Reconstructing balloon trajectories in the tropical stratosphere with a hybrid model using analysed fields

Abstract

AbstractA hybrid Lagrangian‐Eulerian model of particle motion is employed to reproduce the observed trajectories of three constant level balloon flights along the 0.0987 kg m−3 and 0.110 kg m−3 isopycnic surfaces during the EQUATEUR experiment, conducted by Centre Nationale d'Etudes Spatiales in the summer of 1998. The hybrid model employs velocity, temperature and geopotential fields as provided by the European Centre for Medium‐Range Weather Forecasts analysed data for several three week periods of the flights. The balloon in the model is advected primarily by the ambient air velocity to which a correctional velocity, calculated on the basis of a Lagrangian dynamical model, is added. The latter is driven by the Montgomery stream function gradient, interpolated to the relevant isopycnic surface from the geopotential and temperature standard fields, and Coriolis and drag forces. This correctional velocity is free from the continuity constraint satisfied by the air flow. Altogether, the model employs two physical parameters: the Rayleigh friction coefficient and the fraction (weight) of correctional velocity in the advection. To quantify the difference between the calculated and observed trajectories we compute the temporal mean of the separation between the two trajectories. On time‐scales on the order of three weeks the best calculated trajectories are obtained when the correctional velocity fraction in the advection is between 5 and 25%, and these best fitting trajectories have a mean separation that is down to 20% of that of the advection by the ambient air velocity, only. The implication of our results is that, although a balloon is advected primarily (75 to 95%) by the ambient air velocity, the small deviations from this velocity result in a significantly more accurate calculated trajectory over periods of three weeks. When a trajectory remains in one hemisphere for the entire three‐week period, the best fitting fraction of the correctional velocity is smaller and the Rayleigh friction coefficient is larger than in the case when the trajectory straddles the equator, but the improvement factor does not differ significantly.