Impact of tundra snow layer thickness on modelled radar backscatter

Microwave radar backscatter within an Arctic tundra snowpack is strongly influenced by spatial variability of the thickness of internal layering; often comprising two dominant snow layers (basal depth hoar overlain by wind slab). Determining the relative proportions of depth hoar and wind slab from a snowpack of a known depth may help our future capacity to invert forward models of electromagnetic backscatter to improve simulations of snow water equivalent.

Extensive measurements of snow microstructure were made within Trail Valley Creek, NWT, Canada in April 2013 and March 2018 at 54 pit and 9 trench locations (trench extents ranged between 5m to 50m). Ground microstructure measurements included traditional stratigraphy, near infrared stratigraphy, Specific Surface Area (SSA), and density. Trench measurements of stratigraphy showed the mean proportion of depth hoar was just under 30% of total snow depth and the mean proportion of wind slab was consistently greater than 50%, which showed an increasing trend with increasing total snow depth. Measurements of density and SSA in each major snowpack layer were distinctly clustered allowing application of median values as important parameters in the snow microwave radiative transfer (SMRT) model to estimate radar backscatter.

SMRT was used to quantify how spatial variability in the measured microstructural parameters of a synthetic snowpack affect one-dimensional retrievals of snow water equivalent (SWE). A three layer snowpack retrieval experiment over snow depths typical of Arctic tundra demonstrated that a SSA 10% smaller than the median of the depth hoar layer, the most important layer for scattering, was required to retrieve SWE within a ±30 mm SWE error budget. The impact of SSA variability in wind slab was smaller and negligible for fresh snow.