Resources, Conservation & Recycling 161 (2020) 104896

Contents lists available at ScienceDirect

Resources, Conservation & Recycling

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

Full length article

Reconciling Diverging Views on Mineral Depletion: A Modified Cumulative T

Availability Curve Applied to Copper Resources
Emilio Castillo1,2, , Roderick Eggert1
⁎

1
Division of Economics and Business, Colorado School of Mines, 816 15th Street, Golden, CO 80401, USA
2
Chilean Copper Commission (COCHILCO), Agustinas 1161, Santiago, Chile

ARTICLE INFO ABSTRACT

Keywords: Mineral depletion is a perennial concern in natural resources management. Despite a multiplicity of perspectives,
peak minerals most analyses fall largely into one of two groups. Studies in the first group take the fixed quantity of any material
mineral depletion in the earth as the starting point for analysis, focus primarily on physical stocks and flows of mineral resources,
cumulative availability curve and assess when society will run out of a resource or when production and use will peak (the fixed-stock
copper
viewpoint). Studies in the second group start by noting the heterogeneous nature of mineral resources and our
incomplete knowledge of the quantity and quality of minerals in the earth, focus on the dynamic nature of
mineral availability, and assess the ability of society to adjust to mineral depletion at existing mines though
mineral exploration and mine development, technological innovation, and substitution (the opportunity-cost
viewpoint). This paper seeks to reconcile these diverging perspectives by developing a modified cumulative
availability curve (CAC) – which combines physical stocks and flows (from the fixed-stock perspective) with
geologic stock uncertainty, different demand scenarios and extraction costs (from the opportunity-cost per-
spective). When applied to copper resources, the modified CAC suggests that copper demand is likely to be
satisfied from known deposits until about 2075. Thereafter, society's ability to discover and develop previously
unknown deposits and improve the efficiency of their production - as well as efforts at substitution and im-
proving recycling and re-use of copper - will importantly influence whether copper demand is satisfied and at
what copper price.

1. Introduction Policy Commission, 1952). In the 1970s, attention returned to energy

and minerals, in this case concerns about reliability of foreign sources
Minerals have been an intrinsic part of human development, from of oil and some nonfuel minerals, the finite supply nature of natural
the stone age to what is being called the silicon age of the 21st century. resources, as well as long-term adequacy of nonrenewable resources
Minerals stand as major inputs in technological development and in the generally and implications for economic growth (Meadows et al., 1972;
improvement of human living standards. Given the broad use of mi- Smith, 1979). In each case, society adjusted to perceived scarcities
nerals in modern society, long-term mineral depletion is a critical issue through one or a combination of three approaches: developing new
for policy makers and society. sources of mineral and energy resources, using resources more effi-
Concern about depletion of mineral resources is not new. It has been ciently and developing substitute materials or sources of energy.
a recurring societal concern. In the middle 1800s, for example, W.S. A number of recent studies focus on the relationships between and
Jevons explored the exhaustion of coal in Britain (Jevons, 1865). In the among mining, sustainable development and long-term resource de-
late 1800s and early 1900s, dramatic increases in the use of fertilizers in pletion. Dubiński (2013) argues that rational acquisition and use of
commercial agriculture led to fears about the adequacy of naturally natural resources are central to sustainable mineral development.
occurring sources of nitrogen, which at the time came from animal Mudd (2013) revisits original assumptions in “The Limits to Growth”
dung, urine and Chilean nitrate deposits (Hager, 2009; Smil, 2000). In (1972), highlighting that, despite increasing extraction costs, the finite
the years following World War II, concerns focused on a wide range of nature of minerals is still unclear. Ali et al. (2017) take advantage of
mineral raw materials and energy, given the significant use of mineral peak models of resource extraction to suggest that mineral depletion –
and energy resources during the war (The U.S. President's Materials as reflected by declining production and use of copper - may start to

⁎
Corresponding author: Tel: +1-720-243-7157
E-mail addresses: [email protected] (E. Castillo), [email protected] (R. Eggert).

https://doi.org/10.1016/j.resconrec.2020.104896
Received 7 October 2019; Received in revised form 27 March 2020; Accepted 17 April 2020
0921-3449/ © 2020 Elsevier B.V. All rights reserved.
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

occur sooner than expected, urging policy makers to take coordinated reserves and resources to build their models. Demand in both cases is
and planned actions at a global scale, intervening in mineral markets to assumed as perfectly price inelastic (that is, consumption is not affected
promote sustainable development goals. In response, by price changes) and increasing. Section three proposes methodolo-
Tilton et al. (2018) criticize both the use of peak models to assess long- gical changes for supply and demand in the CAC to reconcile differences
term mineral availability and depletion and the call for strong gov- with peak models. Finally, the last section applies the modified CAC to
ernment intervention to mitigate future shortages beyond interventions copper resources, suggesting that known reserves and resources are
to control environmental damages and to correct the tendency of pri- responsive enough to meet demand until 2075. This result highly
vate actors to underinvest in technological innovation from society's contrasts with peak copper models, which signal peak copper before
perspective. 2050.
Broadly speaking, approaches to assessing the threat of mineral
depletion fall into two categories: “fixed stock” and “opportunity cost” 2. Two models for mineral depletion
approaches, where consensus is still hard to reach in even basic defi-
nitions (Segura-Salazar and Tavares, 2018). The fixed-stock approach is Peak models and the CAC differ fundamentally in the unit of ana-
dominantly a physical view of natural resource scarcity, while the op- lysis. The former focus on physical units, such as tonnage, kilograms or
portunity-cost approach is largely economic (Barbier, 2019). Reflecting grades, to estimate availability. On the other hand, the latter is pri-
the physical approach, Gordon et al. (2006) argue that physical stocks marily concerned with costs and prices, economic resources that society
of metals will be unable to sustain the modern quality of life, suggesting needs to give up for an additional unit of mineral resources (Tilton and
an absolute scarcity of minerals. Moreover, prices will not result in real Guzman, 2016, chap. 9). In practice, there are two main similarities and
warnings since price trends are not different between abundant and two differences between the two approaches. The approaches are si-
scarce minerals and markets do not reflect external mining costs milar in that supply drives long-term depletion and lack of complete
(Henckens et al., 2016b). In this context, peak models arise as an in- geologic information is a major limitation. The approaches are different
creasingly preferred approach for assessing mineral depletion using in that peak models use peak production as a clear signal of impending
different physical stock measures (Ali et al., 2017; Calvo et al., 2017; depletion, while CAC does not have a clear counterpart as an indicator
Northey et al., 2014; Prior et al., 2012). Reflecting the opportunity-cost of depletion. Moreover, CAC models acknowledge that markets will
approach, many mineral economists argue that peak models for mineral adapt as depletion drive prices up, either encouraging exploration or by
resources are inherently flawed and, as an alternative, suggest a cu- demand substitution. This adaptive process is not part of peak models, a
mulative availability curve (CAC) to assess long-term availability common critique from the opportunity-cost advocates.
(Tilton, 2018; Tilton et al., 2018; Tilton and Lagos, 2007).
The fixed-stock and opportunity-cost approaches differ significantly
2.1. Peak models
in the inferences they draw on how to respond to the threat of mineral
scarcity. While the former urges society to change the pace of devel-
The physical view usually turns to Hubbert's theory of peak pro-
opment, the latter claims that prices and costs will provide incentives to
duction to assess mineral depletion. Peak models are based on a bell-
offset depletion. The diverging positions appear in virtual stalemate:
shaped profile in the extraction of a fixed stock (Hubbert, 1956)1. The
economists would disregard physical models that do not include how simplicity of the theory is the main driver behind its widespread ap-
markets respond to scarcity, and natural scientists do not feel comfor-
plication, despite critiques from opportunity-cost supporters. There is
table waiting for markets to provide signals to inform decisions taken only one input that defines peak production: The Ultimate Recoverable
by governments and private actors.
Resource (URR) defines the entire supply over time in peak models. The
The purpose of this paper is to review the two approaches for as- URR is an assumed estimate of the total mineral resources that society
sessing the threat of mineral depletion and develop a reconciled
will recover from mineral deposits, both historically and into the future
methodological approach. Toward this end, we build a formal model for (Giurco et al., 2012). In its simplest form (Calvo et al., 2017), supply at
long-term supply and demand for minerals. The model allows us to be
every period t (St) is given by equations (1) and ((2), where b and tpeak
explicit about assumptions behind peak models and CAC applications. (the peak year) are the regression parameters to be adjusted to fit his-
Then, we suggest modifying the CAC in two main aspects to help re-
torical data.
concile these approaches. First, we propose to emphasize separate de-
mand scenarios for more-developed and less-developed regions.
2
1 t tpeak
URR 2
Second, we include different supply units for reserves, resources and St = e b
b (1)
undiscovered deposits, incorporating economic and geologic stock un-
certainty. We finally apply the modified CAC to copper resources and URR
Stpeak =
compare our results with previous peak models. b 2 (2)
The CAC has fewer empirical applications compared to peak models
Peak models implicitly assume that other determinants of supply
(Calvo et al., 2017; Tilton, 2018). Since CAC's theoretical development
(price, technology, exploration or input costs) are irrelevant to assess
in 1987 only a few authors have studied long-term depletion for specific
long-term depletion (Tilton, 2018). In practice, peak models allow
resources like lithium (Yaksic and Tilton, 2009), tellurium
different URR scenarios, but changing the URR does not generate much
(Woodhouse et al., 2013), oil (Aguilera, 2014; Aguilera et al., 2009),
change in the peak year tpeak (Northey et al., 2014; Sverdrup et al.,
thorium (Jordan et al., 2015), indium (Lokanc et al., 2015), gallium
2014). Many authors have tried to incorporate more variables into the
(Frenzel et al., 2016) and iron ore (Jasiński et al., 2018). In practice,
simplest Hubbert model. Sverdrup et al. (2014) develop a complex
there is not a clear point in the CAC to mark depletion, contrasting with
world dynamic model combining population, recycling, markets and
the “peak” in peak models. Despite previous applications, no author has
mining supply. In their model, increasing demand after the peak should
clearly formalized the general methodological assumptions of the CAC,
be met by increasing recycling rates. Giurco et al. (2012) propose a
nor dedicated significant effort to demand's role. Additionally, most
detailed assessment at a mine level, where production decisions follow
CAC applications have been focused on minor minerals, which may
a trapezoid-shaped production profile over time. Northey et al. (2014)
hinder its application to more economically important minerals like
copper.
Section two reviews the main studies using peak models and CACs. 1
Several studies have tried to assess Hubbert's initial prediction about peak
We find that extraction costs driving long-term mineral depletion is oil (Bardi, 2019; Chapman, 2014; Criqui, 2013; Graefe, 2009; Sorrell et al.,
common in both approaches. Both rely on detailed information on 2010). We do not attempt to disprove or confirm Hubbert's hypothesis.

2
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

build on the Giurco et al. (2012) research, including ore grade as the assumed to change in response to income and population growth (gt)
main driver defining what deposits will be mined first. Quantity de- f
and material substitution efforts (st). Is it expected that s < 0 , because
manded in peak models is not a relevant variable as long as it is greater more substitution effort should decrease metal consumption. Ad-
than or equal to production from the peak function. The demand con- ditionally, demand can also change over time as an effect of changes in
dition is guaranteed by having a non-decreasing per capita demand. In consumer preferences or new technology.
the long-term, increasing recycling and exploration is expected to fill
the gap between increasing demand and constrained peak supply CACt (p) = Ri, t (p)
i (3)
(Northey et al., 2014; Sverdrup et al., 2014).
Modifying the simplest peak model does not result in major changes Dt = f (gt , st , t ) (4)
in results, generating a sense of theory robustness. For example, in the
The CAC indicates that future price is determined by future ex-
case of peak copper models, previous literature agrees on a copper
traction costs, and demand will define how quickly we deplete non-
shortage over the next 20 to 30 years, slightly dependent on the URR
renewable stocks (Tilton and Guzman, 2016, chap. 9). The CAC ap-
value (Bardi and Pagani, 2007; Giurco et al., 2012; Laherrére, 2010;
proach assesses depletion in three steps. First, expected future demand
Northey et al., 2014; Sverdrup et al., 2014). Therefore, there is little to
is summed from the current period (t0) up to a time in the future (tf).
gain by adding more dynamic components to the simplest peak model
Second, the price required to supply that amount from current reserves
(Calvo et al., 2017).
is obtained using the inverse of the CAC. Finally, if the resulting price is
Critiques of peak models have been widely discussed by supporters
in the steeper section of the CAC following equation (5) and price needs
of the opportunity-cost paradigm. First, peak models do not consider
to greatly increase to stimulate further supply, then there is a major
the effect of technology increasing availability of reserves, resources
threat from depletion (Tilton, 2018). However, moving into a steeper
and undiscovered deposits (Kharitonova et al., 2013). Second, peak
section of the CAC should also trigger discoveries and substitution ef-
models usually ignore market behavior such as, for example, that peak
forts (Tilton and Guzman, 2016, chap. 9).
production for minerals may occur for political, health or environ-
mental restrictions, or because demand falls, rather than due to physical tf
depletion, and that materials intensity of use declines as countries de- CAC 1
CAC Dt
velop (Criqui, 2013; Crowson, 2011; Ericsson and Söderholm, 2013; p t0
(5)
Tilton and Guzman, 2016, chap. 9). Third, geological stock uncertainty
created by lack of discovery data and unknown deposits does not affect The opportunity-cost paradigm does not generate complete con-
the dynamic behavior of peak models. Additionally, peak models er- sensus on how to assess mineral depletion. First, critics are less con-
roneously consider reserves and resources as fixed stock measures fident in the role of markets providing necessary incentives to offset
(May et al., 2012; Meinert et al., 2016; Wellmer and Scholz, 2018). As a depletion (Tilton, 2002) – importantly due to the arguably high ex-
result, peak model assumptions are generally strong, questionable and ternal social and environmental costs of mining that are not fully in-
biased towards a more pessimistic prediction of mineral depletion ternalized by markets (Segura-Salazar and Tavares, 2018). Second, as-
(Tilton, 2018). Nevertheless, peak models have an impact on the sci- suming that geologically scarce minerals are more likely to face
entific community concerned about sustainable development in the depletion than geologically abundant minerals, then price trends should
mineral industry (Ali et al., 2017; Calvo et al., 2017) or when it comes be different between them. However, price trends are not different
to think strategically about resources (Chapman, 2014). between geologically abundant and scarce minerals, failing to signal
mineral depletion (Henckens et al., 2016b). The implication is that
2.2. The Cumulative Availability Curve responding to mineral depletion requires international coordination to
assure supplies of geologically scarce minerals for future generations
The opportunity-cost approach focuses on what society needs to (Henckens et al., 2018, 2016a). Third, the opportunity-cost paradigm
give up for an additional unit of a mineral resource. The cumulative does not seem concerned about exponential growing demand and
availability curve (CAC) is an alternative to assessing mineral depletion overstates the role of technology offsetting depletion (Gordon et al.,
along with trends in real prices, real extraction costs and the value of 2006; Humphreys, 2013).
marginal reserves. The CAC represents all known geological sources for Despite critiques, the CAC is useful to assess potential depletion
a mineral and their full extraction costs over all time. In the CAC, de- shortages (Yaksic and Tilton, 2009). Figure 1 illustrates the applic-
mand determines how quickly we extract available mineral resources ability of the CAC. For example, Figure 1 a) indicates that minerals
(Tilton and Guzman, 2016, chap. 9). Mineral depletion will drive prices without discontinuities in the CAC and gradual rising costs are unlikely
up, discouraging consumption, increasing substitution and recycling, to suffer from rapid scarcity as demand expands; rather society is likely
and encouraging new supply sources by exploration or technology. It is to see scarcity emerging and allowing time for response through, for
worth noting that the CAC uses all available information at a given example, technological innovation, new supplies, or substitution).
time, assuming that other non-price supply determinants remain fixed Nonetheless, discontinuous jumps or sharp increases in the CAC, as
(Tilton, 2002). The CAC implies that known lower-cost deposits will be depicted in Figure 1 b), c), and d) suggest that scarcity may emerge
mined first in the long-term. The concept is common both in peak suddenly, catching society off guard with insufficient time to respond
models and the CAC, but extraction cost information is not available for appropriately. Figure 1 d) shows a rising slope, becoming inelastic as
uncertain sources, like resources or undiscovered deposits. To solve the we approach to stock depletion (Aguilera, 2014).
data problem, peak models assume that extraction costs rise because of The number of actual CAC studies is limited, likely due to data
a decline in a deposit ore grade (Henckens et al., 2016b; Northey et al., limitations (Tilton and Guzman, 2016, chap. 9). Additionally, the CAC
2014; Prior et al., 2012; Vieira et al., 2012). Nonetheless, a declining is not expected to provide insights about how quickly society consumes
ore grade is a poor depletion predictor because of the endogenous re- available stocks to point expected depletion. However, the CAC is a
lationship between ore grade and economic conditions (Crowson, 2012; more general and flexible approach than peak models, allowing the
Tilton and Guzman, 2016, chap. 7; West, 2011; Wood consideration of different factors affecting mineral depletion
Mackenzie, 2015). (Tilton, 2018).
Formally, equation (3) indicates that the CAC is a stock variable,
adding up all known deposits or mineral sources i (Ri) at a time t, as a 3. Building consensus in depletion models
function of the price to incentivize their extraction (p), keeping all other
supply determinants fixed. Equation (4) reflects demand, which is The analysis of peak models and CAC applications supports two

3
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

Figure 1. Illustrative cases of the CAC. (a) Slowly rising slope due to gradual increase in costs. (b) Discontinuity in slope due to jump in costs. (c) Sharply rising slope
due to rapid increase in costs. (d) Rising slope rapidly becomes inelastic at maximum cumulative output constraint. Source: Tilton and Skinner (1987) and
Aguilera (2014).

Figure 2. Copper consumption for countries in more developed regions, 1950-2017. Points mark maximum and minimum per capita copper consumption every year
and lines represent the interquartile range. Weighted average considers total consumption of more developed regions weighted by their population.

major common concepts: supply and extraction costs drive depletion in (Meinert et al., 2016). A reconciled view of mineral depletion should
the long-term, and there is uncertain geological information. In peak consider geological stock uncertainty based on the resource classifica-
models, rising and then declining production (over the longer term) is tion and their different expected extraction costs. We consider the CAC
the result of gradual increases in the in the production cost of the re- as a more general approach, and it can be modified to include both
source (Bardi and Pagani, 2007). Ore grade decline is also used as a views and to provide guidance about future mineral availability and
proxy to reflect increasing production costs and environmental impacts their expected extraction costs (Henckens et al., 2016b; Tilton and
(Northey et al., 2014; Sverdrup et al., 2014). Therefore, even if market Guzman, 2016, chap. 9).
prices are not a satisfactory measure of scarcity, both sides agree that Consensus requires building on critiques arising from each para-
extraction costs from various sources at current technology may pro- digm. First, the opportunity-cost paradigm highlights the uncertain
vide useful insights about future mineral availability (Henckens et al., consumption forecast, changing demand with the level of development
2016b; Tilton, 2018). Geological information relies on data on reserves and the demand response to offset mineral depletion (Crowson, 2011;
and resources, which are stock concepts; these data often are mis- Tilton and Guzman, 2016, chap. 2). Second, fixed-stock supporters
understood to mean all the mineral resource there is, which is incorrect criticize the opportunity-cost paradigm's inability to pinpoint mineral

4
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

depletion, heavily relying on technological change and market forces. k represents the average growth rate. Additionally, demand in more
Then, it is required to find a clearer way to mark depletion in the CAC. developed countries needs to include changing effects driven by tech-
In the following sections we provide a framework to modify the CAC nology and substitution on the material composition of products
in two main aspects: demand and supply. First, the demand side should (Tilton and Guzman, 2016, chap. 2). Modelling those effects in the
consider a changing per capita demand as countries develop in the long- frontier is challenging considering the disruptive effect of technology
term. Second, on the supply side, geological stock uncertainty requires and consumer preferences in the long-term. We propose to follow
modifying the CAC to include a certainty level on extractable or re- simple ARIMA models to model per capita consumption in more de-
coverable resources, based on the geological classification for mineral veloped regions.
resources. Both modifications are necessary to construct a clear point to The second modification requires demand to adapt as we deplete
mark depletion in the CAC. known mineral stocks and prices increase. In the previous section, we
model demand at any time as a function of income or population
3.1. Demand growth and substitution effort taken in the past (equation (4)). We
explicitly consider substitution efforts as a function of past prices,
s
Long-term metal consumption in economic and peak models relies where substitution efforts increase with an increase in prices ( p > 0) .
t 1
to a significant extent on variants of the intensity-of-use hypothesis Therefore, demand will evolve over time according to equation (7). The
(IU), either analyzing metal consumption by unit of income (GDP) or first term in equation (7) refers to the change in demand as a result of
per capita. The original IU hypothesis suggests that metal consumption increasing income or population (IU hypothesis). The second term re-
changes with economic development and per capita income (Tilton and flects the offsetting effect of substitution. The last term indicates that
Guzman, 2016, chap. 2). The ratio of metal consumption to total in- demand may suffer additional unpredictable shocks over time.
come is expected to follow an inverted U shape, which is supported by D s p
historical evidence (Warell and Olsson, 2009). In another study, = fg ·g + fs · · t 1
+ ft
t pt 1 t (7)
Sverdrup et al. (2014) acknowledge that copper per capita consumption
depends on per capita GDP, but they consider a stable demand at ap- We expect that prices in the long-term are given by the CAC func-
proximately 10 kg/person when income per capita exceeds $25,000. tion as indicated in section 2.2. Then, we obtain a more useful ex-
However, they do not consider that IU for individual countries may pression for t 1 ,
p
t
decline on average as per capita GDP increases (Cuddington and
Zellou, 2013). Similarly, Müller et al. (2011) find a saturation con- CAC 1 t 1
t 0 Dj
pt 1 Dt 1 Dt 1
sumption level between 8 to 12 tons of iron per capita when studying t
= t
= = CAC
CAC t 1 (pt 1)
CAC 1
future availability of iron. p t 0 Dj p

In the case of copper, Figure 2 shows historical data of metal per pt 1 pt 1 Dt 1

= ·
capita consumption following an inverted U path in more developed CAC p
(pt 1)· t 1 p, CAC CAC
p Dt 1
regions2. The trend suggests that global per capita metal consumption
will not increase unbounded over the next century and instead is likely Where ɛp, CAC denotes the percentage change in cumulative resources
to converge at the average level of developed countries. because of one percentage change in prices. Replacing the previous
Analyzing per capita metal consumption provides an approach to result in equation (7) generates equation (8) below.
developing a reconciled view on mineral depletion, given its simplicity Dt 1
and use in previous applications. The methodology can be refined to Dt s p p, s ·s · CAC
= fg · g + fs · · t 1 + ft = fg ·g + fs · + ft
provide a more robust analysis in two ways. First, differentiate con- t pt 1 p, CAC p, CAC (8)
sumption from more and less developed countries in the analysis, de-
The term ɛp, s represents the price elasticity of substitution efforts.
fining various demand scenarios based on different saturation points.
From equation (8), substitution efforts become more important as we
Separating more and less developed countries allows consumption
approach the inelastic section of the CAC and can dominate as mineral
patterns to change as countries develop. Besides, demand scenarios
depletion increases over time. As discussed by Tilton and Guzmán
provide a range of variability for unexpected future outcomes. Second,
(2016, chap. 2), consumption may change in ways impossible to know
demand should respond to mineral scarcity as prices rise. Future market
due to substitution and changes in preferences. Material substitution in
behavior may respond in ways impossible to know but we can infer
response to prices may take a decade or more to materialize, suggesting
when substitution efforts are likely to trigger reductions in demand.
that price elasticity of substitution effort could be closer to zero in the
Even if metal demand is inelastic in the short-term with respect to price
medium-term (Smith and Eggert, 2018). Nevertheless, we can obtain a
changes, we should observe a greater response in the long-term.
conservative depletion measure from the CAC, if we omit substitution
Following previous work using variants of the IU hypothesis
effects while prices are able to provide additional geological resources.
(Cuddington and Zellou, 2013; Sverdrup et al., 2014), the convergent
As we deplete deposits, long-term prices increase, and we move to a
hypothesis implies that, on average, less developed regions change their
more inelastic section of the CAC. When ɛp, CAC becomes less than one
consumption level to approach more developed regions. We propose to
estimate the growth of the per capita consumption (CCLD, t ) from less (ɛp, CAC < 1), there is little room to provide minerals from known
developed regions as proportional to current consumption and the gap sources and substitution effort should dominate. We can mark depletion
with more developed regions, as indicated by equation (6). at the point where ɛp, CAC equals one, meaning that one percent increase
in price generates less than a one percent increase in available re-
CCLD, t sources.
CCLD, t = k· CCLD, t · 1
CCMD (6)
3.2. Supply
Where CCLD, t and CCLD, t indicate the change and current per capita
consumption from less developed regions at time t, CCMD the stable
The misconception of reserves and resources as fixed stocks is a
consumption level (or saturation point) in more developed regions and
common critique of peak models. Mineral resources are defined as a
known concentration of minerals in such form and amount that eco-
2
For consistency, we use the UN convention for more developed regions, nomic extraction of a commodity is currently or potentially feasible.
including Europe, Northern America, Australia/New Zealand and Japan Reserves are a subset of resources - the portion could be economically
(United Nations 2017). extracted at the time of determination (U.S. Bureau of, 1980). In this

5
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

Table 1 As indicated in equation (10), depletion analysis requires ac-

Scenarios of stable per capita copper consumption in less developed regions. knowledging that information changes over time. At every period, re-
Scenario kg/person Description serves are depleted through consumption and a fraction of resources
become reserves through mine planning and investment (αt). Ad-
1 5.31 ARIMA forecast in more developed regions ditionally, a fraction of undiscovered deposits generates reserves de-
2 5.46 40th quantile from the 2000 – 2017 period
pending on exploration effort (βt).
3 6.76 50th quantile from the 1950 – 1969 period
4 7.06 60th quantile from the 2000 – 2017 period dRvt
5 9.03 50th quantile from the 1970 – 1999 period = Dt + t · Rct + t ·UDt
dt (10)

In any given time, most available information to construct the CAC

sense, if we take reserves and resources as fixed geological units, we comes from reserves, representing a higher certainty level about po-
will be incorrectly measuring future availability, not considering how tential extraction. As reserves decrease, prices increase, and resources
they change with economic conditions (Crowson, 2011; Meinert et al., become more attractive. Then, long-term depletion analysis should in-
2016). For example, peak models have adjusted their measures of the clude resources as a future mineral source. Unfortunately, there is little
Ultimate Recoverable Resources (URR). Remaining copper stocks using research about conversion of resources to reserves over time
URR estimates increase from around 1,400 million tonnes (Northey et al., 2014). Without much information about resources, we
(Laherrére, 2010), to 1,800 million tonnes (Henckens et al., 2014; penalize them by giving a higher extraction cost, reflecting what is
Northey et al., 2014) and more recently to 2,100 million tonnes expected from equation (10). Assuming βt equals zero provides a con-
(Calvo et al., 2017). The changing URR is a major source of uncertainty servative estimate for depletion. In this way, we focus on known de-
in peak models, and their total stock estimates are usually too low posits rather than unknown information. The modified CAC contains
(Tilton, 2018). the available mineral stock under current information, extraction
The CAC reflects full extraction costs of all available deposits over techniques and economic conditions. Moreover, extraction costs ideally
all time. Equation (9) separates the CAC at any given time t in reserves should account for negative externalities like environmental pollution
(Rvt(p)), resources (Rct(p)) and undiscovered deposits (UDt(p)). These or other social impacts related to mining. It been shown that increasing
measures can change over time, but the CAC should remain as a stock extraction costs due to environmental and social consequences are al-
measure. It is only possible to generate an imperfect CAC based on ready occurring in the mining industry (Prior et al., 2012). However, as
available information of known deposits. with technological change, changing (more stringent) regulatory fra-
meworks will generate shifts in the CAC in unknown ways, but still
CACt (p) = Rvt (p) + Rct (p) + UDt (p) (9)
preserving the shape of the CAC (Tilton et al., 2018).

Figure 3. Per capita copper consumption in less developed regions, 1950-2100. Each line denotes a different stable consumption level from 5.31 to 9.03 kg/person.
Historical data is marked in white circles. Every scenario is calibrated using maximum likelihood, varying the stable consumption level and defining consumption in
1950 as the initial condition. The solid line shows the scenario of 6.76 kg/person.

6
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

Figure 4. a) Modified CAC separating copper reserves and resources. b) Modified CAC integrating copper reserves and resources. The vertical axis in each graph
indicates the estimated extraction cost in different countries. The horizontal axis indicates remaining contained copper from known deposits.

Assuming that current geological information accounts for all de- environmental considerations, raising concerns about using physical
posits in the earth's crust is misleading, but we also need to take ad- measures like tonnes or grades to make inferences about future copper
vantage of existing data to analyze mineral depletion. Modifying the supply (Mudd and Jowitt, 2018).
CAC for different stocks (reserves, resources and undiscovered deposits)
provides a first approach to include the effect of changing information 4. Applying the Modified Cumulative Availability Curve to copper
over time. However, there is still enormous potential to support con- resources
sumption with undiscovered deposits. For example, deep underground
copper deposits are estimated to contain more than 40 times the most A common or reconciled view of mineral depletion should combine
optimistic URR estimate (Kesler and Wilkinson, 2008; Meinert et al., the strong and simple depletion aspects of peak models with the eco-
2016). Global copper resources have been continuously increasing, nomic scarcity created by increasing costs. In any economic analysis,
despite the increase in production and decline in ore grades supply and demand are defined simultaneously and relate to each other
(Mudd et al., 2013). The increase in copper resources is not only driven through the market price. However, as we note, long-term forecasts of
by exploration, but also by economic, social, political and supply and demand are greatly dependent on unpredictable variables.

7
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

Figure 5. Modified CAC for reserves, resources and undiscovered deposits. The vertical axis in each graph indicates the estimated extraction cost in different
countries. The horizontal axis indicates total remaining contained copper tonnage, separating known (reserves and resources) and unknown (undiscovered) deposits.
The segmented vertical line indicates total current resource and reserves (3,020 million tonnes), representing not a physical barrier, but a discovery barrier.

Thus, as recommended by Rosenau, et al. (2009), we propose different Copper Mine Costs Workbook Q42016, considering total extraction
demand scenarios and assess how known deposits meet these scenarios. costs (C3) for each country (Wood Mackenzie, 2017). Average resource
extraction costs are estimated for each country considering the last cost
4.1. Data sources quartile (25% highest extraction costs) and using the highest world
average quartile for those without mining information. Unknown and
Previous literature on mineral depletion focuses on how copper undiscovered deep underground deposits are estimated to have condi-
demand increases as population grows (Northey et al., 2014; tions similar to current high-cost underground operations. It is worth
Sverdrup et al., 2014). Our population forecast by country is based on noticing that our estimate of undiscovered deposits is purely theoretical
World Population Prospects considering constant-fertility and constant- and only provides a reference figure for contained copper in the earth's
mortality prepared by the Population Division of the Department of crust. Nevertheless, as a conservative measure, we consider the highest
Economic and Social Affairs in the United Nations (United cost quartile from underground mines as the approximate extraction
Nations, 2017). Total copper consumption per country from 1950 to cost for undiscovered underground copper deposits.
2017 comes from the historical archive of the Chilean Copper Com-
mission (COCHILCO, 2018) and is included in the online appendix. 4.2. Demand
Copper deposits are divided into reserves, resources and un-
discovered deposits. Reserve data from countries are based on Per capita consumption in more developed regions is a non-sta-
Mudd and Jowitt (2018), accounting for approximately 641 million tionary series (we fail to reject the null hypothesis of unit root at the
tonnes. Current reserves and resources were estimated at 3,020 million 10% level) and our best estimate for future per capita consumption in
tonnes based on Mudd and Jowitt (2018)3. Following Kesler and more developed regions is 5.31 kg/person, based on an ARIMA (0,1,0)
Wilkinson (2008), estimated recoverable underground copper deposits model4. To estimate future per capita consumption in less developed
(known and unknown) are about 89,000 million tonnes. These un- regions, we apply equation (4) with various stable consumption levels
known (undiscovered) deposits are expected to have similar ore quality by distinguishing three main periods in more developed countries. First,
as known underground deposits, therefore can be extracted with ex- a reconstruction period from 1950 to 1969 in the decades immediately
isting methods. following World War II. Then an industrialization and consumption
Average extraction costs of reserves are based on Wood Mackenzie's
4
The ARIMA (0,1,0) purges serial correlation in the residuals and it is pre-
3
Resources from international waters (14.98 million tonnes) were excluded ferred using information criteria (AIC and BIC) compared to ARIMA (1,1,0),
from the analysis as extraction costs cannot be clearly estimated. ARIMA (0,1,1) and ARIMA (1,1,1).

8
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

Figure 6. Price elasticity of the modified CAC for copper reserves and resources, 2018-2100. The vertical axis indicates the price elasticity of quantity supplied based
on the modified CAC. Each line illustrates different consumption scenarios.

period between 1970 to 1999. Finally, a digitalization period beginning demand scenario at 14.5 percent, the average global value from 2013 to
in 2000. Per capita copper consumption between 1970 and 1999 in- 2017 (World Metal Statistics, 2018).
creased at every quantile with respect to consumption in the 1950 –
1969 period. For example, 50 percent of the population in more de- 4.3. Supply
veloped regions consumed on average less than 6.76 kg/person in the
1950 – 1969 period and the same percentage consumed less than 9.03 A typical CAC represents remaining mineral tonnage in the hor-
kg/person in the 1970 – 1999 period. Based on these periods, four izontal axis and the expected cost to provide them on the vertical axis.
potential stable consumption levels are defined (Table 1): The 50th In the modified cumulative availability curve (MCAC), we represent
quantile of the first period, the 50th quantile of the second period, the different supply sources according to their certainty level in reserves,
40th quantile of the last period and the 60th quantile of the last period5. resources and undiscovered deposits. Despite having different economic
These values represent a wide range of growth alternatives for less definitions, reserves and resources represent known copper deposits. As
developed regions. Additionally, we include a scenario defined by the mentioned in section 4.1, we penalize resources by assigning them the
consumption forecast in more developed regions. Figure A1 in the ap- top quartile in extraction cost from each country. The MCAC for re-
pendix provides further details of consumption distributions. serves and resources is presented in Figure 4. The top figure illustrates
Every scenario considers the 1950 consumption level as the required that reserves are only a fraction of known deposits but are also expected
initial condition, and we perform maximum likelihood to calibrate the to be mined first. Nevertheless, as reserves decrease, prices should en-
growth parameter that better fits historical data. Figure 3 shows that courage extraction from other known available deposits in the long-run.
models do not greatly differ to explain historical data, representing the Given the previous expected behavior, Figure 4b) represents a preferred
past exponential behavior. MCAC for long-term availability from known deposits.
Projected accumulated consumption from 2018 to 2100 appears in Underground unknown deposits are a potential source for minerals.
the appendix (Figure A2). In 2050, accumulated total demand (in- Nevertheless, we lack confidence about their location and extraction
cluding demand for recycled products) is expected to reach 1,193-1,551 costs. As previously stated, to undiscovered deposits we assign the ex-
million tonnes; and 4.788-7,136 million tonnes by 2100. Consumption traction costs from the highest quartile from current underground
from less developed regions represents 87 – 90 percent of total con- mines. The assumption aims to provide a measure of future mineral
sumption by 2050 and 95 – 97 percent of total consumption by 2100 sources, considering the lack of information about discovery costs.
(Figure A3). The recycling rate is assumed to remain fixed in each Nonetheless, it is not clear that we can simply include undiscovered
deposits in the MCAC with reserves and resources. The discovery bar-
rier is marked by the vertical line in Figure 5. In practice, the figure
5
The 40th and 60th quartile in the last period are expected to give a wider serves two main purposes. First, it illustrates that undiscovered deposits
range of scenarios for the most recent data from more developed countries. can greatly contribute to decrease future depletion if their geological

9
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

characteristics are similar to known underground deposits. The shape of increasing) part of the CAC, assuming that substitution efforts are mild
the curve suggests that we are unlikely to experience rapid economic before that point. These modifications consider major critiques elabo-
depletion. Second, the figure puts in perspective that physical depletion rated by each paradigm. A significant concern comes from increasing
should not be a primary concern for copper. Instead, mineral resource environmental and social concerns about mining activity, which are not
management can benefit from expanding our current understanding of directly included in the model. It can be expected that more stringent
deep underground deposits and reserves conversion rate to avoid a regulation and social demands make mining more challenging, being
discovery barrier. expressed as higher extraction costs, but still preserving the shape of the
CAC.
4.4. Copper long-term availability Implementing the modified CAC (MCAC) for copper suggests that
current reserves and resources can meet world consumption until 2075
The modified CAC offers two main insights related to physical maintaining current recycling rates. This is a more optimistic view than
availability of known deposits and their economic scarcity. First, the peak models, where peak production should occur before 2050.
availability from known deposits indicates that meeting demand by Additionally, the MCAC model highlights that physical depletion is not
2050 requires around one third of current reserves and resources, and the main issue when analyzing future copper supply. The conversion
around 88 percent by 2075. Nevertheless, the supply source composi- rate from resources to reserves is more important in the medium-term
tion is not homogenous across time. This is illustrated in Figure A4. For and deposit discovery rate is more important in the long-term.
example, in 2050, 71 percent of known reserves and 25 percent of The MCAC for copper does not indicate an insurmountable barrier
known resources need to be extracted. After 2050 we observe a change after 2075. Future mineral depletion depends on two main unknown
in the slope of reserves consumption, indicating a heavier reliance on variables requiring further research. First, the conversion rate from
resources. This indicates that, in the medium to long term, the con- resources and undiscovered deposits to reserves represents an under-
version rate from resources to reserves has a greater importance than developed topic than can improve the dynamic aspect of the depletion
physical depletion. In the most conservative demand growth scenario, analysis. Second, we are mostly unaware of society's ability to sub-
current known deposits are insufficient only after 2074. stitute metals in the long-term as a response to increases in prices. Both
As demand increases, deposits are depleted starting from those with aspects, related to supply and demand for minerals, will improve MCAC
lower costs. In this sense, economic scarcity appears as we move to the outcomes and policy guidance about mineral sustainability.
inelastic section of the modified CAC, indicating that an increase in Finally, urging mineral-rich countries to take protective measures
price generates a less than proportional increase in quantity supplied. about their resources seems not advisable based on the copper MCAC.
We approximate elasticity by fitting a polynomial function to the MCAC Delaying peak production can lock natural capital, taking away the
and calculating the price elasticity as the percentage change in quantity opportunity to support their economic growth. Additionally, restricting
supplied every year as a response of a one percent change in price mineral production will only increase the threat of substitution, po-
( % P ) . According to Figure 6 and depending on the demand scenario, tentially destroying markets for a country's mineral assets. In this sense,
% Q

we would expect to achieve the inelastic part between 2066 and 2076. research in new exploration techniques to find deep-covered deposits
The result contrasts with depletion suggested by copper peak models, and in improving the efficiency of mineral processing and extractive
where supply fails to meet demand by 2033-2050. Additionally, the metallurgy, as well as policy initiatives to incentivize exploration in
estimate from the MCAC is still conservative, because as we approach to technically-challenging and riskier deposits, are valuable to reduce
the inelastic section, markets will create strong incentives to search for potential impacts from resource depletion.
undiscovered deposits and substitute materials.
CRediT author statement
5. Conclusions
Emilio Castillo: Conceptualization, Methodology, Software, Formal
Mineral depletion is one of the main debate topics in mineral re- Analysis, Investigation, Writing - Original Draft, Visualization
sources management for sustainable development. Historically there Roderick Eggert: Conceptualization, Methodology, Writing -
have been two major approaches to assessing depletion: a fixed-stock Review & Editing, Supervision
paradigm and an opportunity-cost paradigm. The former is based on the
self-evident fact that the earth is finite and so are the minerals con-
Declaration of interests
tained in it, mostly relying on peak models to analyze mineral deple-
tion. The latter supports that economic behavior may change long be-
fore actual physical depletion takes place; the analysis is not so much The authors declare that they have no known competing financial
about the extraction of fixed stocks as it is about of what society needs interests or personal relationships that could have appeared to influ-
to give up for them. The opportunity-cost paradigm prefers to analyze ence the work reported in this paper.
mineral depletion based on the cumulative availability curve (CAC).
In this paper, we review applications of peak models and the CAC, Acknowledgements
finding that both approaches consider that supply and extraction costs
drive depletion in the long-term, and both acknowledge that geological This work was supported in part by CONICYT – 72180034 and the
information is at best uncertain. As a more general framework, we Critical Materials Institute, an Energy Innovation Hub funded by the US
modify and formalize the cumulative availability curve, including dif- Department of Energy, Advanced Manufacturing Office. We thank the
ferent demand scenarios for less- and more-developed countries and Chilean Copper Commission (COCHILCO) for providing the data sup-
considering uncertainty in geological information. In this way, we porting the analysis in this study. We thank two anonymous reviewers
propose a depletion measure as we move to the inelastic (steeply whose comments have greatly improved this manuscript.

Appendix A

The appendix lists relevant figures not included in the main text.

10
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

Figure A1. Per capita copper consumption in more developed regions, 1950 - 2017. Each curve indicates the distribution of per capita copper consumption in more developed
regions in every period. The vertical axis represents the percentage of the population in more developed regions consuming less than the corresponding value in the horizontal
axis. For example, 50 percent of the population in more developed regions consumed on average less than 6.76 kg/person in the 1950 – 1969 period.

Figure A2. Accumulated world copper consumption, 2018 to 2100. Accumulated demand forecast adds up yearly refined total consumption starting from 2018. Different lines
represent different growth scenarios for less developed regions. In more developed regions, every scenario considers a per capita consumption of 5.31 kg/person.

11
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

Figure A3. Participation in the world copper consumption by region, 2018 to 2100. Participation is estimated according to consumption in any given year. Different
lines represent different growth scenarios for less developed regions. In more developed regions, every scenario considers a per capita consumption of 5.31 kg/
person.

Figure A4. Copper extraction from reserves and resources, 2018-2100. The vertical axis indicates the percentage extracted over time for reserves and resources (left),
reserves (middle) and resources (right) over time. Each figure illustrates different consumption scenarios. Vertical lines appear as references for years 2050 and 2075.

12
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

References diffusion of porphyry copper deposits. Geology 36, 255–258. https://doi.org/10.

1130/G24317A.1.
Kharitonova, M., Mikhailov, A., Matsko, N., 2013. Influence of the time factor on the
Aguilera, R., 2014. Production costs of global conventional and unconventional petro- availability of deposits of nonferrous metals. Resour. Policy 38, 490–495. https://doi.
leum. Energy Policy 64, 134–140. https://doi.org/10.1016/j.enpol.2013.07.118. org/10.1016/j.resourpol.2013.06.006.
Aguilera, R., Eggert, R., Lagos, G., Tilton, J., 2009. Depletion and the Future Availability Laherrére, J., 2010. Copper Peak [WWW Document]. Oil Drum Eur. URLhttp://europe.
of Petroleum Resources. Energy J 30, 141–174. theoildrum.com/node/6307.
Ali, S.H., Giurco, D., Arndt, N., Nickless, E., Brown, G., Demetriades, A., Durrheim, R., May, D., Prior, T., Cordell, D., Giurco, D., 2012. Peak Minerals: Theoretical Foundations
Enriquez, M.A., Kinnaird, J., Littleboy, A., Meinert, L.D., Oberhänsli, R., Salem, J., and Practical Application. Nat. Resour. Res. 21, 43–60. https://doi.org/10.1007/
Schodde, R., Schneider, G., Vidal, O., Yakovleva, N., 2017. Mineral supply for sus- s11053-011-9163-z.
tainable development requires resource governance. Nature 543, 367–372. https:// Meadows, Donella, Meadows, Dennis, Randers, J., Behrens, W., 1972. The Limits to
doi.org/10.1038/nature21359. growth; a report for the Club of Rome's project on the predicament of mankind.
Barbier, E.B., 2019. From Limits to Growth to Planetary Boundaries: The Evolution of Universe Books, New York.
Economic View on Natural Resource Scarcity. Seminar, Payne Institute for Public U.S. Bureau of Mines and U.S. Geological Survey, 1980. Principles for Resource/Reserve
Policy, Colorado School of Mines, Golden, CO. Classification for Minerals. URL https://pubs.usgs.gov/circ/1980/0831/report.pdf.
Bardi, U., 2019. Peak oil, 20 years later: Failed prediction or useful insight. Energy Res. Lokanc, M., Eggert, R., Redlinger, M., 2015. The Availability of Indium: The Present,
Soc. Sci. https://doi.org/10.1016/j.erss.2018.09.022. Medium Term, and Long Term. URL https://www.nrel.gov/docs/fy16osti/62409.
Bardi, U., Pagani, M., 2007. Peak Minerals [WWW Document]. Oil Drum Eur. URLhttp:// pdf.
europe.theoildrum.com/node/3086. Meinert, L.D., Robinson, G.R., Nassar, N.T., 2016. Mineral resources: Reserves, peak
Calvo, G., Valero, Alicia, Valero, Antonio, 2017. Assessing maximum production peak and production and the future. Resources 5, 14. https://doi.org/10.3390/
resource availability of non-fuel mineral resources: Analyzing the influence of ex- resources5010014.
tractable global resources. Resour. Conserv. Recycl. 125, 208–217. https://doi.org/ Mudd, G.M., 2013. The limits to growth and “finite” mineral resources: Re-visiting the
10.1016/J.RESCONREC.2017.06.009. assumptions and drinking from that half-capacity glass. International Journal of
Chapman, I., 2014. The end of Peak Oil? Why this topic is still relevant despite recent Sustainable Development 204–220. https://doi.org/10.1504/IJSD.2013.056562.
denials. Energy Policy 64, 93–101. https://doi.org/10.1016/J.ENPOL.2013.05.010. Mudd, G.M., Jowitt, S.M., 2018. Growing global copper resources, reserves and produc-
COCHILCO, 2018. Anuario de estadísticas del cobre y otros minerales. Santiago. tion: Discovery is not the only control on supply. Econ. Geol. 113, 1235–1267.
Criqui, P., 2013. Peak Oil: Myth or Impending Doom? In: Dannreuther, Roland, https://doi.org/10.5382/econgeo.2018.4590.
Ostrowski, W. (Eds.), Global Resources: Conflict and Cooperation. Palgrave Mudd, G.M., Weng, Z., Jowitt, S.M., 2013. A Detailed Assessment of Global Cu Resource
Macmillan, pp. 187–205. https://doi.org/10.1057/9781137349149. Trends and Endowments. Econ. Geol. 108, 1163–1183. https://doi.org/10.2113/
Crowson, P., 2012. Some observations on copper yields and ore grades. Resour. Policy 37, econgeo.108.5.1163.
59–72. https://doi.org/10.1016/J.RESOURPOL.2011.12.004. Müller, D.B., Wang, T., Duval, B., 2011. Patterns of iron use in societal evolution. Environ.
Crowson, P.C.F., 2011. Mineral reserves and future minerals availability. Miner. Econ. 24, Sci. Technol. 45, 182–188. https://doi.org/10.1021/es102273t.
1–6. https://doi.org/10.1007/s13563-011-0002-9. Northey, S., Mohr, S., Mudd, G.M., Weng, Z., Giurco, D., 2014. Modelling future copper
Cuddington, J.T., Zellou, A.M., 2013. A simple mineral market model: Can it produce ore grade decline based on a detailed assessment of copper resources and mining.
super cycles in prices? Resour. Policy 38, 75–87. https://doi.org/10.1016/j. Resour. Conserv. Recycl. 83, 190–201. https://doi.org/10.1016/j.resconrec.2013.10.
resourpol.2012.09.003. 005.
Dubiński, J., 2013. Sustainable Development of Mining Mineral Resources. J. Sustain. Prior, T., Giurco, D., Mudd, G., Mason, L., Behrisch, J., 2012. Resource depletion, peak
Min. 12, 1–6. https://doi.org/10.7424/jsm130102. minerals and the implications for sustainable resource management. Glob. Environ.
Ericsson, M., Söderholm, P., 2013. Mineral Depletion and Peak Production. In: Chang. 22, 577–587. https://doi.org/10.1016/J.GLOENVCHA.2011.08.009.
Dannreuther, Roland, Ostrowski, W. (Eds.), Global Resources: Conflict and Rosenau-Tornow, D., Buchholz, P., Riemann, A., Wagner, M., 2009. Assessing the long-
Cooperation. Palgrave Macmillan, pp. 222–231. https://doi.org/10.1057/ term supply risks for mineral raw materials-a combined evaluation of past and future
9781137349149. trends. Resour. Policy 34, 161–175. https://doi.org/10.1016/j.resourpol.2009.07.
Frenzel, M., Ketris, M.P., Seifert, T., Gutzmer, J., 2016. On the current and future 001.
availability of gallium. Resour. Policy 47, 38–50. https://doi.org/10.1016/j. Segura-Salazar, J., Tavares, L.M., 2018. Sustainability in the minerals industry: Seeking a
resourpol.2015.11.005. consensus on its meaning. Sustainability. https://doi.org/10.3390/su10051429.
Giurco, D., Mohr, S., Mudd, G., Mason, L., Prior, T., 2012. Resource Criticality and Smil, V., 2000. Enriching the Earth: Fritz Haber, Carl Bosch, and the Transformation of
Commodity Production Projections. Resources 1, 23–33. https://doi.org/10.3390/ World Food Production. MIT Press.
resources1010023. Smith, B.J., Eggert, R.G., 2018. Costs, Substitution, and Material Use: The Case of Rare
Gordon, R.B., Bertram, M., Graedel, T.E., 2006. Metal stocks and sustainability. Proc. Earth Magnets. Environ. Sci. Technol. 52, 3803–3811. https://doi.org/10.1021/acs.
Natl. Acad. Sci. U. S. A. 103, 1209–1214. https://doi.org/10.1073/pnas. est.7b05495.
0509498103. Smith, V.K. (Ed.), 1979. Scarcity and Growth Reconsidered. Johns Hopkins University
Graefe, L., 2009. The Peak Oil Debate. Atlanta, GA. Press for Resources for the Future, Baltimore.
Hager, T., 2009. The Alchemy of Air: A Jewish Genius, A Doomed Tycoon, and the Sorrell, S., Speirs, J., Bentley, R., Brandt, A., Miller, R., 2010. Global oil depletion: A
Scientific Discovery That Fed the World But Fueled the Rise of Hitler. Broadway review of the evidence. Energy Policy 38, 5290–5295. https://doi.org/10.1016/j.
Books. enpol.2010.04.046.
Henckens, M.L.C.M., Driessen, P.P.J., Ryngaert, C., Worrell, E., 2016a. The set-up of an Sverdrup, H.U., Ragnarsdottir, K.V., Koca, D., 2014. On modelling the global copper
international agreement on the conservation and sustainable use of geologically mining rates, market supply, copper price and the end of copper reserves. Resour.
scarce mineral resources. Resour. Policy 49, 92–101. https://doi.org/10.1016/j. Conserv. Recycl. 87, 158–174. https://doi.org/10.1016/j.resconrec.2014.03.007.
resourpol.2016.04.010. The U.S. President's Materials Policy Commission, 1952. Resources for Freedom.
Henckens, M.L.C.M., Driessen, P.P.J., Worrell, E., 2014. Metal scarcity and sustainability, Washington D.C.
analyzing the necessity to reduce the extraction of scarce metals. Resour. Conserv. Tilton, J., 2002. On Borrowed Time? Assessing the Threat of Mineral Depletion. Resources
Recycl. 93, 1–8. https://doi.org/10.1016/J.RESCONREC.2014.09.012. for the Future.
Henckens, M.L.C.M., Ryngaert, C.M.J., Driessen, P.P.J., Worrell, E., 2018. Normative Tilton, J., Guzman, J.I., 2016. Mineral Economics and Policy. RFF Press. https://doi.org/
principles and the sustainable use of geologically scarce mineral resources. Resour. 10.4324/9781315733708.
Policy 59, 351–359. https://doi.org/10.1016/j.resourpol.2018.08.007. Tilton, J., Skinner, B., 1987. In: Mclaren, D., Skinner, B. (Eds.), Resources and World
Henckens, M.L.C.M., van Ierland, E.C., Driessen, P.P.J., Worrell, E., 2016b. Mineral re- Development, Wiley.
sources: Geological scarcity, market price trends, and future generations. Resour. Tilton, J.E., 2018. The Hubbert peak model and assessing the threat of mineral depletion.
Policy 49, 102–111. https://doi.org/10.1016/j.resourpol.2016.04.012. Resour. Conserv. Recycl. https://doi.org/10.1016/j.resconrec.2018.08.026.
Hubbert, M.K., 1956. Nuclear energy and the fossil fuels, in: Drilling and Production Tilton, J.E., Crowson, P.C.F., DeYoung, J.H., Eggert, R.G., Ericsson, M., Guzmán, J.I.,
Practice 1956. San Antonio, TX, pp. 7–25. Humphreys, D., Lagos, G., Maxwell, P., Radetzki, M., Singer, D.A., Wellmer, F.W.,
Humphreys, D., 2013. Long-run availability of mineral commodities. Miner. Econ. 26, 2018. Public policy and future mineral supplies. Resour. Policy 57, 55–60. https://
1–11. https://doi.org/10.1007/s13563-013-0033-5. doi.org/10.1016/j.resourpol.2018.01.006.
Jasiński, D., Meredith, J., Kirwan, K., 2018. The life cycle impact for platinum group Tilton, J.E., Lagos, G., 2007. Assessing the long-run availability of copper. Resour. Policy
metals and lithium to 2070 via surplus cost potential. Int. J. Life Cycle Assess. 23, 32, 19–23. https://doi.org/10.1016/j.resourpol.2007.04.001.
773–786. https://doi.org/10.1007/s11367-017-1329-4. United Nations, 2017. World Populations Prospects.
Jevons, W.S., 1865. The Coal Question; An Inquiry Concerning the Progress of the Nation, Vieira, M.D.M., Goedkoop, M.J., Storm, P., Huijbregts, M.A.J., 2012. Ore grade decrease
and the Probable Exhaustion of Our Coal Mines. Macmillan & Co., London and as life cycle impact indicator for metal scarcity: The case of copper. Environ. Sci.
Cambridge. Technol. 46, 12772–12778. https://doi.org/10.1021/es302721t.
Jordan, B.W., Eggert, R.G., Dixon, B.W., Carlsen, B.W., 2015. Thorium: Crustal abun- Warell, L., Olsson, A., 2009. Trends and Developments in the Intensity of Steel Use : An
dance, joint production, and economic availability. Resour. Policy 44, 81–93. https:// Econometric Analysis. Paper presented at Securing the Future and 8th ICARD,
doi.org/10.1016/j.resourpol.2015.02.002. Skellefteå, Sweden. URL https://www.diva-portal.org/smash/get/diva2:1001923/
Kesler, S.E., Wilkinson, B.H., 2008. Earth's copper resources estimated from tectonic FULLTEXT01.pdf.

13
E. Castillo and R. Eggert Resources, Conservation & Recycling 161 (2020) 104896

Wellmer, F.-W., Scholz, R.W., 2018. Peak gold? Not yet! A response to Calvo et al. (2017). Woodhouse, M., Goodrich, A., Redlinger, M., Lokanc, M., Eggert, R., 2013. The Present,
Resour. Conserv. Recycl 134, 313–314. https://doi.org/10.1016/J.RESCONREC. Mid-Term, and Long-Term Supply Curves for Tellurium; and Updates in the Results
2017.11.015. from NREL's CdTe PV Module Manufacturing Cost Model.
West, J., 2011. Decreasing Metal Ore Grades: Are They Really Being Driven by the World Metal Statistics, 2018. Copper Worldbook.
Depletion of High-Grade Deposits? J. Ind. Ecol. 15, 165–168. https://doi.org/10. Yaksic, A., Tilton, J.E., 2009. Using the cumulative availability curve to assess the threat
1111/j.1530-9290.2011.00334.x. of mineral depletion: The case of lithium. Resour. Policy 34, 185–194. https://doi.
Wood Mackenzie, 2017. Copper Mine Costs Workbook Q42016. org/10.1016/j.resourpol.2009.05.002.
Wood Mackenzie, 2015. Copper grade evolution and interpretation. London.

14