Identification of coastal water quality by statistical analysis methods in Daya Bay, South China Sea

In this paper, cluster analysis (CA), principal component analysis (PCA) and the fuzzy logic approach were employed to evaluate the trophic status of water quality for 12 monitoring stations in Daya Bay in 2003. CA grouped the four seasons into four groups (winter, spring, summer and autumn) and the sampling sites into two groups (cluster DA: S1, S2, S4-S7, S9 and S12 and cluster DB: S3, S8, S10 and S11). PCA identified the temporal and spatial characteristics of trophic status in Daya Bay. Cluster DB, with higher concentrations of TP and DIN, is located in the western and northern parts of Daya Bay. Cluster DA, with the low Secchi, is located in the southern and eastern parts of Daya Bay. The fuzzy logic approach revealed more information about the temporal and spatial patterns of the trophic status of water quality. Chlorophyll a, TP and Secchi may be major factors for deteriorating water quality.     

Simulation of the Urban Heat Island Phenomenon in Mediterranean Climates

An intelligent data-driven method is used in the present study for investigating, analyzing and quantifying the urban heat island phenomenon in the major Athens region where hourly ambient air-temperature data are recorded at twenty-three stations. The heat island phenomenon has a serious impact on the energy consumption of buildings, increases smog production, while contributing to an increasing emission of pollutants from power plants, including sulfur dioxide, carbon monoxide, nitrous oxides and suspended particulates. The intelligent method is an artificial neural network approach in which the urban heat island intensity at day and nighttime are estimated using as inputs several climatic parameters. Various neural network architectures are designed and trained for the output estimation, which is the daytime and nighttime urban heat island intensity at each station for a two-year time period. The results are tested with extensive sets of non-training measurements and it is found that they correspond well with the actual values. Furthermore, the influence of several input climatic parameters measured at each station, such as solar radiation, daytime and nighttime air temperature, and maximum daily air temperature, on the urban heat island intensity fluctuations is investigated and analyzed separately for the day and nighttime period. From this investigation it is shown that heat island intensity is mainly influenced by urbanization factors. A sensitivity investigation has been performed, based on neural network techniques, in order to adequately quantify the impact of the above input parameters on the urban heat island phenomenon.     

The temporal variation of indoor pollutants in photocopying shop

The paper includes the identification of the main factors responsible for the temporal variations of indoor pollutants during three daily intervals in a photocopying shop. The measurements of concentration levels of total volatile organic compounds, ozone, carbon monoxide, carbon dioxide, nitrogen dioxide, ammonia, perchloroethylene and non-methane hydrocarbons were performed. The individual concentrations of target pollutants were subjected to principal component analysis (PCA) using a software XLSTAT 2014.1.10. Pearson correlation model indicated the relatively weak correlation between the investigated pollutants in a photocopying environment. PCA extracted three principal components (PCs) from the indoor air pollution data set. Obtained PCs explained 56.72 % of the total variance. The summarized biplots showed which pollutants are responsible for photocopying indoor pollution per sampling day/sampling point/time interval/number of measurement. The results pointed out that the main PCs were related to the usage of toners, electrostatic discharge, heating of photocopiers as well as general intensifying of photocopying processes.     

基于RS-PCA-GA-SVM的砂土液化预测方法研究

砂土液化是一种危害性比较大的自然灾害,对砂土液化进行判定预测在地质灾害防治领域中有重要的研究意义。通过粗糙集理论(Rough Set,RS)对影响砂土液化的6个初始评价指标(包括震级、土深、震中距、地下水位、标贯击数和地震持续时间)进行属性约简,去掉冗余或干扰信息,得到基于4个核心预测指标的数据集。通过主成分分析法(Principal Component Analysis,PCA)从核心评价指标中提取出主成分,采用支持向量机(Support Vector Machine,SVM)对数据集进行训练,用遗传算法(Genetic Algorithm,GA)优化参数,建立砂土液化的RS-PCA-GA-SVM预测模型。并结合砂土液化实际数据将预测结果与基于Levenberg-Marquardt算法改进的BP神经网络模型(LM-BP)的预测结果做比较。实例计算表明:基于RS-PCA-GA-SVM模型得到的砂土液化预测结果精度较LM-BP神经网络有很大的提高,判别结果与实际情况比较吻合,可在实际工程中应用。     

Time series analysis and forecasting for air pollution in small urban area: an SARIMA and factor analysis approach

Despite the existing public and government measures for monitoring and control of air quality in Bulgaria, in many regions, including typical and most numerous small towns, air quality is not satisfactory. In this paper, factor analysis and Box–Jenkins methodology are applied to examine concentrations of primary air pollutants such as NO, NO₂, NO_x, PM10, SO₂ and ground level O₃ in the town of Blagoevgrad, Bulgaria within a 1 year period from 1st September 2011 to 31st August 2012, based on hourly measurements. By using factor analysis with PCA and Promax rotation, a high multicollinearity between the six pollutants has been detected. The pollutants were grouped in three factors and the degree of contribution of the factors to the overall pollution was determined. This was interpreted as the presence of common sources of pollution. The main part of the study involves the performance of time series analysis and the development of univariate stochastic seasonal autoregressive integrated moving average (ARIMA) models with recording on a hourly basis as seasonality. The study also incorporates the Yeo–Johnson power transformation for variance stabilizing of the data and model selection by using Bayersian information criterion. The obtained SARIMA models demonstrated very good fitting performance with regard to the observed air pollutants and short-term predictions for 72 h ahead, in particular in the case of ozone and particulate matter PM10. The presented statistical approaches allow the building of non-complex models, effective for short-term air pollution forecasting and useful for advance warning purposes in urban areas.     

Groundwater Quality Modeling with a Small Data Set

Seventeen groundwater quality variables collected during an 8‐year period (2006 to 2013) in Andimeshk, Iran, were used to implement an artificial neural network (NN) with the purpose of constructing a water quality index (WQI). The method leading to the WQI avoids instabilities and overparameterization, two problems common when working with relatively small data sets. The groundwater quality variables used to construct the WQI were selected based on principal component analysis (PCA) by which the number of variables were decreased to six. To fulfill the goals of this study, the performance of three methods (1) bootstrap aggregation with early stopping; (2) noise injection; and (3) ensemble averaging with early stopping was compared. The criteria used for performance analysis was based on mean squared error (MSE) and coefficient of determination (R²) of the test data set and the correlation coefficients between WQI targets and NN predictions. This study confirmed the importance of PCA for variable selection and dimensionality reduction to reduce the risk of overfitting. Ensemble averaging with early stopping proved to be the best performed method. Owing to its high coefficient of determination (R² = 0.80) and correlation coefficient (r=0.91), we recommended ensemble averaging with early stopping as an accurate NN modeling procedure for water quality prediction in similar studies.     

云南星云湖水质变化及其人文因素驱动力分析

星云湖目前存在水污染加重、富营养化进程加快、水体功能受损等问题.以星云湖为研究对象,根据星云湖2005-2015年的水质数据、社会经济统计数据和遥感影像图,运用目视解译、叠加分析、污染足迹模型及主成分分析法,分析了星云湖流域近10年以来水质变化趋势、入湖河流污染物污染足迹及其人文因素驱动力.结果表明:(1)水质数据趋势表明,从月变化看,3月份水质最好,9月份水质最差;从年变化看,2005-2015年间,2008年水质状况最好,2014年的水质状况最差,从2008-2014年水质持续变差,到2015年好转.(2)2015年有机物、氮和磷的污染足迹分别为583.26、705.88和494.11 km~2.污染足迹前4位的入湖河流依次为:大街河东西大河东河渔村河东西大河西河,占星云湖流域总污染足迹的66.21%.污染程度大的大街河、东西大河和渔村河周边土地利用类型为水田、旱地和村庄.(3)星云湖水质影响因素第1主成分(总人口、播种面积、农村人口、化肥使用量、农膜使用量、大牲畜存栏量)与农村生活和农业面源污染有关;第2主成分(人均GDP、第一产业产值、第二产业产值、第三产业产值)与社会经济发展有关.因此,星云湖流域水质变化的人文因素驱动力为农村生活和农业面源污染类和社会经济发展类,其中第1主成分的贡献率是84.389%,农村生活和农业面源污染是水质变化的主要驱动力.     

Modeling and Prediction of Monthly Total Ozone Concentrations by Use of an Artificial Neural Network Based on Principal Component Analysis

In the work discussed in this paper we considered total ozone time series over Kolkata (22°34′10.92″N, 88°22′10.92″E), an urban area in eastern India. Using cloud cover, average temperature, and rainfall as the predictors, we developed an artificial neural network, in the form of a multilayer perceptron with sigmoid non-linearity, for prediction of monthly total ozone concentrations from values of the predictors in previous months. We also estimated total ozone from values of the predictors in the same month. Before development of the neural network model we removed multicollinearity by means of principal component analysis. On the basis of the variables extracted by principal component analysis, we developed three artificial neural network models. By rigorous statistical assessment it was found that cloud cover and rainfall can act as good predictors for monthly total ozone when they are considered as the set of input variables for the neural network model constructed in the form of a multilayer perceptron. In general, the artificial neural network has good potential for predicting and estimating monthly total ozone on the basis of the meteorological predictors. It was further observed that during pre-monsoon and winter seasons, the proposed models perform better than during and after the monsoon.     

An empirical analysis of the impact of a bus rapid transit system on the concentration of secondary pollutants in the roadside areas of the TransJakarta corridors

A bus rapid transit (BRT) system began operation in Jakarta City, Indonesia, in January 2004 and led to a modal shift from private to public modes of transport. This modal shift from car and motorcycle to BRT reduced the emission intensity of primary pollutants, such as NO_x and CO. We applied a combined structural equation model and an artificial neural network to evaluate the impact of the BRT system on the concentration of secondary pollutants in the roadside areas in the BRT corridors. An empirical analysis was carried out using data collected at five continuous ambient air quality monitoring stations located near to the BRT TransJakarta corridors in 2005. The establishment of our structural equation model gives a better understanding of the cause–effect relationship among the factors influencing roadside ambient air pollution, and was useful in simplifying the complexity of our artificial neural network model for predicting the modal shift’s impact on the PM₁₀ values and concentration of O₃. The introduction of the BRT system, and the modal shift it produced, had a greater influence on rapidly decaying pollutants, such as PM₁₀, than on O₃ because of the exposure to near-source microenvironments, such as the roadside of the TransJakarta corridors.     

Simulation of Air Quality using an ISCST3 Dispersion Model

Urban air quality is an issue of major concern across many cities in India. In particular, high levels of particulate matter (both SPM and RSPM) are responsible for noncompliance to air quality standards. Air quality modeling is an effective tool to simulate the air quality of a region and to predict air quality concentrations under different scenarios. Kanpur city which is a top‐ten urban conglomerate in India (based on population) is chosen for the application of the ISCST3 model and simulation of air quality. Sectored emission loads are estimated for transport, industrial, power, and domestic sectors, which provide an estimate of the major contributors to air pollution with specific reference to particulate matter, which is a major pollutant of concern. A detailed scenario analysis is carried out to estimate the changes in emissions that would take place due to various interventions. Dispersion modeling is carried out using the ISCST3 model, to estimate the concentrations of SPM all over the city under different scenarios. Emission inventory and meteorological data served as input to the model, and the air quality is predicted for various seasons and intervention scenarios. The modeled values for the scenario without intervention results in an underestimation of 48%, which is due to unaccountable or unidentified sources, trans‐boundary movement of SPM, and model calibration errors. To overcome the error, the model is calibrated with the observed values and results are obtained for other scenarios using the calibration factor. The paper demonstrates only the research direction currently used to simulate air quality in Indian cities. However, further refinement and research is required before it could be used for more accurate predictions.     

基于PCA的地震官方微博影响力评价研究

随着政务微博用户规模及影响力的不断提升,微博作为地震部门传播平台,在地震信息传播方面发挥着巨大的作用。本文在充分考虑地震部门行业特点的基础上,对30个地震官方微博数据进行收集,利用主成分分析法(PCA)归纳出3种地震微博的主要影响指标,即服务力主成分、交互力主成分和创作力主成分,并由此构建出地震官方微博影响力评估指标体系,在此基础上计算得出各地震官方微博的主成分指标得分和影响力综合得分,最后根据得分情况对地震官方微博影响力提出了具可行性的提升策略。     

基于PCA-RBF神经网络的航空飞行几何参数拟合

采用地面异常线圈对直升机时域航空电磁探测系统进行标定时,发射-接收线圈姿态的变化将导致实测数据产生误差,影响标定的精度.本文基于时间域航空电磁系统,计算了发射-接收线圈姿态任意变化时异常线圈的电磁响应,提出了主成分分析-径向基神经网络(PCA-RBF)的拟合算法,采用主成分分析法提取飞行几何参数的贡献率,利用径向基神经网络法对电磁响应进行了测线剖面的批量数据拟合,并对理论仿真和河南桐柏直升机飞行试验数据进行拟合分析,单一异常体理论数据的绝对误差平均值小于20nV·m-2,双异常体理论数据绝对误差平均值为160nV·m-2.野外实测数据在异常线圈中心位置的拟合相对误差小于1%,整条剖面测线的拟合相对误差小于±6%,平均值为2.5%.结果表明PCA-RBF拟合算法能够较好地实现航空电磁系统飞行参数的拟合,为航空电磁系统海量实测数据的快速处理提供了新方法.     

Patterns of remotely sensed floodplain saturation and its use in runoff predictions

Principal components analysis (PCA) is applied to a time series of European Remote Sensing (ERS) synthetic aperture radar (SAR) scenes of the Alzette River floodplain (Grand‐Duchy of Luxembourg). These images cover markedly different hydrological conditions during several winter seasons in order to enable the examination of the decrease of the radar backscattering signal during drying‐up phases following important flood events. At the floodplain scale, with homogeneous land use and constant topography, the first principal components (PCs) are mainly dominated by the variance related to the changing areas. The PCs are thus mainly controlled by subsurface and surface water dynamics. The field observations of a densely equipped piezometric network in the floodplain are used to calculate a mean soil saturation index (SSI) continuously. A classification scheme, based on the PCs and k‐means algorithm, leads to the segmentation of the floodplain into several hydrological behaviour classes with distinctive responses versus changing moisture conditions. To validate this classification method with ground‐based estimations, the relation between the mean backscattering values of microplots within each PCA‐derived hydrological class and the water table measurements, expressed by means of the SSI, is evaluated. Results show that each class of microplots is characterized by the slope of the ‘backscattering–SSI’ function and by the SSI threshold value at which groundwater resurgence appears. The water ponding implies very low signal return due to the specular backscattering effect on the water surface. Based on established relationships between measured initial water table depths, runoff coefficients and rainfall‐induced water table rises, these results are used to discuss the potential of SAR‐derived information in flood management applications. Copyright © 2005 John Wiley & Sons, Ltd.     

Using multivariate statistical analysis of groundwater major cation and trace element concentrations to evaluate groundwater flow in a regional aquifer

Groundwater samples were collected from 11 springs in Ash Meadows National Wildlife Refuge in southern Nevada and seven springs from Death Valley National Park in eastern California. Concentrations of the major cations (Ca, Mg, Na and K) and 45 trace elements were determined in these groundwater samples. The resultant data were subjected to evaluation via the multivariate statistical technique principal components analysis (PCA), to investigate the chemical relationships between the Ash Meadows and Death Valley spring waters, to evaluate whether the results of the PCA support those of previous hydrogeological and isotopic studies and to determine if PCA can be used to help delineate potential groundwater flow patterns based on the chemical compositions of groundwaters. The results of the PCA indicated that groundwaters from the regional Paleozoic carbonate aquifers (all of the Ash Meadows springs and four springs from the Furnace Creek region of Death Valley) exhibited strong statistical associations, whereas other Death Valley groundwaters were chemically different. The results of the PCA support earlier studies, where potentiometric head levels, δ¹⁸O and δD, geological relationships and rare earth element data were used to evaluate groundwater flow, which suggest groundwater flows from Ash Meadows to the Furnace Creek springs in Death Valley. The PCA suggests that Furnace Creek groundwaters are moderately concentrated Ash Meadows groundwater, reflecting longer aquifer residence times for the Furnace Creek groundwaters. Moreover, PCA indicates that groundwater may flow from springs in the region surrounding Scotty's Castle in Death Valley National Park, to a spring discharging on the valley floor. The study indicates that PCA may provide rapid and relatively cost‐effective methods to assess possible groundwater flow regimes in systems that have not been previously investigated. Copyright © 1999 John Wiley & Sons, Ltd.     

Water quality assessment of the Huaihe River segment of Bengbu (China) using multivariate statistical techniques

In the study, multivariate statistical methods including principal component analysis (PCA)/factor analysis (FA) and cluster analysis (CA) were applied to analyze surface water quality data sets obtained from the Huaihe River segment of Bengbu (HRSB) and generated during 2 years (2011–2012) monitoring of 19 parameters at 7 sampling sites. The results of PCA for 7 sampling sites revealed that the first four components of PCA showed 94.89% of the total variance in the data sets of HRSB. The Principal components (Factors) obtained from FA indicated that the parameters for water quality variations were mainly related to heavy metals (Pb, Mn, Zn and Fe) and organic related parameters (COD, PI and DO). The results revealed that the major causes of water quality deterioration were related to inflow of industrial, domestic and agricultural effluents into the Huaihe River. Three significant sampling locations—(sites 2, 3 and 4), (sites 1 and 5) and (sites 6 and 7)—were detected on the basis of similarity of their water quality. Thus, these methods were believed to be valuable to help water resources managers understand complex nature of water quality issues and determine the priorities to improve water quality.     

A comparative study of three neural network forecast combination methods for simulated river flows of different rainfall—runoff models

Abstract

The performances of three artificial neural network (NN) methods for combining simulated river flows, based on three different neural network structures, are compared. These network structures are: the simple neural network (SNN), the radial basis function neural network (RBFNN) and the multi-layer perceptron neural network (MLPNN). Daily data of eight catchments, located in different parts of the world, and having different hydrological and climatic conditions, are used to enable comparisons of the performances of these three methods to be made. In the case of each catchment, each neural network combination method synchronously uses the simulated river flows of four rainfall—runoff models operating in design non-updating mode to produce the combined river flows. Two of these four models are black-box, the other two being conceptual models. The results of the study show that the performances of all three combination methods are, on average, better than that of the best individual rainfall—runoff model utilized in the combination, i.e. that the combination concept works. In terms of the Nash-Sutcliffe model efficiency index, the MLPNN combination method generally performs better than the other two combination methods tested. For most of the catchments, the differences in the efficiency index values of the SNN and the RBFNN combination methods are not significant but, on average, the SNN form performs marginally better than the more complex RBFNN alternative. Based on the results obtained for the three NN combination methods, the use of the multi-layer perceptron neural network (MLPNN) is recommended as the appropriate NN form for use in the context of combining simulated river flows.     

Application of MODIS satellite products to the air pollution research in Beijing

The direct correlation between NASA MODIS aerosol optical depth (AOD) products and the air pollution index (API) in Beijing was found relatively low based on the long-term comparison analysis. The correlation improved to some extent after taking account of the seasonal variation of scale height and the vertical distribution of aerosols. The correlation coefficient further improved significantly after considering the influencing factor of Relative Humidity (RH). This study concluded that satellite remote-sensing could serve as an efficient tool for monitoring the spatial distribution of particulate pollutants on the ground-level, as long as corrections have been made in the two aforementioned processes. Taking advantage of the MODIS information, we analyzed a pollution episode occurring in October 2004 in Beijing. It indicated that satellite remote-sensing could describe the formation process of the ground-level pollution episode in detail, and showed that regional transport and the topography were crucial factors to air quality in Beijing. The annual averaged distribution in the urban area of Beijing and its surroundings could be also obtained from the high-resolution retrieval results, implicating that high-resolution satellite remote-sensing might be potential in monitoring the source distribution of particulate pollutants.     

Modeling the Effects of Meteorological Factors on SO2 and PM10 Concentrations with Statistical Approaches

Air quality has been deteriorated seriously in urban areas as a result of increasing anthropogenic activities. Meteorological conditions affect air pollution levels in the urban atmosphere significantly due to their important role in transport and dilution of the pollutants. This paper aims to investigate usability of some promising statistical methods for examining the impacts of metrological factors on SO₂ and PM10 levels. Data were collected from city centre of Kocaeli in winter periods from 2007 to 2010 as pollutant concentrations increase in winters due to expanding combustion facilities. Results of bivariate correlation analysis showed that humidity and rainfall have remarkable negative correlations with the pollutants. Multiple linear regression models and artificial neural network (ANN) models were used to predict next day's PM10 and SO₂ levels. In regression models calculated R² values were 0.89 and 0.75 for PM10 and SO₂, respectively. Among the various architectures, single layer networks provided better performance in ANN applications. Highest R² values were obtained as 0.89 and 0.69 for PM10 and SO₂, respectively, by using appropriate networks.