基于决策树技术及在线监测的水质预测

利用北方某城市水源的水质在线监测系统,建立了基于决策树技术,具有较强可视性和实际应用,以及能预测次日源水中叶绿素水平的决策树模型.该模型将某城市水源在线监测的溶解氧和太阳辐射照度数据转换计算为每日平均标准偏差及均值,并与每日定时取样测定的叶绿素含量一起作为预测因子,通过将115组数据的前100组数据作为训练集建立预测次日叶绿素水平决策树模型,并采用后15组数据进行模型的仿真预测检验,结果只有3 d的预测出错,预测准确率达80%.并讨论了模型建立对数据的要求及解读预测规则等问题.     

2016年秋季新乡市空气质量模式预报效果评估

基于新乡市空气质量数值预报平台,采用相关系数（r）、标准化平均偏差（NMB）等统计指标,系统评估2016年秋季新乡市嵌套网格空气质量预报模式（NAQPMS）和通用多尺度空气质量模式（CMAQ）的预报效果,对比分析2套模式不同预报时效和不同水平分辨率的空气质量等级预报准确率。结果显示：2套模式均较好地表征了各主要污染物的浓度变化特征,2套模式的等级预报准确率高于60%,其中CMAQ对中度及重度的预报等级准确率达到70%。对比模式24、48、72 h 3种预报时效效果,24 h预报时效的统计数据最优,说明24 h预报时效模拟结果可作为业务预报重要的支撑。     

Evaluation of the temporal scaling variability in forecasting ground-level ozone concentrations obtained from multiple linear regressions

Ozone is a highly unpredictable pollutant which severely affects living conditions in urban and surrounding areas in the Mediterranean basin. This secondary pollutant periodically reaches extremely high concentrations, damaging human health. Multiple linear regression has been widely used in previous works due to the fact that it is a simple and versatile method for forecasting ozone concentrations. However, these models usually prove their validity using fulfillment of statistical constraints, ignoring other intrinsic characteristics existing in the time series, such as the temporal scaling behavior and the data distribution over different time scales. In previous works, it has been demonstrated that observed ozone time series are of a multifractal nature, meaning that the data distribution can be described by using the multifractal spectrum. This work focuses on the capacity of a forecasting model to reproduce the scaling features existing in an observed time series when several chemical and meteorological explanatory variables are introduced following the stepwise procedure. A comparison between the observed spectrum and the simulated ones for each step is used to check which explanatory variables better reproduce the multifractal nature in real ozone time series. It has been confirmed that a model with few explanatory variables allows reproducing the multifractal nature in the simulated time series with an acceptable accuracy without compromising the values of the coefficient of determination and root-mean-squared error, which were used as performance indicators.     

最优化集成方法在城市臭氧数值预报中的应用研究

基于多模式(NAQPMS、CMAQ、CAMx、WRF-Chem)空气质量数值预报业务系统的滚动预报结果,结合站点观测资料,评估了最优化集成方法在城市臭氧数值预报中的可行性和预报效果。一年的评估结果表明:当训练期为15 d时,最优化集成方法能够得到相对较好的结果。总体而言,最优化集成方法对城市臭氧浓度变化趋势和浓度水平的预报效果明显优于单个模式,且在大部分城市优于多模式的最优预报,其预报值和观测的相关系数提高0.11以上,均方根误差降低约10μg/m~3;该方法对城市臭氧污染等级的预报能力也明显优于单个模式,特别是轻、中度污染。此外,在模拟偏差较大的城市,最优化集成方法对预报效果的改进更为显著;在模拟偏差较小的城市,该方法仍可进一步提升预报效果。     

Soil water content forecasting by ANN and SVM hybrid architecture

Soil water content prediction is essential to the development of advanced agriculture information systems. Because soil water content series are inherently noise and non-stationary, it is difficult to get an accurate forecasting result. Considering the problems, in this paper, a novel hybrid learning architecture is proposed according to divide-and-conquer principle, the forecasting accuracy is improved. This novel hierarchical architecture is composed of ANN (Kohonen neural network) and SVM (support vector machine). The Kohonen network is used as a classifier, which partitions the whole input space into several distinct feature regions. Then, the best SVM predictor combined with an appropriate kernel function can be achieved for correspondence regions. The experimental results based on the hybrid model exhibit good agreement with actual soil water content measurements and outperform ANN and SVM single-stage models.     

基于OPAQ的城市空气质量预报系统研究

空气质量预测在国内的关注度日益提高,传统的空气质量预测系统通常运用数值化学传输模型,利用物理方程来计算污染物的扩散、沉降和化学反应。而化学传输模型的预测准确性很大程度上需要依赖详细的污染源排放信息和气象模型的输出结果。基于统计模型的OPAQ空气质量预报业务系统,采用人工神经网络算法,可预测各污染物的日均值或日最大值。并对北京空气质量预报的结果进行了评价,OPAQ空气质量预报业务系统对空气质量预测的准确性较高,能够利用较低的计算资源得到较为准确的预测结果。与数值预报相比,OPAQ空气质量预报业务系统不需要大量的基础数据作为输入,可弥补数值预报的不足,并成为数值预报的有力补充。     

Performance assessment of geospatial simulation models of land-use change—a landscape metric-based approach

Performance evaluation is a critical step when developing land-use and cover change (LUCC) models. The present study proposes a spatially explicit model performance evaluation method, adopting a landscape metric-based approach. To quantify GEOMOD model performance, a set of composition- and configuration-based landscape metrics including number of patches, edge density, mean Euclidean nearest neighbor distance, largest patch index, class area, landscape shape index, and splitting index were employed. The model takes advantage of three decision rules including neighborhood effect, persistence of change direction, and urbanization suitability values. According to the results, while class area, largest patch index, and splitting indices demonstrated insignificant differences between spatial pattern of ground truth and simulated layers, there was a considerable inconsistency between simulation results and real dataset in terms of the remaining metrics. Specifically, simulation outputs were simplistic and the model tended to underestimate number of developed patches by producing a more compact landscape. Landscape-metric-based performance evaluation produces more detailed information (compared to conventional indices such as the Kappa index and overall accuracy) on the model’s behavior in replicating spatial heterogeneity features of a landscape such as frequency, fragmentation, isolation, and density. Finally, as the main characteristic of the proposed method, landscape metrics employ the maximum potential of observed and simulated layers for a performance evaluation procedure, provide a basis for more robust interpretation of a calibration process, and also deepen modeler insight into the main strengths and pitfalls of a specific land-use change model when simulating a spatiotemporal phenomenon.     

Modified grey model for estimating traffic tunnel air quality

This study compared three forecasting models based on the mean absolute percentage errors (MAPE) of their accuracy in forecasting air pollution in a traffic tunnel: the Grey model (GM), the combination model used four sample point and five sample point prediction with GM (1,1)(GM(1,1)^4 + 5), and the modified grey model (MGM). An MGM was combined using the four points of the original sequence using the original grey prediction GM (1,1) for short-term forecasting. The proposed method cannot only enhance the prediction accuracy of the original grey model, but can also solve the jump data forecasting problem something for which the original grey model is inappropriate. The MAPE was applied to the models, and the MGM found the proposed method to be simple and efficient. The MAPE of MGM, calculated over 3 h of forecasts, were as follows: 10.12 (Upwind), 10.07 (Middle) and 7.68 (Downwind) for CO; 10.79 (Upwind), 6.05 (Middle) and 5.98 (Downwind) for NO _x , and 11.67 (Upwind), 7.32 (Middle) and 4.56 (Downwind) for NMHC. The MGM model results reveal that the combined forecasts can significantly decrease the overall forecasting error. Results of this demonstrate that MGM can accurately forecast air pollution in the Kaohsiung Chung–Cheng Tunnel.     

Climate dynamics and extreme precipitation and flood events in Central Europe

Past changes and possible future variations in the nature of extreme precipitation and flood events in Central Europe and the Alpine region are examined from a physical standpoint. An overview is given of the following key contributory physical processes: (1) the variability of the large-scale atmospheric flow and the associated changes of the North-Atlantic storm track; (2) the feedback process between climate warming and the water cycle, and in particular the potential for more frequent heavy precipitation events; and (3) the catchment-scale hydrological processes associated with variations in major river flooding events and that are related to land-use changes, river training measures, and shifts in the proportion of rain to snowfall. In this context an account is provided of the possible future forecasting and warning methodologies based upon high-resolution weather prediction and runoff models. Also consideration is given to the detectability of past (future) changes in observed (modeled) extreme events. It is shown that their rarity and natural fluctuation largely impedes a detection of systematic variations. These effects restrict trend analysis of such events to return periods of below a few months. An illustration using daily precipitation from the Swiss Alps does yield evidence for pronounced trends of intense precipitation events (return period 30 days), while trends of stronger event classes are not detectable (but nevertheless can not be excluded). The small detection probability for extreme events limits possible mitigation of future damage costs through an abatement of climate change alone, and points to the desirability of developing improved early forecasting/warning systems as an additional no-regret strategy. This revised version was published online in July 2006 with corrections to the Cover Date.     

Frequency Analysis for Precipitation Events and Dry Durations of Virginia

Stormwater Control Measures (SCMs) are widely used to control and treat stormwater runoff pollution. The first step in SCM design is to evaluate the precipitation patterns at a site. SCMs are normally designed using a storm with a specific return period. A robust design process that uses frequency distribution of precipitation and monitoring of performance could improve our understanding of the behavior and limitations of a particular design. This is not the current norm in the design of SCMs. In this research, frequency analyses (FA) of precipitation events was conducted using hourly precipitation data from 1948 to 2010 for eight sites representing the four major physiographic regions of Virginia. The available data were treated using an inverse distance method to eliminate missing gaps before processing into events determined by minimum inter-event times. FA was at each site to develop frequency plots of precipitation and dry duration. FA of the eight locations indicates a range of rainfall depths from 22.9 mm in Bristol to 35.6 mm in Montebello, compared to the nominal “water quality storm” of 25.4 mm (i.e., 1 in.). Similarly, for dry duration, for a 10 % exceedance probability, the range is from 16.8 days in Richmond and Norfolk to 19.5 days in Montebello. Dry duration provides guidance for vegetation selection, which is important for some SCMs. The degree of variability in both parameters argues for consideration of site-specific information in design. FA was also used to provide guidance to improve monitoring programs. Monte Carlo simulations demonstrated that performance monitoring programs applied in different regions would likely encounter more than 30 % of precipitation events less than 6.35 mm, and 10 % over 25.4 mm under various sampling regimes. The percentages of precipitation events encountered in the Coastal Plain and Piedmont regions are not impacted by sampling regimes, however the Blue Ridge Mountains and Valley and Ridge regions are likely impacted. Anticipating event occurrences improves the chances of implementing a successful monitoring program. The use of these results could enhance the performance of SCMs with consideration of local conditions for both monitoring SCMs and their design basis.     

Grey Box and Component Models to Forecast Ozone Episodes: A Comparison Study

For the purpose of short-term forecasting of high ozone concentration episodes stochastic models have been suggested and developed in the literature. The present paper compares the quality of forecasts produced by a grey box and a component time-series model. The summer ozone patterns for three European urban areas (two continental and one mediterranean) are processed. By means of forecast performance indices according to EC and WHO guidelines, the following features of the models could be found: The grey box model is highly adaptive and produces forecasts with low error variance that increases with the time horizon of forecast. The component model is more 'stiff' that results in a higher forecast-error variance and poorer adaption in detail. The forecast horizon, however, could be enlarged with this model. The accuracy of predicting threshold exceedance is similar for both models. This can be understood from the assumption of a cyclical time development of ozone that was made for both models.     

Using improved neural network model to analyze RSP,NOx and NO2 levels in urban air in Mong Kok,Hong Kong

As the health impact of air pollutants existing in ambient addresses much attention in recent years, forecasting of airpollutant parameters becomes an important and popular topic inenvironmental science. Airborne pollution is a serious, and willbe a major problem in Hong Kong within the next few years. InHong Kong, Respirable Suspended Particulate (RSP) and NitrogenOxides NO_x and NO₂ are major air pollutants due to thedominant diesel fuel usage by public transportation and heavyvehicles. Hence, the investigation and prediction of the influence and the tendency of these pollutants are ofsignificance to public and the city image. The multi-layerperceptron (MLP) neural network is regarded as a reliable andcost-effective method to achieve such tasks. The works presentedhere involve developing an improved neural network model, whichcombines the principal component analysis (PCA) technique and theradial basis function (RBF) network, and forecasting thepollutant levels and tendencies based in the recorded data. Inthe study, the PCA is firstly used to reduce and orthogonalizethe original input variables (data), these treated variables arethen used as new input vectors in RBF neural network modelestablished for forecasting the pollutant tendencies. Comparingwith the general neural network models, the proposed modelpossesses simpler network architecture, faster training speed,and more satisfactory predicting performance. This improvedmodel is evaluated by using hourly time series of RSP, NO_x and NO₂ concentrations collected at Mong Kok Roadside Gaseous Monitory Station in Hong Kong during the year 2000. By comparing the predicted RSP, NO_x and NO₂ concentrationswith the actual data of these pollutants recorded at the monitorystation, the effectiveness of the proposed model has been proven.Therefore, in authors' opinion, the model presented in the paper is a potential tool in forecasting air quality parameters and hasadvantages over the traditional neural network methods.     

On the Use of MRF/AVN Global Information to Improve the Operational Air Quality Model OPANA

Operational air quality models have become an important tool to assist the decision makers in European Environmental Offices at different levels: cities, regional and state. Because of the important advance on computing capabilities during the last few years the possibility of incorporating the complex research and academic mesoscale air quality models under routine operational basis has become a reality. OPANA model is the operational version of the research model ANA (Atmospheric mesoscale Numerical pollution model for regional and urban Areas). This model is a limited area model (mesoscale beta) and the capability to extend the prediction horizon is limited unless proper boundary conditions are provided during long simulations. In this contribution we show how AVN/MRF (NOAA) vertical numerical meteorological soundings are incorporated to the OPANA system by using JAVA technology. This new feature helps to keep the air quality model into medium power workstations and the performance is improved accordingly. This technology avoids running mesoscale models over larger areas (continental scale) to accordingly increase the forecasting temporal horizon.     

广西重点城市空气质量预报系统

介绍了应用数值预报和统计预报相结合的集成预报技术建立的广西重点城市空气质量预报系统。指出该系统可根据影响大气污染物的天气系统类型修正预报结果，提高预报的准确率；建立的气象和环保部门的预报会商平台，使气象与环保部门实现了每天异地技术会商、信息资源共享等功能；系统的运行环境要求较低，只需终端用户具有拨号或专线上网或局域网联网功能；系统能用于多城市空气质量预报，对城市环境天气服务中的中小尺度短期(24h-48h)的天气预报也有参考价值。     

2019年江苏省PM2.5和O3多模式集合预报算法效果评估

基于江苏省重污染天气监测预报预警系统多模式预报结果,分析了不同数值模式对江苏省13个城市细颗粒物(PM_2.5)和臭氧(O₃)的预报偏差特征,发展了多模式集合预报算法,并对其进行了评估。结果表明,相较于单一数值模式,集合预报算法显著改善了PM_2.5和O₃预报的准确率,其对江苏省PM_2.5和O₃空气质量分指数等级的预报准确率超过了80%。就江苏省整体而言,PM_2.5集合预报的准确率相比最优单一数值模式提升了6%。O₃浓度较低时,集合预报能有效改善各模式存在的高估现象。但受限于目前的校正策略,出现高浓度O₃污染时,集合预报对预报效果的提升相对有限。     

Development of a model to assess environmental performance, concerning HSE-MS principles

The main objective of the present study was to develop a valid and appropriate model to evaluate companies’ efficiency and environmental performance, concerning health, safety, and environmental management system principles. The proposed model overcomes the shortcomings of the previous models developed in this area. This model has been designed on the basis of a mathematical method known as Data Envelopment Analysis (DEA). In order to differentiate high-performing companies from weak ones, one of DEA nonradial models named as enhanced Russell graph efficiency measure has been applied. Since some of the environmental performance indicators cannot be controlled by companies’ managers, it was necessary to develop the model in a way that it could be applied when discretionary and/or nondiscretionary factors were involved. The model, then, has been modified on a real case that comprised 12 oil and gas general contractors. The results showed the relative efficiency, inefficiency sources, and the rank of contractors.     

美国加州南岸地区空气质量预报系统运行管理与借鉴

借鉴美国加州南岸空气质量预报经验,特别是其运行管理模式对于现阶段中国城市环境空气质量预测预报具有极高的参考价值。简要介绍了加州南岸空气质量管理局(SCAQMD)空气质量预报业务流程、技术方法、信息发布方式和预报效果评估策略。建议从4个方面借鉴SCAQMD经验:(1)建立相对完善的空气质量预报体系,分层级自上而下指导预报工作;(2)选择合适的空气质量预报方法,综合统筹成本-效益评估工作;(3)评估预报效果进而提高预报准确率;(4)在优化空气质量预报能力的基础上,有序开展预警协同控制工作。     

Prediction of near-surface soil moisture at large scale by digital terrain modeling and neural networks

The capability of Artificial Neural Network models to forecast near-surface soil moisture at fine spatial scale resolution has been tested for a 99.5 ha watershed located in SW Spain using several easy to achieve digital models of topographic and land cover variables as inputs and a series of soil moisture measurements as training data set. The study methods were designed in order to determining the potentials of the neural network model as a tool to gain insight into soil moisture distribution factors and also in order to optimize the data sampling scheme finding the optimum size of the training data set. Results suggest the efficiency of the methods in forecasting soil moisture, as a tool to assess the optimum number of field samples, and the importance of the variables selected in explaining the final map obtained.     

On the evaluation of box model results: the case of BOXURB model

In the present paper, the BOXURB model results, as they occurred in the Greater Area of Athens after model application on an hourly basis for the 10-year period 1995-2004, are evaluated both in time and space in the light of observed pollutant concentrations time series from 17 monitoring stations. The evaluation is performed at a total, monthly, daily and hourly scale. The analysis also includes evaluation of the model performance with regard to the meteorological parameters. Finally, the model is evaluated as an air quality forecasting and urban planning tool. Given the simplicity of the model and the complexity of the area topography, the model results are found to be in good agreement with the measured pollutant concentrations, especially in the heavy traffic stations. Therefore, the model can be used for regulatory purposes by authorities for time-efficient, simple and reliable estimation of air pollution levels within city boundaries.     

Investigation of impact of environmental changes on precipitation pattern of Pakistan

In this paper, variability in precipitation pattern of Pakistan due to environmental and climatic changes has been studied. Maps have been generated to depict global precipitation variation. Precipitation data of 25 stations of Pakistan have been used. These data were taken from Meteorological Department, Islamabad, Pakistan. The results of two global climate models, namely Australia’s Commonwealth Scientific and Industrial Research’s third generation general circulation model and National Center for Atmospheric Research’s first generation precipitation circulation model for A2 scenario have been applied to investigate the changes. It is observed that precipitation pattern will change significantly in the future. The occurrence of precipitation in all seasons for Pakistan is expected to increase with almost uniform distribution across a season. Average annual precipitation of the country will undergo an increase in the range of +57 to +71 % as compared to average of the base period.