the use of space borne differential SAR interferometry
Hentschel, T., Hruschka, F., Priester, M., 2002. Global report on artisanal and small-scale for systematic monitoring and failure forecast of alpine landslides. EarthArXiv 1–20.
mining. Report commissioned by the Mining, Minerals and Sustainable Development Maus, V., Giljum, S., da Silva, D.M., Gutschlhofer, J., da Rosa, R.P., Luckeneder, S.,
of the International Institute for Environment and Development. Download from. Gass, S.L., Lieber, M., McCallum, I., 2022. An update on global mining land use.
http://www.iied.org/mmsd/mmsd_pdfs/asm_global_report_draft_jan02.pdf. on, 20 Scientific data 9 (1), 1–11.
(08), 2008. McKenna, P.B., Lechner, A.M., Phinn, S., Erskine, P.D., 2020. Remote sensing of mine site
Hilson, G., 2002. Small-scale mining and its socio-economic impact in developing rehabilitation for ecological outcomes: a global systematic review. Remote Sens.
countries. Nat. Res. Forum. 26, 3–13. (Basel) 12 (21), 3535.
Hilson, G., 2020. ‘Formalization bubbles’: a blueprint for sustainable artisanal and small- McQuilken, J., Hilson, G., 2016. Artisanal and Small-Scale Gold Mining in Ghana.
scale mining (ASM) in sub-Saharan Africa. The Extractive Industries and Society 7 Evidence to Inform an ‘Action Dialogue’. IIED, London.
(4), 1624–1638. Miserendino, R.A., Bergquist, B.A., Adler, S.E., Guimarães, J.R.D., Lees, P.S., Niquen, W.,
Hilson, G., 2024. Formalizing artisanal and small-scale gold mining in sub-Saharan Velasquez-López, P.C., Veiga, M.M., 2013. Challenges to measuring, monitoring, and
Africa: Reviving a deteriorating policy dialogue. In: Routledge Handbook of Natural addressing the cumulative impacts of artisanal and small-scale gold mining in
Resource Governance in Africa. Routledge, pp. 257–274. Ecuador. Resources Policy 38 (4), 713–722.
Hilson, G., Maconachie, R., 2020. Artisanal and small-scale mining and the sustainable Monserrat, O., Crosetto, M., Luzi, G., 2014. A review of ground-based SAR interferometry
development goals: opportunities and new directions for sub-Saharan Africa. for deformation measurement. ISPRS J. Photogramm. Remote Sens. 93, 40–48.
Geoforum 111, 125–141. Moomen, A.-W., Jensen, D., Lacroix, P., Bertolotto, M., 2019. Assessing the policy
Hilson, G., McQuilken, J., 2014. Four decades of support for artisanal and small-scale adoption and impact of geoinformation for enhancing sustainable mining in Africa.
mining in sub-Saharan Africa: a critical review. The Extractive Industries and Society J. Clean. Prod. 241, 118361.
1 (1), 104–118. Moomen, A.-W., Lacroix, P., Benvenuti, A., Planque, M., Piller, T., Davis, K., Miranda, M.,
Hilson, G., Goumandakoye, H., Diallo, P., 2019. Formalizing artisanal mining ‘spaces’ in Ibrahim, E., Giuliani, G., 2022. Assessing the applications of earth observation data
rural sub-Saharan Africa: the case of Niger. Land Use Policy 80, 259–268. for monitoring artisanal and small-scale gold mining (ASGM) in developing
Hilson, G., Sauerwein, T., Owen, J., 2020. Large and artisanal scale mine development: countries. Remote Sens. (Basel) 14 (13), 2971.
the case for autonomous co-existence. World Dev. 130, 104919. Morales-Barquero, L., Lyons, M.B., Phinn, S.R., Roelfsema, C.M., 2019. Trends in remote
Hilson, G., Bartels, E., Hu, Y., 2022. Brick by brick, block by block: building a sustainable sensing accuracy assessment approaches in the context of natural resources. Remote
formalization strategy for small-scale gold mining in Ghana. Environ. Sci. Policy Sens. (Basel) 11 (19), 2305.
135, 207–225. Moreira, A., Prats-Iraola, P., Younis, M., Krieger, G., Hajnsek, I., Papathanassiou, K.P.,
Ibrahim, E., Lema, L., Barnabé, P., Lacroix, P., Pirard, E., 2020. Small-scale surface 2013. A tutorial on synthetic aperture radar. IEEE Geoscience and remote sensing
mining of gold placers: detection, mapping, and temporal analysis through the use of magazine 1 (1), 6–43.
free satellite imagery. International Journal of Applied Earth Observation and Mutanga, O., Kumar, L., 2019. Google earth engine applications. MDPI, p. 591.
Geoinformation 93, 102194. Ngom, N.M., Mbaye, M., Baratoux, D., Baratoux, L., Catry, T., Dessay, N., Faye, G.,
Isidro, C., McIntyre, N., Lechner, A., Callow, I., 2017. Applicability of earth observation Sow, E., Delaître, E., 2020. Mapping artisanal and small-scale gold mining in Senegal
for identifying small-scale mining footprints in a wet tropical region. Remote Sens. using Sentinel 2 data. GeoHealth 4 (12) e2020GH000310.
(Basel) 9 (9). https://doi.org/10.3390/rs9090945. Nyamekye, C., Ghansah, B., Agyapong, E., Kwofie, S., 2021. Mapping changes in
Jaillon, A., De Brier, G., 2019. Mapping Artisanal Mining Sites in the Western Central artisanal and small-scale mining (ASM) landscape using machine and deep learning
African Republic. IPIS. November.

10
I. Nursamsi et al. Science of the Total Environment 951 (2024) 175761

algorithms.-a proxy evaluation of the 2017 ban on ASM in Ghana. Environmental Smith, V., Portillo-Quintero, C., Sanchez-Azofeifa, A., Hernandez-Stefanoni, J.L., 2019.
Challenges 3, 100053. Assessing the accuracy of detected breaks in Landsat time series as predictors of
Obodai, J., Adjei, K.A., Odai, S.N., Lumor, M., 2019. Land use/land cover dynamics using small scale deforestation in tropical dry forests of Mexico and Costa Rica. Remote
landsat data in a gold mining basin-the Ankobra, Ghana. Remote Sensing Sens. Environ. 221, 707–721.
Applications: Society and Environment 13, 247–256. Snapir, B., Simms, D.M., Waine, T.W., 2017. Mapping the expansion of galamsey gold
Rocca, F., Prati, C., Monti Guarnieri, A., Ferretti, A., 2000. SAR interferometry and its mines in the cocoa growing area of Ghana using optical remote sensing.
applications. Surv. Geophys. 21, 159–176. International Journal of Applied Earth Observation and Geoinformation 58,
Rustad, S.A., Østby, G., Nordås, R., 2016. Artisanal mining, conflict, and sexual violence 225–233.
in eastern DRC. The Extractive Industries and Society 3 (2), 475–484. Stoll, E., Roopsind, A., Maharaj, G., Velazco, S., Caughlin, T.T., 2022. Detecting gold
Schoepfer, E., Kranz, O., Addink, E., Coillie, F., 2010. Monitoring Natural Resources in mining impacts on insect biodiversity in a tropical mining frontier with SmallSat
Conflict Using an Object-Based Multi-Scale Image Analysis Approach. Proceedings of imagery. Remote Sensing in Ecology and Conservation 8 (3), 379–390.
GEOBIA. Tang, L., Werner, T.T., 2023. Global mining footprint mapped from high-resolution
Schwartz, F.W., Lee, S., Darrah, T.H., 2021. A review of the scope of artisanal and small- satellite imagery. Communications Earth & Environment 4 (1), 134.
scale mining worldwide, poverty, and the associated health impacts. GeoHealth 5 (1) Wagner, L., Hunter, M., 2020. Links between Artisanal and Small-Scale Gold Mining and
e2020GH000325. Organized Crime in Latin America and Africa. Illegal Mining: Organized Crime,
Seccatore, J., Veiga, M., Origliasso, C., Marin, T., De Tomi, G., 2014. An estimation of the Corruption, and Ecocide in a Resource-Scarce World, pp. 77–104.
artisanal small-scale production of gold in the world. Sci. Total Environ. 496, World-Bank-Group, 2019. Forest-Smart Mining: Identifying Good and Bad Practices and
662–667. Policy Responses for Artisanal and Small-Scale Mining in Forest Landscapes. World
Shen, L., Gunson, A.J., 2006. The role of artisanal and small-scale mining in China’s Bank.
economy. J. Clean. Prod. 14 (3–4), 427–435. Zvarivadza, T., 2018. Artisanal and small-scale mining as a challenge and possible
contributor to sustainable development. Resources Policy 56, 49–58.

11