An Experimental Study of Gravity-Inertial Waves and Wind-Induced Kelvin-Type Upwellings in a Rotating System

Abstract

Large-amplitude gravity-inertial waves have been observed in stratified seas. These waves, of a shorter period than the Coriolis period, are linked to a gust of wind. Numerous theories have been proposed to explain these waves but, up to now, they have not been verified experimentally in a rotating system. On the large rotating platform in Grenoble, the author has built a large tank (8 m ? 2 m ? 0.6 m) equipped with a wind tunnel. By strictly controlling the main parameters (speed of rotation of the platform, and thus Coriolis period, intensity and duration of the impulsional wind, height and density of the two fluid layers) he has studied and analyzed the variations in height of the interface between the two fluids. Here, the author's work is placed in the general context of various studies of this phenomenon, and some analytical developments are presented within the general hypothesis first used by Crépon (1969a,b). The experimental facilities are described briefly and a preliminary explanation is given of the phenomena occurring in the tank when the wind suddenly blows. The variations in height of the interface are analyzed and the existence of a gravity-inertial wave linked to the impulsional character of the wind is proved. The period of this wave, which is shorter than the Coriolis period, depends on the dimensions of the tank and on the phase speed of the baroclinic mode, and may be predicted by the simple model designed. The wave is progressive, first appearing at the sides of the channel. At the beginning of the motion, it can be shown that the wave amplitude is independent of the observation point, and depends on the wind intensity and duration. Wave amplitude is found to be maximum when wind duration is equal to half the period predicted by the tank model. Looking at the entire interface, there are two opposite corners where Kelvin-type interface variations occur, propagating along the longitudinal sides of the tank and giving rise to a Poincaré-Kelvin amphidromy. The existence of such particular points has been predicted by recent analytical developments of Crépon and Richez (1981).