Predicting tree distributions in an East African biodiversity hotspot: model selection, data bias and envelope uncertainty

The Eastern Arc Mountains (EAMs) of Tanzania and Kenya support some of the most ancient tropical rainforest on Earth. The forests are a global priority for biodiversity conservation and provide vital resources to the Tanzanian population. Here, we make a first attempt to predict the spatial distribution of 40 EAM tree species, using generalised additive models, plot data and environmental predictor maps at sub 1 km resolution. The results of three modelling experiments are presented, investigating predictions obtained by (1) two different procedures for the stepwise selection of predictors, (2) down-weighting absence data, and (3) incorporating an autocovariate term to describe fine-scale spatial aggregation. In response to recent concerns regarding the extrapolation of model predictions beyond the restricted environmental range of training data, we also demonstrate a novel graphical tool for quantifying envelope uncertainty in restricted range niche-based models (envelope uncertainty maps). We find that even for species with very few documented occurrences useful estimates of distribution can be achieved. Initiating selection with a null model is found to be useful for explanatory purposes, while beginning with a full predictor set can over-fit the data. We show that a simple multimodel average of these two best-model predictions yields a superior compromise between generality and precision (parsimony). Down-weighting absences shifts the balance of errors in favour of higher sensitivity, reducing the number of serious mistakes (i.e., falsely predicted absences); however, response functions are more complex, exacerbating uncertainty in larger models. Spatial autocovariates help describe fine-scale patterns of occurrence and significantly improve explained deviance, though if important environmental constraints are omitted then model stability and explanatory power can be compromised. We conclude that the best modelling practice is contingent both on the intentions of the analyst (explanation or prediction) and on the quality of distribution data; generalised additive models have potential to provide valuable information for conservation in the EAMs, but methods must be carefully considered, particularly if occurrence data are scarce. Full results and details of all species models are supplied in an online Appendix.     

环境监测数据管理系统的开发

在论述开发环境监测数据管理系统重要性的基础上,根据环境监测站的主要工作职能,确定了环境监测数据管理系统的功能,进行了系统总体设计和程序设计,开发出了操作方便、灵活实用的环境监测数据管理系统。     

飞机铝合金结构腐蚀损伤预测方法对比性研究

依据飞机LD2结构材料的加速腐蚀试验数据,提出了数据拟合、神经网络、时间序列3种腐蚀损伤预测方法,并对3种预测方法的基本原理、预测精度及预测外延性进行了对比分析。研究发现神经网络和时间序列方法预测精度较高;时间序列方法的预测外延性较好,能够以较高精度预测未来一段时期腐蚀损伤的发展趋势及损伤程度。因此,应根据研究需要选用合适的方法进行腐蚀损伤预测。     

Emergency evacuation model assuming leakage of toxic substances in a chemical plant

Preparedness of emergency evacuation for the leakage of toxic substances in chemical plants is very important in order to reduce damage. In order to implement an emergency evacuation properly, it is necessary to comprehensively and concretely determine the conditions of the leakage and atmospheric conditions and predict the consequences of the dispersed gases. Repeated training for emergencies is also essential. In order to realize effective evacuation, a prediction model of the evacuation area that anyone can use to obtain the same results both accurately and promptly is developed in the present study. The prediction model is designed such that the wind speed and atmospheric conditions are automatically set, and the leakage rate is the only input parameter, so that anyone can use the model easily. In addition, the model can also predict the atmospheric parameters for up to 3 h and can calculate the evacuation distance so that smooth evacuation can be achieved for changing atmospheric conditions. Finally, the evacuation area is defined by statistically analysed wind fluctuations, and a series of emergency evacuation measures is implemented.     

Performance of the National Water Model in Iowa Using Independent Observations

This paper explores the performance of the analysis‐and‐assimilation configuration of the National Water Model (NWM) v1.0 in Iowa. The NWM assimilates streamflow observations from the United States Geological Survey (USGS), which increases the performance but also limits the available data for model evaluation. In this study, Iowa Flood Center Bridge Sensors (IFCBS) data provided an independent nonassimilated dataset for evaluation analyses. The authors compared NWM outputs for the period between May 2016 and April 2017, with two datasets: USGS streamflow and velocity observations; Stage and streamflow data from IFCBS. The distribution of Spearman rank correlation (r_s), Nash–Sutcliffe efficiency (E), and Kling–Gupta efficiency (KGE) provided quantification of model performance. We found the performance was linked with the spatial scale of the basins. Analysis at USGS gauges showed the strongest performance in large (>10,000 km²) basins (r_s = 0.9, E = 0.9, KGE = 0.8), with some decrease at small (<1,000 km²) basins (r_s = 0.6, E = ?0.25, KGE = ?0.2). Analysis with independent IFCBS observations was used to report performance at large basins (r_s = 0.6, KGE = 0.1) and small basins (r_s = 0.2, KGE = ?0.4). Data assimilation improves simulations at downstream basins. We found differences in the characterization of the model and observed data flow velocity distributions. The authors recommend checking the connection of USGS gauges and NHDPlus reaches for selected locations where performance is weak.     

基于专利数据分析的污染地下水修复技术发展现状及趋势分析

为明确地下水污染修复技术在世界范围的研究布局和发展态势,本研究基于智慧芽全球专利数据库对1989-2018年地下水污染修复技术专利进行计量学统计,并探讨该技术的研究布局、研究重点及授权趋势。结果显示,从申请量来看,世界范围内地下水污染修复技术涉及的申请量逐渐增长,但绝对申请量和专利授权比相对较低,地下水污染修复技术研发的重视程度依然不够,有待进一步加强;中国是该技术领域主要专利申请国,也是最大的技术来源国;科研单位在该领域的研发实力显著高于企业机构。基于以上研究,认为今后应加强研发,并且创造高效经济的地下水污染修复技术,注重科研机构与企业单位合作及已有专利成果的转化,促进专利的市场应用,使其发挥更大的社会和经济效益。     

三维变分在PM2.5重污染数值模拟中的应用研究

为了探讨三维变分法(3DVAR)对成渝城市群冬季PM2.5重污染模拟的改善效果,采用3DVAR对成渝城市群2017年12月至2018年1月的空气质量数值模拟结果进行资料同化,对比评估嵌套网格空气质量预报模式(NAQPMS)原始数据与同化再分析数据的准确率,并分析成渝重污染特征。研究结果显示,3DVAR在PM2.5、PM10和NO2的同化实验中均取得较好的改善效果,成渝地区检验站点各污染物相关系数(r)的平均提升比例依次为44%、90%和332%,r改善的站点占检验站点总数的比例分别为98%、100%和82%;检验站点均方根误差(RMSE)的平均下降比例分别为15%、37%和31%,RMSE改善的站点占检验站点总数的比例为65%、98%和84%。与原始模拟结果相比,同化结果能够更准确地反映成渝地区冬季重污染期间的PM2.5和PM10空间分布特征。     

Development of injury risk models to guide CT evaluation in the emergency department after motor vehicle collisions

Abstract

Objective: The clinical evaluation of motor vehicle collision (MVC) victims is challenging and commonly relies on computed tomography (CT) to detect internal injuries. CT scans are financially expensive and each scan exposes the patient to additional ionizing radiation with an associated, albeit low, risk of cancer. Injury risk prediction based on regression modeling has been to be shown to be successful in estimating Injury Severity Scores (ISSs). The objective of this study was to (1) create risk models for internal injuries of occupants involved in MVCs based on CT body regions (head, neck, chest, abdomen/pelvis, cervical spine, thoracic spine, and lumbar spine) and (2) evaluate the performance of these risk prediction models to predict internal injury.

Methods: All Abbreviated Injury Scale (AIS) 2008 injury codes were classified based on which CT body region would be necessary to scan in order to make the diagnosis. Cases were identified from the NASS-CDS. The NASS-CDS data set was queried for cases of adult occupants who sought medical care and for which key crash characteristics were all present. Forward stepwise logistic regression was performed on data from 2010–2014 to create models predicting risk of internal injury for each CT body region. Injury risk for each region was grouped into 5 levels: very low (<2%), low (2–5%), medium (5–10%), high (10–20%), and very high (20%). The models were then tested using weighted data from 2015 in order to determine whether injury rates fell within the predicted risk level.

Results: The inclusion and exclusion criteria identified 5,477 cases in the NASS-CDS database. Cases from 2010–2014 were used for risk modeling (n?=?4,826). Seven internal injury risk models were created based on the CT body regions using data from 2010–2014. These models were tested against data from 2015 (n?=?651). In all CT body regions, the majority of occupants fell in the very low or low predicted injury rate groups, except for the head. On average, 57% of patients were classified as very low risk and 15% as low risk for each body region. In most cases the actual rate of injury was within the predicted injury risk range. The 95% confidence interval overlapped with predicting injury risk range in all cases.

Conclusion: This study successfully demonstrated the ability for internal injury risk models to accurately identify occupants at low risk for internal injury in individual body regions. This represents a step towards incorporating telemetry data into a clinical tool to guide physicians in the use of CT for the evaluation of MVC victims.     

Fatal crash between a car operating with automated control systems and a tractor-semitrailer truck

Abstract

Objective: The objective of this investigation was to evaluate the interaction between an SAE level 2 automated vehicle and the driver, including the limitations imposed by the vehicle on the driver.

Methods: A case study of the first fatal crash involving a vehicle operating with an automated control system was performed using scene evidence, vehicle damage, and recorded data available from the vehicle, and information from both drivers, including experience, phone records, computer systems, and medical information, was reviewed.

Results: System performance data downloaded from the car indicated that the driver was operating it using the Traffic-Aware Cruise Control and Autosteer lane-keeping systems, which are automated vehicle control systems within Tesla’s Autopilot suite. As the car crested the hill, a tractor trailer began its left turn onto a crossing roadway. Although reconstruction of the crash determined that there was sufficient sight distance for both drivers to see each other and take action, neither responded to the circumstances leading to the collision. Further, based on the speeds of the vehicles and simulations of the truck’s path, the car driver had at least 10.4?s to detect the truck and take evasive action. Neither the car driver nor the Autopilot system changed the vehicle’s velocity.

?At the time of the crash, the system performance data indicated that the last driver interaction with the system was 1?min 51?s prior when the cruise control speed was set to 74?mph. The driver was operating the vehicle using the Autopilot system for 37 of the 41?min in the last trip. During this period, the vehicle detected the driver’s hands on the steering wheel for a total of 25?s; each time his hands were detected on the wheel was preceded by a visual alert or auditory warning.

Conclusions: The National Transportation Safety Board (NTSB) determined that the probable cause of the Williston, Florida, crash was the truck driver’s failure to yield the right of way to the car, combined with the car driver’s inattention due to overreliance on vehicle automation, which resulted in the car driver’s lack of reaction to the presence of the truck. Contributing to the car driver’s overreliance on the vehicle automation was the car’s operational design, which permitted the driver’s prolonged disengagement from the driving task and his use of the automation in ways inconsistent with guidance and warnings from the manufacturer.     

Estimating Root Zone Soil Moisture at Continental Scale Using Neural Networks

This study investigates the feasibility of artificial neural networks (ANNs) to retrieve root zone soil moisture (RZSM) at the depths of 20 cm (SM20) and 50 cm (SM50) at a continental scale, using surface information. To train the ANNs to capture interactions between land surface and various climatic patterns, data of 557 stations over the continental United States were collected. A sensitivity analysis revealed that the ANNs were able to identify input variables that directly affect the water and energy balance in root zone. The data important for RZSM retrieval in a large area included soil texture, surface soil moisture, and the cumulative values of air temperature, surface soil temperature, rainfall, and snowfall. The results showed that the ANNs had high skill in retrieving SM20 with a correlation coefficient above 0.7 in most cases, but were less effective at estimating SM50. The comparison of the ANNs showed that using soil texture data improved the model performance, especially for the estimation of SM50. It was demonstrated that the ANNs had high flexibility for applications in different climatic regions. The method was used to generate RZSM in North America using Soil Moisture and Ocean Salinity (SMOS) soil moisture data, and achieved a spatial soil moisture pattern comparable to that of Global Land Data Assimilation System Noah model with comparable performance to the SMOS surface soil moisture retrievals. The models can be efficient alternatives to assimilate remote sensing soil moisture data for shallow RZSM retrieval.