TL;DR
This paper uses the observed hemispheric difference in cloud droplet number concentration between the pristine Southern Ocean and the polluted Northern Hemisphere to constrain the impact of industrial aerosols on Earth's radiative forcing. The findings suggest that current climate models underestimate preindustrial aerosol concentrations and cloud droplet numbers, leading to a tighter constraint on the aerosol-cloud interaction forcing.
What Problem Does This Paper Address?
A major source of uncertainty in predicting future climate change is the "aerosol-cloud interaction" (ACI) effect. Aerosols, tiny particles in the atmosphere, can influence clouds by changing the number of cloud droplets. More droplets can make clouds brighter, reflecting more sunlight back into space, thus cooling the planet. However, quantifying this cooling effect (known as aerosol radiative forcing) is challenging because we don't fully understand how aerosol concentrations and their impact on clouds have changed since the industrial era. Specifically, there's a significant discrepancy between how climate models simulate preindustrial cloud droplet numbers and what we observe today, especially in pristine regions like the Southern Ocean. This paper aims to bridge this gap by using observed hemispheric contrasts to constrain these changes.
Methods
The authors leverage satellite-derived cloud droplet number concentration ($N_d$) data, particularly focusing on the contrast between the Southern Hemisphere (SH) oceans (considered a proxy for preindustrial conditions) and the Northern Hemisphere (NH) oceans (polluted by industrial aerosols). They compare these observations with simulations from a suite of global climate models (GCMs) from the Aerocom project and a perturbed parameter ensemble (PPE) from the HadGEM3-GA7.1 model.
Key methodological aspects include:
- Calculating Hemispheric Contrast: They define the hemispheric contrast in $N_d$ over oceans as $\Delta N_d(\text{NH-SH})$, the difference between the annual, area-weighted mean $N_d$ in the 30°-60°N latitude band and the 30°-60°S latitude band.
- Using $N_d$ as a Proxy: They use the observed $\Delta N_d(\text{NH-SH})$ to infer the change in global mean $N_d$ from preindustrial (PI) to present-day (PD) conditions, denoted as $\Delta N_d(\text{PD-PI})$.
- Constraining Radiative Forcing: They establish a relationship between $\Delta N_d(\text{NH-SH})$ and the aerosol-cloud interaction radiative forcing ($RF_{aci}$) using the PPE, allowing them to constrain $RF_{aci}$ based on satellite observations of $\Delta N_d(\text{NH-SH})$.
- Source-Sink Budget Model: To understand the processes controlling $N_d$ in the Southern Ocean, they employ a simple aerosol source and sink budget model to assess the impact of precipitation scavenging.
Datasets
- MODIS (Moderate Resolution Imaging Spectroradiometer) Data: Used for satellite-derived cloud droplet number concentration ($N_d$) over oceans from 2003-2015. This is the primary observational dataset.
- Aerocom Phase II Models: A collection of GCM simulations used to compare modeled PI and PD $N_d$ with observations.
- HadGEM3-GA7.1 Development Models: Simulations from a specific climate model, including a million-member perturbed parameter ensemble (PPE), used to explore model uncertainty and establish relationships between $N_d$ and $RF_{aci}$.
- In Situ Data: Aircraft measurements and ground-based observations from Antarctic stations (King Sejong, McMurdo) and other campaigns (RITS, ORCAS, OFCAP) were used to validate satellite-derived $N_d$ in the Southern Ocean.
Key Insights and Ideas
- Southern Ocean as a Preindustrial Proxy: The study strongly supports using the pristine Southern Ocean as a proxy for preindustrial cloud conditions, showing that its $N_d$ patterns are consistent with in situ measurements and are influenced by natural processes like biogenic emissions and sea ice retreat.
- GCMs Underestimate Pristine $N_d$: A significant finding is that most GCMs systematically underestimate summertime $N_d$ in the SH, particularly near Antarctica, where observed values can be as high as those found in polluted NH outflows. This suggests models are missing key aerosol production or processing mechanisms in these regions.
- Hemispheric Contrast as a Powerful Constraint: The observed hemispheric difference in $N_d$ provides a robust observational constraint on the anthropogenic change in $N_d$ and, consequently, on the $RF_{aci}$. This approach bypasses some of the challenges associated with directly estimating preindustrial $N_d$.
- Preindustrial Aerosol Concentrations Likely Higher: The large observed hemispheric contrast implies that preindustrial aerosol concentrations were likely higher than estimated by many current GCMs, especially in the SH.
Key Results
- Constrained $\Delta N_d(\text{PD-PI})$: Based on satellite observations and GCM relationships, the change in oceanic $N_d$ from preindustrial to present-day ($\Delta N_d(\text{PD-PI})$) is constrained to be between 8 cm⁻³ and 24 cm⁻³.
- Constrained $RF_{aci}$: The aerosol-cloud interaction radiative forcing ($RF_{aci}$) since 1850 is constrained to be between -1.2 W⋅m⁻² and -0.6 W⋅m⁻². This range is consistent with previous estimates but provides a tighter bound.
- GCMs Overestimate NH $N_d$ and Underestimate SH $N_d$: Models generally overestimate $N_d$ in the polluted NH and underestimate it in the pristine SH midlatitudes and high latitudes.
- Precipitation Sinks in SH: The study confirms that precipitation scavenging is a significant sink for CCN in the SH mid-latitude storm track, reducing $N_d$, but its impact is weaker poleward of Antarctica.
Limitations and Future Work
- Single-Observation Constraint: The authors acknowledge that using a single hemispheric contrast as a constraint might lead to an overly tight range due to structural uncertainties in GCMs. Future work should combine this constraint with other observational datasets and model behaviors.
- Model Representation of Pristine Regions: The significant underestimation of $N_d$ in pristine regions by GCMs highlights a critical need for improved representation of aerosol production and cloud microphysical processes in these environments. This includes better understanding of biogenic emissions (like DMS), new particle formation, and sea spray processes.
- Equifinality in Models: The study points to potential "equifinality" issues in GCMs, where different combinations of parameters can lead to similar outcomes, making it hard to pinpoint specific process errors.
- Future Directions: The authors propose that future GCM evaluations should incorporate the contrast between pristine and polluted regions as a key test of realism, in addition to evaluating predicted $N_d$ in anthropogenically perturbed regions. Further investigation into the specific aerosol parameters driving the observed $N_d$ in the SH is also suggested.
Summary updated