2020

Causes of Higher Climate Sensitivity in CMIP6 Models

M. D. Zelinka, et al.

Observation and Context

Scientists use complex computer programs called global climate models (GCMs) to estimate “Equilibrium Climate Sensitivity” (ECS). ECS is a measure of how much the Earth’s surface temperature will warm if the amount of carbon dioxide (CO₂) in the atmosphere doubles. Historically, climate models estimated that this warming would likely fall between 1.5 K and 4.5 K. However, when the newest generation of models — called CMIP6 — was released, scientists noticed that several of them predicted much higher warming, with some exceeding 4.5 K. This surprising observation prompted researchers to investigate what physical changes in the models were causing this sudden jump in predicted warming.

Hypothesis

The researchers hypothesized that the higher climate sensitivity in CMIP6 models is primarily driven by stronger positive cloud feedbacks. Specifically, they proposed that as the Earth warms, the models predict a greater reduction in low-level cloud cover and a weaker increase in the water content of those clouds. Because clouds act like mirrors that reflect sunlight back into space, having fewer or thinner clouds allows the planet to absorb more solar energy, amplifying global warming.

Experiment and Methodology

To test this hypothesis, the authors compared data from 27 CMIP6 models against 28 older CMIP5 models. They analyzed simulated experiments where atmospheric CO₂ was abruptly quadrupled and held constant.

• Feedback Separation: They used “radiative kernels” — computational tools that helped them isolate and calculate how much individual factors (like temperature, water vapor, surface reflectivity, and clouds) contribute to the planet’s overall warming response.
• Cloud Analysis: They used the “approximate partial radiative perturbation” (APRP) technique to study shortwave (sunlight-reflecting) cloud feedbacks.
• Cloud Separation: They separated low-altitude clouds from high-altitude clouds using statistical techniques to pinpoint exactly where the model changes were occurring.

Results and Data

The research yielded several key pieces of data:

• Higher Warming Estimates: The average predicted warming (ECS) rose from 3.3 K in CMIP5 to 3.9 K in CMIP6, with some new models reaching up to 5.6 K.
• The Culprit is Low Clouds: Non-cloud feedbacks (like water vapor and ice melting) remained virtually unchanged. Instead, the stronger warming was almost entirely due to a more positive shortwave low cloud feedback.
• Southern Ocean Impact: This feedback change was strongest in the extratropics (regions between 30° and 60° latitude, especially over the Southern Ocean). Here, warmer sea surface temperatures caused a sharper decrease in low-cloud cover and a weaker increase in cloud water compared to older models.

Conclusion and Climate Impact

The study concluded that the higher climate sensitivity in CMIP6 is driven by stronger positive low-cloud feedbacks, particularly in colder, mixed-phase cloud regions (which contain both ice and liquid water). The newer models represent these mixed-phase clouds more realistically based on recent observational data, showing that they hold more liquid water than previously thought.

If these high-sensitivity models are correct, it means our planet is more sensitive to greenhouse gases than we hoped. This would require humanity to implement much stricter carbon-reduction targets to avoid crossing dangerous global temperature thresholds.