Research Area 3: Emerging Technologies Enabling Advanced Marine Science

Technological and methodological innovation is the foundation for scientific advancements, accelerating and improving our ability to observe, model, and predict the ocean. Novel autonomous observing platforms, integrated remote sensing technologies, innovative molecular biological approaches, significantly growing computational resources, challenges and opportunities of an exponentially growing data volume, or the use of machine learning and artificial intelligence, demonstrate that research in methodology and technology will in the future create an enormous benefit for environmental science. Research area 3 therefore addresses the central challenge of knowledge generation through integrated observation, advanced modelling, and intelligent data use. By linking technological innovation with scientific application, RA3 provides the enabling framework that underpins progress in RA1 and RA2 and strengthens IOW’s role in international observing systems, modelling initiatives, and data infrastructures. Its research activities are consolidated in three highly interconnected and interdisciplinary research themes:

• New observation technologies
• Scientific model development
• Data integration

Here is a short list of the most recent publications of the research area and recently funded projects. The full list can be found under RA3 publications and RA3 Projects, respectively.

New observation technologies

Despite major advances in observational technologies, our ability to measure key ecosystem processes and budgets remains limited. Progress requires not only improved measurement tools that perform reliably in challenging environments, but also smarter, adaptive strategies to ensure observations are made at the right place and time. Expanding the range of observed parameters, including emerging biological indicators based on environmental DNA, is equally important for a more integrated ecosystem understanding. Our approach combines the development of new sensing technologies, including remote optical sensing, improved sampling and analytical methods, and optimized observation frequency, resolution, and spatial coverage. Data processing procedures and calibration routines that ensure observation quality and address measurement uncertainties in a transparent and reproducible way will be automated for a wide array of parameters. Furthermore, suitable integration mechanisms for data from heterogeneous sources will be designed. A key goal is to reduce critical gaps in current observation systems. Smart combination of data streams from different European research infrastructures (ICOS, Euro-ARGO, Copernicus) with tailored data from individual projects will be a crucial component of this effort. Additionally, the systematic integration of machine learning and modelling into observation systems represents a key innovation pathway.

In particular for the dynamic shallow water coastal zone, technological advances are needed to capture its temporal and spatial variability. There, physical, chemical, and biological processes interact across scales from micrometers to basins and from seconds to seasons and strong waves, currents, and limited accessibility make the measurement of such processes extremely challenging. Long-term observations therefore require robust instrumentation capable of withstanding storms, while intense summer biofouling further complicates sustained deployments. A central achievement of RA3 is the establishment of an integrated observing systems for the coastal ocean. Specifically, a flexible, online-capable mooring system was developed within the S2B project designed for harsh near-shore conditions.

Scientific model development

To improve the representation of Earth system complexity in Baltic Sea models, we develop an efficient, high-resolution multi-model coupling framework that consistently captures ocean processes and their interactions with the atmosphere, land, and hydrology. Building on modelling systems co-developed at IOW (ERGOM, GOTM, GETM, IOW-ESM), this approach integrates advanced numerical methods, analytical theory, and process-based understanding into a flexible regional Earth system modelling framework. A key focus is the development of multi-scale coupled physical–biogeochemical models that resolve interactions between high-resolution coastal zones and lower-resolution open-sea domains. This includes advanced nesting and grid-refinement techniques, enabling explicit simulation of coastal dynamics and their feedbacks on basin-scale processes. Enhanced representations of turbulence closure, surface-wave effects, and boundary-layer dynamics, together with improved descriptions of water mass transformation and vertical mixing, will further strengthen process realism. A modular coupling architecture, featuring exchange-grid flux calculations and interactive boundary conditions, ensures numerical robustness, flexible component integration, and consistent atmosphere-ocean interactions, while supporting regional dynamical downscaling of global climate projections. Recently introduced machine-learning-assisted calibration techniques further systematically integrate in-house observational systems into the tuning process, improving the accuracy and consistency of the coupled modelling framework. In addition, RA3 incorporates innovative multivariate statistical and machine learning approaches to better represent ecosystem structure and function in both benthic and pelagic environments, explicitly accounting for strong spatial and temporal heterogeneity. These developments are supported by dedicated infrastructure for software and data management, enabling seamless integration of observational and model data. This will enhance predictive capabilities, improve scenario projections, and accelerate ecosystem model development. At the same time, model outputs will inform the identification of observational gaps and the optimization of monitoring strategies, fostering a tightly coupled, iterative advancement of observation and modelling systems.

Data integration

Over the past decade, the amount and diversity of scientific data being produced in all disciplines of IOW’s research increased significantly. This trend is expected to continue with the development of new observation technologies and modelling approaches, allowing for data science applications that leverage data sets available at IOW and elsewhere at various spatial and temporal scales within a joint analytical framework across disciplinary boundaries. We will focus on the adaptation and implementation of targeted and exploratory data science applications, like machine learning and artificial intelligence and on the development of approaches for the re-use (i.e. discovery and integration) of heterogeneous data sets in this context. These approaches will exploit such data resources for the purpose of pattern recognition to provoke innovative research questions (and solutions) that only become apparent due to the available data volume. This includes also the development of decision support tools and AI-supported early warning systems for coastal health risks . A key element for the success of such research endeavors is a common information infrastructure for the FAIR management of scientific data. Building on the current research data infrastructure, future developments will address the challenges related to (i) the integration of data generated by novel technologies, including ‘omics methods, into existing and new data structures and scientific procedures, (ii) data quality assurance including provenance information and (iii) the ability to find and re-use existing data, which requires intelligent storage strategies as well as richly described contextual parameters and metadata. Such integrated data infrastructures and FAIR and sustainable data workflows simplify both institutional data management as well as the archiving and provision of data and data products to the scientific community and general public. Active membership within the NFDI consortia NFDI4Biodiversity and NFDI4Earth will enhance large-scale data integration, metadata harmonization, and accessibility, ensuring that heterogeneous datasets can be efficiently reused for interdisciplinary research.