The results obtained in Ajaccio are very homogeneous,
even more when the ocean dynamics correction is applied.
Indeed, the dispersion of the mean bias estimates decreases
from 0.3 cm to 0.1 cm when the correction is used. In both
cases, it is exactly in the order of the standard error of the
bias estimates, which means that the biases are coherent
wherever they are estimated around the calibration site.
Nevertheless, the estimates are more than 1 cm higher than
the values in Senetosa, when no ocean dynamics correction
is used. It must be noticed that, due to the gap of one year
and a half in the Ajaccio data (from May 2008 to September
2009), which corresponds to the first 43 cycles of the
Jason-2 mission, the bias estimates at this calibration site
are computed with nearly two times less data than in Senetosa.
The analysis of the bias time series at the various
crossover points, at both calibration sites, shows that the
temporal variations of the bias estimates are very coherent
from one site to the other, particularly in terms of trend
(not shown here). The bias of 1 cm between the two sites
is clearly visible and is consequently not due to the fact that
less data are used for the computation in Ajaccio. It is more
probably due to the fact that the tide gauge position is
rather sheltered in the Ajaccio harbour and the instrument
measures specific and local ocean dynamics.