Cross-Domain Virtual Laboratories on Blue-Cloud 2026: Shared Technologies and Platform Lessons Learned

Blue-Cloud 2026 delivers a transversal suite of Virtual Laboratories (VLabs) implemented on a common Virtual Research Environment (VRE) in D4Science to enable end-to-end, FAIR-aligned marine data science across disciplines: from coastal physics and extremes to biogeochemistry, indicators, and fisheries. Rather than isolated demonstrators, the VLabs share reproducible platform patterns that standardize how users discover data, subset and authenticate to external providers, run analytics, and publish reproducible output. 

Across VLabs, a common interaction model combines (i) gateway-based identity and access, (ii) curated “VRE folders” that distribute ready-to-run notebooks and resources, and (iii) interactive web dashboards for parameter selection (space/time/variables), pre-flight checks (coverage and overlap), and guided execution. A shared technical backbone supports transparent data acquisition and processing: connection to federated repositories and services (e.g., Copernicus Marine, EMODnet, thematic nodes), automated subsetting, and workflow steps that harmonize formats, apply quality control, manage gaps, and generate analysis-ready datasets. Several VLabs implement the same methodological “building blocks” in different contexts: variational mapping/interpolation (DIVAnd) for gridded fields, model–observation fusion, and standardized production of map/time-series outputs (NetCDF plus figures/HTML).

A second cross-cutting layer is scalable computation. While notebooks remain central for transparency and education, compute-heavy workflows increasingly migrate to shared cloud analytics services (e.g. CCP Analytics Engine) - which include delegation of compute-intensive routines to optimized backend implementations - to (a) reduce local data dependencies through remote subsetting, (b) reuse cached intermediate products, (c) support larger spatio-temporal domains, and (d) generate interactive deliverables (e.g., Plotly dashboards), alongside archival outputs. This pattern is exemplified by indicator services (MHW, OHC, TRIX, SSIv2) but is transferable to other VLabs with large datasets or reproducible executions.

Transversal lessons learned include: (1) interoperability hinges on early harmonization (units, grids, metadata, vocabularies) and “best-practice” preprocessing embedded in the VRE; (2) user trust improves when workflows expose logs, provenance, and configuration exports for audit and reproducibility; (3) robust operations require resilience to upstream outages, authentication variability, and evolving toolchains; and (4) modular design (shared UI patterns, reproducible processing kernels, and standardized outputs) accelerates expansion to new regions, variables, and communities. Collectively, the Blue-Cloud 2026 VLabs demonstrate how a unified VRE can operationalize cross-domain marine analytics, translating distributed infrastructures into consistent user experiences and reproducible digital workflows.