Surface Meltwater Routing Through the Supraglacial-Proglacial River System on the Northwestern Greenland Ice Sheet

Ya Li,Kang Yang,Laurence Smith,Xavier Fettweis,Shuai Gao,Wensong Zhang
DOI: https://doi.org/10.5194/egusphere-egu21-3625
2021-01-01
Abstract:Large, complex supraglacial river networks are widely distributed on the northwestern Greenland Ice Sheet (GrIS) each summer. Owing to the absence of moulins and crevasses on the ice surface, meltwater is continuously routed on the ice surface by supraglacial river networks to feed proglacial rivers on land. This continuous supraglacial-proglacial river system controls the magnitude and timing of surface meltwater runoff on the northwestern GrIS but remains poorly studied. In this study, we first mapped the supraglacial-proglacial river system across the Inglefield Land on the northwestern GrIS during 2016–2019 melt seasons using ninety Sentinel-2 and forty-five Landsat-8 images. Then, we proposed two quantitative river metrics, i.e., surface meltwater area fraction and proglacial river width, to quantify the seasonal and annual evolutions of the supraglacial-proglacial river system. Next, we correlated these satellite-derived river metrics with surface meltwater runoff estimated by two Surface Mass Balance (SMB) models (MAR v3.11 and MERRA-2), and estimated the optimal meltwater routing lag times. Our results showed that: (1) two satellite-derived river metrics, surface meltwater area fraction and proglacial river width, are strongly and positively correlated, indicating that the northwestern GrIS supraglacial-proglacial river system can efficiently route surface meltwater from the ice surface to the proglacial zone; (2) these two satellite-derived river metrics are also positively correlated with simultaneous surface runoff simulated by MAR and MERRA-2, indicating that SMB models can capture the general runoff pattern but exhibit considerable discrepancy with satellite observations; and (3) delayed MAR surface runoff better match two satellite-derived river metrics than simultaneous MAR surface runoff, and the optimal lag times are both two days, suggesting that supraglacial routing accounts for most of the lag time whereas rapid proglacial routing accounts for short lag time. Overall, the northwestern GrIS supraglacial-proglacial river system is a unique and efficient meltwater routing system, and multi-temporal satellite observations of this river system raise prospects for directly estimating surface meltwater runoff on the poorly-studied northwestern GrIS.
What problem does this paper attempt to address?