基于支持向量回归机的南黄海浒苔分布面积预测模型

绿潮作为一种新型的海洋灾害,已经引起了各个国家的重视.依据2012年南黄海海域浒苔遥感监测分布面积数据,选取了温度、天气状况、风向、风力、浪高5种影响浒苔扩散的气候因子,建立了基于SVR的浒苔分布面积预测模型,并与经典的最近邻点插值模型、线性插值模型、3次样条函数插值模型和分段3次Hermite插值模型进行了回归效果的对比.分析结果表明,基于SVR的浒苔分布面积预测模型能够为浒苔遥感数据的插补提供一种方法,且回归效果优于传统的回归方法,为浒苔的防治提供辅助决策信息.     

Experience in quantitative analysis of floristic structure and diversity in Samarskaya Luka

Quantitative methods have been applied to the study of floristic structure and diversity in Samarskaya Luka. The results show that Samarskaya Luka is a floristically heterogeneous area in which six elementary floras can be distinguished. This area has been assessed as a floristically autonomous formation, and its floristic representativeness was estimated.     

Epidemiologic Analysis of Factors Associated with Local Disappearances of Native Ranid Frogs in Arizona

Abstract:  We examined factors that may independently or synergistically contribute to amphibian population declines. We used epidemiologic case–control methodology to sample and analyze a large database developed and maintained by the Arizona Game and Fish Department that describes historical and currently known ranid frog localities in Arizona, U.S.A. Sites with historical documentation of target ranid species ( n = 324) were evaluated to identify locations where frogs had disappeared during the study period (case sites) and locations where frog populations persisted (control sites). Between 1986 and 2003, 117 (36%) of the 324 sites became case sites, of which 105 were used in the analyses. An equal number of control sites were sampled to control for the effects of time. Risk factors, or predictor variables, were defined from environmental data summarized during site surveys and geographic information system data layers. We evaluated risk factors with univariate and multifactorial logistic-regression analyses to derive odds ratios (OR). Odds for local population disappearance were significantly related to 4 factors in the multifactorial model. Disappearance of frog populations increased with increasing elevation (OR = 2.7 for every 500 m, p < 0.01). Sites where disappearances occurred were 4.3 times more likely to have other nearby sites that also experienced disappearances (OR = 4.3, p < 0.01), whereas the odds of disappearance were 6.7 times less (OR = 0.15, p < 0.01) when there was a source population nearby. Sites with disappearances were 2.6 times more likely to have introduced crayfish than were control sites (OR = 2.6, p = 0.04). The identification of factors associated with frog disappearances increases understanding of declines occurring in natural populations and aids in conservation efforts to reestablish and protect native ranids by identifying and prioritizing implicated threats.     

Prenatal prediction of childhood-onset spinal muscular atrophy (SMA) in Turkish families

Childhood-onset spinal muscular atrophy (SMA) is one of the most common neurodegenerative genetic disorders. SMN1 is the SMA-determining gene deleted or mutated in the majority of SMA cases. There is no effective cure or treatment for this disease yet. Thus, the availability of prenatal testing is important. Here we report prenatal prediction for 68 fetuses in 63 Turkish SMA families using direct deletion analysis of the SMN1 gene by restriction digestion. The genotype of the index case was known in 40 families (Group A) but unknown in the remaining 23 families (Group B). A total of ten fetuses were predicted to be affected. Eight of these fetuses were derived from Group A and two of these fetuses were from Group B families. Two fetuses from the same family in Group A had the SMNhyb1 gene in addition to homozygous deletion of the NAIP gene. One fetus from Group A was homozygously deleted for only exon 8 of the SMN2 gene, and further analysis showed the presence of both the SMN1 and SMNhyb1 genes but not the SMN2 gene. In addition, one carrier with a homozygous deletion of only exon 8 of the SMN1 gene was detected to have a SMNhyb2 gene, which was also found in the fetus. To our knowledge, these are the first prenatal cases with SMNhyb genes. Follow-up studies demonstrated that the prenatal predictions and the phenotype of the fetuses correlated well in 33 type I pregnancies demonstrating that a careful molecular analysis of the SMN genes is very useful in predicting the phenotype of the fetus in families at risk for SMA. Copyright © 2002 John Wiley & Sons, Ltd.     

人工神经网络在预测深基坑周边地表沉降变形中的应用研究

深基坑开挖引起的周边地表变形预测是一个复杂非线性问题,引起地表沉降的影响因素很多,各因素之间呈高度的非线性关系。传统的基坑用边地表沉降变形预测方法存在着一定的局限性,其预测精度有待提高,而人工神经网络是一种多元非线性动力学系统,可以灵活方便地对多成因的复杂未知系统进行高度建模,实现全面考虑各种主要影响因素的深基坑周边地表沉降变形预测。本文介绍了误差反向传播(BP)网络模型的结构、学习过程及其算法的改进,径向基函数(RBF)网络模型的结构及其学习过程;分析了影响深基坑开挖周边土体沉降变形的主要影响因素;以25个基坑工程的地表沉降实测资料为训练样本,建立了11个输入影响因素的BP神经网络模型和RBF神经网络模型,通过对样本的学习训练过程及对5个检验样本的预测精度,说明了人工神经网络用于预测基坑周边地表沉降的可行性和准确性。     

企业安全文化评估与企业安全行为的质化研究

通过总结国内外各种不同的安全文化概念,归纳分析了国内外较为流行的近10种安全文化建立模式。在研讨文献的基础上,通过深入访谈、焦点小组与参与观察,归纳构造了一套共11个维度,总题项为152题的安全文化评估量表。通过对选定企业的实证问卷调查,进行统计分析,采用多元逐步回归分析等方法对安全文化各评核因素及整体安全文化进行预测,并以百分数呈现其整体的安全文化现状水平。通过基于扎根理论的各种质化研究方法,对案例企业的安全文化建设和企业安全行为进行研究,得到了7点关于企业安全行为的研究发现,并提出了相应的改进建议。     

焦化废水处理过程中固相物质的形成及处置方法评价

焦化废水处理过程中产生的焦油、污泥和结晶杂盐等固相物质,既有资源属性,又有污染特性,但目前缺乏基于能源、经济及环境影响方面的评估.本研究阐述了3类固相物质的形成机制,建立了质量当量计算及处置方法评价模型.以宝武集团韶关钢铁股份有限公司焦化厂（二期）焦化废水处理工程的A/O/H/O（厌氧/好氧/水解/好氧）流化床工艺作为考察对象,利用工程运行参数和水质统计数据进行固相物质的产量推算,结果发现,焦油、物化污泥、生物污泥（含水率为80%）和工业杂盐的产率分别为0.186、5.80、4.24和1.97 kg·m^-3.通过处置方法评价模型明确了焦油焚烧、污泥热解、结晶杂盐分盐提纯后工业应用是最佳处置方案,在60 m³·h^-1废水处理规模的固相物质处置过程中,每年约产生1177 MWh的能源,获得135.0万元的经济效益,排放627.0 t CO₂,表明能源回收、经济效益和环境影响的协同存在.     

国内外饮用水余氯限值及监测方法的研究进展

新型冠状病毒肺炎疫情期间大规模使用含氯消毒剂,其残留可能对水环境及人体健康造成影响.我国饮用水水源地质量标准并未设置余氯项目及其浓度限值,且缺乏统一的余氯现场快速分析方法标准.为公共卫生事件发生期间的水质余氯监测与评价提供参考,对国内外饮用水标准余氯限值、实验室标准分析方法、现场快速分析方法等进行汇总分析,结果表明:①不同国家和地区以及WHO（World Health Organization,世界卫生组织）在饮用水标准中分别设置了出厂水中余氯限值（范围为0.1~2.0 mg/L）、管网末梢水中余氯限值（范围为0.1~1.8 mg/L）及饮用水中余氯最大允许浓度（范围为4~5 mg/L）.②比色法、容量法因其具有反应迅速且稳定、准确度及精密度较高等优点而成为国内外实验室主要标准或推荐分析方法,高效液相色谱法检出限最低,灵敏度最高,可用于余氯痕量分析.③余氯现场快速分析方法多以比色法为主,在线监测方法多为电化学方法,但缺乏统一的标准方法.研究显示,国外饮用水标准中余氯最大允许浓度为5 mg/L,WHO推荐高风险环境下的管网末梢余氯浓度最低为0.5 mg/L,建议尽快开展水质余氯现场监测方法标准化研究.      

随机振动响应峰值因子的预计理论研究

简要介绍了超越概率理论、超越频次理论、损伤等效理论和功率谱密度(PSD)的时域拟合理论等4种常见的峰值因子预计理论,并基于三角级数提出了一种新的预计理论。结合试飞加速度数据样本,对比分析超越频次理论、PSD时域拟合理论和三角级数理论的预估精度。研究表明,上述4种常见的预计理论本质上属于统计学理论;PSD时域拟合理论预计的峰值因子波动较大,峰值因子与归一化次数满足高斯分布;三角级数理论的预估精度较高,但缺乏离散峰个数的合理判据。     

基于FLUS模型的湖北省生态空间多情景模拟预测

改革开放以来,中国经济在飞速发展的同时,生态环境问题日益严峻。为保障国家和地区的生态安全,对未来生态空间进行模拟预测十分必要。在长江大保护和长江经济带绿色发展背景下,以湖北省为研究区,利用FLUS模型基于湖北省2010年、2015年土地利用数据及包含自然和人文因素的15种驱动因子数据,对2035年的湖北省生态空间进行模拟预测。结果表明：利用2010年土地利用现状模拟出的2015年湖北省土地利用变化情况,总体精度达到0.976,Kappa系数达到0.961,模拟精度较高。设置的生产空间优先、生活空间优先、生态空间优先以及综合空间优化4种不同情景,基本满足未来湖北省不同发展导向的需求。从地貌单元角度来看,在不同情景下,湖北省生态空间主要分布于湖北省边陲四大山区,中部江汉平原生态空间零星分布。从数量规模上来看,不同情景下各个用地类型数量规模差异较为明显,生产空间优先情景下耕地面积增加1216 km²,生活空间优先情景下城镇用地规模增加5959 km²,生态空间优先情景下生态空间用地增长722 km²,综合空间优化情景下生态空间用地规模变化更趋于平缓。从生态空间变化分布来看,四大山区的生态空间变化不大,但中部江汉平原生态空间变化较为明显,其中从行政区划上来看,变化范围主要分布于武汉城市圈、襄阳市、宜昌市中西部地区及随州市中部地区。总而言之,FLUS模型对于湖北省生态空间模拟的适用性较好,多情景模拟结果可为湖北省未来国土空间规划及未来生态空间管控提供多角度、多方向的政策决策参考。