Understanding Biases in ICESat-2 Data Due to Subsurface Scattering Using Airborne Topographic Mapper Waveform Data

The process of laser light reflecting from surfaces made of scattering materials that do not strongly absorb at the wavelength of the laser can involve reflections from hundreds or thousands of individual grains, which can introduce delays in the time between light entering and leaving the surface. These time-of-flight biases depend on the grain size and density of the medium, and thus they can result in spatially and temporally varying surface height biases estimated from laser altimeters, such as NASA's ICESat-2 (Ice Cloud, and land Elevation Satellite-2) mission. Modeling suggests that ICESat-2 might experience a bias difference as large as 0.1–0.2 m between coarse-grained melting snow and fine-grained wintertime snow (Smith et al., 2018), which exceeds the mission's requirement to measure seasonal height differences to an accuracy better than 0.1 m (Markus et al., 2017). In this study, we investigate these biases using a model of subsurface scattering, laser altimetry measurements from NASA's ATM (Airborne Topographic Mapper) system, and grain size estimates based on optical imagery of the ice sheet. We demonstrate that distortions in the shapes of waveforms measured using ATM are related to the optical grain size of the surface estimated using optical reflectance measurements and show that they can be used to estimate an effective grain radius for the surface. Using this effective grain radius as a proxy for the severity of subsurface scattering, we use our model with grain size estimates from optical imagery to simulate corrections for biases in ICESat-2 data due to subsurface scattering and demonstrate that, on the basis of large-scale averages, the corrections calculated based on the satellite optical imagery match the biases in the data. This work demonstrates that waveform-based altimetry data can measure the optical properties of granular surfaces and that corrections based on optical grain size estimates can correct for subsurface-scattering biases in ICESat-2 data.