As a check on the total system accuracy of the TOPEX/POSEIDON satellit
e, a southern hemisphere verification study was undertaken in Bass Str
ait (41 degrees S) immediately adjacent to the Southern Ocean, An acou
stic tide gauge and two pressure gauges were installed on the southern
side of Bass Strait near where descending pass 88 crosses the Tasmani
an coast, The tide gauge benchmark height was estimated relative to th
e laser tracking station at Orroral Valley and the very long baseline
interferometry station near Hobart using Global Positioning System sur
veying techniques, In the TOPEX/POSEIDON geophysical data records, fla
gs on the altimeter data began to be set 15-20 km from land and little
useful altimeter data were available within about 13 km of the coast,
The acoustic tide gauge heights were corrected for changes in tidal p
hase and amplitude between the coast and the offshore subsatellite poi
nt using results from a numerical model, The altimeter correction for
atmospheric water vapor calculated from the TOPEX microwave radiometer
data and the insitu precipitable water data agreed to 2 cm RMS, with
no significant bias, The RMS variability between the insitu sea level
estimates and the TOPEX/POSEIDON estimates was less than 3.5 cm, As pa
rt of this variability is caused by inaccuracies in the estimates of t
he sea level variations between the coast and the subsatellite point,
the result indicates that the precision of the TOPEX/POSEIDON sea leve
l height measurement system, including the time dependent orbit errors
, the ionospheric corrections and the wet and dry tropospheric path de
lays, is less than 3.5 cm over Bass Strait, considerably below the ori
ginal system specifications of 13 cm. Our estimate of the altimeter bi
as (including geographically correlated orbit error) is -25 +/-10 cm,
consistent with the results at the NASA (Harvest Platform) and Centre
National d'Etudes Spatiales (Lampedusa) calibration sites, However, ac
curate estimation of this total altimeter bias is confounded by the la
ck of altimeter data close to land and in particular, limitations in t
he available local geoid data, The altimeter bias shows no significant
drift at this site,