🎓 CoMet Training Workshop – ARIA workshop 2026
📅 Date: Wednesday 15th April
🕰️ Time: 09:15–12:30
📍 Location: National Physical Laboratory
On this page you’ll find links to the training materials.
This session covers key concepts around uncertainties, with guided exercises using the CoMet toolkit. You will:
- Gain a conceptual overview of uncertainties in Earth Observation data processing.
- Learn how to use the CoMet tools in practical workflows.
- Apply methods through interactive notebooks hosted on Google Colab.
- 📃 Uncertainties 101: A short introduction to key concepts, why they matter, and how CoMet helps with uncertainty handling.
Exercises
Links and descriptions of the hands‑on exercises for this training session.
Exercise 1: Uncertainty Propagation Basics (Click here to open exercise)
• Simple measurement functions
• Manual specification of input uncertainties
• MC and LPU uncertainty propagation methods
Exercise 2: Error Correlation in EO Datasets (Click here to open exercise)
• Simple measurement functions
• Propagating error correlation information
• Random and systematic uncertainties
Exercise 3: From Spectrometer Measurements to NDVI with Uncertainty Propagation (Click here to open exercise)
• Spectrometer example from demo and uncertainty tree diagram session
• Propagate uncertainties from raw data, to reflectance and NDVI
• Random and systematic uncertainties
Exercise 4: Multi-Dimension Datasets (Click here to open exercise)
• Use obsarray to store error-correlation information for multi-dimensional measurement datasets - such as from Earth Observation.
• Propagate uncertainties from these datasets through measurement functions using punpy.
(if time allows) Exercise 5: Work on your own use case of HYPERNETS (Click here to open exercise)
• Get familiar with a sample EO data (HYPERNETS)
• Use the previous exercises to add uncertainties to the HYPERNETS data processing chain
• Generate and interpret uncertainty-aware outputs
✅ Solutions
Exercise solutions will be added after the workshop.