Understanding climate change using high performance computing and machine learning


What scientific, industrial, and social challenges can HPC face?  

Given its interdisciplinary nature and its ability to process massive amounts of data and perform complex calculations, HPC is essential to address a wide range of scientific, industrial, and societal challenges

1. Earth Sciences and Climate

The HPC forms the basis of the study and prediction of the climate since it allows to make more precise meteorological predictions and, in real-time, to predict and manage large-scale natural disasters, such as devastating hurricanes, or to study the behavior of the oceans. In Europe Between 1970 and 2012, extreme weather events claimed 149,959 lives and cost € 270 billion in economic damage. 

Weather forecasting is highly dependent on numerical simulation in supercomputers. The more powerful the supercomputer, the more accurately and in advance climatologists can predict the magnitude and trajectory of storms and floods, thereby facilitating decisions such as activating early warning systems in time to evacuate. The population and save human lives.

HPC technologies also allow for simulations related to climate change with increasing resolution (for example, in the study of the oceans’ behavior) and more precise monitoring of land resources’ evolution. Furthermore, HPC allows us to learn more about the geophysical processes and the Earth’s interior structure, which helps us better understand natural disasters such as earthquakes. 

To create maps of seismic risk, seismologists use data on the ground movement recorded by more than 10,000 seismometers installed worldwide; however, this massive amount of data can only be handled with powerful HPC infrastructures.

2. Health, demographic change, and well-being

HPC is a fundamental instrument when designing renewable energy parks or high-performance photovoltaic materials, testing new types of more efficient materials for solar panels, and optimizing turbines for the production of electricity. For example, precision in wind resource assessment, wind farm design, and short-term and micro-scale wind simulations to forecast daily energy production allows predicting wind farms’ commercial viability. 

HPC is an indispensable tool for the wind energy production industry, especially in areas where the terrain has complex characteristics. Thanks to some energy, today’s nuclear power plants could be replaced by a safer, more environmentally friendly, and virtually inexhaustible energy source. In today’s experimental fusion reactors, HPC is being used to simulate and manage fusion plasma behavior, especially instabilities, turbulent transport, plasma-wall interaction, or heating.

3. Health, demographic change, and well-being

HPC is a driving force for the development of new forms of medicine. Personalized and precision medicine relies heavily on HPC to process information about the patient’s genes, proteins, and environment to prevent, diagnose, as well as treat disease. For example, each disease has its genetic makeup when it comes to cancer, making each tumor’s cells and tissues unique with specific tendencies and vulnerabilities. Personalized and precision medicine will assign each patient the most appropriate treatment and respond to their specific needs.

The early detection of rare diseases is a challenge that HPC can address more efficiently. Thanks to HPC technologies, diagnostics, and tests that now take weeks could be completed in just a few days.HPC also enables faster and more efficient analysis of genomic sequences. There are approximately 4 100 genetic diseases that affect humans, and this type of disease constitutes one of the main causes of infant mortality.

In the field of biomolecular research, HPC is also used to investigate the dynamics of biomolecules and proteins in human cells, critical for more efficient treatment of autoimmune diseases and cancer and diabetes. In brain research, as in the flagship FET initiative ‘Human Brain Project’, HPC is used for high-resolution, multiscale modeling and simulation to understand the human brain’s structure and function.

Last but not least, HPC is the cornerstone for new drug development. The time required to develop a new drug ranges from ten to seventeen years, and this is compounded by the increase in the cost of the process, which makes it increasingly unaffordable for both companies and patients. However, HPC can significantly speed up the testing of candidate drug molecules. It can also help reorient existing drugs for the treatment of new diseases. All this will be beneficial for the treatment of patients and will also significantly reduce the costs of the process.

4. Food security, sustainable agriculture, marine research, and the bioeconomy

HPC is needed to developing more sustainable agriculture by optimizing food production, analyzing factors that contribute to sustainability, and monitoring pests, disease control, and the effects of pesticides. There are HPC-based applications that use, for example, radio frequency identification (RFID) tags that can automatically store and download a set of data on the bullet’s moisture percentage, weight, and GPS position. 

In the future, there will be a wide display of micro-labels the size of a soil particle to measure things such as humidity or disease pressure and even determine whether or not the crop is ready for harvest. HPC technologies can also contribute to the more efficient management of agricultural and water resources and help vulnerable communities in a region by improving their management and drought response capacities.

The Future of Climate Change Science is in the Cloud

When comparing the impact of climate in data centers, cloud providers’ advantages and on-premises are clear across power mix, energy efficiency and resource utilization. A typical large-scale cloud provider achieves at least 65 percent server utilization rates compared to 15% on-premises, which means as companies move to the cloud, they typically provide fewer than ¼ of the servers compared to that for on-premises. 

Furthermore, a typical on-premises data center is 29% less efficient in its use of power than a typical large-scale cloud provider that makes use of a world-class facility designs, workload-optimized equipment and cooling systems.

Add to needing fewer, more power-efficient servers that giant cloud providers like AWS use. A power mix that is 28 percent less carbon intense than the global average, and cloud customers can end up with a reduction in carbon emissions of 88%.

Climate change is one of the most difficult problems of our time, and if a meaningful solution is not found, the consequences can have repercussions for us and our children’s future. Making change in this arena will require a joint effort across academia, government, industry, nonprofits, and other society. It will require innovation, ingenuity, and scale. High power computing and machine learning on the cloud will serve as the key to unlocking scientific insights into understanding and combating climate change.

Need Help?

Add a strong one liner supporting the heading above and giving users a reason to click on the button below.