INTRODUCTION

1 Although exploration and mining can be traced back to Neolithic times, the science of ore deposits is still a relatively recent field of study. Mining was one of the first areas worthy of intellectual investigation by Renaissance scholars, yet economic geology– in the present sense of the term– did not emerge until the end of the 19^th century in either Europe or the United States. The origins of the science revolved around two nuclei: the North American school, from which sprang the Economic Geology Publishing Company founded in 1905 (Skinner 2005), and the European school, from which the term ‘ metallogeny' was coined by de Launay (1913), the same year that W. Lindgren published his landmark book entitled, Mineral Deposits. Within a period of 100 years, a significant body of knowledge has emerged and economic geology has become an established university discipline. Many countries, like Canada, now require a degree to become a professional geoscientist, and links between economic geology and other geoscience disciplines continue to be forged around the world.

2 Geoscientific progress has been most remarkable since World War II, although we still lack the hindsight needed to fully appreciate this fact. The transformation of conceptual insights into mining discoveries is a slow process, and it generally takes several decades for the real economic impact of innovations in fundamental science to materialize. Regardless, the field of economic geology has expanded from one of descriptive methodology to an explanatory and predictive science applied to mineral resource exploration and utilization.

3 Several publications collectively synthesize the post-1980s understanding of economic geology. Following Routhier's (1980) essay on predictive metallogeny, contributions have come from Nicolini (1990) and Pélissonnier (2001) from France; Mitchell and Garson (1981) and Edwards and Atkinson (1986) from England; Lunar and Oyarzun (1991) from Spain; Hutchinson (1982), Sawkins (1984), Eckstrand et al. (1995), Kirkham et al. (1995) and Laznicka (1985) from Canada; Guilbert and Park (1986), Kesler (1994) and Misra (2000) from the United States; Pirajno (1992, 2000) and Robb (2004) from South Africa; Solomon and Groves (1994) from Australia; and Dardenne and Schobbenhaus (2001) from Brazil.

4 The study of ore deposits is primarily an applied science and is thus influenced by other discoveries in theoretical science as well as changes– often rapid– in related industries. The last twenty years were marked by economic globalization accompanied by the reorganization of political structures, notably those of Russia, China and South Africa. Such sweeping globalization caused profound changes in the basic funding structure and economic landscape of mineral exploration and development, influenced by fluctuating and unpredictable metals markets. These factors resulted in the perceived necessity for mining companies, even the largest, to regroup into increasingly larger entities. These changes in the industry were coincident with a radical reduction in the size of government geoscience programs in an attempt to reduce deficits. Combined with the gains in efficiency accorded by technological advancements, the net effect has been a worldwide drop in the number of mineral industry personnel from 18 million to 15 million in only one decade (Anonymous 2004).

5 The decline in the number of experienced exploration geologists in recent years has produced a precarious situation with negative repercussions for a profession that is already undervalued by the public, particularly after the Bre-X stock scandal in the late 1990s and the poor environmental reputation in most developed countries. Globalization has also led to job specialization at the international level. Also witnessed, is the demise of the traditional mining industry in parts of many industrialized countries, as much in Europe as in North America. Historical base-metal districts– once the economic backbone in frontier regions– have now closed: Noranda (Québec), Sullivan (British Columbia), Les Malines and St-Yrieix (France), Laisvall (Sweden), Leadville (Colorado), Touissit-Bediane (Morocco), and Tsumeb (Namibia). Figure 1 shows that a large number of the mines in developed countries, which were the focus of metallogenic studies in the 1970s, are now closed. Exploration in developed countries holds little promise, even if the geological potential is excellent, because there is slim prospect that a discovery can be turned into a mine despite the fact that recycling cannot meet either present or future resource needs. Instead, explorationists have concentrated their efforts in sparsely populated regions in the north (Scandinavia, Canada) and south (Spain, Mexico, Australia). These efforts have resulted in the discovery of new ore deposits, particularly in South America but also in the Pacific region and likely soon in China. Western Africa's potential is slowly emerging, whereas the former Soviet world waits in the wings.

6 This essay summarizes the prominent shifts and new ideas in metallogeny, which have emerged during the past 20 years and have been heralded as major advances in the field. The focus is on contributions published in French and English. It is an exercise that can only be incomplete: more than 20,000 articles have been published since 1980. Only the most pronounced trends are emphasized and the reader is asked to excuse any omissions, which are inevitable in such a short text. Progress in the field of mineral deposits will be summarized first, followed by an overview of the most significant advances in the key aspects of metallogeny: depositional conditions, mode of transport, and sources of ore elements.

Display large image of Figure 1

THE SCIENCE OF MINERAL DEPOSITS

7 The two terms ‘ metallogeny' and ‘ gitology' are used to describe the science of mineral deposits (Nicolini 1990). ‘ Metallogeny' was widely adopted by the French after being introduced by de Launay (1913); however, the word took on a genetic connotation and became fairly restricted to mineralogical and physio-chemical approaches. ‘ Gitology' was a term invented near the end of the 1960s by geologists from France's Bureau de recherches géologiques et minières to describe a more naturalistic approach based on geological context and descriptive classification (Nicolini 1970; Rabinovitch 2000).

8 Many deposit types have been studied in detail using significantly more quantitative methods that integrate an ever-increasing number of parameters. Some of the deposit types familiar today were, in fact, only recently recognized or re-interpreted: copper porphyries (Lowell and Guilbert 1970); Carlin-type disseminated gold (Radtke and Dickson 1976); volcanogenic massive sulfides (Parmentier and Spooner 1978); unconformity-type uranium (Hoeve and Sibbald 1978); sedimentary-exhalative deposits (Large 1980); diamondiferous kimberlites (Haggerty 1986); iron-oxide Cu-Au-U deposits (Hitzman et al. 1992); and breccia-hosted platinum-group-element deposits (Lavigne and Michaud 2001). It is for this reason that the reference volume for the 1980s– Economic Geology's 75th Anniversary Volume (Skinner 1981)– does not contain separate chapters for such currently important deposits as epithermal invisible gold, auriferous shear zones (orogenic gold), diamonds, hydrothermal platinum group elements, unconformity-type uranium, tantalum pegmatites, or iron oxide-copper-gold. The emergence of these deposit types was largely driven by the search for high-value commodities, like gold, platinum group elements and diamonds. The conceptualization of a new deposit type is undoubtedly one of the most important factors from an economic perspective: most mining companies will develop their exploration strategy based on a type-deposit approach and the most common geological and/or tectonic setting for an ore deposit.

9 The classification of deposits has evolved considerably during the past 20 years. There are, of course, a number of different approaches based on descriptive or genetic criteria, and whether the classification of the deposit is from an internal (mineralogical) or an external perspective (geological context). Routhier (1969) warned of the risks of using poorly controlled parameters that can lead to inconsistencies. Since that time, the classification of deposits has been progressively standardized around several main geologic and geodynamic themes. In some cases, new classifications have been elegantly reconciled with existing mineralogical and volcanological parameters (Sillitoe et al. 1996). Table 1 presents the classification system used at the Université du Québec à Montréal (UQAM) and the current level of understanding attained for each deposit type. During the past few decades, the ability to determine genetic age (i.e. the development of isotopic geochronometers) was one of the most significant factors in deducing ore genesis and has helped mediate battles between syngenetic and epigenetic proponents. The Deposit Modeling Program supported by IUGS and UNESCO (Cox and Singer 1986; Kirkham et al. 1995; Seal and Foley 2004) popularized the North American approach and eventually squeezed out Soviet concepts. In the future, the improved documentation of deposits that have yet to be discovered in the relatively poorly documented environments of Africa and Asia will undoubtedly lead to further revision of existing classifications.

10 The distribution of ore deposits has always occupied a prominent place in economic geology studies given the significant economic advantage for anyone who can predict the location of the next big mineable deposit. Routhier (1980) proposed a belt-style distribution for major mineral deposits; however, it was only after the development of geographic information systems in the 1990s that the analysis of ore deposit distribution became a true discipline, combining ore deposit models with empirical observations and supported by statistical analysis (Carlson 1991; Bonham-Carter 1994; Agterberg 1995). Many of the findings helped to scientifically confirm the existence of districts dominated by specific metals (districts that had been informally recognized since ancient times) and to define the exclusionary relationships between certain ore families. Nonetheless, much remains to be accomplished in this area, particularly in our understanding of the vertical distribution of deposits within continents, and their link to the orogenic cycle. It is reasonable to expect that crustal architecture (in terms of thermal structure, permeability, redox barriers, etc.) acts as an important control on the sources of ore elements and their mechanisms of transfer.

11 Metallogenists long ago demonstrated the existence of distinct ore forming periods, and the study of metallogenic epochs has been an area of major recent progress. Classic Soviet studies systematically linked metallogenic periods to orogenic phases (Smirnov 1988); yet, it has become apparent that there are discrepancies between the productivity of certain orogenic belts, which are difficult to explain (Goldfarb et al. 2001). New insights into mantle behaviour may hold the answer by revealing that convection periodically affects the upper mantle and, less commonly, the entire mantle; it is also evident that this process produces complex distribution patterns over time (Larson 1991; Ernst and Buchan 2001; Abbott and Isley 2002). It is now understood that the mantle can provoke sudden accelerations in plate movements and that mantle convection is not regular; it is accompanied by the abrupt arrival of hot deep material representing plumes that form hot spots and possibly cause continental rupture. Plumes were more abundant during the Archean and thus influenced the plate motions of that era (Condie 2001).

12 An ever-increasing number of deposit classes can be linked to these mantle processes: mantle-derived magmas that form layered intrusions (e.g. platinum, chromite, titaniferous magnetite and vanadium ores at Bushveld) and continental flood basalts (e.g. copper-nickel-PGE ores at Noril'sk). Ultramafic igneous rocks from the Labrador Trough (Québec) and the Thompson nickel belt (Manitoba) have been associated with a mantle super-plume event at 1.9 Ga (Condie 2001; Hulbert et al. 2005), and high-T, Mg-rich komatiites– the typical Archean expression of extensive partial melting– are associated with Ni-Cu deposits.

13 Anorogenic magmatism is linked to lithospheric extension, hotspots and intraplate rifting and various types of mineralization associated with this style of magmatism, including Sn, Nb, Ta, U, Th, F and Be in anorogenic granites (Sawkins 1984), and Olympic Dam-type Cu-Au-U iron-oxide deposits. Plume-related crustal magmatism can also exert a major influence on mineralizing processes. The giant massive sulfide deposits of Kidd Creek are genetically associated with the emplacement of a high-temperature volcanic pile of komatiites and rhyolites (Barrie1999), and several F-Ba vein systems have been linked to rifting, such as the Rio Grande rift and the Rhine Graben, although such rift systems do not necessarily require deeply rooted mantle processes (Hamilton 2003).

14 Deposit size is a key economic factor. Size distribution has been studied by a number of USGS researchers, beginning with the pioneering work of Cox and Singer (1986). Two distribution patterns have emerged: fractal and exceptional. A fractal distribution is observed for most hydrothermal deposits, reflecting continuity and similarity in the processes that created both small and giant deposits (Whiting et al. 1993). If mineral deposits are the result of growth and preservation (Veizer et al. 1989), then the dynamics of these fractal systems suggest that their size and distribution depend heavily on the rate of accumulation of ore, which is an almost unknown parameter for ore deposits at the present time. Exceptional distribution is observed in deposit systems that appear to be truly exceptional in size and scale, and must reflect extreme events in the Earth, like meteoritic impacts (Ni-Cu in Sudbury: Naldrett 2004; possible U in Athabaska: Duhamel et al. 2004), superplume events (Larson 1991), or major climatic events (e.g. bauxite, manganese, iron). These deposits represent scalar breaks in both spatial (referring to size) and temporal distribution patterns.

Table 1: Main types of mineral deposits using the UQAM classification scheme.

Display large image of Table 1

MECHANISMS OF FORMATION

15 Three factors must be considered in the formation of mineral deposits: the source(s) of the ore elements, the transport mechanism of those elements (silicate melt or hydrothermal fluid), and the mode of deposition. These components constitute a system for which an external source of energy acts as the engine that drives the processes. Although the formation of a deposit begins at the source, our understanding of the processes involved usually operates in reverse by first deciphering the depositional conditions and then working backward to the source(s) of the ore elements.

16 Knowing the age of a deposit relative to its surroundings is typically the key factor in determining the mechanism of formation. The careful use of radiogenic isotope geochronometers represented a significant advancement in this area and resolved the considerable speculation surrounding the petrogenesis of various deposits, including Carlin-type deposits in Nevada (Tretbar et al. 2000); gold in Witwatersrand, South Africa (Kirk et al. 2002); laterites in New Caledonia (Samama 1986); vein-type fluorite deposits in France (Jébrak 1984; Marignac and Cuney 1999); and the multiple events that led to iron-rich formations (Powell et al. 1999).

METALLOGENY AND GEODYNAMICS

35 During the last two decades, the scale of metallogenic studies has continued to grow, covering increasingly vast regions of the Earth. Fortunately, the development of geodynamics as a discipline has led to an improved understanding of a number of deposit types. For example, it was determined early on that exhalative deposits follow a geodynamic pattern, where copper deposits are oceanic and lead-zinc deposits are more continental (Hutchinson 1982; Large 1992). Vein-type gold deposits were also integrated into their seismological and orogenic contexts (Cox et al. 2000; Goldfarb et al. 2001). Yet this approach could not explain the temporal distribution of deposits, i.e. why particular epochs, like the late Archean or Cretaceous, show a distinct tendency to have more deposits than others.

36 Two important fields of study, climate-related ore formation and ore deposit research in the context of mantle dynamics, have emerged in the last ten years, which will provide the keys to answering some of the remaining dilemmas. The past decade saw a strong reversal of actualism in favour of a more complex vision of the climatic system. It began more than 20 years ago when the classifications of Gross (1980) and James and Trendall (1982) helped us understand that Lake Superior-type iron formations are evidence of an atmosphere that became increasingly oxidizing during Proterozoic time– the great rust age that remains to this day the best example of atmospheric pollution by living organisms– and from this realization was born the field of atmospheric and oceanic geology (Berner 2001; Holland 2002). Periods of cooling have also been recently linked to oceanic geo-chemistry and the formation of deposits (e.g. manganese). We can now tackle the issue of continental climates and contemplate the consequences of a 70°C average atmospheric temperature in Archean time.

37 Since the beginning of plate tectonic theory, it has been accepted that plate movement was surficial and that stationary hot spots existed, most likely rooted at great depth. The last few years have witnessed some important advances in this field (Courtillot et al. 2003; Anderson 2006), and it now appears that it may be necessary to invoke two styles of convection to account for deep plutonism: fairly surficial convection that represents the causal mechanism for plate tectonics, and deeper convection marked by sub-stationary hot spots capable of transferring material from the core-mantle boundary.

38 The early work on hot spots concentrated on long intra-oceanic chains, like the Hawaiian-Emperor seamount chain, which proved to be fairly barren in terms of economic mineralization. The ban on mineral exploration in Yellowstone National Park in the United States also hindered the recognition of a link between epithermal mineralization, geothermal gradient and hot spots. Nevertheless, the relationship between hot spots and mineralization was eventually established and documentation is continually improving (Pirajno 2000, 2004; Fig. 3). Hot spot basaltic magmatism is now seen as the most likely cause of giant komatiitic Ni-Cu deposits in the Archean, and Norilsk-type deposits in the Paleozoic (Yakubchuk and Nikishin 2004). Concentrations of platinum group elements and chromite accumulate in the plutonic chambers underlying these immense mafic complexes. If the hot spot causes melting of continental crust, a volcano-plutonic system will develop (like that of Yellowstone) generating both high and low-sulfidation epithermal mineralization with the probability of additional porphyry systems at depth. It was even proposed at one point that the Yellowstone hot spot caused regional hydrothermal fluid flow that produced the deposits of the Carlin district (Glen and Ponce 2002), but isotopic ages subsequently refuted this hypothesis. If the region is submerged during hot spot activity, then volcano-sedimentary deposits (the submarine equivalent of epithermal deposits) may develop in association with high-temperature rhyolites erupting over the hot spot (Barrie and Hannington 2000).

39 With this in mind, it seems that giant mineral deposits could be indicators of extraordinary terrestrial conditions – the result of global-scale events. Moments in Earth's history marked by superplume episodes completely disrupted the planetary landscape-from the bottoms of the ocean right to the atmosphere-leading to the formation of a variety of deposits. It is a fact that invalidates any purely actualistic notions (Condie 2001).

Display large image of Figure 3

DISCUSSION

40 Until the middle of the 20^thcentury, economic geology was basically empirical, as evidenced by haphazard mining discoveries and perplexed musings on the oddity of sufide minerals in silicate gangue. Deposits were seen as localized events, a view that prevented any real understanding of their underlying patterns.

41 It is evident that ore deposit formation can be linked to the six stages (magmatism, erosion, transport, deposition, lithification and metamorphism) of the rock cycle (Fig. 4). The link between ore deposits and orogeny was popular in the 1930s, whereas the link with sedimentology was the main focus of the 1960s. The role of plate tectonics in ore-forming processes was a hot topic beginning in the 1970s (Sillitoe 1972; Mitchell and Garson 1981; Sawkins1984) and not only established a new concept of mountain building, but also a real understanding of oceanic processes and hydrothermal systems. It was only at the end of the 1980s that the role of the climate began to be truly recognized and taken into account (Samama 1986), and finally, the new millennium signals a new era that will focus on the role of the mantle (Pirajno 2000, 2004; Ernst and Buchan 2001).

42 Empirical observations of ore deposits began in the 16^thcentury (Rabinovitch 2000), but it was only in the 19^thcentury that a genetic classification stage finally superseded purely descriptive work. During the second half of the 20^thcentury, we embarked on a new phase of ore deposit science when theoretical studies were reconciled with the pragmatic needs of the mineral exploration industry. Simulations of ore-forming processes were developed, including geochemical simulations that predict the composition of fluids and rocks in equilibrium (Barnes 1997), thermal simulations that estimate the duration of the convective phase (Barrie 1999), and geometric simulations that will ultimately allow us to predict the size of a deposit (Oliver et al. 1999).

43 One day it will be possible to simulate the complete formation of an orebody, and in so doing, predict its location in real space. The evolution of economic geology from description to simulation (Fig. 2) represents a fairly classic epistemological history. The history of climatology is similar in many ways, moving from the description of clouds to forecasting weather systems. However, the uncertainties that are the hallmark of field-based sciences still remain (Stengers 1994).

Display large image of Figure 4

44 Economic geology in the 21^stcentury will continue to show progress in our understanding of increasingly large systems, in our analysis of central issues in ore deposit genesis, and in the development of analytical tools that out perform their predecessors. In response to the challenges faced by the industry, it will be necessary to tie mineral systems to exploration methods (McCuaig and Hronsky 2000); analyzing the geometry of fluid transport systems in greater detail is one such example. Simulations will lead to predictions, undoubtedly still modest in scope yet continually evolving thanks to increasing computing power. And today's unconventional deposits, discovered through serendipity, will become the classic deposit of the future. The dialogue between explorationists and metallogenists is certainly here to stay!