Numerical investigation of an oceanic resonant regime induced by hurricane winds

Abstract

The oceanic mixed layer (OML) response to an idealized hurricane with different propagation speeds is investigated using a two-layer reduced gravity ocean model. First, the model performances are examined with respect to available observations relative to Hurricane Frances (2004). Then, 11 idealized simulations are performed with a Holland (Mon Weather Rev 108(8):1212–1218, 1980) symmetric wind profile as surface forcing with storm propagation speeds ranging from 2 to 12 m s⁻¹. By varying this parameter, the phasing between atmospheric and oceanic scales is modified. Consequently, it leads to different momentum exchanges between the hurricane and the OML and to various oceanic responses. The present study determines how OML momentum and heat budgets depend on this parameter. The kinetic energy flux due to surface wind stress is found to strongly depend on the propagation speed and on the cross-track distance from the hurricane center. A resonant regime between surface winds and near-inertial currents is clearly identified. This regime maximizes locally the energy flux into the OML. For fast-moving hurricanes (>6 m s⁻¹), the ratio of kinetic energy converted into turbulence depends only on the wind stress energy input. For slow-moving hurricanes (<6 m s⁻¹), the upwelling induced by current divergence enhances this conversion by shallowing the OML depth. Regarding the thermodynamic response, two regimes are identified with respect to the propagation speed. For slow-moving hurricanes, the upwelling combined with a sharp temperature gradient at the OML base formed in the leading part of the storm maximizes the oceanic heat loss. For fast propagation speeds, the resonance mechanism sets up the cold wake on the right side of the hurricane track. These results suggest that the propagation speed is a parameter as important as the surface wind speed to accurately describe the oceanic response to a moving hurricane.

Appendix

1.1 Impact of the wind field asymmetry on the OML response

An additional simulation has been conducted to quantify the effect of an asymmetric component in the surface wind field on the oceanic response. This asymmetry depends directly on the hurricane propagation speed. As a consequence, we investigate the case of the fastest moving storm (U _H = 12 m s⁻¹) to check the range of validity of our results.

The surface wind field is multiplied by (1 + c cos θ), where θ is the horizontal azimuth (increasing counterclockwise from east) and c = U _H/2V(r), where V(r) is the surface wind speed at a radial distance r from the storm center. For U _H = 12 m s⁻¹ and V _max = 50 m s⁻¹, we obtain c = 0.12 which is comparable with the asymmetry factors used in previous studies (Chang and Anthes 1978; Price 1981; Greatbatch 1983). For distances greater than RMW, c increases because of the decrease of V(r), and the asymmetric component becomes more important on each side of the storm track. In Fig. 16, this new asymmetric wind field is compared to the symmetric one.

Fig. 16

Symmetric and asymmetric surface wind fields (m s⁻¹) with 10-m s⁻¹ isocontours

When applying the asymmetric factor, the wind speed reaches 56 m s⁻¹ on the right side of the hurricane and decreases to 44 m s⁻¹ on the left side of the hurricane. The extensions of the 20 and 10-m s⁻¹ isotachs are multiplied by a factor 2 on the right-hand side of the hurricane.

The impact of this asymmetric wind field on the SST pattern is presented in Fig. 17 in terms of SST difference between the asymmetric and the symmetric case. On the right-hand side of the storm, the wind asymmetry intensifies the maximum cooling by 0.4°C between 200 and 350 km (i.e., 3.3 and 5.8 RMW). This means that the perturbation induced by the asymmetric wind field is mostly located outside the studied region. Between 0 and +3 RMW, the temperature difference is equal or less than 0.3°C, which is small compared to the cooling induced by the symmetric wind field (2.2°C). As a consequence, the maximum cooling location remains at the same place in both cases. This difference is twice as small as the cooling difference observed between the simulations with U _H = 10 and 12 m s⁻¹. On the left side of the storm, the cooling decreases by 0.2°C between 100 and 200 km (1.6 and 3.3 RMW). This difference is small compared to the 1.6°C cooling observed in this region with the symmetric wind field. The central region temperature (between −2 and +2 RMW) difference induced by the asymmetric wind field is equal or less than 0.1°C. Finally, the SST response extends slightly more on the right side of the storm than in the symmetric case, but the cooling differences are small compared to that induced by varying the propagation speed.

Fig. 17

SST difference between the symmetric and the asymmetric cases (°C) for U _H = 12 m s⁻¹

The impact of the asymmetric wind field on the MKE pattern is presented in Fig. 18 in terms of MKE difference between the asymmetric and the symmetric case. The maximum MKE difference, which is less than 0.15 m² s⁻², is located between 150 and 300 km (i.e., 3.3 and 5.8 RMW) on the right-hand side of the storm. This difference is small compared to the maximum MKE produced by the symmetric wind field between 50 and 100 km (i.e., 0.8 and 1.7 RMW) which is about 0.65 m² s⁻². The left-hand side of the hurricane is not affected in terms of MKE by the asymmetric wind field because of the very small amount of MKE in this region. The MKE distribution in the central region (−100 km < x < 100 km) is also not modified by the asymmetric wind field (<0.05 m² s⁻²). Globally, we observe a slightly more biased MKE distribution toward the right of track, but it does not modify the characteristics of the MKE pattern observed in the symmetric case.

Fig. 18

MKE difference between the symmetric and the asymmetric cases (m² s^-2) for U _H = 12 m s⁻¹

These results are in good agreement with the previous findings of Chang and Anthes (1978), Price (1981), and Greatbatch (1983) who showed that the wind asymmetric component has a small impact on the OML response. Moreover, we found that the impact of the asymmetric wind field decreases when the storm translation speed decreases. Furthermore, when decreasing U _H, the cooling induced by the resonance mechanism and nonlinear dynamics strongly increases, while the cooling induced by the wind field asymmetry decreases. Hence, we can conclude that the asymmetric wind field has a limited impact on the OML response and can be ignored without compromising the validity of our results.