Atmospheric processes and their representation in Global Climate Models (GCMs)

There are many diverse and complex interactions within the earth system which control the behaviour of the large outlet glaciers of Greenland. One such interaction is between the atmosphere and the surface of the ice. The atmospheric controls on the surface melt rates have great importance for the ice dynamics: supraglacial meltwater penetrates into the interior of the glacier via water channels and contributes to mechanical and thermodynamic processes that affect the stability of the entire glacier. Subproject 8 focusses on the atmosphere and surface mass balance of the 79N region. Surface mass loss and runoff of melt water are key contributors to the loss of ice in Greenland in the last few decades. Surface mass balance is closely linked to the weather and climate of a region, however relatively little was known about the key atmospheric conditions of 79 North Glacier when we started our studies.

Contacts: Dr. Jenny Turton & Prof. Dr. Thomas Mölg

What was the main question you aimed to answer during the first project phase?

The main aims of GROCE phase 1 can be summarised as (1) investigating the main mesoscale processes influencing the climate of the region and (2) estimating the surface mass balance of the most recent past. Data from automatic weather stations, ERA-Interim reanalysis product, the Weather Research and Forecasting (WRF) atmospheric model and COSIPY surface mass balance model were used to complete these aims (more information on the methods can be found here). 

What were your main results?

  1. Using observational data and reanalysis products, we identified that katabatic winds and warm-air advection events have a large influence on the atmospheric variability during winter, and regularly raise the air temperature by 10°C or more in less than 24 hours (Figure 1). We also analysed the annual air temperature and identified a 3°C warming since 1979, in agreement with, but exceeding the global warming rate. The results were published here: https://doi.org/10.1175/MWR-D-18-0366.1.
  2. We optimised the Polar modified Weather Research and Forecasting (WRF) atmospheric model using weather stations, and then simulated the meteorological conditions for five years from 2014 to 2018 (Figure 2). Daily average values of key meteorological variables are archived online on an open access repository for others to use  and can be accessed here: https://doi.org/10.17605/OSF.IO/53E6Z. Evaluation of the model output against observation sites and a description of available data are published here: https://doi.org/10.5194/essd-12-1191-2020. The model output is currently being used in the second phase of GROCE to assess the relationship between atmospheric processes and supraglacial lake (melt lakes on the surface of the glacier) distribution.
  3. We used the atmospheric model output at 1 km horizontal resolution as input to a surface mass balance model. Using the ‘COSIPY’ mass balance model, we now have 5 years of hourly surface mass balance estimates at 1 km resolution for the 79N and NEGIS region. The results of this are currently under review at the Journal of Glaciology. 

 

What are your goals for GROCE-2?

Out first aim for GROCE-2 is to investigate the atmospheric drivers of supraglacial lake (melt water lakes on the surface of the glacier) development and drainage, in order to provide insight into the amount of freshwater entering and leaving the glacier system. Close collaboration with subproject 7 will continue to fulfil this aim. 

Our second main aim is to assess a critical resolution for global climate models to accurately represent key atmospheric processes (i.e., the local energy balance, main meteorological variables and mesoscale processes) in the region. Using our atmospheric model output from its three horizontal resolutions (1km, 5km 25km), reanalysis data at a range of horizontal resolutions, and Global Circulation Models at approximately 100 km resolution, we will assess which atmospheric processes are well represented at these scales and which require high-resolution models to accurately simulate them.