Public Research Seminar by Earth, Ocean and Atmospheric Sciences (EOAS) Thrust, HKUST (GZ) - Global Cloud Radiative Feedback: SST Pattern Formation, Percentage Emergent Constraint, and AI-UAV Observing System

This seminar will present research progress in the following three areas: 
1) The South Pacific surface warming is interactively shaped by shift and intensification of its subtropical anticyclone；
2) Patterns of percentage cloud change effectively constrain cloud feedback；
3) Cloud AI Multi-drone Observing System (CAMOS)

Projected sea surface temperature (SST) warming into the 21st century exhibits robust spatial patterns, e.g., a southeastern minimum (SEM) in the subtropical Pacific. Here, we discover a novel air-sea interaction for its formation, associated with the southwestward shift and intensification of the South Pacific Subtropical Anticyclone (SPSA). They initialize and propagate anomalous southeasterly winds to cool the entire SEM via latent evaporative forcing enhanced by cloud radiative feedback. Triggered by the tropical expansion and relative cooling of the Southern Ocean, this interactive process is dominated by the relative mean cooling in the South Pacific rather than by specific SST patterns.

The SST patterns can perturb clouds, which substantially influence the radiative balance of climate system with complex ongoing changes. This provides the greatest uncertainty when projecting the magnitude of future global warming, yet efforts to provide a tight constraint on cloud radiative feedback have been insufficient. Here, we identify percentage change in cloud fraction driven by sea surface temperature patterns, to be key to height variation of the tropical clouds. This inspires an emergent constraint on climate model projections with recent satellite measurements, effectively reducing the intermodel ranges by 59% for cloud feedback and 33% for surface warming.

Ultimately, projection uncertainty of cloud change and radiative feedback comes from poorly understood convective dynamics due to the lack of 3-D observations of parcel motions. Here, we introduce a cutting-edge cloud observing system, adopting robot-vision technology aboard multiple drones. This can provide perspective, persistent, and portable records of convection dynamical field, endorced by a stereo motion-tracking of cloud parcels. Now the device is under calibration and will be soon applied to scientific research, with potentials to build a global observation network for cloud dynamics aiding rigorous convective parameterization and reliable climate projection.