In principle, any topography change in the footprint of the radar altimeter should alter the returned waveform. This means that alterations in the subglacial lakes presentation on the surface, or the movement of crevasses should be captured by the waveform deformation. If we can sucessfully understand and model how that surface deformation influences the change in the waveform, it may be possible to look at the historic record and extract surface changes from the waveforms themselves. If we can do that, we can look at how subglacial lakes changed as far back as the 1990's.
To better understand Antarctic ice-sheet processes, we need to understand how small scale features (such as subglacial lakes or crevasses) were changing multiple decades ago. Currently we are unable to do this, as the data record is only high resolution enough after 2012 or so. However, due to satellites such as ERS1/2 and ENVISAT we have a plethora of radar altimetry data from the early 1990's with high spatial coverage.
So far, we have successfully created a GPU-based forward model which simulates radar altimetry for ERS1/2 and ENVISAT. We have modeled surface deformation based on true subglacial lake deformation values as estimated via laser altimetry. We then compared this to the true data.
Below is the paper I presented at IGARSS 2024, which discusses just the model and does not compare it to true data with deformations.