The study "Radiative Energy Flux Variation from 2001–2020" published in MDPI's Atmosphere claims that declining outgoing shortwave radiation (due to cloud changes) is the primary driver of Earth's warming from 2001–2020 — not the greenhouse effect. This conclusion rests on a fundamental confusion of cause and effect. The observed shortwave trends in the CERES data are fully consistent with mainstream climate science: greenhouse gas forcing in the longwave range, combined with climate feedbacks, produces precisely the observed pattern — large SW changes and virtually no net LW change at the top of atmosphere. The study was criticized by leading climate scientists, including NASA-GISS Director Gavin Schmidt, as "rather poorly argued and seriously confused," and has not been cited or replicated in any high-ranking journal.
The paper's core claim is that declining outgoing shortwave radiation (SW, +1.42 W/m²) is the "major heating cause," while increasing longwave emission (LW, −1.1 W/m²) largely compensates the classical greenhouse effect. The authors conclude that cloud changes — not greenhouse gases — are the primary cause of warming.
This argument contains a category error that has been identified from multiple directions. In 2022, Gavin Schmidt demonstrated on RealClimate, using a simple two-layer radiative model, that a purely longwave greenhouse gas forcing combined with SW and LW feedbacks produces exactly the observed CERES pattern: Greenhouse gases initially reduce outgoing LW radiation, but the Planck response (higher surface temperature → more LW emission) and the lapse-rate feedback nearly fully compensate this LW effect. Simultaneously, positive SW feedbacks (fewer low clouds, less sea ice, rising water vapor concentrations) cause a decrease in reflected SW radiation. The net result: large SW changes, minimal LW change at TOA — even though the original forcing lies entirely in the longwave range.
Raghuraman et al. (2021, Nature Communications) confirmed this quantitatively using the GFDL-AM4 model: the probability that the observed EEI trend is explainable by internal variability alone is less than 1%. Forcings and feedbacks add up in the SW range but compensate each other in the LW range — the exact opposite of what a naive reading of the CERES data suggests.
The "And Then There's Physics" analysis (2021) also referenced Trenberth & Fasullo (2009) and Donohoe et al. (2014), who showed that energy accumulation in the climate system under rising greenhouse gases manifests itself primarily as increased absorbed solar radiation — a result known for over a decade.
The raw data in the study are not the problem — the CERES EBAF Ed4.1 observations are robust and confirmed by leading scientists such as Norman Loeb (CERES project leader). Loeb et al. (2021, Geophysical Research Letters) likewise found a doubling of the Earth Energy Imbalance from ~0.5 W/m² (2005) to ~1.0 W/m² (2019), driven primarily by an increase in absorbed solar radiation of 0.9 ± 0.3 W/m², partially offset by +0.4 ± 0.25 W/m² more outgoing LW radiation.
The critical issue is interpretation: Loeb attributes the EEI increase to a mixture of anthropogenic forcing and internal variability (PDO phase shift), both acting in the same direction. Dübal and Vahrenholt, however, treat all observed cloud-related radiative changes as independent causes of warming, without distinguishing among three possible sources:
Hodnebrog et al. (2024, Communications Earth & Environment) quantified the aerosol contribution alone at 0.2 ± 0.1 W/m² per decade — approximately 38% of the total EEI trend. This is anthropogenic forcing, not evidence against greenhouse gas theory.
1. The "Cloudy Areas" claim about an undetectable greenhouse effect is a misinterpretation of the Cloud Radiative Effect (CRE = All-Sky minus Clear-Sky). This standard methodology isolates the net effect of clouds on the radiation budget but says nothing about whether greenhouse gases are active in cloudy regions. CO₂ and water vapor absorb and emit radiation regardless of cloud presence. The fact that LW CRE trends are positive (greater LW transmissivity in cloudy areas) does not mean the greenhouse effect is "compensated" there — it primarily reflects changed cloud heights and coverage as feedback.
2. The application of the Beer-Lambert law to the atmosphere is physically inadequate. The Beer-Lambert law describes absorption in a homogeneous medium without emission. The atmosphere, however, has a temperature gradient, and absorbed LW radiation is re-emitted at the local temperature. The correct treatment requires the Schwarzschild equation (absorption and emission). Furthermore, CO₂ absorption is saturated only at the band center near 15 μm, not at the spectral wings — additional CO₂ broadens the effective absorption band, as captured by the logarithmic forcing relationship (Myhre et al. 1998).
3. The claim of a medieval OHC maximum 600 ZJ above the present value likely relies on Rosenthal et al. (2013, Science), but is severely misinterpreted. Rosenthal measured temperature proxies in Pacific intermediate waters (500–1000 m) at a few sites — extrapolation to global ocean heat content is methodologically unjustified. The authors themselves cautioned against such generalizations. Moreover, the modern rate of OHC change is, according to Rosenthal, 15 times faster than any previous rate over the past 10,000 years.
4. The reference to Svensmark's cosmic ray hypothesis invokes a theory largely refuted by the CLOUD experiment at CERN. While CLOUD confirmed that ions can slightly enhance aerosol nucleation, for the atmospherically most relevant nucleation pathway (sulfuric acid + dimethylamine) the ion contribution is negligible (Almeida et al. 2013, Nature). The entire causal chain from nucleation through CCN growth to cloud formation has not been demonstrated (Pierce 2017, JGR). The IPCC AR6 assesses the evidence as insufficient to attribute significant climate forcing to this mechanism.
5. The AMO correlation with "heating phases" is problematic for several reasons. Mann et al. (2021, Science) demonstrated that the AMO as an internal oscillation likely does not exist: the apparent ~60-year signal in the industrial period is an artifact of competing greenhouse gas warming and time-varying sulfate aerosol cooling, produced by linear detrending. In the pre-industrial period, volcanic forcing pulses project onto the 50–70-year frequency. CMIP5 control simulations without external forcing show no robust spectral peak at the AMO period. Using the AMO to explain warming trends is therefore circular — the "AMO signal" is itself largely caused by the anthropogenic forcing that the authors seek to minimize.
6. The enthalpy reconstruction back to 1750 rests on extremely uncertain foundations. Reliable instrumental OHC data exist only from ~1960 onward; before 1870 there is virtually no global SST coverage. Zanna et al. (2019, PNAS) reconstructed OHC only back to 1871 with substantial uncertainties (±91 ZJ). Any reconstruction before 1850 requires massive assumptions about ocean circulation and equilibrium states that are not met — the ocean was still responding to the transition from the Medieval Warm Period to the Little Ice Age.
IPCC AR6 Chapter 7 (Forster et al. 2021) assesses the net cloud feedback as positive with high confidence (+0.42 W/m²/°C), meaning clouds amplify GHG-driven warming. A negative net cloud feedback is rated as "very unlikely." The total anthropogenic effective radiative forcing since 1750 amounts to +2.72 [1.96 to 3.48] W/m², dominated by CO₂ (+2.16 W/m²), partially offset by aerosols (−1.3 W/m²). Solar activity contributes a mere ~0.01 W/m² over the entire industrial period.
Raghuraman et al. (2021) estimate the anthropogenic contribution to the observed EEI trend at 65–135% (66% confidence interval). Notably, a 2025 update extending the Dübal-Vahrenholt methodology through 2024 (Harris, Substack) shows that Clear-Sky LW radiation is now decreasing — the classic GHG signal is becoming dominant and undermines the original paper's narrative.
The journal Atmosphere has an Impact Factor of 2.3 and ranks in the lower third of atmospheric science journals (Q2–Q3) — compared to Nature Climate Change (IF ~12), Geophysical Research Letters (IF ~4.6), or Journal of Climate (IF ~3.9). The submission-to-publication time was only 30 days (submitted September 1, 2021; accepted October 1, 2021; published October 5, 2021).
MDPI as a publisher faces substantial criticism: Finland's Publication Forum downgraded 193 MDPI journals to the lowest quality tier in 2024. The Norwegian Scientific Index created a dedicated "Level X" category. Mass resignations of editors at Nutrients (2018) and Vaccines (2021) followed allegations that MDPI pressured acceptance of low-quality manuscripts. Approximately 88% of all MDPI articles appear in Special Issues, the number of which exploded from 6,756 (2020) to 39,587 (2021).
The paper's own open peer review reports confirm the quality issues: Reviewer 1 called the central claim (clouds rather than GHGs as the cause) "seems weak to me" with only a single sentence of justification. Reviewer 2 noted the topic was "extensively investigated by several authors, especially the team that generates the dataset" and questioned genuinely new findings. Reviewer 3 flagged poor figure quality and language problems. The paper would very likely have failed rigorous peer review at a top journal (Nature, JGR, Journal of Climate).
Fritz Vahrenholt (b. 1949) holds a doctorate in chemistry, is an honorary professor of chemistry at the University of Hamburg, former SPD Environment Senator in Hamburg, and former CEO of Deutsche Shell, REpower Systems, and RWE Innogy. He co-authored Die kalte Sonne (The Cold Sun, 2012), which promotes the Sun as the primary driver of climate change and predicted a cooling of 0.2–0.3°C by 2035 — a prediction already falsified. He is a trustee of the Global Warming Policy Foundation (GWPF), a signatory of the CLINTEL declaration "There is no climate emergency," and a regular speaker at EIKE events. He had no peer-reviewed publications in climate science prior to collaborating with Dübal.
Hans-Rolf Dübal (b. 1955) likewise holds a doctorate in physical chemistry, with research stints at ETH Zurich and the University of Wisconsin, followed by over 30 years in the chemical industry. He published in 2020 the book Klima: Auf dem Wege zu meiner eigenen Meinung (Climate: Toward My Own Opinion) through self-publishing (Books on Demand), which discusses Svensmark and solar variability. No publication history in climate science prior to 2021.
Both authors have published exclusively in low-ranking journals and disseminate their results primarily through climate-skeptic blogs (NoTricksZone, WUWT, Judith Curry's Climate Etc., Die kalte Sonne).
The study by Dübal and Vahrenholt describes real CERES observations that are confirmed by mainstream research — particularly the dominance of SW trends in the EEI. However, the interpretation of these data is fundamentally flawed. The central error consists of treating observed cloud changes as independent causes of warming, rather than recognizing them for what they largely are: feedbacks and responses to greenhouse gas forcing and anthropogenic aerosol reductions. This error was already explained in the foundational literature (Trenberth & Fasullo 2009) and refuted by detailed attribution studies (Raghuraman et al. 2021). The paper's additional claims — Beer-Lambert saturation argument, medieval OHC maximum, AMO heating phases, Svensmark hypothesis — are all problematic and contradict the current state of research. The combination of the authors' lack of domain expertise, publication in a journal of contested quality, and dissemination through climate-skeptic networks presents a consistent picture: a paper that reinterprets existing data to fit a preconceived narrative, rather than interpreting them according to scientific standards.