WEBGIS技术在地震前兆台网中的应用

使用WEBGIS技术,结合地震前兆台网的实际情况,设计完成地震前兆台网的GIS展示工具.该工具以互联网为媒介,使用地理空间数据库存储的全国台站数据,对全国范围内的物理场分布在网络客户端进行发布,方便地震前兆台网用户获取数据,同时丰富地震前兆数据的展示方式.     

利用支持向量机和高斯过程回归测定水库诱发的地震(英文)

水库诱发地震震级(M)的预测是在地震工程中的一项重要任务。本文采用支持向量机(SVM)和高斯过程回归(GPR)模型根据水库的参数预测了水库诱发地震震级(M)。综合参数(E)和最大的水库深度(H)作为支持向量机和高斯过程回归模型的输入参数。我们给出一个方程确定水库诱发地震震级(M)。将本文开发的支持向量机和建立的高斯过程回归方法与人工神经网络(ANN)方法相比。结果表明,本文研发的支持向量机和高斯过程回归方法是预测水库诱发地震震级(M)的有效工具。     

Retrieval algorithm for microwave surface emissivities based on multi-source, remote-sensing data: An assessment on the Qinghai-Tibet Plateau

The Qinghai-Tibet Plateau plays a very important role in studying severe weather in China and around the globe because of its unique characteristics. Moreover, the surface emissivities of the Qinghai-Tibet Plateau are also important for retrieving surface and atmospheric parameters. In the current study, a retrieval algorithm was developed to retrieve the surface emissivities of the Qinghai-Tibet Plateau. The developed algorithm was derived from the radiative transfer model and was first validated using simulated data from a one-dimensional microwave simulator. The simulated results show good precision. Then, the surface emissivities of the Qinghai-Tibet Plateau were retrieved using brightness temperatures from the advanced microwave-scanning radiometer and atmospheric profile data from the moderate resolution imaging spectroradiometer. Finally, the features of the time and space distribution of the retrieved results were analyzed. In terms of spatial characteristics, a spatial distribution con- sistency was found between the retrieved results and surface coverage types of the Qinghai-Tibet Plateau. In terms of time characteristics, the changes in emissivity, which were within 0.01 for every day, were not evident within a one-month time scale. In addition, surface emissivities are sensitive to rainfall. The reasonability of the retrieved results indicates that the algorithm is feasible. A time-series surface emissivity database on the Qinghai-Tibet Plateau can be built using the developed algorithm, and then other surface or atmospheric parameters would have high retrieval precision to support related geological re- search on the Qinghai-Tibet Plateau.     

Environmental magnetic studies of sediment cores from Gonghai Lake: implications for monsoon evolution in North China during the late glacial and Holocene

Environmental magnetic studies were conducted on a 9.42-m-long sediment core from Gonghai Lake, North China. Radiocarbon dating indicates that the record spans the last 15,000 cal year BP. The principal magnetic mineral in the sediments is pseudo-single domain magnetite of detrital origin with minimal post-depositional alteration. Although the variations in the concentration of detrital magnetic minerals and their grain size throughout the core reflect inputs from both soil erosion and eolian dust, it is shown that their climatic and environmental significance changes with time. In the lowermost part of the core, ~15,000–11,500 cal year BP, the magnetic minerals were supplied mainly by bedrock erosion, soil erosion and dust input when climate ameliorated after the cold and dusty last glacial maximum. The increasing magnetic susceptibility (χ) in this interval may indicate a combination of changes in the lake environment together with catchment-surface stabilization and a decreasing proportion of dust input. In the central part of the core, ~11,500–1,000 cal year BP, the detrital magnetic minerals mainly originated from dust inputs from outside the catchment when the lake catchment was covered by forest, and catchment-derived sediment supply (and thus the lake sediment accumulation rate) were minimal. The generally low concentration of magnetic minerals in this part of the core reflects the highest degree of soil stability and the strongest summer monsoon during the Holocene. In the uppermost part of the core, the last ~1,000 years, detrital magnetic minerals mainly originated from erosion of catchment soils when the vegetation cover was sparse and the sediment accumulation rates were high. Within this part of the core the high magnetic susceptibility reflects strong pedogenesis in the lake catchment, and thus a strong summer monsoon. This scenario is similar to that recorded in loess profiles. Overall, the results document three main stages of summer monsoon history with abrupt shifts from one stage to another: an increasing and variable summer monsoon during the last deglacial, a generally strong summer monsoon in the early and middle Holocene and a weak summer monsoon in the late Holocene. The results also suggest that different interpretational models may need to be applied to lake sediment magnetic mineral assemblages corresponding to different stages of environmental evolution.     

基于均值漂移的粒子滤波算法设计及其在导航数据处理中的应用

针对标准粒子滤波算法中存在的计算量大和粒子的权值退化的缺陷,将均值漂移算法和PF算法进行融合,设计基于均值漂移搜索算法的粒子滤波新算法。该算法仍遵从粒子滤波算法的计算框架,基本原理是利用MS算法对粒子的聚类作用,将均值漂移思想融合到粒子滤波算法的重要性采样过程中,对粒子集进行确定性搜索,使每个粒子收敛于局部最优值,这样粒子的状态表示更接近真实的状态分布,因此只需较少的粒子数便可达到未嵌入MS的使用大量粒子数的粒子滤波状态估计的性能,从而在缓解粒子的权值退化的同时提高粒子滤波算法的实时性。大量的数值试验和对GPS/DR组合导航数据处理的结果验证了该方法的有效性。     

精密单点定位用户自主式完备性监测算法

精密单点定位PPP是当前GNSS高精度定位中的关键技术之一,使用的PPP采用扩展卡尔曼滤波估计,未知参数包括站点坐标,接收机钟差,对流层延迟以及虚拟未知数。在QR奇偶检校法的基础上,重点考察设计矩阵向量间的相关距离,将其作为粗差探测和识别的研究对象。通过向量相关距离时间序列,可以区分单个粗差和多个粗差的粗差集。提出精密单点定位的RAIM算法,解决了精密单点定位中的质量控制问题,使得多个粗差的识别更加清晰和快捷。     

自适应换站算法及其在长距离机载GPS动态相对定位中的应用

提出采用＂自适应换站＂法来解决单基线处理模式难以满足长距离动态定位的问题。该方法的主要思想是：整个解算过程始终保持单基线的动态定位模式,当发现流动站与原参考站间的距离大于流动站与其他参考站间的距离时,自动更换周围最近的参考站作为新参考站;对换站前后的数据设置一定时间的重叠段,利用等价消参法将换站前后的数据进行严格的自适应融合,以保持整个解算过程的连续性。通过对实测机载数据的解算,证明该方法不仅可以克服因载体飞行距离过远而导致的定位精度下降,而且能够避免因换站所引起的前后解的不连续,计算结果表明该方法的定位精度约为2 cm。     

超高阶次勒让德函数递推计算中的压缩因子和Horner求和技术

摘要：超高阶次（如2000阶次）缔合勒让德函数值的递推计算,在接近两极时达到极大的数量级（超过10的数千次方）,这导致现有递推方法在计算缔合勒让德函数值及其导数值时失效。我们通过插入压缩因子技术,提出一个修改的递推算法,并结合使用Horner求和技术计算球谐级数的部分和。试验表明,该算法至少可以计算到3600完全阶次的球谐级数式,优于现有结果。     

基于R树索引的三维场景细节层次自适应控制方法

针对大规模三维城市建模需要,介绍一种基于三维R树索引的多细节层次（简称LOD）管理方法,从叶节点层向根节点自动生成LOD场景,并设计实现LOD检索的算法。通过试验分析,证明本文的LOD定义参数能够定量控制三维场景中的渲染目标数目,进而实现三维场景的自适应可视化方法,尤其适合于建筑物和树木类型的地物目标。     

多普勒雷达降水反演及其与改进的区域气候模式RIEMS的同化研究

本博士论文在前人构建的陆面水文过程模型TOPX和区域气候模式RIEMS的耦合模式的框架基础上,针对耦合模式中RIEMS对降水和蒸散发的模拟精度较低,RIEMS和TOPX模式之间尺度不匹配等几个核心问题进行了深入细致的研究:对天气雷达资料的定量估测降水方面的研究获得了对沂沭河流域最佳的Z-R关系,并通过雷达雨量计联合校正法得到较高精度的面降雨量;采用集合卡尔曼滤波同化算法对雷达反演的面雨量及区域气候模式RIEMS的降水输出进行了数据同化方案研究,获得了更好的降水模拟效果;在RIEMS和水文模型TOPX构成的耦合模型中加入同化方案,实现了流域降水的实例模拟研究,结果表明使用同化方案明显改善了水文模拟效果。