Substantial Cloud Brightening from Shipping in Subtropical Stratocumulus Clouds

The influence of aerosol particles on cloud reflectivity over the oceans remains one of the largest sources of uncertainty in our understanding of anthropogenic climate change. Commercial shipping constitutes a large and relatively concentrated aerosol perturbation in a meteorological regime where clouds have a disproportionally-large effect on climate. Yet, to date, studies have been unable to detect climatologically-relevant cloud radiative effects from shipping on regional or global scales, despite models indicating a sizable negative forcing.

We attribute a significant increase in cloud reflectivity to enhanced cloud droplet number concentrations within a major shipping corridor in the southeast Atlantic. Prevailing winds constrain emissions around the corridor, which cuts through a climatically-important region of expansive low-cloud cover. We use universal kriging, a classical geostatistical method for spatial interpolation, to estimate counterfactual fields of cloud properties in the absence of shipping.

Enhancements of sulfate mass, cloud droplet number concentration, and cloud albedo and reductions in effective radius in the stratocumulus cloud deck are all statistically significant and consistent with a dominant Twomey effect. Increasing cloud brightness is partially compensated by decreasing liquid water path in the afternoon. There are no consistent changes found in cloud amount.

Extrapolation of these results globally leads to a radiative forcing estimate of -0.06 W/m², which is anticipated to diminish as new fuel-sulfur regulations come into force in 2020. This global radiative forcing estimate is two orders of magnitude larger than previous observationally-derived values but toward the lower end of estimates from global climate models. We argue that the cloud changes and resulting effective radiative forcing from shipping in this region can serve as a strong constraint for evaluating both the Twomey effect and cloud adjustments in models.