| The contaminated sites are a critical environmental issue of concern all over the world. Especially some large-scale industrial contaminated sites which are the key of object of environmental supervision and risk management of many countries. In recent years, following the rapid economic development and the adjustment of industrial structure, the pollution industrial plants were commonly closed or moved out in economic developed cities, the quantity and the degree of contamination of the remaining industrial contaminated sites appeared to be rising. Soil contamination of industrially contaminated sites has received wide attention in China, which not only directly affects soil physicochemical properties and the environment, but also threatens human health in the contaminated area. The industrial contaminated sites restricted the urban land resource of the sustainable development severely. The regional soil environment is an important part of the whole ecological environment system. The original ecological function and system balance of the regional soil environment was destroyed by industrial contaminated sites, the contaminant in the soils threaten organisms and human health through migration and transformation. The present of the industrial contaminated sites has an important impact on sustainable development of regional economy, land resources reuse, laws and regulations and the relevant industry standards, which has aroused the Chinese government department’s great attention. The environmental risk management of the contaminated sites in China lack of perfect laws and management system compared to the developed country. The theoretical basis and technological needed to perfect immediately. According to the management and application requirements in China, the new theory and technology research was indispensable.Knowledge of the spatial distribution of available pollutants is critical for risk control, remedial strategy development and for determination of the remediation boundary and soil volume. The pollution characteristics of contaminated sites are different from the general non-point source pollution affected by human disturbance and accumulative releasing actor, the hotspots are existed in the local region of contaminated sites. The commonly used spatial interpolation methods can provide an unbiased prediction with minimum variance for the content of a given pollutant, but the interpolation techniques all have a smoothing effect, which underestimates the local high values and overestimates the local low values. The uncertainty in determination of contamination boundary at contaminated sites is affected by spatial interpolation methods. In order to test the influence of different prediction models on the determination of contamination boundary in contaminated sites, a large-scale coking plant contaminated was selected as an example, based on the concentration data of114soil samples, the statistical characteristic of soil sample concentration was analyzed by the multivariate statistical analysis, trend analysis and spatial local variance analysis. The spatial autocorrelation was used to analyze the statistical structure characteristics of samples and hotspots in contaminated site. The typical spatial interpolation models, indicator kriging of the non-parametric geostatistics, Inverse Distance Weighting model (IDW), Johnson’s ordinary lognormal kriging model (OLKM), and Combination Prediction Model (CPM) were employed to compare their efficiencies and precisions in determining site contamination boundary. The3-d visualization method was used to show the3-D spatial distribution of PAHs in different stratum. The main results of the research as follow:(1) The descriptive statistical analysis results indicated that∑PAH levels varied significantly and the data was severely skewed. Correlation matrix (CM) and principal component analysis (PCA) were applied to the data processing. Two principle components (PCs) were extracted, which were representative indexes of the whole site pollution data. PAHs distributions showed a low-high-low trend, in both east-west and north-south directions. The variation coefficient for the site was higher in the center, north-west and south-west directions. This study will significantly contribute to and act as a reference for further investigations into this site’s pollutant sources and subsequent site remedial strategies.(2) The identification of contamination "hotspots" are an important indicator of the degree of contamination in localized areas, which can contribute towards the re-sampling and remedial strategies used in the seriously contamination areas. Accordingly,114surface samples were assessed for16PAHs and analyzed using spatial autocorrelation techniques. The results showed that the Global Moran’s I statistics indicated that the significant autocorrelations were detected and the autocorrelation distances of six indicator PAHs were750m,850m,1200m,850m,750m and1200m, respectively; There were visible high-high values (hotspots) clustered in the mid-bottom part of the site through the Local Moran’s I index analysis. Hotspot identification and spatial distribution results can play a key role in contaminated site investigation and management.(3) The PAHs’ spatial distribution probability in surface-soil (0-50cm) was studied through the indicator kriging of the non-parametric geostatistics; the map of probability distribution with a contaminant target was plotted over the entire site. Results indicated that the indicator semivariograms were stable after the conversion of sample data, but the poor correlation of spatial samples was observed due to the spatial variability. In this site, the distribution of four PAHs’ contamination probability has a similar characteristic, and the areas of probability more than45%are mainly distributed in production process workshops for coking, gas purification, tar products etc, of the central, northwest and southeast site with serious contamination, while the areas of less than45%are mainly distributed in coal preparation, gas purification workshops of the southwest and northeast site. Based on the above analysis, we can draw a conclusion that the forecast probability results are consistent with of the occurrence and pollution resource distribution, which is important for defining the remediation boundary and contaminated soil volume.(4) Three spatial interpolation models, Inverse Distance Weighting model (IDW), Johnson’s ordinary lognormal kriging model (OLKM), and Combination Prediction Model (CPM) were employed to compare their efficiencies and precisions in determining site contamination boundary. A recommended value0.4mg.kg-1for BaP was used as the reference standard based on the Beijing Screening Levels for Soil Environmental Risk Assessment of Sites. The Results showed that the contamination areas calculated by IDW, OLKM, and CPM were70.15%,44.78%and57.06%, respectively. The CPM was found to be more accurate than the other two models in predicting the actual pollution situations of the contaminated site. The area on the top-right corner of the site with less sampling points, and the area in the mid-bottom part with high levels of contamination showed the largest standard errors based on the prediction standard errors map. This study gives great contributions and useful references for dealing with uncertainty in determining remediation boundary of contaminated sites.(5) In order to determine the PAHs polluted boundary and soil volumes need to be restored, and different prediction models had the uncertainty influence on determining the polluted boundary, thus a large domestic coking contaminated site was selected, four3D interpolation models of Krig-3D, IDW-Shepard, IDW-(Franke/Nielson), and Nearest Neighbor were employed to compare their reliability and uncertain in predication. A significant difference of model results was observed, and cross-validation test indicated that the Krig-3D was more accurate in predicting the actual pollution situations. Referring to the remediation goal, the polluted soil volumes calculated by four3D interpolation model were8.51×105,5.62×105,7.12×415,1.09×106m3, respectively. This study based on the typical field site gives great contributions for analysis of pollution spatial distribution characteristics and determination of remediation soil volumes.The possible innovation of this research as follows:(1) Due to the lack of analysis of data statistical characteristics, spatial characteristics and hotspots identification, the environment investigate result has uncertainty and error in the process of contamination spatial distribution estimate, risk assessment and remediation. The paper aimed to provide scientific data analysis methods for organic pollutant sample concentration in contaminated sites.(2) The Combination Prediction Model was introduced to the data of severely skewed and high peak values based on the data analysis and compare different interpolation model. The uncertainty area and the factors of uncertainty were assessed in the process of spatial distribution estimated. |
PDF Full Text Request