Assessment of radar‐derived snow depth over Arctic sea ice

Knowledge of contemporaneous snow depth on Arctic sea ice is important both to constrain the regional climatology and to improve the accuracy of satellite altimeter estimates of sea ice thickness. We assess new data available from the NASA Operation IceBridge snow radar instrument and derive snow depth estimates across the western Arctic ice pack using a novel methodology based on wavelet techniques that define the primary reflecting surfaces within the snow pack. We assign uncertainty to the snow depth estimates based upon both the radar system parameters and sea ice topographic variability. The accuracy of the airborne snow depth estimates are examined via comparison with coincident measurements gathered in situ across a range of ice types in the Beaufort Sea. We discuss the effect of surface morphology on the derivation, and consequently the accuracy, of airborne snow depth estimates. We find that snow depths derived from the airborne snow radar using the wavelet‐based technique are accurate to 1 cm over level ice. Over rougher surfaces including multiyear and ridged ice, the radar system is impacted by ice surface morphology. Across basin scales, we find the snow‐radar‐derived snow depth on first‐year ice is at least ∼60% of the value reported in the snow climatology for the Beaufort Sea, Canada Basin, and parts of the central Arctic, since these regions were previously dominated by multiyear ice during the measurement period of the climatology. Snow on multiyear ice is more consistent with the climatology. Novel processing technique applied to Operation IceBridge snow radar dataSnow depth uncertainty as a function of sea ice surface topography is analyzedContemporary Arctic snow depth on basin scale compared to snow climatology