Human-induced environmental alterations via climate change, land use change, or eutrophication have pushed freshwater ecosystems beyond crucial thresholds, resulting in widespread oxygen depletion – anoxia. The consequences include biodiversity loss, water quality deterioration, and toxic algal blooms. To explore past eutrophication events over the last 12 000 years, we study lake sediments which contain a wealth of information about past environmental conditions. We scan sediment cores using high-resolution hyperspectral imaging (HI) and identify algal or biogenic pigments through their absorbance. Pigments in turn reflect changes in aquatic productivity and anoxia in the lake over time. HI produces large datasets whose analysis requires a complex computational workflow. To replace slow, proprietary software currently used for that task, the Institute of Geography (GIUB) and the Data Science Lab (DSL) have been developing an alternative Python-based software leveraging open-source libraries (dask, scikit-learn, spectral etc.) and data formats (zarr). The entire workflow is implemented as a set of interactive plugins for the fast multi-dimensional visualization software napari. In an effort to promote open-science and foster data exchange, the software is being made available on the data science platform Renku in collaboration with the Swiss Data Science Center. The project was presented at this years Data4Sciences Conference in Bern.
