INTRODUCTION

Decline of the success ratio (Gabelman, 1974) implies that uranium explorationists have become relatively complacent in applying the exploration technology which was developed through discovery of the comparatively easily recognizable sandstone uranium deposits of the Colorado Plateau. As the proportion of those remaining to be discovered in plateau sandstones has decreased, this technology appears increasingly inadequate (Gabelman, 1975) for recognizing more complexly concealed uraniferous sandstone deposits and other less conventional types of uranium deposits. The stagnation of technology and discovery suggested the wisdom of reexamining the entire possible range of uranium localization and all the mobilization and fixation processes, as well as the kinds of targets which are most ppropriate at higher uranium prices. This review identifies several underemphasized and newly recognized processes as well as several types of preferred radioelement concentrations which deserve stronger emphasis in exploration.

SELECTION OF EXPLORATION TARGETS

The seeming contradiction between low reserves and exploration failure on one hand, and abundant low-grade resources on the other, and between stagnated uranium-exploration geology and extensive knowledge of the low-grade resources justifies some review of the types of uranium localization and of the corresponding selection of exploration targets.

The exploration technology for a natural commodity such as uranium in the United States depends on industry to provide the empirical data (the deposits themselves) on which the technology is developed (Gabelman, 1974). Yet industry has not been disposed to perform or subsidize the geologic studies. This contrasts with the ready availability of surface geologic data which can be reinterpreted constantly to keep pace with the progressing science. At the surface, geologic science is never stale for lack of data. Most commonly, ore reserves are known only from subsurface penetration whereas submarginal resources are quantified almost entirely from surface data. As with orebodies, the proportion of total resources known from surface data may be small but because this part nevertheless is l rge in itself, and in the absence of economic operating experience, there is a tendency to consider it all available.

End_Page 1993------------------------------

Therefore, with this seeming abundance, geology seldom has been used to reinterpret the high-cost resources in advance of need, and those studies performed are more commonly academic than economic. By normally confining itself to currently economic ore deposits, industry excludes itself from studies of high-cost resources, leaving this field to government. However, government has found it difficult to perform these studies in the light of realistic long-range economics. When shortage calls for exploitation of unused high-cost types of resources, industry first must relate its geology to economic viability. At such times industry always seems to prefer to visualize exploration targets in terms of the known deposits being exploited, rather than of the untouched available low-grade depos ts. This choice appears illogical if the shortage is expected to cause early price rises, as has been the case recently with uranium. However, this can be explained by the also threatened cost of remodeling plants and processes to handle new types of ore, and the threat of having these revised operations wiped out by later discoveries of "conventional" deposits producible with older lower cost technology.

Therefore, the stagnation of geologic technology refers specifically to the arena in which conventional deposits are sought. Only when these prove sufficiently difficult or costly to find will industry tackle the new resources.

UNCONVENTIONAL RESOURCES

A review of the type-cost balance of known high-cost uranium resources is worthwhile to provide perspective for discussion of the selection of the geologic type of exploration target, and the burden placed on an apparently stagnated guiding technology. The following correlations are based on the Atomic Energy Commission (AEC) cost categorization of unused long-range resources, made in the middle 1960s. The original costs have been escalated arbitrarily 40 percent by the writer, rounded, and placed in parentheses for comparison.

Deposits economically viable at $8 ($11) per pound of U₃O₈ in concentrate, the criterion during the latter part of the Federal purchasing program, became known as "conventional" deposits simply because they characterized the mining industry. These were sandstone impregnations (95 percent of reserves) most containing above 50 tons (45 t) U₃O₈ and averaging 0.25 percent U₃O₈ in grade; and veins (5 percent of reserves) with as low as 5 tons (4.5 t) U₃O₈, most averaging above 0.5 percent U₃O₈. The prices of $6.70 to below $5, prevailing during the Federal stretch-out and phase-out program, and during the initial private market caused the size and grade cutoffs for conventional dep sits to rise. This stimulated "high grading," or "picking the eyes" out of orebodies, all of which are characterized by grade gradients decreasing radially from the highest grade "eyes." The later recognition of impending shortage in the power market induced the opposite practice in AEC ore estimation of raising the cost ceiling to $10 ($14), and in company mining of lowering cutoff grade to allow extraction of marginal low-grade conventional ore, by mixing with central high grade. However, the additional reserves made available by extending orebody margins proved relatively insignificant because of the highgrading during the low-price period and because the actual price rise was postponed together with the increased demand. When the price rise materialized it was too late to extend many orebody margins. Therefore, as shortage loomed, the choices of the resource industry became intensive exploration for conventional $10 ores, or switching to high-cost known, but unused resources. For long-range planning in government, consideration was forced toward the high-cost resources in types of deposits that could be termed exotic in relation to the geology and exploitation practice of conventional deposits. That the operating and economic characteristics of these deposits never had been determined by exploitation earned for them the designation "unconventional" resources.

Most of the presently identifiable unconventional resources became known long ago through academic geologic investigation, during the great uranium prospecting boom which swept the country, or as by-products of investigations for other metals. These were categorized by their production costs in increments of $10-$15 per pound U₃O₈, $15-$30, $30-$50, $50-$100, and $100-$500. The $500-limit was chosen because it represented the ceiling-fuel cost for the economic operations of breeder reactors, as estimated during the 1960s without the benefit of operating experience. These cost categories since have been invalidated by inflation, but the relation of the sequence of resource types to rising cost remains. The expected linear direct relation of cost to reserve size an indirect relation of cost to grade does not apply here because of the great difference in the expectable quantity of each of the different resource types. The relations also contain a significant lesson in the geology of uranium and the mechanisms of its migration. The cost-availability relation also varies with local geographic and operating characteristics. Between categories of resource type, however, the total-resource-quantity

End_Page 1994------------------------------

curve rises stepwise, emphasizing the great difference in the size of the respective resources. In sequence of increasing exploitation cost, the resource types visualized in 1967 by the AEC were:

1. $8-10 ($11-14) conventional orebody margins and conventional low-grade bodies.
2. $10-15 ($14-21) refractory mineral placers (mostly monazite).
3. $10-15 ($14-21) marine phosphates.
4. $10-15 ($14-21) uraniferous coals, lignites, and oils.
5. $15-30 ($21-42) rhyolitic tuffs.
6. $15-30 ($21-42) black carbonaceous shale.
7. $30-50 ($42-70) igneous rocks with ± 15 ppm U and 45 ppm Th.
8. $50-100 ($70-140) sea water.
9. $100-500 ($140-540) all granites (average 4.5 ppm U, 15 ppm Th).

Wide ranges of cost and availability are recognized within each of the resource-type categories; thus there is a mix of types available at any level of economics. Yet the categories still indicate broadly that the type of resource exploited should be expected to change with changing economics.

The monazite placers are known only in a few localities (South Carolina, Georgia, Idaho, Egypt, India), strangely in areas otherwise devoid of known conventional uranium deposits and together are volumetrically insignificant. This illustrates that only the most refractory uranium minerals escape complete destruction during transport and most uranium is highly dispersed in stream and sea waters.

Virtually all marine collophanes contain very small quantities of uranium universally distributed in the rock, locked in the lattice, and demonstrating well the chemical selective extraction of uranium from sea water. However, the grade of 0.00X to 0.0X percent U₃O₈ is so small that production of phosphate for uranium alone has not been economically feasible. Further, the rock has greater value as phosphate. The technical feasibility of extracting, at prices of $10-15, uranium from that phosphate converted to phosphoric acid was established in pilot plants in the 1960s but the incentives to do so have been insufficient until price rose in the middle 1970s. However, the size of available uranium resource is limited to the size of treble-superphosphate industry. Fo the rest of the century the domestic quantity available was judged at less than 20,000 short tons (18,000 t) U₃O₈ in 1967, and at 70,000 tons (63,000 t) in 1973 (OECD, 1973).

Uraniferous carbonaceous shales ranging from 50 to 500 ppm U are well known throughout the world, the Chattanooga (±65 ppm) of the United States being an excellent example. Uranium extraction from shale has been studied for many years and Sweden conducted a pilot operation in the 1960s, establishing a cost of about $11/pound U₃O₈. France and several other countries now are initiating operations. Other than the low grade, the main problems are economic release of uranium from the carbon and the environmental problems of large open-pit mines. However, the obvious reserves in continuously uraniferous extensive beds are tremendous and are a great attraction for easy alleviation of the uranium shortage.

The shale resource also illustrates excellently the geologic mechanisms of syngenetic extraction of uranium from sea water by adsorption on carbon particles (Conant and Swanson, 1961). It emphasizes the large quantity of uranium which finds its way into sea water. The prevalence of uraniferous shales and phosphates over uraniferous detrital sediments demonstrates that uranium in marine water may represent most of what has migrated into the crust.

Despite the well-known ability of carbon to attract and fix mobile uranium, the well-studied coals and oils were found to represent an insignificant resource. The ability of coal to extract uranium from aqueous solution was found by Moore (1954) to be inversely proportional to the rank of the coal and to reach a maximum in subbituminous coal and lignite. This factor may be controlled by the decreasing coal permeability with increasing rank, as much as by the increasing loss of chemical reactivity. Low-rank coals, then, act as sponges to migrating uranium and, because they are present in extensive shallow sheets and are actively mined, were considered as potential sources of by-product uranium. The grade of uraniferous lignite has been found to range locally to as high as O.X percent U SUB>3O₈ and in North Dakota lignites were mined for uranium alone. However, exploitation usually has proved uneconomic, mostly because of burning and handling problems, but also because the beds are too thin. Oil has proved singularly barren of uranium, possibly because of the natural immiscibility of oil and aqueous uranium fluids and the lack of geologic mechanisms for emulsification. However, the residue remaining in flushed oil fields is an effective uranium precipitant, and migratory plant-derived humates are considered the ore precipitants at Grants, New Mexico.

The highest U-Th concentrations in plutonic rocks are in syenites and alkaline granites where 100 ppm U commonly is reported. The Th/U ratio in these rocks is relatively constant at 3 or 4 to 1. Textures and mineralogy indicate that the thorium

End_Page 1995------------------------------

and uranium were fixed late in the igneous cycle ranging from magmatic to hydrothermal.

The Illimausaq alkaline complex of Greenland with local concentrations of as much as 1,200 ppm U (Sorensen, 1970) is the best example. The most uraniferous alkaline rocks also are generally small deposits, and are those in which the uranium, locked in refractory minerals, is expensive to extract.

The larger and less uraniferous granites, however, contain some uranium in more easily leachable form (Brown and Silver, 1956) suggesting them as the more practical sources of high-cost uranium. The Conway granite of New Hampshire, containing 45 ppm Th and 15 ppm U, was investigated intensively in this regard (Adams et al, 1963). In 1965 this resource was considered available at $30-50 per pound. Feasible extraction metallurgy was devised and it was expected that this resource could be developed ahead of the shales. The resource is so large that almost any need could be satisfied. At such fuel costs, breeding also could become economic and a credit therefore be earned by the thorium.

The final resort in terms of minimal grade of resource would be the extraction of uranium from sea water; however, this source was considered ahead of the granites because of the low potential extraction costs that have been claimed. The supply is obvious and unlimited in terms of human need. However, grade (3 parts per billion) is so low that water would have to be processed in tremendous quantities to produce a significant quantity of uranium. A further complication is that a continually new water supply unmixed with already processed water is needed, thus relegating processing sites to localities of significant tidal flow, pumping or to moving shipborne plants in the open ocean. Still the technical simplicity of extraction and the size and availability of resource always have combi ed to make recovery attractive, and periodically new attempts are reported. The separation process is basically one of ion exchange. Uranyl ions can be absorbed selectively by hydrated titanium oxide, and eluted with a variety of agents inducing ammonium or sodium bicarbonate (Llewelyn, 1975). The United Kingdom Atomic Energy Authority announced in 1964 (Davies et al, 1964) a projected recovery cost of $11 to $22 per pound U₃O₈ for a 1,300-short-ton U₃O₈/year plant at the Menai Strait in the United Kingdom but review in the late 1960s by Oak Ridge National Laboratory (Harrington et al, 1974) indicated a realistic cost not related to the Menai site of closer to $50. In 1972 Pakistan (Khan, 1972) considered recovering uranium from sea water proce sed in desalination reactors. In 1974 the British (Taylor and Walford, 1974, in Llewelyn, 1975) began a series of energy-cost and feasibility analyses. More recently the Japanese (Kawada, 1975) investigated the economic feasibility of extraction from large volumes of sea water, which are handled primarily as coolants for steam-power plants. Bettinali and Pantanetti (1975) of Italy also concluded that the process was feasible at recovery costs of $62-120/pound U₃O₈, in the same range as updated estimates by the United Kingdom and the United States. A contemporary updating review by Llewelyn (1975) considered the emerging concern for environmental protection as an increased cost factor, and compared the cost (in terms of energy) of the materials to produce a sea-water uranium-power system, with the energy the system could produce. For a pumped scheme the gain factor is 6.4 and for a tidal scheme 11.7. Thus Llewelyn concluded that either scheme was economically feasible at suitable sites. Still this exercise does not compare the relative availability and cost of sea-water uranium with uranium from other sources. Nevertheless, the same problem of extremely low recovery rate would seem to render the production relatively insignificant in terms of demand.

A fundamental reluctance to reorient exploration toward higher cost, even in periods of conventional reserve shortage, as stated, derives from the constant threat of discovery of new low-cost resources. The conversion of technology and equipment is very costly, and once locked into a high-cost, unconventional resource, an operation could not compete with ample supplies of conventional resources. Although many companies might venture experimentarily past the exploration frontier, most keep an anchor foot in the present world by maintaining healthy exploration for conventional deposits. It is with this emphasis in mind, and because conventional exploration has not been sufficiently productive in the new uranium boom, that the writer considers a review of all possible uranium mobilizatio , migration, and fixation mechanisms opportune. It is hoped that such a review will suggest some available geologic evidence that could overcome the present stagnation of conventional exploration or orient attention toward lowest cost unconventional resources. The current resource-evaluation program of the Energy Research and Devel. Administration recognizes this strong emphasis of industry on the conventional-deposit end of the type-cost range, and for several years has published figures only for the categories below $30 per pound U₃O₈ (ERDA Grand Junction Office,

End_Page 1996------------------------------

1975; Hetland, 1975). The program recognizes inflation by transferring reserves to successively higher categories, but estimates of the inflation rate are excluded.

New unsuspected types of unconventional resources have been recognized in several metal fields (copper, gold, lead, and uranium) as a result of statistical or research-type exploration. Such discoveries are surprises in the sense that the type of deposit discovered was not visualized as a target. Often the new type of resource seriously upsets the existing cost-type-availability categorization of resources and the utilization schedule. Such an event already has occurred in the uranium industry with recognition of the Rossing deposit (von Backstrom, 1970) of South-West Africa as a large significant unconventional resource economic at $10 per pound or less. This long-known deposit always had been considered a pegmatite, and therefore submarginal at $8. Pegmatites were one of the first t pes of uranium localization known. Their uranium long had been considered the product of the residual postmagmatic fluid acting in a time and physico-chemical realm intermediate between the deuteric-reaction-replacement phase and the magmatic-hydrothermal-mineralization phase. The realm is distinctive enough to have become known as the pegmatitic phase, characterized by insertion of highly volatile fluid along folia and fractures (provided these can exist) with expansive growth and zoning of the body by replacement of walls and already formed pegmatite. Exotic minerals collect at scattered centers, many of the volatile ions from large volumes of fluid. Uranium minerals, mostly refractory, form in this manner. They are rich in uranium, but are scattered so unpredictably and sparsely that earch for them, or extraction of the entire pegmatite on their behalf, never has been economic.

Recognition simply by thorough statistical ground penetration demanded for Rossing the apparent contradiction of the rule. However, close inspection of literature reveals that it may not be a typical pegmatite, which opens an entire new field for consideration of exploration targets. The writer had suggested (1969) that the somewhat similar Bancraft deposits of Ontario are not true pegmatites, but rather replacement disseminations produced by hypothermalism during the hydrothermal rather than the pegmatitic stage, as are the so-called "porphyry" copper replacement disseminations. Similar observations were made during 1974 in northeast Brazil where the writer recognized at least five separate stages of pulse-like late-magmatic injections punctuated and followed by several stages of peg atitic and hypothermal replacement. One and possibly three of these stages produced exotic-mineral-replacement disseminations.

The Rossing experience demonstrates for the first time in uranium exploration that the large, low-grade, high-temperature-replacement disseminations of thorium and uranium can constitute economic deposits of such large uranium content as to rebalance the distribution of world uranium reserves substantially in their favor. As a result since 1970 the replacement dissemination (also popularly called the "porphyry" or "Rossing" type) has become one of the most sought targets. This development illustrates that although the knowledge of unused types of high-cost resources may appear to be extensive it is still far from representative, and that it still is possible to discover new types of deposits as well as new conventional deposits. It is possible also that a seemingly unreasonable demand can be met suddenly and easily from an unexpected source.

FUNDAMENTAL RADIOELEMENT DISTRIBUTION

Although the least abundant of the three significant natural heat-producing radioelements (potassium, thorium, and uranium listed in decreasing abundance and order), uranium has the greatest heat-producing capacity (Barth, 1952; Heier and Billings, 1972), and its abundance is indicated reasonably by heat flow.

From a cosmic standard believed most represented by carbonaceous chondrites (Taylor, 1964) and a terrestrial standard most approximated by the mantle (Clark and Ringwood, 1964; Ringwood, 1969; Rogers and Adams, 1969a, b), all three radioelements have migrated upward during crustal evolution. Selective enrichment in sialic crust is demonstrated by both greater abundance (Rogers and Adams, 1969a, b; Ronov and Yarashevsky, 1969) and heat-producing capacity. That the mantle still emanates these radioelements is indicated by equal heat flow in oceanic and continental crust (Verhoogen, 1954; Lee and Uyeda, 1965). Oceanic crust is temporal by virture of sea-floor spreading and has a maximum residence time of approximately 80 × 10^-6 years (Dietz, 1961; Hess, 1962; Wyllie, 1971 , so that all oceanic radioelements eventually are delivered to orogenes at continental margins. The equality of oceanic and continental heat flow, considering past selective enrichment of continental crust, suggests that a substantial part of oceanic radioelements is recycled to the mantle along Benioff subduction zones. The amount of radioelements in oceanic tholeiite and ophiolites (Gast,

End_Page 1997------------------------------

1968; Rogers and Adams, 1969a, b; Ronov and Yaroshevsky, 1969) is about 1 ½ orders of magnitude less than for continental basement. Therefore, for heat flows to be equal, the portion of radioelements emanated into oceans must have exceeded that deposited in oceanic rock by more than 1 ½ times and must have been emanated in volatile or liquid phases into sea water or oceanic sediments. Uranium in sea water currently is so low as to imply extraction by marine sediments and eventual delivery to the orogenes.

CRUSTAL URANIUM MOBILIZATION

Selective mobilization of crustal uranium and thorium occurs mostly in continental-margin orogenes. Here, sedimentary and previously metamorphosed rocks are metamorphosed progressively as they are incorporated into the mobile belt. Regional metamorphism is a gradual metasomatic fractionation of mafic and sialic components by liberated and mobilized rock fluids. In comparison with unmetamorphosed equivalents, the progressively lower abundance of uranium, together with many other easily volatilized elements, in more metamorphosed rocks indicates that these elements were mobilized in fluid or volatile phases. Their distribution in tectonically controlled zones (Gabelman, 1968, 1969; Gabelman and Krusiewski, 1968) suggests that they are driven forelandward along the same metamorphic inten ity-temperature gradient to seek new positions of temperature-controlled stability. The gradient can be considered as a superimposed reversal of the metamorphic-intensity gradient.

URANIUM MOBILIZATION BY METAMORPHISM-ANATEXIS

Advanced products of regional metamorphism in the orogene root are amphibolites and granites in closely proximate bodies. In relatively shallow sialic layers these may be final products, and commonly constitute the batholithic zones. Deeper, as the average density of the medium increases, there is a growing tendency for the lighter isolated sialic fraction to melt and rise as granite magma, enhancing batholithization in the rear half of the mobile belt. Nearer the foreland, shallow sialic crust of granodioritic average composition, even though fractionated by metamorphism, is dragged ever deeper by subduction and arrives in a higher density layer where all the sialic fractions are melted by anatexis. By virtue of its lower density, this anatectic sialic product also should rise to pro uce granodioritic magma. This magma intrudes and differentiates through the calc-alkaline series.

THORIUM-URANIUM MOBILIZATION IN MAGMAS AND THEIR DERIVATIVES

Although strong metamorphic or magmatic differentiation has produced moderately undersaturated alkaline granites from crustal sial (Wyllie, 1971), most strongly undersaturated rocks are believed to be derived directly from the upper mantle by volatile-driven rapid intrusion (Harris, 1967; Ringwood, 1969; Wyllie, 1970, 1971). They differentiate through an independent undersaturated series of more limited range.

Magmatic intrusion, emanation of pneumatolytic or hydrothermal differentiates, or tectonic elevation of orogene roots are the only means for radioelements to arrive in shallow continental crust.

Thorium exhibits only the +4 valence state in nature whereas uranium commonly assumes the +4 and +6 states (Katz and Rabinowitch, 1951; Kirk and Othmer, 1955a, b; McKelvey et al, 1955) permitting its oxidation and reduction (McKelvey et al, 1955). Thorium and uranium are geochemically similar and mobile at high temperatures in the +4 valence state probably because of similar ion sizes (Goldschmidt, 1954) and proximate atomic configurations. However, at lower temperatures thorium is relatively immobile whereas uranium is mobilized and fixed easily by the oxidation-reduction mechanism, so the two elements become separated. The degree of their spatial association is a valid genetic-temperature index. Thorium and uranium in a consistent and uniformly distributed crystalline rock associate ratio of about 3.5 to 1 (Rogers and Adams, 1969a, b) become increasingly abundant near the silicic end of all the magmatic differentiation series. This increasing abundance illustrates the preference of thorium and uranium for the more volatile phases. The association and increasing abundance persist into the postmagmatic and hydrothermal stages where uniform lithologic disseminations give way to local concentrations as replacements of suitably reactive rocks or precipitates from fluids. Thorium becomes relatively immobile below the hypothermal stage and thus is separated from uranium, which continues to concentrate to a maximum at surficial temperature.

Thorium and uranium in the strongly undersaturated (subsilic) alkaline magma series are two to three times more abundant than in the saturated granitic and granodioritic series mentioned previously (Rogers and Adams, 1969a, b). This difference again is attributed to the enrichment of these elements in the volatile phases during fractionation to form alkaline magma in the mantle (Wyllie,

End_Page 1998------------------------------

1970, 1971), as well as to the flight of these elements down metamorphic-intensity gradients during regional metamorphism in the orogene (Gabelman, 1968, 1970; Gabelman and Krusiewski, 1968). The more saturated mantle magmas which produce basalt and peridotite could have been robbed of most of their thorium and uranium prior to their transfer to the crust. The more radioelement-rich less saturated magmas are believed to have been emplaced mostly as volatile-rich emanations directly from the mantle (Wyllie, 1970). The persistence of a high radioelement content in the youngest alkalic rocks, as well as the heat-flow balance, here are considered evidence of significant quantities of thorium and uranium still in the mantle and of the continuation of the process of their delivery to the cr st.

THORIUM-URANIUM MINERALS CRYSTALLIZED FROM MAGMAS

Minerals formed by thorium and uranium together at high temperatures typically are stable complex silicates and multiple oxides involving a variety of other uncommon or exotic metals including rare earths, zirconium, niobium, and titanium (George, 1949; Gabelman, 1970). Thorium alone forms simple silicates and oxides at virtually all surficial and intermediate temperatures (Gabelman, 1970).

LIBERATION OF THORIUM AND URANIUM FROM MAGMATIC ROCKS

Thorium and uranium are liberated from crystalline rocks being destroyed at the surface by erosion or being leached in the water-saturated zone by suitably corrosive fluids.

Placer deposits formed by sediment transport and deposition contain essentially only the stable high-temperature complex thorium-uranium minerals in which thorium dominates at ratios of at least 3.5:1. The low uranium content of these minerals plus their refractory nature and the small size of the placers render normal placer deposits a relatively insignificant uranium resource. Such deposits are characterized further by a predominance of ferrous oxide iron, and subordination of simple uranium oxide and silicate. The quartz-pyrite-uraninite-pitchblende conglomerates, which do constitute a major uranium resource, are unusual special cases further characterized by absence or subordination of ferrous oxide iron and refractory thorium-uranium-rare earth minerals. These require nearby prol fic pitchblende-rich quartz veins in a readily destructible greenschist matrix (Roscoe, 1969) and a reducing terrestrial atmosphere (Roscoe, 1969), to preserve the sulfides. As such the pyritic conglomerates cannot qualify for comparison with "ordinary" placers derived from destroyed granitic terrains. Less stable, simple uranium minerals usually are oxidized and dissolved during erosion and transport over all but very short distances, and the uranium is added to the meteoric hydrologic realm. Thus, the most important accumulations of uranium in surficial or shallow rocks form by hydrologic or pneumatolytic processes.

Purely pneumatolytic mineralization processes, including those of uranium, are relatively unimportant in the epizone because most volatile elements or ions arriving in this environment as gases from any source are soluble in water and must enter the terrestrial hydrosphere. However, gases appear to have been important ingredients of mineralizing fluids. Probably most such gases arrive as solfateric emanations with actual or implied volcanic affiliation, although commonly there is no direct association with lavas. Uranium has been detected in volcanic gases by Stoiber and Rose (1968) and Rose et al (1970). Gases such as carbon dioxide or hydrogen sulfide evolved by purely surficial chemical or biologic processes are of very localized distribution and believed by the writer to be little related to the introduction and regional distribution of uranium, although they can influence its local fixation (Jensen, 1958).

URANIUM MINERALIZATION IN THE HYDROSPHERE

Surficial, vadose, and saturated layers constitute three subrealms of the terrestrial hydrosphere. The saturated layer or zone is underlain by a dry layer of unknown, but no doubt variable, depth, in which most fracture porosity and considerable lithologic porosity are closed by high lithostatic pressure, and to which meteoric water access is limited to a few preferred stratigraphic or structural aquifers where flow is channeled restrictively. Surficial water is without question meteoric (Graf et al, 1965; White, 1969). In the vadose and saturated layers water also has been proved to be mostly meteoric but differs from surficial water by virtue of mixing with connate water (ultimately also meteoric) or hypogene fluids (gases or liquids). Hydrochemistry suggests that most ground water s ultimately meteoric (White, 1969). Both connate and hypogene waters by definition are distinctive only genetically and where their sources can be identified. Although they undoubtedly modify the character of meteoric ground water at least by contributing to its

End_Page 1999------------------------------

salinity, once mixed with meteoric water they lose their identity.

POTENCY AND PENETRABILITY OF URANIUM MINERALIZING FLUIDS

Pure water is a powerful solvent, given adequate time, and brines are more powerful. Therefore the remobilization of relatively unstable elements and minerals in the vadose layer is a valid process of lateral secretion (Gabelman, 1976, in press). Selected fracture or pore-aquifer systems extend the remobilization possibility into the underlying dry realm. Therefore, without conclusive means of qualitatively distinguishing many key elements according to their source, the best way to recognize mineral deposits formed by supergene, laterogene, or hypogene processes operating at shallow depths and low temperatures, is by the geometry of their controlling hydrologic systems (Gabelman, 1976, in press) and by the degree of compatibility between the host-rock changes produced by the mineraliz tion and those typical of the general movement of meteoric ground waters through the area. The negligible effect typically produced by relatively fresh meteoric water is well demonstrated by the lack of significant epigenetic mineral deposits in active aquifer systems. Mineralization visibly produced by cold brines is also negligible. However, rising temperature increases mineralization dramatically, and the mineralization capacity of hot brines is orders of magnitude greater than that of hot fresh water.

The effects of temperature can be deduced from the distribution of mineral deposits within a district and from the distribution of districts (Gabelman, 1961, 1968; Gabelman and Krusiewski, 1968). Although the different sources of volatiles and salts are not easily recognizable, the geometry of their distribution in relation to areal or regional geologic features can suggest supergene, stratigraphic, or hypogene sources. Fluid flow and mineralization in the hydrosphere obviously are restricted to physical channels. Pervasive alteration and metallization in the dry layer, where physical openings are much less common, formed mostly bodies of near-vertical longest dimension. Even stratigraphic aquifer-controlled pervasive mineralization bodies in this layer are isolated areally enough to mply strong high-angle movement of fluids capable of reacting only with certain lithologies. The pervasive penetration of rocks of low permeability suggests a high volatile content of mineralizing fluids and at least moderately elevated temperature. The near vertical orientation and the isolation of many pipe-like bodies of pervasively altered rock imply a deep origin of fluids.

PROBABILITY AND PREVALENCE OF LATERAL URANIUM SECRETION

The amount of uranium supplied to hydrosphere processes by leaching of crystalline rocks--indicated by abnormally low uranium abundance in visibly and locally altered rocks--should be suggested by the amount of uranium remaining in the rocks, contrasted with a norm (Rogers and Adams, 1969b; Sterling and Malan, 1970) established by the regional average content in fresh rocks of the same type. Not all granites contain easily leachable (Brown and Silver, 1956) simple minerals resulting from endogenesis. Most uranium and thorium are in refractory minerals, where they remain unless leached by fluids made strongly corrosive by heat or the addition of volatiles. Such fluids alter rocks strongly, and the distribution of altered areas and their geometric relation to other geologic features ind cate that the fluids and their actions are localized by tectonic, structural, and stratigraphic features in that order of decreasing magnitude-order of control. Although the granite family is a great potential supplier of leachable uranium, many centers of such strong alteration also are centers of tectonic or volcanic activity, raising the possibility of juvenile additions to leaching hydrosphere fluids.

Such leaching is demonstrable only at small localized sites of rock alteration, and most sialic rocks available for observation still contain uranium in amounts typical of the average for each rock type investigated. Subsurface leaching of oxidized but uneroded crystalline rocks, therefore, appears to have liberated locally only amounts of uranium too small to account for concentrations in the same vicinity. Yet uranium liberated in this manner is assumed widely to have been the main source of epigenetic uranium in sandstone.

Destruction by surficial erosion undoubtedly has liberated enormous amounts of uranium from granites and syenites, and in terms of only quantity this uranium would be a more logical source for epigenetic deposits in arkosic sandstones. However, related uranium concentrations indicate that the proportion of eroded uranium which enters the ground-water system is small, and that which is reprecipitated in aquifers is smaller still. Whereas the abundance of uranium in granitic rocks is on the order of X parts per million that in average surface waters is X (10^-3) ppm, and that in underground waters is X(10^-4) ppm. As expected, uranium which is strongly dispersed in granites is dispersed further by liberation into the hydrosphere. Obviously, most of the uranium libera ed by erosion remains in surficial runoff and in solution reaches most distant depositional basins

End_Page 2000------------------------------

where it is fixed syngenetically in shale. There is no realistic way to estimate the part of superficially transported uranium which enters the ground-water system. The strong limitation on size and distribution of uranium deposits even within host aquifers of the best uranium districts, however, suggests that uranium generally dispersed in aquifer fluids is not easily nor uniformly extracted, but that special conditions of both fluids and aquifer are required.

Rhyolitic tuffs and arkoses also are popular suspected sources of uranium in sandstone impregnations (Densen and Gill, 1956; Eargle and Weeks, 1962; Grutt, 1972) and a chemical lateral-secretion mechanism has been demonstrated (Gruner, 1956). However, for large quantities of uranium such as represented by the Gas Hills, even greater problems are presented than for secretion from granites, and this interpretation therefore is hardly permissive rather than conclusive. These rocks are logical sources because of the well-known high uranium abundance in the glass component of the tuff and in the potash feldspars of arkose; both rocks are simply extended forms of granite. However, in the fresh state these rocks still display their normal high uranium content so they have not released signif cant uranium by leaching. The source thus is shifted to tuffs and arkoses that have been altered selectively by corrosive fluids, or destroyed by erosion. Some clay beds, or the argillaceous component of other clastic beds, such as the Jurassic Morrison sandstones, have been considered the equivalent of destroyed uraniferous tuffs. Without isotopic, or other suitable distinctions for the uranium in each environment, geographic and geometric conformance between distribution patterns of uraniferous source (including those destroyed) and host again is the only way to establish a genetic connection. As with the granitic rocks, the conformance of patterns leaves much to be desired, and implies the importance, if not dominance, of one or more processes other than simple destructive secretion. ssuming this was an important process, however, another problem is still the concentration of liberated uranium into deposits in a regime of general dispersion. The stratigraphic thinness of most tuff and arkose horizons implies that beds covering very extensive areas must have been destroyed to supply known uranium concentrations. Further, the well-known result of erosion and transport in surface hydrodynamic systems is dilution.

As with granites the logical source areas for uranium leached from relatively thin layers of tuffs and arkose, then, are reduced to the local areas that visibly are altered by corrosion and depleted of their uranium. These areas are also commonly the matrices of districts of concentrated uranium deposits. Here the mechanism is much more credible if the agent is conceded to be a concentrated corrosive fluid. However, the relative-size problem remains. In view of the dispersion from leaching, and the small part of pregnant leach liquor that would be channeled past any particular point, the altered areas do not appear large enough to have furnished all the uranium in the deposits. Tuff and arkose leaching, therefore, here are considered relatively minor processes of uranium mineralizatio .

TECTONIC-SEDIMENTATION CYCLE

Thorium-uranium remobilization, as part of the tectonic-erosion-sedimentation-anatexis cycle, occurs only in and next to orogenes. Sediments deposited on distant continental interiors or shelves which thereafter become stable may be removed from the cycle for long intervals of geologic time. Sediments carried into marginal-marine geosynclines are carried back into the orogene by sea-floor spreading, and reenter the orogenic-magmatic-sedimentational cycle.

RELATIVE IMPORTANCE OF URANIUM MINERALIZATION PROCESSES

Separated from thorium, uranium can arrive in the shallow low-temperature environment from juvenile or remobilized crustal sources. The oxidation-reduction and multiple-migration-accretion mechanisms are valid means of concentration by lateral secretion. However, the lack of geometric-geographic relation of apparently supergene concentrations to the most potentially uraniferous source rocks suggests that these mechanisms are not the dominant processes controlling the regional distribution of productive sandstone-uranium districts. The remobilization mechanism may serve mostly to obscure the original controls over uranium introduction. Uranium apparently does not travel great distances from centers where it is introduced into the upper crust. This is suggested by several factors. Parts of uranium deposits may be out of equilibrium but whole large deposits usually are in balance (Gabelman, 1975), as are districts, indicating that if entire districts represent only remobilized uranium, both parents and daughters equally were remobilized, or remobilization occurred so slowly that balance was not distributed. Also most significant districts exist at sites which are special in a tectonic, stratigraphic, or structural sense (Gabelman and Krusiewski, 1963, 1968; Gabelman, 1975) and which are critical because of large-scale fracture permeability.

End_Page 2001------------------------------

PROMISING LITTLE-CONSIDERED PROCESSES

A few of the most promising processes which may have created economically significant uranium concentrations but which have been used little in exploration include:

1. Crystallization-dissemination of thorium-uranium minerals in late-stage alkalic agpaitic and miascitic syenites, or volcanic equivalents.

2. Deuteric or hypothermal replacement-dissemination by thorium-uranium minerals in late-phase alkaline granites to form "porphyry" disseminations.

3. Direct diapiric emplacement of carbonatites (and their close relatives) or their pneumatolytic or hydrothermal modification.

4. Diatreme feeding of uraniferous mantle volatiles or fluids into structurally prepared ground, or directly into favorable host lithologies.

5. Feeding of mantle- or crustal-derived volatiles or fluids through taphrogenic intersections into favorable lithologies.

6. Hypogene laterogenesis in geopressured basins.

MOST PROMISING EXPLORATION TARGETS

The types of uranium deposits most likely to provide tonnages commensurate with higher level demand, closest to the conventional end of the cost-type spectrum, and for which guiding discovery technology is closest to viability are the ones which economics and haste would dictate should be sought first. In order of decreasing grade and increasing size and ease of identification the environments considered most promising by the writer are as follows:

1. Uranium concentrations in: (a) nonfluvial residual-oil-bearing sandstones, (b) lignitic sandstones near coalfields, (c) volatile piercement pipes, (d) low-temperature ends of mineralization and metamorphic gradients.

2. Uranium dispersions in: (a) undersaturated alkaline intrusives, (b) oversaturated alkaline end products of the calc-alkaline magmatic series, (c) small lake basins, (d) lignites, and (e) Chattanooga Shale.

CONCLUSIONS

The supply of untapped uranium resources is not so limited as the exploration record of the last few years would suggest. Rather the supply available to society at any time is determined by the need-cost balance, to which available supply can be adjusted by sliding economics. The uranium industry is demonstrating it is not yet ready to venture into the high-cost arena of unconventional, but more abundant resources, beyond the first step to the next level of lower grade, higher cost, maximum-amenability type of deposit represented by the Rossing replacement dissemination.

Additional steps toward exploitation of unconventional deposits can be postponed further by addressing little considered, but valid processes of uranium mobilization which could have formed conventional deposits in geologic environments not now being investigated. Review of the entire spectrum of mobilization and fixation processes indicates that the discovery of significant conventional deposits is still a very fertile field.