Critical Review In-Use Stocks of Metals: Status and Implications M I C H A E L D . G E R S T † A N D T . E . G R A E D E L * ,‡ Environmental Engineering Program, School of Engineering and Applied Science, and Center for Industrial Ecology, School of Forestry and Environmental Studies, Yale University, New Haven, Connecticut 06511
Received February 11, 2008. Revised manuscript received July 10, 2008. Accepted July 21, 2008.
The continued increase in the use of metals over the 20th century has led to the phenomenon of a substantial shift in metal stocks from the lithosphere to the anthroposphere. Such a shift raises social, economic, and environmental issues that cannot be addressed without quantifying the amount of stock of ″metal capital″ utilized by society. Estimation of the inuse stock of metals has occurred for at least 70 years, with over 70% of the publications occurring after the year 2000. Despite the long history, this is the first critical review to consolidate current findings, critique methods, and discuss future avenues of research. Only aluminum, copper, iron, lead, and zinc have been studied to any extent. Nonetheless, it is clear that for the more-developed countries, the typical per capita in-use metal stock is between 10 and 15 t (mostly iron). Comparison of the per capita stocks in more-developed countries with those in less-developed countries suggests that if the total world population were to enjoy the same per capita metal stock levels as the more-developed countries, using a similar suite of technologies, the amount of global in-use metal stocks required would be 3-9 times those existing at present.
Introduction Determining the in-use stock of a particular resource appears to be a straightforward concept. As with most concepts that have an application to multiple scientific disciplines, however, defining and reviewing the concept and measurement of a stock turns out to be complex and challenging. Matters are not improved when narrowed to the specific discipline in which this paper resides, material flow analysis. Only a decade ago, a comprehensive review of the historical influences of societal metabolism found the concept of in-use stock to be ill-defined (1, 2), even though estimations of the in-use stock of metals can be found as far back as the 1930s (3). Since then, the literature on material flow analysis and in-use stocks of materials has experienced exponential growth. For in-use stocks of metal, the increase in attention can largely be traced to the effect of Baccini and Brunner (4) on bringing the concept of measuring in-use stocks of materials to a wider audience. Previous to their work, quantification of in-use stocks of metal was sporadic (5-13) and largely functioned as a novelty within the mining engineering and mineral economics communities. Thus, a long-standing, but recently burgeoning body of literature on the in-use stock of metals now exists. A survey * Corresponding author phone: (203) 432-9733; fax: (203) 4325556; e-mail:
[email protected]. † School of Engineering and Applied Science. ‡ School of Forestry and Environmental Studies. 7038
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008
of the literature (54 studies) shows that in-use metal stock estimation has the potential to address several interesting and pertinent questions concerning the future of anthropogenic metal use and associated environmental impacts, such as the following. (1) Are there patterns in the way different societies or nations use, accumulate, and discard metal that is relevant to investigating future demand scenarios? (2) Does the comparative quality and amount of metal in natural and anthropogenic stocks change how we should think about the sources of future metal supply? Should we think about large-scale mining of cities instead of virgin ore? (3) How will the scale of environmental impacts from dissipative uses (corrosion, wear, etc.) change with increasing population growth and affluence? (4) Given that metal is ultimately discarded, do waste management systems currently have the necessary information and capacity to handle waste from the metal cycle? Answering these questions with rigor is hindered because no comprehensive critical review of in-use metal stock estimates exists to provide an illumination of methodological commonalities, as well as to critique the current state-of-the artsa necessity for a relatively young and quickly expanding field. We herein address this gap by presenting a critical review of the literature on in-use metal stocks. First, a formal definition of in-use metal stocks is described, followed by a discussion of the estimation methodology. Next, an overarching view of the extant in-use metal stock estimates is given, with an accompanying discussion of the treatment of uncertainty and validation. Then, future avenues of research are laid out in the context of potential users of in-use stock information. Finally, a simple application of the extant inuse metal stock data is explored by comparing the amount of in-use metal stock required for less-developed countries to acquire the same amount of per capita in-use metal stock currently employed in more-developed countries.
Formal Definition of In-Use Stocks Put simply, material flow analysis (MFA) characterizes and quantifies flows of materials into, out of, and through a system of interest, balancing all flows by conservation of mass. By this description, MFA is certainly not a new concept; FischerKowalski (1, 2) highlights the influence upon it of such wellestablished fields as physics, economics, and cultural anthropology. Moreover, MFA’s novel contribution is not its mathematical formulation, but its application of pre-existing scientific paradigms to a previously uninvestigated spatial and temporal scale: the intergenerational temporal scale of 10.1021/es800420p CCC: $40.75
2008 American Chemical Society
Published on Web 09/04/2008
the anthropogenic (i.e., human-made) world as an embedded system within the natural world. Framing the contribution of MFA in this manner suggests that the choice of scale, level, and extent is central to the field. (These terms, often misused, are carefully defined by Gibson et al. (14).) In defining a general theory of stocks, Faber and others (15) touch on the importance of these factors in terms of set theory. Their observation is that the placement of the observer of the system and the choice of time extent define the distinction between stock and flow variables. Placement implies a spatial scale; customary levels along this scale for in-use stock determinations are cities, countries, regions, and the entire planet, and this choice establishes the system boundary. The operable definition of stock that follows is that a given quantity of mass is considered stock if and only if, from the perspective of the observer, the mass does not cross a system boundary during the entire time period extent of interest. During the next period of observation this mass can remain as stock or leave the system and become outflow for that period of time. Conversely, a quantity of mass that remains outside the system boundary during a time period is considered stock outside the system. During the next period of observation this mass can remain as stock or enter the system and become an inflow for that period of time. In terms of units of measurement, stock is a level variable (i.e., kg), while flow is a rate variable (i.e., kg/unit of time). By precise choice of the system boundary and temporal extent, the generalized definition of stock can be transformed into a specific definition of the in-use stock of metals. Thus, for any given discrete time period, in-use stock is defined as the matter within any final commodity (an ″economic entity of matter with a positive or economic value″ (16)) that is used by a human population during the entire time period. A final commodity is one whose form requires no further alteration before being delivered to the end-user. For example, a house provides the service of shelter, which people use during an entire time period. The materials that become the house are commodities, but not final commodities, because they require further alteration and/or integration. If the time period chosen is sufficiently long, then these intermediate commodities are clearly not stock of any kind and are an inflow. Once the house is demolished, the house ceases to exist and the contained materials are no longer in-use stock. The discarded flow of materials can be reused or recycled back into a final commodity, which is subsequently input into in-use stock. If this route is not chosen, discarded flow is deposited in stocks of discarded materials or is released across the system boundary into the natural world. The metal portion of in-use stock can be defined in two ways. If an individual element is specified, the in-use stock of metal refers to the total mass of that element, regardless of its chemical form. If a metal alloy is specified, the in-use stock of metal refers to the total mass of that alloy (including all its constituent elements).
Methodology Because scale, level, and extent are important to defining in-use stock, they are important as well to the methodology. Different choices for temporal extent could yield very different estimates of in-use stock. Take, for example, the two widely divergent extents of 1 min and 100 years, with a constant system boundary set at the geopolitical borders of a typical city and an initial in-use stock of zero. For an extent of 1 min, most metal-containing final commodities will satisfy the inuse stock condition of remaining within the city during the entire time period. In contrast, few metal-containing final commodities (with the exception of some long-lived build-
ings) will remain in-use during the entire time period of 100 years. Thus, the 1 min extent will yield an in-use stock which is nearly equal to the input at the beginning of the period, while the 100 year extent will yield an in-use stock of nearly zero. For reasons that will be elaborated upon later, most studies either explicitly or implicitly choose an extent of one year. System boundaries must also be chosen with care. Typically, system boundaries are interpreted to be spatial and correspondent with a predefined geopolitical boundary (e.g., city or country). Spatial boundaries are convenient from a data collection standpoint and also allow for straightforward normalization of in-use stock estimates. Normalization by, for example, people or area within a system boundary is often desirable when the relative states of different systems are compared. However, using strict geopolitical boundaries to define a system can have unintended, and sometimes difficult to reconcile, consequences. Consider again the definition proposed for in-use stock: the matter within any final commodity that is used by a human population during the entire time period. Under the previous example of using a city as a system boundary, a power plant or an industrial enterprise outside the city would not be considered in-use stock even though it is used by the city population in the form of electricity or employment, respectively. Is the system boundary of a city or any other geopolitical border therefore inappropriate? The answer depends on the problem for which the in-use stock estimate is used. If one seeks to create a map of in-use stock for the purpose of providing insight into future scrap flows, then a geopolitical or any other sort of spatial boundary seems appropriate. Depending on the spatial resolution, in-use stock estimates could be normalized by areassomething akin to creating a map of density or spatial concentration (see ref 17 for a recent example). However, if one seeks to infer conclusions about population affluence via in-use stock, a system (such as city borders) that is too small might be inappropriate for some metals. For example, a small spatial boundary might be appropriate for a metal that is mostly contained in residential buildings, but not for metals which have significant stocks in infrastructure or manufacturing equipment that could be outside the city border. Even if a spatial boundary as large as a country is chosen, stocks in a different country or continent, such as manufacturing equipment, could still be thought of as being used by the population in the system boundary. One answer is to consider this a ″hidden″ stock, much as the concept of hidden flows has been developed in the material flow analysis literature (18). Once the appropriate parameters have been chosen, the next step is to define an estimating procedure. This quantification is a considerable challenge, because there are no convenient, regularly collected data that can be drawn upon. Instead, methods initially developed 70 years ago (and described below) must be employed for the purpose. For estimation of in-use metal stock, the earliest example appears to be that of Bain (3), who produced the simplest example of what would later become known as ″top-down″ in-use stock estimation. Top-down estimations take information regarding flows and infer in-use stock by the cumulative difference between inflow and outflow. Mathematically, if St is stock at time t, then in discrete time steps T
St )
∑ (inflow - outflow ) + S t
t
(1)
o
To
where To is the time of the initial time step, T is the current time step, and So is the extant stock at the initial time step. Typically, the range from To to T is 50-100 years or longer. VOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7039
This yields the result that St is much larger than So, making the contribution of So negligible and therefore unnecessary to include in practice. As in ref 3, the simplest application of the top-down methodology involves collecting inflow data on metal production per year. Outflow is determined by multiplying either the current inflow or past inflows by a constant multiplier between 0 and 1. The physical interpretation of the multiplier is that it is the fraction of inflow that is discarded. It is a rather simplistic model of discard from in-use stock. However, this simplifying assumption is generally necessary due to a lack of data on outflows. In contrast, more recent top-down methods are utilizing increasingly complex methods. With the help of computing power and better access to data, contemporary top-down studies have been able to disaggregate metal production into inflow of specific final commodity categories and then model discard by defining final commodity lifetime functions. While an improvement, this method is still completely dependent on inflow data, because historical outflow data are poor to nonexistent. Future efforts to collect historical data on outflows would allow for a well-needed empirical check on the results created by discard models (see ref 19 for an example). The types of available data have a significant effect on the system boundary defined for top-down analyses. Time steps are typically one year, because most applicable data are available as per year flows. The data used for inflow are gathered from government documents, technical literature, expert elicitation, and industry trade organizations. Little or no discussion is generally spent on the relative reliability of these various sources. Only a few years after publication of ref 3, a different method introduced by Ingalls (5) appeared in the literature. This method, now termed ″bottom-up″, takes a strategy opposite that of top-down methods because it gathers information on stock variables to estimate in-use stock and (if desired) infer the behavior of flows. In its simplest form, estimating in-use stock via the bottom-up method is represented by A
St )
∑N m it
it
(2)
i
where Nit is the quantity of final commodity i in use at time t, mit is the metal content of in-use final commodity i, and A is the number of different types of final commodities. More complicated versions of the bottom-up method retain the same formulation of eq 2, but allow for more precise definitions of the commodity categories and metal contents. The types of available data also have a significant effect on the system boundaries of bottom-up approaches. As with the top-down method, most data collected are interpreted as being representative of a one year time period. In this case though, the data are of stock existent during one year instead of the rate of flow per year. Also in similarity with the data collected for the top-down approach, data collected for the bottom-up approach are often constrained by geopolitical boundaries. However, this constraint is often not as strict for bottom-up methods, because stock-relevant data (e.g., houses or cars) are frequently available at the city/town or lower levels of aggregation. In recent years, the use of GIS data on public infrastructure and other stocks of interest has improved the overall accuracy and opened the door to spatial resolution of bottom-up stocks. In these cases, where high spatial resolution of Nit is acquired, precision is limited by metal content determinations. For many new final commodities entering use, such as machinery, electronics, and cars, metal content can be measured or obtained from manufacturers. Commodities 7040
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008
such as buildings are more difficult because they are seldom mass-produced. Moreover, using the metal content of new commodities to infer the metal content of in-use stock can be problematic, because in-use stock contains an amalgam of stock vintages that have been accumulated over time. For example, applying the copper content of a newly built house to a house built in the 1950s would yield a misleading result. Alleviation of this data problem is difficult, because it requires either a statistically significant regime of sampling of the metal content of in-use stock or a significant historical investigation of engineering design plans for various commodities (see ref 20 for an approach to this problem). To summarize, although both methods seek to measure the same quantity, the total in-use stock, they work in fundamentally different ways, and current data sets and technology are more appropriate for some goals than others. Inherent in the top-down method is evolution over time. As a result, that approach is often employed where inflow/ outflow time series data are available. Because the top-down method is, to some extent, a more derived estimate of stock than the bottom-up method, it could be viewed as less precise. However, the bottom-up method is hampered by an inability to count everything in use and by uncertainties as to the metal content of many of the units being included in the assessment. This tradeoff must be considered in the context of the study goal. If a more detailed and spatially explicit analysis is desired, then the bottom-up method is likely to be more desirable. If less detail is acceptable, and a larger temporal and spatial scale is the focus, then the top-down method may be the more appropriate method to employ.
In-Use Stock Estimates: A Status Report With the definition of in-use stock and estimation methodologies in hand, comparison of metal in-use stock estimates from the literature can now be made. A review of the scholarly literature, as well as industry publications and presentations, yielded a total of 54 studies of metal in-use stocks. The publication dates range from 1932 to 2007, with 70% of those publications occurring after the year 2000. A total of 24 different metals (22 elements and 2 alloys) have been addressed, with a total of 124 metal stock in-use estimates. The individual data assembled in our review are too numerous to display herein. Instead, we include in the Supporting Information two tables: a specification of the principal final goods categories in which each metal resides when in use (information necessary for bottom-up studies) and a listing of all the extant per capita in-use stock determinations, by locale, applicable year, amount, and reference source. Interested readers may access this information for the individual values. We content ourselves here with a presentation and discussion of the overall picture shown by the data. Because the collected studies present estimates of different metals with varying system boundaries, we chose to normalize by population because interesting conclusions about affluence, spatial scale, and the level of metal in-use stock can be drawn when in-use stock is discussed on a per capita basis. The results of our literature review are summarized in Table 1. Copper, lead, zinc, and iron are the metals for which estimations have most frequently been made. Geographically, an overwhelming majority of the estimates concern moredeveloped countries. On a global basis, top-down estimates for in-use stocks exist for only five elements (4, 21-25). The per capita stock of iron is largest, as befits its high rate of flow into use. Aluminum’s stock is slightly larger than that of copper, and that for lead is significantly lower. There is a quite small stock estimate for cadmium; it is seriously out of date and cannot be regarded with confidence.
TABLE 1. Extant In-Use Metal Stock Estimationsa metal
number of estimates
fraction of all estimates (%)
aluminum antimony cadmium chromium cobalt copper gold iron lead magnesium manganese mercury molybdenum nickel palladium platinum rhodium silver steel stainless steel tin titanium tungsten zinc
9 1 3 3 1 34 2 13 20 1 1 1 1 3 2 2 1 2 1 5 2 1 1 14
7.4 0.8 2.5 2.5 0.8 27.0 1.6 10.7 16.4 0.8 0.8 0.8 0.8 2.5 1.6 1.6 0.8 1.6 0.8 4.1 1.6 0.8 0.8 11.5
global per capita stock 80 40 35-55 2200 8
110
MDC per capita stockb 350-500 1 80 7-50 1 140-300 35-90 7000-14000 20-150 5 100 10 3 2-4 1-4 1-3 0.2 13 7085 80v180 3 13 1 80-200
LDC per capita stockc 35
30-40 2000 1-4
15
20-40
a
The years of the determinations vary, but most are for the period 2000-2006. The units of per capita stock are kilograms of metal in most cases, but grams of metal for cadmium, gold, mercury, palladium, platinum, rhodium, and silver. b The more-developed countries (MDCs) used in this calculation are Australia, Canada, the European Union EU15, Sweden, Switzerland, Japan, New Zealand, and the United States (altogether about 860 million people in 2005). c The less-developed countries (LDCs) used in this calculation consist of all countries except those in the ″more-developed″ category (altogether about 5620 million people in 2005).
Few stock estimates are available on a regional basissonly aluminum for Europe (26), copper for North America (27) and Europe (19), and stainless steel (28) for several regions (these are top-down estimates). The stock evaluations are all quite recent and appear reasonably reliable. On a country basis, few data exist except for Japan (the most extensive in terms of elemental diversity), the United States, Australia, and several European countries. Urban-level evaluations have been performed for several metals and several cities, all on a bottom-up basis. The locales are diverse: Stockholm (29), Vienna (30), New Haven (31, 32), Cape Town (33, 34), Beijing (35), and Sydney (17). There are still too few bottom-up studies to evaluate the accuracy of this approach, although where comparisons of the results can be made, they appear reasonable. At the city level, it is possible to allocate the stock on a spatial basis, and this has been done for Cape Town and Sydney. For the MDCs, there is sufficient information in Table 1 to compute an average person’s allocation of in-use metal stock. Doing so demonstrates the dependence of the MDC lifestyle on a substantial stock of metal, a perspective especially useful because few individuals are aware of their share of metals in infrastructures such as communications, rail networks, and power generation and distribution systems, as well as in their personal homes and vehicles. We find that the average MDC citizen’s in-use metal stock is between 10 and 15 t. Of this amount, five metalssiron, aluminum, copper, zinc, and manganesesmake up more than 98%. One feature of the data is that there appear to be significant differences in the per capita stock of more-developed and less-developed countries (Table 1). Additionally, when measured at different spatial levels, the urban residents appear to have larger per capita stock than do rural residents, although this is an observation that needs input from future research on urban/rural in-use stocks. Given the obvious wealth discrepancies (and perhaps cultural differences as
well), these results seem qualitatively reasonable, but much work remains to be done to better understand stocks from a spatial (or spatial analytical level) perspective. An additional factor is that taking into account an entire nation may increase the per capita in-use stock due to infrastructure and other in-use final commodities that occur only in sparsely populated areas. Such final commodities may include ships, large trucks, heavy industrial equipment, offshore drilling equipment, military hardware, and aircraft. For perhaps only copper, iron, aluminum, and lead, and only for the more-developed countries, can we feel we have enough information to give us a reliable estimate of the stocks of metal in use. These are generally the most widely employed metals, however. For many others, the order of magnitude of their in-use stocks can readily be appreciated from the information assembled in this paper. In the case of silver, a single estimate for Japan is much lower than a rather old global estimate; we suspect the latter may be unreliable. However, what is clearly lacking from Table 1 and from many of the individual studies is a quantification of the uncertainty of the in-use stock estimates. A comparison to another scientific field, estimating the elemental composition of the continental crust, is especially useful here because estimating the elemental composition of the continental crust can be thought of as a natural analogue to estimating in-use stock. For this purpose, we employ Rudnick and Gao’s review (36) of the relevant literature. In this work, the authors compile data to create an authoritative account of the average elemental concentration of the upper, middle, and lower continental crusts. Along with average concentration values, corresponding standard deviations are givensa quantification of uncertainty. Comparison of the Rudnick and Gao study and those of in-use stock estimates suggests several reasons for the relative lack of uncertainty analysis. The first is the relative volume and maturity of the VOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7041
literature. Rudnick and Gao cite several hundred sources in their review, which are in turn based upon multiple sets of experimental data and literature citations. This creates a significant data set upon which classical statistical analysis can be applied. In contrast, our review has located only 54 studies, many of which cover only one or two metals, and few of which make any attempt to quantify uncertainty. In addition, several have not been published in the peerreviewed literature. A second reason for the difference in treatment of uncertainty is firmly rooted in methodological differences. At the core of the individual geochemical investigations are sampling protocols that involve taking direct measurements of the crust. If undertaken properly, sampling is repeated frequently enough to establish a statistically meaningful concentration value along with its accompanying uncertainty. As discussed in the methodology section of the present paper, in-use stock estimates are mostly reliant on published data concerning inflows, commodity lifetimes, metal content, and stock of metal-containing commodities. Frequently these data do not explicitly quantify uncertainty, and crosschecking across multiple sources is often not an option due to limited data availability. In a methodological context, inability to cross-check sources makes validation of the results difficult. As an alternative, one could make educated guesses as to the shape and degree of uncertainty and then infer through Monte Carlo analysis the uncertainty of the final in-use stock estimate (see ref 20, for example). However, educated guesses of first-order uncertainty (the parameters of the probability density function) are often subject to significant secondorder uncertainties (the uncertainty of the probability density function shape). Despite the problems of assuming degrees of uncertainty, a clear improvement of most in-use stock estimates would be a more directed approach to uncertainty analysis, especially if the estimates are to have policy relevance.
Potential Utility of In-Use Stock Determinations Quantification of in-use stocks of metals, while of sometimes academic interest, has thus far demonstrated rather little usefulness despite its 70 year history. A look back at Kenneth Boulding’s classic essay, ″The economics of the coming spaceship Earth,″ reveals that the relationship between inuse stocks and environmentally relevant flows was starting to be brought to attention in the 1950s and 1960s (37). In his essay, Boulding places the idea of stocks as central to the paradigm shift from a high-waste to a low-waste economy: “The essential measure of the success of the economy is not production and consumption at all, but the nature, extent, quality, and complexity of the total capital stock... what we are primarily concerned with is stock maintenance, and any technological change which results in the maintenance of a given total stock with a lessened throughput is clearly a gain.” Considering that the diffusion of scientific knowledge can take many decades (38), perhaps the lack of perceived usefulness is the result of a quickly maturing field that has not been explored in ways easily understood as relevant to decision-makers. Thus, although it may be flows that directly cause environmental burdens, a lack of understanding of the relationship of metal flow to changes in in-use metal stock is problematic from the perspective of providing information to decision-makers. With this in mind, it is now prudent to ask what groups of decision-makers might be expected to benefit from this information and why they have not done so. Mining Industries. The mining industries extract ore minerals. Their interest is primarily in the future demand for virgin metals, a demand that is enhanced by increasing per 7042
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008
capita resource intensity throughout the world and reduced by recycled scrap that can substitute for virgin metal. Perspectives potentially useful to these industries include metal in-use stock estimates to measure the stock of metal required to deliver any given service to a population and the creation of scenarios of potential metal demand based on different assumptions of technology choice, population growth, and other relevant parameters. Failure to assemble and publicize existing in-use stock estimates, a lack of metal demand scenarios, and the absence of stock discard scenarios have to date prevented the metal production industries from utilizing in-use stock information in these ways. Metal Production Industries. These industries transform impure metal from either ore or scrap into metal of desired purity. They could benefit from scenarios of discards from stock in use, especially if the form of the metal (alloy, coated metal, etc.) were part of the scenario. Current in-use stock studies provide the basis for scenario development, but the work remains to be done. Waste Management and Scrap Industries. The function of the waste management and scrap industries is to collect metals at discard and then either sort them for sale to the secondary metals industry or inter them in landfills for longterm containment. The factors determining which path a metal might take include the concentration and speciation of metal in discarded commodities, the ease of separation and concentration, and by what manner the metal is discarded (separated or mixed). Thus, waste management and scrap industries have as their basic inputs materials which are spatially heterogeneous and that embody a significant amount of uncertainty with regard to material content and timing of discard. Here, in-use stock information in itself is not of value, but discard scenarios linked to in-use stock with relatively high temporal and spatial resolution could be. Public Health and Environmental Agencies. Unlike the potential users discussed above, who are concerned with the quantity and form of metals recovered at discard, these agencies are concerned with the quantity and form of metals that are discarded or dissipated and not recovered. As with the waste management industries, they would value discard scenarios with high spatial resolution to predict effects on public health and/or the environment. Public Policy Organizations. These organizations are diverse, with many different goals. Economic and national security policy-makers are often concerned with having adequate scrap supply for certain metals in the case of disruption of trade in either metal commodities or final goods containing metals. Environmental policy-makers, at least in the context of resource efficiency, are concerned with promoting the environmental benefits of metal recycling. The organizational diversity is reflected in the uses the organizations might make of in-use stocks information or of the products of such information. As a consequence, they would value scenarios at a variety of temporal and spatial scales and with details on the physical form of discards. What emerges from a review of potential users of information on in-use metal stocks is that there are several major avenues of research that need to be explored to increase applicability of the science of in-use stocks to decisionmakers. The first is a greater disaggregation of final commodity categories into more specific final commodity or constituent intermediate commodities. For bottom-up estimates, disaggregation is not so much an issue because of the type of data collected. For top-down estimates, however, further disaggregation can present significant data challenges because of the method’s time-series nature. In either case, more specificity in the types of final commodities which are in-use and eventually discarded is crucial to providing useful
FIGURE 1. (a) 2005 production of metals, in order of absolute amount. (b) Number of in-use stock determinations for the metals of (a). information to decision-makers. For example, contrast the usefulness of the information of just stock and flow of copper in buildings to information on intermediate commodities contained in a building, such as wiring or piping. Each intermediate commodity has different properties that necessitate different decisions or policies that would no doubt be more effective than a broad decision or policy on just buildings. The second avenue of research to be explored is that of spatially explicit in-use stock estimation. Van Beers and Graedel (17) have shown that relatively fine spatial resolution can be obtained through the utilization of GIS and spatially explicit stock and population data. However, no studies to date have been able to exploit the combination of GIS and remote sensing in the determination of in-use metal stock. Such work could be advantageous, as the results could be linked with other spatial attributes useful to decision-makers. The last avenue of research pertains to how to combine the positive attributes of both top-down and bottom-up methodologies. Clearly, bottom-up methodologies excel at providing a more detailed picture of in-use stock, while topdown methodologies include the element of time-dependent stock dynamics. One possible way, as illustrated by Wittmer and Lichtensteiger (20), is to estimate stock at multiple periods in the past. Inflows and discards can then be calculated by using the vintage structure of in-use stock as well as final commodity lifetimes. This is a promising construct that could prove especially amendable to producing future in-use stock scenarios.
Discussion A relationship that is of interest to explore is whether the effort expended on in-use stock determinations of individual metals is related to their relative intensity of production. We address this issue in Figure 1, where at the top of the figure we plot in order of absolute magnitude the 2005 production (39) for 32 metals. Below that display, we plot the number of stock determinations in the same sequential order. It is immediately apparent that the relationship between the two is weak. Relative to its production, rather little attention has been given to aluminum. Even more dramatic is the neglect of no. 4 manganese (one stock determination) and no. 7 silicon (no determinations). Other elements whose stock is completely unstudied include no. 16 arsenic, no. 18 vanadium, no. 19 niobium, no. 20 lithium, and no. 24 yttrium. In
FIGURE 2. Estimates of current stock in use of five metals in MDCs, LDCs, and the world, together with the estimated in-use stock that would be required were all people in the world to have the same per capita in-use stock as those in MDCs. contrast, no. 8 lead has more stock determinations than any other metal except copper. If we are to regard in-use stocks as the ″mines of the future″ (40), it would seem appropriate to invest some effort in estimating the in-use stocks of a wider range of metals than is currently the case and doing so in a variety of different countries and regions. In-use stock determinations vividly demonstrate the linkage between metal stock and the state of technological development of a country or region. For the metals where in-use stock has been determined in both developed and developing countries or regions, the stock in the former is typically 5-10 times higher on a per capita basis. An interesting exception to the trend is silver. Per capita in-use silver stocks tend to be much larger in the developing world because of high levels (relative to those in the developed world) of silver used in jewelry and decorations. This leads us to ask what the stock would need to be if the entire global population were to be receiving the services provided by metal stocks at the same level as the developed world and with the same technology. Figure 2 presents a preliminary exploration (due to limited stock data) of this question for the year 2000 (approximately the year for which most of the stock determinations apply) for the five metals for which data permit: aluminum, copper, iron, lead, and VOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7043
zinc. The result is that the in-use stock of metal would have to increase by a factor of between 3 and 9. This is a significant amount that will ultimately lead to increases in exploration, mining of lower grade ores, material substitution, and recycling. The ways in which these options are utilized, as well as technological change, will result in a variety of environmental impacts. Public and private policy, among other things, will influence these future choices and could greatly benefit from knowledge of in-use stock dynamics. All potential applications of in-use stock information require significant improvements in the science of producing in-use stock estimates. From the preceding overview of the literature and potential end-users, a few commonalities for improvement have emerged. The first is that the overall methodology would benefit from a blending of the top-down and bottom-up methods, incorporating the detail of bottomup methods with the temporal aspect of top-down methods. The second is that regional specificity must increase. A large majority of the studies reviewed in this paper are from developed countries, whose values are not representative of developing countries. The third is that spatial resolution and greater disaggreation of final goods categories will become increasingly important for some end-users. Finally, the fourth common focus for future research is a more directed approach to quantifying and communicating the uncertainty and validation of the results. These four research foci are, of course, important to different applications, but as a whole will serve to concentrate research time and resources where they will make the highest impact. In summary, the present work brings together the extant information on in-use stocks of metals. In some cases, a good deal of consistency is shown to exist. In other cases, it is clear that the stocks have yet to be well characterized. A review of potential users makes it clear that in-use stock information by itself has limited utility. It is only when it is used to generate scenarios of future use intensity, discard, and reuse, with good spatial and temporal resolution and final commodity disaggregation, that its value is manifest. Such data and scenarios are now beginning to appear. In providing perspective and relative magnitudes, the results are of inherent use as they stand, but they point to a large challenge aheadsto do a better job of evaluating stocks and their rates of growth and decay and to use that information to make informed inferences about the future. Only by so doing can we have a fully adequate picture of the rich ″anthropogenic mines″ that have the potential to be tapped as sources of metal for the uses of modern society.
Acknowledgments We thank Robert Gordon, Daniel Mu ¨ ller, Seiji Hashimoto, Gerd Rebernig, and three anonymous reviewers for providing relevant literature and for helpful discussions and comments. The contribution of M.D.G. was supported by the Grainger Foundation.
(4) (5)
(6)
(7)
(8)
(9)
(10)
(11) (12) (13) (14)
(15) (16) (17)
(18)
(19)
(20) (21) (22) (23)
Supporting Information Available Two tables containing background information on different metal uses and lifetimes and all of the in-use stock estimates used for this paper. This material is available free of charge via the Internet at http://pubs.acs.org.
(24) (25) (26)
Literature Cited (1) Fischer-Kowalski, M. Society’s metabolism: The intellectual history of materials flow analysis, part I, 1860-1970. J. Ind. Ecol. 1998, 2, 61–78. (2) Fischer-Kowalski, M.; Hu ¨ ttler, W. Society’s metabolism: The intellectual history of materials flow analysis, part II, 19701998. J. Ind. Ecol. 1999, 2, 107–136. (3) Bain, H. F. The rise of scrap metals. In Mineral Economics; Lectures under the Auspices of the Brookings Institution, 1st ed.; 7044
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008
(27)
(28) (29)
(30)
Tryon, F. G., Eckel, E. C., Eds.; McGraw-Hill: New York, 1932; p 331. Baccini, P.; Brunner, P. H. Metabolism of the Anthroposphere; Springer-Verlag: Berlin, 1991; p 157. Ingalls, W. R. The Economics of Old Metals Especially Copper, Lead, Zinc, and Tin; Mining and Metallurgical Society of America, Columbia University Club: New York, 1935; pp 75-95. Miller, H. J. The supply and industrial applications of scrap metals. Proceedings of the United Nations Scientific Conference on the Conservation and Utilization of Resources, Lake Success, NY, 1949; United Nations Department of Economic Affairs: Lake Success, NY, 1949. Merrill, C. W. The accumulation and conservation of metalsin-use. Proceedings of the United Nations Scientific Conference on the Conservation and Utilization of Resources, Lake Success, NY, 1949; United Nations Department of Economic Affairs: Lake Success, NY, 1949. Brown, H. The Challenge of Man’s Future; An Inquiry Concerning the Condition of Man during the Years That Lie Ahead; Viking Press: New York, 1954; p 290. Merrill, C. W. The significance of the mineral industries in the economy. In Economics of the Mineral Industries; A Series of Articles by Specialists, 1st ed.; Robie, E. H., Ed.; American Institute of Mining, Metallurgical, and Petroleum Engineers: New York, 1959; p 755. Merrill, C. W. The significance of the mineral industries in the economy. In Economics of the Mineral Industries; A Series of Articles by Specialists, 2nd ed.; Robie, E. H., Ed.; American Institute of Mining, Metallurgical, and Petroleum Engineers: New York, 1964; p 787. McMahon, A. D. Copper: A Materials Survey; United States Bureau of Mines: Washington, DC, 1965; p 340. Sousa, L. J. The U.S. Copper Industry: Problems, Issues, and Outlook; United Stated Department of the Interior: Washington, DC, 1981; p 86. Chapman, P. F.; Roberts, F. Metal Resources and Energy; Butterworths: London, 1983; p 238. Gibson, C. C.; Ostrom, E.; Ahn, T. K. The concept of scale and the human dimensions of global change: A survey. Ecol. Econ. 2000, 32, 217–239. Faber, M.; Frank, K.; Klauer, B.; Manstetten, R.; Schiller, J.; Wissel, C. On the foundation of a general theory of stocks. Ecol. Econ. 2005, 55, 155–172. Brunner, P. H.; Rechberger, H. Practical Handbook of Material Flow Analysis; Lewis Publishers: New York, 2003; Vol. 1, p 336. van Beers, D.; Graedel, T. E. Spatial characterisation of multilevel in-use copper and zinc stocks in Australia. J. Clean. Prod. 2007, 15, 849–861. Hammond, A.; Adriaanse, A.; Bringezu, S.; Moriguchi, Y.; Rodenburg, E.; Rogich, D.; Schu ¨ tz, H. Resource Flows: The Industrial Basis of Industrial Economies; World Resources Institute: Washington, DC, 1997; p 50. Ruhrberg, M. Assessing the recycling efficiency of copper from end-of-life products in western Europe. Resour., Conserv. Recycl. 2006, 48, 141–165. Wittmer, D.; Lichtensteiger, T. Exploration of urban deposits: Long-term prospects for resource and waste management. Waste Manage. Res. 2007, 25, 220–226. European Aluminum Association. Aluminum Recycling: The Road to High Quality Products; Brussels, 2004. Kapur, A. The Future of the Red Metal: Scenario Analysis and Implications for Policy. Ph.D. Dissertation, Yale University, New Haven, CT, 2004. Jolly, J. L. W. The US Copper-Base Scrap Industry and Its Byproducts: An Overview; United States Geological Survey: Washington, DC, 1999. Lichtensteiger, T. The Petrologic Evaluation; Springer-Verlag: Berlin, 2002. Mao, J. S.; Graedel, T. E. Lead in-use stocks: A dynamic analysis, submitted for publication. European Aluminum Association. Global Aluminum Recycling: A Cornerstone of Sustainable Development; Brussels, 2006. Spatari, S.; Bertram, M.; Gordon, R. B.; Henderson, K.; Graedel, T. E. Twentieth century copper stocks and flows in North America: A dynamic analysis. Ecol. Econ. 2005, 54, 37–51. Teroerde, F. ELG Metals, Duisburg, Germany. Personal communication, 2006. So¨rme, L.; Bergba¨ck, B.; Lohm, U. Century perspective of heavy metal use in urban areas. A case study in Stockholm. Water, Air, Soil Pollut.: Focus 2001, 1, 197–211. Obernosterer, R.; Brunner, P. H. Urban metal management:
(31)
(32) (33) (34) (35)
The example of lead. Water, Air, Soil Pollut.: Focus 2001, 1, 241–253. Drakonakis, K.; Rostkowski, K.; Rauch, J.; Graedel, T. E.; Gordon, R. B. Metal capital sustaining a North American city: Iron and copper in New Haven, CT. Resour., Conserv. Recycl. 2007, 49, 406–420. Rostkowski, K.; Rauch, J.; Drakonakis, K.; Reck, B.; Gordon, R. B.; Graedel, T. E. ″Bottom-up″ study of in-use nickel stocks in New Haven, CT. Resour., Conserv. Recycl. 2007, 50, 58–70. Van Beers, D.; Graedel, T. E. The magnitude and spatial distribution of in-use copper stocks in Cape Town, South Africa. South Afr. J. Sci. 2003, 99, 61–69. van Beers, D.; Graedel, T. E. The magnitude and spatial distribution of in-use zinc stocks in Cape Town, South Africa. Afr. J. Environ. Assess. Manage. 2004, 9, 18–36. Zhang, Z. Spatial analysis of in-use copper stocks in Beijing, China. In Cities as Opportunities; Bai, X., Graedel, T. E., Morishima, A., Eds.; Cambridge University Press: Cambridge, U.K., 2008.
(36) Rudnick, R. L.; Gao, S. Composition of the continental crust. In Treatise on Geochemistry; Holland, H. D., Turekian, K. K., Eds.; Elsevier: Amsterdam, 2003; Vol. 3, pp 1-64. (37) Boulding, K. The economics of the coming spaceship Earth. In Environmental Quality in a Growing Economy; Jarrett, H., Ed.; Johns Hopkins Press: Baltimore, MD, 1966; pp 3-14. (38) Bettencourt, L. M. A.; Kaiser, D. I.; Kaur, J.; Castillo-Cha´vez, C.; Wojick, D. E. Population modeling of the emergence and development of scientific fields. Scientometrics 2008, 75, 495– 518. (39) United States Geological Survey. Mineral Commodity Summaries; Department of the Interior: Washington, DC, 2007. (40) Kapur, A.; Graedel, T. E. Copper mines above and below the ground. Environ. Sci. Technol. 2006, 40, 3135–3141.
ES800420P
VOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7045
