For each mission, the altimeter sea surface heights were
computed at the high frequency rate (20 Hz for the Jason-1
and Jason-2 missions, 18 Hz for the Envisat mission). No
specific editing was done to the data during the computation
of the SSH. Nevertheless, a filtering process was necessary to
reduce the instrumental noise in the high frequency measurements
at the crossover points where the bias was estimated.
To that purpose, for each cycle, the altimeter SSH was computed
along the track over a few tens of kilometres around
the considered crossover point. A sea level anomaly (SLA)
was estimated at each of the points and for each cycle, as
the difference between the SSH and the mean sea surface profile
of the same mission. An editing was then performed on
these anomalies, with a maximum value threshold of
50 cm, which is representative of the SLA in this part of
the Mediterranean Sea. The data beyond the median value
of the SLA at a given point with a 50 cm tolerance margin
were also rejected. Finally, the SLA that were three times larger
than the standard deviation of the SLA at a given point
were also eliminated. The last step consisted in averaging the
SLA over the portion of the track around the crossover
point, for each cycle. Then, these smoothed SLA were added
to the mean sea surface at the crossover point, which provided
the smoothed altimeter SSH time series to be used
for the comparison with the in situ data.