Tropical cyclone intensification

Earlier studies indicated that environmental factors alone (e.g., sea surface temperature) cannot differentiate the tropical cyclones (TCs) with different intensification rates. Given this, we made efforts to extract additional environmental information, i.e., synoptic flow patterns, to improve the rapid intensification (RI) forecasts. Six RI flow patterns were identified based on all TCs formed in the South China Sea between 1981 and 2011. Most RI cases happened in fall, when vertical wind shear (VWS) is low, and thereby agrees with previous understanding. One exception occurs in the summer monsoon season, when VWS is the strongest climatologically (see SU1 flow pattern in the left figure). 

The SU1 pattern drew our attention, and we analyzed one representative case, Typhoon Vicente (2012), in the following several publications.

Chen, Wang, and Zhao, MWR, 2015

Strong vertical wind shear can tilt the tropical cyclone (TC) vortex and induce dry air intrusion into core circulations, which are typically not ideal for TC rapid intensification (RI). This study revealed the "secrets" of weak TCs to resist vertical wind shear by generating a more aligned reformed inner vortex (or new center). Two-stage alignment of the reformed inner vortex were found (see left figure). Vertical alignment associated with the reformed vortex alleviates the dry air intrusion above the boundary layer and leads to the formation of a nearly saturated core at RI onset.  These findings have inspired many observational and modeling studies focusing on similar phenomenon afterward.

Chen, Wang, Fang, and Xue, JAS, 2018; Chen, Zhang, and Marks, GRL, 2019

Accurately predicting the timing of rapid intensification (RI) is crucial for hazard management. From the inner-core perspective, RI is typically preceded by the formation of a vertically-aligned, nearly-saturated, and compact core. We aimed to identify universally effective RI indicators under different sea surface temperatures (SST) by analyzing numerical simulations of Typhoon Mujigae (2015), which underwent RI over anomalously warm SSTs. We found that the vortex alignment, albeit necessary, is not an effective RI indicator under different SSTs, while a more immediate cause of RI is the formation of a strong/compact inner core with high precipitation symmetry. 

Chen, Xue, and Fang, JAS, 2018

A "comma-like" convective precipitation shield (CPS) wrapping around the tropical cyclone (TC) center to form a eyewall is a classic indicator of rapid intensification under vertical wind shear. This study reveals that the eyewall formation is through continuous inward rebuilding of convective rainbands at the leading edge of the CPS.  While boundary layer flushing of downdraft-cooled parcels was perceived to hinder TC intensification, we found that these downdraft-cooled parcels, surprisingly, contribute to the development of upshear deep convection after an efficient boundary layer recovery over warm sea surface temperatures. The inward-rebuilding pathway also illuminates why deep convection is preferentially located inside the radius of maximum wind of sheared TCs undergoing RI.

Chen, Gu, Zhang, Marks, Rogers, and Cione, JAS, 2021

Ensemble numerical simulations with perturbations added to the initial TC vortex structure and intensity can provide valuable insights into the sensitivity of TC forecasts under marginally favorable environments. We examined two subsets of ensemble members of Hurricane Gonzalo (2014), referred to as early-RI and late-RI members, which display significant differences in RI onset timing under the influence of a nearby upper-tropospheric trough and an associated dry-air intrusion. 

• In early-RI members, the initial vortex is stronger and more vertically-aligned. The TC vortex is able to fend off the dry-air intrusion relatively unscathed.

• In late-RI members, mid-upper-tropospheric dry-air intrusion depletes moist static energy from the TC core, while mesoscale subsidence beneath the dry-air intrusion acted to dry a deep layer of the troposphere. The inner-core convection is disrupted and the core becomes more tilted, which further delay the RI onset timing. These findings advance our understanding toward the role of mid-upper ventilation in TC intensity change: it could be more detrimental than we thought!!

Fischer, Reasor, Tang,  Corbosiero, Torn, and Chen, MWR, 2023

A consensus machine learning (CML) model for tropical cyclone (TC) intensity-change forecasting, especially for rapid intensification (RI), was developed. This CML model was built upon selected classical machine learning models, with 21 or 34 RI-related predictors extracted from the 2018 version of HWRF. Key takeaways include:

1) Compared to the traditional key predictors from Statistical Hurricane Intensity Prediction Scheme (SHIPS) model, this study includes the inner-core predictors derived from the HWRF and highlights the importance of inner-core relative humidity in predicting RI (INRH in the left figure). 

2) We found that the CML model has satisfactory performance on RI predictions compared to the operational models. CML reached 56% POD and 46% false alarm ratio (FAR), while the operational models had only 10%–30% POD but 50%–60% FAR (see left figure). 

Ko, Chen, Kubat, and Gopalakrishnan, WAF, 2023