High-resolution numerical simulation of summer wind field comparing WRF boundary-layer parametrizations over complex Arctic topography: case study from central Spitsbergen

The Weather Research and Forecasting (WRF) mesoscale model was run in three different configurations over Svalbard Archipelago and compared with the 12‑day summer measurements of surface wind characteristics at three sites along the western coast of Petuniabukta, central Spitsbergen. For studying wi...

Full description

Bibliographic Details
Published in:Meteorologische Zeitschrift
Main Authors: Kamil Láska, Zuzana Chládová, Jiří Hošek
Format: Article in Journal/Newspaper
Language:English
Published: Borntraeger 2017
Subjects:
Online Access:https://doi.org/10.1127/metz/2017/0796
https://doaj.org/article/cef0af4137e248849a213af5c9666819
Description
Summary:The Weather Research and Forecasting (WRF) mesoscale model was run in three different configurations over Svalbard Archipelago and compared with the 12‑day summer measurements of surface wind characteristics at three sites along the western coast of Petuniabukta, central Spitsbergen. For studying wind patterns over complex topography, we chose the following sites for their differing terrain elevation and local surface characteristics: a raised marine terrace (15 m a.s.l.), the foreland of Hørbyebreen Glacier (67 m a.s.l.) and the top of Mumien Peak (773 m a.s.l.). The WRF simulations were conducted using three boundary layer (BL) parameterization schemes: the Yonsei University (YSU), the Mellor-Yamada-Janjic (MYJ) and the Quasi-Normal Scale Elimination (QNSE), with 1‑km horizontal resolution of the inner domain. The WRF simulations agreed fairly well with the surface wind observations taken at all the sites. For wind speed, the mean correlation coefficients between the modelled and observed data ranged from 0.56 to 0.67. The best results across all the stations were found for the QNSE parameterization scheme, with a bias of 0.1 m s−1. The wind speed simulations were sensitive to the geographical location and elevation of the stations. All the parameterization schemes had difficulties in capturing the surface wind field in the narrow valley near the Hørbye foreland station, while satisfactory estimates were found at the top of Mumien Peak and Terrace station, located in a wide part of the fjord. The WRF estimates proved to be highly sensitive to the large-scale forcing, as was documented in the cyclonic circulation patterns with strong northerly winds. This led to an overestimation of the modelled wind speed on the leeward slopes of the highest peaks in the study area. The model results tended to be underestimated during the anticyclonic situations with the geostrophic winds < 4 m s−1 from the northwest and east sectors.