85 
GEOSAT MEAN SEA SURFACE 
ERM 1 TO 12 
60 
30 
0 
-30 
-60 
-110 -50 0 50 90 
Figure 5 - Global Mean Sea Surface as computed from GEOSAT observations collected over the first 
six months of the Exact Repeat Mission (ERM 1 to 12: November 1986 to May, 1987). 
al., 1988; Marsh et al., 1989). The orbit error of the GEM- 
T1-derived GEOSAT ephemerides is estimated to be at the 
1-m level, while the respective accuracy of the GEM-T2 
improved ephemerides is estimated to be at the 50 cm level 
or better (Haines et al., 1990). 
The algorithm portion of the operational system currently 
being developed at CCS includes: 
• the development of software to make an assessment of 
the accuracies of the latest computed orbits of GEOSAT, 
and to study the stability of sea surface solutions through 
comparisons with the latest available high degree and order 
geoidal solutions (e.g. OSU89); and 
• the design of optimum methodologies for the 
simultaneous crossover and collinear arc solutions, which 
are necessary in order to remove various unmodelled 
significant errors (e.g. orbit, ocean dynamic state errors 
etc.) inherent in the altimetry observations. 
Along with precise orbit information, an enhanced 
processing of radar altimeter data allows the generation of 
sea surface heights on a global or regular scale with high 
spatial and temporal resolution. Figure 5 illustrates a 
typical solution for the global mean sea surface using 
roughly six months of data from the first twelve GEOSAT 
17-day ERMs (November 8, 1986 to May 30, 1987). 
Subsequent analyses with oceanographic and geodetic data 
provides the additional capability to derive improved 
models of the marine geoid and to estimate large scale 
features of the sea surface topography. 
The primary goal of the current studies is to provide a 
homogeneous intermediate-wavelength (100-500 km) and 
long-wavelength (>500 km) gravity database over the 
global oceans, particularly for the support of: • 
• mesoscale oceanographic tasks (e.g. detection of current 
boundaries and mesoscale eddies); 
• generating improved maps of geoid height, gravity 
anomalies, and deflections of the vertical; and 
• the definition of specific geophysical/geologic areas for 
more detailed follow-up gravity surveys or resource 
exploration studies. 
Figure 6 illustrates typical world maps of altimetry-derived 
Significant Wave Heights (SWH) computed over two 
adjacent 5-day periods during ERM 1. They clearly portray 
the early November rough seas in both the middle North 
Atlantic and the Southern seas. The maps also reveal that 
most of the oceans experience SWHs less than 4 metres. In 
addition, like many other 5-day average maps we have 
computed, they clearly slow that the sea state is 
predominantly latitudinally structured and that is has long 
wavelength nature. Christou (1990) in an extensive 
treatment of GEOSAT data to study the space-time ocean 
variability and its effects on the length of day has 
demonstrated even more clearly the capability of satellite 
altimetry to depict related space/time ocean surface current 
variability. Figure 7 is just a "snap-shot" of zonally 
averaged sea-level-slope time series within a longer 
(2-year long) record of the mean surface ocean current 
velocity fields (ibid). Each time series represents the zonal 
average of current velocity fluctuations with respect to the 
2-year long mean (Nov. '86 to Nov. '89) for every 1° 
latitude band between 65° S and 65° N. The most obvious 
observation that can be drawn from Figure 7 is that most of 
the zonal velocity variability is confined within the north 
and south tropical zones (i.e., between 23° S and 23° N), 
while significant variability is observed in the northern 
mid-latitudes (i.e., between 30° N and 50° N) and southern 
higher latitudes (i.e., between 50° S and 60° S). The 
northern mid-latitude zonal velocity fluctuations are likely 
related to the large mid-latitude oceanic gyres (i.e. Gulf