基于深度学习的大数据空气污染预报

为了更好地反映环境污染变化趋势,为环境管理决策提供及时、全面的环境质量信息,预防严重污染事件发生,开展城市空气质量预报研究是十分必要的.本文针对环境大数据时代下的城市空气质量预报,提出了一种基于深度学习的新方法.该方法通过模拟人类大脑的神经连接结构,将数据在原空间的特征表示转换到具有语义特征的新特征空间,自动地学习得到层次化的特征表示,从而提高预报性能.得益于这种方式,新方法与传统方法相比,不仅可以利用空气质量监测、气象监测及预报等环境大数据,充分考虑污染物的时空变化、空间分布,得到语义性的污染物变化规律,还可以基于其他空气污染预测方法的结果(如数值预报模式),自动分析其适用范围、优势劣势.因此,新方法通过模拟人脑思考过程实现更充分的大数据集成,一定程度上克服了现有方法的缺陷,应用上更加具有灵活性和可操作性.最后,通过实验证明新方法可以提高空气污染预报性能.     

数据挖掘在预测组织事故防控效果中的应用*

为进一步探索数据挖据技术在组织事故预防工作中的融入性与适用性，基于24Model构建事故预控基础模型,通过预测准确率数值及接受者操作特性曲线（ROC曲线）对比分析随机森林（RF）、支持向量机(SVM)、决策树（DT）与神经网络（NN）4种方法对组织事故防控效果的预测性能。结果表明:针对事故率控制（Y1）、职业危害预防（Y2）、财产损失3类预测目标（Y3），RF方法均能达到较高的准确率及稳定性，具有较优的预测性能。根据特征重要度（FI）排序，明确对组织事故水平影响最显著的因素为安全实践活动认知（SC5）及安全管理程序文件（SMS3），FI值均大于0.150 0。研究结果可为有效预测组织事故防控效果提供方法依据，同时为企业安全工作的规划设计提供思路。     

基于大数据技术构建生态环境信息化体系及其重点应用研究

本文首先分析了生态环境信息化体系的主要构成要素,其后基于大数据技术的生态环境信息化体系重点领域应用进行了深入研究。     

基于第二次全国污染源普查质量保证的思考及体系建设初探

通过对清查数据的结果分析,确定第二次全国污染源入户普查的对象,基于企业填报、现场入户、资料收集审核、数据核查等具体工作的实施与开展,思考和归总普查工作中影响数据质量的现实问题和难点,初步构建污染源普查质量保证体系,以期为全国污染源普查数据质量保证提供有益参考。     

探究环境监测数据在环境执法中的应用

伴随经济发展的迅速,人们对环境越来越重视,国家也制定了相应的环境执法与保护措施,保障环境质量。在环境执法的过程中,必须严格依照环境监测的数据,从而保障信息的准确性,使环境执法策略的开展能够顺应时代发展,符合目前发展的现状。因此,环境执法中环境监测数据占据重要的位置。基于此,本文分析了环境执法中环境监测数据的应用,笔者根据自身经验提出相应的建议。     

基于平衡计分卡的大气环境监测数据共享网站绩效评估研究

基于平衡计分卡原理,结合网站特点,建立了大气环境监测数据共享网站绩效评估模型。以共享网站的战略愿景为模型框架的中心,以财务、顾客、内部程序、学习与成长等4个基本维度为一级指标,构建了绩效评估指标体系,并采用层次分析法确定了各指标的权重。利用该模型对长三角区域空气质量预测预报系统和宁波市空气质量预报预警平台的大气数据共享模块进行实证研究,绩效评估结果均为良好。     

Injury risk functions for frontal oblique collisions

Objective: The objective of this article was the construction of injury risk functions (IRFs) for front row occupants in oblique frontal crashes and a comparison to IRF of nonoblique frontal crashes from the same data set.

Method: Crashes of modern vehicles from GIDAS (German In-Depth Accident Study) were used as the basis for the construction of a logistic injury risk model. Static deformation, measured via displaced voxels on the postcrash vehicles, was used to calculate the energy dissipated in the crash. This measure of accident severity was termed objective equivalent speed (oEES) because it does not depend on the accident reconstruction and thus eliminates reconstruction biases like impact direction and vehicle model year. Imputation from property damage cases was used to describe underrepresented low-severity crashes―a known shortcoming of GIDAS. Binary logistic regression was used to relate the stimuli (oEES) to the binary outcome variable (injured or not injured).

Results: IRFs for the oblique frontal impact and nonoblique frontal impact were computed for the Maximum Abbreviated Injury Scale (MAIS) 2+ and 3+ levels for adults (18–64 years). For a given stimulus, the probability of injury for a belted driver was higher in oblique crashes than in nonoblique frontal crashes. For the 25% injury risk at MAIS 2+ level, the corresponding stimulus for oblique crashes was 40 km/h but it was 64 km/h for nonoblique frontal crashes.

Conclusions: The risk of obtaining MAIS 2+ injuries is significantly higher in oblique crashes than in nonoblique crashes. In the real world, most MAIS 2+ injuries occur in an oEES range from 30 to 60 km/h.     

Connecting the Dots: Public Engagement with Environmental Data

Air pollution and other environmental hazards are often imperceptible and need to be made publicly visible. The paper argues for the importance of visualizations in drawing public attention to imperceptible hazards and in providing the public with access to empirical data describing the risks. It also argues for critical inquiry into hazards’ selective visibility as it is produced by visualizations. The impact of visualizations and their selective visibility are considered through the example of a public art project called Particle Falls installed in 2014 in Pittsburgh, a city with a long history of both ignoring air pollution and working to ameliorate this problem. I examine the impact and selective visibility of Particle Falls by considering the underlying production of data, as well as context and support systems for this visualization, and by comparing it with other visualizations of local air quality.     

Before the crash: A systematic analysis of police narratives to determine the prevalence of pre-crash factors in fatality road transport crashes

Abstract

Objectives: The objectives of this study were to identify the prevalence of pre-crash factors that were present in fatal road transport crashes for the deceased and counterpart road user.

Methods: The study is a retrospective population-based case series study of transport-related deaths reported to coroners in Australia from 2013 to 2016. Data was extracted from the National Coronial Information System.

Results: In total, 6,137 fatality crashes occurred during the study period. Police reports were available for 5,523 crashes (89.9%). The most frequently reported pre-crash factors reported behaviour specifically drivers (e.g., driving without a license or while license was disqualified). Presence of intoxicating substances were also reported in the deceased and counterparts. Analyses of toxicology reports are continuing to determine if rates are comparable to level of use in community.

Conclusions: Coronial report provide detailed information that may be pertinent to understanding and potentially preventing crashes. Also emerging from the data is the extent of pre-crash factors that relate to illegal or deviant behavior and whether these are contextual or contributory factors.     

Applying Statistical Causal Analyses to Agricultural Conservation: A Case Study Examining P Loss Impacts

Estimating the effect of agricultural conservation practices on reducing nutrient loss using observational data can be confounded by factors such as differing crop types and management practices. As we may not have the full knowledge of these confounding factors, conventional statistical meta‐analysis methods can be misleading. We discuss the use of two statistical causal analysis methods for quantifying the effects of water and soil conservation practices in reducing P loss from agricultural fields. With the propensity score method, a subset of data was used to form a treatment group and a control group with similar distributions of confounding factors. With the multilevel modeling method, data were stratified based on important confounding factors, and the conservation practice effect was evaluated for each stratum. Both methods resulted in similar estimates of the conservation practice effect (total P load reduction avg. ~70%). In addition, both methods show evidence of conservation practices reducing the incremental increase in total P export per unit increase in fertilizer application. These results are presented as examples of the types of outcomes provided by statistical causal analyses, not to provide definitive estimates of P loss reduction. The enhanced meta‐analysis methods presented within are applicable for improved assessment of agricultural practices and their effects and can be used for providing realistic parameter values for watershed‐scale modeling.