Geographical and echo corrections
On the left, the Antarctica's map of the RMS of the surface height difference (dH) at the crossover computed using ENVISAT cycles from 9 to 67 after validation process, on the right the same but with geographical and echo corrections applied.
The orbit of satellite altimeters in repeat mode is constrained to fit in a limited range of across track repeatposition. For example ERS in 35d repeat mode is constrained to be in a +-1km around the nominal track. This mean that if we plot the repeat tracks nadir position we obtain a distribution of measurements in a 2km wide band. This size is similar in size to the first impact of the radar altimeter echo. Thus over ocean most studies consider it exact repeat and are not affected by across track variations in topography which are negligible. Over land surfaces, it is not the case. First if the across track slope is 1m/km (that is a small value over lands and very average over ice sheets) a displacement across track induces a 1m difference in the measurement. If the topography is not a simple slope but includes curvature and undulations at small scale, the effect becomes rapidly anoying for the interpretation of time series. We fit a diquadratic form with the whole repetitive point in order to remove this effect.
Another error when using radar altimetry over ice caps and radar penetrating surfaces is dues to changes in the ratio volumes/surfaces with time depending on the surface state. The measured height is then variable according to the surface state (or microroughness) variations or other volume echo intensity variations (linked to temperature changes impacting the medium’s absorption). The used retracking (ENVISAT ICE-2 or equivalent for ERS) describes the altimeter return through 4 main parameters : The height (H), the intensity of the Backscatter (Bs), the Leading edge width of the echo (LeW) and the Trailing edge Slope of the echo (TeS). The last 3 parameters describe the shape of the radar echo and help detect physical measurement induced variations of the height measurement. We use the time series of the 3 parameters to estimate this impact and correct for it in order to recover height variations the most possible representative of the actual surface height variations.
For more details, please have look here.