Idealized WRF-ROMS Coupled Model Simulation Study: Effect of Atmosphere-Ocean Coupling on Tropical Cyclone Intensity


Student thesis: Doctoral Thesis

View graph of relations


Related Research Unit(s)


Awarding Institution
  • Chung Leung Johnny CHAN (Supervisor)
Award date8 Mar 2017


Atmospheric and oceanic conditions are both important for tropical cyclone (TC) intensity change. However, the air-sea interaction processes and the underlying mechanisms are not completely understood. This study therefore aims to better understand the physical processes of air-sea interaction and their impact on TC intensity and structure. Two components of a coupled Ocean-Atmosphere-Wave-Sediment Transport modeling system (COAWST), the atmosphere model Weather Research and Forecasting (WRF) and the ocean model Regional Ocean Model System (ROMS), are adopted in this study.

In Part I, idealized simulations are conducted to investigate the relationships between TC intensity changes and ocean mixed layer depth (MLD) as well as translation speed. The coupled WRF-ROMS model can well simulate TC intensity changes and sea surface temperature (SST) cooling induced by the TCs under different ocean MLD and background flow conditions. Surface heat flux is reduced due to the SST decrease, which subsequently inhibits TC development. A thinner MLD and a slower speed produce a more symmetric cold wake under the TC inner core, which can effectively weaken the TC. The SST has a non-linear decrease with increasing translation speed under the same MLD condition. A heat budget analysis supports that upwelling plays a more important role in changing the temperature within the upper-ocean when the TC moves slowly (e.g. 1 m s-1). However, upwelling only dominates the subsurface temperature changes when the translation speed increases so that a smaller SST cooling occurs. A translation speed of 3 m s-1 is found to be the optimal speed for TC development when the MLD is thick enough (50 m and 100 m in this study). A larger basic flow (> 3 m s-1) produces a stronger asymmetry in the vertical velocity and results in a weaker secondary circulation, and hence inhibits TC intensification. For a shallow MLD of 20 m, a 5 m s-1 translation speed can mitigate more cooling effect and becomes more favorable for TC maintenance. This result suggests the competing effects of atmospheric and oceanic conditions in determining TC intensity.

The purpose of Part II is to investigate the effects of the vertical wind shear (VWS), the upper ocean condition and their combined impact on TC intensity variation. A series of idealized simulations are conducted with easterly shears of 2, 4, 6, 8 m s-1 and MLDs of 20, 50 and 100 m. In the uncoupled experiment, a larger VWS results in a weaker TC which is similar to the result in previous studies. The stronger relative flow associated with larger VWS causes larger asymmetries in convection within the eyewall, which weakens the secondary circulation of the TC and thus inhibits the intensification.

For the coupled experiments, strong ocean response is induced. Over a 20 m MLD, all TCs weaken dramatically after 12 h. The smaller shear leads to a faster decaying rate of the TC because of weaker secondary circulation resulted from the net effect of SST cooling and VWS. The SST cooling suppresses the upward motion in the eyewall, especially in the rear-right quadrant with respect to the TC track. Therefore, the shear-induced downdraft can destroy the eyewall more easily in the coupled experiment compared to in the uncoupled experiment. The intensity evolution over 50 m MLD is similar to that over 20 m MLD. The negative ocean feedback plays a decisive role in determining the TC intensity when the MLD is shallow.

For the MLD of 100 m, all the TCs can further intensify for 12 or 24 h and then weaken slightly. The VWS is more important in modulating TC intensity when the OHC is 100m. The TC in the 2 m s-1 shear environment is still the weakest one during the decaying period due to the larger SST cooling compared to that in other experiments. It is noted that a 4 m s-1 shear seems be optimal for TC intensification. This is because the larger VWS induces a more asymmetric convection and diabatic heating patterns, which inhibits TC intensification in the 6 and 8 m s-1 shear environment.

Two wind profiles with the same shear magnitude of 8 m s-1 but different vertical wind distribution are used to compare their effect on TC intensity. There is almost no intensity difference in both the uncoupled and coupled experiments with these two shear profiles. The deep convection of the intense initial vortex may minimize the impact of the shallow convection induced by the asymmetric wind at surface. Hence, there is no large intensity difference between these two cases.

In Part III, the ocean response and intensity changes under different initial TC size (in terms of the radius of maximum wind and the radius of gale-force wind (R17)) and ocean conditions are studied. Consistent with previous studies, the initial larger vortex still has a larger size throughout the life cycle in both the uncoupled and coupled experiments within the 72 h simulation. The local maximum SST cooling is similar in each experiment, which is likely due to the similar maximum surface wind (MSW). However, the area-averaged cooling varies depending on different TC size. A larger TC produces a larger area-averaged SST cooling because of the larger winds at the same radius from TC centre. Although a larger area-averaged SST cooling is found in the larger TC case, the intensities show small difference. This is likely because of the extreme large SST cooling induced by the stationary vortex.

It is also found that the SST changes contribute little to TC size (R17) change, which demonstrates that thermodynamic process does not contribute as much on R17 change as the dynamic process proposed (e.g. angular momentum transport) in previous studies.

In summary, the results show that the combined impact of the atmospheric condition and oceanic condition is important in affecting TC structure and intensity. The air-sea interaction process is complex, which cannot be parameterized in a model without coupling. Therefore, it is necessary to utilize the fully coupled model to investigate the TC structure and intensity changes.