Climate Change Report published with Met ÉireannPosted on 30 September 2013
ICHEC is delighted to announce that Mr. Alastair McKinstry and Dr. Paul Nolan, climate scientists at ICHEC, have worked in collaboration with Met Éireann and a number of national research centres to publish a climate change report.
The report, Ireland's climate: the road ahead provides an overview of the expected changes in the future Irish climate due to global warming. The report was launched on September 25th at the Botanic Gardens, Glasnevin, Dublin, Ireland. The timing of the report is particularly appropriate as it coincides with the launch of the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change. While AR5 addresses climate change on global scales, this report looks closer to home and provides an up-to-date assessment of Irish climate trends.
Global simulations were performed using EC-Earth, a model consisting of an atmosphere-land surface model coupled to an ocean and sea-ice model. Met Éireann and ICHEC were the two Irish institutions involved in both the development of the model and the running of simulations on the ICHEC and ECMWF supercomputers. The future simulations run from 2006 to 2100 based on the RCP4.5 (low to medium) and RCP8.5 (high) greenhouse gas emission scenarios. In addition, ICHEC acted as the data hub for the EC-Earth model data and facilitated the sharing of the data with international researchers via the CMIP5 project.
Mean global land temperatures are expected to rise by 2.7 degrees for the period 2071-2100 under a medium-low emission scenario (RCP4.5) and by up to 4.6 degrees under a high emission scenario (RCP8.5). Warming is greatest at high latitudes, leading to an accelerated loss of Arctic sea-ice cover.
Changes in 2071-2100 mean 2m temperature [°C] relative to 1961-1990 for (a) DJF RCP4.5, (b) JJA RCP4.5, (c) DJF RCP8.5 and (d) JJA RCP8.5
The computational cost of running complex global climate models increases rapidly with the level of climate detail required. To achieve reasonable execution speeds, the model grid is set relatively coarse in comparison with operational weather forecast models. The relatively coarse grid (125km) used in the EC-Earth global simulations underestimates extremes and local effects, which are important for precipitation and wind modelling in particular.
To capture local climate details it is necessary to downscale the global model data to a finer grid over Ireland. This work was started at UCD and continued by researchers at ICHEC using the Max Planck Institute's ECHAM5 global climate model (GCM), the UK Met Office's HadGEM2-ES GCM, the CGCM3.1 GCM from the Canadian Centre for Climate Modelling and the EC-EARTH GCM. These global model data were downscaled using three different Regional Climate Models (RCMs): COSMO-CLM versions CLM3 & CLM4 and WRF. The COSMO-CLM simulations were run at 50km, 18km, 7km and 4km resolutions. The WRF simulations were run at 54km, 18km and 6km resolutions. To address the issue of model uncertainty, a large ensemble of RCM simulations were run. This was a substantial computational task and required extensive use of the ICHEC supercomputer systems over 3 to 4 years.
To create a large ensemble, the 4km, 6km and 7km resolution data were regridded to a common 7km grid over Ireland. The simulations were carried out using RCP4.5 and B1 greenhouse gas scenarios to create a "medium-low emission" ensemble while RCP8.5, A1B and A2 simulations were used to create a "high emission" ensemble. This Multi-Model Ensemble (MME) approach enables the uncertainty in RCM projections to be quantified, providing a measure of confidence in the predictions.
Projected mid-century temperature changes for the high emission ensemble; (a) the top 5% of summer highest daytime temperatures and (b) the lowest 5% of winter night-time temperatures.
The animation above presents an ICHEC visualization of Met Éireann EC-EARTH data. The annual 2m temperature change with respect to the 30-year baseline (1961-1990) is presented for both the RCP4.5 (medium) and RCP8.5 (high) scenarios.
The second visualization presents the Annual Sea Surface Temperature (SST) and Sea Ice Fraction for March and September. Because of the high heat capacity of water, the SST lags behind atmospheric temperatures. Hence, the artic sea ice is at a maximum during March and minimum during September.
Model of collective online behaviour of Facebook users
|ICHEC and NewsWhip take on Viral stories
A collaboration conducted to predict how viral a story will be
|ICHEC's partnership with Opel Ireland (General Motors)
Springboard course on HPC systems administration, open for applications and starts Aug
|Fionn most power efficient supercomputer on Top 500
Fionn is now at position 478 on the latest Top500 performance list. However in terms of power efficiency, it is the number 1 x86_64 system on the list.
|NVIDIA's Women Who CUDA Campaign
The work of computational scientist at ICHEC is highlighted in the Women Who CUDA
|ICHEC’s collaboration with NTNU
Collaborative work: "Exploiting HPC Tools and Techniques" in the META magazine
|International HPC Summer School 2014
1-6 June, Budapest, Hungary
|NUI Galway Joins Intel Parallel Computing Centre Programme to Boost Parallel Application Performance in High Performance Computing
Irish Centre for High End Computing (ICHEC) will host the research that has likely applications in materials science, meteorology and data analytics.