Critical Review of Environmental Forensic Techniques: Part II

A multitude of forensic techniques are available for age dating and source identification, including aerial photography interpretation, corrosion models, the commercial availability of a chemical, chemical associations with discrete types of equipment, chemical profiling, degradation models and contaminant transport models. The success of these techniques in environmental litigation and their applicability to a particular fact situation is rarely discussed in the literature. When these techniques are introduced as scientific evidence, their governing assumptions and the adequacy of the underlying data are rigorously scrutinized and often, successfully challenged. The purpose of this paper is to review selected forensic techniques and discuss their merits so that the user can select the technique or combination of techniques most appropriate for the factual elements of the case.     

Nestmate recognition and the genetic relatedness of nests in the ant Formica pratensis

Genetic relatedness of the mound-building ant Formica pratensis was determined by means of microsatellite DNA polymorphism, and its impact on nestmate recognition was tested in a population in Southern Sweden (Oeland). Recognition between nests was measured by testing aggression levels between single pairs of workers. The genetic distances of nests (Nei's genetic distance) and the spatial distance of nests were correlated and both showed a strong relation to the aggression behavior. Multiple regression analysis revealed a stronger impact of genetic relatedness rather than spatial distances on aggression behavior. Neighbouring nests were more closely related than distant nests, which may reflect budding as a possible spreading mechanism. The genetic distance data showed that nestmate recognition was strongly genetically influenced in F. pratensis. Received: 2 October 1997 / Accepted after revision: 10 January 1998     

乌梁素海富营养化状态的模糊模式识别研究

本文利用模糊模式识别交叉迭代模型对分布于乌梁素海各水功能区的21个水质监测点,2009年5月～10月的富营养化等级进行了模糊模式识别研究;并利用Arcview的空间分析功能,绘制富营养化等级识别结果的空间分布月变化Grid图。将富营养化状态识别结果及其分布区域与乌梁素海同期水质监测指标浓度变化及其分布区域进行对比分析及验证。识别结果与实际水质状况相符,验证了该方法的有效性。     

地下水主要组分水化学异常识别方法对比：以柳江盆地为例

地下水化学组分的异常识别是构建水化学背景值及开展人类活动影响程度识别量化的重要基础.以往提出的基于5种水化学图的主要组分异常识别方法取得了良好效果,但是该方法考虑的水化学图种类过多,计算复杂.为简化水化学图法,本文以柳江盆地为例,尝试采用Durov图替代5种水化学图,对地下水主要组分进行异常识别.为此,分别对比分析了水化学图与数理统计法组合出的7种异常识别方法的剔除效果,并进行方法可靠性检验.结果表明,数理统计法与水化学图法结合识别地下水异常相对于两者单独使用可以取得更好的识别效果;其中3σ准则+5种水化学图法和3σ准则+Durov图法对地下水主要组分异常识别效果最好;并证实采用Durov图可以有效替代5种水化学图进行地下水主要组分的异常识别,既保留了水化学识别异常的科学性,又大大简化了异常识别计算的过程.     

基于机器学习和非参数估计的PM2.5风险评估

为开展区域风险评估,融合手机信令、气象和地理信息等多源数据,引入随机森林机器学习、非参数估计分位数图示法和非监督学习K-mean等方法,构建了区域PM_2.5风险评估及特征识别评价框架,在南京市区以0.3km分辨率网格为基础单元开展了案例研究.结果表明,该技术既可有效模拟PM_2.5浓度时空分布,十折交叉验证R²达到0.76,证明了准确度较高,并基于此识别出4种主要污染特征;也可有效捕捉短期人口流动导致的风险,在污染浓度不变的情况下短期人口流动会导致风险增加0.30~0.97倍.综合PM_2.5浓度和人口分布,识别出4种主要暴露风险模式,其中,研究区域6.5%的面积为高风险地区,23.0%的面积为低风险地区.“十四五”期间应加快现代科学技术在环境保护领域的应用,实施网格化和差异化的风险控制政策,维护人群健康.     

空气质量综合评价的基于熵权的属性识别模型

文章通过将熵值法与属性识别模型相结合,建立了基于熵权的属性识别模型来进行空气质量综合评价,从而避免了权重确定的主观性问题。并结合实际说明了建模过程,得出了更符合实际的评价结果。     

Just sustainabilities and local action: evidence from 400 flagship initiatives

Just sustainabilities has emerged as a powerful discourse to guide local action towards sustainability. As an overarching discourse, it prescribes four policy principles: (1) addressing well-being and quality of life; (2) meeting the needs of present and future generations; (3) enabling justice and equity in terms of recognition, process, procedure, and outcome; and (4) living within ecosystem limits. Following previous calls for engaging public and private actors in just sustainabilities, this paper inquiries about the extent to which these principles can be realistically integrated in local environmental governance. A database of 400 sustainability initiatives in more than 200 cities in all world regions is analysed to examine whether just sustainabilities principles are already enshrined, explicitly or implicitly, in local sustainability initiatives. This analysis suggests that, in this sample, there is a significant deficit in terms of addressing the principles of justice and equity, and ecosystem limits. However, the data also suggest that local action may already be delivering some aspects of just sustainabilities, even if this is not always explicit. The paper concludes with a call for a coordinated effort to translate a just sustainabilities discourse to local actors leading action on the ground.     

钻井过程人员异常行为视频智能识别系统

针对钻井现场视频监控人工值守工作量大、异常行为识别不及时、海量视频数据缺乏实时报警等问题,结合钻井作业操作规程建立井控坐岗监测、起下钻、甩钻具等钻井作业场景典型异常行为判别规则,在基于卷积神经网络的目标检测方法基础上,引入SIFT方法和基于强化学习的困难样本筛选策略,提高了算法识别的准确率与鲁棒性。测试结果表明,在像素分辨率1 920×1 080的情况下,该方法整体识别准确率达到85%以上,平均识别速度为135 ms/帧。在此基础上,开发了钻井过程人员异常行为智能识别系统,试点应用结果表明,该系统能够对接入的视频流进行实时检测,及时、准确识别出坐岗人员擅自离岗、起下钻及甩钻具过程人员危险站位等异常行为,有效提升钻井现场安全管理水平。     

属性识别理论在综合评价城市污泥堆肥腐熟度中的应用

基于属性识别理论,以含水率、挥发固体、粪大肠菌值、发芽指数作为腐熟程度的判别指标体系,并根据信息熵客观计算权重,建立属性识别模型,并用此模型综合评价城市污泥堆肥的腐熟度.实例应用表明,属性识别理论应用于污泥堆肥腐熟度的评价是可行的,而且评价结果比较合理.