UROP Project
Scalability of a Random Forest-based shallow landslide hazard mapping approach
Contact
Key Info
Basic Information
- Project Offer-Number:
- 1195
- Category:
- RWTH UROP, UROP Sustainability, UROP International
- Field:
- Applied Geosciences
- Faculty:
- 4
- Organisation unit:
- Chair of Methods for Model-based Development in Computational Engineering
- Language Skills:
- English
- Computer Skills:
- Programming, preferably in Python; preferably experience in git
MoveOn
The probability of occurrence of shallow landslides in space and time can be effectively communicated through hazard maps. Hazard mapping can be conducted on different spatial scales, like regional, local or global scale. Each scale has different requirements and challenges that need to be considered and taken care of. The scalability of a shallow landslide hazard mapping workflow will be investigated in this project. An already existing flexible workflow can be used for this purpose. A procedure will be defined on how a high-quality and high-resolution hazard map on different spatial scales can be created by combining different landslide inventories. Transferability of an existing shallow landslide prediction model will be assessed to check if the combination of various landslide inventories from different regions also allows reliable hazard prediction in areas where inventories are not available, and therefore are often neglected.
Task
• Research on available landslide inventories • Preparation of the inventories for use in hazard mapping workflow • Investigation of the predictive quality on different spatial scales
Requirements
• Programming experience, preferably in Python • Basic geoscientific knowledge • Preferably basic data science knowledge