TL;DR
This study systematically analyzes how effectively different climate forcings (like CO2, methane, aerosols, and solar radiation) change global temperatures across multiple climate models. The findings show that "Effective Radiative Forcing" (ERF) is a much better predictor of temperature change than older methods, and that while efficacies are generally close to one for ERF, there's significant model-to-model variation, especially for aerosols. This variation could introduce a small but notable uncertainty in historical climate sensitivity estimates.
What Problem Does This Paper Address?
Understanding how different factors influence Earth's temperature is crucial for predicting future climate change and its sensitivity to greenhouse gas emissions. A core concept in climate science is "radiative forcing," which measures the initial energy imbalance caused by a perturbation (like increased CO2). However, the climate system doesn't just respond to the initial forcing; it also undergoes "rapid adjustments" (changes in clouds, water vapor, etc.) and slower "feedbacks" (like ice-albedo changes).
Historically, "instantaneous radiative forcing" (IRF) was used, but it doesn't account for these rapid adjustments. This led to the idea of "forcing efficacy"—the notion that different forcings might have different impacts on global temperature even for the same radiative forcing. For example, some studies suggested aerosols might be more effective at cooling than CO2 is at warming. This paper addresses the question of how consistent these efficacies are across different climate models and forcing agents, and whether the choice of forcing definition significantly impacts our understanding of historical climate change and climate sensitivity.
Methods
The researchers utilized the Precipitation Driver and Response Model Intercomparison Project (PDRMIP) dataset, which consists of idealized single-forcing experiments from 11 global climate models. They calculated "efficacies" using several different definitions of radiative forcing, including:
- Instantaneous Radiative Forcing (IRF): The initial energy imbalance at the top of the atmosphere or tropopause.
- Stratospheric Temperature Adjusted Radiative Forcing (RFstrat): IRF minus rapid adjustments except for stratospheric temperature changes.
- Fixed SST Effective Radiative Forcing (ERFsst): The change in radiative flux with fixed sea surface temperatures, accounting for rapid adjustments in the atmosphere and land.
- Adjusted Effective Radiative Forcing (ERFssta): ERFsst adjusted for land surface temperature changes using radiative kernels.
- Regression Effective Radiative Forcing (ERFreg): Derived from a linear regression of the energy balance against temperature change.
Efficacy was then computed as the ratio of the temperature response to a given forcing to the temperature response to a CO2 forcing, normalized by their respective forcings. They also employed a simple impulse response model to simulate historical temperature changes, comparing scenarios with and without accounting for differing forcing efficacies.
Datasets
The study primarily uses data from the Precipitation Driver and Response Model Intercomparison Project (PDRMIP). This project provides output from 11 global climate models for various idealized single-forcing experiments. The specific forcings include:
- Doubling CO2 concentration (2xCO2)
- Tripling methane concentration (3xCH4)
- Five times sulfate aerosol concentrations/emissions (5xSO4)
- Ten times black carbon aerosol concentrations/emissions (10xBC)
- A two-percent increase in insolation (2%SOL)
Additional regional aerosol forcing experiments were also analyzed. The historical forcing data used in the impulse response model is an updated version from Myhre et al. (2013).
Key Insights and Ideas
- ERF is King: Effective Radiative Forcing (ERF), which accounts for rapid atmospheric adjustments, is a significantly better predictor of global temperature change than older methods like instantaneous radiative forcing (IRF). Efficacies calculated using ERF are much closer to unity across different forcings.
- Aerosols are Tricky: While ERF generally brings efficacies closer to one, there's still a notable intermodel spread, particularly for aerosol forcings (sulfate and black carbon). This suggests that how different models represent aerosol processes and their interactions with clouds is a major source of uncertainty.
- Location Matters (Maybe): The study investigated whether the geographic distribution of sulfate aerosol forcing affects its efficacy. While some models showed regional differences, the overall multimodel analysis found no robust evidence that the location of sulfate aerosol forcing significantly alters its efficacy.
- Slow Start for Aerosols: Aerosol forcings (which are generally cooling) tend to cause a slower initial temperature response compared to CO2 forcing, though they eventually converge.
- Historical Impact is Subtle: When using a simple impulse response model to reconstruct historical temperature trends, accounting for different forcing efficacies had a surprisingly small impact on the overall historical GSAT trend compared to assuming all forcings behave like CO2.
Key Results
- Multimodel mean efficacies based on ERF are generally close to one for methane, sulfate, black carbon, and insolation.
- Intermodel spread in efficacies is significant, especially for aerosols. For example, the efficacy of 10xBC (black carbon) ranged from 0.36 to 1.33 across models.
- ERFssta and ERFsst (ERF with fixed SST, adjusted for land temperature) were the best predictors of climate response, showing efficacies closest to unity.
- Solar forcing (2%SOL) consistently showed efficacies less than one across models (0.77–0.95).
- Methane efficacy was variable, with multimodel means close to one but a wide range (0.56–1.15).
- Historical GSAT trend estimates were only slightly affected by accounting for differing forcing efficacies, with a potential bias of ±0.4 K in equilibrium climate sensitivity to CO2 doubling due to intermodel spread.
Limitations and Future Work
- Large Intermodel Spread: The significant intermodel spread, particularly for aerosol forcings, highlights the need for further research into the physical reasons behind these differences. This spread limits the robustness of conclusions drawn from single-model studies.
- Large Perturbations: The study used very large perturbations (e.g., 5xSO4, 10xBC). While the temperature response is assumed to scale linearly, future work could explore whether these findings hold for more realistic, smaller forcing changes.
- Model Differences: The study acknowledges that models differ in their treatments of aerosol microphysics, cloud adjustments, and whether they use emissions or prescribed concentrations for aerosols. These differences contribute to the model spread.
- Volcanic Forcing: The impulse response model had to approximate volcanic forcing using the 2%SOL response curve, as PDRMIP did not include volcanic experiments. This is a simplification.
- Future Work: The authors suggest future studies should focus on understanding the physical differences driving the model spread in efficacies to determine robust variations versus signal-to-noise issues. Long model integrations would also be beneficial for improving the signal-to-noise ratio.
Summary updated