SMAP卫星的RBF神经网络海表盐度遥感反演

针对传统海表盐度遥感反演精度不高、影响因素较少等问题,本文基于SMAP（Soil Moisture Active Passive）卫星L2C（Level 2 C）数据、Argo（Array for Real-time Geostrophic Oceanography）数据和其他辅助数据,以太平洋部分海域（160°E～120°W,0°～30°N）为研究区域,综合考虑海面粗糙度以及白冠覆盖率等参量,利用径向基神经网络建立RBF亮温增量模型,并对平静海面亮温进行修正,然后基于Meissner-Wentz介电常数模型得到反演后的盐度值。验证结果表明：模型预测盐度和SMAP卫星盐度相对于Argo实测盐度的均方根误差分别为0.4和0.5,平均绝对误差分别为0.3和0.4。实验证明,利用RBF神经网络建立的亮温增量模型可以提高海表盐度反演的精度,对海表盐度反演具有实用意义。     

基于SMOS卫星数据的BP神经网络盐度反演模型

海表面盐度SSS(Sea Surface Salinity)是研究大洋环流和海洋对气候影响的重要参量、是决定海水基本性质的重要因素之一。卫星微波遥感可以满足盐度研究过程中大范围、连续观测的需要。目前,由欧洲空间局设计开发的SMOS(Soil Moisture and Ocean Salinity)卫星于2009年发射成功,并且根据它的观测数据和物理机制反演出了海表面盐度的相关产品。但结果显示,在某些近海岸区域(如中国南海海域)受陆地RFI等诸多因素的影响,基于卫星遥感物理机制反演得到的盐度产品的精度较低。本文的主要目的是利用中国海洋大学"东方红2"科学考察船的走航数据、SMOS卫星数据,针对中国南海海域提出了用BP神经网络预测海表面盐度的方法,并用实测Argo浮标、WOA13的盐度数据对模型进行适应性评估。结果表明,模型产品相对于"东方红2"实测盐度数据的均方根误差(RMSE)是0.21,而SMOS的SSS1产品、SSS2产品和SSS3产品的精度分别为1.90、1.93和1.91。同时,在验证数据集中,模型预测数据相对于Argo浮标实测盐度数据的均方根误差(RMSE)是0.50,而SMOS的SSS1产品、SSS2产品和SSS3产品的精度分别为1.83、1.83和1.84。此模型具有良好的适应性和泛化能力,为海表面盐度的反演和预测提供了一个不依赖于物理机制的新方法。     

基于Argo的SMAP与SMOS盐度产品质量评估与比对

以2016年SMAP 8日均、月均海表盐度(Sea Surface Salinity, SSS)产品与SMOS日均、月均SSS产品为研究对象,基于Argo浮标实测盐度数据与Argo月均盐度网格产品,进行了质量评估,并开展了SMAP和SMOS盐度遥感产品间的交叉比对。结果表明:SMAP 8日均、月均SSS产品均比SMOS日均、月均SSS产品更接近于Argo盐度数据;SMAP SSS产品在12个月里的标准差均小于SMOS SSS产品,SMAP SSS产品年均偏差与标准差在大部分海域小于SMOS SSS产品,除高纬度地区,4种SSS产品与Argo数据的平均偏差和标准差随纬度变化浮动不大;SMAP 8日均、月均与SMOS日均、月均产品的年平均SSS空间分布特征一致,SMAP 8日均与SMOS日均、SMAP与SMOS月均SSS产品间的年平均偏差在全球范围内总体在-0.50～0.50,标准差总体为0～1.00。     

海表面盐度的统计预测模型

海表面盐度SSS(Sea Surface Salinity)是研究大洋环流和海洋对全球气候影响的重要参量。海表面盐度卫星遥感探测可以满足大范围、连续观测的研究需要,是获取该参量的有效手段。2009年欧洲空间局发射了SMOS(Soil Moisture and Ocean Salinity)卫星,并且根据卫星观测的数据反演出海表面盐度的相关产品,但是产品的精度还有待于进一步的提高。本文利用多元线性回归的统计分析方法,针对SMOS卫星相关数据(观测亮温数据和辅助数据)建立一种全新的不依赖于物理机制的海表面盐度统计模型。本文针对太平洋中部提出的统计模型计算的盐度点对点的精度为0.165 5psu,1°×1°月平均精度为0.106 3psu,而SMOS卫星Level 2盐度产品的精度分别为0.585 5和0.181 9psu。同时将模型应用到验证数据集,得到了点对点精度为0.224 2psu,进一步说明模型具有很好的适应性和泛化能力。     

基于风带的海表面亮温增量统计模型

海表面盐度(Sea Surface Salinity, SSS)是研究大洋环流和海洋对全球气候影响的关键参数之一。目前借助卫星遥感技术获取全天候和连续的SSS是最有效的方法,但是SSS的反演精度在大部分海域达不到预期目标。众所周知,海表面亮温是反演SSS的关键因素之一,海面粗糙度导致了亮温增量的产生,亮温正演模型的误差会影响盐度反演的精度。本文首次提出了依据6个风带划分全球海域,利用Argo实测盐度数据、SMOS卫星数据和相关辅助数据,通过LASSO统计方法在各风带覆盖的海域构建了一个全新的二次曲线亮温增量模型,再通过贝叶斯迭代反演算法计算出了各个海域的SSS产品。与Argo实测SSS对比,新模型下6部分海域反演SSS的绝对平均误差分别为0.76、0.88、0.93、0.92、1.28和1.21,均显著优于修正前(SMOS L2 SSS)产品的误差(0.98、1.61、2.82、1.50、2.35和3.13)。     

基于现场观测资料的卫星海表盐度网格化产品误差分析与质量评估

自欧洲土壤湿度和盐度卫星SMOS和美国宝瓶座盐度卫星Aquarius相继发射之后,多个数据中心发布了两颗卫星的海表盐度网格化产品,其中包括法国海洋研究院SMOS卫星数据小组发布SMOS Locean L3盐度产品、西班牙巴塞罗那专家中心发布SMOS BEC L4盐度产品和美国宇航局喷气动力实验室发布AquariusV3.0 CAP L3盐度产品。本文利用精确盐度现场观测资料从产品精度和模拟海洋现象能力两个方面对以上3种产品质量进行了评估。研究表明:(1) 在精度方面,与盐度现场资料相比,Aquarius CAP 产品质量最高,产品盐度偏差和均方根误差全年稳定且偏差较小,部分海域达到了设计精度;SMOS两种卫星产品在全球海域偏差较不稳定,个别月份出现异常偏差值;SMOS产品在低纬和开阔海域的数据质量相对较高,但在高纬海域仍存在较大误差,需要进一步提升;(2) 在刻画海洋现象方面,Aquarius产品在热带太平洋较好刻画了淡池东缘盐度锋,SMOS BEC产品的刻画能力次之,SMOS Locean产品在热带太平洋充满了小尺度噪音,描述物理现象方面表现偏差。     

SMOS卫星盐度数据在中国近岸海域的准确度评估

盐度是描述海洋的关键变量,对海表面盐度进行观测可以推进对全球水循环的理解。本文的主要目的是在中国近海海域对SMOS卫星盐度数据进行准确度评估。主要方法是将SMOS卫星L2海洋盐度数据产品(V317)与实测ARGO数据和走航数据进行匹配,并采用统计学的方法对SMOS卫星数据准确度进行评估。结果表明:匹配数据的线性关系不显著,SMOS卫星盐度数据(V317)在南海和东海的均方根误差分别约为1.2和0.7,应用海表面粗糙度修正模型得到的3组海表盐度数据准确度都相对较低,尤其在近岸强风场区域,海表盐度卫星数据相对于实测数据偏高,这可能是由于海表粗糙度和陆地射频干扰(RFI)作用影响的结果;SMOS卫星数据在东海的均方根误差比南海高0.5左右,这可能是由于东海海域为相对开阔海域,受陆地RFI影响相对南海较小;在中国近岸海域,应用SSS1和SSS3模型得到的盐度数据准确度相对较高,可以对模型进行地球物理参数修正,进行局地化改进,预计可以提高近岸海域盐度反演的准确度。     

基于SMOS卫星数据的海表面盐度模型

海表面盐度是研究海洋对全球气候影响以及大洋环流的重要参量之一,而卫星遥感技术是获取海表面盐度数据的最有效方法.目前,L波段的SMOS和Aquarius/SAC-D遥感卫星正在用于探测海表面盐度,并根据卫星观测数据和物理机制反演出海表面盐度的产品.但在某些近陆地区域,由于淡水流入及陆地射频(RFI)等因素影响,卫星反演盐度的产品精度较低.文中利用“东方红2号”科学考察船的实测数据、SMOS卫星数据,首次针对中国南海海域提出了用贝叶斯网络模型计算海表面盐度,并用验证数据集(实测Argo盐度)对模型进行适应性评估.经过计算,模型误差和验证误差分别为0.47 psu和0.45 psu,而相应的SMOS Level 2产品的精度分别为1.90 psu和1.82 psu.此模型为海表面盐度的计算提供了一个新方法.     

海表面盐度的高精度预测模型

为了建立高精度的海洋表面盐度预测模型,采用BP神经网络的方法,针对SMOS卫星level 1C级亮度温度数据和辅助数据建立了一种海表面盐度预测模型,以ARGO浮标观测值作为海表盐度实测值来检验新模型预测结果的准确度,同时利用验证集对模型的精度进行验证。结果表明:通过新模型预测的海表盐度(SSS0)比SMOS卫星的3个粗糙度模型盐度产品(SSS1,SSS2,SSS3)精度高;SSS0,SSS1,SSS2,SSS3与ARGO浮标实测盐度(SSS ARGO)的均方根误差分别为0.8473,2.0417,2.0288和2.0805,平均绝对误差分别为0.7553,1.4226,1.4216和1.4566,SSS0与SSS ARGO的均方根误差和绝对平均误差值都明显小于SSS1,SSS2和SSS3与SSS ARGO的;由此可见,建立的海表盐度预测模型精度较高。新模型为海表盐度的反演算法提供了新思路。     

基于Aquarius 卫星数据的孟加拉湾海表盐度分析

海洋表面盐度(Sea Surface Salinity,SSS)是海洋的重要物理和化学参量,SSS的时空分布与全球大洋环流和水汽循环密切相关。本文基于美国国家航空航天局(NASA)发射的Aquarius卫星3 a的SSS遥感数据,给出了孟加拉湾及其附近海域海表盐度的空间分布特征,并重点分析了影响孟加拉湾海表盐度变化的可能因素。研究结果从一个侧面说明了利用Aquarius卫星遥感观测海洋大尺度盐度变化的可行性。     

Sea Levels and Coastal Inundation Due to Tropical Cyclones in Indian Coastal Regions of Andhra and Orissa

Coastal inundation associated with extreme sea levels is the main factor which leads to the loss of life and property whenever a severe tropical cyclonic storm hits the Indian coasts. The Andhra and Orissa coasts are most vulnerable for coastal inundation due to extreme rise in sea levels associated with tropical cyclones. Loss of life may be minimized if extreme sea levels and associated coastal flooding is predicted well in advance. Keeping this in view, location specific coastal inundation models are developed and applied for the Andhra and Orissa coasts of India. Several numerical experiments are carried out using the data of past severe cyclones that struck these regions. The simulated inland inundation distances are found to be in general agreement with the reported flooding.     

Sea Levels and Western Boundary Flow Near a Tectonic Zone in the Northern Pacific

This article concerns an interrelation between the sea levels and the western boundary flow near a tectonic boundary in a local zone in the Northwestern Pacific. In this zone, sea level variations at stations located on the coast facing the Pacific are studied to find the interrelation between variations of the Kurosio flow as an index of the distance of the flow axis off a specific coast. The result is discussed after data processing of the monthly means of the sea levels, and a notice is taken of variations caused by active crustal upheavals during a seismic event, a local earthquake.     

Role of Siliceous Fossil Dissolution in Downcore Variations of Grain Size and Water Content: Western Margin of Clarion-Clipperton Fracture Zone, NE Equatorial Pacific

Grain size and water content in box-core sediments from the Clarion-Clipperton fracture zone (C-C zone) in the northeast equatorial Pacific were analyzed in detail to understand the downcore variations across a hiatus between Quaternary and Tertiary layers. Grain-size distributions in the topmost core sediments show two modes: a coarse mode (peaked at 50 μm) and a fine mode (at 2-25 μm). The coarse mode disappears gradually with depth accompanied by the dissolution of siliceous fossil tests, whereas the fine mode coarsens due to the formation of authigenic minerals. Water content increases abruptly across a color boundary between an upper pale brown layer and a lower dark brown layer that is the hiatus between Quaternary and Tertiary layers. Abundant smectites and microvoid molds, which are created by the prolonged fossil dissolution in the underlying sediment, are attributed for the abrupt downcore variation of water content. Overall variations in grain size and water content in the topmost core sediments in the western C-C zone are possibly constrained by the dissolution of biogenic siliceous fossils. Variations in geotechnical properties related to these changes must be considered in the design of nodule collectors.     

A Nondestructive Technique for Predicting the In Situ Void Ratio for Marine Sediments

This study tests the hypothesis that the in situ void ratio of surficial marine sediments may be predicted from shear wave velocity-depth data with a reliability equal to that of other methods currently available. Shear wave velocity is fundamentally controlled by the number of grain-to-grain contacts per unit volume of material and by the effective stress across those contacts. In this study, three previously established empirical formulae are used to predict void ratio from velocity-depth data. Field data were acquired along a transect off the northern Californian coast across which water depth increased from 35 to 70 m and seafloor sediment type varied from sand to silty-sand, respectively. A towed seafloor sled device was used to collect shear wave refraction data, and a marked, systematic decrease in velocity was observed along the line, ranging from 35-70 m/s for the coarse, near-shore material to 25-40 m/s for the finer, offshore deposits. Void ratios predicted from these velocities were compared with data measured directly from box-core samples. Of the formulae used for prediction, two agree remarkably well with the control data. Both predicted and control values increase from 0.6-0.8 for the sandy material to 1.1-1.5 for the silty-sand. Thus, this study does not disprove the hypothesis set and demonstrates the potential of field shear wave velocity-depth data as a means of delineating spatial variation in void ratio for surficial marine sediments in a remote, nondestructive manner.     

Particulate n-alkanes and fatty acids in the Changjiang river system

Particulate samples were collected from the Changjiang river system during a flood period, in May 1997, and POC, stable isotope and lipids associated with particles were examined. Results showed the decrease (0.84% ～ 1.88%) of organic carbon content from the upper reaches to the estuary.δ13C values of particulate organic carbon was in the range of -24.9×10-3 to -26.6×10-3, which were close to the isotopic signature of continental C3 vegetation. Total particulate n-alkanes concentrations varied from 1.4 to 10.1μg/dm3,or from 23.7 to 107μg/g of total suspended matter. Fatty acids were present in all the samples, from 1.4 to 5.4μg/dm3, with saturated and unsaturated straight-chain and branched compounds in the carbon number range from C12 to C30. Both δ13C and the ratio of carbon content to nitrogen content indicate the predominance of terrestrial inputs (soil organic matter) among the particles. The biomarker approach has been used to identify the relative portion of terrigenous and autochthonous fraction in the particulate samples. The distribution of fatty acids suggests a striking phytoplanktonic and microbial signal in most particle samples. The terrestrial alkanes are used to estimate the contribution of terrestrial inputs along the mainstream.     

Three-Dimensional Geographically Weighted Inverse Regression (3GWR) Model for Satellite Derived Bathymetry Using Sentinel-2 Observations

Current trends of development of satellite derived bathymetry (SDB) models rely on applying calibration techniques including analytical approaches, neuro-fuzzy systems, regression optimization and others. In most of the cases, the SDB models are calibrated and verified for test sites, that provide favorable conditions for the remote derivation of bathymetry such as high water clarity, homogenous bottom type, low amount of sediment in the water and other factors. In this paper, a novel 3-dimensional geographical weighted regression (3GWR) SDB technique is presented, it binds together methods already presented in other studies, the geographically weighted local regression (GWR) model, with depth dependent inverse optimization. The proposed SDB model was calibrated and verified on a relatively difficult test site of the South Baltic near-shore areas with the use of multispectral observations acquired by a recently launched Sentinel-2 satellite observation system. By conducted experiments, it was shown that the proposed SDB model is capable of obtaining satisfactory results of RMSE ranging from 0.88 to 1.23[m] depending on the observation and can derive bathymetry for depths up to 12m. It was also shown, that the proposed approach may be used operationally, for instance, in the continuous assessment of temporal bathymetry changes, for areas important in the context of ensuring local maritime safety.     

Geotechnical Properties of Surface Sediments in the INDEX Area

As a part of the environmental impact assessment studies, geotechnical properties of sediments were determined in the Central Indian Basin. The undrained shear strength and index properties of the siliceous sediments were determined on 20 box cores of uniform dimension collected from various locations in five preselected sites. The maximum core length encountered was 41 cm and most of the sediments were siliceous oozes consisting of radiolarian or diatomaceous tests. The shear strength measurements revealed that surface sediments deposited in recent times (0-10 cm) have a shear strength of 0-1 kPa; this value increases with depth, reaching 10 kPa at 40 cm deep. Older sediments have greater strength because of compaction. Water content varies in the wide range of 312-577% and decreases with depth. The clay minerals such as smectite and illite are dominant and show some control over water content. Wet density, specific gravity, and porosity do not indicate any notable variation with depth, thereby indicating a uniform, slow rate of sedimentation. The average porosity of sediments is 90.2%, specific gravity 2.18, and wet bulk density 1.12 g/cm 3 . Sediments exhibit medium to high plasticity characteristics, with the average plasticity index varying between 105% and 136%. Preliminary studies on postdisturbance samples showed an increase in natural water content and a decrease in undrained shear strength of sediments in the top 10- to 15-cm layer.     

Ferromanganese Nodules and their Associated Sediments from the Central Indian Ocean Basin: Rare Earth Element Geochemistry

The rare earth element (REE) distribution in nine deep-sea ferromanganese nodules and their associated siliceous sediments from the Central Indian Ocean Basin (CIOB) have been studied to elucidate the REE relationship among them. Total REE concentration varies from 398-928 ppm in the nodules and 137-235 ppm in the associated sediments, suggesting two- to four-fold enrichment in the nodules compared to associated sediments. REE of nodules and their associated sediments show a positive correlation, suggesting REE are supplied from a common source such as seawater. The positive correlation between REE of nodules and sediments from the CIOB is contrary to the competitive scavenging of REE between nodules and sediment in the equatorial Pacific Ocean. REEs in the nodules are carried by Fe, P, and Ti, whereas in the sediment they are carried by P and Mn phases. A similar REE fractionation pattern with middle REE enrichment over heavy and light REE in both the nodules and their associated sediment suggest fractionation is independent of REE abundance and their carrier phases.     

Marine Minerals in the Mexican Pacific: Toward Efficient Resource Management

Vast reserves of marine minerals are known to occur within the Mexican Exclusive Economic Zone. These reserves include: (1) heavy mineral placers rich in titanium and iron along the coast of San Antonio del Mar, Baja California and the southern States of Oaxaca and Guerrero; (2) phosphorites, which represent the only source for superphosphate fertilizers along the Pacific coast off the Baja California peninsula and the Gulf of Tehuantepec; and (3) manganese oxide deposits in nodule and crusts located adjacent to the Clarion-Clipperton Zone and containing high concentrations of Cu + Ni + Co (up to 1.9%), which are essential for the steel and super-alloy industries. Few but important efforts toward surveying these deposits have been carried out by Mexican and international researchers. Relevant data generated by these investigations are reviewed here and put together in a single document. Adequate knowledge of Mexican mineral resources is essential for the development of management strategies when offshore mining starts during this century. Thus, the review discusses the economics inherent to ocean mining in Mexico, setting baselines for future exploration and development activities in the Mexican Exclusive Economic Zone.     

Enhanced Acoustic Backscatter Due to High Abundance of Sand Dollars, Dendraster excentricus

Detailed acoustic surveys of benthic sediments were conducted in July 1995 and September 1998 in the vicinity of Humboldt Bay, California. During these surveys, a band of enhanced acoustic backscatter was observed offshore from the bay entrance, approximately parallel to the isobaths, in water depths ranging from 16-24 m. In order to assess the cause of the increase in backscatter levels, a more comprehensive study was conducted in August and September 1999 using 100 kHz side-scan sonar, bottom grab sampling and underwater video recording. New observations indicated that a dense population of sand dollars ( Dendraster excentricus ) coincided with the enhanced backscatter band. Compared to the two previous acoustic studies, the central section of the band expanded westward by 180 m and the southern section of the band shifted eastward by 160 m, possibly resulting from a change in the biological or physical factors which influence the location and breadth of sand dollars. The relationship between high sand dollar abundance and enhanced acoustic backscatter was further verified in the nearshore region off Samoa Beach California, where a dense, banded population of sand dollars was previously observed. Video footage confirmed the presence of a band of sand dollars, also nominally parallel to the isobaths, in water depths of 8-15 m. A band of enhanced backscatter coincided with the dense sand dollar population. The identification of dense aggregations of sand dollars through enhanced acoustic backscatter could lead to the use of acoustic techniques to study sand dollar distributions and abundance.