At short time scales, Argo data have been used to study the evolution of near-surface temperature and salinity beneath tropical cyclones. The data show clear temperature differences left and right of the cyclone track, but produce conflicting patterns of salinity change. Monsoons and ENSO events dominate the low-latitude seasonal/inter-annual ocean-atmosphere variability. Argo data, when combined with TAO/Triton tropical buoy array data, extend the mapping of tropical Pacific Ocean structures, and are also used in ENSO forecast systems. Argo profiles have also revealed the Arabian Sea space-time response during the summer monsoons of 2002 and 2003.

Many results focused on exploration of the circulation and the definition of the properties and abundance of winter-formed mode waters in mid-latitude ocean basins. This could even be done in areas such as the Okhotsk Sea, where there is extensive ice cover in some years. Twelve operational analysis/forecast centers routinely use Argo data, and through GODAE, are routinely producing ocean state products. Improvements in ocean predictions from assimilating Argo data were demonstrated at the workshop. These give an exciting foretaste of the likely impact of the full Argo array when combined with remote sensing data.

Although the Argo array is not yet complete, its impact on global-scale problems can already be seen from studies such as Willis et al. [2003], where Argo allows heat and fresh water storage in the ocean to be estimated. Its advantage over ship-based observations is the uniform geographical and seasonal distribution, depth penetration (deeper than the typical 750 m of XBTs), and data volume (Figure 3).

Argo data have been shown to be of high enough quality to document changes over almost 20 years in subsurface salinity across the south Indian Ocean, an area where climate change models predict that anthropogenic change will be most easily detected [Banks and Wood, 2002]. However, applications of this type highlight the need for caution in adjusting float data to climatological values.

While much emphasis is placed on Argo profile data, velocity estimates both at depth and from the floats' time at the surface have been used for the global estimation of inertial oscillation statistics to reveal hitherto unsuspected subsurface circulation patterns. When combined with surface drifter data and T/S profiles, they can be used to derive velocity field estimates on basin (and ultimately global) scales throughout much of the water column.