Investigating hurricane boundary layer turbulent structure in airborne radar data and the MPAS model

The tropical cyclone (TC) boundary layer is the interface where energy is transferred from the ocean surface to the vortex aloft and plays a critical role in the TC intensification. Therefore, the structure of the boundary layer will highly influence the evolution and predictability of the TC. The TC boundary layer is fundamentally turbulent and driven largely by shear instabilities that create coherent turbulent structures (CTSs). Previous studies have found that CTSs seem to be different from classic roll vortices in the Ekman layer and are involved in energy backscatter to larger scales via wave-wave interactions (Guimond et al., 2018; Li and Pu, 2023; Sroka and Guimond, 2021). Using airborne radar observation data, Guimond et al. (2018) show that CTSs were involved in the eyewall replacement cycle of a strong TC, with radial wavelengths of 1-3 km, and depths of 1.5 km. In addition, these CTSs could be related to the potential vorticity waves generated by the boundary layer shear instabilities. By conducting a kinetic energy budget, Sroka and Guimond (2021) further found that CTSs and their associated sub-filter-scale energy transfer will impact on the TC boundary layer through wave-wave nonlinear interactions, which can transfer energy into the TC mean flow through wave-mean flow interactions. Using the WRF-LES model to examine the CTSs in a real-case hurricane, Li and Pu (2023) found that CTSs are only present near the eyewall region and highly related to inflow convergence, which transfers horizontal turbulent energy to the vertical component.

The role of CTSs in shaping the TC structure and development remains an active area of research. Previous studies mostly analyzed the momentum field, and we expect that these processes couple strongly to the thermodynamic variables. It is known that the thermodynamic evolution will feedback strongly onto the larger scales of the TC vortex. This feedback is a possible route for predictability beyond what one would expect from the local influence of the CTSs.

To investigate this hypothesis we will use nonlinear data assimilation (DA) to examine the nonlinear interaction of the observed CTSs with the vortex thermodynamics. This allows us to investigate the influence of certain complex structures on the evolution of the system, which is difficult using standard model sensitivity experiments. To use DA in a complex, nonlinear system such as a TC, it needs to preserve approximate nonlinear dynamical and thermodynamical balance. Hence, the DA method itself has to be nonlinear. Recently, fully nonlinear DA techniques for high-dimensional geophysical systems have been developed (Pulido and Van Leeuwen.,2019; Hu and Van Leeuwen, 2021) and are now available in the DA framework JEDI.

The objectives of this study, funded by NASA via the GESTAR II program, are to investigate how to use fully nonlinear DA methods to integrate turbulence resolving radar observations into the numerical model, examine the influence of CTSs on the predictability of a hurricane by performing data denial experiments with the nonlinear DA method and different observational data sets, and to analyze the dynamical pathways from the CTSs to the large-scale vortex during intensification. Specifically, we will assimilate the Imaging Wind and Rain Airborne Profiler (IWRAP) data of Hurricane Teddy (2020), which is a downward-pointing, dual-frequency airborne Doppler radar that is ideal for measuring CTSs in the TC boundary layer (Guimond et al. 2014). The high sampling rate of IWRAP allows mapped radar reflectivity and calculation of the three Cartesian wind components over the full 3D scan volume with approximately 125 m horizontal and 30 m vertical grid spacing. Furthermore, we will assimilate high-resolution satellite observations to observe the larger-scale upper structure of the Hurricane.

For the numerical model, we will use the state-of-the-art Model for Prediction Across Scales (MPAS) in a regional set up, with a 10 km resolution at the lateral boundaries, zooming in to 200 m in the inner part of the domain. The model is initialized with ERA5 data and spun up for 12 hours. The high resolution will permit the CTSs we see in observations to develop, and allows for assimilating the high-resolution radar observations. The inner high-resolution area is elongated in the direction of the TC track to maintain high resolution during TC movement. For the DA method, we use the Particle Flow Filter in the JEDI data-assimilation system. This filter is fully nonlinear, appropriate for the highly nonlinear Hurricane physics in the boundary layer, and the physics of convective towers, and its connection to Hurricane intensity. The working of the filter is similar to that of ensemble 3DVars, with that difference that the ensemble members communicate with one another during the minimizations, and the prior does not have to be Gaussian (Hu and Van Leeuwen, 2021). In the presentation we will report on our findings in this ambitious project.