An Assessment of the Errors in Lagrangian Velocity Estimates Obtained by FGGE Drifters in the Labrador Current

Abstract

Drogued floats have been widely used to develop information about the Lagrangian structure of ocean circulation. However, the degree to which the various designs actually follow water parcels has been the subject of considerable debate in the literature. The design developed for the First GARP (Global Atmospheric Research Program) Global Experiment (FGGE) program (basically, a spar buoy and window-shade drogue with ARGOS navigation) has been in routine use for a decade in the Labrador Current, and a large dataset has been accumulated. The methods and results of a set of experiments conducted to directly measure the relative velocity of the drogue through the water are summarized. Current components relative to the drifter were measured by two electromagnetic current meters, and the local weather and sea state were observed by a drifting meteorological buoy. Two uninstrumented drifters and an undrogued but weighted float were deployed in conjunction with the main drifter to provide an independent estimate of the drifter performance. Three deployments were successfully conducted during periods in which the wind speed varied from 0 to 10 m s?1 with wave heights up to 3.5 m. The authors found no significant differences in the Lagrangian velocities of the FGGE drifters due to the presence of the current meters but significantly greater slip of the meteorological buoy and an undrogued drifter. These observations are consistent with a simple model of drifter performance. The current meter measurements showed relative velocities of up to 0.1 m s?1 with vertical shear across the drogue of order 10?2 s?1 with considerable high-frequency variability. Regression analysis of the data averaged in 20-h bins to remove serial correlation showed significant linear relationships between wind speed and slip in all experiments and a significant dependence on the shear in one. When all the data was combined, a linear model of the slip based on wind alone explained 60% of the observed variance and a best estimate of the slip at 10 m s?1 wind speed of 0.06 m s?1.