Randy Barnhill | Barnhill Industrial Group | Substack: AI for Engineers Version 2.0 — October 2026 (supersedes May 2026 edition) Methodology document for the Earth Desk project
Every number in this paper traces to a named public source, and the source's own publication date is given alongside it. Coverage is uneven. The United States, the EU, Australia, and a handful of other countries publish detailed, audited data on reserves, production, water, and air. Much of the world does not. Where data is thin, the figures here are best estimates built from what is available, and they are labeled as estimates. They will be replaced as better information is published. Bad data is no data; a number without a source and a date is not presented here.
Reserves (economically extractable now, at current prices and technology) are distinguished throughout from resources (geologically present but not yet economical). The ledger uses reserves. Where the first edition of this paper mixed the two, it has been corrected.
The static lifetime index (reserves ÷ annual production) is a snapshot under current load. It is a warning flag, not a countdown: reserves grow with exploration and price, and shrink with demand. Compounding demand collapses static lifetimes far faster than the headline numbers suggest, and that is the point of running this as a live meter rather than a one-time table.
We are not running out of the Earth. We are running through the easy, concentrated, low-cost fraction of what it holds: the geological inheritance of 4.5 billion years of pressure, chemistry, and time. The constraint is gradient, not volume. What remains tends to be deeper, lower-grade, more remote, or concentrated in fewer countries.
The most widely quoted static lifetimes come from the Energy Institute Statistical Review: roughly 54 years of oil, 49 years of natural gas, and 139 years of coal at then-current production.
Correction and caveat. Those ratios are computed on proved-reserve figures last updated for year-end 2020. The Energy Institute has not published refreshed global reserve totals since; its reserves methodology has been under review, and the 2026 Review (published June 30, 2026) still carries the 2020 reserves base. Meanwhile production has not stood still: the 2026 Review reports that in 2025 coal, oil, and gas all reached record highs, with oil production up 3.5%. Applied to 2025 production, the oil ratio is closer to 50 years than 56. These figures should be read as "2020 reserves against today's burn rate," and the site will say so on the meter.
The real constraint: gradient, not volume. The energy return on investment (EROI) of oil was on the order of 100:1 in the 1930s. Conventional oil today returns roughly 10:1 to 20:1; some unconventional sources such as oil sands are as low as 3:1 to 5:1. Oil doesn't run out; it runs expensive. The energy transition accelerates the economic cliff before the geological one arrives.
Phosphorus is one of the three essential fertilizer nutrients and has no substitute in agriculture. Nitrogen can be pulled from the atmosphere; phosphorus comes almost entirely from mining phosphate rock.
Reserve picture (USGS Mineral Commodity Summaries, February 2026): - World reserves: 73 billion tons of phosphate rock - World production 2025: 250 million tons (up 5% from 239 million tons in 2024) - Static lifetime at flat production: approximately 290 years - Morocco and Western Sahara: 50 billion tons, about 68% of world reserves - China: 110 million tons produced in 2025 (44% of world output) against 3.4 billion tons of reserves, a 31-year static lifetime for the world's largest producer - World resources: more than 300 billion tons, much of it on continental shelves and seamounts
Correction. The first edition cited "peak phosphorus before 2040." That claim originates in work predating the 2010–2011 USGS revision that raised Morocco's reserve estimate roughly tenfold, and USGS's own current assessment is that there are no imminent shortages of phosphate rock. The honest framing is not a near-term peak but three structural facts: concentration (one country holds two-thirds of reserves), the depletion rate of the largest producer (China, ~31 years), and declining ore grade in legacy producers (U.S. production has held near 20 million tons for years as Florida reserves decline and P2O5 content falls, driving rising imports).
Positive wildcard. A 2025 study (Scholz et al.) estimated that phosphorus already stored in croplands and grasslands but not readily plant-available amounts to 32–41% of 2020 geologic reserves, equivalent to 146–186 years of fertilizer application at 2020 rates. If agronomy can unlock legacy soil phosphorus, the runway extends.
Policy marker. Phosphate rock was added to the U.S. Critical Minerals List in November 2025.
No copper, no grid, no motors, no electronics.
Reserve picture (USGS MCS, February 2026): - World reserves: 980 million tons - World mine production 2025: 23 million tons - Static lifetime: approximately 43 years - Identified resources (as of the 2015 USGS global assessment): 1.5 billion tons unextracted; undiscovered resources estimated at a further 3.5 billion tons
Correction. The first edition computed a 60–70 year lifetime from the 1.5-billion-ton identified-resource figure. That number is resources, not reserves. On reserves the static lifetime is about 43 years.
The ore grade problem. Ore grades have declined from more than 2% copper in the early 20th century to under 0.5% at many major mines, meaning more than 200 tons of rock processed per ton of copper, with energy intensity rising as grade falls. USGS reported that 2025 U.S. output was reduced by lower ore grades at multiple mines.
Demand. The IEA projects copper demand growth on the order of 30% by 2040 under stated policies, driven by grids, electrification, and renewables.
Policy marker. Copper was added to the U.S. Critical Minerals List on November 7, 2025 (90 FR 50494), alongside lead, potash, rhenium, silicon, silver, boron, metallurgical coal, phosphate rock, and uranium.
REEs are not geologically rare; they are rare in minable concentrations and rare outside Chinese processing.
Reserve picture (USGS MCS, February 2026): - World reserves: more than 85 million tons REO - World mine production 2025: 390,000 tons - China: 44 million tons of reserves (about 52% of world) and 270,000 tons production (about 69% of world mine output; China's figure is a quota and excludes undocumented production) - Static lifetime on reserves alone: more than 200 years; volume is not the constraint
Correction. The first edition gave China's reserve share as 38%. That was against an older, larger world total; USGS has since revised the world figure to >85 million tons, which raises China's share to about half.
Export controls — status as of early 2026. In April 2025 China added controls on seven heavy rare earths (samarium, gadolinium, terbium, dysprosium, lutetium, scandium, yttrium). In October 2025 it extended controls to europium, holmium, erbium, thulium, and ytterbium. In November 2025 China suspended the October controls for one year; the April controls remain in force, with general export licenses being issued to selected exporters. The leverage is real and demonstrated, but the first edition's "bans active" language is now overstated and the site will track status as it changes.
Lead time. New rare earth projects require roughly 15–20 years from discovery to output: 3–5 years permitting, 2–4 years construction, 5–8 years for separation and processing capacity. Demand is projected to grow 50–60% by 2040. The supply concentration risk is structural, not cyclical.
Reserve picture (USGS MCS, February 2026): - World reserves: 37 million tons lithium content (revised up from 28 million) - World production 2025: approximately 290,000 tons excluding the U.S. (up 31% from 222,000 tons in 2024) - World consumption 2025: 263,000 tons (up 20% from 220,000 tons) - Static lifetime at 2025 production: approximately 128 years - Measured and indicated resources: about 150 million tons - Batteries: 88% of end use
Correction. The first edition generalized one year's 29% growth into "30% annually" and a 2.4-year doubling. Growth of 30% per year doubles demand in about 2.6 years, but no one should project a single year's growth as a constant; the IEA's own scenarios have lithium demand growing several-fold by 2040, not 30% compounded indefinitely. The point stands in a weaker form: static lifetime is meaningless under compounding demand.
Status. The near-term market is oversupplied; prices fell 31% in 2025 before recovering in the second half on EV and grid-storage demand. The IEA projects a deficit in the 2030s under high EV adoption. Chile, Argentina, and Bolivia hold the largest brine resources (Argentina 28 Mt, Bolivia 23 Mt, Chile 13 Mt of measured and indicated resource).
Reserve picture (USGS MCS, February 2026): - World reserves: 12 million tons - World mine production 2025: 310,000 tons - Static lifetime: approximately 39 years - Congo (Kinshasa): 230,000 tons produced (73% of world), 6 million tons reserves (50%) - Indonesia: 44,000 tons (14%), the fastest-growing producer - Identified terrestrial resources: about 25 million tons; an estimated 5 billion tons in seafloor polymetallic nodules
Correction. The first edition used ~8 million tons reserves and ~175,000 tons/yr production (a 45-year lifetime); both are superseded.
Leverage demonstrated. In February 2025 the DRC banned cobalt exports to address oversupply and low prices. In October 2025 the ban was replaced by quotas: 18,125 tons for the rest of 2025 and up to 96,600 tons per year in 2026–2027, including 9,600 tons for national strategic reserves. That quota is roughly 40% of 2025 DRC output.
Substitution. LFP (lithium iron phosphate) chemistry eliminates cobalt and already holds significant market share in China. The substitution pathway exists; retooling supply chains takes years.
Reserve picture (USGS MCS, February 2026): - World reserves: 610,000 tons - World mine production 2025: 26,000 tons (up from 25,300) - Static lifetime on reserves: approximately 23 years - World consumption 2025: 35,700 tons, meaning mine supply covers about 73% of demand; the balance comes from recycling and stock drawdown - More than two-thirds of silver is produced as a byproduct of lead-zinc, copper, and gold mining
Correction and reframing. The first edition led with "peak silver 2025–2030." Peak-production dates come from a single modeling lineage (Sverdrup et al.) and are contested. The verified, stronger facts are a 23-year reserve lifetime and a persistent supply deficit: USGS cites the continued deficit as the reason prices rose for 11 straight months in 2025 to a November high of $53.60/oz. Silver was added to the U.S. Critical Minerals List in November 2025.
Industrial collision. Photovoltaics account for about 15% of U.S. silver use and electrical/electronics 25%. Silver loadings per PV cell are falling, which partly offsets volume growth, but the demand base is now industrial, not ornamental.
Gallium is recovered almost entirely as a byproduct of bauxite processing; there are no gallium mines.
All mineral figures: USGS Mineral Commodity Summaries, February 2026 (2025 production estimates). Fossil fuel ratios: Energy Institute, reserves base year-end 2020.
| Resource | World Reserves | 2025 Production | Static Lifetime | Primary Constraint | Data Tier |
|---|---|---|---|---|---|
| Oil | 2020 base | record 2025 | ~50–54 yr | Economic before geological | Annual; reserves stale |
| Natural gas | 2020 base | record 2025 | ~49 yr | Transition pressure | Annual; reserves stale |
| Coal | 2020 base | record 2025 | ~139 yr | Climate constraint first | Annual; reserves stale |
| Phosphate rock | 73 Bt | 250 Mt | ~290 yr | Morocco 68% of reserves; no substitute | Annual |
| Copper | 980 Mt | 23 Mt | ~43 yr | Ore grade decline + demand surge | Annual |
| Lithium | 37 Mt | 290 kt | ~128 yr | Demand growth; deficit 2030s | Annual |
| Cobalt | 12 Mt | 310 kt | ~39 yr | DRC 73% of output; quotas active | Annual |
| Rare earths | >85 Mt REO | 390 kt | >200 yr | China ~52% reserves, ~69% output; controls partly active | Annual |
| Silver | 610 kt | 26 kt | ~23 yr | Mine supply ~73% of demand | Annual |
| Gallium | byproduct; >1 Mt in bauxite | 900 t | n/a | China 99% of output | Annual |
Every resource above depends on fresh water. You cannot mine, smelt, grow food, cool a power plant, or run a chemical process without it.
The satellite verdict (Abdelmohsen et al., Geophysical Research Letters, May 2025). Using NASA GRACE and GRACE-FO gravimetry from April 2002 to October 2024, researchers at Arizona State University documented: - Total terrestrial water storage loss in the Colorado River Basin: 52.2 ± 4.0 km³ (about 13.8 trillion gallons) - Groundwater share of that loss: 65%, or 34.3 ± 9.2 km³ (about 9 trillion gallons) - Lower Basin (Arizona, California, Nevada): 36.0 km³ lost, 71% from groundwater, depleting at 1.15 km³/yr - Upper Basin: 14.6 km³ lost, 53% from groundwater, depleting at 0.35 km³/yr - Annual groundwater loss averages more than 1.2 million acre-feet, about four times Las Vegas's annual Colorado River entitlement
Correction. The first edition's "13 trillion gallons of groundwater" was the total water-storage figure; groundwater alone is about 9 trillion gallons. The groundwater loss since 2002 is comparable to the full capacity of Lake Mead.
Groundwater depletion in the basin is running at more than twice the rate of surface reservoir losses. Lake Mead's bathtub ring is the visible story; the aquifer story is worse. The study's authors note that in large parts of the basin groundwater pumping remains unregulated, and that only a small fraction of Arizona's area falls under any groundwater management regime.
The root cause is agriculture, not cities. Agriculture accounts for the large majority of basin withdrawals, and a majority of that irrigates livestock feed — alfalfa, hay, corn silage. Phoenix and Tucson as urban systems can be made sustainable. Cities don't grow alfalfa.
Surface flow. The Colorado's flow has declined roughly 20% over the past century, with projections of a further decline of up to 30% by mid-century from warming and reduced snowpack. As surface water shrinks, groundwater pumping rises, which is the depletion spiral.
Fossil aquifers are geological savings accounts that took tens of thousands of years to fill. We are treating them as checking accounts. The instrument to measure the balance has existed since 2002. The policy response has lagged by decades. That is a sensor and governance failure, not just a resource failure.
Coverage note. The Colorado Basin is among the best-instrumented aquifer systems on Earth. GRACE sees the whole planet, but ground-truthed, regulated, published aquifer data of this quality exists for very few regions. The Earth Desk water layer will start where the data is real.
There is a version of this paper that ends in crisis and another that ends in systems redesign. The difference is recycling, as hard-nosed supply-chain engineering under permanent scarcity pressure.
The scorecard (rates are global estimates from IEA, WEF, and industry bodies; the reporting year varies and is noted on the site per metric):
| Material | Recycled share of supply | Energy saved vs. virgin | Status |
|---|---|---|---|
| Steel | >85% | ~74% | Mature; the model for others |
| Aluminum | ~26% recycled share of supply | ~95% | Economics overwhelming; collection lagging |
| Copper | ~33% (down from ~37% in 2015) | ~85% emissions reduction | Moving backward |
| Lithium (batteries) | ~20% of available feedstock | significant | Early; EOL flood 2035+ |
| Cobalt/nickel (batteries) | >40% of available feedstock | significant | Growing from low base |
| Phosphorus (wastewater) | <5% | n/a; no substitute | Critical failure; struvite recovery proven |
| E-waste | ~22% | high | ~64 Mt generated (2022); ~78% not formally recycled |
| Silver (U.S.) | ~11% of apparent consumption | high | USGS 2025 |
What the numbers mean. Steel shows what mature, industrial-scale recycling looks like after decades of investment. Aluminum recycling uses about 5% of the energy of primary production, yet the recycled share is still only about a quarter. Copper tells the most important story: the recycled share fell from 37% in 2015 to 33% in 2023 even as demand accelerates. In the U.S., scrap contributed about 30% of copper supply in 2025 (USGS), but most of that is new manufacturing scrap, not post-consumer recovery.
Battery metals. Recovery against available feedstock surged in 2023, but feedstock is currently manufacturing scrap. From roughly 2035 onward, end-of-life EV and storage batteries become the dominant feedstock. The recycling capacity must be built a decade before the flood arrives, which means investing ahead of economic justification. Markets will not build what they cannot yet justify. Policy and capital allocation must lead.
Phosphorus. Struvite precipitation from municipal wastewater is demonstrated at commercial scale. Global recovery remains below 5% of consumption. We mine phosphate rock from Morocco, run it through soil and food, and flush it into the ocean. No other resource failure is this structurally simple to fix and this neglected.
The market. The metal recycling market was roughly $300 billion in 2025 and is projected to grow at about 8.5% per year through 2033. That lags the demand curves it needs to offset.
The engineering argument. Recycling is not competing with mining; it is replacing mining at lower cost, lower energy, lower geopolitical exposure, and lower environmental load. The reasons it has not displaced more primary production are infrastructure gaps, weak collection frameworks, and commodity prices historically too low to justify capital. All three are changing. The nations and companies that build closed-loop recovery infrastructure in the next decade will carry a structural material cost advantage. This is not a sustainability pitch; it is a competitive engineering argument and should be made in exactly those terms.
The technician's reading of this data is not doom. It is constraint-driven design pressure, the oldest discipline in the field.
Every scarcity above is already generating substitution pressure: LFP eliminates cobalt, sodium-ion may ease lithium, optical fiber displaced copper in telecom. One constraint has no substitution pathway: phosphorus. You cannot photosynthesize without it.
The deeper pattern is geographic concentration overlaid on demand growth. The second insight is that recycling rate is the hidden variable almost every static analysis ignores. Silver's above-ground societal stock exceeds annual mine output by orders of magnitude. The difference between civilizational crisis and managed transition may live entirely in the recycling buildout.
The substrate is finite. The ledger is real. The engineering response is systems design: closed-loop, constraint-grounded, materially honest. Not panic. Not denial. Design.
| Item | v1 | v2 | Source |
|---|---|---|---|
| Copper lifetime | 60–70 yr (used resources) | ~43 yr (reserves 980 Mt / 23 Mt) | USGS MCS 2026 |
| Cobalt reserves / production | 8 Mt / 175 kt / 45 yr | 12 Mt / 310 kt / ~39 yr | USGS MCS 2026 |
| Lithium reserves | 28 Mt / 100+ yr | 37 Mt / ~128 yr | USGS MCS 2026 |
| Silver | "Peak now"; 560 kt | 610 kt / ~23 yr; deficit framing | USGS MCS 2026 |
| Phosphate | 72 Bt / 220 Mt / 300 yr; Morocco 70%; peak by 2040 | 73 Bt / 250 Mt / ~290 yr; Morocco 68%; "no imminent shortage" per USGS | USGS MCS 2026 |
| REE China reserve share | 38% | ~52% (44 Mt of >85 Mt) | USGS MCS 2026 |
| Gallium / REE export bans | "active" | Gallium ban to U.S. lifted Nov 2025 (1 yr); Oct 2025 REE controls suspended 1 yr; Apr 2025 controls remain | USGS MCS 2026 |
| Fossil fuel R/P | attributed to EI 2025 | EI reserves base is year-end 2020; production record in 2025 | EI Review 2026 |
| Colorado Basin groundwater | 13 trillion gal | 9 trillion gal groundwater (34.3 km³); 13.8 trillion total storage | Abdelmohsen et al. 2025 |
| Lithium doubling | 2.4 yr at 30%/yr | 2.6 yr; single-year growth not a trend | arithmetic |
| Critical minerals list | "2025" | Final list Nov 7, 2025; added copper, silver, phosphate rock, others | 90 FR 50494 |
| Source | Covers | Cadence | Access |
|---|---|---|---|
| USGS Mineral Commodity Summaries 2026 (Feb 2026), pubs.usgs.gov/periodicals/mcs2026 | All mineral reserves/production | Annual (Jan/Feb) | PDF + data release |
| Energy Institute Statistical Review of World Energy 2026 (June 30, 2026) | Oil/gas/coal production, consumption; reserves (2020 base) | Annual (June) | PDF + XLSX |
| EIA Weekly Petroleum Status Report; EIA STEO | U.S. and world oil/gas consumption | Weekly / monthly | API |
| IEA Global Critical Minerals Outlook 2025 | Demand projections | Annual | |
| IEA Recycling of Critical Minerals (2024) | Recycling rates | One-off | |
| Abdelmohsen et al. 2025, GRL 52(10), doi:10.1029/2025GL115593 | Colorado Basin groundwater | Monthly GRACE updates via NASA | Open |
| USBR Lower Colorado water operations | Lake Mead / Powell levels | Daily | Public |
| Scholz et al. 2025 (phosphorus legacy stocks) | Soil phosphorus | One-off | Journal |
| WEF Copper Recycling Report 2024 | Copper recycled share | One-off | |
| Global E-waste Monitor 2024 (UNITAR/ITU) | E-waste | ~biennial |