High-resolution wave measurements at intermediate water depth are required to improve coastal impact modeling. Specifically, such data sets are desired to calibrate and validate models, and broaden the insight on the boundary conditions that force models. Here, we present a wave data set collected in the North Sea at three stations in intermediate water depth (6–14 m) during the 2021/2022 storm season as part of the RealDune/REFLEX experiments. Continuous measurements of synchronized surface elevation, velocity and pressure were recorded at 2–4 Hz by Acoustic Doppler Profilers and an Acoustic Doppler Velocimeter for a 5-month duration. Time series were quality-controlled, directional-frequency energy spectra were calculated and common bulk parameters were derived. Measured wave conditions vary from calm to energetic with 0.1–5.0 m sea-swell wave height, 5–16 s mean wave period and W-NNW direction. Nine storms, i.e., wave height beyond 2.5 m for at least six hours, were recorded including the triple storms Dudley, Eunice and Franklin. This unique data set can be used to investigate wave transformation, wave nonlinearity and wave directionality for higher and lower frequencies (e.g., sea-swell and infragravity waves) to compare with theoretical and empirical descriptions. Furthermore, the data can serve to force, calibrate and validate models during storm conditions. Dataset: https://doi.org/10.4121/233f11ff-7804-4777-8b32-92c4606e56d8 Dataset License: CC-BY 4.0.

We propose two end-to-end neural configurations for language diarization on bilingual code-switching speech. The first, a BLSTM-E2E architecture, includes a set of stacked bidirectional LSTMs to compute embeddings and incorporates the deep clustering loss to enforce grouping of languages belonging to the same class. The second, an XSA-E2E architecture, is based on an x-vector model followed by a self-attention encoder. The former encodes frame-level features into segmentlevel embeddings while the latter considers all those embeddings to generate a sequence of segment-level language labels. We evaluated the proposed methods on the dataset obtained from the shared task B in WSTCSMC 2020 and our handcrafted simulated data from the SEAME dataset. Experimental results show that our proposed XSA-E2E architecture achieved a relative improvement of 12.1% in equal error rate and a 7.4% relative improvement on accuracy compared with the baseline algorithm in the WSTCSMC 2020 dataset. Our proposed XSA-E2E architecture achieved an accuracy of 89.84% with a baseline of 85.60% on the simulated data derived from the SEAME dataset.