Glider sampling simulations in high-resolution ocean models

Abstract

Idealized simulations of autonomous underwater glider sampling along sawtooth vertical–horizontal paths are carried out in two high-resolution ocean numerical models to explore the accuracy of isopycnal vertical displacement and geostrophic velocity profile estimates. The effects of glider flight speed, sampling pattern geometry, and measurement noise on velocity profile accuracy are explored to inter-pret recent full-ocean-depth Deepglider observations and provide sampling recommendations for glider missions. The average magnitude of velocity error profiles, defined as the difference between simulated glider-sampled geostrophic velocity profile estimates and model velocity profiles averaged over the spatial and temporal extent of corresponding simulated glider paths, is less than 0.02 m s21 over most of the water column. This accuracy and the accuracy of glider geostrophic shear profile estimates are de-pendent on the ratio of mesoscale eddy to internal wave velocity amplitude. Projection of normal modes onto full-depth vertical profiles of model and simulated glider isopycnal vertical displacement and geostrophic velocity demonstrates that gliders are capable of resolving barotropic and baroclinic structure through at least the eighth baroclinic mode.