Detection of lakes over the Greenland Ice Sheet

Project type: proof-of-concept study

Region: Greenland 

Duration: approx 2 years (September 2020 - November 2022)

Collaboration partner: Technical University of Danmark (DTU)

Introduction

The mass of the Greenland Ice Sheet is steadily decreasing, and the ice loss has accelerated over recent decades. About 60% of the mass loss is attributed to changes in surface mass balance, e.g. the difference between snowfall and surface melt, with the rest is attributed to dynamic mass loss caused by changes in ice flow towards the ocean. 

Added flux of meltwater with nutrients, sediments, and solutes from the ice sheet to the ocean not only alters the local marine ecosystem, but also affects deep-water formation and potentially the global ocean circulation system. Finally, the mass loss from the ice sheet increases the global sea level, with significant ecological and societal consequences. 

The overall aim of the ESA-funded POLAR+ 4DGreenland project is to advance the current state of knowledge on the hydrology of the Greenland Ice Sheet, by capitalising on the latest advances in Earth Observation (EO) data.

Results

Within the project S&T has evaluated the capability of AI (more specifically Deep Learning) to learn to identify supraglacial lakes using optical imagery (Sentinel-2), with the hope to open up to the opportunity for large scale autonomous mapping programs in the future. Here is a sneak peak of the results: 

Interested? We could make something cool for you too, just contact us!