Geomatics 2024 · updated through Jul 2026
Why they say radiative forcing is the wrong ledger: temperature tracks absorbed sunlight, not the TOA energy imbalance, and Equation 16 has no greenhouse-gas term.
Observed trend
0.27 K/dec
Eq. 16 trend
0.38 K/dec
ASR vs GSAT R²
0.84
EEI vs GSAT R²
0.45
λ (Eq. 15b)
0.297
Δ albedo
-0.77 pp
The claim
Nikolov & Zeller’s argument is not that Earth has not warmed. It is that the IPCC-style ledger — greenhouse forcing, top-of-atmosphere energy imbalance, stored heat — is the wrong way to count. Surface temperature tracks absorbed sunlight. The leftover “imbalance” is what you get after the atmosphere adiabatically attenuates longwave at TOA.
Textbook forcing narrative
Prediction: GSAT should track EEI, and the temperature equation needs a greenhouse-forcing term.
Nikolov & Zeller
Prediction: GSAT should track ASR, not EEI. Sensitivity to sunlight is modest (~0.30 K per W m⁻²).
Absorbed SW vs GSAT
R² 0.84
The sunlight diagnostic — it works.
Apparent EEI vs GSAT
R² 0.45
The forcing diagnostic — it does not.
Same months, same GSAT. Absorbed sunlight explains 84% of the variance; the energy-imbalance residual explains 45%. That is a 1.9× gap. If radiative forcing were the causal ledger, the right-hand scatter would be the tight one.
01 · Missing term
Equation 16 is only TSI and Bond albedo. There is no ΔF_CO₂, no optical depth, no EEI. It still tracks the GISTEMP–NOAA average (R² 0.84) from Mar 2000 through Jul 2026.
02 · Wrong correlation
A forcing theory needs temperature to follow the TOA imbalance. In this record it follows the absorbed shortwave instead. TSI alone is a ripple (R² 0.10); albedo-driven ASR carries the trend.
03 · Wrong sensitivity
Once the energy that actually changed is ASR (1.27 W m⁻²/decade), the implied sensitivity is 0.21 K per W m⁻² empirically, 0.297 from Eq. 15b. That is a low-gain sunlight response, not a high-gain CO₂ forcing response inferred from a smaller ΔF.
04 · Misread leftover
Surface σT⁴ changes more than TOA longwave. TEFAC = 0.477: only 47.7% of the surface flux change arrives at TOA as OLR. The rest is adiabatic attenuation with height. Calling that remainder “radiative forcing” or “trapped heat” is the category error.
Fig. 9
Observed 0.27 K/decade · Modeled 0.38 K/decade · R² = 0.84
Equation 16 with measured TSI and a CERES-style albedo path versus the GISTEMP–NOAA average through July 2026.
Textbook warming story
Twenty-first-century warming cannot be reconstructed without greenhouse-gas radiative forcing. CO₂ is the necessary cause; solar terms are a small residual.
Why they say this opposes it
The black curve is only albedo and TSI — no ΔF_CO₂, no optical depth, no EEI. It still tracks the observed series. If a greenhouse term were required, the solar-only model would miss the trend. It does not.
Universal response of global surface air temperature to simultaneous changes in TSI (Δs) and Bond albedo (Δα), differentiated from the planetary RATE law. No greenhouse-gas term is present.
ΔTt = Tb [(1 + Δs/Sb)^0.25 × (1 − Δα/(1 − αb))^0.25 − 1]
Independent check
Recent global temperature surge intensified by record-low planetary albedo. Science 387, 68–73 (5 Dec 2024). A Science paper from AWI and ECMWF, using CERES and ERA5, finds the same thing Nikolov & Zeller put at the centre of Equation 16: planetary albedo fell to a record in 2023, mostly from low clouds, and that extra absorbed sunlight is what closed a ~0.2 K hole in the 2023 temperature budget. The overlap is the albedo. The fork is what that albedo is allowed to mean.
doi:10.1126/science.adq7280 · arXiv:2405.19986 · climatology 2001–2022
2023 GMST vs PI
1.48 K
their figure
Unexplained gap
~0.2 K
after GHG + El Niño
CERES ASR 2023
+1.82 W m⁻²
vs 2001–2022
This atlas ASR 2023
+2.52 W m⁻²
same climatology
Their Fig. 2 — GSAT and absorbed sunlight
Anomalies vs 2001–2022. Grey band is the 2023–24 El Niño. Red marks 2023.
Albedo and EEI on their climatology
2023 is the lowest albedo year in this atlas (2023), matching their CERES minimum. EEI rises with ASR — the same leftover they call imbalance.
Their Fig. 1b — 2023 temperature contributions
Dark bars are Goessling et al.’s published 2023 increments. Light bars are this atlas’s Eq. 16 split for 2023 versus the same 2001–2022 mean, where we have a term.
Annual albedo, 2001–2026
Both papers: 2023 is the floor. This atlas then shows a partial rebound in 2024–26; Goessling’s published window ends with the 2023 record.
Nikolov & Zeller (Geomatics, 2024) and Goessling, Rackow & Jung (Science, 2024) are not the same paper. They share a CERES-era observation: Earth’s Bond albedo dropped, the drop is clouds not the solar constant, and 2023 is the record. That is an independent verification of the albedo change. They split on diagnosis.
| Observable | Nikolov & Zeller 2024 | Goessling et al. 2024 |
|---|---|---|
| Planetary albedo in 2023 | Record low in this series (2023; α ≈ 0.284, -0.69 pp vs 2001–2022) | Record low since at least 1940. Mean albedo ≈ 0.29. CERES 2023 is the satellite minimum. |
| Absorbed shortwave | +2.52 W m⁻² in 2023 vs 2001–2022 (this atlas, 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 lowered the albedo | Cloud and surface Bond albedo — ENSO pulses plus a secular decline. TSI is a ripple. | Reduced low-cloud cover in the northern mid-latitudes and tropics (2023: -1.5% ; 2013–2022: -1.27%/decade). Polar ice is only ~12% of the ASR warming. |
| TSI / solar cycle | Near-zero secular TSI; solar-only term 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 is the leftover after TEFAC attenuates surface longwave. It tracks GSAT worse than ASR. | CERES EEI 2023 +0.97 W m⁻² vs 2001–2022, driven mainly by the ASR rise — same energy they book as albedo. |
| What the albedo is allowed to mean | Eq. 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 (+0.07 K) are counted. They still sit inside a GHG + feedback + aerosol + variability frame, and ask whether the low-cloud trend is variability, aerosol cleanup, or an emerging cloud feedback. |
Shared result
Two groups, two journals, two methods (RATE-law derivative vs CERES/ERA5 energy budget), one albedo fact: 2023 absorbed more sunlight because the planet was darker, mainly from fewer low clouds in the northern mid-latitudes and tropics. Polar ice is a sideshow. TSI is a sideshow. That is the verification.
The fork
Goessling et al. treat the albedo drop as the term that intensifies a greenhouse-plus-El-Niño budget — a 0.2 K patch on top of anthropogenic warming. Nikolov & Zeller treat the same drop as the budget: Equation 16 does not need the greenhouse term. Same CERES darkening; opposite place in the causal order.
Download the series plotted above (this atlas, rebased to their 2001–2022 mean) and the 2023 budget numbers transcribed from the paper. Native CERES EBAF and ERA5 files stay at NASA and Copernicus — those portals are linked, not mirrored.
Retrieval notes
Time series here are this atlas rebased to Goessling’s 2001–2022 mean so the years line up; they are not a re-release of the CERES EBAF or ERA5 files used in Science. Contribution bars mix their published 2023 increments with this atlas’s Eq. 16 split for the terms we actually model. 2023 ASR in this reconstruction is +2.52 W m⁻² versus their CERES +1.82 W m⁻² — same sign and order, different product and smoother.
Nikolov & Zeller commentary
Ned Nikolov & Karl F. Zeller, 29 Mar 2025. Zenodo 10.5281/zenodo.15107068, CC-BY-4.0. They do not dispute the CERES albedo drop. They dispute the arithmetic that turned it into a 21% side-show on a greenhouse budget. Below: their claims, then what the Science paper would look like if Goessling et al. adopted them.
G2024, they write, “attributed the 2023 record heat anomaly primarily to anthropogenic forcing (~78.6%) while ascribing only ~21.4% … to natural drivers.” CERES, on a matching 21st-century baseline, “show that the 2023 global heat anomaly was entirely caused by a decrease of Earth’s albedo mostly due to a reduction of low-level clouds over the oceans.”
G2024 split of 1.48 K vs 1850–1900
Solar 0.25 K + El Niño 0.07 K on a pre-industrial stack. That is the 16.9% / 78.6% split they contest.
NZ correction: Dec 2020 – Dec 2023 only
Same CERES window as G2024’s solar experiment. Observed +0.43 K; Eq. 16 +0.44 K. No leftover for a greenhouse residual in that window.
The satellite result would get stronger. The attribution and the Paris/ECS close would not survive. That is the opposite of a minor erratum: it keeps the albedo and rewrites what the albedo is allowed to mean.
01 Baseline
G2024
G2024 mixed CERES ASR/ISR anomalies vs 2001–2022 with GMST vs 1850–1900, then divided 0.25 K of solar by 1.48 K of pre-industrial warming (16.9%).
NZ fix
Use 21st-century anomalies for both radiation and temperature. A solar forcing measured after 2000 cannot be asked to explain a rise that started in 1850.
If adopted — reverses attribution
The 1.16 K “residual anthropogenic” term vanishes. Over Dec 2020–Dec 2023 the 13-month means move 0.43 K observed vs 0.44 K from Eq. 16. Solar forcing accounts for the surge, not 17% of a 150-year stack.
02 Counterfactual model
G2024
The two-layer energy-balance model already contains greenhouse forcing and CMIP5-tuned feedbacks. Zeroing ASR after Dec 2020 then calling the leftover “not solar” smuggles the conclusion into the tool.
NZ fix
Replace the EBM with a solar-only response (Eq. 16): TSI and Bond albedo, no CO₂ term. Compare that curve to independent surface series.
If adopted — reverses attribution
The paper would test solar forcing instead of assuming it is a residual. NZ’s CERES-era fit is R² 0.89 over 2000–2024 and 0.93 over Dec 2020–Jul 2024, with a 6-month lag of temperature behind albedo.
03 EEI as heat trap
G2024
G2024 treats the CERES Earth energy imbalance as greenhouse heat trapping and a driver of the surge.
NZ fix
Read EEI as the leftover after surface flux is adiabatically attenuated with height (TEFAC), a by-product of ASR-driven warming, not its cause.
If adopted — withdraws speculation
The sentence that EEI “brings us closer to Paris targets” would be withdrawn. The satellite EEI time series can stay; its causal role cannot.
04 Why the clouds fell
G2024
Only three mechanisms are considered: internal variability, aerosol cleanup, and a greenhouse-forced low-cloud feedback. Cosmic-ray / solar-wind cloud papers (2013–2023) are uncited.
NZ fix
The decline is directional over decades (against pure variability). Temperature lags albedo by 5–6 months through the record (against aerosol cleanup and a temperature-led cloud feedback).
If adopted — strengthens observation
The mechanism paragraph would stop treating greenhouse warming as the default parent of the cloud trend. The observational claim — low clouds, oceans, northern mid-latitudes and tropics — is unchanged and actually cleaner.
05 Climate sensitivity
G2024
G2024 speculates that if the albedo crash is feedback plus aerosol, ECS sits at the high end of current ranges.
NZ fix
If the albedo drop is externally paced, the ECS construct itself is the wrong sensitivity. The relevant number is λ to absorbed sunlight (~0.30 K per W m⁻²), not λ to CO₂ ΔF.
If adopted — withdraws speculation
The closing “hotter faster / Paris closer” paragraph would not survive. The paper would report a low-gain solar sensitivity instead of an inflated greenhouse one.
06 Window
G2024
The analysis is 2021–2023. The CERES record is 24 years.
NZ fix
Run ASR vs GMST over the whole CERES era, as NZ2024 did.
If adopted — strengthens observation
2023 becomes the latest pulse of a multidecadal albedo decline, not a one-off intensification of anthropogenic warming. That is a stronger paper, not a weaker one — the satellite result gets a longer spine.
Keeps
Record-low 2023 Bond albedo. Low-cloud decline over oceans, northern mid-latitudes and tropics. Polar ice as a sideshow. TSI as a ripple. CERES as the instrument of record.
Reverses
“~78.6% anthropogenic.” The 1.16 K residual. The two-layer EBM counterfactual. EEI as a greenhouse heat trap. 2023 as a one-off intensification of a 150-year CO₂ trend.
Improves
Title becomes causal, not “intensified by.” The full CERES era is the sample. Sensitivity is λ to absorbed sunlight. Cloud mechanism has to face the 5–6 month lag and the uncited solar-wind / GCR literature.
Numbers in the left-hand bars are transcribed from the commentary’s reading of G2024 (1.48 K vs 1850–1900; 0.22 K albedo, 0.027 K ISR, 0.07 K El Niño, 1.16 K residual). Right-hand bars are their CERES-era result (0.43 K observed, 0.44 K Eq. 16, 6-month lag, four-dataset surface average). This atlas’s own reconstruction is close but not identical; the commentary is the source for this section. CC-BY-4.0.