Knowledge Resource Center for Ecological Environment in Arid Area
| 中国北方荒漠化遥感动态监测与定量评估研究 | |
| 其他题名 | Monitoring and Assessment of Desertification from Remote Sensing in the Northern China |
| 郭强 | |
| 出版年 | 2018 |
| 学位类型 | 博士 |
| 导师 | 付碧宏 |
| 学位授予单位 | 中国科学院大学 |
| 中文摘要 | 荒漠化是全球陆地生态环境系统所面临的最为严峻的问题之一，它严重制约着干旱-半干旱区经济社会的可持续发展。中国作为受荒漠化影响最严重的国家之一，国土面积的三分之一受到荒漠化影响。为了更好地对我国北方荒漠化动态过程进行空间监测，深入了解荒漠化发展的驱动机制，本研究基于多时相遥感数据建立荒漠化遥感监测评估模型，从区域尺度到全国尺度对中国北方荒漠化进行了动态监测与定量评估。论文的研究形成了一套从荒漠化时空动态变化监测到驱动力定量评估的方法，可以为荒漠化防治的决策提供重要的科学数据和强有力的技术支撑。主要内容和结论如下：（1）基于Thornthwaite公式法计算湿润指数进行的荒漠化气候区划依赖于于我国地面气象站点数据，没有考虑到具体的植被覆盖以及地形地貌等特征。本文采用长时间序列归一化植被指数Normalized Difference Vegetation Index（NDVI）以及高程数据反演新的湿润指数。根据计算的湿润指数进行荒漠化气候区划，改进了现有气候区划的一些缺陷。结果表明，中国北方荒漠化潜在发生区域面积为364.3万km2，约占我国陆地国土面积的37.8%。（2）参考前人荒漠化监测指标的研究，从荒漠化表象以及本质两方面考虑，选取修改型土壤调整植被指数Modified Soil-Adjusted Vegetation Index（MSAVI）、地表反照率Albedo和裸土指数Bare Soil Index（BSI）三个遥感监测指标，利用Landsat和Moderate-Resolution Imaging Spectroradiometer（MODIS）数据分别从区域尺度和全国尺度计算荒漠化监测指标，并对计算的荒漠化遥感监测指标进行了分析。（3）以鄂尔多斯高原为典型区，建立了基于Logistic回归模型、决策树模型和随机森林模型的荒漠化遥感监测模型。通过对比验证，得出决策树模型精度最高。基于Landsat计算的监测指标，用决策树模型对鄂尔多斯高原2000、2006、2010、2015年的荒漠化现状进行监测分析。结果显示，在2000-2015年期间鄂尔多斯高原的荒漠化呈现恢复的趋势。（4）利用Carnegie-Ames-Stanford Approach（CASA）模型计算鄂尔多斯高原长时间序列的Net Primary Productivity（NPP），在此基础上利用Potential NPP （PNPP）、Human Appropriation NPP（HANPP）分别代表气候变化和人类活动在生态环境变化中的定量作用，建立了荒漠化影响因素定量评估模型。对鄂尔多斯高原的研究表明，人类活动，尤其是中国北方实施的生态恢复工程是2000-2015年荒漠化恢复的主要原因。气候变化，主要是降水和气温的波动在2006-2010年期间主导了荒漠化退化。在2000-2006年以及2010-2015年期间，气候变化和人类活动在荒漠化退化过程中都扮演了重要的角色。（5）利用中分辨率的MODIS数据和决策树模型，从全国尺度对2001-2015年中国北方荒漠化现状和动态变化进行了遥感监测，并对荒漠化影响因素进行了定量的评估。研究结果表明，在2001-2015年的时段内，中国北方荒漠化发展呈现明显恢复的趋势，荒漠化区域总面积从180.32万km2减少到166.81万km2，荒漠化强度指数从0.2107下降到0.1888。对荒漠化驱动因素的定量评估结果表明，在2001-2005年的期间内，气候变化和人类活动在荒漠化发展过程中都扮演了重要的角色。在2005-2015年期间，气候变化是导致中国北方荒漠化总体呈现恢复趋势的主要因素。（6）从全国尺度上来讲，降水和气温的波动是荒漠化发展的主要控制因素，主导着一定时段内的恢复和退化。然而，在区域尺度上，人工造林、飞播造林、封山（沙）育林以及退耕还林、退牧还草等人类活动可以有效地促进当地生态环境恢复和改善。为了控制我国北方荒漠化发展趋势，促进生态环境和社会经济的可持续发展，各级政府应该发挥在政策方面的引领作用，持续推进和落实我国北方实施的退牧还草、退耕还林、三北防护林以及天然林保护工程等一系列生态修复和建设措施，实现生态环境和社会经济的可持续发展。 |
| 英文摘要 | Desertification is among the most serious problem facing the global terrestrial ecosystem which gravely hinder the sustainable development of economy and society in arid and semi-arid areas. China is one of the most seriously affected countries by desertification in the world with almost a third of its land area is under threat. In order to monitor the spatial-temporal development of desertification in Northern China, and further reveal the driving mechanisms behind the phenomenon, this research established a monitoring and evaluation model by the aid of satellite image time series. We established a scheme for monitoring the spatial-temporal variations of desertification and quantitatively assessing the draving factors in the desertification process, which may be beneficial for the alleviation of desertification. Main conclusions of this research are as follows:(1) The desertification climatic regionalization based on Thornthwaite method is totally dependent on data from ground-based meteorological stations, without considering the vegetation, topographic and geomorphic characteristics. By contrast, this study retrieved the moist index based on its relationship with NDVI and elevation, further zoning the climatic areas, which may counteract the disadvantages of the existing desertification climatic regionalization. The result of the new regionalization shows that the potential range of desertification occupies 3.643 million km2, and comprises 37.8% of land area in China.(2) Based on analysis and comparison of current monitoring methods, we selected three indicators (MSAVI, Albedo, BSI) considering the superficial phenomenon and nature of desertification process. We analysed the spatial characteristics of the indicators by using the Landsat and MODIS satellite images.(3) We established three remote sensing monitoring model based on logistic regression model, decision tree model and random forest model in the Ordos Plateau. Through comparison and analysis, the decision tree was showed to be the most accurate among the three models. Using Landsat data and decision tree model, we acquired the desertification status of Ordos Plateau of 2000, 2006, 2010 and 2015. The result shows that the desertification in Ordos has been through a recovery trend during 2000 to 2015.(4) We retrieved the NPP of Ordos using the CASA model, and established a model for evaluating the desertification driving forces using PNPP and HANPP representing the quantitative roles of climate change and human activities respectively during desertification process. Our result shows that human activities was the main cause of recovery during 2000 to 2015. Climate change was the dominant factor in degradation the period of 2006 to 2010. During 2000 to 2006 and 2010 to 2015, however, both climate change and human activities played important roles in the expansion of desertification.(5) Using MODIS data and decision tree model, we monitored the status and dynamics of desertification of Northern China during 2001 to 2015 on a national scale, and further assessed the percentile roles of different factors. During 2001 to 2015, the desertification status of Northern China has been through a recovery trend, with the desertification area decreased from 1.8032 million km2 to 1.6681 million km2, and the DI (desertification intensity) decreased from 0.2107 to 0.1888. During 2001 to 2005, climate change and human activities both played important roles in the reversion and degradation. However, during the period of 2005 to 2015, climate change was the main reason for desertification reversion.(6) On national scale, climate change including the variations of temperature and precipitation is the dominant factor in the desertification process. On regional scale, however, human activities such as reforestation and afforestation can improve the local ecological environment effectively. Therefore, the ecological programs must be continuously implemented in order to promote harmonious development between nature and human society in Northern China. |
| 中文关键词 | 荒漠化 ; 评估 ; 遥感 ; 气候变化 ; 人类活动 |
| 英文关键词 | Desertification assessment remote sensing climate change human activities |
| 语种 | 中文 |
| 国家 | 中国 |
| 来源学科分类 | 地图学与地理信息系统 |
| 来源机构 | 中国科学院遥感与数字地球研究所 |
| 资源类型 | 学位论文 |
| 条目标识符 | http://119.78.100.177/qdio/handle/2XILL650/288199 |
| 推荐引用方式 GB/T 7714 | 郭强. 中国北方荒漠化遥感动态监测与定量评估研究[D]. 中国科学院大学,2018. |
| 条目包含的文件 | 条目无相关文件。 | |||||
| 个性服务 |
| 推荐该条目 |
| 保存到收藏夹 |
| 导出为Endnote文件 |
| 谷歌学术 |
| 谷歌学术中相似的文章 |
| [郭强]的文章 |
| 百度学术 |
| 百度学术中相似的文章 |
| [郭强]的文章 |
| 必应学术 |
| 必应学术中相似的文章 |
| [郭强]的文章 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
