Open literature for interoperability in climate & atmospheric sciences

An annotated reading map for seven workshop topics

4TU.ResearchData · Training resources

2026-10-16

00 · Foundations of interoperability

Guiding question: What does interoperable add beyond discoverable and downloadable data?

Start here · essential

• The FAIR Guiding Principles
  Why accessible data still need shared representations and meanings.

• Pythia Foundations
  Accessible, hands-on geoscience introduction to NetCDF, CF and Xarray.

Explore further · optional / instructor

• EOSC Interoperability Framework Instructor
  Connects technical and semantic dimensions to broader infrastructures.

• FAIR Data in the IPCC Assessment Cycle Optional
  Real climate-science example of sharing and reusing data.

Teaching connection: Assign the FAIR article and relevant Pythia chapters; use the IPCC example for motivation.

Full source annotations: 00 · Foundations of interoperability

• The FAIR Guiding Principles for Scientific Data Management and Stewardship — Essential. Introduces FAIR and I1–I3; use to discuss why findable and accessible datasets need not be interpretable or directly comparable. Source: https://www.nature.com/articles/sdata201618
• EOSC Interoperability Framework: Report from the EOSC Executive Board Working Groups FAIR and Architecture — Instructor. Provides a broader interoperability framework, including technical and semantic as well as organizational dimensions; not the same taxonomy as the course. Source: https://op.europa.eu/en/publication-detail/-/publication/d787ea54-6a87-11eb-aeb5-01aa75ed71a1
• Pythia Foundations — Essential. Open geoscience computational textbook; use selected NetCDF/CF and Xarray chapters instead of assigning the entire book. Source: https://foundations.projectpythia.org/
• Implementing FAIR Data Principles in the IPCC Seventh Assessment Cycle: Lessons Learned and Future Prospects — Optional. Shows why FAIR data stewardship and reuse matter in a real climate-science assessment workflow. Source: https://journals.plos.org/climate/article?id=10.1371/journal.pclm.0000533

These notes are visible in Reveal.js speaker view rather than the slide itself.

01 · Structural interoperability

Guiding question: How do shared data models and schemas help software understand a dataset’s layout?

Start here · essential

• The NetCDF User’s Guide
  Primary reference for dimensions, variables, attributes and storage.

• NetCDF in Python: From Beginner to Pro
  Guided inspection and creation of NetCDF and CF metadata.

Explore further · optional / instructor

• Xarray: N-D Labeled Arrays and Datasets Optional
  Links the NetCDF model to practical labelled arrays in Python.

• The CF Data Model and cf-python Instructor
  Clarifies how the CF data model relates to its NetCDF encoding.

• Frictionless Table Schema Optional
  Shows why CSV benefits from explicit data types and constraints.

• WMO Manual on Codes (GRIB/BUFR) Instructor
  Alternative structured encodings used in meteorology.

Teaching connection: Show how identical arrays may have different storage arrangements and metadata.

Full source annotations: 01 · Structural interoperability

• The NetCDF User’s Guide — Essential. The authoritative explanation of NetCDF data models, dimensions, variables, attributes, encoding, and chunking. Source: https://docs.unidata.ucar.edu/nug/current/
• NetCDF in Python: From Beginner to Pro — Essential. Practical, freely available course with NetCDF file inspection, CF metadata creation, cell methods, and file organization. Source: https://nordatanet.github.io/NetCDF_in_Python_from_beginner_to_pro/intro.html
• xarray: N-D Labeled Arrays and Datasets in Python — Optional. Peer-reviewed conceptual bridge between the multidimensional NetCDF data model and its use in Python. Source: https://doi.org/10.5334/jors.148
• A Data Model of the Climate and Forecast Metadata Conventions (CF-1.6) with a Software Implementation (cf-python v2.1) — Instructor. Excellent conceptual treatment of the distinctions and relationships between the NetCDF representation and the CF data model. Source: https://gmd.copernicus.org/articles/10/4619/2017/
• Table Schema Specification — Optional. Demonstrates why CSV alone lacks a fully specified machine-readable schema and how field types and constraints can be described. Source: https://specs.frictionlessdata.io/table-schema/
• Manual on Codes (WMO-No. 306), Volume I.2: Machine-Readable Codes — Instructor. Domain-specific counterexample illustrating structured meteorological encodings such as GRIB and BUFR beyond NetCDF. Source: https://community.wmo.int/site/knowledge-hub/programmes-and-initiatives/wmo-information-system-wis/about-manual-codes-volume-i2/latest-version

These notes are visible in Reveal.js speaker view rather than the slide itself.

02 · Semantic interoperability

Guiding question: How can scientists distinguish identical variable labels from identical scientific meaning?

Start here · essential

• CF Metadata Conventions 1.13
  Normative rules for coordinates, units, bounds and scientific interpretation.

• CF Standard Name Table
  Practice matching physical quantities with agreed names and units.

• Semantic Interoperability in Environmental Sciences
  Survey of semantic artefacts, vocabularies and governance.

• IOOS Compliance Checker
  Validate CF/ACDD conventions; validation is not scientific equivalence.

Explore further · optional / instructor

• Attribute Convention for Data Discovery Optional
  Dataset-level discovery metadata that complement CF variable metadata.

• CfRadial Optional
  Domain-specific CF extension for radar and lidar observations.

• NERC Vocabulary Server Optional
  Controlled terms and machine-accessible vocabularies.

• Xarray with MetPy Optional
  Meteorological example of software interpreting units and coordinates.

Teaching connection: Ask learners to compare long_name, standard_name, units, coordinates, and cell methods.

Full source annotations: 02 · Semantic interoperability

• NetCDF Climate and Forecast (CF) Metadata Conventions (Version 1.13) — Essential. Primary normative reference on standard_name, coordinates, units, calendar, cell_methods, bounds, and semantic context. Latest stable release at research date. Source: https://cfconventions.org/Data/cf-conventions/cf-conventions-1.13/cf-conventions.html
• CF Standard Names and Other CF Controlled Vocabularies (Standard Name Table v95) — Essential. Use as a lookup exercise for physical quantities and canonical units; versioned separately from CF Conventions. Source: https://cfconventions.org/vocabularies.html
• Assessing Semantic Interoperability in Environmental Sciences: Variety of Approaches and Semantic Artefacts — Essential. Comparative survey of 540 environmental semantic artefacts; illuminates vocabularies, ontologies, formal representations, and governance. Source: https://www.nature.com/articles/s41597-024-03669-3
• Attribute Convention for Data Discovery (ACDD), Version 1.3 — Optional. Shows how discovery-oriented, dataset-level descriptive metadata complement CF variable-level descriptions. Source: https://wiki.esipfed.org/Attribute_Convention_for_Data_Discovery_1-3
• CfRadial: NetCDF CF Conventions for Radial Coordinate Data for RADAR and LIDAR — Optional. Radar-specific extension of NetCDF/CF, closely matched to the lesson’s precipitation and IDRA radar use case. Source: https://ncar.github.io/CfRadial/
• NERC Vocabulary Server: Controlled Vocabulary Collections — Optional. A domain semantic-resource registry including versioned CF standard names and machine-accessible terminology. Source: https://vocab.nerc.ac.uk/collection/
• IOOS Compliance Checker Documentation — Essential. Practical validation resource for CF and ACDD; explain that syntactic conformance is not proof of full scientific comparability. Source: https://ioos.github.io/compliance-checker/
• xarray with MetPy Tutorial — Optional. Meteorological use case showing how CF-aware software can interpret units and coordinates and why metadata quality matters. Source: https://unidata.github.io/MetPy/latest/tutorials/xarray_tutorial.html

These notes are visible in Reveal.js speaker view rather than the slide itself.

03 · Technical interoperability: OPeNDAP

Guiding question: How does a common data-access protocol connect independently developed clients and servers?

Start here · essential

• OPeNDAP User Guide
  Client–server data access, metadata queries and selective retrieval.

• THREDDS Data Server
  Exposes data collections through multiple interoperable services.

Explore further · optional / instructor

• DAP4 Specification Instructor
  Protocol architecture distinct from specific server software.

• NASA OPeNDAP Constraint Expressions Optional
  Concrete URL requests for subsetting remote datasets.

Teaching connection: Compare a DAP request with a full file download; separate the protocol from THREDDS.

Full source annotations: 03 · Technical interoperability: OPeNDAP

• OPeNDAP User Guide — Essential. Explains client–server separation, DDS/DAS (DAP2), constraint expressions, and server-side selective retrieval. Source: https://opendap.github.io/documentation/UserGuideComprehensive.html
• The Data Access Protocol Version 4 (DAP4) Specification — Instructor. Normative DAP4 architecture and protocol reference; particularly useful for distinguishing the DAP protocol from its server implementations. Source: https://opendap.github.io/dap4-specification/DAP4.html
• What Is the THREDDS Data Server? — Essential. Explains how THREDDS exposes one data collection through several standard data-access services, including OPeNDAP. Source: https://docs.unidata.ucar.edu/tds/current/adminguide/what_is_the_tds.html
• OPeNDAP User Guide: Constraint Expressions — Optional. Concrete HTTP URL and constraint-expression examples for extracting subsets rather than downloading whole datasets. Source: https://www.earthdata.nasa.gov/fr/engage/open-data-services-software/earthdata-developer-portal/opendap/user-guide

These notes are visible in Reveal.js speaker view rather than the slide itself.

04 · Technical interoperability: APIs

Guiding question: What does agreeing on a web API guarantee—and what does it not guarantee?

Start here · essential

• OGC API: Environmental Data Retrieval Workshop
  Hands-on environmental data discovery and spatial subsetting.

Explore further · optional / instructor

• Federated Earth Observation Processing Optional
  openEO and standardized APIs across independent providers.

• OGC API: Environmental Data Retrieval 1.2 Instructor
  Published specification for environmental data access services.

• OGC API: Common Optional
  Shared API components and machine-readable interfaces.

Teaching connection: Ask whether a valid API response guarantees scientifically comparable datasets.

Full source annotations: 04 · Technical interoperability: APIs

• Federated and Reusable Processing of Earth Observation Data — Optional. Open-access research article examining openEO and interoperable workflows across providers; a case study showing what standardized APIs enable and where data/processing semantics still matter. Source: https://www.nature.com/articles/s41597-025-04513-y
• OGC API - Environmental Data Retrieval: Workshop Module — Essential. A guided exercise on standardized environmental web APIs, discovery, and location-specific requests. Source: https://ogcapi-workshop.ogc.org/api-deep-dive/environmental-data-retrieval/
• OGC API - Environmental Data Retrieval Standard, Version 1.2 (OGC 19-086r9) — Instructor. The published 2026 standard for machine-to-machine spatiotemporal discovery and subsetting; complements DAP rather than replacing it. Source: https://docs.ogc.org/is/19-086r9/19-086r9.html
• OGC API - Common: Overview — Optional. Provides a clear example of machine-readable API descriptions, shared paths and responses, and standardized client interoperability. Source: https://ogcapi.ogc.org/common/overview.html

These notes are visible in Reveal.js speaker view rather than the slide itself.

05 · Cloud-native data layouts

Guiding question: What changes when the same scientific arrays move from files to chunks and object storage?

Start here · essential

• Cloud-Optimized Geospatial Formats Guide
  Clear comparisons of chunked layouts, object stores and formats.

• Virtual Zarr Cookbook
  Use Kerchunk/VirtualiZarr without rewriting source files.

• Basics of Virtual Zarr Stores
  Step-by-step example with a downscaled climate dataset.

Explore further · optional / instructor

• Zarr Core Specification v3.1 Instructor
  Defines arrays, groups, chunks and metadata: not scientific meaning.

• Pangeo Forge Optional
  Reproducible pipelines for analysis-ready, cloud-optimized datasets.

• Kerchunk in Practice Optional
  Access to CMIP6-related NetCDF through virtual references.

Teaching connection: Compare a NetCDF file with chunked Zarr or a virtual Zarr view of existing data.

Full source annotations: 05 · Cloud-native data layouts

• Cloud-Optimized Geospatial Formats Guide — Essential. Illustrated open guide comparing cloud-friendly storage layouts; start with the Zarr chapter and object-storage concepts. Source: https://guide.cloudnativegeo.org/
• Zarr Core Specification, Version 3.1 — Instructor. Defines Zarr arrays, groups, metadata documents, chunk grids and object stores; distinguish the format from the scientific semantic conventions. Source: https://zarr-specs.readthedocs.io/en/latest/v3/core/index.html
• Pangeo Forge: Crowdsourcing Analysis-Ready, Cloud Optimized Data Production — Optional. Peer-reviewed open-access analysis of reproducible pipelines for publishing analysis-ready geoscientific datasets in cloud-optimized formats; connects storage layout to reuse and computational workflows. Source: https://www.frontiersin.org/journals/climate/articles/10.3389/fclim.2021.782909/full
• Virtual Zarr Cookbook (Kerchunk and VirtualiZarr) — Essential. Executable notebooks showing how to expose existing NetCDF/HDF data as virtual Zarr without rewriting source files. Source: https://projectpythia.org/kerchunk-cookbook/README.html
• Basics of Virtual Zarr Stores — Essential. One-file walkthrough using a concrete downscaled climate dataset; especially close to the course’s Kerchunk exercise. Source: https://projectpythia.org/kerchunk-cookbook/notebooks/foundations/kerchunk-basics/
• Kerchunk in Practice — Optional. Practical notebook for virtual access to publicly available CMIP6-related NetCDF data. Source: https://guide.cloudnativegeo.org/kerchunk/kerchunk-in-practice.html

These notes are visible in Reveal.js speaker view rather than the slide itself.

06 · Interoperable infrastructure in the AI era

Guiding question: Which kinds of interoperability must be in place before scientific datasets become AI-ready?

Primary reading · instructor context

• Building Blocks for AI-Ready Geospatial Data Instructor
  A discussion paper connecting semantics, structure and services to AI.

Complementary reading

• Recommendations for FAIR Data Services Optional
  FAIR infrastructure design: APIs, metadata, ontologies and repositories.

Teaching connection: Discuss which requirements are solved by storage alone and which need governed semantics.

Full source annotations: 06 · Interoperable infrastructure in the AI era

• Building Blocks for Interoperable, AI-Ready Geospatial Data: The Next Chapter of Spatial Data Infrastructures and Geospatial Ecosystems — Instructor. Recently published discussion paper directly distinguishing semantic, structural, and technical layers for AI-readiness. It is a proposal, not a normative OGC standard. Source: https://docs.ogc.org/techpaper/26-050/26-050.html
• Recommendations for Services in a FAIR Data Ecosystem — Optional. Explains the role of APIs, domain ontologies, and repositories in machine-actionable FAIR research infrastructures. Source: https://www.sciencedirect.com/science/article/pii/S2666389920300696

These notes are visible in Reveal.js speaker view rather than the slide itself.