Multivariate statistics and geostatistical analyzes of metal elements in soil at waste disposal site in Khulna. Moreover, the spatial distribution of metal elements in the soil is necessary to explore their extent.
CONCLUSION AND RECOMMENDATIONS 149-150
Multivariate Statistics and Geostatistical Analyses of Metal Elements in Soil of Waste Disposal Site in Khulna
Declaration
Approval
Board of Examiners
Acknowledgement
Abstract
Contents
LIST OF TABLES
LIST OF ILLUSTRATIONS
Nomenclature
- Background Information
- Problem Statement and Justification
- Contribution to Knowledge
- Significance of the Study
- Scope and Limitations
- Outline of the Thesis
Develop correlations between metallic elements and their potential sources of contamination in landfill soil. The distributions of metallic elements in and around the soil were also shown for different interpolation techniques.
Chapter I represented a general knowledge on the background of waste disposal sites, MSW, contaminated soil, metal element and the possible sources of contamination of soil
The fabricated surface with the least error provided more clear visualization of distribution of metallic elements in and around the soil of the waste dump. The chapters depicted the knowledge of origin and generation of metallic elements and the sources of the contamination of soil from the selected waste disposal site.
This chapter mainly deals with MSW disposal facilities, MSW and the impacts of disposal sites on.
Introduction
The literature review presented in this chapter has been compiled from available previous research reports and technical articles related to this topic. After preparing a review of the literature related to heavy metals and soil, this chapter also discusses a general discussion of conventional and multivariate statistical analysis.
Municipal Solid Waste
- Overall Impacts of MSW Disposal Site
The practice of landfill system as a method of household waste disposal is rarely adopted in many developing countries (Oyeku and Eludoyin, 2010). Therefore, heavy metal contamination occurs in and around the bottom of the domestic wastewater landfill.
Context of Heavy Metal
- Providence of Heavy Metals in Soil and Environment
- Behavior of Heavy Metals in Soil
- Consequences of MSW, Soil and Heavy Metal in Disposal Site
Industries such as electroplating, ceramics, glass, mining, and battery manufacturing are considered major sources of heavy metals in local water systems, causing heavy metal contamination of groundwater. Total heavy metal levels showed a trend in the relationship between soil metal concentration and long-term irrigation.
Present Scenario of MSW Management and Disposal Facilities in Bangladesh
Very little household waste is recovered or salvaged for recycling, depending on its market value. Khulna City Corporation (KCC), community-based organizations (CBOs) and non-governmental organizations (NGOs) account for only 42% of the total municipal waste generated, while the rest remains unmanaged.
Distribution Pattern of Heavy Metals in Soil
In the United States, soil map units are grouped based on the basic principles of Soil Taxonomy (Soil Survey Staff, 1999). These maps can be used to predict and model properties that are not explicitly classified in the map itself. A case study of Beijing, China by Zou et al. 2015) found that the maximum distribution of heavy metals was in the northwest for Cu, Cd, Pb and Zn; to the southeast for As; mainly at the urban limit for Hg; and mainly in the southwest for Cr and Cd.
Conventional Statistics
In the current study, a normality test was conducted using SPSS to test the accuracy of the statistical analysis. Power is the most common measure of the value of a test for normality, namely the ability to detect whether a sample comes from a non-normal distribution (Conover, 1999). 2012) conducted an analysis on descriptive statistics of heavy metal concentrations in agricultural soils for As, Cd, Cu, Hg, Pb and Zn, respectively.
Multivariate Statistics
Multivariate and geostatistical analyzes of the spatial distribution and origin of heavy metals in the agricultural soil in Shunyi, Beijing, China. To provide comprehensive understanding of heavy metal distributions in the agricultural soil of the Beijing. Multivariate statistical analysis has been widely used for source apportionment of metals in soil and water in different parts of the world.
Geostatistical Analysis
A research by Johnston et al. 2001) identified RBF fits a surface through the measured sample values while minimizing the total curvature of the surface. Semi-variogram models are the spatial structure of the regionalized variable and provide weight information to the kriging algorithm for interpolation (Yan et al., 2015). Multivariate and geostatistical analyzes of the spatial distribution and origin of heavy metals in the agricultural soils in Shunyi.
Introduction
The performance of these interpolation techniques was assessed based on indices such as mean absolute percentage error (MAPE), relative improvement (RI) and goodness of prediction (G-value). It has an area of 4394.45 km² and is bounded on the north by Jessore district and Narail district, on the south by the Bay of Bengal, on the east by Bagerhat district, and on the west by the Satkhira district. This MSW contains a large amount of metal elements that come into direct contact with the environment.
Location and Soil Conditions of Waste Disposal Site
Urban development is dribbled into neighboring areas in the North and West, resulting in a large amount of waste generation. The selected waste disposal site, Rajbandh is the only certified waste disposal site in Khulna, shown in Figure 3.1. Based on the aforesaid authentifications, a comprehensive study of the distribution of metallic elements in soils accreting to the vicinity of the Rajbandh tailings site has become inevitable.
Soil Sampling
The first sampling site, BH-1, is located in the center of the landfill. Adequate care was taken to remove all loose material, debris, coarse aggregates from the bottom of the excavated pit. The soil samples were collected from the bottom of the borehole by digging out the soil manually using hand shovels.
Laboratory Investigations
- Acid Digestion
On the other hand, the first borehole of rainy season (BH-41) is about 30 m from BH-1 which is the center of the site and maintains a gradual addition of about 15 m in the selection of other next boreholes. After the digestion procedure was performed, metal element concentrations in this digested solution were determined using atomic absorption spectrophotometer (AAS) and the amount of each heavy metal was deduced from the calibration graph. The concentration of the metal elements of Al, Fe, Mn, Cr, Cu, Pb, Zn, Ni, Cd, As, Co, Sb, Sc and Hg in mg/kg was measured in the laboratory.
Descriptive Statistics
- Normality Test
- Shapiro-Wilk Test
- Kolmogorov–Smirnov test
- Conventional Statistics
If the normality test fails, we can say with 95% confidence that the data does not fit a normal distribution. Where F is the theoretical cumulative distribution of the tested distribution, which must be a continuous distribution and must be fully specified (i.e., location, extent, and shape parameters cannot be estimated from the data. It may be noted here that If the significance (p) value of the K-S test is greater of 0.05, the data is normal.
Multivariate Statistics
- Principal Component Analysis
The null hypothesis (H0) and alternative hypothesis (H1) of the significance test for correlation were expressed according to a two-tailed test. Determining the distance between the objects of classification by applying some similarity measure, e.g. Euclidean distance or correlation coefficient;. Performing appropriate linking between the objects by some of the clustering algorithms such as single, mean or centroid linking;.
Geostatistical Analysis
The quality of the interpolation was evaluated by examining the RMSPE of cross-validation and validation: the smaller the RMSPEs, the better the interpolation (Brovelli et al., 2011). The quality of the interpolation was judged by examining the RMSPE of cross-validation and validation: the smaller they are, the better the interpolation (Brovelli, 2011). The ordinary kriging method is the most general and widely used of the kriging methods, which incorporates the statistical properties of the measured data (spatial autocorrelation).
Assessment of Method Performance
Predictions are unbiased, with a mean prediction error close to 0, and standard errors are accurate, indicated by a root-mean-square standardized prediction error close to 1 (Brovelli et al., 2011). A study by Xie et al. 2011) found that the prediction does not deviate much from the measured values, indicated by the root-mean-square error and average standard error that are as small as possible. To analyze the effect of model parameters on pollution estimation, ordinary kriging, IDW with power 1 to 5, LP with order 1 to 3 and RBF five kernel functions of CRS, IMQ, MQ, ST and TPS were selected.
Artificial Neural Networks
When the neural network tends to overlap, this error increases and the weights are determined based on the minimum error (Rooki et al., 2011). It was collected depending on the complexity of the problem and the amount and nature of the learning data. The error histogram presented for the training data would provide additional verification of the network performance, which shows the distribution of residuals between the targets and the network output (Shahin et al., 2004).
General
- Normality Test
Descriptive statistics analysis was performed to investigate the quantitative distribution of metal elements presence in the soil of the landfill in a manageable form. First, the normality of metallic elements was performed using the Kolmogoroph-Smironov (K-S) test, Shapiro-Wilk (S-W) test and Normal QQ Plot. In addition, the metal elements of Al, Ca, Cd, Fe, K, Ni, Pb, Sb, Sc, Sr and Ti in soil for dry season were normally distributed indicating that the null hypothesis was accepted at the significance value (p. ) greater than 0.05.
Q plot (Cd)
From the K-S test, for an alpha level of 0.05, a set of data with a significance value (p) of less than 0.05 would reject the null hypothesis that the data sets were some form of normally distributed. The result obtained from both the non-parametric normality test of K-S and S-W for all investigated metal elements in soil for both dry and rainy seasons is shown in Table 4.1.
Q plot (Ni)
Q plot (Pb)
Q plot (Zn)
- Conventional Statistics
- Dry Season
- Rainy Season
- Seasonal Comparison of the Concentration of Metal Elements
- Multivariate Statistics
- Principal Component Analysis
- Cluster Analysis: Agglomerative Hierarchical Clustering
- Cluster Analysis: Artificial Neural Network
- Geostatistical Analysis
- Deterministic Methods
- Assessment of Method Performance
- Mean Absolute Percentage Error
- Goodness of Prediction
- Relative Improvement
- Artificial Neural Network
- Mean Standard Error
- Regression Coefficient
- Error Histogram
- General
- Comparison with Previous Studies
Descriptive statistics of metal elements in the soil for the dry period of the waste disposal site are given in Table 4.2. Descriptive statistics of metal elements in the soil for the rainy season of the landfill are given in Table 4.3. In this study, the association patterns of metallic elements in landfill soils were presented.
Different interpolation methods of ArcGIS showed spatial distribution of metal elements in the soil of the waste disposal site with prediction errors. In addition, IDW and RBF provided accurate prediction of spatial distribution of metal elements in soil.
Beijing suburbs
Conclusion
From normal QQ plot, it can be concluded that almost all the metal elements in soil for both the dry and rainy seasons are normally distributed, except As. The generation sources of all studied metal elements in soil obtained from PCA were completely consistent with AHC for both the dry and rainy seasons. The predicted values of metal elements in soil from ANN were almost the same as those obtained from laboratory.
Recommendation for Further Study
Current status of municipal solid waste management in Bangladesh, Waste-The Social Context, Canada, Vol.16:1-10. Evaluation of the accuracy of geostatistical methods for the zonation of heavy metals in the soils of urban-industrial areas. Analysis of Spatial Variations and Sources of Heavy Metals in Agricultural Soils of Beijing Suburbs, Vol.
