A High-Resolution Record of Surface Melt on Antarctic Ice Shelves using Multi-Source Remote Sensing Data and Deep Learning

The dataset provided in this repository corresponds to the original data used in the publication by De Roda Husman et al. (2023) titled "A High-Resolution Record of Surface Melt on Antarctic Ice Shelves using Multi-Source Remote Sensing Data and Deep Learning" (DOI to be announced!). <br>The dataset, named UMelt, contains a comprehensive surface melt record for all Antarctic ice shelves. It offers a high spatial resolution of 500 meters and a high temporal resolution of 12 hours, covering the period from 2016 to 2021. Our methodology relies on the utilization of a deep learning model known as U-Net, which integrates microwave remote sensing observations from three sources: Sentinel-1, Special Sensor Microwave Imager/Sounder (SSMIS), and Advanced Scatterometer (ASCAT). <br>The data is available for download in two formats: 1. "Timeseries": This format provides the data at a twice-daily resolution, allowing for detailed analysis over time.2. "MeltFraction": This format offers a yearly, summed product, providing a consolidated representation of the melt fraction.<br>Feel free to access and explore the dataset to gain valuable insights into surface melt dynamics on Antarctic ice shelves.

本仓库提供的数据集对应De Roda Husman及其团队2023年发表的学术论文所使用的原始实验数据，该论文标题为《利用多源遥感数据与深度学习获取南极冰架表面融化高分辨率记录》，其DOI待公布！
本数据集命名为UMelt，涵盖了所有南极冰架的完整表面融化观测记录。其空间分辨率达500米，时间分辨率为12小时，时间覆盖范围为2016年至2021年。本研究采用的方法依托U-Net深度学习模型，该模型整合了三类数据源的微波遥感观测数据：哨兵1号（Sentinel-1）、专用传感器微波成像仪/声测仪（Special Sensor Microwave Imager/Sounder, SSMIS）以及先进散射计（Advanced Scatterometer, ASCAT）。
该数据提供两种下载格式：1. 时间序列（Timeseries）格式：该格式以每日两次的分辨率提供数据，支持精细化的长时间序列分析。2. 融化分数（MeltFraction）格式：该格式为年度累加产物，可直观呈现整体的融化分数情况。
欢迎使用并探索本数据集，以期获取关于南极冰架表面融化动态的宝贵研究见解。