Using rank-order geostatistics for spatial interpolation of highly skewed data in a heavy-metal contaminated site

Heavy metals concentration in soils across the conterminous USA: Spatial prediction, model uncertainty, and influencing factors

Assessment and proper management of sites contaminated with heavy metals require precise information on the spatial distribution of these metals. This study aimed to predict and map the distribution of Cd, Cu, Ni, Pb, and Zn across the conterminous USA using point observations, environmental variables, and Histogram-based Gradient Boosting (HGB) modeling. Over 9180 surficial soil observations from the Soil Geochemistry Spatial Database (SGSD) (n = 1150), the Geochemical and Mineralogical Survey of Soils (GMSS) (n = 4857), and the Holmgren Dataset (HD) (n = 3400), and 28 covariates (100 m × 100 m grid) representing climate, topography, vegetation, soils, and anthropic activity were compiled. Model performance was evaluated on 20 % of the data not used in calibration using the coefficient of determination (R2), concordance correlation coefficient (ρc), and root mean square error (RMSE) indices. Uncertainty of predictions was calculated as the difference between the estimated 95 and 5 % quantiles provided by HGB. The model explained up to 50 % of the variance in the data with RMSE ranging between 0.16 (mg kg-1) for Cu and 23.4 (mg kg-1) for Zn, respectively. Likewise, ρc ranged between 0.55 (Cu) and 0.68 (Zn), respectively, and Zn had the highest R2 (0.50) among all predictions. We observed high Pb concentrations near urban areas. Peak concentrations of all studied metals were found in the Lower Mississippi River Valley. Cu, Ni, and Zn concentrations were higher on the West Coast; Cd concentrations were higher in the central USA. Clay, pH, potential evapotranspiration, temperature, and precipitation were among the model's top five important covariates for spatial predictions of heavy metals. The combined use of point observations and environmental covariates coupled with machine learning provided a reliable prediction of heavy metals distribution in the soils of the conterminous USA. The updated maps could support environmental assessments, monitoring, and decision-making with this methodology applicable to other soil databases, worldwide.

Screen for low-arsenic-risk rice varieties based on environment-genotype interactions by using GGE analysis

Arsenic (As) accumulation in rice is a global health concern that has received increased attention in recent years. In this study, 12 rice genotypes were cultivated at four As-contaminated paddy sites in Taiwan. According to the different crop seasons and As levels in the soil, the sites were further divided into 18 environmental conditions. For As in soils, results showed that 67% of the studied environments were likely to represent As contamination. For As in rice, the mean total As concentration in brown rice grains ranged from 0.17 to 0.45 mg kg-1. The analysis of variance for the environment effect indicated that grain As concentration was mainly affected by the environmental conditions, suggesting that there was a remarkable degree of variation across the trial environments. According to the combination of the GGE biplot and cumulative distribution function of order statistics (CDFOS) analysis, five genotypes-TCS17, TCS10, TT30, KH139, and TC192-were regarded as stable, low-risk genotypes because the probability of grain As concentration exceeding the maximum permissible concentration (MPC) was lower for these genotypes across all environmental conditions. Particularly, TCS17 was recommended to be the safest rice genotype. Thus, grain As levels in the selected genotypes were applied to assess the health risk to Taiwanese residents associated with As exposure through rice consumption. Results showed that the upper 75th percentile values of the hazard quotient were all less than unity. This suggested that the health risk associated with consuming the selected rice genotypes was acceptable for most of the residents. The methodology developed here would be applicable to screen for stable, low-As-risk rice genotypes across multiple field environments in other regions or countries.

Effectiveness of predicting the spatial distributions of target contaminants of a coking plant based on their related pollutants

The prediction accuracy of the spatial distribution of soil pollutants at a site is relatively low. Related pollutants can be used as auxiliary variables to improve the prediction accuracy. However, little relevant research has been conducted on site soil pollution. To analyze the prediction accuracy of target pollutants combined with auxiliary pollutants, Cu, toluene, and phenanthrene were selected as the target pollutants for this study. Based on geostatistical analysis and spatial analysis, the following results were obtained. (1) The reduction in the root mean square errors (RMSEs) for Cu, toluene, and phenanthrene with multivariable cokriging was 68.4%, 81.6%, and 81.2%, respectively, which are proportional to the correlation coefficient of the relationship between the auxiliary pollutants and the target pollutants. (2) The RMSEs calculated for the multivariable cokriging were lower than those obtained by only combining one related pollutants, and two co-variables should be better. (3) The predicted results for Cu, phenanthrene, and toluene and their corresponding related pollutants are more accurate than the results obtained not using the related pollutants. (4) In the interpolation process, the RMSEs for Cu, toluene, and phenanthrene with multivariable cokriging basically increase as the neighborhood sample data increases, and then they become stable. (5) When 84, 61, and 34 sample points were removed, the RMSEs for Cu, toluene, and phenanthrene, respectively, with multivariable cokriging were close to the RMSEs of the target pollutants based on the total samples. The results are of great significance to improving the prediction accuracy of the spatial distribution of soil pollutants at coking plant sites.

A quantitative review and meta-analysis on phytoscreening applied to aquifers contaminated by chlorinated ethenes

Applications and acceptance of phytoscreening, i.e., the use of trees as screening tools for underground contamination, are still limited in many countries due to the lack of awareness of application policies, the intrinsic qualitative nature of the technique, and the paucity of critical analyses on available data. To date, the conditions influencing the effectiveness of the technique have been descriptively discussed, yet rarely quantified. This review will contribute to filling this knowledge gap, shedding light on the most suitable approaches to apply phytoscreening. The focus was placed specifically on chlorinated ethene compounds since they are among the main organic contaminants in groundwater and have been the most studied in the field of phytoscreening. Chlorinated ethenes' behavior and biodegradation potential largely depend on their physicochemical properties as well as the hydrogeological features of the system in which they migrate. Besides, their fate and transport in surface ecosystems are still poorly understood. Here, phytoscreening data from sites contaminated by chlorinated ethenes were extracted from relevant literature to form a global-scale database. Data were statistically analyzed to identify the major drivers of variability in tree-cores concentration. Correlation between tree-core and groundwater concentration was quantified through Spearman's rank coefficients, whilst detectability potential was determined based on tree-cores showing non-detection of contaminants. The influence on such parameters of factors like contaminant properties, hydrogeology, tree features, and sampling/analytical protocols was assessed. Results suggest that factors controlling plant uptake and contaminant phytovolatilization regulate correlation and detectability, respectively. Conditions increasing the correlation (e.g., sites with shallow and permeable aquifers) are recommended for phytoscreening applications aimed at mapping and monitoring contaminant plumes, whereas conditions increasing detectability (e.g., sampling tree-cores near ground level) are recommended to preliminary screen underground contamination in poorly investigated areas.

Status assessment and probabilistic health risk modeling of metals accumulation in agriculture soils across China: A synthesis

Heavy metal accumulation in agriculture soils is of particular concern in China, while the status and probabilistic health risks of metal contamination in Chinese agriculture soils have been rarely studied at the national scale. In this study, we compiled a database of heavy metal concentrations in Chinese agriculture soils and selected six heavy metals for pollution assessment and risk screening: arsenic (As), cadmium (Cd), chromium (Cr), nickel (Ni), lead (Pb) and Zinc (Zn). Monte Carlo simulation was employed to assess the probabilistic health risks, the associated uncertainties, as well as variations in toxicity parameters, ingestion rate and body weight. Results indicated that the concentrations of Cd were elevated above their reference standard and Cd had the highest mean geo-accumulation index (I_geo) of 1.79. Moreover, the mean hazard index (HI) through exposure to six heavy metals was 1.85E-01 and 2.87E-02 for children and adults, respectively, with 2.2% of non-cancer risks for children that exceeded the guideline value of 1. In contrast, 95.0% and 90.0% of the total cancer risks (TCR) through exposure to six heavy metals for children and adults, respectively, exceeded the guideline value of 1E-06. Six metals were ranked based on their percent of risk outputs exceeding the guideline values. Arsenic had the high exceedance of both cancer and non-cancer risks, while both Cr and Cd were metals with high concern that had high exceedance of cancer risk. Sensitivity analyses indicated that metal concentrations and ingestion rate of soil were the predominant contributors to total risk variance. Overall, the adverse health risks induced by exposure to heavy metals contaminated farmland were elevated. Results from this study may provide valuable implications for public health professionals and policy-makers to design effective strategy to manage nation-wide farmland and reduce heavy metal exposure.

Geostatistical interpolation of available copper in orchard soil as influenced by planting duration

Mapping the spatial distribution of available copper (A-Cu) in orchard soils is important in agriculture and environmental management. However, data on the distribution of A-Cu in orchard soils is usually highly variable and severely skewed due to the continuous input of fungicides. In this study, ordinary kriging combined with planting duration (OK_PD) is proposed as a method for improving the interpolation of soil A-Cu. Four normal distribution transformation methods, namely, the Box-Cox, Johnson, rank order, and normal score methods, were utilized prior to interpolation. A total of 317 soil samples were collected in the orchards of the Northeast Jiaodong Peninsula. Moreover, 1472 orchards were investigated to obtain a map of planting duration using Voronoi tessellations. The soil A-Cu content ranged from 0.09 to 106.05 with a mean of 18.10 mg kg^-1, reflecting the high availability of Cu in the soils. Soil A-Cu concentrations exhibited a moderate spatial dependency and increased significantly with increasing planting duration. All the normal transformation methods successfully decreased the skewness and kurtosis of the soil A-Cu and the associated residuals, and also computed more robust variograms. OK_PD could generate better spatial prediction accuracy than ordinary kriging (OK) for all transformation methods tested, and it also provided a more detailed map of soil A-Cu. Normal score transformation produced satisfactory accuracy and showed an advantage in ameliorating smoothing effect derived from the interpolation methods. Thus, normal score transformation prior to kriging combined with planting duration (NSOK_PD) is recommended for the interpolation of soil A-Cu in this area.

Spatial pattern of heavy metals accumulation risk in urban soils of Beijing and its influencing factors

Accumulations of heavy metals in urban soils are highly spatial heterogeneity and affected by multiple factors including soil properties, land use and pattern, population and climatic conditions. We studied accumulation risks of Cd, Cu, Pb and Zn in unban soils of Beijing and their influencing based on the regression tree analysis and a GIS-based overlay model. Result shows that Zinc causes the most extensive soil pollution and Cu result in the most acute soil pollution. The soil's organic carbon content and CEC and population growth are the most significant factors affecting heavy metal accumulation. Other influence factors in land use pattern, urban landscape, and wind speed also contributed, but less pronounced. The soils in areas with higher degree of urbanization and surrounded by intense vehicular traffics have higher accumulation risk of Cd, Cu, Pb, and Zn.

Accuracy and uncertainty analysis of soil Bbf spatial distribution estimation at a coking plant-contaminated site based on normalization geostatistical technologies

Data distribution is usually skewed severely by the presence of hot spots in contaminated sites. This causes difficulties for accurate geostatistical data transformation. Three types of typical normal distribution transformation methods termed the normal score, Johnson, and Box-Cox transformations were applied to compare the effects of spatial interpolation with normal distribution transformation data of benzo(b)fluoranthene in a large-scale coking plant-contaminated site in north China. Three normal transformation methods decreased the skewness and kurtosis of the benzo(b)fluoranthene, and all the transformed data passed the Kolmogorov-Smirnov test threshold. Cross validation showed that Johnson ordinary kriging has a minimum root-mean-square error of 1.17 and a mean error of 0.19, which was more accurate than the other two models. The area with fewer sampling points and that with high levels of contamination showed the largest prediction standard errors based on the Johnson ordinary kriging prediction map. We introduce an ideal normal transformation method prior to geostatistical estimation for severely skewed data, which enhances the reliability of risk estimation and improves the accuracy for determination of remediation boundaries.

Assessment of lead pollution in topsoils of a southern Italy area: Analysis of urban and peri-urban environment

Exposure to lead (Pb) may affect adversely human health. Mapping soil Pb contents is essential to obtain a quantitative estimate of potential risk of Pb contamination. The main aim of this paper was to determine the soil Pb concentrations in the urban and peri-urban area of Cosenza-Rende to map their spatial distribution and assess the probability that soil Pb concentration exceeds a critical threshold that might cause concern for human health. Samples were collected at 149 locations from residual and non-residual topsoil in gardens, parks, flower-beds, and agricultural fields. Fine earth fraction of soil samples was analyzed by X-ray Fluorescence spectrometry. Stochastic images generated by the sequential Gaussian simulation were jointly combined to calculate the probability of exceeding the critical threshold that could be used to delineate the potentially risky areas. Results showed areas in which Pb concentration values were higher to the Italian regulatory values. These polluted areas were quite large and likely, they could create a significant health risk for human beings and vegetation in the near future. The results demonstrated that the proposed approach can be used to study soil contamination to produce geochemical maps, and identify hot-spot areas for soil Pb concentration.

Mapping and determinism of soil microbial community distribution across an agricultural landscape

Despite the relevance of landscape, regarding the spatial patterning of microbial communities and the relative influence of environmental parameters versus human activities, few investigations have been conducted at this scale. Here, we used a systematic grid to characterize the distribution of soil microbial communities at 278 sites across a monitored agricultural landscape of 13 km². Molecular microbial biomass was estimated by soil DNA recovery and bacterial diversity by 16S rRNA gene pyrosequencing. Geostatistics provided the first maps of microbial community at this scale and revealed a heterogeneous but spatially structured distribution of microbial biomass and diversity with patches of several hundreds of meters. Variance partitioning revealed that both microbial abundance and bacterial diversity distribution were highly dependent of soil properties and land use (total variance explained ranged between 55% and 78%). Microbial biomass and bacterial richness distributions were mainly explained by soil pH and texture whereas bacterial evenness distribution was mainly related to land management. Bacterial diversity (richness, evenness, and Shannon index) was positively influenced by cropping intensity and especially by soil tillage, resulting in spots of low microbial diversity in soils under forest management. Spatial descriptors also explained a small but significant portion of the microbial distribution suggesting that landscape configuration also shapes microbial biomass and bacterial diversity.

Contrasting spatial patterns and ecological attributes of soil bacterial and archaeal taxa across a landscape

Even though recent studies have clarified the influence and hierarchy of environmental filters on bacterial community structure, those constraining bacterial populations variations remain unclear. In consequence, our ability to understand to ecological attributes of soil bacteria and to predict microbial community response to environmental stress is therefore limited. Here, we characterized the bacterial community composition and the various bacterial taxonomic groups constituting the community across an agricultural landscape of 12 km(2) , by using a 215 × 215 m systematic grid representing 278 sites to precisely decipher their spatial distribution and drivers at this scale. The bacterial and Archaeal community composition was characterized by applying 16S rRNA gene pyrosequencing directly to soil DNA from samples. Geostatistics tools were used to reveal the heterogeneous distribution of bacterial composition at this scale. Soil physical parameters and land management explained a significant amount of variation, suggesting that environmental selection is the major process shaping bacterial composition. All taxa systematically displayed also a heterogeneous and particular distribution patterns. Different relative influences of soil characteristics, land use and space were observed, depending on the taxa, implying that selection and spatial processes might be differentially but not exclusively involved for each bacterial phylum. Soil pH was a major factor determining the distribution of most of the bacterial taxa and especially the most important factor explaining the spatial patterns of α-Proteobacteria and Planctomycetes. Soil texture, organic carbon content and quality were more specific to a few number of taxa (e.g., β-Proteobacteria and Chlorobi). Land management also influenced the distribution of bacterial taxa across the landscape and revealed different type of response to cropping intensity (positive, negative, neutral or hump-backed relationships) according to phyla. Altogether, this study provided valuable clues about the ecological behavior of soil bacterial and archaeal taxa at an agricultural landscape scale and could be useful for developing sustainable strategies of land management.

Spatial variability of soil total and DTPA-extractable cadmium caused by long-term application of phosphate fertilizers, crop rotation, and soil characteristics

Increasing cadmium (Cd) accumulation in agricultural soils is undesirable due to its hazardous influences on human health. Thus, having more information on spatial variability of Cd and factors effective to increase its content on the cultivated soils is very important. Phosphate fertilizers are main contamination source of cadmium (Cd) in cultivated soils. Also, crop rotation is a critical management practice which can alter soil Cd content. This study was conducted to evaluate the effects of long-term consumption of the phosphate fertilizers, crop rotations, and soil characteristics on spatial variability of two soil Cd species (i.e., total and diethylene triamine pentaacetic acid (DTPA) extractable) in agricultural soils. The study was conducted in wheat farms of Khuzestan Province, Iran. Long-term (27-year period (1980 to 2006)) data including the rate and the type of phosphate fertilizers application, the respective area, and the rotation type of different regions were used. Afterwards, soil Cd content (total or DTPA extractable) and its spatial variability in study area (400,000 ha) were determined by sampling from soils of 255 fields. The results showed that the consumption rate of di-ammonium phosphate fertilizer have been varied enormously in the period study. The application rate of phosphorus fertilizers was very high in some subregions with have extensive agricultural activities (more than 95 kg/ha). The average and maximum contents of total Cd in the study region were obtained as 1.47 and 2.19 mg/kg and DTPA-extractable Cd as 0.084 and 0.35 mg/kg, respectively. The spatial variability of Cd indicated that total and DTPA-extractable Cd contents were over 0.8 and 0.1 mg/kg in 95 and 25 % of samples, respectively. The spherical model enjoys the best fitting and lowest error rate to appraise the Cd content. Comparing the phosphate fertilizer consumption rate with spatial variability of the soil cadmium (both total and DTPA extractable) revealed the high correlation between the consumption rate of P fertilizers and soil Cd content. Rotation type was likely the main effective factor on variations of the soil DTPA-extractable Cd contents in some parts (eastern part of study region) and could explain some Cd variation. Total Cd concentrations had significant correlation with the total neutralizing value (p < 0.01), available P (p < 0.01), cation exchange capacity (p < 0.05), and organic carbon (p < 0.05) variables. The DTPA-extractable Cd had significant correlation with OC (p < 0.01), pH, and clay content (p < 0.05). Therefore, consumption rate of the phosphate fertilizers and crop rotation are important factors on solubility and hence spatial variability of Cd content in agricultural soils.

Pollutant plume delineation from tree core sampling using standardized ranks

There are currently contradicting results in the literature about the way chloroethene (CE) concentrations from tree core sampling correlate with those from groundwater measurements. This paper addresses this issue by focusing on groundwater and tree core datasets in CE contaminated site, Czech Republic. Preliminary analyses revealed strongly and positively skewed distributions for the tree core dataset, with an intra-tree variability accounting for more than 80% of the total variability, while the spatial analyses based on variograms indicated no obvious spatial pattern for CE concentration. Using rank transformation, it is shown how the results were improved by revealing the initially hidden spatial structure for both variables when they are handled separately. However, bivariate analyses based on cross-covariance functions still failed to indicate a clear spatial correlation between groundwater and tree core measurements. Nonetheless, tree core sampling and analysis proved to be a quick and inexpensive semi-quantitative method and a useful tool.

Altered vegetative assemblage trajectories within an urban brownfield

Recognizing the growing importance of both structure (maintenance of biodiversity) and function (fostering natural cycles) of urban ecologies, we examine coarse scale (herbaceous, shrub and forest) beta guild trajectory in an urban brownfield. The distribution of the pioneer forest assemblage dominated by Betula populifolia Marsh. and Populus spp. could be correlated positively with total soil metal load (arsenic, cadmium, chromium, copper, lead, zinc, lead and vanadium),whereas herbaceous and shrub guilds were negatively correlated. Distinct assemblage development trajectories above and below a critical soil metal threshold are demonstrated. In addition, we postulate that the translocation of metals into the plant tissue of several dominant species may provide a positive feedback loop, maintaining relatively high concentrations of metals in the litter and soil. Therefore assembly theory, which allows for the development of alternate stable states, may provide a better model for the establishment of restoration objectives on degraded urban sites.

Spatial uncoupling of biodegradation, soil respiration, and PAH concentration in a creosote contaminated soil

Hotspots and coldspots of concentration and biodegradation of polycyclic aromatic hydrocarbons (PAHs) marginally overlapped at the 0.5-100 m scale in a creosote contaminated soil in southern Sweden, suggesting that concentration and biodegradation had little spatial co-variation. Biodegradation was substantial and its spatial variability considerable and highly irregular, but it had no spatial autocorrelation. The soil concentration of PAHs explained only 20-30% of the variance of their biodegradation. Soil respiration was spatially autocorrelated. The spatial uncoupling between biodegradation and soil respiration seemed to be governed by the aging of PAHs in the soil, since biodegradation of added 13C phenanthrene covaried with both soil respiration and microbial biomass. The latter two were also correlated with high concentrations of phospholipid fatty acids (PLFAs) that are common in gram-negative bacteria. However, several of the hotspots of biodegradation coincided with hotspots for the distribution of a PLFA indicative of fungal biomass.

Combining a finite mixture distribution model with indicator kriging to delineate and map the spatial patterns of soil heavy metal pollution in Chunghua County, central Taiwan

This study identifies the natural background, anthropogenic background and distribution of contamination caused by heavy metal pollutants in soil in Chunghua County of central Taiwan by using a finite mixture distribution model (FMDM). The probabilities of contaminated area distribution are mapped using single-variable indicator kriging and multiple-variable indicator kriging (MVIK) with the FMDM cut-off values and regulation thresholds for heavy metals. FMDM results indicate that Cr, Cu, Ni and Zn can be individually fitted by a mixture model representing the background and contamination distributions of the four metals in soil. The FMDM cut-off values for contamination caused by the metals are close to the regulation thresholds, except for the cut-off value of Zn. The receiver operating characteristic (ROC) curve validates that indicator kriging and MVIK with FMDM cut-off values can reliably delineate heavy metals contamination, particularly for areas lacking background information and high heavy metal concentrations in soil.

Validation of Bayesian kriging of arsenic, chromium, lead, and mercury surface soil concentrations based on internode sampling

Bayesian kriging is a useful tool for estimating spatial distributions of metals; however, estimates are generally only verified statistically. In this study surface soil samples were collected on a uniform grid and analyzed for As, Cr, Pb, and Hg. The data were interpolated at individual locations by Bayesian kriging. Estimates were validated using a leave-one-out cross validation (LOOCV) statistical method which compared the measured and LOOCV predicted values. Validation also was carried out using additional field sampling of soil metal concentrations at points between original sampling locations, which were compared to kriging prediction distributions. LOOCV results suggest that Bayesian kriging was a good predictor of metal concentrations. When measured internode metal concentrations and estimated kriged values were compared, the measured values were located within the 5th-95th percentile prediction distributions in over half of the internode locations. Estimated and measured internode concentrations were most similar for As and Pb. Kriged estimates did not compare as well to measured values for concentrations below the analytical minimum detection limit, or for internode samples that were very close to the original sampling node. Despite inherent variability in, metal concentrations in soils, the kriged estimates were validated statistically and by in situ measurement.

Mapping field-scale spatial patterns of size and activity of the denitrifier community

There is ample evidence that microbial processes can exhibit large variations in activity on a field scale. However, very little is known about the spatial distribution of the microbial communities mediating these processes. Here we used geostatistical modelling to explore spatial patterns of size and activity of the denitrifying community, a functional guild involved in N-cycling, in a grassland field subjected to different cattle grazing regimes. We observed a non-random distribution pattern of the size of the denitrifier community estimated by quantification of the denitrification genes copy numbers with a macro-scale spatial dependence (6-16 m) and mapped the distribution of this functional guild in the field. The spatial patterns of soil properties, which were strongly affected by presence of cattle, imposed significant control on potential denitrification activity, potential N(2)O production and relative abundance of some denitrification genes but not on the size of the denitrifier community. Absolute abundance of most denitrification genes was not correlated with the distribution patterns of potential denitrification activity or potential N(2)O production. However, the relative abundance of bacteria possessing the nosZ gene encoding the N(2)O reductase in the total bacterial community was a strong predictor of the N(2)O/(N(2) + N(2)O) ratio, which provides evidence for a relationship between bacterial community composition based on the relative abundance of denitrifiers in the total bacterial community and ecosystem processes. More generally, the presented geostatistical approach allows integrated mapping of microbial communities, and hence can facilitate our understanding of relationships between the ecology of microbial communities and microbial processes along environmental gradients.

Soil metal concentrations and productivity of Betula populifolia (gray birch) as measured by field spectrometry and incremental annual growth in an abandoned urban Brownfield in New Jersey

A forested brownfield within Liberty State Park, Jersey City, New Jersey, USA, has soils with arsenic, chromium, lead, zinc and vanadium at concentrations above those considered ambient for the area. Using both satellite imagery and field spectral measurements, this study examines plant productivity at the assemblage and individual specimen level. Longer term growth trends (basal area increase in tree cores) were also studied. Leaf chlorophyll content within the hardwood assemblage showed a threshold model for metal tolerance, decreasing significantly beyond a soil total metal load (TML) of 3.0. Biomass production (calculated with RG-Red/Green Ratio Index) in Betula populifolia (gray birch), the co-dominant tree species, had an inverse relationship with the Zn concentration in leaf tissue during the growing season. Growth of B. populifolia exhibited a significant relationship with TML. Assemblage level NDVI and individual tree NDVI also had significant decreases with increasing TML. Ecosystem function measured as plant production is impaired at a critical soil metal load.

On-site gamma dose rates at the Andreeva Bay shore technical base, northwest Russia

The spent nuclear fuel (SNF) and radioactive waste (RAW) storage facility at Andreeva Bay shore technical base (STB) is one of the largest and most hazardous nuclear legacy sites in northwest Russia. Originally commissioned in the 1960s the facility now stores large amounts of SNF and RAW associated with the Russian Northern Fleet of nuclear powered submarines. The objective of the present study was to map ambient gamma dose rates throughout the facility, in particular at a number of specific sites where SNF and RAW are stored. The data presented here are taken from a Norwegian-Russian collaboration enabling the first publication in the scientific literature of the complete survey of on-site dose rates. Results indicate that elevated gamma dose rates are found primarily at discrete sites within the facility; maximum dose rates of up to 1000 microSv/h close to the ground (0.1m) and up to 3000 microSv/h at 1m above ground were recorded, higher doses at the 1m height being indicative primarily of the presence of contaminated equipment as opposed to ground contamination. Highest dose rates were measured at sites located in the immediate vicinity of buildings used for storing SNF and sites associated with storage of solid and liquid radioactive wastes. Elevated dose rates were also observed near the former channel of a small brook that became heavily contaminated as a result of radioactive leaks from the SNF storage at Building 5 starting in 1982. Isolated patches of elevated dose rates were also observed throughout the STB. A second paper detailing the radioactive soil contamination at the site is published in this issue of Journal of Environmental Radioactivity.

On-site radioactive soil contamination at the Andreeva Bay shore technical base, Northwest Russia

The radioactive waste (RAW) storage site at Andreeva Bay in the Russian Northwest has experienced radioactive contamination both as a result of activities carried out at the site and due to incidents that have occurred there in the past such as accidental releases of radioactive materials. The site is an interesting case study for decommissioning due to the extremely large amounts of radioactivity present at the site and the conditions under which it is stored; very little has been previously published in the scientific literature about this site. This paper complements the paper describing dose rates at Andreeva Bay which is published in this issue of Journal of Environmental Radioactivity by the same authors. This study presents new data related to the activity concentrations of (137)Cs and (90)Sr in surface soils and measurements of alpha- and beta-particle fluxes taken at different areas around the site. Limited data on 60Co is also presented. The results of the study indicate that the main areas of site contamination are associated with the former spent nuclear fuel storage facility at Building 5, due to accidental discharges which began in 1982. Substantial contamination is also observed at the solid radioactive waste storage facilities, probably due to the ingress of water into these facilities. More than 240 samples were measured: maximum contamination levels were 1 x 10(6)Bq/kg (137)Cs (mean value 4.1 x 10(5)Bq/kg) and 4 x 10(6)Bq/kg (90)Sr (mean value 1.2 x1 0(5)Bq/kg). Localised patches of alpha and beta contamination were also observed throughout the site.

Soil metal concentrations and vegetative assemblage structure in an urban brownfield

Anthropogenic sources of toxic elements have had serious ecological and human health impacts. Analysis of the soil samples from a brownfield within Liberty State Park, Jersey City, NJ, USA, showed that arsenic, chromium, lead, zinc and vanadium exist at concentrations above those considered ambient for the area. Accumulation and translocation features were characterized for the dominant plant species of four vegetative assemblages. The trees Betula populifolia and Populus deltoides were found to be accumulating Zn in leaf tissue at extremely high levels. B. populifolia, P. deltoides and Rhus copallinum accumulated Cr primarily in the root tissue. A comparison of soil metal maps and vegetative assemblage maps indicates that areas of increasing total soil metal load were dominated by successional northern hardwoods while semi-emergent marshes consisting mostly of endemic species were restricted primarily to areas of low soil metal load.

Additional sampling based on regulation threshold and kriging variance to reduce the probability of false delineation in a contaminated site

Kriging-based delineation when used to determine a cost-effective remediation plan should be based on the spatial distribution of the pollutant. This study proposed an adaptive cluster sampling (ACS) approach based on the regulation threshold and kriging variance for additional sampling to improve the reliability of delineating a heavy-metal contaminated site. A reliability index for reducing the probability of false delineation was used to determine the size and configuration of additional samples. A data set of Ni concentrations in soil was used for illustration. The results showed that the additional sampled observations during ACS were clustered where the Ni concentrations were close to the regulation threshold of 200 mg kg(-1), and were located where the first-phased sampling density was low. Compared with a simple random sampling (SRS), the relative frequency of misclassification over the whole study area (RFMW) using ACS in a 100 replicates simulation was lower when the same sample number of pooled data was used. In addition, the spatial distribution of the local misclassification rate (LMR) showed that the area with a high-valued LMR could be reduced and that the LMR gradients in the region could be lowered by using ACS instead of SRS. The above results suggest that the proposed ACS approach could improve the reliability of kriging-based delineation of heavy-metal contaminated soils.

Adaptive sampling based on the cumulative distribution function of order statistics to delineate heavy-metal contaminated soils using kriging

Correctly classifying "contaminated" areas in soils, based on the threshold for a contaminated site, is important for determining effective clean-up actions. Pollutant mapping by means of kriging is increasingly being used for the delineation of contaminated soils. However, those areas where the kriged pollutant concentrations are close to the threshold have a high possibility for being misclassified. In order to reduce the misclassification due to the over- or under-estimation from kriging, an adaptive sampling using the cumulative distribution function of order statistics (CDFOS) was developed to draw additional samples for delineating contaminated soils, while kriging. A heavy-metal contaminated site in Hsinchu, Taiwan was used to illustrate this approach. The results showed that compared with random sampling, adaptive sampling using CDFOS reduced the kriging estimation errors and misclassification rates, and thus would appear to be a better choice than random sampling, as additional sampling is required for delineating the "contaminated" areas.

Geostatistics and GIS: tools for characterizing environmental contamination

Geostatistics is a set of statistical techniques used in the analysis of georeferenced data that can be applied to environmental contamination and remediation studies. In this study, the 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene (DDE) contamination at a Superfund site in western Maryland is evaluated. Concern about the site and its future clean up has triggered interest within the community because residential development surrounds the area. Spatial statistical methods, of which geostatistics is a subset, are becoming increasingly popular, in part due to the availability of geographic information system (GIS) software in a variety of application packages. In this article, the joint use of ArcGIS software and the R statistical computing environment are demonstrated as an approach for comprehensive geostatistical analyses. The spatial regression method, kriging, is used to provide predictions of DDE levels at unsampled locations both within the site and the surrounding areas where residential development is ongoing.

Spatial interpolation methods for nonstationary plume data

Plume interpolation consists of estimating contaminant concentrations at unsampled locations using the available contaminant data surrounding those locations. The goal of ground water plume interpolation is to maximize the accuracy in estimating the spatial distribution of the contaminant plume given the data limitations associated with sparse monitoring networks with irregular geometries. Beyond data limitations, contaminant plume interpolation is a difficult task because contaminant concentration fields are highly heterogeneous, anisotropic, and nonstationary phenomena. This study provides a comprehensive performance analysis of six interpolation methods for scatter-point concentration data, ranging in complexity from intrinsic kriging based on intrinsic random function theory to a traditional implementation of inverse-distance weighting. High resolution simulation data of perchloroethylene (PCE) contamination in a highly heterogeneous alluvial aquifer were used to generate three test cases, which vary in the size and complexity of their contaminant plumes as well as the number of data available to support interpolation. Overall, the variability of PCE samples and preferential sampling controlled how well each of the interpolation schemes performed. Quantile kriging was the most robust of the interpolation methods, showing the least bias from both of these factors. This study provides guidance to practitioners balancing opposing theoretical perspectives, ease-of-implementation, and effectiveness when choosing a plume interpolation method.

Using sequential indicator simulation to assess the uncertainty of delineating heavy-metal contaminated soils

Mapping the spatial distribution of soil pollutants is essential for delineating contaminated areas. Currently, geostatistical interpolation, kriging, is increasingly used to estimate pollutant concentrations in soils. The kriging-based approach, indicator kriging (IK), may be used to model the uncertainty of mapping. However, a smoothing effect is usually produced when using kriging in pollutant mapping. The detailed spatial patterns of pollutants could, therefore, be lost. The local uncertainty of mapping pollutants derived by the IK technique is referred to as the conditional cumulative distribution function (ccdf) for one specific location (i.e. single-location uncertainty). The local uncertainty information obtained by IK is not sufficient as the uncertainty of mapping at several locations simultaneously (i.e. multi-location uncertainty or spatial uncertainty) is required to assess the reliability of the delineation of contaminated areas. The simulation approach, sequential indicator simulation (SIS), which has the ability to model not only single, but also multi-location uncertainties, was used, in this study, to assess the uncertainty of the delineation of heavy metal contaminated soils. To illustrate this, a data set of Cu concentrations in soil from Taiwan was used. The results show that contour maps of Cu concentrations generated by the SIS realizations exhausted all the spatial patterns of Cu concentrations without the smoothing effect found when using the kriging method. Based on the SIS realizations, the local uncertainty of Cu concentrations at a specific location of x', refers to the probability of the Cu concentration z(x') being higher than the defined threshold level of contamination (z(c)). This can be written as Prob(SIS)[z(x')>z(c)], representing the probability of contamination. The probability map of Prob(SIS)[z(x')>z(c)] can then be used for delineating contaminated areas. In addition, the multi-location uncertainty of an area A,delineated as contaminated based on the probability map of Prob(SIS)[z(x')>z(c)], can be calculated to assess the reliability of delineation. Multi-location uncertainty refers to the probability of Cu concentrations in several locations, x'(1), x'(2), em leader, x'(m,) in the area A, being higher than the threshold (z(c)) as denoted by Prob(SIS)[z(x'(1))>z(c), z(x'(2))>z(c), em leader, andz(x'(m))>z(c)] or Prob(SIS)[z(A)>z(c)]. The multi-location uncertainty Prob(SIS)[z(A)>z(c)], obtained from the SIS, can be used to assess the reliability of delineation for regions suspected of contamination, (A), which has been delineated as contaminated. Reliance on this information facilitates the decision making process in determining which areas are contaminated and require cleanup action.

Implications of the spatial variability of landfill emission rates on geospatial analyses

Accurate methods quantifying whole landfill surface flux of methane are important for regulatory and research purposes. This paper presents the results from the analysis of chamber measurements utilizing geospatial techniques [kriging and inverse distance weighting (IDW)] to arrive at an estimation of the whole landfill surface flux from the spatially distributed chamber measurement points. The difficulties in utilizing these methods will be discussed. Methane flux was determined on approximately 20 m grid spacing and variogram analysis was performed in order to model spatial structure, which was used to estimate methane flux at unsampled locations through kriging. Our analysis indicates that while the semi-variogram model showed some spatial structure, IDW was a more accurate interpolation method for this particular site. This was seen in the comparison of the resulting contour maps. IDW, coupled with surface area algorithms to extract the total area of user defined contour intervals, provides a superior estimate of the methane flux as confirmed through the methane balance. It is critical that the results of the emissions estimates be viewed in light of the whole cell methane balance; otherwise, there is no rational check and balance system to validate the results.