The drift of an Argo float is determined by three factors: the ocean current at its parking depth, the surface current during the 6–12h period while the float transmits its data, and the current profile during ascent and descent. The relative contributions of each factor to the overall drift are difficult to determine. Ichikawa et al. (2001) estimate that the drift at the parking depth is overestimated by 10–25% as a result of effects of current shear during ascent and descent and uncertainties from the determination of the exact location where the float surfaces (which may happen some time before a satellite passes and picks up the signal). A very basic estimate of the surface and parking depth contributions to the overall drift can be obtained by assuming that surface currents are an order of magnitude larger than currents at 1000– 2000m depth. This corresponds roughly to the ratio of the time when the float drifts at its parking depth to the time of measurement and data transmission, suggesting that surface and parking depth drift contribute about equally to the overall drift.