高分多模卫星大气校正总体设计与在轨验证

高分多模卫星搭载中国首台民用大气同步校正仪,可以对同平台的高分辨率相机观测区域进行时间同步、视场覆盖的大气参数测量,实现对高分辨率图像的大气校正。文章论述了高分多模卫星大气同步校正的总体设计思路,并给出了时空同步、多谱段、多偏振通道探测体制的设计方案和地面验证结果。高分多模卫星入轨以后,对大气同步校正仪的大气参数反演效果和高分辨率图像的校正效果进行了分析,结果表明利用大气同步校正仪测量数据对大气气溶胶光学厚度（AOD）和大气水汽含量（CWV）的反演结果可信,对高分辨率图像有较好的大气校正效果。     

由现场离水辐亮度估算黄海透明度几种方法的比较

本文根据黄海现场离水辐亮度及遥感反射比,采用４种统计算法估算海水透明度（Ｓｅｃｃｈｉ　Ｄｉｓｃ　Ｄｅｐｔｈ）。前２种算法中海水透明度作为因变量,两不同波段辐亮度之比及两辐亮度之和与第三者之比分别作为自变量,进行一元回归分析,前者相关系数ｒ２为０．９３３５。后者相关系数ｒ２为０．９４８３。第３种算法中,所有波段的遥感反射比做因子分析形成的主因子作为自变量,透明度作为因变量进行多元回归分析,子样全相关系数ｒ２为０．９６４９。第４种算法中,遥感反射比作为自变量,透明度作为因变量进行逐步回归分析,复相关系数ｒ２为０．９８１５。计算结果表明,多元统计分析的估算精度高于一元回归分析。     

用EOS/MODIS资料反演积雪深度参量

利用EOS/MODIS可见光、近红外及短红外多通道资料以及新疆地区积雪深度气象台站实测资料等,在考虑积雪性质包括积雪粒子相态、积雪年龄等的差异以及积雪区的下垫面条件包括地表粗糙度、土地覆盖类型等的不同的情况下进行积雪分类,在此基础上,建立EOS/MODIS积雪深度反演模型,实现深度在30 cm以内的积雪深度反演的主要原理、思路及方法,并对模型的反演结果进行了验证。结果表明,利用该模型对30 cm以内的积雪进行深度反演计算,其精度能达到80%以上。     

Field Studies of Dynamic Compaction on Marine Deposits

Marine deposit ground usually need significant improvement before the construction of civil structures in coastal areas due to the poor mechanical properties of soils. Dynamic compaction (DC) is a widely used technique in such projects. In this study, a case history of DC tests in sandy soils with a weak embedded layer is introduced. Two series of DC tests—single point tests and impact zone tests—were applied to test zones with similar geological conditions to investigate the effect of energy level on the depth of improvement (DI). The highest energy used is up to 15000 kN · m. Field measurements were conduct before and after DC in each series to validate the effectiveness of improvement, including crater settlement, excessive pore pressure, and acceleration measurement for single point tests, and the pressure meter and CPT tests for impact zone tests. For single point tests, the effectiveness of improvement increases as the energy level increases to 12000 kN · m. Further increase of compaction energy does not have an effect on settlement, pore pressure, or ground acceleration. For impact zone tests, the energy level does not show a positive correlation with the DI, mainly due to the presence of an embedded weak layer.     

Bathymetric distribution patterns of Southern Ocean macrofaunal taxa: Bivalvia, Gastropoda, Isopoda and Polychaeta

The aim of this study is to compare the depth distributions of four major Southern Ocean macrobenthic epi- and infaunal taxa, the Bivalvia, Gastropoda, Isopoda, and Polychaeta, from subtidal to abyssal depth. All literature data up to summer 2008, as well as the unpublished data from the most recent ANDEEP I–III (Antarctic benthic deep-sea biodiversity: colonisation history and recent community patterns) expeditions to the Southern Ocean deep sea are included in the analysis. Benthic invertebrates in the Southern Ocean are known for their wide bathymetric ranges. We analysed the distributions of four of the most abundant and species-rich taxa from intertidal to abyssal (5200 m) depths in depth zones of 100 m. The depth distributions of three macrofaunal classes (Bivalvia, Gastropoda, Polychaeta) and one order (Isopoda) showed distinct differences. In the case of bivalves, gastropods and polychaetes, the number of species per depth zone decreased from the shelf to the slope at around 1000 m depth and then showed stable low numbers. The isopods showed the opposite trend; they were less species rich in the upper 1000 m but increased in species numbers from the slope to bathyal and abyssal depths. Depth ranges of families of the studied taxa (Bivalvia: 31 families, Gastropoda: 60, Isopoda: 32, and Polychaeta: 46 families) were compiled and illustrated. At present vast areas of the deep sea in the Southern Ocean remain unexplored and species accumulation curves showed that only a fraction of the species have been discovered to date. We anticipate that further investigations will greatly increase the number of species known in the Southern Ocean deep sea.     

基于CyGNSS数据的土壤水分与植被光学厚度反演研究

本文提出了一种仅基于CyGNSS数据,能够同时反演土壤水分与植被光学厚度的方案,该方案使用了神经网络与暴力穷举算法.首先考察了2018年以及2020年的数据,并对结果进行了验证.通过分析发现反演结果与参考数据展现了良好的一致性.土壤水分的反演结果与2018年和2020年的测试数据比较,其相关系数分别高达0.86和0.84,均方根误差分别为0.064和0.071 cm³/cm³;对于植被光学厚度,2018年与2020年的相关系数均为0.98,均方根误差分别为0.079和0.084.研究结果表明,CyGNSS可作为一种新型且独立的泛热带土壤水分与植被光学厚度反演手段.     

Study of the spatial and historical distribution of sediment inorganic contamination in the Toulon bay (France)

This study presents trace elements levels in surface and deep sediments of the Toulon bay (SE France) subjected to anthropogenic inputs (navy base, harbors, etc.). The studied elements (As, Cd, Cr, Cu, Hg, Ni, Pb, Zn) are defined as priority contaminants in aquatic systems. Fifty-five points scattered on the entire bay were sampled, allowing the determination of contaminants distribution with a high resolution. Several approaches were used to assess the degree of contamination and the potential toxicity of the Toulon bay sediments: comparison to the French legislation, surface-weighted average metal concentrations, enrichment factors (EF), geoaccumulation indices (Igeo), trace element stock calculation and comparison to sediment quality guidelines. A principal component analysis was performed to reveal common behavior of the studied contaminants. Results demonstrated the very high contamination of the small bay, especially in Hg (EF up to 1500), Cu, Pb and Zn, with export to the large bay further governed by hydrodynamics.     

地球静止卫星和网格化站点支持下的PM2.5精细制图

许多城市建立的相对稠密的网格化监测站点,为精细化监管城市空气质量奠定了基础。本文选用徐州市网格化监测数据、地球静止卫星Himawari-8/AHI及COMS/GOCI的表观反射率和气溶胶光学厚度数据、气象和其他辅助数据,开展了徐州地区0.005°空间分辨率网格的PM_2.5浓度精细化制图研究。本文使用了极端梯度提升（XGBoost）、随机森林（RF）及时空加权回归（GTWR）等3种方法,并选用多种特征参数组合进行对比分析。综合分析模型精度和过拟合程度,结果表明XGBoost模型表现最好,其R²为0.90,RMSE为11.65 μg/m³。进一步将本文结果与国控站点、清华大学的TAP数据集和马里兰大学的CHAP数据集的对比分析,结果表明基于网格化站点的PM_2.5制图结果能更好地反映城市内部不同区域的PM_2.5浓度分布差异性,弥补因国控站点稀疏带来的缺陷,更好地服务于城市空气质量精准管控。     

接收函数反演上地幔速度结构和间断面的剥壳遗传算法

探讨了使用台站接收函数反演得到上地幔速度结构和间断面深度的一种方法——剥壳遗传算法. 由地表至深部分成几个深度段，逐段反演速度结构，并在接收函数曲线上剥离浅处间断面的多次反射震相和反射转换震相对深部间断面的一次透射转换相的影响. 在每一深度段使用浮点编码、算术杂交和非均匀变异的遗传反演方法；在10次独立反演后，依据适应度选出50个优秀模型，对其加权平均得到该段的最终速度结构. 所形成的方法和程序，用iasp’91模型进行数值检验之后，又反演了海拉尔台的观测接收函数，给出台下深至700km的速度结构和主要间断面埋藏深度及速度跃变量.     

重力数据三维共轭梯度聚焦反演及应用

基于共轭梯度算法对欠定线性目标函数进行求解。 为改善目标函数的多解性、 消除多余构造信息影响, 引入粗糙度系数矩阵; 为克服“趋肤效应”, 更好地反映地质体的真实形态, 在模型目标函数中引入深度加权函数; 为更好的反映地质体的某些尖锐构造和边界, 本文对目标函数添加了基于最小支撑泛函的聚焦反演约束。 通过对多种模型的计算, 验证了该方法具有较好的有效性和稳定性, 并将该方法应用于实际重力资料地下密度反演中去, 得到了较好的反演结果。