Structural characterisation of mangrove forests achieved through combining multiple sources of remote sensing data
Structural characterisation of mangrove forests achieved through combining multiple sources of remote sensing data
复制标题
DOI：
10.1016/j.rse.2019.111543
复制
发表时间：
2020-02
影响因子：
13.5
通讯作者：
R. Lucas;R. Van De Kerchove;Viviana Otero;D. Lagomasino;L. Fatoyinbo;H. Omar;B. Satyanarayana;F. Dahdouh-Guebas
中科院分区：
文献类型：
--
作者：
R. Lucas;R. Van De Kerchove;Viviana Otero;D. Lagomasino;L. Fatoyinbo;H. Omar;B. Satyanarayana;F. Dahdouh-Guebas
关于红树林范围、年龄、结构和生物量的时间信息为理解这些生态系统在其提供的服务方面的作用（例如，与碳储存、保护生物多样性有关）提供了重要贡献，特别是考虑到人类活动和自然事件和过程影响的多样性。这项研究以马来西亚半岛霹雳州的马当红树林保护区（MMFR）为重点，旨在从星载光学和合成孔径雷达（SAR）中检索有关红树林生物物理特性的全面信息，以支持更好地了解其在管理环境中的动态。 1988年至2016年（29年）期间，通过结合陆地卫星来源的归一化水分指数（NDMI）和日本L波段合成孔径雷达（SAR）数据的时间序列，至少每年估算一次森林年龄。 NDMI 进一步用于检索冠层覆盖率 (%)。评估了干涉航天飞机雷达地形任务 (SRTM) X/C 波段 (2000)、TanDEM-X 波段 (2010-2016) 和立体 WorldView-2 立体 (2016) 数据在估计冠层高度 (CH) 中的作用，并使用预先建立的异速生长从中导出地上生物量 (AGB, Mg ha−1)。虽然 L 波段 HH 和 HV 数据在大约 8-10 年的生长后随着 AGB 的增加而增加，但由于森林更新中的砍伐和木质材料、树木在大约 15 年和 20 年间稀疏以及大约 20 年生长后 L 波段 SAR 数据的饱和，不同数量的死亡木本碎片的混合散射，检索受到了影响。参考立体 Phantom-3 DJI 立体图像来支持冠层覆盖 (CC) 的估计和卫星衍生 CH 的验证。 AGB 估计值与地面测量值进行了比较。利用与森林年龄的关系，对 Landsat 或 L 波段 SAR 观测的每个日期的 CH 和 AGB 进行估计，并且显示 L 波段 SAR 的时间趋势可以有效跟踪砍伐和再生的序列。由此，定义了收获周期的四个阶段。该研究提供了有关MMFR红树林生物物理特性和生长动态的新信息、未来监测活动的投入以及促进更好地描述和绘制全球红树林地区的方法。
Temporal information on mangrove extent, age, structure and biomass provides an important contribution towards understanding the role of these ecosystems in terms of the services they provide (eg, in relation to storage of carbon, conservation biodiversity), particularly given the diversity of influences of human activity and natural events and processes. Focusing on the Matang Mangrove Forest Reserve (MMFR) in Perak Province, Peninsular Malaysia, this study aimed to retrieve comprehensive information on the biophysical properties of mangroves from spaceborne optical and Synthetic Aperture Radar (SAR) to support better understanding of their dynamics in a managed setting. For the period 1988 to 2016 (29 years), forest age was estimated on at least an annual basis by combining time-series of Landsat-derived Normalised Difference Moisture Index (NDMI) and Japanese L-band Synthetic Aperture Radar (SAR) data. The NDMI was further used to retrieve canopy cover (%). Interferometric Shuttle Radar Topographic Mission (SRTM) X/C-band (2000), TanDEM-X-band (2010–2016) and stereo WorldView-2 stereo (2016) data were evaluated for their role in estimating canopy height (CH), from which above ground biomass (AGB, Mg ha− 1) was derived using pre-established allometry. Whilst both L-band HH and HV data increased with AGB after about 8–10 years of growth, retrieval was compromised by mixed scattering from varying amounts of dead woody debris following clearing and wood material within regenerating forests, thinning of trees at~ 15 and 20 years, and saturation of L-band SAR data after approximately 20 years of growth. Reference was made to stereo Phantom-3 DJI stereo imagery to support estimation of canopy cover (CC) and validation of satellite-derived CH. AGB estimates were compared with ground-based measurements. Using relationships with forest age, both CH and AGB were estimated for each date of Landsat or L-band SAR observation and the temporal trends in L-band SAR were shown to effectively track the sequences of clearing and regeneration. From these, four stages of the harvesting cycle were defined. The study provided new information on the biophysical properties and growth dynamics of mangrove forests in the MMFR, inputs for future monitoring activities, and methods for facilitating better characterisation and mapping of mangrove areas worldwide.