Cloud Computing Is Revolutionizing Climate Change Research—While Adding to the Problem

Forget Blair and Serena: The world’s most influential frenemies are cloud computing and climate change.

Cumulus: The cloud’s computing power can supercharge climate change research, which involves identifying trends in increasingly large, complex data sets. It’s also helped democratize access to machine learning tools.

• Researchers lean on ML algorithms to pull insights from rainfall patterns, satellite imagery, ocean temperatures, and more.
• A University of Oxford team uses Amazon Web Services to understand ship exhaust’s effects on actual clouds, reports the WSJ.

Cumulonimbus: The data centers that enable cloud computing eat up enormous amounts of energy—much of which comes from fossil fuels. By 2030, data center demand could reach 13% of the world’s total electricity consumption.

• Training one natural language processing pipeline, including tuning and experimentation, can emit over 78,000 lbs of CO2 equivalent, according to one paper. That's more than the average human's climate change impact over a two-year period.

The massive amounts of data to sort through will only increase as we inch closer to climate change’s “point of no return.” The National Ocean and Atmospheric Administration alone collects ~20 terabytes of data daily.

Potential bright spot: According to Microsoft’s Project Natick, storing data centers below the ocean could be instrumental in reducing their energy consumption.