求解GM(1,1)模型新总体最小二乘算法

在GM(1,1)模型中系数矩阵和观测向量都是由原始序列组成的。系数矩阵中同样是有误差的,与观测向量中的误差一样,亦来源于原始序列,即它们误差同源。不同位置的相同元素应该有相同的改正数,采用传统总体最小二乘求解则不能达到此目的。针对这一缺陷,推导了一种新的总体最小二乘算法;并且通过算例验证了新方法的可行性和有效性。     

水体悬浮颗粒物中溶解态黑碳(DBC)的含量估算——以渤海为例

溶解态黑碳(DBC),通常被认为是黑碳或降解产物中可溶解于水的组分.作为一类具有高疏水性稠环芳香结构的物质, DBC完全可能以吸附在悬浮颗粒物或颗粒态有机物的形式存在(PDBC),并且不等同于黑碳颗粒物(PBC).然而,目前环境样品中测量黑碳的主流方法,尤其是颗粒物的处理方式,并未对这一部分黑碳进行测定.因此,自然界中DBC的量在很多研究中有可能被低估,使得目前关于DBC地球化学行为以及定量研究结果的可靠性存疑.本研究基于之前在渤海工作的详细数据,包括PBC、 DBC,以及颗粒有机碳含量、悬浮颗粒物浓度、多环芳烃含量,通过对比实测多环芳烃固液分布情况,圈定出渤海表层水体中PDBC固液分布系数(Kd)大致在0.018～0.072之间,含量在1.637～6.449μg C L^-1之间.同时,估算出渤海水体悬浮颗粒物上PDBC储量(2.049～8.194Gg),低于PBC储量(15.16Gg),却同属一个数量级,该结果表明, PDBC是渤海水体环境中DBC的重要存在形式.明确DBC在水体中的吸附解吸附过程是厘定DBC源汇定量关系以及完善海洋DBC迁移转化过程的重要一环.     

基于机器学习的区域工程地质分层思路与方法研究

传统的地质分层方法通常依赖于人工解释和经验判断,存在过程繁琐、工作量大、受人为因素影响较大等缺点。本文旨在研究基于机器学习的区域工程地质分层的思路与方法,将地层分层问题转换为地质体空间单元的序列到序列预测任务及地质属性特征分类任务,以提高地质分层的准确性和效率。本文在对工程地质原始数据开展深入研究的基础上,提出采用深度学习方法对静力触探试验数据进行地质分层的训练和预测,通过分类算法对取土孔土样进行分层,并定量评价了相关算法的适用性。通过对比传统方法和机器学习方法的结果,验证了机器学习在地质分层中的优势。本研究为区域工程地质研究提供了一种新的地质分层思路,具有重要的实践意义。     

海洋信息通信网不同VPN实例之间 互通问题的研究

主要对海洋信息通信网中MPLS VPN技术进行了研究,通过分析海洋信息通信网网络现状、MPLS VPN技术在海洋信息通信网中的架构设计,以及RD值与RT值在VPN实例中的重要作用,以实验的方式,对不同VPN实例的Import RT和Export RT匹配策略进行了验证。     

2019年《地理学报（英文版）》办刊进展

6月20日,科睿唯安发布了2019年《期刊引证报告》(Journal Citation Reports,JCR),该报告共收录全球科技期刊及人文社科期刊(SCI、SSCI)12515种,其中中国大陆224种。JCR公布的数据显示:《地理学报(英文版)》(Journal of Geographical Sciences,JGS)2018年度影响因子为2.347,在JCR收录的12515种期刊中排名第4361位.     

基于多元分析的地铁车站深基坑形变规律研究

鉴于深基坑变形监测中观测数据相互关联影响,结合地下管线与桩体位移等形变数据,将多元线性回归模型与灰色模型GM(1,N)应用于深基坑形变数据预测。本文根据深基坑监测基础数据进行多种形变数据的规律性分析;应用两种分析模型预测了五种基坑形变数据,将预测结果与实测值进行对比分析。结果表明,基坑监测设计合理,开挖期间形变不大、较稳定;比较两种模型,随时间生成动态参数的GM(1,N)模型具有更小的残差,能较好地预测基坑变形规律,可以用于相关项目的实施与数据分析。     

基于GIS地理空间关系的消防审查知识图谱构建

BIM具有精细三维建模的优点,能够通过计算机平台展示消防设备的具体细节,GIS具有强大的空间分析功能,可以反映出BIM模型内构件的空间关系^[1]。随着消防审查技术要求的日益提高,审查人员在消防设备的运营和管控中,大都缺少对消防设备地理空间信息属性的审查环节,这无疑会对审查结果的准确性造成一定影响。本研究针对上述问题,提出将建筑信息建模(BIM)技术与地理信息系统(GIS)技术相结合,实现这两种技术的相辅相成,并引入了知识图谱来改良建模。该方法以《建规》为基础语料库,通过改良的Methontology法,实现半自动化的图谱本体构建;接着采用Stanford大学开发的DeepDive新型数据管理系统,实现消防规范的特征关系抽取;最后在KGB(Knowledge Graph Builder)智能开发平台完成知识图谱的构建。该研究的意义在于一方面能够促进建筑行业与GIS技术的结合,另一方面也能够推动知识图数据库与文本规则条例的融合配对,实现综合楼宇消防构件的实景三维自动化审查。     

《应用海洋学学报》征稿简则

《应用海洋学学报》是国家海洋局第三海洋研究所、中国海洋学会、福建省海洋学会主办,中国生态学学会海洋生态专业委员会协办,国家海洋局主管,国内外公开发行的全国性综合性海洋科学学报(季刊,国内统一刊号:CN35-1319/P,国际标准刊号:ISSN2095-4972)。《应用海洋学学报》是中文核心期刊、中国科技核心期刊、中国科学引文数据库(CSCD)收录期刊、中国学术期刊(光盘版)收录期刊和中国核心期刊(遴选)数据库收录期刊,也是中国、美国、英国、日本以及联合国等十几个科技期刊数据库的收录期刊。     

Numerical simulation and preliminary analysis of typhoon waves during three typhoons in the Yellow Sea and East China Sea

In this study,typhoon waves generated during three typhoons(Damrey(1210),Fung-wong(1416),and Chan-hom(1509))in the Yellow Sea and East China Sea were simulated in a simulating waves nearshore(SWAN)model,and the wind forcing was constructed by combining reanalyzed wind data with a Holland typhoon wind model.Various parameters,such as the Holland fitting parameter(B)and the maximum wind radius?,were investigated in sensitivity experiments in the Holland model that affect the wind field construction.Six different formulations were considered and the parameters determined by comparing the simulated wind results with in-situ wind measurements.The key factors affecting wave growth and dissipation processes from deep to shallow waters were studied,including wind input,whitecapping,and bottom friction.Comparison with in-situ wave measurements suggested that the KOMEN scheme(wind input exponential growth and whitecapping energy dissipation)and the JONSWAP scheme(dissipation of bottom friction)resulted in good reproduction of the significant wave height of typhoon waves.A preliminary analysis of the wave characteristics in terms of wind-sea and swell wave revealed that swell waves dominated with the distance of R to the eye of the typhoon,while wind-sea prevailed in the outer region up to six to eight times the R values despite a clear misalignment between wind and waves.The results support the hypothesis that nonlinear wave-wave interactions may play a key role in the formation of wave characteristics.     

Contribution of trimethylamine N-oxide on the growth and pressure tolerance of deep-sea bacteria

Trimethylamine N-oxide(TMAO) is widely dispersed in marine environments and plays an important role in the biogeochemical cycle of nitrogen. Diverse marine bacteria utilize TMAO as carbon and nitrogen sources or as electron acceptor in anaerobic respiration. Alteration of respiratory component according to the pressure is a common trait of deep-sea bacteria. Deep-sea bacteria from dif ferent genera harbor high hydrostatic pressure(HHP) inducible TMAO reductases that are assumed to be constitutively expressed in the deep-sea piezosphere and facilitating quick reaction to TMAO released from ?sh which is a potential nutrient for bacterial growth. However, whether deep-sea bacteria universally employ this strategy remains unknown. In this study, 237 bacterial strains affliated to 23 genera of Proteobacteria,Bacteroidetes, Firmicutes and Actinobacteria were isolated from seawater, sediment or amphipods collected at dif ferent depths. The pressure tolerance and the utilization of TMAO were examined in 74 strains. The results demonstrated no apparent correlation between the depth where the bacteria inhabit and their pressure tolerance, regarding to our samples. Several deep-sea strains from the genera of Alteromonas, Halomonas,Marinobacter, Photobacterium, and Vibrio showed capacity of TMAO utilization, but none of the isolated Acinebacter, Bacillus, Brevundimonas, Muricauda, Novosphingobium, Rheinheimera, Sphingobium and Stenotrophomonas did, indicating the utilization of TMAO is a species-speci?c feature. Furthermore, we noticed that the ability of TMAO utilization varied among strains of the same species. TMAO has greater impact on the growth of deep-sea isolates of Vibrio neocaledonicus than shallow-water isolates. Taken together, the results describe for the ?rst time the TMAO utilization in deep-sea bacterial strains, and expand our understanding of the physiological characteristic of marine bacteria.