Mangrove Forests Land Area and Crown Surface Cover Area Data

Mangrove ecosystems play a dominant role in global tropical and subtropical coastal wetlands. Remote sensing has a central role in mangrove conservation, as it is the preferred tool for monitoring spatiotemporal distribution changes. Through Landsat remote sensing image data, this study employed support vector machine (SVM) machine learning and Res-UNet deep learning to monitor the changes of mangrove forests and crown surface cover were monitored in Hainan Island from 1991 to 2021. Additionally, based on the crown surface cover area of mangrove forests in Hainan Island, the influencing mechanisms were analyzed via dynamic changes and landscape patterns. The following is the data package of our research: 1. The data in the folder (Res-UNet Deep Learning Classification) the classification results of Res-UNet deep learning algorithm, as well as its accuracy verification results. 2. The data in the folder (Influential Mechanisms) contains the total population, urban population, and rural population, GDP and gross output fishery value and climatic data and Pearson correlation analysis index for Hainan Island during the study period. The total population, urban population, and rural population, GDP, gross output fishery value, and shelter forests planting area of Hainan Island were obtained from the Annual Statistical Report of Hainan Province, and the climate data is downloaded from WorldClim data website (https://www.worldclim.org/data/index.html). 3. The data in the folder (Landscape Patterns) is the landscape pattern index and the annual change rate of mangrove forest crown surface cover in Hainan Island. 4. The data in the folder (SVM Machine Learning Classification) is the classification results of SVM machine learning algorithm, as well as its accuracy verification results. 5. The data in the folder (Ground Survey) are the distribution range of mangrove forests and the distribution of dominant mangrove tree species obtained by the team members in the ground survey in 2020.红树林生态系统在全球热带、亚热带滨海湿地中占据主导地位。遥感技术在红树林保护中具有核心作用，是监测其时空分布变化的首选工具。本研究依托Landsat遥感影像数据，采用支持向量机（Support Vector Machine, SVM）机器学习与Res-UNet深度学习模型，对1991至2021年海南岛红树林及其冠层地表覆盖的变化进行监测。此外，本研究基于海南岛红树林冠层地表覆盖面积，通过动态变化与景观格局分析其影响机制。 本研究的数据集详情如下： 1. Res-UNet深度学习分类文件夹：包含Res-UNet深度学习算法的分类结果及其精度验证结果。 2. 影响机制文件夹：包含研究时段内海南岛的总人口、城镇人口、乡村人口、国内生产总值（GDP）、渔业总产值、气候数据及皮尔逊相关分析指标。其中总人口、城镇人口、乡村人口、GDP、渔业总产值及防护林种植面积数据来源于《海南省统计年鉴》，气候数据下载自WorldClim数据网站（https://www.worldclim.org/data/index.html）。 3. 景观格局文件夹：包含海南岛红树林冠层地表覆盖的景观格局指数及其年变化率。 4. SVM机器学习分类文件夹：包含支持向量机机器学习算法的分类结果及其精度验证结果。 5. 地面调查文件夹：包含研究团队2020年实地调查获取的红树林分布范围及优势红树林树种分布数据。