In addition, even if the offshore estimates are lower than
the absolute bias, the dispersion of the mean biases (0.6 cm)
obtained with the regional method is close to the mean
standard error of the mean regional bias (0.2 cm). It means
that the bias estimates are consistent individually, at the
various crossover points, but also in terms of mean and dispersion.
The differences from one point to the other are
partly due to the differences in the mean sea surface profiles
at the crossover points, but also probably to the land contamination
in the altimeter footprint. It would be interesting
to do the same analysis with a retracking method
dedicated to coastal areas, but no dataset of this kind is
available for the Jason-1 mission at the moment, to our
knowledge.