The estimation of «global» sea level change: a problem of uniqueness