h1

%0 Thesis
%A Simson, Anna
%T Reusability in cryospheric sciences: fundamental concepts and case studies
%I Rheinisch-Westfälische Technische Hochschule Aachen
%V Dissertation
%C Aachen
%M RWTH-2025-06267
%P 1 Online-Ressource : Illustrationen
%D 2025
%Z Veröffentlicht auf dem Publikationsserver der RWTH Aachen University
%Z Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2025
%X Reusability of research products in form of digital resources such as data sets and modeling software is of utmost importance to science. It makes research transparent, sustainable, and accelerates the scientific endeavor. Over the past few decades, repositories for sharing research products have been established, and standards for the consistency of digital representations of resources have been developed. The aim is to facilitate the reuse of data and software. Since 2016, guidelines for Findable, Accessible, Interoperable and Reusable research products have been formulated, widely known as the FAIR Principles. The FAIR vision describes a future automated state in which humans assign tasks like data analysis and integration to machines. In light of these FAIR developments, one might think that reuse scenarios ranging from model validation based on existing measurement data, to model coupling using existing modeling software are smoothly executable. In reality, despite FAIR's focus on machines, reuse is still carried out mostly manually, i.e., humans find, integrate, and analyze existing research products, and reuse efforts remain time consuming and ineffective. Cryospheric sciences, which study all occurrences of frozen water on Earth, would particularly benefit from highly reusable research products. Measurements of a constantly shrinking cryosphere cannot be repeated, in-situ measurements during polar expeditions are costly, and cryospheric modeling software must be easily coupled with models from other geoscience domains. A holistic understanding of the cryosphere is essential to predict and mitigate the impacts of climate change. This requires the combination of cryospheric data and models at different scales, and their seamless, interdisciplinary integration with data and models from other geoscience domains. Therefore, the reusability of cryospheric research products and their current leverage in reuse scenarios needs to be investigated in depth. In this thesis, I investigate the reusability of cryospheric research products based on two case studies. Cryospheric Case Study I is a physics-based process model for coupled water vapor transport and settling in the snowpack. The model is characterized by modularity and extensibility. The reuse potential of the model's software is highlighted at the example of a real-world reuse scenario in form of a model comparison. Cryospheric Case Study II describes the approach followed to compile sea ice core measurement data into a comprehensive and analysis-ready database. This case study demonstrates the challenges encountered when manually harmonizing and combining distributed and heterogeneous data sets. Both case studies put a special focus on the transparency of the method and the reusability of the generated research products. The case studies effectively demonstrate that the reuse of cryospheric data and software is not trivial and not yet executable by machines alone. The sharing of cryospheric research products follows individual preferences, and the products lack standardization of data and metadata as well as quality information, both of which affect their understandability and interoperability. The case studies demonstrate: (1) the need to clearly communicate reuse needs in the form of reuse scenarios; (2) the large discrepancies between the many challenges faced when manually reusing a resource and the FAIRness of the resource, which reflects its machine reusability; and (3) the challenges experienced in manual reuse will be inherited by machines. In the future, reuse scenarios should be documented more effectively to represent the reuse perspective in a systematic way. A major focus should be on improving reusability with transparent and comprehensive documentation of data and software with metadata, the development of community-agreed standards for terminology and formats, and the prioritization and documentation of data and software quality. Such developments will benefit manual reusers, support the development of technologies that enable autonomous reuse, and it will facilitate the combination of small, heterogeneous, and distributed data sets. To automate reuse scenarios, future research should specifically investigate the application of large language models, including exploring their limitations.
%F PUB:(DE-HGF)11
%9 Dissertation / PhD Thesis
%R 10.18154/RWTH-2025-06267
%U https://publications.rwth-aachen.de/record/1015222