Careers Denmark
Browse opportunities

Digital & Data

PhD scholarship in Physics-Informed Machine Learning for Greenland’s Firn Tipping Point - DTU Space

Danmarks Tekniske Universitet (DTU) Kgs. Lyngby, Denmark Posted 29 Oct 2026 Deadline 29 Nov · 12 days left Employer's career siteOn-siteTemporaryFull timeStudent / internEnglish

Summary

A PhD student will develop physics-informed machine-learning models combining satellite data and firn models to predict when Greenland’s firn may stop storing meltwater. The role includes detecting tipping-point signals, collaboration with international partners, and teaching.

The fields are read by AI from the ad. If the employer filled in a field themselves, we use theirs.

Opportunity summary

Are you interested in turning natural phenomena into equations and code, and then applying them to big open questions in climate science? If so, we are looking for a PhD student with a background in physics and/or mathematics to help predict when Greenland’s porous near-surface firn layer, which currently stores much of the ice sheet’s meltwater, loses that capacity, and what that means for sea-level rise.

You will build models that combine physics, machine learning, and satellite data, and apply them to climate scenarios from the EU’s Destination Earth initiative. You will join a diverse and collaborative group at DTU Space and develop skills in scientific computing and machine learning that are valued in both research and industry.

No prior experience in glaciology or climate science is needed; we will help you learn the field.

Responsibilities and qualifications

In the FirnTip project, you will develop a framework that predicts when and where Greenland’s firn stops being able to store meltwater, a possible climate tipping point. You will work closely with your supervisors at DTU Space as well as international partners.

Your main tasks will be to

  • Identify which climate factors control when firn stops absorbing meltwater
  • Combine satellite-derived firn density with firn models (data assimilation)
  • Develop machine-learning models (e.g., graph neural networks) that obey physical conservation laws.
  • Detect early-warning signals of tipping points in future climate scenarios
  • Spend 3-6 months aboard visiting international partners, planned in collaboration with you
  • Take part in teaching and knowledge-sharing as part of the DTU PhD program

To do well in the role, we think you need:

  • A master’s degree in physics, mathematics or a field with a comparable mathematical foundation
  • Experience writing scientific code, e.g., from your thesis or projects
  • Good knowledge of at least one of: differential equations, numerical methods, statistics, or machine learning

This is an extract. Read the whole advert at Employer's career site