节点文献
北疆区域积雪深度变化的遥感监测研究
Remote Sensing Monitoring of Snow Depth Change in North Part of Xinjiang
【作者】 魏玥;
【作者基本信息】 新疆师范大学 , 地图学与地理信息系统, 2010, 硕士
【摘要】 本论文以分析反演积雪覆盖与雪深为目标,利用积雪观测、实测资料以及相应时次的MODIS卫星资料,以GIS技术手段的测试反演分析方法,运用系统的遥感科学理论和数值分析模拟方法,建立MODIS资料的积雪覆盖率与深度的反演模型,探讨了积雪深度变化和气候之间的相互反馈及影响作用。论文资料选取北疆地区20个基本气象台站1971~2006年36年的年最大积雪深度的实测资料,建立年最大积雪深度时间序列;根据2004-2006年度冬季伊犁河谷地区积雪站资料和2003-2006年度冬季相关气象台站雪深资料进行了分析。针对典型地区积雪深度和密度状况2009-2010年冬季进行了多次野外实测,共获取127个样本数据,为模型研建与验证提供了第一手宝贵资料。同时收集了对应的MODIS晴空探测资料,在MODIS多光谱多波段资料处理分析的基础上,针对雪深反演自动判识这一难点,利用数学统计方法结合不同目标物的光谱特性知识库,同时考虑下垫面条件(包括地表积雪覆盖度、反射率、积雪密度)等多种因子和季节性等对积雪深度分布的影响,实现云、雪、地表、水体等各种目标物的自动判识和积雪分类。分析研究采取多学科交叉融合,从野外实验、数据收集到模型模拟和区域集成的技术路线。通过对区域气候变化、积雪观测事实分析,以山区积雪制图为基础,将积雪覆盖信息和地理信息数据在GIS空间分析技术支持下进行多元数据融合、栅格运算、积雪覆盖度计算等多种处理,制作出观测区和自然流域的积雪产品,建立EOS/MODIS积雪遥感监测与GIS应用技术的积雪深度反演模型。结果表明北疆地区季节性积雪年际波动不大,年际变化的区域差异不显著,相对比较稳定。在区域气候变暖的背景下,该地区季节性积雪与青藏高原积雪和两极地区的冰盖变化趋势一样,通过平均差值法、最小二乘法和自回归滑动平均法三种统计模式检验,北疆地区季节性积雪随着区域气候的变暖而有所增加。季节性积雪的长期变化趋势是冬季气温、降水长期波动变化的结果,季节性积雪的年际变化与冬季平均气温的波动呈弱的负相关,与冬季降水的波动呈显著的正相关关系。通过利用MODIS卫星资料的分析和建立的反演模型,基本了解北疆区域积雪的积累和分布的规律,为新疆积雪的定性、定量及动态变化分析提供重要技术手段,为新疆气象部门拓展服务领域,指导农牧业生产和积雪的防灾减灾决策提供直观和及时的服务产品。
【Abstract】 Aiming for retrieving snow cover and snow depth, based upon the observed snow data and corresponding MODIS satellite data, utilizing GIS retrieval methods, systematic remote sensing theory and numerical simulation method, the paper developed the retrieval model of snow cover and depth of MODIS data, discussed the feedback and influence between snow depth and climate change.The paper selected the observed maximum snow depth data of 20 primary weather stations in north part of Xinjiang from 1971 to 2006, set up the time series of maximum snow depth; analyzed the data from Tianshan Snow-Cover and Avalanche Research Station and the concerned weather stations in Yili valley. Aiming for the snow depth and snow density in typical area, we did field observation in the winter of 2009 and 2010, took 127 samples and offered the first-hand data for model construction and verification. We collected MODIS data in clear days. Aiming for the difficulty in automatically identification of snow depth retrieval, combining statistical methods with spectrum features bank of various objects, considering the influence of substrate condition (including snow cover, albedo and density) and season on snow depth, the paper realized snow automatic classification and identification of cloud, snow, land surface and water.The study adopted the routine of interdisciplinarity from field experiment, data collection to model simulation and regional integration. Through the analysis on regional climate change and facts of snow observation, based on mapping of mountainous snow, the paper dealed with the data with the methods of integration of multi-data, grid calculation and snow cover calculation under the support of GIS spatial analysis technique, produced the snow product of observed area and natural catchments, and set up the snow depth retrieval model of EOS/MODIS snow remote sensing monitoring and GIS application.It indicated that the inter-annual change of seasonal snow in north part of Xinjiang was stable, with the indistinctive regional difference. Under the background of regional climate warming, the tendency of seasonal snow of the study area was consistent with that of Tibetan Plateau and icecap of polar area. Through three statistical tests of mean subtraction, least square, and auto-regressive moving-average, it was found that the seasonal snow in north part of Xinjiang increased under the regional warming background. The long-term change of seasonal snow was associated with the long-term fluctuation of winter temperature, and precipitation. The inter-annual change of seasonal snow change was weakly negative correlated with mean winter temperature and significantly positive correlated with winter precipitation.Through the analysis of MODIS satellite data and retrieval model, we understood the discipline of snow accumulation and distribution in north part of Xinjiang, offered critical ways to the qualitative, quantitative and dynamic analysis on snow in Xinjiang, expanded the service field of Xinjiang meteorological section, and provided the visual and timely service product for the decision-making on agriculture and animal husbandry production and the prevention and reduction of snow disaster.
【Key words】 snow depth; snow density; MODIS data; snow depth retrieval; remote sensing monitoring;