NHM–SMAP: spatially and temporally high-resolution nonhydrostatic atmospheric model coupled with detailed snow process model for Greenland Ice Sheet

• Published in: The Cryosphere
• Main Authors: M. Niwano, T. Aoki, A. Hashimoto, S. Matoba, S. Yamaguchi, T. Tanikawa, K. Fujita, A. Tsushima, Y. Iizuka, R. Shimada, M. Hori
• Format: Article in Journal/Newspaper
• Language: English
• Published: Copernicus Publications 2018
• Subjects: geo; envir; Greenland; Ice Sheet; The Cryosphere
• Online Access: https://doi.org/10.5194/tc-12-635-2018; https://www.the-cryosphere.net/12/635/2018/tc-12-635-2018.pdf; https://doaj.org/article/77234e68d45443529101f4a58950aa81

To improve surface mass balance (SMB) estimates for the Greenland Ice Sheet (GrIS), we developed a 5 km resolution regional climate model combining the Japan Meteorological Agency Non-Hydrostatic atmospheric Model and the Snow Metamorphism and Albedo Process model (NHM–SMAP) with an output interval of 1 h, forced by the Japanese 55-year reanalysis (JRA-55). We used in situ data to evaluate NHM–SMAP in the GrIS during the 2011–2014 mass balance years. We investigated two options for the lower boundary conditions of the atmosphere: an offline configuration using snow, firn, and ice albedo, surface temperature data from JRA-55, and an online configuration using values from SMAP. The online configuration improved model performance in simulating 2 m air temperature, suggesting that the surface analysis provided by JRA-55 is inadequate for the GrIS and that SMAP results can better simulate physical conditions of snow/firn/ice. It also reproduced the measured features of the GrIS climate, diurnal variations, and even a strong mesoscale wind event. In particular, it successfully reproduced the temporal evolution of the GrIS surface melt area extent as well as the record melt event around 12 July 2012, at which time the simulated melt area extent reached 92.4 %. Sensitivity tests showed that the choice of calculation schemes for vertical water movement in snow and firn has an effect as great as 200 Gt year−1 in the GrIS-wide accumulated SMB estimates; a scheme based on the Richards equation provided the best performance.