Ecologically Sustainable RDM Practices

In a time when human-induced climate change is contributing to rising global temperatures, the responsible use and conservation of natural resources, including energy and water, is increasingly important. Sustainable practices have a role to play across all areas of society; from business and infrastructure to research and our daily lives.

Research Data Management (RDM) can also contribute. Although data may seem intangible, storing, processing, transferring and analysing research data relies on physical infrastructure that consumes energy and other resources. By considering ecological sustainability when planning how research data are managed, we can help reduce unnecessary resource consumption while continuing to support high-quality, reproducible and FAIR research.

So, what can we do in CRC1382?

Photo: Unsplash

Green IT Practices

  • Compress Data

  • Reduce file sizes where appropriate to decrease storage requirements and data transfer. Use lossless compression when the original information must be preserved. 7-Zip is a free, open source tool that can be used to compress files.

  • Avoid Unnecessary Duplicate Storage

  • Before creating another copy, check whether the data are already stored within the CRC1382 data infrastructure. Checking the DataHub for data products to avoid multiple uncontrolled copies of the same data. To learn more about the DataHub check out Newsletter 21!

  • Data Management Plans

  • Consider ecological sustainability when planning, collecting, processing, storing, sharing and archiving research data. Not every temporary or intermediate file needs to be preserved indefinitely. Decide early which raw data, processed data and analysis outputs are necessary for reproducibility, reuse and long-term preservation.

  • Use AI Consciously

  • Choose an AI model that is appropriate for the task. Smaller and more energy-efficient models may be sufficient for routine tasks and can require considerably fewer computational resources. AI Energy Score is a site which evaluates the energy efficiency of AI models doing similar tasks, such as text generation, and provides a star rating with 5 stars indicating the most energy efficient and 1 star indicating least energy efficient. This tool is something that can help researchers to choose energy appropriate models for their needs.

  • Reuse Data

  • Before generating new data, check whether suitable data already exist within the CRC1382 or in external repositories. Data product cards within the CRC1382 DataHub complete with metadata make existing datasets easier to discover and reuse.

Ecologically sustainable RDM is not about simply storing less. It is about using data, storage and computational resources deliberately throughout the research data lifecycle.