ArticlePDF Available

DAMSAT: An eye in the sky for monitoring tailings dams

Abstract and Figures

During the past decade, there have been a number of catastrophic tailings dam failures. Affordable monitoring systems, as well as methods to assess the risk posed to communities living downstream of these structures, are needed. In recent years the availability and accuracy of remote sensing information has increased, whilst its cost has decreased.. This paper provides an overview of DAMSAT, a web-based system that brings together Earth Observation and other data to help governments and mining companies monitor tailing dams, and estimate the downstream risks they pose. The methods developed are being piloted in Peru at a number of tailings dams, with the overall goal of improving the decision making process and sharing of information with respect to managing these structures. Engagement with Peruvian stakeholders has shown that DAMSAT provides tools that can help government authorities both reduce the risks and increase the sustainability of mining.
Content may be subject to copyright.
DAMSAT: An eye in the sky for monitoring tailings dams
Darren Lumbroso1, Marta Roca Collell1, Gregor Petkovšek1, Mark Davison1, Ye Liu1, Craig
Goff1, Mark Wetton1
1HR Wallingford, Howbery Park, Wallingford, Oxfordshire OX10 8BA, UK
This is a version of an article published in a special issue of the Mine Water
and the Environment Journal on tailings dam in October 2020
The final authenticated version is available online at:
During the past decade, there have been a number of catastrophic tailings dam failures. Affordable
monitoring systems, as well as methods to assess the risk posed to communities living downstream of
these structures, are needed. In recent years the availability and accuracy of remote sensing information
has increased, whilst its cost has decreased.. This paper provides an overview of DAMSAT, a web-
based system that brings together Earth Observation and other data to help governments and mining
companies monitor tailing dams, and estimate the downstream risks they pose. The methods developed
are being piloted in Peru at a number of tailings dams, with the overall goal of improving the decision
making process and sharing of information with respect to managing these structures. Engagement with
Peruvian stakeholders has shown that DAMSAT provides tools that can help government authorities
both reduce the risks and increase the sustainability of mining.
Transparency in the extractive sector is needed to improve governance, reduce corruption, prevent
environmental harm, and capture the benefits of mining within countries whose economies are reliant
on it (Ayee et al. 2011; Condon 2017; Leonard 2017; Sachs and Warner 2001; Palú and Julien 2019).
The origins of the movement towards transparency date back to the late 1990s and the need to put an
end to resource conflict and, what has come to be known as, the “resource curse” (Sachs and Warner
2001). The sudden discovery of precious resources in low income countries can trigger conflict when
the wealth generated is concentrated in the hands of a small number of people (Dale 2019; McDevitt
2017); this has led to violence in a number of countries including Mozambique, Nigeria, Sierra Leone,
and Peru. A review of 52 empirical studies found that industrial mining was frequently linked to increased
poverty in low income countries (Gamu et al. 2015). Over the long term, research has found a strong
relationship between mineral wealth and a decline in democracy in low income countries with fragile
governments (Dale 2019). Much of the solution to the resource curse can be found in increased
transparency and accountability (Dale 2019).
Many governments in low income, and some high income countries, lack the technical expertise and
resources that large mining companies have (Aguirregabiria and Luengo 2016; Franks 2015). This
makes it challenging for environmental regulators and governments to hold mining companies to
account. The wage differences between government and multinational mining corporations, as well as
the industry-wide dependence on specialist environmental engineering consultancies, inhibits
transparency (Franks 2015). This often prevents effective government and third-party evaluation of the
environmental risks posed by mines (De Carvalho 2019; Franks 2015). Governments also face the
challenge of overseeing many mines, often in remote locations, with limited numbers of staff, who are
often less experienced than those employed by the mines they are regulating (Lumbroso et al. 2019).
Large multinational firms often account for the greatest volume of revenue from mining in low income
countries; however, they often do not account for most of the mining footprint. Instead, a multitude of
“junior” mining companies work on small to medium sized sites, operating on exceptionally tight budgets,
with limited reputational risk (Ledwaba and Nhlengetwa 2016). The junior mining sector faces a number
of challenges including access to finance for capital intensive prospective field mining projects, as well
as very high economic and technical risks (Mulaba-Bafubiandi and Singh 2018).
In 2001, the International Commission on Large Dams (ICOLD) criticised the mining industry’s record
on tailings dam failures (ICOLD 2001); however, in the past 20 years the number of significant tailings
dams accidents has remained high. A spate of catastrophic tailings dam failures in recent years have
killed thousands of people, done great environmental harm, and damaged people’s livelihoods
(Lumbroso et al 2019). On 25 Jan. 2019, the Barragem 1 tailings dam, which stored 11.7 million m3 of
tailings at the Mina Cόrrego de Feijão in the Minas Gerais region of Brazil, failed, resulting in the deaths
of about 300 people and the pollution of hundreds of kilometres of river downstream of the site
(Cambridge and Shaw 2019; Vergilio et al. 2020). This was the fifth major failure of a tailings dam in this
region of Brazil in the past 18 years (Cambridge and Shaw 2019). In 2015, the Fundão tailings dam
failed. It was also located in Minas Gerais and resulted in 19 fatalities, pollution of the water supply for
hundreds of thousands of people, as well as a 20% increase in dengue fever in the affected communities
(Agurto-Detzel et al. 2016; Nishijima and Rocha 2020). Planning for this disaster was found to be
inadequate, with no response plan in place, no effective alarm system, and no quantitative or qualitative
description of the worst-case scenarios (De Carvalho 2019).
The Brazilian Government has recently passed legislation that requires regular inspection of tailings
dam, which have been “effectively self-auditing to date and the regulator is recognised as having
inadequate resources to provide oversight of the new regulations given the large stock of at-risk tailings
dams in Minas Gerais” (Cambridge and Shaw 2019). In addition, in many countries, environmental
regulators often do not know where all the tailings dams are located. Recent work by Oxford University,
which used remote sensing data together with machine learning, has to date found three major tailings
dams that were previously unknown to the Brazilian National Audit Office (O’Neill 2019).
There are also thousands of abandoned tailings dams throughout the world. These are often the result
of junior mining operations, which if they do not “strike it rich” before their debt is due, may have no
choice but to file for bankruptcy and abandon their mining operations, leaving a legacy of “orphaned”
tailings dams (Lumbroso et al. 2019). There is also some evidence to suggest that the data on tailings
dam failures is incomplete, with many smaller accidents going unreported for a variety of reasons,
including because mine managers are afraid of taking legal responsibility (Bowker, and Chambers 2016,
2017; Davies 2002;).
To improve the transparency of mining operations, there is a need to reduce the cost and improve the
way in which many tailings dams are monitored, especially where there are thousands of these
structures spread over a wide area. Over the last 10 years, there has been an increase in the coverage
and accuracy of remote sensing data from satellites (UN 2017), a trend that should continue (Marcuccio
et al. 2019). This paper presents a system, known as DAMSAT, which uses a variety of Earth
Observation (EO) data to improve the transparency and accountability of the mining industry with
regards to tailings dam management. DAMSAT is an applied research project, which was funded by a
grant from the U.K. Space Agency’s International Partnership Programme, aims to demonstrate the
advantages of using EO based technologies. A consortium of organisations, named in the
acknowledgements, have collaborated to develop methods based on remote sensing data to monitor
and assess the risks posed by various aspects of tailings dams. These methods are currently being
piloted in Peru. This paper summarizes these methods and how they can be used to improve risk
assessments and tailings dam monitoring.
Types of tailing dams and their potential problems
There are three main ways in which tailings dams are constructed (Fig. 1):
1. Upstream construction This is the most cost effective method, due to the minimal amount of
material needed for the initial construction and subsequent raisings; however, they are unforgiving
structures” vulnerable to failure as a result of any one or combinations of improper design, construction,
and operation (Martin and McRoberts 1999).
2. Downstream construction This method minimises the chance of breaching due to its structural
stability because no wet tailings are stored below the embankment. It is also the most expensive
construction method because it requires a larger area and more material.
3. Centreline construction This is an intermediate solution in terms of costs and volume. In this
method, the centre lines of the embankments coincide as the tailings dam is raised.
Figure 1. The main methods of construction for tailings dams (Adapted from Vick 1990)
The possible problems related to the above structures include damage to their physical or chemical
stability due, for example, to slope failures or seepage. In the case of tailings dams sites, breaching is
also a catastrophic type of failure. If any of the three types of tailings dams shown in Fig. 1, fail they can
release both tailings and contaminated drainage downstream. The release of tailings downstream,
which is likely to have catastrophic impacts can take place in different ways depending on the mode of
failure and the type of structure involved. Figure 2 provides an overview of the main ways in which
tailings dams can fail and provides an overview of what EO-based methods can monitor.
Tailings are deposited in the tailings pond as a slurry and contain a certain amount of water. If released
downstream as a result of a dam failure, the tailings will generate a mudflow. At closed and abandoned
sites, tailings have, in general, a much lower content of water and therefore, if the sites fail, the material
released takes the form of an unsupported mass moving downward. At abandoned sites, the tailings
may not be sealed or have any protection around them and the action of rainfall or runoff over the
deposits can erode the material and mobilize tailings downstream.
Another typical release from tailings sites is contaminated drainage, a mix of water and pollutants, which
constitutes one of the most serious environmental problems in the mining industry. The most common
contaminated drainage is acid mine drainage (AMD), acidic water resulting from the exposure of
sulphides in the tailings to oxygen and water causing the generation of sulphuric acid. Depending on
the characteristics of the tailings though, alkaline contaminated drainage is also possible.
Effective monitoring of tailings storage facilities can result in early warnings of potential failures or the
leaching of pollutants, which can subsequently minimize or eliminate the social, environmental, and
economic losses they produce. Figure 3 shows, as an example, how the development of seepage from
a tailings dam can lead to breaching. By detecting seepage soon after it occurs (e.g. at time T1 in Fig.
3), the amount of damage can be minimized. However, if the seepage is detected later (e.g. at time T2
in Fig. 3) the losses will be greater. If the seepage is not detected, the structure may fail catastrophically
leading to a significant increase in losses. Table 1 summarises the results of monitoring and, depending
on the level of detected alerts, from Level 1 (low) to Level 3 (high), the types of actions that could be
Figure 2. The main tailings dam failure modes, warning signs and what EO-based methods can monitor
(Adapted from Martin and Davies 2000)
Table 1. Examples of early actions that can be taken at tailings storage facilities as a result of effective
Types of Actions
Defect detected at time T1 owing
to monitoring, see Fig. 3
Increase monitoring and surveillance activities taking
more detailed and focussed measurements in
highlighted areas
Change mining operations (e.g. adjusting tailings
deposition location)
Undertake minor repair works if needed
Stop mining production
Reassess stability with more detailed data
Preventative closure of water sources downstream to
avoid contamination risk
Undertake repair works if needed
Defect detected at time T2 owing
to monitoring, see Fig. 3
Activate emergency plans including:
o Alert and evacuation of the population at risk
o Closure/protection of main infrastructure
Figure 3. The effect of early action as a result of effective monitoring in reducing the damage caused
by tailings dams
State of the art remote sensing technology and how it can be used to monitor tailings dams
Most of the recent tailings dam failures could have been prevented, either by improvements in their
design or by better operational management (Bowker and Chambers 2017). However, as stated above,
many regulatory bodies in low income countries do not have the resources to adequately monitor tailings
dams and mining operations. Although it is not possible to determine the actual level of monitoring for
tailings dams by mining companies worldwide, it appears that many of the large tailings dams are
monitored using, at the minimum, regular visual inspections. In most cases, there is also some form of
in-situ instrumentation (e.g. piezometers, drain flow meters; Lumbroso et al. 2019). However, the rate
of failure of tailings dams owned by large mining companies has remained constant, or in some parts of
the world increased in the past 20 years, despite there being no obvious barriers for these firms to use
the best available monitoring techniques (Lumbroso et al. 2019).
There are also a very large number of small, unmonitored operating and non-operating tailings dams
(Lumbroso et al. 2019). In the past 30 years, there has been an increased number of mines operating
on low profit margins by less experienced and less financially stable miners. This poses a difficult
challenge to significantly reducing tailings dam failures; hence, there is an acute need to monitor junior
mining operations (Lumbroso et al. 2019). Recent work has shown that a low-cost tool, which uses EO-
based information, could help improve the capacity of environmental regulators and non-governmental
organisations (NGOs) to monitor tailings by improving their oversight of both large and small mines
(Chetty 2013; Hui et al. 2015; Thomas et al. 2019).
EO-based data provides a means by which monitoring of tailings dams could be improved. EO-based
monitoring methods are unlikely, in the foreseeable future, to make in situ monitoring methods
redundant; however, they do provide regulators and civil society organisations with an eye in the sky via
which the risks posed by tailings dams can be prioritised and, in some cases, predicted in advance,
allowing them to be mitigated. Table 2 provides an overview of the EO-based data that has been used
in DAMSAT, what aspects of tailings dams it can be used to monitor, together with a summary of its
advantages and disadvantages.
Table 2. An overview of Earth Observation (EO) based data that can be used to monitor tailings dams.
Data provided
Use of EO-
technology with
tailings dams
Advantages and disadvantages of
the EO technology
Radar (InSAR)
displacement of the
dam or structure to
an accuracy of a
few mm.
Assessment of
water in the pond.
Digital terrain
model (DTM)
creation 0.5 m
resolution DTMs
with 10 cm vertical
resolution and
approaching 3.5 m
with no ground
Monitoring of
deformation of the
Monitoring of
movement of the
natural ground in key
areas of risk, such as
at the dam
abutments or the
reservoir rim.
Assessment of
tailings and water
areas in the ponds.
Provides a method for a relatively
limited staff to remotely monitor
thousands of points on a tailings dam to
assess if there have been significant
deformations. Allows monitoring of
abandoned, uninstrumented tailings
dams in remote locations to be
Freely available InSAR has limited
accuracy, making it more challenging to
depict “abnormal” displacements.
More accurate, commercially available
InSAR information is relatively
satellite data
identification and
with increasingly
short return
periods combined
with high
resolution of up to
Leaching of tailings
from the structure
indicated by
declining vegetation
health or increasing
iron oxide traces.
Changes in seepage
rates that could
Thousands of tailings dams in remote
locations can be monitored effectively
by a small number of people.
Cloud cover can hamper and limit
image availability. The temporal
resolution of days rather than a few
40 cm from low
earth orbit; optical
data provides
synoptic, accurate
and fresh context
mapping of an
entire site at very
low cost per
square km,
generally from a
single shot.
contaminated land
indicators such as
vegetation die
back and water
discolouration can
be determined
from these.
detection: high,
almost daily revisit
of sites affords
opportunity to
monitor landscape
changes for
detecting and
mapping rapidly
evolving processes
and activities.
indicate internal
erosion issues in the
dam body. Measures
width of the beach
formed as tailings
are deposited, which
keeps the
supernatant pond
away from the dam
and the phreatic
surface low, thus
reducing the risk of
failure. Slope failure
detection, especially
relevant for closed
sites, using a
combination of
parameters (e.g. soil
muddiness, surface
features, slope
hours may also be a constraint for
some users. When used to assess
leaching from tailings dam, the size of
the anomaly needs to bigger than the
pixel resolution of the EO data
GNSS can be
used to
compute three-
motion vectors of a
number of points
using standard in-
situ located GNSS
receivers to an
accuracy of a few
Monitoring of
deformation of key
specific points of the
structure in real time
(requires a base
station to be
GNSS has a high degree of accuracy,
i.e. a few mm, and provides real-time
displacement measurements and
changes in levels.
Expensive to purchase and install.
Capital and installation costs can range
between ≈ U.S. $60,000 and $120,000
per installation. Equipment needs to be
installed in a secure location because it
can be easily vandalised. The GNSS
equipment can only be monitored
where it is installed and requires a Wi-
Fi connection and continuous power
Global weather
Forecasts of
climate variables
such as rainfall
and temperature
from, e.g. the
European Centre
for Medium-range
Forecasts of runoff
that can be used to
assess the
probability of
overtopping, or a
potential dangerous
increase in the
moisture content of
Provides lead times between 10 and 15
days of potential extreme hydrological
events that could affect a tailings dam.
Some weather forecasts are
probabilistic, which allows degree of
uncertainty to be assessed
(ECMWF) or the
Global Forecast
System (GFS)
produced by the
National Centers
for Environmental
Prediction. The
GFS forecasts are
available at ≈50
km or 0.5 degree
resolution. The
ECMWF forecast
are available at
≈18 km or 0.2
degree of
the dam body that
could affect stability.
The results of rainfall-runoff models
driven by weather models have a high
degree of uncertainty, especially in
areas where no data is available to
calibrate them. More accurate weather
forecasts (such as those provided by
ECMWF) are more expensive than
those provided for free by the GFS.
Depending on the variables required,
the cost can be U.S. $10,000 to
$50,000 per annum. Ideally, observed
flows and rainfall data are used to
calibrate the hydrological model.
(Source: Adapted from Lumbroso et al. 2019)
DAMSAT functionality and modules
The DAMSAT system is a project being piloted in Peru. From the commencement of the project, key
Peruvian stakeholders, including the Ministry of the Environment, Ministry of Energy and Mines, the
Agency for Environmental Protection and Enforcement, the National Water Authority, and others were
engaged with to elicit their user requirements related to the monitoring of tailings storage facilities, as
well as helping to design the different modules of the system. Many of these organisations reported that
they have limited resources and are not able to visit tailings dam sites frequently (Lumbroso et al. 2019).
The stakeholders’ requirements for an EO-based monitoring system for tailings dams are summarised
Inform assessments of tailings dam safety and performance in accordance with design criteria and
Assist in identifying and forecasting when there is a high probability of failure
Forecast and identify potential pollution incidents
Effectively communicate the degree of risk posed by tailings dams so that appropriate actions are
implemented at national and local government levels and also at-risk communities
The system contains a number of modules, which together with their sources of information and
functionality, are summarised in Table 3.
Table 3. DAMSAT modules including sources of information and functions
Sources of information
Sentinel 2 satellite data with 10
m resolution
High resolution satellites
available in the area
Allows visual inspections of particular sites.
Compares current image with historical images
to assist users detect interesting changes on
InSAR Sentinel-1 satellite data
InSAR high resolution satellites
available in the area
GNSS in-situ sensors
Provides an indication of displacements at the
structure and in nearby points.
Sentinel 2, with a 10m
High resolution satellites
available in the area
Analysing sequences of optical data for signs of
iron oxide or vegetation health variations which
could indicate locations of potential pollution
National Oceanic and
Atmospheric Administration
(NOAA) global forecast system
ECMWF weather forecast
Provides forecasts up to 10 days into the future
of (1) precipitation (2) runoff into the reservoir (3)
water level change at the pond.
Output from dam break and
inundation models
Characteristics of buildings,
people, and road networks
downstream of the dam, as
well as the output from
inundation model
The emergency management module uses dam
break modelling, inundation modelling and the
agent-based Life Safety Model to estimate the
risk to people, evacuation times, and emergency
management strategies (e.g. the location of safe
havens, improved early warnings), which could
reduce risks to people.
The system runs on a secure cloud-based platform and brings together data from a number of EO data
sources, weather forecasts, and on-the-ground sensors. An internet browser based interface allows
users to view the outputs of the modules and update operational parameters such as alert thresholds.
Figure 4 shows the data flows, processes, objects and stores for the EO-based tools and modules
included within DAMSAT. The sections below provide summaries of the EO methods underpinning each
of the different DAMSAT modules.
Figure 4. An overview of the data flows, processes, objects, stores and modules included within
Monitoring of Surface Movements using Space-borne Synthetic Aperture Radar Interferometry
(InSAR) data
Space-borne Synthetic Aperture Radar Interferometry (InSAR) is increasingly used to monitor wide-area
to local scale ground motions. InSAR is non-intrusive and does not require any ground installations. The
shorter the revisit time of the satellite, the more accurate the analysis of the temporal evolution of the
deformation of a structure is likely to be. The repeat cycle of satellites with InSAR is sufficient to allow
the changes in ground movements of tailings dams to be monitored almost daily at most mining locations
(e.g. Sentinel-1A has a revisit time of 12 days; Tomás et al. 2014).
InSAR data is freely available globally from a number of sources (e.g. Sentinel-1) and has been used to
measure deformations in buildings and opencast mines for over a decade (Herrera et al. 2007).
Commercially available InSAR, such as COSMO-SkyMed, provides more accurate measurements of
surface deformation (e.g. to a few millimetres); however, such sources can be costly. The system has
been designed so that it provides the user with ability to use a range of InSAR information sources. The
rate of change of the deformation of a tailings dam over time, based on InSAR data, can be used to help
to predict if, and in which location, the structure has the highest probability of failing. In addition to ground
movements, InSAR can, in some cases, offer the ability to remotely monitor the extent of larger tailings
pond water levels.
Monitoring of Surface Movements using Global Navigation Satellite System (GNSS) Technology
Global Navigation Satellite System (GNSS) technology provides real-time, highly accurate
measurements of surface deformations in all weather conditions. GNSS technology has been used to
measure deformations of dams in China, the USA, and Spain (Scaioni et al. 2018; Xiao et al. 2019).
DAMSAT has the ability to integrate readings from GNSS. GNSS-based solutions require a base station
to be installed, as well as monitoring antenna and support boxes on the structure at the locations where
information on movements are required. Typical costs of GNSS equipment and installation are between
U.S. $60,000 and $120,000 (Lumbroso et al. 2019). GNSS equipment only provides real time
measurements of deformations at the points on the dam where ground stations are located, with data
being collected every minute, or less if required.
Detection of leaching and pollution
The use of spectral data and remote sensing for detecting minerals and analysing mining impact is well
documented. Alexander et al. (1973) described the use of ERTS-1 (later called Landsat) multispectral
data to evaluate the damage caused by surface mining. The contamination caused by leaching from
tailings dam leakage can be detected using EO data by using the health levels of the vegetation
downstream of the dam or by monitoring the presence of iron oxide in the environment. The method
used in DAMSAT detects anomalies based on changes in the normalised difference vegetation index
(NDVI) or in the soil’s iron oxide content.
The use of changes in EO-based NDVI values has been used to detect pollution from mining operations
(Alonzo et al., 2016; Hejmanowska et al. 2016; Sonwalka et al. 2010). Remote sensing derived data
sets such, as NDVI, indicate the greenness of the vegetation. A significant drop in the vegetation’s NDVI
downstream of a tailings dam can be observed, compared to an average year.
The Sentinel-2 satellite provides information on soil attributes including levels of iron oxide, which has
been shown to have a good correlation with potentially toxic elements being present (Van der Werff and
Van der Meer 2015; Gholizadeh et al 2018). Hence, changes in iron oxide concentrations downstream
of a tailings dam can also indicate that potentially toxic substances are seeping through the dam.
The change detection algorithm is based on a comparison of the image for the selected date with at
most three co-registered previous images, where all images satisfy a cloud condition criterion (i.e. less
than 75% of the image is covered by clouds). For NDVI, there is the additional criterion of the maximum
time span between the analysed and comparison images of three months. This is to prevent change
detection being triggered by seasonal changes in vegetal cover. The reason for using three comparison
images rather than just one is to reduce falsely detecting pollution incidents due to ground or
atmospheric changes similar to that of the target species.
For all images, an intensity corresponding to the targeted vegetation species is first calculated. For
vegetation, this is:
where: NIR is the reflectance in the near infrared band (Band 08 of Sentinel-2) and RED is the
reflectance in the red band (Band 04).
For iron oxide, the respective intensity of the correlation coefficient between the observed spectrum and
spectra of the iron oxide signatures reported in the literature (Seifi et al. 2019), as well as those sampled
from the Brumadinho tailings dam failure, is used. For all pixels not covered by clouds, the difference in
the parameter value between the selected image and each of the comparison images is computed to
obtain a scalar (grey scale) difference image (one for each comparison image). The next step is to run
the principal component analysis on this difference image to extract the pixels where the probability of
change exceeds the pre-determined threshold. (For examples of using this method for unsupervised
change detection in satellite images, see Celik 2009.) If the probability for the same pixel exceeds the
threshold in at least two comparisons (or one if the pixel is covered by clouds), then the algorithm raises
a warning with respect to where changes in NDVI or in the presence of iron oxide has occurred. Changes
in these variables are likely to be “non-natural” if they have an artificial shape and are local in scale.
Hydro-meteorological forecasting
Short term weather forecasts from the European Centre for Medium Range Weather Forecast (ECMWF)
and The National Oceanic and Atmospheric Administration (NOAA) Global Forecasting System (GFS)
are coupled to hydrological models, which allow inflows towards tailing dams to be estimated. Existing
drainage systems to divert runoff from the catchment around the dams are also considered in the
estimates. The inflows and rainfall are also used to estimate the likely increase of water levels at the
pond. Estimations are quantified with a lead time of up to 10 days. The forecast inflows are updated
automatically every six hours, together with the forecast rainfall and water levels in the tailings pond.
These probabilistic forecast are displayed as shown in Fig. 5.
Figure 5. An example of the probabilistic forecasts of rainfall intensity, runoff, and water levels for a
tailings dam site in Peru
Emergency planning
Agent-based models, such as the life safety model (LSM) used in DAMSAT, provide information to
support emergency planning. The LSM models the behaviour of people, cars, and buildings downstream
of a tailings dam and their interaction with the mud flow following a breach. The LSM allows the following
to be estimated:
The number of people likely to die or be injured as a result of a tailings dam failure
The number of buildings that are likely to be destroyed
The time required for people to evacuate the area at risk on foot and in vehicles
The effect of measures such as improving early warnings and evacuation routes on reducing the
risk to people (Di Mauro and Lumbroso 2008; Lumbroso and Davison 2018)
The LSM has been used to analyse the risk to people from dam failures and floods worldwide (Johnstone
2012; Lumbroso and Davison 2018). The model has been validated using data from the 1959 Malpasset
Dam disaster in France and used for emergency planning for flood events worldwide (Lumbroso and
Tagg 2011; Lumbroso et al. 2011). The vulnerability functions for people and buildings in the LSM were
originally developed based on the characteristics of floodwater, but these can be modified to consider
that tailings dam failures produce a mudflow which, unlike floodwater, is non-Newtonian in nature. The
LSM was used to model the Brumadinho tailings dam failure in Brazil that occurred in January 2019
(Lumbroso et al. 2020). The LSM was run for a number of scenarios and estimated that between 216
and 345 died. The actual number of fatalities was about 300 people (Keaveny 2019; Mining Journal
2020). The LSM allows the effects of improved warnings on loss of life and evacuation times to be
assessed. Research on the Brumadinho case study found that even if a warning had been provided just
as the dam failed, the number of fatalities could have been significantly reduced (Lumbroso et al. 2020).
The LSM uses the water or mudflow depths and flow velocities from a two dimensional mud and flow
inundation model as input. To determine the mud flow being released during the failure, a dam breach
model called EMBREA-MUD has been specifically developed for tailings dams (see Petkovšek et al.
2020 for more details). EMBREA-MUD is based on a predictive dam breach model developed over a
number of years (see Mohamed et al. 2002; Morris 2011). This model simultaneously computes the
outflow of water and tailings from a tailings dam and the corresponding growth of the opening. Tailings
outflows are represented by a separate non-Newtonian viscous layer, as well as a water layer. The
tailings dam model was verified using empirical data from laboratory tests, as well as data from the
Mount Polley tailings dam failure in Canada in 2015 and the Merriespruit tailings dam disaster that
occurred in South Africa in 1994 (Petkovšek et al. 2020).
Application of DAMSAT in Peru
The DAMSAT system is currently being piloted in the Cajamarca and Pasco regions of Peru to help
infrastructure and environmental regulators improve their monitoring of both active and inactive tailings
dams. Cajamarca is home to a number of large gold and copper mines, while Pasco is a region with a
long mining tradition. Peru has hundreds of active tailings dams and thousands of abandoned tailing
deposits storing chemicals and waste that have the potential to pollute the downstream environment.
Over the past decade, there have been a number of tailings dam incidents in Peru. Most recently, in
July 2019, it was reported that a spill from a tailings dam at the Cobriza copper mine led to the Mantaro
River being severely polluted with cyanide up to 375 km downstream from the dam (Mining Journal
2019; Petley 2019) and in 2010 a tailings dam failure at Huancavelica polluted some 110 km of the
Escalera and Opamayo Rivers (Minerals Policy Institute 2014).
The impacts of mining on water quality and environmental health originate primarily from AMD and the
escape of ancillary products. In July 2008, Peru declared a state of emergency at a mine near Lima over
concerns that a tailings dam, weakened by seismic activity and subterranean water filtration, could
release arsenic, lead, and cadmium into the main water supply for the capital (Beddington and Williams
2008). It has been estimated that over 13 billion m3 of effluent from mining operations are released
annually into Peru’s water environment (Beddington and Williams 2008). This section provides details
of how some of the modules have been implemented in Peru. The names and exact locations of some
of the examples have not been provided owing to reasons of confidentiality.
Monitoring Tailings Dam Displacements using Sentinel-1 InSAR Data and GNSS Monitoring Equipment
The use of Sentinel-1 InSAR data has been piloted on seven large (i.e. higher than 15 m), tailings dams
in Peru. Figure 6 shows an example of Sentinel-1 InSAR displacement results for one of these dams for
a 15 month period between Aug. 2018 and Oct. 2019.
Figure 6. Example of Sentinel-1 InSAR displacement results for a tailings dam in Peru. Top: Maps of
cumulative line-of-sight (LOS) displacement of a dam and the surroundings between Aug. 2018 and
Oct. 2019. Bottom: Displacement time series graphs for the points within the red and green boxes shown
on the top left map. Negative LOS decrease corresponds to movement in the direction away from the
satellite. This figure contains modified Copernicus Sentinel data
The InSAR results are displayed in several ways on the web-based system. The main method of display
is a map, making it easy for users to visualise areas of the structure that may be problematic with regards
to possible abnormal deformations. A “hotspotter” summarises the information from all the points, which
helps to identify these areas. The increase in the line of sight (LOS) displacement is also plotted against
time (see Fig. 6). Work carried out on the failure of the tailings dam at the Cadia gold mine in New South
Wales, Australia has indicated that using InSAR satellite data together with the inverse of velocity
, the timing of the failure could have been predicted almost one month in advance (Carlà et al
2019; Tre-Altimira 2018).
In addition, GNSS monitoring stations have been installed on a tailings dams in Peru (Fig. 7). The GNSS
monitoring stations provide real-time information on tailings dam movements with a high degree of
accuracy (e.g. a few millimetres); however, they are relatively expensive (see Table 2). The monitoring
equipment comprises a SUMMIT box, which is a computer running proprietary software, combined with
three GPS satellite signal monitoring units, called GNSS monitoring and reference units. The SUMMIT
boxes were installed in buildings on or near the site to protect them from the weather. They require a
continuous supply of power, together with a connection to the internet to allow remote user monitoring.
On each site, one of the GNSS monitoring units was designated as the reference unit, and installed at
a location that does not move relative to the dam. The other two GNSS units, referred to as the GNSS
The inverse velocity method is used to calculate the rate of deformation of the slope of the dam (velocity) and to plot the
inverse of the rate of deformation against time (i.e. the inverse velocity against time). As the velocity or rate of deformation
increases, the inverse will tend towards zero, which is when failure occurs. For more details of this method, see Carlà et al.
monitoring units, were installed at locations where movements are anticipated. The GNSS antennas
were mounted in a way that provides them with a sufficiently clear view of the surrounding sky (Fig. 8).
Figure 7. GNSS monitoring station installed (at the front) and building with the SUMMIT box on the right
on a tailings deposit in Peru
Figure 8. Schematic diagram showing the setup of the GNSS monitoring stations and SUMMIT box
A module has been included in the system that generates warnings when certain deformation or
movement thresholds are reached. The users can manage those alerts, change their status, and make
comments. The system is flexible so that it can be modified for different structures. An example of the
web-based alert system is shown in Fig. 9.
Figure 9. Example of the web-based warning system alerting users to possible issues
Assessment of leaching using optical Imagery
The leaching module uses optical imagery. The two methods by which pollution is detected have been
discussed above: direct pollution detection using iron oxide as a proxy for leaching, and indirect pollution
detection using NDVI to monitor the changes to the health of the vegetation adjacent and downstream
of a tailings dam. The leaching module uses optical data from Sentinel-2 over seven bands and has
been applied to two tailings dams in Peru. An example of its application to a tailings dam in Peru is
shown (Fig. 10).
Emergency planning for tailings dams in Peru
The Peruvian stakeholders are interested in the risks posed by tailings dams, both in terms of loss of
life and other damage, so that their limited resources can be targeted at the structures that pose the
greatest risk. The LSM has been used on a number of tailings dam sites in Peru. The mudflow from a
breach of the dam is estimated using EMBREA-MUD and the resulting mud hydrograph is used to drive
a two dimensional model of the mudflow, the results of which are used in the LSM. Data on the number
of people, buildings, and the road network downstream of the dam are taken from openly available data
or digitised from maps. The LSM model is run for several scenarios, such as:
No warning
Siren based warning at the time of failure
Siren based warning 5 minutes before failure
Siren based warning 15 minutes before failure
The uncertainty in the response of people to hearing the warning siren is assessed by simulating delay
times in the response of people to the warning of 5 and 15 minutes. The results for a specific tailings
dam in the Pasco Region of Peru are summarised in Table 4, showing that the estimated number of
fatalities (around 84% of those at risk) are very high when a warning is not issued. The LSM results
show that a warning issued 10 minutes before the failure can significant reduce the risk to people, if they
know how to respond.
Figure 10. Example of change in NDVI for a tailings dam in Peru on 12 Jan. 2019, based on a
comparison between the first clear day post-incident image and three pre-incident images. The spill
occurred on 16 December 2018 from the left side of the dam.
Table 4. The potential number of fatalities caused by the failure of a dam in the Pasco Region for a
range of warning and response times
Time the warning was issued
Number of fatalities
Five minutes before people
15 minutes before people
No warning issued
Warning issued at the time the
dam fails
5 minutes before failure
10 minutes before failure
15 minutes before failure
The LSM has been used to assess the risk to people at a number of tailings dams in Peru because it
allows regulators to quantify the risk they pose to people and to use the information to develop
emergency evacuation strategies, such as informing those at risk to escape to local safe zones above
the mudflow inundation zone.
Earth observation techniques have allowed stakeholders in Peru to monitor tailings dams and save
valuable time in the decision making process and sharing of information, thereby reducing the risks
posed by tailings storage facilities in the country. Assessment times can be reduced when trying to
detect problems and visual inspection programmes can be more informed using the information provided
by these techniques. Risk-based approaches to inspection and monitoring of tailings dams based on
live EO-based data provide valuable information that complements historical date information, which in
most cases is collected from sporadic site inspections that may be out of date.
There is a consensus amongst Peruvian stakeholders that EO-based methods, whilst not a substitute
for in situ measurements, can be used by Government authorities to improve dam monitoring and reduce
the risks posed by the failures of these structures. The different methods incorporated in the DAMSAT
modules can be applied to tailings or water retaining dams anywhere in the world. The methods
developed have been piloted on several large tailings dams in Peru and it is expected that the services
provided will be expanded to help monitor all the tailings dams in Peru, as well as in other countries.
Low income countries, which are often subject to political instability, as many resource-driven economies
are, could benefit quickly from many of the EO methods developed as part of this work. The work carried
out to date indicates that EO-based monitoring of tailings dams can help decision-making that could
help reduce the risks to people and the environment posed by these structures.
Acknowledgments: The authors acknowledge the work of Dr Eleanor Ainscoe who carried out the
analysis of the InSAR data for tailings dams in Peru. The authors also would like to acknowledge that
this work was made possible as part of the DAMSAT project, which was funded by a grant from the UK
Space Agency’s International Partnership Programme. The team working on the project is led by HR
Wallingford and comprises Telespazio VEGA UK, Siemens Corporate Technology, Smith School of
Enterprise and the Environment at the University of Oxford, Satellite Application Catapult, Oxford Policy
Management, Ciemam SAC, Peruvian National Foundation for Hydraulics Engineering, and School of
Hydraulic Engineering at the National University of Cajamarca. More information on this work is available
Aguirregabiria, V, Luengo A (2016) A micro-econometric dynamic structural model of copper mining decisions.
Available at:
Agurto-Detzel H, Bianchi M, Assumpção M, Schimmel B, Collaço C, Ciardelli JR, Barbosa J, Calhau J (2016) The
tailings dam failure of 5 November 2015 in south-east Brazil and its preceding seismic sequence, Geophys
Res Lett 43: 49294936, doi:10.1002/2016GL069257
Alexander SS, Dein D, Gold DP (1973) The use of ERTS-1 MSS data for mapping strip mines and acid mine
drainage in Pennsylvania. Proc, Symp on Significant Results Obtained from Earth Resources Technology
Satellite-1: Vol 1, Sect A: National Aeronautics and Space Administration SP327, pp 569-575
Alonzo M, Van Den Hoek J, Ahmed N (2016) Capturing coupled riparian and coastal disturbance from industrial
mining using cloud-resilient satellite time series analysis. Sci Rep-UK 6, 35129.
Ayee J, Soreide T Shukla GP; Le TM (2011) The political economy of the mining sector in Ghana, Policy research
working paper no. WPS 5730. World Bank
Beddington, A, Williams, M (2008) Water and mining conflicts in Peru, Mountain R&D Vol 28 No 3/4 AugNov
2008: 190195 doi:10.1659/mrd.1039
Bowker LN, Chambers DM (2016) Root causes of tailings dam overtopping: the economics of risk and
consequence. Proc, 2nd International Seminar on Dam Protection against Overtopping, Colorado State Univ,
Fort Collins
Bowker LN, Chambers DM (2017) In the dark shadow of the supercycle tailings failure risk and public liability reach
all-time highs. Environments 4(4):75; doi:10.3390/environments4040075
Cambridge C, Shaw M (2019) Preliminary reflections on the failure of the Brumadinho tailings dam in January
2019. Dams Reserv 29(3): 113-123
Carlà T, Intrieri E, Di Traglia F, Nolesini T, Gigli G, Casagli N (2017) Guidelines on the use of inverse velocity
method as a tool for setting alarm thresholds and forecasting landslides and structure collapses. Landslides
14(2): 517534,
Carlà, T, Intrieri, E, Raspini, F, Bardi, F, Farina, P, Ferretti, A, Colombo, D, Novali, F, Casagli, N (2019)
Perspectives on the prediction of catastrophic slope failures from satellite InSAR, Sci Rep-UK 9:14137
Celik, T (2009) Unsupervised change detection in satellite images using principal component analysis and k-
means clustering. IEEE Geoscience and Remote Sens Letters, Vol. 6, No. 4, October 2009
Chetty, P. (2013) Monitoring of mine tailings using satellite and lidar data, South African Surveying and Geomatics
Indaba (SASGI) Proceedings 2013 Stream 1:
Condon, M (2017) Citizen science, data transparency and they mining industry Na Res and Env Vol 32, Number
2, Fall 2017
Dale, L (2019) Foiling the resource curse, The Environment Forum, May-June, 2019
Davies, MP (2002) Tailings impoundment failures: are geotechnical engineers listening, Geotechnical News, vol.
20, no. 3, pp3136
De Carvalho, DW (2019) The ore tailings dam rupture disaster in Mariana, Brazil 2015: What we have to learn
from anthropogenic disasters, Natl Resour J, Vol 59, Issue 2, Summer, 2019
Di Mauro, M, Lumbroso, D (2008) Hydrodynamic and loss of life modelling for the 1953 Canvey Island flood,
Proceedings of FLOODrisk 2008, Keble College, University of Oxford, UK, 30 September to 2 October 2008
Franks, DM (2015) Mountain movers: Mining, sustainability and the agents of change, First edition, Routledge
Gamu, J, Le Billon, P, Spiegel, S (2015) Extractive industries and poverty: A review of recent findings and linkage
mechanisms, The Extractive Industries and Society, Vol 2, Issue 1, January 2015, pp162-176
Gholizadeh, A, Saberioon, M, Ben-Dor, E, Borůvka, L (2018): Monitoring of selected soil contaminants using
proximal and remote sensing techniques: Background, state-of-the-art and future perspectives, Critic
Reviews in Environ Sci and Tech, DOI: 10.1080/10643389.2018.1447717
Hejmanowska, B, Glowienka, E, Michalowska, K (2016) Free satellite imagery for monitoring a reclaimed sulphur
mining region Tarnobrzeg, Poland, 2016 Baltic Geodetic Congress (BGC Geomatics), Gdansk, 2016, pp.
134-139, doi: 10.1109/BGC.Geomatics.2016.32.
Herrera, G, Tomás, R, Lopez-Sanchez, JM, Delgado, J, Mallorqui, JJ, Duque, S, Mulas, J (2007) Advanced
DInSAR analysis on mining areas: La Union case study (Murcia, SE Spain). Eng Geo. 90 (34): 148159.
Hui, S., Charlebois, L, Sun, C (2015) Real-time monitoring for structural health, public safety, and risk management
of mine tailings dams, Canadian J of Earth Sciences Available at:
International Commission On Large Dams (ICOLD) (2001) Tailings dams: Risk of dangerous occurrences, lessons
learnt from practical experiences, United Nations Environmental Programme (UNEP), Division of Technology,
Industry and Economics (DTIE) and International Commission on Large Dams (ICOLD), Paris, France,
Bulletin 121, 2001
Johnstone, WM (2012) Life safety modelling framework and performance measures to assess community
protection systems: Application to tsunami emergency preparedness and dam safety management, Doctor
of Philosophy, University of British Columbia, Canada, October, 2012
Keaveny, P. (2019) Brumadinho dam collapse: Mining industry needs radical change to avoid future disasters, 8
March 2019, The Conversation, available at:
Ledwaba, P and Nhlengetwa, K (2016) When policy is not enough: prospects and challenges of artisanal and
small-scale mining in South Africa, J of Sus Dev Law and Policy, Vol 7, No 1 (2016) /African Journals On
Leonard, L (2017) State governance, participation and mining development: Lessons learned from Dullstroom,
Mpumalanga, Politikon, South African J of Politic Studies Vol 44, 2017 - Issue 2
Lumbroso, D, Sakamoto, D, Johnstone, WM, Tagg, AF, Lence, BJ (2011) The development of a Life Safety Model
to estimate the risk posed to people by dam failures and floods, Dams and Reservoirs, The official J of the
British Dam Society, Vol 21, Issue 1, pp31-43, June 2011
Lumbroso, D, Tagg, A (2011) Evacuation and loss of life modelling to enhance emergency response, published in
the proceedings of the International Symposium on Urban Flood Risk Management Graz, Austria, 21-23
September 2011
Lumbroso, D, Davison, M (2018) Use of an agentbased model and Monte Carlo analysis to estimate the
effectiveness of emergency management interventions to reduce loss of life during extreme floods, J of Flood
Risk Manage, Vol 11, Issue S1, January 2018, ppS419-S433
Lumbroso, D, McElroy, C, Goff, C, Collell, M, Petkovšek, G, Wetton, M (2019) The potential to reduce the risks
posed by tailings dams using satellite-based information, Int J of Disaster Risk Reduction. ISSN: 2212-4209,
Vol38, August 2019
Lumbroso, D, Davison, M, Body, R, Petkovšek, G (2020) Modelling the Brumadinho tailings dam failure, the
subsequent loss of life and how it could have been reduced, submitted, Nat Hazards and Earth Sys Sci
Marcuccio, S, Ullo, S, Carminati, M, Kanoun, O (2019) Smaller satellites, larger constellations: Trends and design
issues for earth observation systems, IEEE Aerospace and Electronic Systems Magazine. 34. 50-59.
Martin, TE, Davies, MP (2000) Development and review of surveillance programs for tailings dams, AGRA Earth
and Environmental Limited, Burnaby, British Columbia, Canada
Martin, TE, McRoberts, EC (1999) Some considerations in the stability analysis of upstream tailings dams in
Proceedings of Sixth International Conference on Tailings and Mine Waste ’99, 24-27 January 1999, Fort
Collins, Colorado, USA
McDevitt, A (2017) Transparency and accountability initiatives in the extractives sector, K4D Helpdesk Report.
Brighton, UK: Institute of Development Studies.
Mohamed, MAAH., Samuels, PG, Morris, MW, Ghataora, GS (2002) Improving the accuracy of prediction of
breach formation through embankment dams and flood embankments. Proceedings of the international
conference on fluvial hydraulics (Riverflow 2002), 3-6 September 2002. Louvainla- Neuve. Belgium
Minerals Policy Institute (2014) Chronology of major tailings dam failures Updated with Mount Polley, 7 August
2014, available at:
Mining Journal (2019) Tailings spill at Doe Run's Cobriza mine,
Mining Journal (2020) March underway to Brumadinho to mark year since fatal dam collapse, 21 January 2020,
available at:
Morris, MW (2011) Breaching of earth embankments and dams. PhD thesis, The Open University, UK
Mulaba-Bafubiandi, AF, Singh, N (2018) Junior mining as innovation entrepreneurship in minerals industry in South
Africa in Proceedings of the International Conference on Industrial Engineering and Operations Management
Pretoria / Johannesburg, South Africa, 29 October to 1 November 2018
Nishijima, M, Rocha, FF (2020) An economic investigation of the dengue incidence as a result of a tailings dam
accident in Brazil, J of Enviro Manage, 253, art. no. 109748;
O’Neill, S. (2019) A machine learning revolution in disaster response, The Alan Turing Institute, available at:
Palú, MC, Julien, PY (2019) A review of tailings dam failures in Brazil, 24 to 28 November 2019, Conference: XXIII
Simpósio Brasileiro De Recursos Hídricosat: Foz do Iguaçú, Brazil, 2019
Petkovšek, G, Hasan, MAAH., Lumbroso, D, Roca Collell, M (2020) A two-fluid simulation of tailings dam
breaching, submitted, Mine Water and the Environ, J of the Intl Mine Water Assoc (IMWA)
Petley, D (2019) Cobriza, Peru: Another significant tailings dam failure, American Geophysic Union (AGU), 16 July
2019 available at:
Sachs, JD, Warner, AM. (2001) The curse of natural resources. European economic review, 45(4), 827-838
Scaioni, M, Marsella, M, Crosetto, M, Tornatore, V, Wang, J (2018) Geodetic and remote-sensing sensors for dam
deformation monitoring, Sensors 2018, 18, 3682
Seifi A, Hosseinjanizadeh M, Ranjbar H, and Honarman M (2019) Identification of acid mine drainage potential
using Sentinel 2a imagery and field data. Mine Water and the Environment 38:707717,
Sonwalkar, M, Fang, L, Sun, D (2010) Use of NDVI dataset for a GIS based analysis: A sample study of TAR
Creek superfund site, Ecol Inform, Vol 5, Issue 6, November 2010, pp484-491
Thomas, A; Edwards, SJ; Engels, J; McCormack, H; Hopkins, V, Holley, R; (2019) Earth observation data and
satellite InSAR for the remote monitoring of tailings storage facilities: a case study of Cadia Mine, Australia.
in: Paterson, AJC and Fourie, AB and Reid, D, (eds.) Proc of the 22nd International Conference on Paste,
Thickened and Filtered Tailings. pp183-195, Australian Centre for Geomechanics (ACG), The Univ of
Western Australia: Perth, Australia
Tomás, R, Romero, R, Mulas, J, Marturià, JJ, Mallorquí, JJ, Lopez-Sanchez, JM, Herrera, G, Gutiérrez, F,
González, PJ, Fernández, J, Duque, S, Concha-Dimas, A, Cocksley, G, Castañeda, C, Carrasco, D, Blanco,
P (2014) Radar interferometry techniques for the study of ground subsidence phenomena: A review of
practical issues through cases in Spain Environ Earth Sci. 71: pp163181
Tre-Altimira (2018) Tailings dam failure Available at: ]
United Nations (UN) (2017) Earth observations for official statistics, Satellite imagery and geospatial data task
team report, 5 December 2017
Van der Werff, H, Van der Meer, F (2015). Sentinel-2 for mapping iron absorption feature parameters. Remote
Sens 7, 1263512653.
Vergilio, C, Lacerda, D, Oliveira, B, Sartori, É, Campos, G, Pereira, A, Aguiar, D, Souza, T, Almeida, M, Thompson,
F, Rezende, C (2020) Metal concentrations and biological effects from one of the largest mining disasters in
the world (Brumadinho, Minas Gerais, Brazil). Sci Rep. 10. 10.1038/s41598-020-62700-w.
Vick, S (1990) Planning, design, and analysis of tailings dams, Published by BiTech, Vancouver, Canada, ISBN:
Xiao, R, Shi, H, He, X, Li, Z, Jia, D, Yang, Z (2019) Deformation monitoring of reservoir dams using GNSS: An
application to south-to-north water diversion project, China in IEEE Access, vol. 7, pp. 54981-54992, 2019
... Following the example ofGunn et al. (2015), detailed daily reports assessing the level of risk for specific areas of TSFs and WRPs could be generated from TL-ERT measurements to support the operators in charge of stability monitoring. As discussed byRobinson et al. (2008a), TL-ERT could be used to "fill the gap" between point sensor measurements and remote sensing imagery, that are currently the two main strategies developed in EWS monitoring programs for mining wastes (e.g.,Li et al. (2020);Lumbroso et al. (2021)). ...
Full-text available
Mining operations generate large amounts of wastes which are usually stored into large-scale storage facilities which pose major environmental concerns and must be properly monitored to manage the risk of catastrophic failures and also to control the generation of contaminated mine drainage. In this context, non-invasive monitoring techniques such as time-lapse electrical resistivity tomography (TL-ERT) are promising since they provide large-scale subsurface information that complements surface observations (walkover, aerial photogrammetry or remote sensing) and traditional monitoring tools, which often sample a tiny proportion of the mining waste storage facilities. The purposes of this review are: (i) to understand the current state of research on TL-ERT for various applications; (ii) to create a reference library for future research on TL-ERT and geoelectrical monitoring mining waste; and (iii) to identify promising areas of development and future research needs on this issue according to our experience. This review describes the theoretical basis of geoelectrical monitoring and provides an overview of TL-ERT applications and developments over the last 30 years from a database of over 650 case studies, not limited to mining operations (e.g., landslide, permafrost). In particular, the review focuses on the applications of ERT for mining waste characterization and monitoring and a database of 150 case studies is used to identify promising applications for long-term autonomous geoelectrical monitoring of the geotechnical and geochemical stability of mining wastes. Potential challenges that could emerge from a broader adoption of TL-ERT monitoring for mining wastes are discussed. The review also considers recent advances in instrumentation, data acquisition, processing and interpretation for long-term monitoring and draws future research perspectives and promising avenues which could help improve the design and accuracy of future geoelectric monitoring programs in mining wastes.
Full-text available
Tailings ponds are one of the three major production facilities in metal mines. The volume of tailings increases year by year, but the storage capacity of existing tailings ponds is limited. Therefore, tailings dams must become more fine-grained and larger. The potential hazard they represent should not be underestimated. This paper reveals the causes and regional distribution patterns of 342 tailings dam failures globally from 1915 to 2021 through statistical analysis. It was found that tailings pond failures occur almost every year, with an average of 4.4 accidents/year (1947–2021). The frequency has been gradually increasing in recent years, and most tailings pond failures are directly related to heavy rainfall or earthquakes. The frequency of tailings pond failures was significantly higher in Asia (21.3%) and the Americas (57.9%), especially in China (n = 43) and the United States (n = 107). Causes of tailings pond failures differed among regions. Most tailings pond failures in Asia and Europe were related to hydroclimate, while those in South America were mainly triggered by earthquakes. This study will provide theoretical data for the pre-design as well as the safe and stable operation of global tailings ponds, which will help to prevent global tailings pond failures.
Full-text available
A 3D tailings dam visualization early warning system was developed based on GIS (geographic information system) combining ARIMA (autoregressive integrated moving average model) and 3S (RS, GIS, GPS) technology for prediction of phreatic line changes and tailing dam deformation. It was applied for monitoring and early warning for the gold–copper tailing dam in Zijinshan Dadongbei tailing pond. The system consists of equipment management, data management, prediction, monitoring and early warning, and 3D visualization modules. It is able to do data management, visualization and disaster prediction, and early warning based on 79 monitoring points of rainfall, infiltration line, and deformation of the tailing dam in the Zijinshan mine. The design and application of the system reflect its features of rich functionality, high practicality, intuitive effect, and high reference value. The system solves the problems of low visualization of monitoring data, poor management of multiple data, and feasible prediction and early warning of point–surface combination. It realizes high-precision prediction of key factors and real-time warning of disaster.
Full-text available
In recent years the number of tailings dams failures has increased. On 25 January 2019, the Brumadinho tailings dam in Brazil suddenly failed, releasing a mudflow over 10 m deep comprising some 107 m3 of mining waste which killed between 270 and 320 people. This paper details the use of an agent-based model, known as the Life Safety Model (LSM), to estimate the risk to people downstream of the Brumadinho tailings dam and to assess if the number of fatalities could have been reduced if a warning had been received prior to or at time the dam failed. The LSM modelling indicates that even if a warning had been issued as the dam failed, the number of fatalities could have been reduced. Agent-based modelling tools such as the LSM can help to inform and improve emergency plans for tailings dams, which will help to reduce the risks posed by them in the future.
Full-text available
The rupture of the Brumadinho mining tailings dam in Brazil is considered one of the largest mining disasters in the world, resulting in 244 deaths and 26 missing people, in addition to the environmental consequences. The present study aims to evaluate the concentrations of multiple elements and the biological effects on water and sediments of the Paraopeba River after the Brumadinho Dam rupture. The tailings are formed by fine particulate material with large amounts of Fe, Al, Mn, Ti, rare earth metals and toxic metals. In the water, the levels of Fe, Al, Mn, Zn, Cu, Pb, Cd and U were higher than those allowed by Brazilian legislation. In the sediments, Cr, Ni, Cu and Cd levels were higher than the established sediment quality guidelines (TEL-NOAA). The differences in metal concentrations in the water and sediments between the upstream and downstream sides of the dam illustrate the effect of the tailings in the Paraopeba River. Toxicological tests demonstrated that the water and sediments were toxic to different trophic levels, from algae to microcrustaceans and fish. The fish exposed to water and sediments containing mine ore also accumulated metals in muscle tissue. This evaluation emphasizes the necessity of long-term monitoring in the affected area.
Conference Paper
Full-text available
Accidents with tailings dams have been occurring worldwide at alarming rates. In the last decade, there were 18 accidents in 12 different countries. However, the last two accidents in Brazil, i.e., the Mariana dam and the Brumadinho dam cases, changed the perception about the safety of this type of the structures in Brazil and around the world. After these accidents, the National Mining Agency in Brazil banished the design of upstream dams and announced a mandatory decommissioning of same type of dams up to 2021. Internationally, the International Committee on Large Dams (ICOLD) announced the publication of technology update on tailings dam design and the Worldwide the International Council on Mining and Metals (ICMM) announced the establishment of an independent panel of experts to develop an international standard for storage tailings facilities.
Full-text available
We demonstrate the potential of satellite Interferometric Synthetic Aperture Radar (InSAR) to identify precursors to catastrophic slope failures. To date, early-warning has mostly relied on the availability of detailed, high-frequency data from sensors installed in situ. The same purpose could not be chased through spaceborne monitoring applications, as these could not yield information acquired in sufficiently systematic fashion. Here we present three sets of Sentinel-1 constellation images processed by means of multi-interferometric analysis. We detect clear trends of accelerating displacement prior to the catastrophic failure of three large slopes of very different nature: an open-pit mine slope, a natural rock slope in alpine terrain, and a tailings dam embankment. We determine that these events could have been located several days or weeks in advance. The results highlight that satellite inSAR may now be used to support decision making and enhance predictive ability for this type of hazard.
Full-text available
Deformation monitoring plays an important role in performance monitoring to ensure the reservoir dams and embankments are functioning as designed. This work should be the first deformation monitoring application of GNSS in China’s huge South-to-North Water Diversion Project. A GNSS deformation monitoring system, equipped with 4G data transmission and automated data processing software, is established at Shuangwangcheng Reservoir, an important regulation control project on the Eastern Route. Precision evaluations of different observation sessions from both GPS and BDS are conducted, and results show that the performance of BDS is comparable to that of GPS, especially for longer observation session solutions. The 1 mm horizontally and 2 mm vertically precisions of daily solutions meet the deformation monitoring requirement of the project. The deformation time series reveal an uneven settlement of Shuangwangcheng Reservoir dam. Causes of deformation are analyzed and the water level change in the reservoir is deemed as the main factor.
Full-text available
In recent years, the measurement of dam displacements has benefited from a great improvement of existing technology, which has allowed a higher degree of automation. This has led to data collection with an improved temporal and spatial resolution. Robotic total stations and GNSS (Global Navigation Satellite System) techniques, often in an integrated manner, may provide efficient solutions for measuring 3D displacements on precise locations on the outer surfaces of dams. On the other hand, remote-sensing techniques, such as terrestrial laser scanning, ground-based SAR (synthetic aperture radar) and satellite differential interferometric SAR offer the chance to extend the observed region to a large portion of a structure and its surrounding areas, integrating the information that is usually provided in a limited number of in-situ control points. The design and implementation of integrated monitoring systems have been revealed as a strategic solution to analyze different situations in a spatial and temporal context. Research devoted to the optimization of data processing tools has evolved with the aim of improving the accuracy and reliability of the measured deformations. The analysis of the observed data for the interpretation and prediction of dam deformations under external loads has been largely investigated on the basis of purely statistical or deterministic methods. The latter may integrate observation from geodetic, remote-sensing and geotechnical/structural sensors with mechanical models of the dam structure. In this paper, a review of the available technologies for dam deformation monitoring is provided, including those sensors that are already applied in routinary operations and some experimental solutions. The aim was to support people who are working in this field to have a complete view of existing solutions, as well as to understand future directions and trends.
On November 2015, the Fundão Tailings Dam, located at Mariana municipality in Brazil, failed. Besides the deaths and injuries, economic losses, pollution and health problems associated to heavy metals in the water, Brazilian municipalities near the accident experienced an increase in the incidence of dengue. Since dengue fever is an insect-borne disease and the mosquito develops where there is stored water, there must be a relationship between the dam accident and the incidence of the disease. The purpose of this study is to test whether there is a causal relationship between the dam accident in Mariana and the number of dengue cases, number of hospitalizations due to dengue, and dengue outbreak in the municipalities affected by the accident. We find evidence that the accident had a positive and statistically significant impact on dengue indicators (for example, the probability of a dengue outbreak increased in 19%), what makes us call attention to another negative externality of tailings dam accidents.
The importance of earth observation (EO) from space is felt today more than ever in many different fields of human activity. Governments, international organizations, military bodies, private industry, and even individuals benefit every day of the products of spaceborne remote sensing technology. In this paper, we present a brief overview of some of the main trends in the EO scenario, focusing on the emergence of a new paradigm for EO space systems namely: the ongoing, disruptive shift from large satellites to constellations of small spacecraft, fueled by the recent introduction of several key technologies, such as for instance electric propulsion. We present the outline of a tool specifically conceived to assist the system architect in the early design phase of an EO constellation of microsatellites equipped with electric thrusters. A case study is presented for the application of a multisegmented constellation, with visual, thermal infrared, and radar components to remotely monitor a national railway network. We conclude that, in spite of the complexity of modern small satellite constellations, preliminary design can be successfully performed in a simplified and effective way.
Secondary iron minerals associated with acid mine drainage (AMD) such as copiapite, jarosite, schwertmannite, goethite, ferrihydrite, and hematite can be generated from pyrite oxidation. This study was an effort to determine the AMD potential of the Darrehzar mine, a porphyry copper mine in the Kerman Cenozoic magmatic arc, using remote sensing and field data. The spectral angle mapper method was applied on Sentinel 2a images to identify AMD minerals and classify the study area. The produced map was verified by field surveys and laboratory analysis of rocks and sediments as well as pH and electrical conductivity measurements of water samples. Jarosite–clay group minerals were detected in the mine pit and in an active waste dump, and jarosite–goethite and goethite–hematite group minerals were identified in inactive waste dumps. Moreover, acidic water was observed in the pit, while the neutral water was where it arrives and discharges from the mine.
Conference Paper
Tailings storage facilities (TSFs) are an essential infrastructure of mineral processing, but they represent a significant physical, chemical and biological hazard and must, therefore, be strictly and responsibly sited, managed and closed. Tailings can, for example, be dispersed by many processes (such as sinkholes, earthquakes, intense rainfall and flood events, and wind), substandard design and construction, and seepage. The stability and behaviour of TSFs needs to be continuously monitored and one highly effective way of doing this is through satellite Earth observation (EO). The EO industry is witnessing a technological revolution. Large and long-lifespan satellite sensors that have been the staple of national space agencies and commercial satellite manufacturers are now being complemented by constellations of low-cost, short-lifespan ‘cube sats’ by companies with the ambition to image the whole earth daily. Satellites with synthetic aperture radar (SAR) sensors are also collecting high volumes of data, with the added benefit of being able to do so day or night and in different weather conditions. The range of data options and capabilities these provide open opportunities for novel data analysis techniques for TSFs. One of these is satellite InSAR (interferometric SAR; a technique used to map millimetric-precision changes in ground height over time), which is already used by mining companies to reduce risk in and of their operations. From monitoring the stability of TSFs, through to assessments of impacts of natural hazards, InSAR allows rapid and accurate targeting of high-risk areas and structures to identify those that require subsequent investigation through ground-based methods. To demonstrate the application of EO data and InSAR in identifying pre- and post-failure mine activities and TSF deformation, the authors will present a case study across Cadia mine, New South Wales, Australia, which had a localised TSF failure on 9 March 2018. The InSAR results presented show that low-magnitude subsidence signals were observed across the TSF dam during the year preceding the collapse. In January 2018 a notable change in behaviour was observed, with a concentrated area of subsidence focused on the region which initially failed on 9 March 2018. Furthermore, post-collapse InSAR measurements show an increased rate of subsidence for regions either side of the failure zone. Review of medium- and high-resolution satellite images show that the failure was phased, with an initial failure and then a subsequent failure at least two days after 9 March 2018. It also highlights what might be construction activity associated with a dam raise prior to failure.