CHIMERA regular vine matrices in Zarr v3 format

This dataset provides a Zarr version 3 representation of the complete CHIMERA atlas of regular vine structure matrices on 4 to 8 labelled nodes. It is derived from the official CHIMERA Python dataset published by 4TU.ResearchData. The dataset contains: - 24 matrices on 4 nodes;- 480 matrices on 5 nodes;- 23,040 matrices on 6 nodes;- 2,580,480 matrices on 7 nodes;- 660,602,880 matrices on 8 nodes. The arrays are named `matrices4` to `matrices8` and have dimensions `vine_id`, `matrix_row`, and `matrix_column`. The `vine_id` coordinate is zero-based within each array. Matrix elements are stored as unsigned 8-bit integers. Zero is a valid structural value; 255 is reserved as the fill value for an uninitialised or unavailable region. The dataset is stored as Zarr v3 using chunks of up to 100,000 matrices and Blosc/Zstandard compression. All 6,636 expected chunks were present and validated when this release was prepared. Version 1.1.0 contains a corrected `matrices5` array. The previous online Zarr representation was found to be incorrect and the complete five-node array was rebuilt on 8 July 2026. This Zenodo deposit is the immutable and citable release. An operational HTTP representation is also available at: http://meteo.unican.es/work/chimera.zarr Source and related works: - Source Python dataset: https://doi.org/10.4121/20655120.v1- CHIMERA collection, version 8: https://doi.org/10.4121/c17b8790-dfd2-4394-854a-7d98fd254c15.v8- Related peer-reviewed publication: https://doi.org/10.1038/s41597-023-02252-6- Access and processing software: https://github.com/SantanderMetGroup/trasgu The scientific creators of CHIMERA are Marcel 't Hart, Oswaldo Morales Napoles, Mojtaba Rajabi, and Gina Torres Alves. This Zarr representation is curated by Valvanuz Fernández-Quiruelas, Universidad de Cantabria. Licence: CC0 1.0 Universal.