Job offer
Role details
Job location
Tech stack
Job description
Job description:PhD position in Deep Learning for Multi-Decadal Remote Sensing of Arctic Lake Changes (m/f/d/x)
Background Arctic permafrost soils contain the largest terrestrial pool of organic carbon (C). Rapid thaw of ice-rich permafrost quickly mobilizes permafrost carbon on short time scales. A large uncertainty persists in projections of C release from abrupt permafrost thaw and climate feedback estimates. Key challenges include assessing permafrost thaw rates under diverse and changing environmental conditions on a panarctic scale, analyzing impacts of varied thaw processes on the heterogeneous permafrost C stock, integrating and scaling such dynamics using process models, and projecting future permafrost C trajectories, climate feedbacks, and remaining C budgets with Earth System Models.
PeTCaT, a new international 5-year research endeavor funded by Schmidt Sciences within their Virtual Institute of Carbon Cycle (VICC) program, brings together partners from Germany, Canada, USA, Sweden, and The Netherlands, and will tackle these challenges by combining expertise on permafrost, soil C and GHG biogeochemistry, remote sensing, process modeling, and global climate modeling. We will address soil carbon and ground ice stocks, rapid thaw process dynamics, the reactivity of rapidly thawing organic C, the consequences for greenhouse gas fluxes, and the global consequences.
At AWI, you will work in a team consisting of senior scientists, PostDocs, as well as another PhD student hired through PeTCaT. You will contribute to Work Package 1 (Remote Sensing of Rapid Thaw Processes). This position is embedded in the Permafrost Remote Sensing group of Prof Dr Guido Grosse, who will be jointly supervising the PhD together with Dr Ingmar Nitze.
Your Tasks
- Project Scope: Conduct a comprehensive, observation-based long-term (since mid-20th century), panarctic analysis of permafrost thaw dynamics within Arctic Permafrost regions, specifically investigating how lake-rich landscapes have evolved over multi-decadal time scales
- Imagery Handling: The work will involve the acquisition, preprocessing, and analysis of historical reconnaissance Earth Observation (EO) imagery, primarily from sources like CORONA, HEXAGON, and potentially aerial imagery, focusing on creating a robust preprocessing pipeline
- AI/ML Feature Extraction: Utilizing Artificial Intelligence/Machine Learning (AI/ML) models, the candidate will automatically identify and extract key permafrost landscape features from the processed historical imagery, with a particular focus on recognizing and mapping features like lakes, but potentially extending also to other features such as thaw slumps
- Workflow Adaptation: The role requires the development and adaptation of workflows to extend the analysis to recent imagery, with the goal of quantifying the previously poorly understood long-term dynamics of permafrost disturbances across the entire circumarctic region
- You will ...
- work in an international team of experts, and use state-of-the-art computational resources, including HPC
- collaborate closely with international team members and project partners across work packages
- present results in scientific meetings and conferences as well as through publications in scientific journals
- be expected to write a cumulative doctoral thesis consisting of at least three research manuscripts
Requirements
- Master's degree in remote sensing, computer sciences, or closely related fields
- Experience in handling, processing, and analyzing large amounts of data, especially large-scale remote sensing datasets
- Demonstrated very good experience in programming and automation of data processing/analysis
- Demonstrated knowledge of AI processing methods and dataset requirements
- Ability and willingness to publish research results in peer-reviewed literature, and present your work in national and international conferences
- Ability and willingness to work focused, independently, and cooperatively in a large team and with international partners
- Willingness to travel internationally and to attend virtual project meetings across different time zones
- Very good englisch knowledge (approximately equivalent to CEFR level C1)
Preferred Qualifications and Skills
- Experience with High Performance Computing (HPC) environments
- Experience in automated image co-registration and rectification pipelines
- Experience with permafrost and its processes and/or landscape disturbance processes
- Experience presenting scientific results at an international level (papers, conferences)
- Advanced software engineering and programming skills