优化K means算法在中国近海气候区划中的应用

利用我国近海多年的风、浪及温湿资料,采用DB指标和最大距离法相结合的优化K-means聚类算法,对冬季我国近海进行气候区划,避免了传统K-means算法中确定聚类数目和初始聚类中心的主观性。结果表明,冬季我国近海可被划分为3个区域:1区主要包括28°N以北的海域以及我国东南沿海的一条狭长区域,2区主要包括台湾周边海域、台湾海峡、巴士海峡以及自越南东南部起由西南部向东北部的南海大部分海域,3区则主要包括台湾岛西南部海域、北部湾、广东沿海以及南海东南部。根据给出的区划指标,分析各分区的气候特征及其对海上舰船航行和出海人员的影响,判定1区为基本适宜区,2区为不适宜区,3区为适宜区。     

顾及目标异质性的极化SAR图像非局部均值滤波

相干斑抑制是极化合成孔径雷达(Pol SAR)图像分析的重要预处理步骤。为了更好地抑制极化SAR图像中的相干斑,本文综合目标的异质性和结构信息,提出基于目标异质性的非局部均值滤波方法。首先利用K分布距离度量目标的异质性,并以异质性为基础,保留图像中的点、线等高异质性目标;然后计算图像块之间的异质性差异,最后将其作为度量非局部均值加权滤波像元相似性的权重系数,实现对Pol SAR图像的相干斑抑制。实验对比结果表明:本文方法能够有效地抑制相干斑,同时对细节信息和极化信息也具有良好的保持性,能够为后续的图像应用提供支持。     

论1530年涠洲岛东南海域地震及其强度

本文论述了1530年10月对广西合浦与广东海康产生震感的地震是一次海上地震,通过特定作图法得到其震中位于涠洲岛东南13km的海中(21°00′N,109°15′E),推测其震中烈度为Ⅶ度强,震级为5 3/4级。根据历史资料的记述,1530年10月地震是在同一地区相隔数小时内连续发生的两次震级相近似的地震。此外,本文对涠洲岛周围海域区之地质构造作了论述,指出这里发生中强及强震的地质构造条件是充分的,今后对该区的地震危险性应给予重视。     

The economics of leadership in climate change mitigation

Which kind of reaction can a nation or group of nations expect when leading by example in climate policy? This synthesis article describes possible positive reaction mechanisms from different fields of economics, some of which have scarcely been linked to climate economics previously. One effect may be behavioural, a reaction motivated by fairness, reciprocity, or norms. Second, other nations may interpret the leader's action as a signal on his preference or the value of the objective and adjust their own policy based on the new information. Third, the leader may provide a service to other nations, which decreases their costs and risks. The followers could benefit by learning successful policies, adopting technologies, and obtaining information on the cost of environmental policy. All of these mechanisms have in common that the leader sets an example with the intention of motivating others to contribute to the public good.

Policy relevance

A large body of both theoretical and empirical evidence shows that leading by example in climate change mitigation by a small group of nations has important potential for motivating other nations to follow. Modern economics has identified a range of mechanisms to explain why simple free-riding is unlikely to dominate the reaction to leadership. One such mechanism is described by behavioural economics. Humans often behave as conditional cooperators, meaning that they are willing to do their bit once a leader has done his. A second mechanism is the transmission of a credible signal that the leader considers climate change mitigation to be important. Finally, the leader gains knowledge, which spills over to other countries and thus moves their cost–benefit ratio in favour of mitigation. This evidence implies that leadership provides a promising alternative to stimulate the global cooperation that will eventually be needed to stabilize the climate.     

Source of and potential bio-indicator for the heavy metal pollution in Ny-(A)lesund, Arctic

Three kinds of tundra plant samples including Dicranum angnstum(a type of boreal bryophyte) , PuccineUia phryganodes (a type of fringy p/ant),Salix polaris (a type of vascular plant) and surface soil were samples in 200 at Ny-Alesund of the Arctic.The levels of eight heavy metal elements (Hg, Pb, Cd, Cu, Zn, Ni, Fe and Mn) and three metal-like dements (As, Se, Sr) in the plant and soil samples of the areas within previous coal mining activities are significantly higher than those of other areas.The relative accumulation of these elements in these tundra plant samples is consistent with the one in the soft samples, especially in the areas affected by previous coal-mining activities.Thus, the pollution is apparently from local coal mining activity.Dicranum angustum has the highest concentrations among those elements, and it can be a good bio-indicator for heavy metal pollution in Ny(A)lesund.Though Ny(A)lesund is less polluted by heavy metal than nearby Northern European human living areas, but much more than the tundras of the Alaska, Greenland and the Antarctic.     

基于非局部总变差的消除不规则采样遥感图像复原方法

提出一种消除不规则采样的遥感图像复原方法,该方法结合了ACT算法和总变差（TV）图像复原模型以达到同时去除多种图像退化因素的目的,并且结合非局部均值（NLM）算子给出了基于NLTV的消除不规则采样遥感图像复原模型,最后使用算子分裂与扩展的坎贝尔投影算法求解模型。实验结果表明,该方法能够有效减少复原图像的阶梯效应并提高复原图像的纹理细节信息。     

成分数据理论和无监督聚类K- means方法提取背景和异常信息——以安徽省兆吉口铅锌矿床为例

地球化学数据是应用地球化学研究的重要组成部分，是化学勘查工作的基础成果。勘查地球化学数据基本上以元素的质量百分浓度 (简称浓度)的形式表达，是典型的成分数据。其表达的是“组分/总体”相对质量贡献信息，而不是绝对的质量变化信息。浓度数据分布在单纯形空间，而不是整个欧式空间。对成分数据进行处理之前，进行适当的对数比值转换处理可以提高其信息表达。本文以安徽省兆吉口铅锌矿床土壤中Pb数据为示范案例，通过对数比值转换方法优化浓度数据的结构以提高相对信息的表达，并利用无监督学习K- means聚类方法根据对数比值转换数据分布空间质心的距离识别背景和异常信息，最后对K- means聚类方法识别的背景和异常与迭代2倍标准差法和浓度- 面积分形分析法进行比较以衡量其表现。结果表明：浓度数据表达的是相对质量信息，而不是绝对质量关系，不同样品间不能通过比较浓度高低推断出质量的多寡关系。对数比值方法可以有效地提高浓度数据的结构和信息表达，K- means方法能够准确识别对数比值转换数据的背景和异常信息，其效果类似浓度- 面积分形分析方法，比迭代2倍标准差法好。     

流体包裹体的合成方法及其对矿床学研究的意义

流体包囊体的合成方法是一种采集高温高压条件下流体代表性样品的技术。它已经用于在较宽的压力-温度-成分范围内确定流体的体积平衡、相平衡性质和温度校正。详细介绍了国内外(包括作者)几种合成流体包囊体的实验方法,给出了实验结果,并简要叙述了它在矿床学研究方面的一些意义。     

Hopkinson压杆对准脆性材料的动态力学实验研究

分离式Hopkinson压杆是对准脆性材料实施高速加载的理想工具,利用该装置可以对准脆性材料实施各种加载方式,测定材料的动力特性参数。对分离式Hopkinson压杆装置的测试原理做了扼要说明,简单介绍了目前用于准脆性材料的测试方法(压缩实验方法和拉伸实验方法),并结合目前的测试技术,对准脆性材料的实验方法提出了一些新的设想。     

Statistical Evaluation of Compositional Differences Between Upper Eocene Impact Ejecta Layers

Melt droplets formed by the impact of a large meteorite impact event(s) have been found in upper Eocene marine sediments from the Atlantic, Pacific, and Indian oceans. The number of discreet impacts that occurred to form these melt droplets can be gleaned by compositionally analysing the droplets and establishing the number of distinct groups. Previous studies have inferred two, three, or more source impact events, although we believe the statistical techniques used to distinguish the groupings are open to criticism. Multivariate and univariate analysis (including discriminant analysis) of the major-element composition of an increased data set of impact melt ejecta droplets have been performed. The results demonstrate that the uppermost ejecta layer is geochemically distinct from other late Eocene melt ejecta. Our statistical analysis suggests two underlying, purportedly stratigraphically separate ejecta layers, possess minimal differences that cannot be distinguished clearly from one another by discriminant analysis, which adds to the plausibility that they have a common source. Finally, our results reveal apparent disparities between the new major-element data from this study and data compiled from existing sources.