World Uranium Reserves

By James Hopf
Nuclear Engineer
November 2004

I recall reading studies in the 1970's that earnestly stated that we only had ~20-30 years of oil and gas reserves left (based on proven reserves, and the rate of consumption back then). In other words, we would have run out already. Needless to say, those predictions didn't pan out. Despite large increases in consumption rates (especially for gas), not only have we not run out, but the estimated reserves have actually gotten larger (as more oil/gas was discovered than was consumed). This is the result of enormous efforts (and expenditures) to explore, locate, and drill for gas and oil. Now the production peak is still one (or several) decades ahead of us [depending on the rate of increasing demand in China and India], and reserves won't run dry for several decades.

In the same vein, we continually hear about how the “proven reserves” of uranium will only last ~50 years at current consumption levels. These estimates, however, have an even weaker basis than the oil/gas estimates of the 1970's, since the amount of effort and expenditure that has been put, as of today, into uranium exploration and development is far smaller than that put into gas and oil exploration, even as of the 1970's. Some have even said that the amount of uranium exploration is more equivalent to that which had been put into oil exploration as of the 1900s (when Western Pennsylvania surface oil was just about all anyone knew about). This is probably an exaggeration, but not to as great an extent as one may think.

Around the end of WII, efforts began to search for uranium and to develop the discovered deposits. Over a short time period, and with relatively little effort, a large number of uranium ore veins, of various grades, were discovered. In the early days, an enormous amount of future nuclear power generation was predicted, and a large number of mining operations were planned and/or developed. There were a few “mother-lode” sites (in Australia, Canada, and to a lesser extent, the US) that had enormous veins of high-grade ore, and which had a very low cost of production. There were also a large number of smaller, lower-grade sites.

Two things happened down the road which greatly reduced the demand for mined uranium, relative to the initial predictions. First of all, the use of nuclear power grew much more slowly than anticipated, due to lower-than-expected growth in electricity demand (after the 1970's), as well as other factors like the glut of cheap natural gas (in the 1990's) and the anti-nuclear movement. The second thing that happened was the nuclear arms reduction treaties, and the resulting decommissioning of nuclear warheads. The highly-enriched (weapons grade) uranium in these warheads can be blended down to make much larger quantities of low-enriched reactor fuel. It is estimated that warhead uranium will provide almost half of the nuclear power plant fuel in the U.S. between 1990 and 2010, thus cutting the demand for mined uranium in half.

As the demand for mined uranium fell substantially (as opposed to growing as expected) the price of uranium ore plummeted. This caused all of the lower-grade ore mines to shut down. Only the few large, ultra-high grade ore sites could produce ore at a low enough price to make a profit. With only a handful of sites in operation, and a large number of known deposits (and even developed mines) not operating due to the low market price, all uranium exploration and development came to a complete halt, as there was simply no reason to look for more uranium, let alone invest significant sums of money to do so.

Lately, the price of uranium ore has increased substantially, from less than $20/kg to over $40/kg today. This is occurring due to the anticipation of the end of the weapons uranium stockpile, along with improved prospects for increased nuclear power generation. As a result of this price increase, several things are happening. There are now numerous reports of old, lower-grade ore sites going into development, as they will now be profitable. Also, many properties with uranium potential are now being explored, and many are being found to contain economically recoverable uranium. Reserves are increasing as we speak.

One important fact that must be understood is that, unlike the gas and oil, the cost of the uranium ore is a negligible fraction of the cost of nuclear power (with almost all of nuclear power cost being in the form of value added by domestic labor). Specifically, at today's price of ~$40/kG of uranium, the ore costs amount to only ~0.1 cents/kW-hr (i.e., only ~2-3% of nuclear's total power cost). The ore cost could increase by a factor of 10 (to ~$400/kg) and nuclear's power cost would only increase by ~1 cent. Thus, whereas gas and oil applications are extremely sensitive to the cost of fuel, and can be rendered uneconomical by even a small increase in fuel price, nuclear power is almost immune to ore price increases. Thus, the maximum price for uranium ore, above which nuclear power would become uneconomical, is extremely high indeed.

If an extremely high ore price is tolerable, then very low grades of uranium ore can be considered as possible reserves. As the permissible ore grade (uranium concentration) goes down, the amount of recoverable uranium (i.e., reserves) goes up exponentially. As is discussed in more detail later, limitless supplies of uranium are present in seawater and in the earth's crust, which can be extracted at some price. The question is how much uranium is available at a cost that doesn't truly price nuclear power out of the market.

The “proven reserve” estimates are flawed for two primary reasons. First of all they do not consider the fact that very little effort, or money, has been put towards uranium exploration thus far. Second, they do not adequately account for the tiny effect that uranium ore price has on final nuclear power price, and the maximum allowable prices that they use to determine “economically recoverable” reserves are far too low.

The effort made thus far in uranium exploration is absolutely negligible compared to the many hundreds of billions (trillions?) of dollars that has been invested in oil and gas exploration, technology development, and extraction, etc… As the history of oil and gas shows, as these investments are made, more and more reserves are found. As discussed earlier, we stopped exploring for new uranium deposits relatively soon after we started looking, since we rapidly found “all we need”, due to sluggish nuclear expansion and the glut of uranium from decommissioned weapons. Now, even the majority of known sites and mines lay idle [in year 2004] due to the low ore price (although this is starting to change).

As the price of uranium ore goes up, significant resources will go into uranium exploration, and many new deposits will be found, including many high-grade ore deposits that were simply never discovered. It is likely that the amount of uranium in yet-to-be-discovered high-grade (low cost) ore deposits greatly exceeds that which exists in currently-known high-grade deposits. In addition to these high-grade deposits, a large number of lower-grade deposits, both currently known and yet to be discovered, will become economical and will be developed. This is what happened with oil and gas, and it is even more clear that this is what will happen with uranium. Given that uranium produces about a million times as much energy as an equivalent mass of oil, gas, or coal, the amount of energy locked up in uranium (in the earth's crust) exceeds that locked up in fossil fuels by several orders of magnitude. This bodes well concerning the amount of uranium that will/can be eventually discovered and developed.

Current estimates of “economically recoverable” reserves apply an upper price/cost limit of $135/kg for uranium ore. This price cutoff does not sufficiently appreciate the lack of effect that ore cost has on power cost. It corresponds to a power price increase of only ~0.25 cents/kW-hr, versus today's $40/kg ore price. Uranium sources that cost up to $500, and perhaps even ~$1000/kg (which would increase nuclear power's cost by 1-2 cents/kW-hr) can still be economic, especially in a CO2-emission-constrained world, and/or a world where gas and oil have started to run out. Even at $1000/kg, advanced nuclear plants should be able to produce power at ~6 cents/kW-hr or less. The cost of power from post-production-peak gas or oil plants, or from coal plants with full CO2 sequestration, is likely to be higher than this. Finally, it should be noted that (as discussed later), at a uranium price of $500-1000/kg, breeder reactors become economical, and the uranium supply effectively becomes infinite.

It must be stressed, however, that the cost of uranium ore is not likely to ever approach $1000/kg. One reference that discusses all the potential sources of uranium on earth is Deffeyes and MacGregor. It lists various formations and minerals, etc.., that contain uranium, the concentration at which the uranium is present, and the estimated cost of extraction. On one end, there are our current high-grade deposits, where we currently get all of our uranium. On the other end of the spectrum is seawater, which has an essentially infinite amount of uranium, but at a very low concentration (and high expected extraction cost). The source with the largest overall quantity of uranium is the granite in the earth's crust. This actually has a much higher concentration than seawater, and has tens or hundreds of thousands of years worth of uranium.

Using the Deffeyes & MacGregor data you can estimate the total reserves of uranium that can be extracted, as a function of the maximum allowable ore cost. As the allowable cost goes up, the potential supply exponentiates. Extrapolations using this data shows that at a (still economical) price of a few hundred dollars per kg of uranium, there is enough recoverable uranium to provide all of our nuclear power needs for several hundred (perhaps 1000) years, even at a greatly increased rate of usage.

And note, this is even for the once-through cycle, which only makes use of the U-235. If we went to breeders, the amount of uranium ore used, per unit of electricity generation, is divided by a factor of 60-70. Not only that, but since 1/60th as much ore is used, the tolerable ore price increases by yet another factor of 60. This, of course, causes another exponential increase in the economically recoverable reserves. If we go with breeders, we have enough economically recoverable uranium to meet all our power needs for tens, probably hundreds of thousands of years. It should be noted, however, that the price of ore will have to go extremely high ($500-1000/kg) before breeding would make economic sense, and this won't happen for well over a century; plenty of time to develop safe, reliable, and economic breeder technology.

In summary, the actual recoverable uranium supply is likely to be enough to last several hundred (up to 1000) years, even using standard reactors. With breeders, it is essentially infinite. Hundreds of thousands of years is certainly enough time to develop fusion power, or renewable sources that can meet all our power needs.

As far as I can tell, none of the nuclear utilities have shown any real concern about long-term uranium supplies, and for good reason. This is basically a non-issue. The 50-year supply “problem” is most often brought up by two groups, both of which have a vested interest. First, there are the groups opposed to nuclear energy, who use these reserves estimates to argue that nuclear power has no long-term future anyway, and is therefore not worthy of significant investment. The second group consists of nuclear engineers and scientists who are devoted to the concept of a closed fuel cycle, where breeders or spent fuel reprocessing (to re-use the uranium and plutonium in spent fuel) is used. They argue that unless these methods are used, nuclear has no long-term future, because standard reactors (using the once-through fuel cycle) only have enough fuel (uranium ore) for a few more decades.

Whatever the merit of these groups' goals, these arguments are based on a false premise. Long-term uranium supplies are simply not a real problem. Even if (in the distant future) uranium ore does get really expensive, market forces, and nuclear technology, are equipped to handle it. Advances in extraction technology, along with higher ore prices, will exponentiate the recoverable reserves. Breeder reactors, which will become more economical in 50-100 years, will eventually appear and eliminate all supply issues. All indications are that we will have plenty of time (50-100 years) to develop such breeder technology, before the cost of ore really starts to impact nuclear economics. This is true even under the highest nuclear power growth scenarios.

—James Hopf, Nuclear Engineer
November 2004

References:
Deffeyes 1978 - Kenneth S. Deffeyes and Ian D. MacGregor, Uranium Distribution in Mined Deposits and in the Earth's Crust. Final Report, GJBX—1(79), Dept of Geological and Geophysical Sciences, Princeton University, Princeton, NJ. (Return to text.)

Deffeyes 1980 - Kenneth S. Deffeyes and Ian D. MacGregor, “World Uranium Resources,” Scientific American Vol 242, No 1, January 1980, pp. 66-76. (Return to text.)

Reference Resources:
Nuclear Power Plant Fuel
World Supply of Uranium
Uranium Reserves — Wikipedia
How long will nuclear energy last?
Nuclear Cycles and Nuclear Resources
How much longer will the world's uranium reserves last?

Generation-IV Roadmap—Report of the Nuclear Fuel Cycle
— From the U.S. Department of Energy (DOE) website.
337 page Adobe PDF document (size: 4 MB)
Chapter One, Section 1.3: Availability of Ore — Fueling Generation IV.
1.3.4 Sustainability Potential “For such recycle based fuel cycles, if exploited fully, at least a millennium of energy supply can be foreseen from the earth’s endowment of economically recoverable uranium ore listed in Table 1.3. Likewise, the energy potential of the earth’s endowment of thorium is greater still. Thus the resource base is sufficient to support the Gen-IV sustainability goal—given that the fuel cycles and reactor types are deployed to fully exploit those resources.”