| Summary: | A significant source of error in the retrieval of sea ice freeboard from pulse-limited radar altimeters arises when scattering from off-nadir leads dominates the power echo causing the onboard tracker to “snag” and overestimate the two-way travel time. This range overestimate translates into an ocean elevation underestimate relative to the ice surface and an overestimate of the sea ice freeboard. We demonstrate using interferometric CryoSat-2 data that it is possible to infer the across-track angle of return to off-nadir leads and their location in the CryoSat-2 footprint and hence correct for the associated range error for the first time. It is found that specular scattering from leads can dominate the radar echo some 1530 m off nadir. Over the region studied, the mean ocean elevation bias is closely associated with the “pulse peakiness” (PP) parameter used for identifying specular waveforms. Considering only the most specular waveforms, the elevation bias was measured to be -1.21 ±0.93 cm. However, lowering the PP threshold includes lower power waveforms originating from higher off-nadir angles, and the elevation bias becomes -4.06 ±1.66 cm. Unaccounted for, these biases represent an ~ 10-40-cm overestimate in ice thickness. Despite the relatively large error on the mean bias, correcting for off-nadir ranging contributes only a small amount to the elevation uncertainty when combined with range and orbit determination uncertainties. We found that making use of CryoSat-2's interferometric mode over sea ice ultimately decreases the uncertainty on the area-averaged ocean elevation by allowing the inclusion of more waveforms in the analysis.
|
|---|