FluvDepoSet

FluvDepoSet is a data set of 3D models of sedimentary deposits from single-threaded, sand-bedded meandering rivers. It aims at providing a training and testing data set for machine learning applications that require geologically plausible distributions of sedimentary heterogeneities.\nEach file includes a single model with three properties: the geological interfaces (topography and basement), the fraction of coarse sediments, and the deposition time of the sediments.\nLineage: Each model was simulated using the landscape evolution model CHILD (github.com/childmodel/child), in particular the stream meander and floodplain deposition modules. The values of the input parameters that vary between each model are given in the attributes of each hdf5 file. All the other input parameters are kept constant and can be found in the supporting document.

FluvDepoSet是一套面向单河道砂质床曲流河沉积的三维模型数据集，旨在为需要具备地质合理性的沉积非均质性分布的机器学习应用提供训练与测试数据。
每个文件包含一个单一模型，具备三项属性：地质界面（地形与基底）、粗沉积物占比以及沉积物沉积时间。
数据集溯源：所有模型均通过地貌演化模型CHILD（github.com/childmodel/child）模拟生成，尤其使用了其河道曲流与漫滩沉积模块。各模型间存在差异的输入参数取值已存储于每个HDF5（Hierarchical Data Format 5）文件的属性字段中，其余所有输入参数均保持固定，相关信息可参阅辅助文档。