I'm excited to announce a new peer-reviewed publication - my 5th as a lead author! This paper, entitled "Tropical Cyclone Wind Shear-Relative Asymmetry in Reanalyses", is out as an Early Online Release preprint in the AMS Journal of Climate, with co-authors (and excellent mentors) Anthony Didlake, Jr. and Colin Zarzycki.
The paper stems from the first half of my postdoc work at Penn State. We evaluated asymmetric tropical cyclone structure in 2 global reanalysis datasets: ECMWF's ERA5 and NOAA's CFSv2. Specifically, we focus on the effects of vertical wind shear, which has long been shown to relate to asymmetries, dynamically and thermodynamically, throughout the storm environment. I like to use the following GIF of Hurricane Sally (2020) to demonstrate asymmetry. Look at how most of the rain is well east of the center of circulation, even though we often think of hurricanes as perfectly circular/symmetric "donuts":
~6-hour animation of radar reflectivity in Hurricane Sally (2020), viewed from the Eglin Air Force Base radar site in Florida. NOTE: The radar site is northeast of the center, so I acknowledge there's a bias here based on how far radars can observe. But trust me, it's asymmetric :)
(Shoutout to former Florida State labmate Mark Nissenbaum for his great radar webpage!)
This study is motivated by the fact that, since they produce decades of global data, reanalyses and climate models are limited in their spatial resolution, and struggle to capture aspects of hurricane structure and behavior. In the last 10-15 years, we have learned plenty about how the physics within these models affects the axisymmetric (average around a circle) structure, so considering asymmetries next provides a useful new perspective to keep pushing model performance and reliability forward. After all, asymmetries are very important to the fine-scale details of hazards like wind, rain, and storm surge, and for general hurricane track and intensity forecasting!
Our results were more encouraging than I initially expected! Within the limitations of datasets this coarse, we found physically-reasonable asymmetries across the board from circulation, to rainfall, to the direction a vortex gets tilted, to lower- and mid-level humidity. We validated many of these findings by comparing directly to Hurricane Hunter radar observations, using the TC-RADAR dataset from (new Miami professor!) Michael Fischer and colleagues at NOAA's Hurricane Research Division.
Basic examples of tropical cyclone asymmetry in the ERA5 reanalysis. This GIF demonstrates how a composite is put together, and shows where low-level cyclonic and inward flow are strongest, along with where rain is heaviest. Related to Figures 2, 3, and 6 in the paper.
In many ways, we tried to keep it simple, showing the average structure across a collection of storm snapshots sorted by intensity and wind shear. But we also dove into the weeds a bit! For example, we looked at how different physics schemes produce rainfall in models like these, found signs of a realistic growth/decay cycle of deep convection around the storm, and used a rather large equation to examine what processes cause intense convection in certain sectors of the storm.
Examples of how rain is produced by different physical parameterizations ("convective" vs. "large-scale cloud") in the two reanalyses, based on Figure 8 of the paper. The rain rates are shaded, while black contours show where convergence of low-level air is taking place, and red contours show areas where CAPE (instability) is highest.
This serves as the "proof of concept" for ongoing and upcoming work, where we'll take these new tools looking at hurricane structure, expand upon them, and apply them to a larger set of climate models. With this, we can address unexplored, in-depth questions like "will hurricanes interact differently with wind shear in the future?" Check out the paper at https://journals.ametsoc.org/view/journals/clim/aop/JCLI-D-23-0628.1/JCLI-D-23-0628.1.xml for more background, results, and visualizations, and look out for the final, cleaned-up version this Fall!
Basic examples of rainfall and low-level wind asymmetry in a high-resolution climate model.
The work was funded by Penn State's College of Earth and Mineral Sciences, under the Dean's Fund for Postdoc-Facilitated Innovation, as well as the U.S. Department of Energy through the HyperFACETS program.
Comments