Global Mangrove Loss Extent, Land Cover Change, and Loss Drivers, 2000-2016

This dataset provides estimates of the extent of mangrove loss, land cover change, and its anthropogenic or climatic drivers in three time periods: 2000-2005, 2005-2010, and 2010-2016. Landsat-based Normalized Difference Vegetation Index (NDVI) anomalies were used to determine loss extent in each period. The drivers of mangrove loss were determined by examining land cover changes using a random forest machine learning technique that considered change from mangrove to wet soil, dry soil, and water at each loss pixel. A series of decision trees used several global-scale land-use datasets to identify the ultimate driver of the mangrove loss. Loss drivers include commodity production (agriculture, aquaculture), settlement, erosion, extreme climatic events, and non-productive conversion. Maps of loss extent per period, mangrove land cover changes, and loss drivers are provided for each of 39 mangrove holding nations.

本数据集提供了三个时间段（2000-2005年、2005-2010年以及2010-2016年）红树林损失程度、土地覆被变化及其人为或气候驱动因素的估算。利用基于Landsat的归一化植被指数（NDVI）异常值，确定了每个时期内的损失范围。通过随机森林机器学习技术，分析土地覆被变化以确定红树林损失驱动因素，该技术考虑了从红树林到湿土、干土和水的转变在每个损失像素点的变化。一系列决策树使用多个全球尺度的土地利用数据集，以识别红树林损失的最终驱动因素。损失驱动因素包括商品生产（农业、水产养殖）、定居、侵蚀、极端气候事件以及非生产性转换。为39个红树林拥有国提供了每个时期的损失范围图、红树林土地覆被变化图和损失驱动因素图。