Effects of cloud vertical structure on the development of tropical cyclones: A case study based on In-Fa (2021)

Abstract

This study conducted WRF simulations of Typhoon In-Fa (2021), which caused significant damage to the eastern China in 2021, to investigate how cloud vertical structure may affect the development of a tropical cyclone (TC). Specifically, the TC was simulated using two cloud-fraction schemes: the default Xu-Randall (XR) scheme and a newly-developed neural Network-based Scale-Adaptive (NSA) scheme. Results show that the NSA scheme simulates a more eastward TC track than the XR scheme for both the pre-landfall and landfall phases and is closer to the observation. The underlying mechanisms differ between the two phases and are closely associated with the TC asymmetric structure and phase evolution. First, the XR scheme simulates larger cloud fractions than the NSA scheme across the entire TC, yielding a stronger longwave cloud radiative effect (LWCRE). This tends to increase the instability and invigorates the convection. Second, the relatively strong convections in the northeast quadrant of the TC cause a horizontally-distributed cloud layer, where the NSA scheme simulates a less-tilted cloud structure and a more pronounced horizontal gradient of LWCRE, which can amplify the secondary circulation.

During the pre-landfall phase, the effects of increased instability and enhanced secondary circulation in the northeast quadrant tend to counterbalance each other, making the instability effect in the southwest quadrant dominate changes of the TC track. This increases the cloud ice water path (IWP) and latent heating in the west of the TC in the XR simulation, shifting the TC track westward. During the landfall phase, cloud fraction decreases throughout the TC, diminishing the instability effect and allowing the secondary circulation effect in the northeast quadrant to become dominant. This increases the IWP and latent heating in the east of the TC in the NSA simulation, shifting the TC track eastward.