This dataset offers a unique, high-fidelity view of the viscous sublayer in a turbulent boundary layer at Reynolds number up to Reθ = 2400. It comprises 16,384 time-resolved snapshots of a three-dimensional flow field, each of size 3482 × 26 × 256, with four fluid‐dynamic quantities (the three velocity components and pressure) stored in double precision. In total, the data form a five-dimensional array of shape (16384, 3482, 26, 256, 4).
All snapshots were generated by direct numerical simulation using the open-source Xcompact3d solver on the UK's ARCHER2 supercomputer. To make the data easy to work with, the raw binary output has been converted to Zarr format, enabling automatic chunking tailored to common access patterns and on-the-fly compression to minimise download times.
In addition to the time-resolved fields, we also provide pre-computed, time-averaged three-dimensional statistical fields (e.g. mean velocities and Reynolds stresses) plus full metadata describing domain dimensions, time-step interval and viscosity. This rich combination of instantaneous and statistical data will support a wide range of applications, from turbulence theory development and experimental design, through training data-driven models, to validating faster, lower-fidelity numerical methods.