Nicola Scafetta writes at  phys.org (site not known for skeptical thinking) Rethinking climate change: Natural variability, solar forcing, model uncertainties, and policy implications.  Exceprts in italics with my bolds and added images.

Current global climate models (GCMs) support with high confidence the view that rising greenhouse gases and other anthropogenic forcings account for nearly all observed global surface warming—slightly above 1 °C—since the pre-industrial period (1850–1900). This is the conclusion presented in the IPCC’s Sixth Assessment Report (AR6) published in 2021.

Figure 3: CMIP6 GCM ensemble mean simulations spanning from 1850 to 2100, employing historical effective radiative forcing functions from 1850 to 2014 (see Figure 1C) and the forcing functions based on the SSP scenarios 1-2.6, 2-4.5, 3-7.0, and 5-8.5. Curve colors are scaled according to the equilibrium climate sensitivity (ECS) of the models. The right panels depict the risks and impacts of climate change in relation to various global Reasons for Concern (RFCs) (IPCC, 2023). (Adapted from Scafetta, 2024).

Moreover, the GCM projections for the 21st century, produced under different socioeconomic pathways (SSPs), underpin estimates of future climate impacts and guide net-zero mitigation strategies worldwide.

The prevailing interpretation is that only net-zero climate policies can keep future climate change-related damages within acceptable limits. Yet such policies carry extremely high economic and societal costs, making it essential to assess whether these certain and immediate costs are fully justified by the current state of climate science.

On the other hand, a closer examination of observational datasets, paleoclimate evidence, and model performance reveals a more intricate picture—one that merits open discussion among students, researchers, and anyone interested in how climate science is evolving.

My study “Detection, attribution, and modeling of climate change: Key open issues,” published in Gondwana Research, examines several unresolved questions in climate detection, attribution, and modeling. These issues concern the foundations of how past climate changes are interpreted and how future ones are projected, and they matter because climate projections influence decisions that will shape economies and societies for decades. [My synopsis: Scafetta: Climate Models Have Issues. ]

A central theme is natural climate variability. Across the Holocene—the last 11,700 years—the climate system exhibited a Climate Optimum (6,000–8,000 years ago) and repeated oscillations: multidecadal cycles, centennial fluctuations, and millennial-scale reorganizations.

Some longer cycles are well known, such as the quasi-millennial Eddy cycle, associated with the Medieval and Roman warm periods, and the 2,000–2,500-year Hallstatt–Bray cycle. These patterns appear in ice cores, marine sediments, tree rings, historical documents, and in both climate and solar proxy records.

Current GCMs, however, struggle to reproduce the Holocene Optimum and these rhythms. They generate internal variability, but not with the correct timing, amplitude, or persistence. When a model cannot capture the natural “heartbeat” of the climate system, distinguishing human-driven warming from background variability becomes challenging. This is particularly relevant for interpreting the warming observed since 1850–1900, because both the Eddy and Hallstatt–Bray cycles have been in rising phases since roughly the 1600s.

A portion of the post-industrial warming could therefore stem from these long natural oscillations, which are expected to peak in the 21st century and in the second half of the third millennium, respectively.

Another key issue concerns the global surface temperature datasets that serve as the backbone of global warming assessment. These datasets are essential but not perfect.

Urbanization, land-use changes, station relocations, and instrumentation shifts can introduce non-climatic biases. Many corrections exist, yet uncertainties persist. Even small unresolved biases can influence long-term trends.

The study highlights well-known discrepancies: satellite-based estimates of lower-troposphere temperatures since 1980 show about 20–30% less warming than surface-based records, particularly over Northern Hemisphere land areas.

Recent reconstructions based on confirmed rural stations also show significantly weaker secular warming. These differences underscore the need for continued scrutiny of observational records.

Solar and astronomical influences represent another area where science is still evolving. The sun varies in ways not fully captured by the simplified irradiance reconstructions used in many models. Multiple lines of evidence indicate that the climate system responds not only to total solar irradiance but also to spectral variations, magnetic modulation, and indirect effects on atmospheric circulation.

These mechanisms are still under investigation, and their representation in models remains incomplete, even though empirical evidence suggests that they may play a dominant role—potentially more influential than the simple total-solar-irradiance forcing currently implemented.

Moreover, despite ongoing controversy surrounding long-term solar variability, current GCMs are typically forced with solar reconstructions that exhibit extremely low secular variability. This helps explain why these models attribute nearly 0 °C of the observed post 1850–1900 warming to solar changes and simultaneously fail to reproduce the millennial-scale oscillations evident in paleoclimate records.

Direct comparisons between GCM global surface temperature simulations and observations also show that the models do not reproduce the quasi-60-year climatic oscillation associated with the 1940s warming period, and they tend to overestimate the warming observed since 1980. This “hot model” problem has been documented in several studies and appears to affect a substantial fraction of current GCMs.

All of this connects to a key parameter in climate science: equilibrium climate sensitivity (ECS). The canonical estimate—around 3 °C for a doubling of CO₂, with a likely range of 2.5–4.0 °C according to the IPCC—derives largely from model-based assessments.

Empirical studies, including those that account more explicitly for natural variability, often suggest lower values, sometimes around 2.2 ± 0.5 °C, or even as low as 1.1 ± 0.4 °C if long-term solar luminosity varies significantly and if additional solar-related mechanisms influence the climate system—mechanisms not included in current models. If ECS is lower than commonly assumed, projected 21st-century warming would be substantially reduced under all SSP scenarios.

The interplay between natural and anthropogenic factors is definitely more nuanced than often portrayed. When empirical models that include natural oscillations are used to project future temperatures, the result is typically moderate future warming rather than extreme trajectories. This raises important questions about the scientific basis for the most aggressive mitigation pathways.

The figure compares the warming expected from GCMs, as assessed by the IPCC, with the associated relative risks, alongside the expectations derived from the empirical modeling proposed in the paper. While net-zero pathways such as SSP1 are considered necessary to meet the Paris Agreement target of limiting global warming to below 2 °C by 2100, empirical considerations suggest that the same target could also be achieved under the far more moderate SSP2 scenario.

This distinction has major global economic implications, because the
prevailing climate-crisis narrative does not appear to be fully supported
by the evidence, and far less costly adaptation strategies could be
more appropriate than highly aggressive mitigation policies.

The study stresses the importance of addressing the key open questions of climate science. Climate policy should be informed by the full spectrum of scientific evidence, including uncertainties and alternative interpretations.