Differences in wind field characteristics between mesoscale and large-eddy simulations of idealized tropical cyclones

Future offshore wind energy deployment is planned in regions prone to tropical cyclones, like the Gulf of Mexico and the southern U.S. states (Musial et al., 2022). The recurrence period for tropical cyclones in these regions may be shorter than the lifetime of wind farms (Hallowell et al., 2018), calling for a more thorough investigation of the hazards associated with tropical cyclones for wind turbines. High-temporal and spatial resolution measurements of the extreme winds in tropical cyclones at turbine heights (between 50 m and 300 m) are extremely limited due to a sparse observational network (Archer et al., 2016). Numerical simulations can provide virtual estimates of the wind conditions in severe storms to improve the understanding of the underlying risk for offshore structures. Mesoscale and large-eddy simulations (LES) show extreme winds can occur in the lowest 200 m (e.g., Li et al., 2021; Stern et al., 2021), increasing the risk for offshore structures. However, the simulated wind fields in the tropical cyclone boundary layer are sensitive to the model setup (e.g., Xu et al., 2021), adding uncertainty to the risk that severe storms pose to wind turbines.

Here, we examine differences in the wind field structure of mesoscale and large-eddy simulations of several tropical cyclones of varying intensity and size to assess the value of the considerable computational expense of LES as compared to mesoscale simulations. We evaluate differences in the mean flow characteristics between the mesoscale and LES domains at turbine heights. We also contrast large-scale roll structures that develop in the mesoscale and LES domains that can affect wind turbines.

Our idealized nested mesoscale-microscale simulations of five tropical cyclones of varying intensity and size use the Weather and Research Forecasting model (WRF). The outer mesoscale domains (d01-d03) use horizontal grid resolution, Δx, of 13.5 km, 4.5 km, and 1.5 km. The inner LES domains (d04 and d05) use horizontal resolution of 166.67 m and 55 m. We modify storm intensity and size by varying surface forcing. We simulate five distinct tropical cyclones by increasing surface temperature between 26°C and 34°C as in Ren et al. (2020). A complete description of the simulation setup and spin-up of the mesoscale and LES domains can be found in Sanchez Gomez et al. (2023).

The size and intensity of each storm varies with surface forcing and grid resolution. Further, the vertical structure of mean (60-min average) tangential and radial winds varies between the mesoscale and LES domains for all tropical cyclones (Figure 1). In general, the LES domains produce larger vertical shear of the horizontal wind speed below 400 m compared to the mesoscale domains. Furthermore, the LES domains generate strong radial inflow close to the surface, which is often not captured by the mesoscale domains (Figure 1 bottom). Large vertical shear of the tangential and radial winds increases wind veer in the tropical cyclone boundary layer, which may increase loads in wind turbines (Robertson et al., 2019).

Just as mean flow characteristics vary with domain configuration, large-scale structures also exhibit differences between the LES and mesoscale domains (Figure 2). The LES domains produce roll structures in the vicinity of the radius of maximum winds that are not present in the mesoscale domain (Figure 2). The dominant horizontal wavelength of these roll vortices remains nearly unchanged between domains d04 and d05 (λ≅ 3 km). However, the rolls extend deeper into the boundary layer for d04 compared to d05 (Figure 2). These roll vortical structures result in local changes in the radial and tangential velocity components, which are also not captured by the mesoscale domains.

In summary, wind field characteristics differ between LES and mesoscale simulations of idealized tropical cyclones. The vertical structure of mean horizontal winds differs between LES and mesoscale simulations. These differences may be attributed to the presence of roll vortices near the eyewall, which are not captured by mesoscale simulations. Validation of LES and mesoscale simulations is vital towards minimizing uncertainty in current numerical simulations of tropical cyclones.

References

Archer, C. L., Colle, B. A., Veron, D. L., Veron, F., & Sienkiewicz, M. J. (2016). On the predominance of unstable atmospheric conditions in the marine boundary layer offshore of the U.S. northeastern coast. Journal of Geophysical Research: Atmospheres, 121(15), 8869–8885. https://doi.org/10.1002/2016JD024896

Hallowell, S. T., Myers, A. T., Arwade, S. R., Pang, W., Rawal, P., Hines, E. M., Hajjar, J. F., Qiao, C., Valamanesh, V., Wei, K., Carswell, W., & Fontana, C. M. (2018). Hurricane risk assessment of offshore wind turbines. Renewable Energy, 125, 234–249. https://doi.org/10.1016/j.renene.2018.02.090

Li, X., Pu, Z., & Gao, Z. (2021). Effects of Roll Vortices on the Evolution of Hurricane Harvey During Landfall. Journal of the Atmospheric Sciences. https://doi.org/10.1175/JAS-D-20-0270.1

Musial, W., Spitsen, P., Duffy, P., Beiter, P., Marquis, M., Hammond, R., & Shields, M. (2022). Offshore Wind Market Report: 2022 Edition.

Ren, H., Dudhia, J., & Li, H. (2020). Large‐Eddy Simulation of Idealized Hurricanes at Different Sea Surface Temperatures. Journal of Advances in Modeling Earth Systems, 12(9). https://doi.org/10.1029/2020MS002057

Robertson, A. N., Shaler, K., Sethuraman, L., & Jonkman, J. (2019). Sensitivity analysis of the effect of wind characteristics and turbine properties on wind turbine loads. Wind Energy Science, 4(3), 479–513. https://doi.org/10.5194/wes-4-479-2019

Sanchez Gomez, M., Lundquist, J. K., Deskos, G., Arwade, S. R., Myers, A. T., & Hajjar, J. F. (2023). Wind fields in Category 1‐3 tropical cyclones are not fully represented in wind turbine design standards. Journal of Geophysical Research: Atmospheres. https://doi.org/10.1029/2023JD039233

Stern, D. P., Bryan, G. H., Lee, C. Y., & Doyle, J. D. (2021). Estimating the Risk of Extreme Wind Gusts in Tropical Cyclones Using Idealized Large-Eddy Simulations and a Statistical–Dynamical Model. Monthly Weather Review, 149(12), 4183–4204. https://doi.org/10.1175/MWR-D-21-0059.1

Xu, H., Wang, H., & Duan, Y. (2021). An Investigation of the Impact of Different Turbulence Schemes on the Tropical Cyclone Boundary Layer at Turbulent Gray‐Zone Resolution. Journal of Geophysical Research: Atmospheres, 126(21). https://doi.org/10.1029/2021JD035327