多机器学习模型下逐小时PM2.5预测及对比分析

为了能及时、准确的估算出PM_2.5浓度及污染等级,分别构建了K最邻近模型（KNN）、BP神经网络模型（BPNN）、支持向量机回归模型（SVR）、高斯过程回归模型（GPR）、XGBoost模型和随机森林模型（RF）6个PM_2.5浓度预测模型,选取江西省赣州市为实验区域,采用2017~2018年逐小时气象站数据、PM_2.5浓度数据和Merra-2再分析数据开展PM_2.5预测实验.结果表明,缺少污染物观测数据时,利用能见度和气象因子等数据也能较好的预测PM_2.5浓度.在PM_2.5浓度预测精度方面,XGBoost模型最高,随机森林模型次之,高斯过程回归模型最差.6个模型的预测精度总体呈现冬季最高,秋季和春季次之,夏季最低.XGBoost模型的PM_2.5污染等级预测准确率高于其他模型,综合准确率达87.6%,并且XGBoost模型具有训练时间短,占用内存小等优点.XGBoost模型的变量重要性结果表明,能见度变量的重要性最高,相对湿度和时间变量次之.本研究可为环境部门准确预测、预报PM_2.5浓度提供参考.     

基于微波辐射计分析低能见度的液态含水量特征

利用广州国家基本气象站2013年11月~2014年4月的能见度（Vis）、相对湿度（RH）、微波辐射计液态含水量（LPR）小时资料和云资料,分析了广州地区能见度的时间变化,以及低能见度时液态含水量变化特征.结果表明广州地区Vis<10km的出现频率为66.37%;Vis ≤ 1km只在RH ≥ 95%情况下出现,1km < Vis ≤ 2km只在RH>90%出现,Vis>5km在RH>95%区间出现频率为0,Vis>15km在RH>90%的出现频率为0;低能见度事件出现的频率高低顺序依次为早 > 晚 > 中,与相应时段低空LPR的平均值高低一致,说明低能见度事件与低空大气中高液态含水量存在一定的正相关;当Vis<10km时,大气液态水含量都处于较高的水平,在3000m以下存在一个大于0.02g/m³的高含水量层,液态含水量出现最大值的高度在1550m左右;当Vis ≥ 10km时,大气液态含水量都处于较低的水平,液态含水量出现最大值的高度大约在2000m,未出现有大于0.02g/m³的情况;广州地面低能见度过程不是低云（或低云接地为雾）过程导致的概率基本在50%以上,Vis ≤ 1km情况下LPR值基本为0（微波辐射计没有监测到液态水）,但由于RH>95%接近饱和且出现高浓度的颗粒物,说明广州地区低能见度（Vis ≤ 1km）应为雾霾混合情景.表明只有联合利用微波辐射计的液态含水量、相对湿度与颗粒物浓度等要素变化特征的综合分析才能为雾、雾霾混合与霾现象分辨提供一定的参考.     

In situ monitoring of atmospheric nitrous acid based on multi-pumping flow system and liquid waveguide capillary cell

In the last four decades, various techniques including spectroscopic, wet chemical and mass spectrometric methods, have been developed and applied for the detection of ambient nitrous acid (HONO). We developed a HONO detection system based on long path photometry which consists of three independent modules i.e., sampling module, fluid propulsion module and detection module. In the propulsion module, solenoid pumps are applied. With solenoid pumps the pulsed flow can be computer controlled both in terms of pump stroke volume and pulse frequency, which enables the attainment of a very stable flow rate. In the detection module, a customized Liquid Waveguide Capillary Cell (LWCC) is used. The customized LWCC pre-sets the optical fiber in-coupling with the liquid wave guide, providing the option of fast startup and easy maintenance of the absorption photometry. In summer 2014, our system was deployed in a comprehensive campaign at a rural site in the North China Plain. More than one month of high quality HONO data spanning from the limit of detection to 5 ppb were collected. Intercomparison of our system with another established system from Forschungszentrum Juelich is presented and discussed. In conclusion, our instrument achieved a detection limit of 10 pptV within 2 min and a measurement uncertainty of 7%, which is well suited for investigation of the HONO budget from urban to rural conditions in China.     

Transparency in global sustainability governance: to what effect?

ABSTRACT

Transparency in environmental governance is no longer an uncontroversial answer to problems of accountability and effectiveness. How to design effective transparency systems and in what policy contexts they are effective remain contested issues. This special section, consisting of this introduction and four research articles, interrogates complex and potentially conflicting links between transparency, accountability, empowerment and effectiveness in environmental governance. Building on existing literature and the four contributions, we discuss persisting diversity in varieties of transparency, the evolving dynamics of commodity chain transparency, and the consequences of emerging novel forms of digitalized transparency. As we show, the contributions to this special section interrogate in novel ways the transformative potential of transparency, through shedding light on the performative effects of transparency in ever more complex environmental governance contexts. These contexts may include, inter alia, the growing ubiquity of traceability in transnational commodity chains, the need for ever more anticipatory (ex-ante) forms of environmental governance, and an ever-broadening quest for digitally monitored environments. In particular, the impacts of the real-time ‘radical’ transparency engendered by use of novel digital technologies remain under-analyzed in the sustainability domain. We conclude by raising several critical concerns that deserve further scientific research and policy debate.     

气质联用法测定空气中挥发性有机物的方法研究

通过美国Entech公司Summa罐采集大气样品,7100型大气预浓缩系统提取和富集样品中的挥发性有机物,省去了繁琐的有机物提取和净化工序,避免和有机溶剂接触,将样品直接输入气质联机进行分析。利用这种方法定性分析了某油罐区的大气成分,结果表明该地区大气中含有成分复杂的环己烷、苯和萘的衍生物,这些化合物的浓度远远大于居民区同类化合物的含量。     

环境空气影响预测评价方法和模式问题探讨

指导环境影响评价的理论和技术方法要求具备较强操作性和实用性,鉴于实践中发现指导环评的相关技术方法存在一定的问题,这些给环评工作者在实际工作中带来了一定的困惑。通过对建设项目环境空气影响预测评价中存在的预测模式和方法两个常见问题进行了剖析,指出日均浓度预测方法和选取条件存在的不确定性及导则推荐的点源有风、小风模式预测结果的不连续性问题,通过提出上述问题为完善相关技术导则和方法修订提供可参考的依据。     

北黄海大气氮氧化物的季节分布

根据2006年夏季2、007年冬春秋四个季节分别对北黄海进行的大面调查,分析研究了四季氮氧化物(NOx)的季节变化和浓度分布。结果显示,大气NOx浓度呈现明显的季节变化,表现为冬季>夏季>秋季>春季,冬季北黄海空气中NOx浓度的平均值分别是春季的4倍、夏季的2倍和秋季的2.1倍。NOx浓度在北黄海的不同区域特征呈现为近岸向远岸逐渐降低的趋势,说明陆源污染物对北黄海NOx分布有较大影响。     

大气环境防护距离和卫生防护距离区别及应用的探讨

针对2009年4月1日起实施的《环境影响评价技术导则大气环境》(HJ2.2-2008)中所提出的"大气环境防护距离"的概念展开详细探讨。结合《制定地方大气污染物排放标准的技术方法》(GB/T13201-91)中提出的"卫生防护距离",从定义、设置前提、计算方法的角度与"大气环境防护距离"进行对比分析,引用建设项目环境影响评价中的案例阐述这两者的主要区别,并对当前执行情况提出建议。     

齐齐哈尔市大气降尘状况及污染分析

以齐齐哈尔市近5年降尘监测点的监测数据资料为依据,分析了城市大气降尘状况及变化趋势,总结出市内不同功能区域降尘污染时空分布特征,提出了降尘污染防治的对策及建议。     

大气环境影响评价新旧导则的对比分析

通过新旧大气环境影响评价导则HJ2.2-2008和HJ/T2.2-93的对比分析,对两个导则中五个方面对应内容的差异逐一进行了解析。以帮助读者了解新旧导则的异同,有利于更加深入地理解和使用新大气环境影响评价导则。