Documenting the progressions of secondary eyewall formations

Intense tropical cyclones can form secondary eyewalls (SEs) that contract towards the storm center and eventually replace the inner eyewall, a process known as an eyewall replacement cycle (ERC). However, SE formation does not guarantee an eventual ERC, and often, SEs follow differing evolutionary pathways. This study documents SE evolution and progressions observed in numerous tropical cyclones, and results in two new datasets using passive microwave imagery: a global subjectively labeled dataset of SEs and eyes and their uncertainties from 72 storms between 2016–19, and a dataset of 87 SE progressions that highlights the broad convective organization preceding and following a SE formation.

The results show two primary SE pathways exist, No Replacement, known as Path 1, and Replacement, known as the Classic Path. Most interestingly, 53% of the most certain SE formations result in an eyewall replacement. The Classic Path is associated with stronger column average meridional wind, a faster poleward component of storm motion, more intense storms, weaker vertical wind shear, greater relative humidity, a larger storm wind field, and stronger cold air advection.

This study highlights a greater number of potential SE pathways exist than previously thought. The results of this study detail several observational features of SE evolution that raise questions regarding the physical processes driving SE formations. Most importantly, environmental conditions and storm metrics identified here provide guidance for predictors in artificial intelligence applications for future tropical cyclone SE detection algorithms.

Methods
The secondary eyewall (SE) labels and SE Progression datasets are collected using three 89–92 GHz passive microwave imagers (GMI, SSMIS, and AMSR2) with horizontal and vertical (H & V) polarization from the Tropical Cyclone PRecipitation, Infrared, Microwave, and Environmental Dataset (TC PRIMED; https://doi.org/10.25921/dmy1-0595). 

Image Preprocessing

The dual-polarization allows for the creation of a single polarization-corrected temperature image, which reduces the impact of low-emissivity surface features, allowing TC internal features, such as SEs, to appear more distinct. The images are storm-centered and interpolated from the final best-track TC centers. Next, we interpolate the microwave images to a polar grid with a grid spacing of 4 km in radius by 10° in azimuth. Since our dataset includes Northern and Southern Hemisphere storms, we orient the images relative to storm motion, which points to the top of the page.

Storm Selection

For the SE labels dataset, we randomly select seventy-two storms with a lifetime maximum intensity greater than or equal to hurricane intensity (33 m s-1) from the North Atlantic, eastern and western North Pacific, and Southern Hemisphere tropical cyclone basins between 2016–19. For the SE Progressions dataset, we select a subset of our storms from the SE labels dataset (32 of 72) with a storm lifetime maximum SE confidence level of three or greater, where confidence ranges from 1 to 5, with a value of 1 being most uncertain.

Secondary Eyewall and Eye Labeling

We define an eye as a circular area or ring of lower Tb near the storm center and an SE as an outer minimum that surrounds a defined eye with four or more octants (i.e., coverage ≥ 50%). The labels use a confidence-based system, where confidence values again range from 1 (least confident 'yes') to 5 (most confident 'yes') and –1 (least confident 'no') to –5 (most confident 'no').

The eye labels are labeled by only one expert. However, the SE labels are labeled subjectively by four experts. A first expert (Expert 1) independently labels the dataset of seventy-two storms. Next, a second expert (Expert 2) independently labels the SE for the same dataset of microwave imagery. The two SE labeled datasets are then placed into three groups: no secondary eyewall, low confidence SEs (i.e., 1, 2, 3), and high confidence SEs (i.e., 4 and 5). Any SE label discrepancies between Experts 1 and 2 that result in different groups are relabeled by two other independent experts (Experts 3 and 4). The resulting confidence level is then the average between Experts 2, 3, and 4. If Experts 1 and 2 agree that a label should result in a low-confidence 'yes', but disagree on the confidence level, the average value between Experts 1 and 2 is taken. However, if the average confidence level is 1.5, then the confidence from Expert 1 is taken. For discrepancies between Experts 1 and 2 within the high-confidence 'yes' group, the confidence of Expert 1 is taken. For the 'no' group, the confidence level is taken from Expert 1. The purpose of this multi-expert labeling system is to address the subjective nature of labeling SEs and the temporal shortcomings of microwave imagery. The most uncertain cases are denoted using low-confidence values in the SE labels dataset.

Secondary Eyewall Progressions Definitions and Stages

Using the SE labels dataset, we construct a subset of storms with a lifetime maximum SE confidence level of 3 or greater as the basis of our SE Progressions dataset. We define a progression as the evolution of a convective organization from one stage to the next stage. Next, we document the convective appearances before, during, and after a SE formation, which are defined using three stages (Entrance Stage, SE Stage, and Exit Stage). The Entrance Stage is between the time of the first microwave image appearance of an outer convective band and the time of SE formation. The SE stage is when an SE is present in the microwave imagery. Lastly, the Exit Stage is the time of the first microwave image when only one eyewall is clearly visible after the appearance of a SE.