Relationship between specular returns in CryoSat-2 data, surface albedo, and Arctic summer minimum ice extent

Specular (mirror-like) reflections in radar altimeter returns are sensitive indicators of flat open water in leads and melt ponds within the Arctic sea ice cover. Here we find increased specular and near-specular returns in CryoSat-2 waveforms as the sea ice cover transitions from a high albedo snow-covered surface to a lower albedo surface dominated by ponds from snow melt. During early melt, mid-May to late June, increases in fractional coverage of specular returns ( F SR ) show spatial correspondence with concurrent decreases in albedo. To examine the utility of F SR we compared its efficacy with that of satellite-derived albedo in forecasting summer minimum ice extent (SMIE). Regression analysis of the area-averaged F SR ( F SR )(2011–2017) shows that ~72% of SMIE variance can be explained by the dates when F SR climbs to 0.5 within two latitudinal bands covering 70–80°N and 80–90°N. The lag between the two crossing dates provides a measure of the relative rate of the poleward progression of melt. Approximately 93% of SMIE variance can be explained by the date when albedo drops to 0.6 in these same latitudinal bands. Standard errors for these regressions are 0.37 and 0.19 . 10 6  km 2 , respectively. Calculating the regression coefficients using only 2011–2016, the 2017 SMIE was forecast with residuals of 0.06 (2% of the total extent) and –0.17 . 10 6  km 2 (4%). Using only 2011–2015 yielded residuals that are less than 0.5 . 10 6  km 2 (~10%) in forecasts of both 2016 and 2017 SMIE, demonstrating the robustness of the regression models. Even though large-scale changes in albedo during summer melt is a characteristic feature of the ice surface, available albedo fields have not been directly used in SMIE forecasts. While this CryoSat-2 record is short, these results suggest that both F SR and albedo could be potentially useful for enhancing forecasts of SMIE.