作者: Xavier Fettweis , Stefan Hofer , Uta Krebs-Kanzow , Charles Amory , Teruo Aoki
关键词:
摘要: Abstract. The Greenland Ice Sheet (GrIS) mass loss has been accelerating at a rate of about 20 ± 10 Gt/yr2 since the end 1990's, with around 60 % this directly attributed to enhanced surface meltwater runoff. However, in climate and glaciology communities, different approaches exist on how model balance (SMB) components using: (1) complex physically-based models which are computationally expensive; (2) intermediate complexity energy models; (3) simple fast positive degree day base their inferences statistical principles highly efficient. Additionally, many these compute SMB based spatial temporal resolutions, forcing fields as well ice sheet topographies extents, making inter-comparison difficult. In GrIS intercomparison project (GrSMBMIP) we address issues by each same data (i.e., ERA-Interim reanalysis) except for two global is limited oceanic conditions, time interpolating all modelled results onto common mask 1 km horizontal resolution period 1980–2012. outputs from 13 then compared over estimates using combination gravimetric remote sensing GRACE measured discharge, cores, snow pits, in-situ observations, remotely sensed bare extent MODerate-resolution Imaging Spectroradiometer (MODIS). Our reveal that mean between 1980 2012 an average 340 ± Gt/yr, but decreased −7.3 Gt/yr2 (with significance 96 %), mainly driven increase 8.0 Gt/yr2 98 %) Spatially, largest spread among can be found margins sheet, highlighting need accurate representation ablation zone processes driving melt. addition, higher density observations required, especially south-east accumulation zone, where reach 2 mWE/yr due large discrepancies snowfall accumulation. Overall, polar regional (RCMs) perform best particular simulating precipitation patterns. other simpler faster have biases order than RCMs remain useful tools long-term simulations. Finally, it interesting note ensemble produces estimate present relative suggesting not systematic models.