FAIR from the Start: Managing Research Data for the LAFI project

Fall School 2025 07th October 2025 LAFI Data Management Terminology, Standards, Process, Tools Fall School 2025 I 07.10.2025 Fall School 2025 07th October 2025 FAIR from the start: Managing Research data 2 1. Standardization 2. Storage, archiving and publication 3. Web-portal for direct access 4. Documentation & tutorials Aims Fall School 2025 07th October 2025 3 LAFI Research Data Management 1. Data Management Plan. 2. Follow FAIR principles from start. 3. Use community standards & metadata. 4. Deposit in trusted repositories with DOIs. 5. Ensure long-term preservation. 1. Improved collaboration and scientific discovery. 2. Data Re-usability, Reproducibility, Verifiability, Citability. 3. Readiness for tool development & AI/ML, LLM applications. Institutional push Community pull Motivation Proposals, Publications & Preservation Fall School 2025 07th October 2025 Frameworks & Tools Untangling structure and clarifying definitions 4 Caveat: CF guidelines, obs4MIPs utilities, and integration with NFDI4Earth evolve constantly. Fall School 2025 07th October 2025 LAFI + NFDI4Earth = FLAIR First-Class Data Services for FAIR Land-Atmosphere Interaction Research 5 Fall School 2025 07th October 2025 FRAMING LAFI RDM Key Takeaways 6 1. LAFI RDM complies with FAIR principles. 2. CF is the accepted community standard for preparing standardised datasets. 3. CF + additional requirements = obs4MIPs →for publishing data through obs4MIPs/ESGF platform. 4. LAFI RDM outputs will be captured through the NFDI4Earth service portfolio. LAFI DMP Fall School 2025 07th October 2025 Nuts & Bolts: netCDF Unpacking the key components 7 Fig: Idealised representation of key netCDF components netCDF = NETwork Common Data Form 1. Self-describing, portable, compact, multidimensional, arrayoriented scientific binary data format. 3. Metadata: •Global attributes →describe the whole dataset. •Variable attributes →describe each data variable •Coordinate attributes →describe axes like time, latitude, longitude. 2. Shape information: •Dimension →defines the shape of variables. •Coordinate variable →values/labels defining the physical context 4. Thus, creating netCDFs from observation and modelling data requires defining these 5 elements from input Fall School 2025 07th October 2025 Nuts & Bolts: Climate Framework (CF) Ensuring CF compliance 8 Fig 1: Concise document which specifies requirements and recommendations. 1. CF makes netCDF files self-describing, machine-readable, and interoperable across tools and projects. 3. obs4MIPs imposes additional restrictions on attribute definitions, including file nomenclature. 2. Flexible, aside from a minimally required defined set of attributes. Scan 4. Creation of CF/obs4MIPs compliant netCDFs requires only metadata to be defined. Fig 2: Idealized example of minimally CF conformal netCDF dataset with quality flags Fall School 2025 07th October 2025 Conversion walkthrough: Doppler Lidar (DL) A quick look at the data 9