Host institution: Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences (CAS), Beijing, China
Primary research location: Laboratoire des Sciences du Climat et de l’Environnement (LSCE), France
Duration: 2 years, with possibility of extension
Supervisors: Dr. Yuanyuan Huang, Dr. Yang Su, and Dr. Philippe Ciais
We invite applications for a Postdoctoral Researcher to develop advanced remote-sensing and artificial-intelligence approaches for mapping soil organic carbon (SOC) stocks and their changes across Europe. The position will be hosted by the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, while the successful candidate will have the flexibility to choose to conduct most of the research at LSCE in France.
Research Background
Soils constitute one of the largest terrestrial carbon reservoirs, and changes in SOC can substantially influence the land carbon balance and climate feedbacks. However, the spatial distribution and temporal dynamics of SOC remain poorly constrained because field observations are spatially sparse and existing products largely focus on static carbon stocks.
The project aims to exploit recent advances in Earth observation, deep learning, and computer vision to improve the spatially explicit monitoring of SOC and its temporal changes across Europe. By integrating satellite observations, soil measurements, environmental covariates, and long-term geospatial datasets, the project will investigate where SOC is increasing or declining and the climatic, ecological, and land-use processes associated with these changes.
Main Tasks
The successful candidate will:
- Develop high-resolution SOC mapping frameworks using multi-source remote-sensing and environmental data;
- Apply machine learning, deep learning, and advanced computer-vision methods to characterize SOC spatial variability and temporal change;
- Integrate satellite time series, soil observations, climate, vegetation, topography, and land-use information;
- Quantify model uncertainty and evaluate SOC estimates against independent observations;
- Analyze the drivers of SOC gain and loss across European ecosystems and land-use systems;
- Prepare high-quality scientific publications and research datasets.
The methodological direction will remain flexible, with opportunities to explore emerging approaches such as foundation models, representation learning, spatiotemporal modelling, and multimodal data fusion.
Candidate Requirements
Applicants should hold, or be close to completing, a PhD in remote sensing, geography, geoinformatics, environmental science, ecology, biology, soil science, Earth system science, computer science, or a related field.
Applicants are expected to have:
- Experience in remote sensing and/or geospatial data analysis;
- Experience with machine learning and/or deep learning;
- Strong programming skills, preferably in Python;
- Experience processing large spatial or Earth observation datasets;
- Good scientific writing and communication skills in English;
- Ability to work independently and collaboratively in an international research environment.
Experience in satellite time-series analysis, PyTorch, Google Earth Engine, large-scale geospatial computing, soil carbon research, or carbon-cycle science would be an advantage.
Application and Contact
Interested candidates are encouraged to send a CV and a brief statement of research experience and interests to:

