Analyzing and comparing ecosystem assessments of reservoirs
A data-centric approach
Publication date
2026-05-07
Document type
Konferenzbeitrag
Author
Kühnert, Christian
Wunsch, Andreas
Ho, Johannes
Holzer, Chiara
Hügler, Michael
Organisational unit
Fraunhofer Institute of Optronics, System Technologies and Image Exploitation (IOSB)
Department Water Microbiology, TZW: DVGW-Technologiezentrum Wasser
Publisher
Universitätsbibliothek der HSU/UniBw H
Book title
Machine learning for cyber physical systems : proceedings of the conference ML4CPS 2026
First page
19
Last page
29
Peer-reviewed
✅
Part of the university bibliography
Nein
File(s)
Language
English
Keyword
Reservoir ecosystem assessment
Microbial fingerprinting ·
Biochemical monitoring
Data-driven analysis
Abstract
Surface water reservoirs play an essential role in drinking water production, hence monitoring the reservoir behavior is crucial. However, monitoring a reservoir ecosystem is hindered by heterogeneous data collection methods and the sometimes complex interpretation of the different measurements. Therefore, this paper first presents a unified data schema for data collection and then showcases analyses such as complementary fingerprinting methods: Microbial fingerprints are calculated using the Shannon and Pielou’s evenness indices from 16S rRNA data, and a biochemical fingerprint is derived using an autoencoder applied to physicochemical parameters. When applied to an actively managed agricultural-forested reservoir and a natural montane-forest reservoir, distinct fingerprint signatures were revealed. Specifically, an elevated reconstruction error and changes in biodiversity coinciding with a cyanobacterial bloom were observed in the managed system, while the montane reservoir maintained stable patterns with natural seasonal dynamics. This proposed framework enables early detection of ecological shifts and supports cross-reservoir assessments for ecosystem monitoring.
Description
This contribution is part of the conference proceedings, which are licensed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/)
Version
Published version
Access right on openHSU
Open access