We have developed a new technique for extracting global mesoscale variability from satellite altimeter profiles having large radial orbit error (~3 m). Long-wavelength radial orbit error, as well as other long-wavelength errors (e.g., tides, ionospheric-atmospheric delay, and electromagnetic bias), are suppressed by taking the derivative (slope) of each altimeter profile. A low-pass filter is used to suppress the short-wavelength altimeter noise (λ 3 km) adjacent to continental shelves, spreading ridges, and oceanic plateaus. Variability is low in shallower areas (>3 km). Along the ACC, the meso-scale variability appears to be organized by the many shallow areas in its path. We do not see convincing evidence that variability is higher downstream from topographic protrusions. Instead, the areas of highest variability occur in the deep basins (> 4 km). ¿ American Geophysical Union 1989