Analysis: Ireland's new supercomputer will help to assess the impact of changing future greenhouse emissions and land use on the climate

By Paul Nolan, Irish Centre for High-End Computing

It can be a challenge following the thicket of acronyms that surround initiatives on climate policy and climate research. The IPCC (Intergovernmental Panel on Climate Change), which issued a special report in October on the state of the global climate, including implications for Ireland, is relatively well known. Another, less well known, is CMIP (Coupled Model Intercomparison Project), of which CMIP6 is in its latest phase.

While the IPCC is regarded as the authoritative voice on climate change issues, it does not engage directly in climate research and its role is to assess the science based on the work of climate scientists. On the other hand, CMIP is an international group co-ordinating the running of computer simulations of the global climate system and comparing the climate forecasts between different computer models. It has played a key role in supporting IPCC in its previous Assessment Reports and is now fully engaged in its sixth phase (CMIP6) to support the IPCC with the next Assessment Report (AR6) due to be released in 2022.

Like the entire world, Ireland’s climate is a complex phenomenon which requires the use of large high performance computing resources to enable forecasting through simulations to predict the climate of the near and far future. At the Irish Centre for High-End Computing (ICHEC), we are working on simulations for CMIP6 in an Environmental Protection Agency (EPA) funded project in collaboration with Met Éireann. The focus of the project will be on assessing the impact of changing future greenhouse emissions and land use on the climate. Over the coming months, this will be greatly helped by the use of the new national supercomputer which will replace our previous supercomputer Fionn which was in operation since 2013.

Time to say goodbye to Fionn

The new machine is substantially more powerful than its predecessor and will will allow us to run our climate models at a finer spatial resolution than previously. This means that the future climate can be simulated more accurately than before. These enhanced capabilities will also allow us to generate an ensemble of climate projections, each differing slightly from the other.

With a large collection of simulations, we can gain an understanding of how much they differ from each other, giving us a measure of confidence in the projections. For this purpose, we will use the EC-Earth Earth System Model, a type of model which simulates the ocean and sea-ice systems of the planet as well as the atmosphere. 

The fundamental objective of CMIP is to improve our understanding of past, present and future climate change. It’s very important that climate model intercomparison projects exist because there isn’t a definitive climate model that faithfully simulates all aspects of the climate. By comparing the outputs of controlled simulations from different models, further insights can be gained on climate linkages and model errors.

These climate simulations require enormous amounts of hard drive space - much more than most other supercomputing projects

The coordination of CMIP6 is very important as it enables the scientific community to focus more effectively on knowledge gaps in the understanding of the Earth’s climate system. In particular, CMIP6 will attempt to address questions around how exactly the Earth responds to increasing greenhouse gas concentrations in our atmosphere. It will also try to understand model biases and determine which models are in agreement on the predictions which are made.

In additional to global climate projections based on alternative scenarios of future emissions and land use changes, CMIP6 includes numerous other model intercomparison projects which provide further refinements to address specific science topics. For example, the Coordinated Regional Downscaling Experiment (CORDEX) project focuses on regional climate projections which provide an important addition to the global suite as they allow for a zoomed in focus on a specific region of interest - Ireland for example.

ICHEC visualisation of Met Éireann climate projection data

Indeed, ICHEC (in collaboration with Met Éireann) will be contributing towards CORDEX as part of a separate EPA-funded project by downscaling the EC-Earth data to provide sharper regional projections for Europe and Ireland. Both the global and regional data will be stored at ICHEC, where it will be published and made available to national researchers and the wider international research community.

The problem of storage

Clearly, participation in CMIP6 is a major undertaking and only large research groups have the necessary resources (such as access to substantial computing power and storage) to engage in the project. In Ireland, we have a big problem to solve with regards to storage space. These climate simulations require enormous amounts of hard drive space - much more than most other supercomputing projects.

Previously our focus in the climate modelling community was on processing power, but now storage is becoming more of an issue. Without storage we cannot keep the results of our simulations for analysis in years to come. This will be a key aspect of ICHEC’s strategy in the future so we can keep enhancing the accuracy of our climate predictions for the nation.

Dr Paul Nolan leads the climate modelling activities at the Irish Centre for High-End Computing


The views expressed here are those of the author and do not represent or reflect the views of RTÉ