Charles Beard


Satellites are used to study various oceanic phenomena, including sea surface temperature, color, and sea height variability (which is related to oceanic currents). In observing sea height variability from radar altimeters, three methods of analysis can be applied. They are the collinear or repeat track method, the cross-over difference method, and the mean sea surface method. All three are designed to remove the effects of geoid undulations from the altimeter records. This paper compares the collinear method with the mean sea surface method in a small geographic area. Both methods analyze the tracks of collinear altimeter data by subtracting a mean pass and a best fit quadratic curve. Geoid variability is effectively removed by subtracting the mean pass, while the quadratic curve removal eliminates remaining long wavelength orbit error and tidal signals. The difference between the two methods is in the mean pass that is subtracted. The collinear mean is computed from a point by point average of the collinear data taken from one month of the SEASAT mission. The mean sea surface method subtracts a mean pass derived from a global mean sea surface developed from cross-over data from the 3.5 year GEOS-3 and the full 3.5 month SEASAT missions. After this processing, the residual data from these methods are compared in the spatial and spectral (wavenumber) domains. In the region of 1 0~2 cycles per kilometer, a spectral analysis will yield energy primarily from oceanic variability and eddy currents. Over all wavenumbers, the two methods compared qualitatively. Quantitatively the mean sea surface residuals had an order of magnitude higher variance than the collinear mean residuals. This was true at all wavenumbers except in the region of 10-2 cycles per kilometer, where the mean sea surface residuals still had more power, but only by a factor of two or three. This is significant because this region is thought to contain the peak amount of power from oceanic signals. With more data it may be possible to completely and accurately define a mean sea surface suitable for the detection of various oceanic phenomena with a single pass of a satellite altimeter. It is concluded from this study that this cannot be done with the presently available mean sea surface.

Library of Congress Subject Headings

Oceanography--Remote sensing; Radio altimeters--Evaluation

Publication Date


Document Type


Department, Program, or Center

Mechanical Engineering (KGCOE)


Sciremammano, Frank

Advisor/Committee Member

Nye, Alan

Advisor/Committee Member

Torok, Joseph


Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: GC10.4.R4 B46 198


RIT – Main Campus