MSc thesis assignment: Machine Learning Fusion of Sentinel 2 and ICESat-2 data for Satellite Derived Bathymetry (SDB)

Contributing to one of the most innovative developments in coastal mapping: creating a fully spaceborne bathymetry solution that fuses ICESat-2 LiDAR with Sentinel-2 multispectral imagery — that is what you will do as a graduation intern at Haskoning. In this assignment, you combine scientific research with direct engineering impact: developing a scalable tool for mapping shallow-water depths worldwide, you work at the forefront of a promising methodology and deliver a result our team can put to immediate use.

What you will do as a MSc graduation intern

Haskoning already has strong in-house expertise in Satellite-Derived Bathymetry (SDB), but we would like you to focus on a promising new direction in this field. This graduation project centers on a machine-learning approach that predicts shallow-water bathymetry by training Sentinel-2 imagery with depth points derived from ICESat-2 photons. This emerging method offers the potential to produce accurate depth estimates without relying on costly in-situ surveys and is increasingly supported by recent scientific literature.

Main responsibilities:

• Investigate the scientific background of SDB, ICESat-2 photon bathymetry, and ML-based optical depth retrieval.
• Develop machine-learning models to predict bathymetry from multispectral inputs, informed by literature such as Random Forest, SVM, and Neural Networks.
• Produce insights on the applicability, accuracy limits, and scalability of the method for engineering use cases.

We encourage you to bring your own academic perspective and critical thinking to this project. As a starting point, we suggest the following direction:

How can Sentinel-2 multispectral imagery and ICESat-2 LiDAR bathymetry be fused using machine learning to produce accurate, scalable, and engineering-ready Satellite-Derived Bathymetry?

This could include:

• Implementing a workflow which can extract and process ICESat-2 data and Sentinel-2 data; Haskoning already has scripts for this that could be used for inspiration.
• Training ML regression models to capture the relationship between optical reflectance and depth.
• Testing the method in one or more coastal pilot areas and assessing environmental factors influencing accuracy, such as turbidity and bottom type.
• Delivering a prototype SDB tool or workflow that can be scaled across projects.

The deliverable is both a strong scientific thesis and, if possible, a practical method that our company can adopt for future use.

Where you will work

At Haskoning, you will join an independent, employee-owned international consultancy that combines engineering, design, and consultancy services with software and technology.

As our new graduation intern, you will become part of the team Coastal & River, Dynamics & Design, which falls under the Advisory Group Maritime and Renewables. Our team combines deep expertise in coastal and river processes with practical engineering design to deliver robust solutions for flood risk management, coastal protection, constructing marinas, ports, and nature-based solutions for shoreline stability. In addition to modelling and design, the team develops and maintains advanced tooling, for example Python-based workflows and reusable analytical frameworks which support efficient and standardized analyses across our company. You will work alongside colleagues who combine coastal and river engineering knowledge with remote sensing and geospatial analytics. It is an ideal environment for bridging science and engineering.

What you bring

You are curious, analytical, and independent, with the drive to carry out a scientifically rigorous thesis that makes a real-world impact. You enjoy taking ownership of an innovative topic, coding data analyses, exploring complex geospatial datasets, and translating scientific insights into practical, usable tools.

Requirements:

• You are a Master’s student in a relevant field, such as Civil/Coastal Engineering, Physical Geography, Geospatial Sciences, Remote Sensing/Earth Observation, or Data Science.
• Familiarity with remote sensing concepts, or eagerness to learn this new field, including reflectance, atmospheric correction, and optical properties of water.
• Basic understanding of machine learning regression methods and their role in scientific modelling.
• Interest in LiDAR, spectral imagery, and coastal environments, or willingness to learn quickly.
• Ability to work systematically, take ownership, document results clearly, and manage a full research cycle, or interest in learning these skills.
• Enthusiasm for satellite data and leveraging it to leave a positive sustainable impact on the world.

What you can expect from us

During your internship at Haskoning, you’ll have the opportunity to grow professionally in an inspiring work environment. We offer support, guidance, and attractive conditions to help you get the most out of your internship:

• Internship allowance of €750 per month (based on a 40-hour workweek).
• Personal guidance from a dedicated mentor.
• Opportunity to contribute to impactful projects.
• Informal and inclusive work culture with social and sports activities.
• Access to Young Haskoning, our network for young professionals, where you can join inspiring events, workshops, and informal gatherings with colleagues.