Francey et al. 1999 · Allison / GASLAB · Koutsoyiannis 2024 · Cohler & Soon 2026
Reproducing the Suess-effect papers from the data out
Francey et al. (1999) and the CSIRO GASLAB flask programme (Allison et al., Cape Grim, CG92) are cited as the fossil-fuel fingerprint in air: falling δ¹³C. Rebuild their record, strip the documented corrections, and the mixing intercept stays near −13‰ — not the −28‰ of coal, oil and gas. The fossil number is a model isoflux, not a measurement.
Observed intercept δI
−13.3‰
Keeling plot, four stations
Fossil-fuel claim
−28‰
Andres / IPCC mix
Industrial Δδ¹³C
~1.6‰
Law Dome 1850 → flasks
Largest ice revision
0.2‰
South Pole firn, ~1900
00 — Introduction
Jonathan Cohler walks Tom Nelson through the paper this lab reconstructs: mass conservation, the Keeling relation, a four-site intercept near −13.2‰, and why that bound leaves fossil carbon a sliver of the net addition — not the −28‰ fingerprint. 1:13:38. The algebra is in §05; the Scripps flasks are in §04.
Each chip starts the talk at that mark. The primary nine are the walkthrough of this lab; the rest is the Coalition critique and the Q&A.
01 — The published curve
Law Dome cores DSS, DE08 and DE08-2 (same ice Etheridge used for CO₂) plus Cape Grim in situ from 1978. Francey et al. 1999 is the first high-precision δ¹³C history; Rubino et al. 2019 — still with Allison and Francey on the author list — is the currently served file. Pre-industrial air sits near −6.5‰. The industrial decline is real in the processed series. The question is what it is a fingerprint of.
Spline is the 50-year CSIRO fit (Rubino 2019). Points are individual ice and firn extractions. Post-1993 δ¹³C is joined with a constant −13.3‰ input — the intercept the modern flasks themselves imply.
Rust dashed line adds the documented 0.2‰ South Pole firn discrepancy around 1900 (Rubino 2013) plus a 0.03‰ scale term — an illustration of how large a “resolved” inter-record hole is next to the Suess decline. CO₂ from Law Dome spline then Cape Grim / Mauna Loa instrumental. The spline is the smoothed ice; stomatal reconstructions of the same centuries are not that flat — Stomata vs Ice.
02 — Manual edits as methods
That sentence is Francey et al. 1999, not a critic. Firn only agreed with Cape Grim after gravity, diffusion, and a latitudinal gradient. Each switch below is a documented term. On = the published atmosphere. Off = add that term back, approximating the less-processed ice. The 1999-scale bulge is the South Pole firn hole they later closed on paper.
Published reconstruction (2019 scale).
Published δI
-12.76‰
This chain δI
-12.76‰
Notice: stripping corrections moves the curve by tenths of a per mil — comparable to Francey’s “decadal steps” — but the intercept remains nowhere near −28‰. The fossil leap is not in these knobs.
δ¹³C, 1500–2025. Ice-cyan is the CSIRO product; rust is your stripped series.
03 — The mixing intercept
A Keeling plot of δ¹³C against 1/[CO₂] has intercept equal to the net isotopic signature of whatever is being added. Koutsoyiannis (2024) recovers −12.9 to −13.3‰ at Barrow, La Jolla, Mauna Loa and the South Pole, with R² ≈ 0.99 on seasonally adjusted air. Fossil carbon is near −28‰. If the rise were a fossil dump, the intercept would sit there — or drift toward it as the fuel mix went gassier. It does not.
Open the axis to −28‰ to see the gap. Seasonal loops at Mauna Loa locally approach −25‰ (photosynthesis). The long-term intercept does not.
Fit intercept
-13.41‰
R²
0.971
End-member mix
-13.42‰
Fossil line
−28.0‰
| Epoch | Years | δI | R² | n |
|---|---|---|---|---|
| Little Ice Age | 1550–1850 | -11.38‰ | 0.402 | 301 |
| Early industrial | 1850–1950 | -10.28‰ | 0.977 | 101 |
| Post-war | 1950–1978 | -14.36‰ | 0.997 | 29 |
| Flask era | 1978–2025 | -13.23‰ | 1.000 | 48 |
Human emissions went from ~0 to 0.5 GtC/yr in the LIA window, 0.5–5 through 1976, and more than 5 GtC/yr after 1977 (Koutsoyiannis Table 2). The intercept does not follow that staircase toward −28‰.
04 — Koutsoyiannis 2024
Koutsoyiannis, D. (2024) re-plots Scripps δ¹³C against 1/[CO₂] at Barrow, La Jolla, Mauna Loa and the South Pole, plus a 1520–1997 proxy. Seasonally adjusted air is linear with intercepts −12.9 to −13.3‰. Fossil carbon is −28‰. Live series here is the Scripps flask pairing the paper used (archive June 2025). NOAA SIL 1990–2014 is kept as a comparison. Cohler & Soon (2026) write the two-endmember bound on this intercept — open §05.
Acquisition log
Still open
Scripps CO₂ Program monthly flasks, archive 06-Jun-2025. PTB, LJO, MLO, SPO — the four sites in the paper. La Jolla is Scripps-only; NOAA has no file.
Seasonally adjusted series is what produced R² ≈ 0.99 in the paper. Open the fossil line. It misses every station.
This lab δI
-13.27‰
Paper Table 4
-13.3‰
R²
0.994
Fossil claim
−28.0‰
Equation 13 in the paper: δI = (Cfδf − C0δ0) / (Cf − C0) on moving windows. Human emissions doubled in this interval. The windows do not march to −28‰.
| Site | Paper δI | Scripps adj. | NOAA adj. | R² | n | Source |
|---|---|---|---|---|---|---|
| Utqiagvik (Barrow), Alaska | -13.2‰ | -13.21‰ | -13.24‰ | 0.992 | 471 | Scripps flasks |
| La Jolla Pier, California | -13.3‰ | -13.33‰ | — | 0.993 | 405 | Scripps flasks (Scripps-only) |
| Mauna Loa Observatory, Hawaii | -13.3‰ | -13.27‰ | -13.19‰ | 0.994 | 517 | Scripps flasks |
| South Pole | -12.9‰ | -12.98‰ | -13.03‰ | 0.993 | 484 | Scripps flasks |
| Proxy, Böhm et al. | -13.3‰ | -12.67‰ | — | 0.978 | 53 | Law Dome as Böhm Fig. 4 |
Proxy subperiods. The intercept does not follow GtC/year toward fossil.
| Band | Years | GtC/yr | n |
|---|---|---|---|
| A | 1520–1898 | 0–0.5 | 16 |
| B | 1899–1976 | 0.5–5 | 27 |
| C | 1977–1997 | >5 | 10 |
δU / δD are biosphere end-members in the paper’s two-season model, not fossil.
| Site | δU | δD | R²% |
|---|---|---|---|
| PTB | -25.4‰ | -27.6‰ | 98.7 |
| LJO | -24.6‰ | -27.6‰ | 97.8 |
| MLO | -21.2‰ | -27.6‰ | 98.1 |
| SPO | -13.2‰ | -27.6‰ | 98.6 |
| PROXY | -13.2‰ | -13.2‰ | 95.9 |
05 — Cohler & Soon 2026
Cohler & Soon, Science of Climate Change 6.4 (2026), pp. 81–99. 81–99 pages, no new measurements. The paper takes the intercept this lab reconstructs — δI = −13.2 ± 0.2‰ from Koutsoyiannis 2024a — and writes the two-endmember arithmetic the official story never substitutes. Fossil carbon of the net addition is at most 26% in the instrumental window, and the five-century constancy of that intercept collapses it to about 1.3%. The −28‰ signature is not in the flask. It is assembled afterwards.
Five stops. The mixer is the argument; the continuity table is the honesty check against this lab.
1
What this paper is
Thirteen pages of algebra on public numbers. No new flasks, no new ice. It writes the theorem this lab measures: the Keeling intercept is the flux-weighted signature of the net addition, and that number is −13‰, not −28‰.
2
Play the mixer
The two-endmember equation has one free knob once δI and δF are fixed: the natural signature δN. Drag it. Fossil dominance never appears inside the physical window.
3
Then freeze 500 years
If the intercept did not walk toward coal while emissions rose ten-thousand-fold, the fossil share of the increment was never large. That is the constancy collapse to f ≈ 0.013.
4
Name the fourth way out
The paper’s dilemma lists three inputs. This lab’s official story uses a fourth: isofluxes in a carbon inversion. That term is not in the flask. It is how −13 becomes −28 on paper.
5
Stay inside this archive
Scripps four-site flasks are in the locker. Böhm has no public table; Köhler is not a clean third replica. Reconstructions are labelled. Download what we have.
Sources and methods
Jonathan Cohler (Lexington, MA) and Willie Soon (EPSS Sopron). Under review 25 August 2026. Acknowledgments: Demetris Koutsoyiannis for the intercept papers this builds on; Claude 4.7 and Grok 4.5 as writing aids. No external funding. No new data were generated.
Three measurement chains, none of them this lab’s flasks: Scripps CO₂ Program δ¹³C at Barrow, La Jolla, Mauna Loa and the South Pole (via Koutsoyiannis 2024a); Caribbean sclerosponge DIC from Böhm et al. 2002; Law Dome ice via Köhler et al. 2006. Endmembers: Andres et al. 2000 fossil mix ≈ −28‰; Mook 1986 / Zhang 1995 ocean-outgassing ≈ −8‰.
Method: derive the Keeling relation from conservation of each isotope (their Eqs. 1–4), read δI off a linear regression of δ¹³C on 1/C, then substitute into the mixture equation. No ocean chemistry, no biosphere parameterisation, no transport scheme.
Related models
Two-endmember mix. The paper’s engine. δI = f·δF + (1−f)·δN is a definition, not a climate model.
Isoflux inversion. The official story’s extra reservoir of free parameters. Not one of the paper’s three inputs. This lab names it in the verdict: it is how −13 becomes −28 without rejecting the flasks.
Residence vs Bern IRF. Independent cross-check in §9.2: RRR τres ≈ 3.8 yr, bomb-¹⁴C τrel ≈ 17.2 yr. Convolution of GCB emissions leaves 39 or 147 GtC fossil in a 907 GtC air reservoir (4.3% / 16.2%) — not the Bern ~50% hangover.
Ice pipeline. Not in this paper. Gravity, firn diffusion, blanks and CG92 live in §02 of this lab.
Keeling 1958 is not new. The substitution into the mixture equation is what the attribution literature skipped.
Conservation of each isotope, small-ratio approximation R ≪ 1, then integration from an initial state (C₀, δ₀) gives the Keeling relation
δ¹³C(C) = δI + (δ₀ − δI) · C₀ / C
A plot of δ¹³C against 1/C is a straight line whose intercept is δI, the flux-weighted signature of Fnet. Four Scripps sites, 1978–present, return δI = −13.2 ± 0.2‰. Böhm’s sclerosponge hovers near −13‰ to 1500. Köhler’s Law Dome Keeling is cited at ~−13‰, r² = 0.96.
Write the net addition as a mix of fossil and everything else:
δI = f · δF + (1 − f) · δN
Solve for the fossil fraction
f = (δI − δN) / (δF − δN) = (13.2 + δN) / (28 + δN)
Bounds: f ≥ 0 forces δN ≥ −13.2‰. Ocean-outgassed CO₂ sits near −8‰ (DIC ≈ +1.5‰ minus ~9‰ gas-DIC fractionation). Across −13.2‰ ≤ δN ≤ −8‰, f sits in [0, 0.26]. The counterfactual δN = 0‰ — heavier than any surface reservoir — still only reaches f = 0.471. Fossil dominance (f above 0.5) needs δN above +1.6‰. The IPCC position f ≈ 1 is the limit δN → ∞.
δI and δF are pinned at the paper’s values. δN is the only free knob. The rust line is fossil dominance. It never enters the teal window.
Observed δI
-13.2‰
Fossil δF
-28.0‰
Natural δN
-8.0‰
Fossil fraction f
0.260
Physical window · f = 0.260 · mix is 26.0% fossil, 74.0% natural
Teal band: admissible δN. Teal dashed: f = 0.26 (ocean-outgassing ceiling). Rust dashed: fossil-dominance threshold f = 0.5. The paper’s Figure 1 is this function; the CSV is the same arithmetic at 0.2‰ steps.
Constancy collapse
If f were large, δI would walk toward −28‰ as fossil emissions rose from essentially zero (1500) to ~10 GtC/yr. The intercept did not walk. At the pre-industrial end, f ≈ 0, so δI(t₀) = δN(t₀) ≈ −13‰. Holding δN there, the modern mix is
fmodern = (−13.2 + 13) / (−28 + 13) ≈ 0.013
That is 1.3%, inside the ±0.2‰ intercept noise. The instrumental band [0, 0.26] collapses to a point. This lab’s honesty check: Böhm is a DIC proxy, not a flask, and we do not have a public table — see the continuity row.
Three inputs, plus one
This lab names a fourth way out, which the paper’s dilemma omits: an isoflux in a carbon inversion. Extra land and ocean terms with free signatures can always make −13‰ air look like −28‰ fuel. That term is not in the flask. It is the assembled signature.
Where the paper and this reconstruction agree, where we reconstruct, where we disagree on reading, and what only this lab unpacks.
| Claim | Paper | This lab | Status |
|---|---|---|---|
| Four-site instrumental intercept | δI = −13.2 ± 0.2‰, Scripps PTB / LJO / MLO / SPO (Koutsoyiannis 2024a) | Scripps flasks (archive Jun-2025): PTB −13.21, LJO −13.33, MLO −13.27, SPO −12.98‰. NOAA SIL comparison: PTB −13.24, MLO −13.19, SPO −13.03, CGO −13.07‰ | match |
| 500-year constancy | Böhm et al. 2002 Caribbean sclerosponge DIC ≈ −13‰ from ~1500, no significant drift | No public digitised table. Labelled reconstruction: mix δI = −13.21‰; OLS on 53 Law Dome-tied points −12.67‰ | reconstruction |
| Law Dome as a third replica | Köhler et al. 2006 Keeling of Law Dome, intercept ~−13‰, r² = 0.96 | Law Dome is in §01–03. Köhler also shows that on long timescales the ocean as a third reservoir biases the intercept toward ~−8.6‰ — not a clean two-endmember flask replica | tension |
| Two-endmember bound f ∈ [0, 0.26] | δI = f·δF + (1−f)·δN with δF ≈ −28‰ and −13.2‰ ≤ δN ≤ −8‰ | Same algebra, same numbers, live mixer in this section. Independent of which flask file you start from once δI sits near −13 | match |
| Constancy collapse f ≈ 0.013 | Pre-industrial f ≈ 0 fixes δN(t0) ≈ −13‰; modern f = (−13.2 + 13)/(−28 + 13) ≈ 0.013 | Holds if Böhm is read as δI, not as a DIC proxy with a fractionation offset. This lab labels that step a reconstruction | reconstruction |
| Three-input dilemma | Reject algebra, or the standard endmembers, or the Scripps / sclerosponge datasets | The official story’s actual escape is a fourth input: land/ocean isofluxes in an inversion. Named in the verdict, not in the paper’s three | this lab adds |
| Ice is not air | Not unpacked. Ice enters only as Köhler’s published Keeling | §02 is this lab’s unique piece: gravity, firn diffusion, latitudinal gradient, blanks, CG92, the 0.2‰ SPO firn revision | this lab adds |
| Residence of fossil carbon | RRR τres ≈ 3.8 yr; bomb-¹⁴C τrel ≈ 17.2 yr; Nfossil(2025) 39 or 147 GtC of 907 GtC (4.3% / 16.2%) | §08: IPCC turnover ~4 yr, Essenhigh 2009, Starr 1993, bomb-¹⁴C in the 5–10 year band. Same chain, different citations | match |
Hosted copies where the fetch succeeded. Publisher links otherwise. Scripps — the paper’s flask source — is now in the locker (four-site pairing, archive June 2025).
Acquisition: Cohler PDF fetched from papers.jcohler.com (665 KB, 13 pp). Köhler 2006 PDF fetched from Copernicus (open access). Böhm 2002 remains paywalled; this lab hosts a labelled reconstruction, not a digitised Fig. 4. Scripps keelinglabsites CSVs acquired (archive 06-Jun-2025) — PTB, LJO, MLO, SPO monthly flasks, paired in §04. Companion CO₂-dynamics page is linked, not mirrored.
07 — Residence time, 2025
Grok 3 beta, Cohler, Legates, Franklin Soon and Willie Soon, Science of Climate Change 5.1 (2025), pp. 13–28. Human emissions are 10 GtC yr⁻¹, about 4% of a 230 GtC yr⁻¹ gross cycle (80 ocean, 140 land, 10 human). Dividing the air by that flux is a residence time of 3.5–4 years. Table 1 sets that next to the bomb-¹⁴C e-folding and against the Bern model’s figure of more than 100 years. Kenneth Richard’s 22 August 2025 note reads the short times as “90% of human emissions since 1750 have already been removed.” The paper’s own sentence is narrower: negligible long-term retention. The calculator is the pulse arithmetic behind the note, not a new fit.
Human share of the cycle
4%
10 of 230 GtC yr⁻¹. Same order as this lab’s gross-flux card.
Mass-balance τ
3.5–4 yr
Atmospheric stock ÷ 230 GtC yr⁻¹. Koutsoyiannis 2024, as cited.
Bomb-¹⁴C midpoint
7.5 yr
Jacobson 2005. Range 5–10 years. A pulse, not the Bern leftover.
Bern model
>100 yr
IPCC AR6. Assumes sink saturation. Not a measured e-folding.
Empirical estimates in ice, the Bern model in rust. The rust bar is drawn at 120 years and labelled “greater than 100”, because the paper does not give a single Bern e-folding. Methods are theirs: spectroscopy, ¹⁴C, radiative transfer, and mass balance on one side; theoretical sink saturation on the other.
Remaining fraction of a pulse is e^(−t / τ). Ninety percent gone means 10% left, so t = −τ ln 0.1. This is not a convolution of the historical emission curve. Cohler & Soon’s published convolution at τ = 3.8 yr — fossil carbon still in the air as molecules — is the 4.3% of the reservoir already in §05.
Half of a pulse gone
2.8 yr
t = τ ln 2.
90% of a pulse gone
9.2 yr
The note’s 90%, for one year’s emissions.
Still there after 100 yr
0.00%
A single exponential. The Bern model is not this curve.
At the paper’s 4-year mass-balance time, 90% of a pulse is gone in 9.2 years. At the ¹⁴C midpoint of 7.5 years, the same mark is 17.3 years. Either way the molecules do not wait out a century. The paper’s other exhibit is the 2020 lockdown: a 0.7 GtC drop, 7% of 2019 human emissions, left no dent in the Mauna Loa curve. Sinks of that size move in months. A 1998 El Niño release of about 5 GtC is larger than half a year of fossil fuel, which is why the growth rate follows temperature.
It is a review. The residence numbers are taken from Koutsoyiannis, Harde, Harde & Salby, and Jacobson. This lab does not refit them. The isotope sentence it relies on is the same −13‰ intercept already measured in §03–§05, now published as Cohler & Soon, SCC 6.4 (2026).
It is not a claim that the concentration rise is zero. A short residence time says a carbon atom leaves quickly. The adjustment time of the concentration is the argument the Bern model is making, and the paper rejects that model’s saturated-sink premise. The two clocks stay labelled.
The same article also scores CMIP runs against UAH and USCRN, and TSI reconstructions against temperature. Those are outside this lab’s isotope brief. They are in the PDF, not re-plotted here.
4 November 2026. The Keeling-relation paper, now SCC 6.4, pp. 81–99. Stable intercept back to about 1500, fossil fraction of the net addition at most 26% and, under five-century constancy, about zero. That section is §05. The journal link is on that page.
22 August 2025. This residence-time review. The 90% sentence is his summary of a 3–7 year turnover applied to emissions since 1750. The table and the pulse clock are the paper. The convolution that turns the clock into a stock is §05’s 4.3%.
08 — Allison / GASLAB flags
Cape Grim is not a random sample of Southern Hemisphere air. Baseline hours require a 190–280° wind sector off the Southern Ocean, low CO, then an IRMS precision cut on the MAT252. CSIRO’s own anniversary plot of the instrument marks retained ◊ against rejected +. Francey 1999 “confirms” the Cape Grim in-situ trend with the flask programme Allison documented for CG92 — without publishing the rejected list. The ensemble below reconstructs that protocol on 1990–2003 (the GASLAB network era circulated as Allison et al. 2003). It is a methods reconstruction, not a leaked sample sheet.
Cyan = baseline retained. Rust = rejected (wind, CO, or IRMS).
Retained
118
Rejected
102
δI baseline
-13.2‰
δI all flasks
-14.0‰
Current fit: -13.16‰. Rejected air is systematically more depleted — local biosphere, not “clean” marine baseline. Keeping only the baseline is a scientific choice. It is also a manual edit of the modern end-member Francey tied the ice to.
09 — The edit ledger
None of these are emails. They are the papers’ own methods. Stacked, they are large enough to sculpt “decadal steps.” None of them produce a −28‰ intercept. That last row is the only one that converts the data into a fossil signature — and it is not a measurement.
| Step | Source | Magnitude | Share of Δδ |
|---|---|---|---|
Gravitational settlingcorrection Heavier ¹³CO₂ settles in the firn. The published ‘atmosphere’ is the ice measurement minus this model term. Firn only matched Cape Grim after this correction. | Francey et al. 1999 1999 | 0.04–0.15‰ | 3–8% |
Firn diffusion / disequilibriumcorrection Diffusive smoothing and atmosphere–firn lag. Largest while CO₂ is rising fast. Depends on a CSIRO firn model that was itself revised in 2013. | Francey 1999; Trudinger model 1999 | 0.02–0.10‰ | 1–6% |
Cape Grim–Antarctica gradientcorrection An assumed south–north isotopic slope is applied so ice matches the Cape Grim in-situ record. The gradient is not measured in the same ice. | Francey et al. 1999 1999 | 0.05–0.15‰ | 3–8% |
‘Complex calibration strategies’scale Francey’s own words. Errors assigned sample-by-sample; core-versus-core, ice-versus-firn and firn-versus-troposphere uncertainties up to ±0.05‰ are listed separately — and then the spline is drawn through the result. | Francey et al. 1999 1999 | ±0.025–0.07‰ per sample | 2–4% |
Extraction blank / BFI biascorrection Cheese-grater dry extraction plus bubble-free ice tests. A laboratory blank of the same order as the ‘decadal steps’ Francey 1999 treated as climate. | Rubino et al. 2013, 2019 2013 | 0.025–0.07‰ | 2–4% |
South Pole firn discrepancy ‘resolved’scale A 0.2‰ mismatch with South Pole firn — a tenth of the entire industrial decline — was closed by re-running gravity, diffusion and the calibration scale. That is a revision, not a new measurement of 1900. | Rubino et al. 2013 2013 | 0.2‰ around 1900 CE | 10–12% |
CG92 → CSIRO2005 scalescale Allison and Francey found a systematic flask versus in-situ δ¹³C difference at Cape Grim ‘of the correct sign and magnitude to resolve the difference’. The modern end of Francey 1999 sits on the earlier CG92 scale. | Allison & Francey 2007 (methods from GASLAB 1999–2003) 2007 | flask–in situ offset, ~0.02–0.05‰ | 1–3% |
Flask flags & baseline wind sectorselection Cape Grim ‘baseline’ is not all air: wind 190–280°, low CO, then IRMS precision flags. CSIRO’s own MAT252 plot marks retained diamonds versus rejected pluses. The 1999 paper does not publish the rejected-sample list. | Allison / CSIRO GASLAB, Cape Grim programme 2003 | ~40–50% of flasks rejected in typical baseline programmes | trend-sensitive |
ICEBASE rule-based selectionselection 25 years of ICELAB–GASLAB measurements now pass through a database that ‘automatically’ corrects and selects. The public files are the output of those rules, not the input. | Rubino et al. 2019 2019 | undisclosed per-sample | unknown |
50-year spline on δ¹³Cspline Francey 1999 advertised a ‘series of steps’ from 1860 to 1960 as enhanced decadal sensitivity. A heavy spline through sparse ice samples will manufacture steps from noise and core offsets. | Rubino et al. 2019 (Francey 1999 also smoothed) 2019 | attenuates variations <20 yr by ~50% | creates ‘steps’ |
Isoflux model (the fossil leap)model The data’s own mixing intercept is about −13‰, stable since the Little Ice Age (Koutsoyiannis 2024). Reaching the fossil-fuel number requires a large, poorly observed isoflux from oceans and soils. That term is a model, not an isotopic fingerprint. | Tans / Keeling / IPCC carbon-cycle inversions 1993 | converts −13.3‰ into −28‰ | 100% of the attribution |
Francey, R.J., Allison, C.E., Etheridge, D.M., Trudinger, C.M., Enting, I.G., Leuenberger, M., Langenfelds, R.L., Michel, E. & Steele, L.P. (1999)
Tellus B 51(2), 170–193
Law Dome ice + Cape Grim in situ/archive show a decline from ~1860, read as the Suess effect — ¹³C-poor fossil carbon entering the air.
Allison, C.E., Francey, R.J. & CSIRO GASLAB (1999–2003 network product)
Baseline Atmospheric Program; CDIAC / GLOBALVIEW-CO₂C13 circulation, 2003
The modern end-member Francey 1999 ‘confirmed’. Flask versus in-situ, baseline selection, and the CG92 calibration that Allison & Francey 2007 later had to verify — and adjust.
Allison, C.E. & Francey, R.J. (2007)
J. Geophys. Res. 112, D21304
Documents a flask–in situ δ¹³C offset ‘of the correct sign and magnitude to resolve the difference’ that had opened in the Cape Grim records used in 1999.
Rubino, M. et al., including Allison & Francey (2013, 2019)
JGR 2013; ESSD 11, 473–492, 2019
Same laboratory, new gravity/diffusion/blank/scale/selection. A 0.2‰ South Pole firn hole around 1900 is closed on the page, not in 1900.
Koutsoyiannis, D. (2024)
Sci 6, 17
Keeling intercepts −12.9 to −13.3‰, stable since the Little Ice Age. Natural T-driven [CO₂] changes >3× human; fossil ≤4% of the gross flux. The 2019 spline’s growth rate tracks temperature.
10 — The Keeling curve
Journals print the Keeling curve as a smooth rise. NOAA’s own file carries both columns: the monthly air, and a deseasonalized fit. Open two or five years and the 6–8 ppm seasonal bite is the biosphere — a larger annual exchange than fossil adds. Fossil GtC changed by a factor of more than three since the 1960s; the sawtooth did not become a fossil staircase. That is the reading below, sourced, not a hidden dump.
In situ monthly · 1958–2026 · 821 months · NOAA deseasonalized fit as the publication smooth
1958–2026
ΔCO₂ in window
112.1 ppm
Seasonal amplitude
5.8 ppm
If fossil all stayed
201.4 ppm
r(growth, fossil)
0.78
NOAA GML Mauna Loa in situ monthly (Keeling / Lan). Ice = monthly (or weekly) air. Ink = the smooth journals print. Rust = world fossil+cement GtC (OWID / GCP), right axis. Scripps weekly timed out; Mauna Loa weekly is NOAA GML.
Fossil carbon rose from about 3 GtC/yr in the 1960s to more than 10 GtC/yr now. If that pulse stayed in the air, each year’s increment would scale with emissions (1 ppm ≈ 2.12 GtC). In this window the observed rise is 112.1 ppm; the fossil total, if it all stayed, is 201.4 ppm. The leftover fraction is the usual airborne identity. The shape is the issue: r(annual growth, fossil GtC) at this station is 0.78. Pinatubo, El Niño, and the seasonal bite do not sit on the rust staircase.
Mean seasonal amplitude here is 5.8 ppm — a biosphere exchange of order 10+ GtC, comparable to or larger than a year of fossil. Zoom to two years. That is what the atmosphere is doing. The publication smooth erases it.
IPCC AR4 itself puts the turnover time of CO₂ in air at about 4 years — the mean wait before a molecule is taken up by ocean or land. The 100-year figure is a different quantity, the adjustment time of a leftover concentration anomaly in a buffered ocean. The bomb-¹⁴C pulse is the experiment: after 1963 the excess ¹⁴CO₂ left the troposphere on an e-folding of order 10 years (Hua et al. 2013; Levin et al.). Essenhigh (2009) and Starr (1993) recover 4–15 years from mass balance. A short cycle for anthropogenic carbon is not a crank number. It is the turnover IPCC already published, plus the bomb tracer.
Expected = mean fossil GtC / 2.12. Ratio is observed growth over that ceiling. A fossil dump that stayed would sit near 1.0 and would track the rust column as emissions tripled. It does not.
| Decade | Observed ppm/yr | Fossil GtC/yr | If all stayed | Observed / stayed |
|---|---|---|---|---|
| 1960s | 0.86 | 3.05 | 1.44 | 0.60 |
| 1970s | 1.22 | 4.71 | 2.22 | 0.55 |
| 1980s | 1.64 | 5.49 | 2.59 | 0.63 |
| 1990s | 1.53 | 6.43 | 3.03 | 0.51 |
| 2000s | 1.91 | 7.89 | 3.71 | 0.51 |
| 2010s | 2.40 | 9.66 | 4.55 | 0.53 |
Suggested, not exclusive: the sources that actually move this curve on the year-to-year and seasonal scales are natural, and anthropogenic CO₂ is cycling through the exchangeable pool on the 5–10 year turnover the bomb pulse and IPCC’s own 4-year number already imply. The 2019 spline and T → CO₂ section test the same claim against temperature. Switch stations — Arctic sawteeth are huge; the South Pole is almost flat seasonally — the fossil overlay does not snap to either.
Acquisition notes: Scripps weekly in situ timed out; substituted NOAA GML weekly Mauna Loa.
11 — Data locker
Every series in this lab is either an acquired public file, a labelled reconstruction, or a failed fetch with the publisher link. Nothing is presented as a leaked CSIRO dump.
Every cleaned CSV in this lab, plus NOAA monthly originals. Event-level flask dumps are listed separately — they are large.
1:13:38 walkthrough of the Keeling-relation paper. Playable as §00 on this page.
Fossil mix −28‰, C3 −27‰, C4 −13‰, ocean evasion −8.5‰, observed intercept −13.2‰. Textbook values, cited in the sources section.
Equation (8) of Cohler & Soon: f = (δI − δN)/(δF − δN) with δI = −13.2‰, δF = −28‰, stepped 0.2‰. Recreated here; not a new measurement.
Mauna Loa in situ plus CCGG flask monthly CO₂. Publication smooth is NOAA deseasonalized (MLO) or a 12-month mean.
NOAA GML weekly Keeling curve. Scripps weekly timed out from this lab.
Reanalysis of Nikolov & Zeller 2024 and Goessling, Rackow & Jung 2024. CERES-era warming tracks Bond albedo (clouds), not ΔF or EEI. Figures and series live in the atlas.
Annual join used in Conclusions: Rubino spline, HadCRUT5, GISTEMP, OWID/GCP fossil (GtC), NOAA global CO₂ growth.
World entity only, 1750–2024. Converted to GtC in the overlay (MtCO₂ / 3664).
PTB/LJO/MLO/SPO = Scripps flasks (archive 06-Jun-2025). CGO = NOAA SIL + CCGG. LJO Table 3/4 reconstruction kept as a labelled leftover.
NOAA BRW flasks, 1990–2014. Comparison series. Paper intercept −13.2‰; Scripps adj. −13.21‰.
NOAA MLO flasks, 1990–2014. Comparison series. Paper −13.3‰; Scripps adj. −13.27‰.
NOAA SPO flasks, 1993–2014. Comparison series. Paper −12.9‰; Scripps adj. −12.98‰.
NOAA CGO flasks — not in the paper’s four sites, included because Allison / GASLAB used this station.
Scripps-only station. NOAA 404. Paired from monthly_flask_c13_ljo.csv and monthly_flask_co2_ljo.csv (archive 06-Jun-2025). Adj. δI = −13.33‰, R² = 0.993. Matches paper −13.3‰.
Leftover. Built from Table 3/4 on a Mauna Loa CO2 backbone before Scripps LJO arrived. Not the live series.
No public digitised table. 53 Law Dome annual points, counts 16/27/10 as in Koutsoyiannis Table 2. Mix δI = −13.21‰. The atmospheric panel only — not sponge DIC.
Rubino et al. 2019 ESSD, NOAA WDS Paleo 25830. Co-authors include Allison and Francey.
Same cores Francey 1999 used (DSS, DE08, DE08-2).
50-year CSIRO spline plus post-1993 constant-input join at δI = −13.3‰.
Keeling / NOAA GML 1958–2026. Acquired 2026-08-27.
CSIRO / BoM Kennaook Cape Grim in situ.
Amplitudes match published Scripps/CSIRO cycles. Used in the interactive Keeling plot. Not raw flask files.
Methods reconstruction of Cape Grim baseline wind sector and CO flags. Not the unpublished rejected-sample list.
Individual SIL flasks with CCGG-style flags (... retained, !.L rejected). Vintage 1990–2014.
Same flag protocol at Utqiagvik. Download the original NOAA text.
Original NOAA SIL event file.
Original NOAA SIL event file.
Four-site paired monthly, archive dated 06-Jun-2025. Source: R. F. Keeling and S. J. Walker. Unfilled column 5 only. PTB, LJO, MLO, SPO.
Original Scripps δ¹³C monthly, Mauna Loa. Archive 06-Jun-2025. Through 2024-12.
The Scripps-only isotope file. NOAA has no LJO. Through 2024-10.
Original Scripps δ¹³C monthly, South Pole. From 1977.
Original Scripps δ¹³C monthly, Point Barrow. From 1982.
Original monthly_flask_co2_{mlo,spo,ljo,ptb}.csv used to pair with δ¹³C. In the zip under raw/.
The concatenated NOAA paleo template as served. Cleaned CSVs above are parsed from the in-app copy of this record.
Papers and live archives, including the ones that failed here.
Five other desks in this project share Law Dome, HadCRUT5 and the same rule that a processed curve is not a measurement. Data and figures live on those pages.
Client-side export of the in-memory Law Dome annual series is also available if a static file 404s.
12 — Sources and sinks
The stamp in the banner is an inventory of coal, oil and gas. The flasks measure a different number. This section asks what published carbon flows actually have δ¹³C near −13‰, how large they are next to the airborne rise, and which bookkeeping step turns that rise into a fossil detection. Temperature — the quantity §14 reads off CERES albedo, not off a greenhouse forcing — is what moves the natural terms.
Observed δI
−13.2‰
Flasks. Not the fuel inventory.
Fossil inventory
−28‰
Andres / δF. Off the intercept by 15‰.
2024 budget hole
−1.7 GtC
GCB: sources 11.6, sinks + air 13.3. The books did not close.
Airborne vs gross
~3%
5.6 GtC yr⁻¹ is 2.8% of a 200 GtC gross cycle, 3.2% of 177.
Gross, not net
Soil respiration sits inside the land return — it is not added on top of GPP. Fossil is the rust bar. Same units, GtC yr⁻¹.
The identity
Airborne growth is half of fossil + land-use only because the ratio is defined that way. 2024, an El Niño year, missed by −1.7 GtC.
2024: fossil 10.3 + land-use 1.3 − air 7.9 − ocean 3.4 − land 1.9 = -1.7 GtC. GCB also says warming itself cut the land sink 25% (0.8 GtC yr⁻¹) and the ocean sink 7.1% (0.20 GtC yr⁻¹) over the decade. That is T → CO₂, inside their models.
The observed intercept is the ice tick. Fossil fuel and C3 leaves are fifteen per mil to its left. C4 crops and a blend of ocean evasion with a little C3 land on it. Natural gas is further left — a larger gas share would have moved δI, and it did not.
Weights are relative. δ ocean = -8.5‰, C3 = -27‰, C4 = -13‰, fossil = -28‰. The ice line is the flask intercept. Crank fossil up and the mix walks toward the stamp.
Mix δ¹³C
-13.4‰
Weighted mean of the four end members.
Offset from δI
-0.2‰
Zero means the mix matches the flasks.
Fossil share of the mix
4%
Cohler’s constancy bound on Fnet is ~1.3%, and at most 26%.
Takahashi et al. 1993: pCO₂ multiplies by e^(0.0423 ΔT) if DIC and alkalinity stay fixed. Pre-industrial 278 ppm. This is a real source and, alone, too small for the whole rise.
Isochemical rise
+12 ppm
One box, alkalinity fixed.
Observed, 1750–2024
+145 ppm
278 → 422.8 ppm, GCB 2025.
Share of the rise
8%
Do not promote this box to the whole ocean.
A warmer skin of seawater outgasses. It does not, by this one-box sum, equal +145 ppm. The larger lever is a few percent of the gross exchange, and the field fact that the growth rate follows temperature rather than the fossil ramp. GCB’s own ocean term from warming is a 0.20 GtC yr⁻¹ weaker sink, not a sign error.
If a fraction of the two-way exchange fails to return, the atmosphere sees a source. 5.6 GtC yr⁻¹ is the decade’s airborne growth.
Implied source
5.6 GtC
Of 200 GtC yr⁻¹.
Airborne, decade
5.6 GtC
GCB 2015–2024.
Versus airborne
1.00×
1.00× is the whole rise, with no fossil term required by the arithmetic.
This is a sensitivity, not a detection. It shows why a net budget of a few GtC cannot by itself name the source: the same airborne number falls out of a 2.8% leak in the gross cycle. Land-use bookkeeping is uncertain by ±0.7 GtC yr⁻¹ on a 1.4 GtC term. Inland waters evade ~2.1 GtC yr⁻¹ (Raymond et al. 2013) and are usually hidden in the land residual.
Uses the warming slider above. Q10 = 2 on the entire ~93 GtC soil flux is an upper picture. The chamber constraint is milder, and it is still the same direction as GCB’s climate cut to the land sink.
Naive Q10 extra
+6.7 GtC
93 × (2^(ΔT/10) − 1). Too clean to be the measurement.
Field slope
+1.6 GtC
~1.6 GtC per °C, Bond-Lamberty 2018 review.
GCB land, climate only
0.8 GtC
25% weaker land sink, 2015–2024. Net, not gross.
Net land-use
1.4±0.7
Deforestation ~1.9, regrowth ~−1.3. Farming’s annual book.
C4 crops — maize, cane, sorghum, millet — photosynthesize at about −13‰, which is the intercept, and Still et al. (2003) put C4 near a fifth to a quarter of terrestrial GPP. Sanderman et al. (2017) estimate a 133 PgC soil-carbon debt from agriculture. That is a stock accumulated over millennia, not this year’s 5.6 GtC; it is listed so it is not silently added to the annual rise. Volcanoes are under 0.1 GtC yr⁻¹ (Gerlach 2011). They are not the residual.
Where the standard account slips
1Inventory in place of the intercept
δF ≈ −28‰ is a fuel assay. δI ≈ −13.2‰ is the regression on the air. Isofluxes are the extra terms that make the first number ‘explain’ the second. They are fitted. The flasks were not.
2An identity booked as a detection
GATM = EFOS + ELUC − SOCEAN − SLAND closes by construction once a sink is the residual, and it still missed 2024 by −1.7 GtC. Closing the books is not a measurement that the rise was fossil.
3Nets quoted, gross fluxes put off stage
GPP ~123, soil respiration ~93 and ocean exchange 54–80 GtC yr⁻¹ are the cycle. Fossil at 9.8 is a few percent of it. Airborne fraction hides that by dividing the rise by anthropogenic emissions only.
4The cause written into the ratio
Airborne fraction ≈ 5.6 / (9.8 + 1.4) = 0.50. The numerator is assumed to be caused by the denominator. A temperature-driven source of the same 5.6 would produce the same fraction.
5The warming assigned to the CO₂, which was assigned to the fuel
Nikolov & Zeller’s Equation 16 reproduces CERES-era temperature from albedo and the solar constant, with no greenhouse term. Goessling et al. find the same darkening and still keep a forcing residual. If the temperature came from the albedo drop, the CO₂ rise is downstream of it.
The chain this lab can actually draw
Measured
Bond albedo down, ASR up
CERES, 2000–2026. Nikolov & Zeller and Goessling agree on the darkening and disagree on the story. The satellite number is albedo.
Temperature
GSAT follows absorbed sunlight
Atlas: ASR vs GSAT R² 0.84, EEI vs GSAT R² 0.45. Equation 16 has no CO₂ term.
Carbon response
Solubility, soil respiration, weaker sinks
Henry’s law, the soil Q10, and GCB’s own 25% / 7% climate cuts. ENSO years (2024 growth 7.9 GtC) are the same process, fast.
Signature
The addition looks like −13‰
Ocean evasion near −8‰, C4 and tropical grasses near −13‰, C3 near −27‰, blended. Not the −28‰ stamp, and not a drift toward gas.
Fossil carbon is real, about 10 GtC yr⁻¹, and it exchanges on the 5–10 year turnover already measured with bomb-¹⁴C. It does not have to be zero to be the wrong signature. The conclusions next door test the increments. The CERES section is the temperature’s source.
Everything plotted above is one of these, or a number transcribed from the paper named in the table. No station was refit for this section.
δ_mix = Σ wᵢ δᵢ / Σ wᵢ
Four end members only: ocean −8.5‰, C3 −27‰, C4 −13‰, fossil −28‰. The weights are the sliders. Distance to the lab intercept is δ_mix − (−13.2). This is the same two-endmember algebra as §05, opened to the reservoirs that actually have those signatures.
ΔpCO₂ / pCO₂ = exp(0.0423 ΔT) − 1
Isochemical seawater (Takahashi et al. 1993). p₀ = 278 ppm. It is not a full-ocean carbon-cycle model: it holds DIC and alkalinity fixed, so it is a ceiling on the solubility piece of one mixed layer, and a floor on nothing.
F = G × ε
G is a gross exchange (200 GtC in this lab’s round numbers, or 123 + 54 from Beer and AR6). ε is the fraction that does not return. F = 5.6 GtC yr⁻¹ when ε is 2.8% of 200, or 3.2% of 177. The soil illustration beside it is Rₛ × (Q₁₀^(ΔT/10) − 1) with Rₛ = 93 and Q₁₀ = 2, shown next to the 1.6 GtC °C⁻¹ field slope so the illustration cannot be mistaken for the database.
Budget imbalance, as GCB defines it: BIM = EFOS + ELUC − GATM − SOCEAN − SLAND. For 2024 that is 10.3 + 1.3 − 7.9 − 3.4 − 1.9 = −1.7 GtC.
Links go to the papers. The two CSVs are the transcription used on this page, not a re-export of the Global Carbon Project workbook.
Fossil 9.8±0.5, land-use 1.4±0.7, air 5.6, ocean sink 3.2±0.4, land sink 2.4±0.8 GtC yr⁻¹. 2024 imbalance −1.7.
Net budget, 2015–2024 and the year 2024. EFOS from energy inventories. ELUC from three bookkeeping models. GATM from the marine-boundary-layer CO₂ network. Sinks from ocean models, fCO₂ products and DGVMs. BIM is the leftover.
Same terms as the paper
The spreadsheet behind the budget. ICOS Carbon Portal release. This lab does not redistribute the workbook; the table below is a short transcription.
Gross ocean–air flux rescaled to 54 PgC yr⁻¹. Decade nets: air 5.1, ocean 2.5, land 3.4 of 10.9 anthropogenic.
Gross ocean exchange and the 2010–2019 net split. Yellow arrows are pre-industrial gross fluxes. Pink arrows are Friedlingstein et al. 2020 nets. Ocean gross = 70 × 17/22 from the revised bomb-¹⁴C inventory.
123 ± 8 PgC yr⁻¹
Terrestrial gross primary production. FLUXNET eddy covariance upscaled with diagnostic models. A two-way flux, not a net source.
Upscalings cluster at 91–94 PgC yr⁻¹; one implies ~1.6 PgC °C⁻¹.
Global soil respiration and its temperature slope. Review of chamber databases (the Bond-Lamberty & Thomson 2010 set and successors). Not independent of that shared archive.
∂ln(pCO₂)/∂T = 0.0423 °C⁻¹. Equatorial Pacific is the big natural source; high latitudes are sinks.
Ocean pCO₂ and the solubility temperature slope. Ship pCO₂ climatology. The 0.0423 factor is the isochemical warming response of surface seawater (alkalinity and DIC held fixed).
About 2.1 PgC yr⁻¹
Inland-water CO₂ evasion. Streams, rivers, lakes and reservoirs. A natural source of the same order as the entire net ocean sink, usually folded into the land residual.
Roughly a fifth to a quarter of terrestrial GPP
Where C4 vegetation sits. Climate and crop maps. C4 δ¹³C is −12 to −14‰ — the intercept, not the fossil stamp.
133 PgC debt over ~12,000 years of land use
Soil carbon lost to agriculture, cumulative. Present soil-carbon map minus a no-agriculture reconstruction. A stock, not an annual flux. Do not add it to this year’s 5.6 GtC.
Deforestation ~1.9 GtC yr⁻¹, regrowth ~−1.3, net ELUC 1.4±0.7. The ±0.7 is half the term.
Farming and forests as gross fluxes. Bookkeeping (BLUE, OSCAR, LUCE), not a flask. Gross harvest fluxes mostly cancel and are not in the net.
Well under 0.1 GtC yr⁻¹
Volcanoes, as a negative control. Emission inventory. Too small, by two orders, to be the airborne rise. Left in so the map does not invent them.
GSAT from TSI and Bond albedo only. No greenhouse term. CERES-era R² 0.84 against ASR.
What set the temperature that then moves CO₂. Equation 16 on CERES albedo and measured solar. Companion atlas, §14.
2023 albedo is the satellite floor. Low-cloud loss, not TSI, not polar ice.
The same darkening, read as clouds. CERES and ERA5. They still book a greenhouse residual. The shared observation is albedo.
| Already shown | What this section adds |
|---|---|
| §03–§05. δI = −13.2‰, fossil fraction of Fnet at most 0.26 and, under constancy, ~0.013. | Names the reservoirs that actually carry −13‰ and −8‰. The mixer is the §05 algebra with four end members. |
| §08. The seasonal bite is larger than a year of fossil, and the curve does not track emissions. | Puts a δ¹³C on that bite: C3 seasonally, C4 and ocean on the long-term intercept. |
| Conclusions. Natural gross flux ~3× the fossil term in Koutsoyiannis’s sense; growth rate follows T. | Shows the arithmetic: a ~3% gross imbalance equals the airborne term the conclusions correlate with T. |
| §14 CERES. Warming tracks albedo, not forcing. | Uses that temperature as the input to Henry’s law, soil respiration and GCB’s own climate-carbon cuts. |
13 — Conclusions
Koutsoyiannis (2024) closes the carbon cycle before he closes the isotope plot: temperature-driven natural [CO₂] changes over the last 65 years are larger than human emissions by a factor of more than three, and fossil carbon is no more than 4% of the gross flux. The sources section just above is where that −13‰ intercept is placed on ocean evasion, C4 crops and soil respiration rather than on the −28‰ fuel stamp. The Rubino 2019 spline — the same CSIRO curve Francey 1999 became — is then read against HadCRUT5 and Global Carbon Project emissions. Growth rate follows temperature; the mixing intercept never leaves −13‰.
Human share of gross flux
5%
Paper: ≤4%. This year: 10.5 GtC vs land 120 + ocean 80.
T-driven natural vs fossil
more than 3×
Koutsoyiannis 2024, last 65 years. Fig. 22: post-1750 additions exceed human by ~4.5.
CO₂ growth vs T / vs E
0.83 / 0.79
NOAA global annual increment, 1959–2025. r with HadCRUT5 vs r with fossil GtC.
Land photosynthesis/respiration ~120 GtC, ocean exchange ~80 GtC, fossil+cement 10.5 GtC. The rust sliver is the anthropogenic term Francey 1999 treated as the whole story.
1960–2024 the atmosphere gained ~234 GtC (315→426 ppm). Fossil released 423 GtC in the same window. The airborne-fraction story assigns ~half the fossil pulse to the air and hides the rest in an isoflux. Koutsoyiannis’s reading: the biosphere, warmed, both emits and takes up far more than that pulse, and the net isotopic signature stays biospheric.
Each series scaled 0–1 over 1850–2025 so the shape can be compared. Levels of CO₂ always track cumulative emissions — both are integrals. The question is which driver the spline and the growth rate actually resemble.
Yes: T → CO₂ is a possible conclusion, and it is the one the causality papers actually test. Scatter concentration against temperature and the line looks like proof. Two integrals always do that. If the ocean and biosphere outgas when warmer, the growth rate should follow T — Henry’s law, soil respiration, El Niño. That is a different graph.
The trap — [CO₂] vs T
1850–2025 levels. r = 0.95. Looks airtight because both series are ramps.
HadCRUT5 T (°C) → CO₂ ppm · n = 176
Process — d[CO₂]/dt vs T
1959–2025. growth ≈ 1.13 + 1.59 × T. r = 0.89.
HadCRUT5 T (°C) → CO₂ growth (ppm/yr) · n = 67
Raw r, growth vs T
0.89
Shared industrial ramp included.
Ramp stripped, vs T
0.37
El Niño wiggles remain with temperature.
Ramp stripped, vs fossil
0.10
Emissions lose the year-to-year structure.
Reconstruct the increment from T and from fossil, same two-parameter OLS. 1998, 2016 and 2024 (El Niño) sit in the ice and paper traces; the rust fossil fit is a smooth ramp that cannot make those spikes.
Lag of the detrended increment versus lag in years. Negative: growth leads. Positive: T (or fossil) leads. Peak at lag 0 yr, r = 0.37 with temperature — contemporaneous, same-year El Niño. Fossil is a flat residual.
T → CO₂ is a possible reading. Two rising ramps are not evidence of it — or of the reverse. This is the chain the graphs above actually support, and the claims they do not.
Constrained by the series in this lab
| Diagnostic | vs temperature | vs fossil | What it means |
|---|---|---|---|
| NOAA CO₂ growth, 1959–2025 | r = 0.830 | r = 0.785 | The increment, not the integral. Temperature wins, as in Koutsoyiannis’s causality papers. |
| Spline CO₂ vs HadCRUT5 | r = 0.950 | r = 0.988 annual / 0.995 cumulative | Levels vs cumulative emissions is near 1 by construction — two ramps. Do not read it as a fingerprint. |
| Spline δ¹³C vs T | r = -0.944 | r = -0.996 | The 2019 decline is a smooth industrial-era ramp. The intercept of that ramp is still −13‰, not −28‰. |
| Growth after stripping the ramp | r = 0.365 | r = 0.101 | Year-to-year wiggles (El Niño). Temperature still organises them. Fossil GtC does not. |
Combined: the carbon cycle is mostly the biosphere answering to temperature; the isotopic signature of whatever is being added to air has been ~−13‰ since the Little Ice Age; the 2019 spline is consistent with that, and is not a chromatogram of coal. The ranked analysis above is what this lab is willing to say — and what it is not.
14 — CERES albedo
A companion atlas re-reads Nikolov & Zeller (Geomatics, 2024) against Goessling, Rackow & Jung (Science, 2024). Two groups, two journals, two methods. They agree on what CERES actually measured for the last 26 years: Bond albedo fell (clouds, not the solar constant), absorbed sunlight rose, and that extra ASR is what moved surface temperature. They split on whether a greenhouse forcing or a cloud feedback is still required. The atlas’s reading: it is not. Open it for the figures; the overlap is summarised here. The source map treats this temperature as the driver of ocean evasion and soil respiration, which is why a −13‰ intercept does not need a −28‰ fuel. Four other desks in this project — stomata versus ice, both poles versus 1988, the instrumental stack, and the ocean heat cycle — apply the same rule. See companion labs.
ASR vs GSAT (atlas)
R² 0.84
EEI vs GSAT (atlas)
R² 0.45
2023 albedo
CERES floor
CERES window
2000–2026
The overlap
Nikolov & Zeller reconstruct GSAT from TSI and Bond albedo only (their Equation 16 has no greenhouse-gas term; R² 0.84, Mar 2000–Jul 2026). Goessling, Rackow & Jung, using CERES and ERA5, find that the 2023 surge left a ~0.2 K hole after anthropogenic warming and El Niño, and that a record-low planetary albedo — fewer low clouds in the northern mid-latitudes and tropics — closed it. Polar ice is a sideshow. TSI is a sideshow. That is an independent verification of the albedo change, not a CSIRO file.
Over the CERES era the directional fact is the same: the planet reflected less sunlight, absorbed more, and warmed. The 2023 floor sits on a multi-annual low-cloud decline (Goessling: −1.27%/decade from 2013–2022), not on a solar-constant step.
The fork
Goessling et al. still sit inside a GHG + feedback + aerosol + variability frame. They treat the albedo drop as a 0.2 K patch that intensifies a greenhouse-plus-El-Niño budget, and they leave open whether the low-cloud trend is internal variability, aerosol cleanup, or an emerging cloud feedback.
Nikolov & Zeller treat the same drop as the budget. Surface temperature tracks absorbed sunlight, not the TOA energy imbalance. Apparent EEI is what you get after the atmosphere adiabatically attenuates longwave (TEFAC 0.477). Sensitivity to ASR is modest (~0.30 K per W m⁻²). If a greenhouse term were required, a solar-plus-albedo reconstruction would miss the 26-year slope. In the atlas, it does not.
Shared CERES-era observables. The diagnosis after the last row is the fork — not the satellite record.
| Observable | Nikolov & Zeller 2024 | Goessling, Rackow & Jung 2024 |
|---|---|---|
| 2023 Bond albedo | Record low in this series (α ≈ 0.284, −0.69 pp vs 2001–2022) | Record low since at least 1940. CERES 2023 is the satellite minimum. |
| Absorbed shortwave (ASR) | +2.52 W m⁻² in 2023 vs 2001–2022 (CERES-style albedo + measured TSI) | +1.82 W m⁻² CERES 2023 vs 2001–2022. Trend 2013–2022: +1.10 W m⁻²/decade. |
| What darkened the planet | Cloud and surface Bond albedo — ENSO pulses plus a secular decline. TSI is a ripple. | Fewer low clouds, northern mid-latitudes and tropics (2023: −1.5%; 2013–2022: −1.27%/decade). Polar ice ~12% of the ASR warming. |
| Solar constant (TSI) | Near-zero secular TSI. Solar-only does not carry the 21st-century slope. | Solar-cycle ISR contributes only +0.03 K in 2023. The surge is not a TSI event. |
| TOA energy imbalance (EEI) | EEI is the leftover after TEFAC attenuates surface longwave. It tracks GSAT worse than ASR (R² 0.45 vs 0.84). | CERES EEI 2023 +0.97 W m⁻² vs 2001–2022, driven mainly by the ASR rise — the same energy booked as albedo. |
| What albedo is allowed to mean | Equation 16: albedo + TSI is sufficient. No greenhouse-gas term. Radiative forcing is the wrong ledger. | Albedo fills the leftover +0.2 K after anthropogenic warming and El Niño. Still a GHG + feedback frame. |
Suggested reading of the overlap: over the CERES record the warming that actually happened tracks a darker planet — a change in reflected sunlight — not a measured greenhouse forcing and not a diagnosed feedback. Goessling et al. confirm the albedo fact in Science; they do not adopt Nikolov & Zeller’s causal order. The atlas is the place that puts both papers on one page.
15 — What follows
This lab does not invent a hidden CSIRO dump. It uses the NOAA Law Dome file (Rubino 2019, Allison and Francey co-authors), Scripps CO₂ Program flasks at Barrow, La Jolla, Mauna Loa and the South Pole (archive June 2025 — the files Koutsoyiannis 2024 actually used), NOAA GML flask δ¹³C at those sites plus Cape Grim as a comparison (1990–2014), Cape Grim and Mauna Loa CO₂, and the methods those groups published. Böhm Fig. 4 is a labelled reconstruction. Every series is downloadable or linked. Cohler & Soon (2026, SCC 6.4) write the two-endmember bound on those intercepts: f at most 0.26 in the instrumental window, collapsing to ≈0.013 if the 500-year constancy holds. Their three-input dilemma omits the isoflux — the official story’s actual fourth way out. See §05.
The official story
Plant photosynthesis discriminates against ¹³C, so coal, oil and gas — buried plants — sit near −25 to −44‰. Burning them should dilute atmospheric δ¹³C. Francey 1999 supplies the 1000-year curve; Allison’s Cape Grim flasks lock the modern end; IPCC inversions close the budget with airborne fraction ~0.45 plus ocean and land isofluxes large enough that the observed −13‰ intercept can still be blamed on −28‰ fuel.
That last clause is the whole trick. The fingerprint is inferred after a model of exchange fluxes that are not measured at the precision of the flasks.
The intercept story
Mass balance on the atmosphere alone — Keeling’s own plot, without a second reservoir of free parameters — gives a net source/sink signature of about −13.3‰ at every background station, and in the ice–firn join back to the 16th century (Koutsoyiannis 2024). Seasonal loops reach −25‰, which is the biosphere, and they do not grow as fossil emissions double.
A genuine fossil takeover would drag the intercept toward −28‰ as the natural-gas share rose. The series does the opposite in the high-emission decades. Koutsoyiannis’s carbon-cycle close is that temperature-driven natural [CO₂] changes already exceed fossil by more than 3×, with human carbon ≤4% of the gross flux — so the 2019 spline is a biosphere-on-temperature curve, not a fuel chromatogram. Two ramps of [CO₂] and T are the misunderstanding; the increment follows T once the ramp is stripped. See Conclusions and T → CO₂. A companion CERES atlas finds the same pattern in the energy budget: 26 years of satellite warming track Bond albedo, not greenhouse forcing — the overlap of Nikolov & Zeller with Goessling, Rackow & Jung. Sister desks take the same rule to ice versus stomata, both poles versus 1988, the instrumental stack, and the ocean heat cycle.
Not a forged chromatogram. A constructed claim: that a processed δ¹³C decline is “the” isotopic signature of fossil fuel. The construction has three layers, all in the open literature.
16 — Companion labs
They are not extra series on these charts — they are the arguments about ice versus stomata, the ocean heat cycle, the 1988 polar forecasts, the instrumental desk, and what CERES actually attributes the last 26 years to. Each desk uses published files, names its reconstructions, and keeps raw and smoothed traces one click apart. This lab is the δ¹³C chapter.
CO₂ proxies
Plant leaves keep a growing-season count of pores. Antarctic ice keeps bubbles of old air. Ice-core CO₂ is heavily smoothed; stomatal series still carry century-scale swings that only meet the ice at the instrumental splice. Kouwenberg’s Roman-era maximum is ~357 ppm against Law Dome ~281 ppm at AD 225. Pre-industrial Law Dome wanders ~15 ppm; Kouwenberg wanders ~106 ppm.
In this lab. The Law Dome spline this lab treats as the 1000-year ceiling is the same smoothed ice. Firn averaging is the CO₂ twin of the gravity / diffusion / blank pipeline on δ¹³C.
Open the deskBoth poles
1988-era scenarios warmed both poles under well-mixed greenhouse gas. The split is not that: Arctic air rose; Antarctic did not. Sea-level rise since 1900 is expansion-scale, not a melt pulse matching those polar-ice forecasts. Albedo — Arctic ice going dark — is a first-order term the equal-poles story never had room for.
In this lab. A well-mixed fossil fingerprint should look the same at both poles, the way a −28‰ isoflux is supposed to look the same at every flask. The polar split is the geographic version of an intercept that never drifts.
Open the deskObservational desk
A stand-alone desk of published measurements: thermometers, Mauna Loa and Law Dome CO₂, ocean / Niño, the two sea-level instruments, sea ice and snow, the solar cycle, and ice-core paleoclimate. Raw and smoothed traces are always one click apart. The charts do not argue a single attribution.
In this lab. HadCRUT5, GISTEMP, Mauna Loa and the Law Dome join used in Conclusions and the Keeling-curve section live here as the full observational stack, not just the slices this lab needs.
Open the deskZettajoule
The deglacial ocean took ~12,000–20,000 ZJ; the Medieval Climate Anomaly sat ~1,500 ZJ above the Little Ice Age trough; the industrial ocean is ~500 ZJ. Gebbie, Haeberli, Baggenstos. Modern excess heat is ~90% in the ocean partly because those ice sheets are already gone. The cycle is real; 427 ppm sits off the ice-core envelope either way.
In this lab. If temperature moves ocean heat and ocean outgassing, the mixing intercept staying near −13‰ is what a natural source looks like. The heat ledger is the energy counterpart of this carbon lab.
Open the deskAlbedo, not forcing
Nikolov & Zeller and Goessling, Rackow & Jung land on the same CERES fact: the last ~26 years of warming track absorbed sunlight from a falling planetary albedo, not the greenhouse forcing diagnostic. Albedo is down ~0.77 pp (record low ~0.284 in 2023). GSAT follows ASR at R² ≈ 0.84 versus ~0.45 for energy-imbalance. They split on story; the satellite number they both use is albedo.
In this lab. Already summarised in §14. The atlas is the place that puts both papers on one page. Same rule as this lab: the quantity that moved is the one that was measured.
Open the desk17 — Glossary
Dotted terms in the text open a short definition on hover; click or tap one to jump here. Each entry has the textbook sense, then how this reconstruction uses it. Nothing here is a substitute for Francey 1999, Allison’s GASLAB methods, Koutsoyiannis 2024, or Cohler & Soon 2026.
50 of 50 terms
Also: delta 13C · d13c · carbon-13
Textbook
The ratio of ¹³C to ¹²C in a sample, reported in per mil (‰) relative to the VPDB standard. More negative means relatively less ¹³C.
In this lab
The quantity Francey 1999 and GASLAB treated as a fossil-fuel fingerprint. This lab tracks the mixing intercept of δ¹³C vs 1/[CO₂], not just the decline.
Also: permil · per mille · ppt
Textbook
Parts per thousand. Isotope ratios are small, so they are written as a deviation from a standard in ‰ rather than as a raw ratio.
In this lab
Industrial-era atmospheric δ¹³C falls about 1.6–2.1‰. Ice-core “corrections” are tenths of a ‰. The fossil claim jumps to −28‰ — a different scale of story.
Also: keeling
Textbook
A plot of δ¹³C against 1/[CO₂]. For two-end-member mixing, points fall on a line whose intercept is the δ¹³C of the added (or removed) carbon.
In this lab
Keeling’s own diagnostic, without a second reservoir of free parameters. Four stations and the 1520–1997 proxy share an intercept near −13.2‰, not −28‰.
Also: keeling intercept · di · end member
Textbook
δI = (C₂δ₂ − C₁δ₁) / (C₂ − C₁). The isotopic signature of the net source/sink implied by two atmospheric states, or by the intercept of a Keeling plot.
In this lab
The number this lab actually measures. Observed ~−13.3‰. The fossil −28‰ is not this intercept; it is imposed later by an isoflux.
Also: suess
Textbook
The decline of ¹⁴C (and, by analogy, ¹³C) in air after the industrial era, attributed to adding ¹⁴C-free / ¹³C-depleted fossil carbon.
In this lab
Cited as the fossil-fuel fingerprint. A declining δ¹³C is real. Calling that decline “the” isotopic signature of coal is the claim this lab unpacks.
Also: isotope flux
Textbook
The product of a carbon flux and its isotopic signature. In budget inversions, land and ocean isofluxes are extra terms that close the isotope accounts.
In this lab
The term that converts an observed −13‰ intercept into a −28‰ fossil story. It is a model identity, not a flask measurement.
Also: firn air
Textbook
Compacted snow that is not yet ice. Air in the pore space still exchanges with the atmosphere until the close-off depth, typically 50–100 m.
In this lab
The join between ice and Cape Grim. Francey 1999 only made firn “agree” with Cape Grim after gravity, diffusion and a latitudinal gradient.
Also: law dome · ice
Textbook
A cylinder of glacier ice whose trapped bubbles are a sample of old air. Age is not the same as the year of snowfall; diffusion and close-off smear it.
In this lab
Law Dome (DSS, DE08) via Rubino et al. 2019 — Allison and Francey co-authors. Ice is not air until a processing chain is applied.
Also: gravity enrichment
Textbook
Heavier molecules (¹³CO₂) settle slightly in a stagnant firn column. The enrichment is roughly proportional to depth and mass difference.
In this lab
Francey 1999 subtracts this model term so the published “atmosphere” is ice minus gravity. On the order of 0.04–0.15‰.
Also: disequilibrium · trudinger
Textbook
Gases diffuse through firn. A changing atmosphere is smoothed and lagged; the bubble is not a snapshot of one year.
In this lab
A CSIRO firn-model term, larger while CO₂ is rising fast. Rubino 2013 revised the same class of model that Francey 1999 used.
Also: cape grim antarctica · gradient
Textbook
CO₂ and δ¹³C vary with latitude because sources and sinks are not uniform. Southern-ocean air is not Antarctic firn air.
In this lab
Francey 1999 applies ~0.08‰ so ice matches Cape Grim (41°S). The gradient is assumed, not measured in the same ice.
Also: cg92 · csiro2005 · scale
Textbook
Successive CSIRO laboratory calibration scales for δ¹³C. Changing the scale shifts every number on that lab’s record by a constant (plus drift terms).
In this lab
Francey 1999’s modern end is CG92. Allison & Francey 2007 document the jump to CSIRO2005 (~0.025‰). A scale is not a measurement of 1900.
Also: allison · csiro gaslab
Textbook
CSIRO’s Global Atmosphere Sampling Laboratory: flasks of air analysed on an IRMS (MAT 252) for CO₂ mole fraction and δ¹³C.
In this lab
Allison’s Cape Grim programme is how Francey 1999 “confirms” the in-situ trend. Baseline hours are a wind-sector and CO cut, not a random sample.
Also: blank · bfi · bubble-free
Textbook
Laboratory contamination or fractionation during gas extraction from ice. Bubble-free ice (BFI) is used as a process blank.
In this lab
Rubino 2013/2019: 0.025–0.07‰ — the same size as “decadal steps” Francey 1999 treated as climate.
Also: rubino · 50-year spline
Textbook
A smooth curve fitted through irregular samples. A 50-year cutoff removes high-frequency noise and, with it, some real variability.
In this lab
The Rubino et al. 2019 Law Dome annual spline — the curve Francey 1999 became. This lab reads it against HadCRUT5 and fossil GtC.
Also: ordinary least squares · regression
Textbook
Ordinary least squares: the unique line that minimises the sum of squared vertical residuals. Slope, intercept and R² follow from that fit.
In this lab
Used for Keeling intercepts and for growth ≈ a + b·T. Two-parameter OLS is a diagnostic, not a structural ocean model.
Also: r2 · pearson · r-squared
Textbook
Pearson r is the linear association of two series (−1 to 1). R² is the fraction of variance that line explains. Both treat any two ramps as “related.”
In this lab
Levels of CO₂ vs cumulative fossil give r ≈ 0.995 because both are integrals. After stripping the ramp, residual growth vs T is 0.37 vs 0.10 for fossil.
Also: residual · ramp
Textbook
Subtract a linear (or other) trend against time so the leftover is the year-to-year wiggle, not the shared industrial-era rise.
In this lab
The test that separates two ramps from a process. Residual CO₂ growth still follows temperature; it does not follow fossil GtC.
Also: airborne fraction
Textbook
The share of emitted fossil CO₂ that remains in the atmosphere, typically quoted near 0.45. The rest is assigned to ocean and land sinks.
In this lab
A budget identity, not a measured partition of this year’s air. It is how inversions hide the rest of the fossil pulse in isofluxes.
Also: gigatonne · gt carbon
Textbook
Gigatonnes of carbon. 1 ppm of atmospheric CO₂ ≈ 2.12 GtC. Fossil numbers in the wild are often Mt of CO₂, not GtC — a factor of 3.664.
In this lab
Land ~120 GtC/yr and ocean ~80 vs ~10.5 GtC fossil this year. Human share of the gross cycle is about 5% (paper: ≤4%).
Also: mole fraction
Textbook
Parts per million by mole: micromoles of CO₂ per mole of dry air. The modern Mauna Loa unit.
In this lab
Law Dome spline plus NOAA global means. Growth rate is Δppm per year — the increment, not the integral.
Also: henry · solubility
Textbook
The amount of gas dissolved in water is proportional to its partial pressure, with a coefficient that falls as temperature rises. Warmer water holds less CO₂.
In this lab
The physical reason d[CO₂]/dt can follow T: a warmer ocean and soil outgas. That is the process the two-ramps plot is not.
Also: el nino · nino · enso
Textbook
The El Niño–Southern Oscillation: a 2–7 year tropical Pacific swing. Warm El Niño years typically raise global temperature and terrestrial CO₂ release.
In this lab
1983, 1987, 1998, 2016, 2024 sit in observed growth and in the T-driven OLS. A fossil-GtC fit is a smooth ramp and cannot make those spikes.
Also: hadcrut · gistemp · temperature
Textbook
The Met Office / CRU global surface-temperature anomaly, 1850–present. GISTEMP is NASA’s related series from 1880.
In this lab
The T in T → CO₂. NOAA growth vs HadCRUT5 is r = 0.83; vs fossil GtC, 0.79. After detrending, only T still organises the wiggles.
Also: photosynthesis · c3 · respiration
Textbook
Plants prefer ¹²CO₂, so organic carbon (and coal, oil, gas made from it) is ¹³C-depleted, typically −20 to −30‰ for C3 biomass. Respiration returns that signature.
In this lab
Seasonal Keeling loops reach ~−25‰ and stay there as fossil emissions triple. The net annual intercept does not walk toward the fuel mix.
Also: minus 28 · andres · fuel mix
Textbook
Coal, oil and gas inherit plant depletion plus further processing. Inventory-weighted fossil CO₂ is often quoted near −28‰ (Andres / CDIAC / IPCC).
In this lab
The claimed fingerprint. It is not the intercept of the flasks or the ice. A genuine fossil takeover would drag δI toward −28‰ as the gas share rose.
Also: flasks · baseline
Textbook
Air collected in a glass flask at a station, later analysed in a laboratory. Selection rules decide which flasks represent “background” air.
In this lab
Cape Grim baseline: 190–280° wind off the Southern Ocean, low CO, then an IRMS cut. Retained vs rejected is a protocol, not a random draw.
Also: insitu · continuous
Textbook
Measured on site by a continuous analyser, as opposed to discrete flasks shipped to a lab.
In this lab
Cape Grim and Mauna Loa in-situ CO₂ lock the modern mole-fraction end. Allison & Francey 2007 discuss flask–in situ offsets of the size that “resolve” mismatches.
Also: gross flux · net flux
Textbook
Gross fluxes are the large two-way exchanges (photosynthesis and respiration, ocean invasion and evasion). Net is the small residual that changes the atmosphere.
In this lab
Fossil carbon is a few percent of the gross cycle. Koutsoyiannis: human ≤4% of the total; T-driven natural [CO₂] changes exceed fossil by more than 3× over 65 years.
Also: inverse · budget inversion
Textbook
A model that infers sources and sinks from observed concentrations (and isotopes), given transport and prior fluxes. Under-determined: extra terms are free.
In this lab
Where the isoflux lives. The inversion can always blame −13‰ air on −28‰ fuel if land/ocean isofluxes are allowed to do the arithmetic.
Also: two ramps · spurious correlation
Textbook
Any two series that both rise (or both integrate a positive quantity) are highly correlated. Correlation of levels is not a test of who causes whom.
In this lab
[CO₂] vs T is r = 0.95; [CO₂] vs cumulative fossil is 0.995. That line is the misunderstanding. The process test is the increment, then the residual.
Also: dco2 · annual increment
Textbook
The year-to-year change in atmospheric CO₂, in ppm/yr. Distinct from the concentration itself, which is the integral of past increments.
In this lab
The diagnostic Koutsoyiannis’s causality papers use. NOAA global increment vs T beats vs fossil; El Niño spikes follow the T-driven OLS.
Also: cgo · baseline station
Textbook
A clean-air station on the northwest tip of Tasmania (41°S), sampling Southern Ocean baseline when the wind is in sector.
In this lab
The modern end Francey 1999 tied the ice to. CG92, GASLAB flasks, and the in-situ record all meet here — after the flags.
Also: mauna loa curve · co2 record
Textbook
The continuous record of atmospheric CO₂ at Mauna Loa started by C.D. Keeling in 1958. Publications usually show a deseasonalized smooth; the monthly (and weekly) air is a sawtooth.
In this lab
Zoom the raw months and the 6–8 ppm seasonal bite is the biosphere, larger than a year’s fossil increment. The ribbon is not a fossil staircase.
Also: turnover · lifetime · e-folding
Textbook
The mean time a CO₂ molecule spends in the air before it is exchanged into ocean or land. IPCC AR4: about 4 years. Distinct from the ‘adjustment time’ of a concentration perturbation.
In this lab
Bomb-¹⁴C and mass-balance (Essenhigh 2009, Starr 1993) put the same quantity in the 5–10 year band. SCC 5.1 (2025) Table 1 puts the empirical estimates at 3–7.5 years and the Bern model above 100. That is a short cycle, not a century hangover.
Also: perturbation lifetime · revelle
Textbook
How long a *concentration* anomaly takes to decay after emissions stop, given the Revelle buffer in the ocean. IPCC quotes decades to centuries. It is a model of the leftover, not a measurement of the sawtooth.
In this lab
The quantity often quoted as ‘CO₂ lasts 100 years.’ It is not the Keeling curve, and it is not the bomb-¹⁴C pulse.
Also: bomb radiocarbon · delta 14c
Textbook
Thermonuclear tests (peak 1962–63) doubled atmospheric ¹⁴C. After the test ban the excess ¹⁴CO₂ declined as it mixed into ocean and biosphere — a dated pulse of CO₂ leaving the air.
In this lab
Hua et al. 2013 / Levin: tropospheric e-folding of order 10 years. An empirical 5–10 year cycle for carbon atoms, not a spline.
Also: sawtooth · seasonal amplitude
Textbook
Land photosynthesis draws CO₂ down in NH summer; respiration returns it in winter. Mauna Loa peak-to-trough is about 6–8 ppm each year.
In this lab
That bite is ~14 GtC exchanged seasonally — more than a year of fossil. Zoom the Keeling curve and that is what you actually see.
Also: ceres ebaF · nasa ceres
Textbook
NASA’s Clouds and the Earth’s Radiant Energy System: satellite radiometers that measure reflected shortwave and outgoing longwave at the top of the atmosphere, from 2000 to present.
In this lab
The 26-year record both Nikolov & Zeller and Goessling, Rackow & Jung read. It is where the albedo drop is measured, not modelled.
Also: planetary albedo · albedo
Textbook
The fraction of incoming solar radiation a planet reflects to space, after all cloud and surface interactions. Earth’s mean is about 0.29.
In this lab
CERES 2023 is the satellite floor (atlas: α ≈ 0.284). The last 26 years of warming track this darkening, not a greenhouse ΔF.
Also: absorbed solar · absorbed sunlight
Textbook
Incoming solar minus reflected shortwave: the sunlight the planet actually keeps. ASR = ISR × (1 − albedo).
In this lab
The quantity Equation 16 uses. Atlas: ASR vs GSAT R² 0.84 over CERES. Goessling 2023: +1.82 W m⁻² vs 2001–2022.
Also: toa imbalance · net flux
Textbook
Net flux at the top of the atmosphere: ASR minus outgoing longwave. IPCC treats a positive EEI as heat going into the ocean.
In this lab
Tracks GSAT worse than ASR (R² 0.45 vs 0.84). Nikolov & Zeller: the leftover after adiabatic attenuation of longwave, not the cause.
Also: forcing · delta F · ghg forcing
Textbook
An imposed change in net TOA flux, usually from greenhouse gases, before the surface temperature has adjusted. The IPCC control knob.
In this lab
The term Equation 16 does not have. Goessling still book most of 2023 as anthropogenic forcing; the atlas’s overlap with them is albedo, not this lab.
Also: two end member · binary mix · mixture equation
Textbook
A mixture of two sources with signatures δA and δB in fractions f and 1−f has signature f·δA + (1−f)·δB. Solving for f is unique once the mix and both endmembers are known.
In this lab
Cohler & Soon apply this to the Keeling intercept: δI = f·δF + (1−f)·δN. With δI = −13.2‰ and δF = −28‰, f cannot be large unless δN is unphysical.
Also: f · fossil share of Fnet
Textbook
The share of a net flux that comes from fossil carbon. Distinct from airborne fraction (a budget leftover) and from the fossil share of the whole atmospheric burden.
In this lab
In Cohler & Soon, f is the fossil share of Fnet — the net addition the intercept sees. Instrumental bound [0, 0.26]; five-century constancy collapses to ≈0.013.
Also: ceratoporella · coralline sponge · böhm
Textbook
A slow-growing calcareous sponge whose aragonite skeleton records the δ¹³C of surrounding seawater DIC, with a small stable fractionation and annual banding that can be dated.
In this lab
Böhm et al. 2002, Caribbean Ceratoporella nicholsoni, ~1500–present. The paper’s 500-year constancy of δI. This lab’s Böhm series is a labelled reconstruction, not a digitised table.
Also: c4 · maize · sugarcane
Textbook
A photosynthetic pathway used by maize, sugarcane, sorghum, millet and many tropical grasses. It discriminates against ¹³C less than C3 plants do, so tissue δ¹³C is about −12 to −14‰.
In this lab
That is the observed mixing intercept. C3 leaves sit near −27‰ with the fossil stamp. C4 does not. Still et al. 2003 put C4 near a fifth to a quarter of terrestrial GPP.
Also: rs · soil co2 · q10
Textbook
CO₂ leaving the soil from roots and microbes. Global upscalings cluster near 90 PgC yr⁻¹. A Q10 of 2 means the flux doubles for a 10 °C rise.
In this lab
Large enough that a 1–2% change matches the airborne rise. The field slope reviewed by Bond-Lamberty (2018) is about 1.6 PgC per °C — same direction, smaller than a naive Q10 on the whole flux.
Also: evasion · outgassing · ocean source
Textbook
CO₂ leaving seawater when surface pCO₂ exceeds the air. Solubility falls as temperature rises (Henry’s law). The equatorial Pacific is the largest natural ocean source; high latitudes are sinks.
In this lab
Evaded CO₂ is about −8 to −9‰, not −28‰. A one-box isochemical warming of 1 °C on 278 ppm gives ~12 ppm, a real piece of the rise and not the whole of it.
Also: bim · budget residual
Textbook
BIM = fossil + land-use − atmospheric growth − ocean sink − land sink. Zero means the five terms close. It is an accounting leftover, not an observed flux.
In this lab
GCB 2025 reports −1.7 GtC for 2024. The books did not close in the year the airborne growth set a record. That is the identity the fossil detection is written in.