Abstract
Abstract. HARMONIE-AROME is a convection-permitting non-hydrostatic model that includes the multi-purpose SURFEX surface model. It is developed for high resolution (1–3 km) weather forecasting and applied in a number of regions in Europe and the Mediterranean. A version of HARMONIE-AROME is also under development for regional climate modelling. Here we run HARMONIE-AROME for a domain over Greenland that includes a significant portion of the Greenland ice sheet. The model output reproduces temperature, wind speed and direction and relative humidity over the ice sheet well when compared with the observations from PROMICE automatic weather stations (AWS) operated within the model domain on the ice sheet (mean temperature bias 1.31 ± 3.6 K) but we identified a much lower bias (−0.16 ± 2.3 K) at PROMICE sites on days where melt does not occur at the ice sheet surface and is thus an artefact of the simplified surface scheme over glaciers in the existing HARMONIE-AROME operational set-up. The bias in summer time temperature also affects wind speed and direction as the dominant katabatic winds are caused by the cold ice surface and slope gradient. By setting an upper threshold to the surface temperature of the ice surface within SURFEX we show that the weather forecast error over the Greenland ice sheet can be reduced in summer when glacier ice is exposed. This improvement will facilitate accurate ice melt and run-off computations, important both for ice surface mass budget estimation and for commercial applications such as hydro-power forecasting. Furthermore, the HCLIM regional climate model derived from HARMONIE-AROME will need to accurately account for glacier surface processes in these regions in order to be used to accurately compute the surface mass budget of ice sheets and glaciers, a key goal of regional climate modelling studies in Greenland.
Highlights
The HARMONIE-AROME model is the current operational weather forecasting model used by the Danish Meteorological Institute and other weather forecasting institutions in Europe and North Africa (Termonia et al, 2017) for numerical weather prediction (Bengtsson et al, 2017)
In order to assess the performance of HARMONIE-AROME over the ice sheet we examined model output and compared it with data from the PROMICE weather station observations (Van As et al, 2011) on the ice sheet at a number of locations around the ice sheet where the domain overlapped with the automatic weather stations (AWS) sites that were active in 2014
Hboomis the height of the PROMICE station measurement boom, U is the wind measured at this height, U (10 m) is the wind scaled to 10 m height, and the roughness length z0 is taken from HARMONIE-AROME to be 0.001 m for snow and ice
Summary
Operational HARMONIE-AROME numerical weather prediction (NWP) domains include both Greenland and Iceland but little or no model evaluation of HARMONIE-AROME has previously been performed over ice covered surfaces. The HCLIM model, a climate variant of HARMONIE-AROME, though with a different dynamical core, is already in use for climate simulations within Europe (Lindstedt et al, 2015). A future target is to apply HCLIM more broadly to domains such as Greenland where glaciation is an important environmental characteristic, in order to accurately characterize the surface mass budget of the ice sheet and peripheral glaciers. In this paper we examine the model performance over glacier surfaces in an operational NWP version of HARMONIE-AROME with the aim of assessing its utility as a climate model for applications in Greenland in the future, including in the HCLIM model format
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
