Observations and high-resolution modeling of tropical cyclone formation: Developing and non-developing cases over the west Pacific

Tropical cyclone (TC) formation depends on the interactive processes on scales ranging from convective scale (~1-10 km) to synoptic or larger scales (~1,000–10,000 km). It has been known that TCs form within large-scale disturbances such as easterly waves and monsoon troughs, which provide a favorable environment for large, persistent convective cloud systems (or cloud clusters) to develop. On the other hand, cloud clusters can modify the large-scale environment through convective upscaling by moistening the deep troposphere. Also, convectively-enhanced vorticity on the mesoscale (e.g., 100-500 km) is favorable for TC genesis. These interactive processes are referred to as the “preconditioning” stage. Eventually, in most cases, a dominant large and persistent convective system—a precursor—forms, which leads to a TC formation.

In this study, an objective cloud cluster tracking method is used to identify the TC precursor clusters using hourly satellite IR images from July-October 2003-2009. Contiguous areas of cloud top temperature colder than 208 K are defined as cloud clusters and are tracked in time. Nearly all significant western North Pacific (WPAC) TCs forming south of 26oN during July-October 2003-2009 are found to have cloud clusters lasting > 15 hours prior to TC genesis. These long-lived cloud clusters are associated with persistent mesoscale (100 – 500 km) areas of stratiform precipitation, which are evident from the TRMM TMI data. There are also long-lived cloud clusters that do not develop into TCs, and some of them occur in environments that are apparently conducive to TC formation. The question is what are the distinct properties of these developing and non-developing clusters as well as in their large-scale environment?

To address this question, we use both observational analysis and high-resolution modeling approaches to provide some insights into the physical processes of the scale interactions in TC genesis. First, we will composite the environmental conditions for developing vs. non-developing cases using the global analysis fields, which can quantify the large-scale “forcing” for TCs. The 7-year record of developing and nondeveloping cloud cluster tracks is used to derive a parametric genesis index. The genesis index can produce a skillful genesis forecast, and it is also useful for the selection of long-lived nondeveloping clusters in favorable environments. Second, we use the high-resolution MM5 and WRF simulations with explicit moist physics to better understand the convective upscaling processes. The chosen developing cases are Typhoons Hagupit (2008), Jangmi (2008), and Choiwan (2009). These developing cases are compared with some non-developing cases for which the model produced long-lived cloud clusters but did not produce a TC. Preliminary model simulation results suggest that a major difference between the developing and non-developing cases is a better defined mesoscale, inertially stable region with enhanced vorticity in the former. Convective cores (e.g, "hot towers") are apparently present and with similar intensity in both cases, but in developing cases latent heat is more efficiently “trapped” near the disturbance center leading to a warm-core structure, allowing the surface pressure to fall and surface winds to increase. It appears that formation of large areas of stratiform precipitation is vital for setting up the "stiff" inertially stable region.