• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >Advancing the quantification of humid tropical forest cover loss with multi-resolution optical remote sensing data: Sampling & wall-to-wall mapping.
【24h】

Advancing the quantification of humid tropical forest cover loss with multi-resolution optical remote sensing data: Sampling & wall-to-wall mapping.

机译：利用多分辨率光学遥感数据促进对潮湿热带森林覆盖率损失的量化：采样和墙到墙映射。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Humid tropical forest cover loss is threatening the sustainability of ecosystem goods and services as vast forest areas are rapidly cleared for industrial scale agriculture and tree plantations. Despite the importance of humid tropical forest in the provision of ecosystem services and economic development opportunities, the spatial and temporal distribution of forest cover loss across large areas is not well quantified. Here I improve the quantification of humid tropical forest cover loss using two remote sensing-based methods: sampling and wall-to-wall mapping. In all of the presented studies, the integration of coarse spatial, high temporal resolution data with moderate spatial, low temporal resolution data enable advances in quantifying forest cover loss in the humid tropics. Imagery from the Moderate Resolution Imaging Spectroradiometer (MODIS) are used as the source of coarse spatial resolution, high temporal resolution data and imagery from the Landsat Enhanced Thematic Mapper Plus (ETM+) sensor are used as the source of moderate spatial, low temporal resolution data. In a first study, I compare the precision of different sampling designs for the Brazilian Amazon using the annual deforestation maps derived by the Brazilian Space Agency for reference. I show that sampling designs can provide reliable deforestation estimates; furthermore, sampling designs guided by MODIS data can provide more efficient estimates than the systematic design used for the United Nations Food and Agricultural Organization Forest Resource Assessment 2010. Sampling approaches, such as the one demonstrated, are viable in regions where data limitations, such as cloud contamination, limit exhaustive mapping methods. Cloud-contaminated regions experiencing high rates of change include Insular Southeast Asia, specifically Indonesia and Malaysia. Due to persistent cloud cover, forest cover loss in Indonesia has only been mapped at a 5-10 year interval using photo interpretation of single best Landsat images. Such an approach does not provide timely results, and cloud cover reduces the utility of map outputs. In a second study, I develop a method to exhaustively mine the recently opened Landsat archive for cloud-free observations and automatically map forest cover loss for Sumatra and Kalimantan for the 2000-2005 interval. In a comparison with a reference dataset consisting of 64 Landsat sample blocks, I show that my method, using per pixel time-series, provides more accurate forest cover loss maps for multiyear intervals than approaches using image composites. In a third study, I disaggregate Landsat-mapped forest cover loss, mapped over a multiyear interval, by year using annual forest cover loss maps generated from coarse spatial, high temporal resolution MODIS imagery. I further disaggregate and analyze forest cover loss by forest land use, and provinces. Forest cover loss trends show high spatial and temporal variability. These results underline the importance of annual mapping for the quantification of forest cover loss in Indonesia, specifically in the light of the developing Reducing Emissions from Deforestation and Forest Degradation in Developing Countries policy framework (REDD). All three studies highlight the advances in quantifying forest cover loss in the humid tropics made by integrating coarse spatial, high temporal resolution data with moderate spatial, low temporal resolution data. The three methods presented can be combined into an integrated monitoring strategy.
机译：湿润的热带森林覆盖率的丧失正威胁着生态系统产品和服务的可持续性，因为广阔的森林地区已被迅速清理为工业规模的农业和人工林。尽管湿润的热带森林在提供生态系统服务和经济发展机会方面很重要，但仍无法很好地量化大面积森林覆盖物损失的时空分布。在这里，我使用两种基于遥感的方法改进了对热带湿润森林覆盖率损失的量化：采样和墙到墙贴图。在所有提出的研究中，将粗略的空间，高时间分辨率的数据与适度的空间，低时间分辨率的数据集成在一起，可以在量化湿热带地区的森林覆盖率损失方面取得进展。来自中分辨率成像光谱仪（MODIS）的图像被用作粗略的空间分辨率，高时间分辨率的数据，来自Landsat Enhanced Thematic Mapper Plus（ETM +）传感器的图像被用作中等空间，低时间分辨率的数据源。在第一项研究中，我使用巴西航天局提供的年度森林砍伐图来比较巴西亚马逊不同采样设计的精度。我表明抽样设计可以提供可靠的毁林估计；此外，以MODIS数据为指导的抽样设计比联合国粮食及农业组织2010年森林资源评估所采用的系统设计可提供更有效的估算。抽样方法（例如已证明的抽样方法）在数据有限的区域是可行的，例如云污染，限制了详尽的映射方法。变化频繁的受云污染的地区包括东南亚岛屿，特别是印度尼西亚和马来西亚。由于持续的云层覆盖，印度尼西亚的森林覆盖率损失仅通过对单个最佳Landsat图像的照片解释以5-10年的间隔进行绘制。这种方法无法提供及时的结果，并且云量会降低地图输出的实用性。在第二项研究中，我开发了一种方法来彻底挖掘最近打开的Landsat档案库以进行无云观测，并自动绘制2000-2005年间苏门答腊和加里曼丹的森林覆盖率损失。在与由64个Landsat样本块组成的参考数据集进行的比较中，我表明，与使用图像合成的方法相比，使用每像素时间序列的方法可以提供多年间隔的更准确的森林覆盖率损失图。在第三项研究中，我使用从粗糙的空间，高时间分辨率的MODIS图像生成的年度森林覆盖率损失图，按年份细分了以多年间隔绘制的Landsat映射的森林覆盖率损失。我进一步分析和分析了按林地使用情况和省份划分的森林覆盖率损失。森林覆盖率损失趋势显示出较高的时空变异性。这些结果强调了年度绘图对于量化印度尼西亚森林覆盖率损失的重要性，特别是考虑到发展中国家政策框架（REDD）不断减少的森林砍伐和森林退化造成的排放。所有这三项研究都突出了通过将粗糙的空间，高时间分辨率数据与适度的空间，低时间分辨率数据相结合而在量化潮湿热带地区森林覆盖率方面取得的进展。提出的三种方法可以组合为一个集成的监视策略。

著录项

  • 作者

    Broich, Mark.;

  • 作者单位

    South Dakota State University.;

  • 授予单位 South Dakota State University.;
  • 学科 Geographic information science and geodesy.;Remote sensing.;Environmental science.
  • 学位 Ph.D.
  • 年度 2010
  • 页码 184 p.
  • 总页数 184
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

  • 入库时间 2022-08-17 11:37:07

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号