Research Topics
Natural Science Topics
As part of three driving natural science questions, DSM is engaged in:
-
Mapping and evaluating the impact of marine munition on the ecosystem and in parallel investigate social concerns and economic possibilities. Since 2016, DSM is strongly involved in munition related questions and established itself as an expert on this on national and international level. Currently marine munition is the dominant research topic of the group, with one large projects on national (CONMAR as part of the DAM sustainMare Mission) and one - maybe two - large EU project (MMinE-SwEEPER; CAMMera – in grant agreement negotiations) being coordinated by us. The group is also partner in several other munition projects (see Projects).
-
Monitoring of deep-sea mining activities with focus on plume monitoring, benthic megafauna and resource estimation. This research is at the core of DSM expertise and has been worked on in the frame of EU (MIDAS) and the JPIO projects Mining Impact I & II. Much of the technological developments are driven by this topic and were important for establishing the group. The group is also part of the MiningImpact III project and it has several publication on this topic (Gazis et al., 2025; Purkiani et al. 2022; Haalboom et al., 2022; Purkiani et al., 2021; Baeye et al., 2022).
-
Quantitative transport of methane bubbles from the seafloor into the atmosphere. This research dates back to the PhD times of Jens Greinert; in the last years the activities around this topic dropped because of upcoming other topics and some people left the group and found a new home at Ghent University (Peter Urban) and the company Develogics GmbH (Mario Veloso, his PhD thesis to the topic). However, tools and expertise are still within the group.
-
Morphology and underlying physical processes of seafloor volcanism and hydrothermalism. Volcanic and hydrothermal processes and their morphological seafloor expressions are complicated and quickly vary in time and space. They form the substrate for unique chemoautotrophic habitats, and have profound impact and importance to humankind both as a hazard and a potential natural resource. In collaboration with colleagues at RD4 and international partners we investigate both constructive and erosive processes connected to Island Arc, Ocean Island and Mid-Ocean-Ridge volcanism.
Engineering Science Topics
In addition to natural science we also work in engineering science topics in the field of marine robotic. Both topics below are closely linked to mapping marine munition and tracking anthropogenic sediment plumes; as part of both topics, software code is published as well:
- SMART magnetic AUVs. This research is currently undertaken within the SAM project, run by Marc Seidel. The project aims at applying on-line magnetic sensor data processing (integrated into our GIRONA 500 AUVs happed during the past BASTA project). In SAM the AUV gets trained to automatically define an ideal survey path for detailed magnetic mapping of magnetic objects, after data of an initial preliminary-survey have been processed online. Such a smart and adaptive survey approach gets better data in a shorter time.
- SMART AUVs and AUV fleets for tracking ocean interfaces (e.g. thermoclines) or sediment plumes. Enabling AUVs to truly autonomously track such marine features is the needed next step in autonomy of such robots. Luckily, DSM works very close with the AUV Team at GEOMAR in using the GIRONA 500 AUVs for our engineering science. Benedict von See is currently finishing his PhD on this (hopefully defending in summer 2025).
-
Modular multi-body remotely operated vehicles. Together with the RD4 Oceanic Volcanism team, we develop MOMO, a multibody remotely operated vehicle consisting of a tow-sled and a small 6000m capable ROV. The goal of the project is to fill a niche left open by both work-class ROVs and AUVs when it comes to a quickly deployable, inexpensive, deep diving platform that offers excellent real-time situational awareness, nested-scale multimodal surveying and high-volume sampling capacities.
-
Economies of scale in optical seafloor surveying. Especially our engagement in the studies of seafloor volcanism require high-resolution, six-degree of freedom mapping techniques that also scale up to large, holistic surveys of entire geologic outcrops across hundreds of meters. Optical (photographic) methods are the choice but they do not readily scale up such as on land or in the air. We therefore design our own camera and lighting hardware combined with state-of-the-art underwater navigation tightly integrated with the robotic platforms of our choice, to tackle the challenge to be precise and accurate across a range of scales.
Data Science, Data Management and Software Development Topics
Data science, data management and software development are very important aspect of the group as such work often links, at least supports our natural and engineering science endeavours. In this respect we work on:
-
Automated Image analyses for habitat mapping purposes. When at sea, we typically acquire several 10,000 images and lost of video footage which needs to be quantitatively analysed. We develop and apply workflows for semi-automated image analyses that are tailored to be used in our habitat mapping purposes and will soonish be integrated in AUVs to enable autonomous adaptive mission execution of AUV surveys. Mbani Benson recently finished his PhD on this aspect.
-
Spatial, quantitative and predictive habitat mapping and resource assessments. For all of our naturel science question, spatial documentation of acquired data and the wish to extrapolate spot observations from e.g. images or sediment cores into space for predictive analyses. AI techniques as Machine Learning and Deep Learning are applied similarly to the automated image analyses. A strong focus is on using such techniques correctly, knowing about their limitations and applicability using detailed statistical analyses of the training data. Iason Gazis recently finished his PhD on this topic in the frame of Mn-nodule resource assessment.
-
Data management as essential backbone for doing good environmental science. Acquiring TB of data, good data management with defined and followed workflows is essential. Already some years ago the DSM group established data management procedures for AUV images and more recently took over the task of leading the data management ambitions of the DAM Mission sustainMare. In this respect we also established several GeoNode instances at GEOMAR for sharing unpublished project data. We do this to allow access within and during project time and have well curated data available for the final data publication e.g. in PANGAEA. DSM has and will continue publishing its data and software code, following FAIR principles.
-
Interactive 4D Visualisation of very large data sets. As part of the HGF funded and internal project Digital Earth (that DSM coordinated) we developed the Digital Earth Viewer for displaying and exploring large 4D data sets (properties of CTD casts, ocean model NetCDFs, GeoTiffs of photomosaics or bathymetry, …). Check it out and let us know any questions, ideas and … bugs.
-
Ocean Machine Vision. In the past, Kevin Köser also worked at DSM and specifically addressed the important aspect of ocean machine vision as part of his research and Emmy-Noether group DEEP QUANTICAMS. Kevin now got a professorship position at CAU for Marine Data Science, we are still cooperating. Much of the generated knowledge is almost daily used by DSM members at sea and in the lab, and is also used by other GEOMAR groups, i.e. the AUV-Team.
-
Visual data exploration using spatially immersive environments. Since 2007, Tom Kwasnitschka has been conducting research and development in the field of spatially immersive domes as a tool for geoscientific research. The current and fourth iteration of GEOMAR’s immersive visualization laboratory, ARENA2, is an internationally competitive and highly versatile visualization environment across all disciplines of ocean science but we focus on virtual geoscientific fieldwork as in the most recent Mardata project of Armin Bernstetter.
Social Sciences and Related Stakeholder Interaction Topics
Social sciences and related stakeholder interaction are a stronghold in our munition projects. We regularly engage with stakeholders from state and national politics and ministries, with people from the industry (with whom we are also working in several projects) and also the public. Our social scientific questions deal with:
-
Testing and advancing stakeholder interactions and deriving information and data from such interactions. Several workshops were needed to learn about indicators for industry and official EOD services, and conceptualising the workshops and analysing the outcomes is part of the scientific work in the CONMAR project.
-
Multi-Criteria analyses for decision making. Concluding which munition pile to remediate first, or if remediation is at all cost-effective, does need an integrated approach of the pros and cons for very different points of view. Such a multi-criteria analyses is currently further developed in the second phase of the CONMAR project, aiming at providing a guidance approach for future national ambitions to remediate (remove and dispose) marine munition from German waters.
-
Provenance in immersive scientific visualization workflows. Deep-sea and seafloor research are carried out remotely (using robots) and they are often time critical, let alone expensive. It is therefore imperative to gain good spatial understanding and situational awareness of the study area, and to agree upon findings in a scientific peer group of researchers. Immersive visualization is thought to have a high potential to catalyze these episthemic processes, but they leave no artifacts that would support the rigorous FAIR scientific documentation process. We therefore work on prerequisites, challenges and opportunities of a visualization work flow that can be traces, exported, and that produces artifacts suitable to be published in supplements and repositories.
-
Professional and societal impact of immersive visualization. Together with several partners at the CAU Kiel (at the Center for Ocean and Society, the Faculty of Engineering, but also the Faculty of Economics), we investigate how immersive visualization may enhance the collaboration and interaction among scientific teams. We also believe that a carefully designed visual language is equally suitable for lay audiences in the field of informal education. In the context of the ARENA2 lab, we therefore produce and disseminate immersive documentaries (e.g. in the fulldome and VR formats) on geosciences and marine science narratives in general, and conduct research on the reception of these media.
