This gridded data set is constructed as described by Reverdin et al. (1994) from current data during the period 1987 to April 1992. We merged the drifts of more than 1000 surface buoys with currents from current-meter morings at 5 sites (4 along the equator at 165E, 170W, 140W and 110W and one at 6N, 140W). The drifters were initially drogued by a sea anchor of various designs centered at 15 meters. The proportion of drifters with drogue present is large near the equator (more than 80%), but diminish to roughly 50% near 20N and 20S. We did not correct for drogue loss, and the systematic error we make on the 15m currents can therefore exceed 5 cm/s near 20N and 20S.
The analysis of an average seasonal cycle is presented with 12 fields of u (zonal velocity) and v (meridional velocity) for each mid-month date on a 5x1 degrees latitude*longitude grid.
Longitude ranges from 122.5E to 82.5W (32 grid points) Latitude ranges from 19.5N to 19.5S (40 grid points) Time ranges from mid-January to mid-DecemberNo analysis is made on land, but values are reported for grid boxes partially over land.
Because the analysis is an optimal fit on a set of specified functions applied locally and which involve a substantial zonal smoothing, the currents near the continents should not be expected to be represented properly. We should also warn that the conditions sampled by the current measurements may not be representative of the long-term conditions. This seems in particular to be the case near the date-line along the equator.
The data are here.
Gilles Reverdin email@example.com