Global Detached Structures Dataset (GDSD)

Here, the Global  Detached Structures Dataset (GDSD) is presented. 

The Global Detached Structures Dataset (GDSD) is presented here as a curated compilation of 376 detached coastal structures from 10 countries worldwide. The dataset includes detailed geometric and hydrodynamic parameters, as well as shoreline response classifications, covering both natural (e.g., hedlands, reefs, islands) and human-made structures (emerged, low-crested, and submerged types). High-resolution orthophotographs, satellite imagery, and multiple digital elevation models (DEMs) form the basis for extracting structural, morphological, and hydrodynamic variables. Where necessary, these data are complemented by values reported in the scientific literature to ensure completeness and consistency.

Specifically designed for machine learning applications, the GDSD facilitates the development and training of predictive models that aim to outperform traditional empirical approaches, enabling more accurate classification of detached structures and improved projections of their morphodynamical impacts on sandy coastlines across diverse geographic settings.

The following table list describes the variables and parameters considered.