High spatial resolution Greenland Ice Sheet mass change linked to Greenland Blocking and North Atlantic Oscillation

The low spatial resolution (300–400 km) of Gravity Recovery and Climate Experience (GRACE) and the 11-month data gap between GRACE and GRACE Follow-On (GRACE-FO) complicate the evaluation of long-term ice mass loss at sub-region scale. Although several studies have downscaled GRACE data for water storage evaluation, efforts to downscale ice mass change are limited. Here, we utilized multiscale geographically-weighted regression (MGWR) to improve the spatial resolution of GRACE-derived ice mass change from 0.25° to 5 km in Greenland Ice Sheet (GrIS). The reconstructed results after downscaling demonstrate consistent temporal variation and reduced noise compared to other reconstruction methods for sub-regions and the entire GrIS. They can capture finer signals of ice mass change while maintaining original spatial characteristics. Validation against free air gravity anomalies changes also demonstrates the superior performance of the downscaled results over the reconstructed CSR Mascon data. For overlapping lines, the mean correlation coefficient (CC) increases from 0.13 to 0.32, with the mean root mean square error (RMSE) decreasing from 1.36 to 1.13. At the intersection points, the mean CC improves from 0.06 to 0.30, and the mean RMSE is reduced from 1.33 to 1.05. They are further used to detect the contributions of Greenland Blocking (GB) and North Atlantic Oscillation (NAO) to high-resolution ice mass change through precipitation and runoff in four sub-periods, during which the driving factors affecting ice mass change differ significantly. Additionally, NAO and GB can influence ice mass change through solid ice discharge by impacting surface temperature and sea-ice concentration.