Journal of Geology & Geophysics
Open Access

Comprehensive Assessment of Heavy Metals in Soils

Satyendra Nath Chakrabartty^{* *}Correspondence: Satyendra Nath Chakrabartty, Indian Ports Association, Indian Statistical Institute, Indian Maritime University, Kolkata, West Bengal, India, Email:

Author info »

Introduction

Pollution of soil by Heavy Metals (HMs) from natural and anthropogenic sources like domestic/industrial waste water effluents, urban and agricultural runoff, fossil fuel combustion, atmospheric deposition, etc. have emerged as a major threat to human society. Major HMs affecting soils are Cadmium (Cd), Chromium (Cr), Copper (Cu), Lead (Pb), Mercury (Hg), etc. Toxicity of metals increase with concentration. Illustrative list of toxic effects of HMs are: Dysfunction of liver and renal problems by Pb, Cd, and Hg; cognitive functions may be affected by Pb, Hg which may introduce neurological changes; exposure to Cd can cause cancer, hypertension, even affect our reproductive system; Hg may result in fetal abnormalities, etc. [1-3]. Agricultural fields have shown significant contamination of HM. For example, predominant HM pollution of agricultural fields in India due to use of sewage water and industrial effluents were found [4,5]. Widespread HM pollution primarily due to rapid industrialization and urbanization has already affected public health, food safety and even the ecological environment.

Assessments of soil pollution due to HMs are either index based or model driven. While the former approaches involve use of a single pollution, extent of pollution, etc. the later approaches include enrichment factors, ecological hazard index, etc. [6], investigated pollution level due to concentrations of HMs using contaminant index (single factor), Nemerow index, spatial distributions of HMs, and assessed environmental health risk utilizing a United States Environmental Protection Agency recommended model [7]. Studies focusing on one specific HM cannot satisfy interests of landowners. Regulatory authorities and researchers who could be interested in approach of aggregation of several measures from individual sampling units to obtain sample information, based on which estimation of population parameters for further analysis and inferences including effectiveness of remedial action [8]. However, no agreed standard exists for evaluation of soil environmental quality [9].

While enrichment factors indicate (extent of increase of an element in sampling unit relative to average natural abundance; requiring selection of a background composition and also a reference element), Ecological risk index reflects the Potential Ecological Risk Index (PERI, or RI), and combine the single index of Ecological Risk Factor ( ERI) and the Toxic Response factor of Individual metal (TRI).

Nemerow Pollution Index (PI_Nemerow) indicates soil quality in terms of the single Pollution Index (PI) and is computed as

Where, PI_max: Maximum value of the single pollution index of all heavy metals

n: number of heavy metals.

Several indices are there for measuring effect of single factor, geo- accumulation, ecological risk and resulting human health risk, etc. Comprehensive assessment requires combining such indices on a methodologically sound approach to get a comprehensive index for analysis of overall pollution status of HMs. Clearly, comprehensive index is a Composite Index (CI) which is a real valued function f whose domain consists of set of individual indices, which could be unit-free or with different units, following different distributions and having different degrees of correlations.

Distributions of concentration of HMs vary for different locations, different sections of soil and are not normal for different sections of soil [10]. Large value of Coefficient of Variation (CV) of a HM shows uneven distribution of the HM concentration.

The paper describes limitations of existing indices and methods of obtaining comprehensive index as combination/aggregation of individual indices and suggests an assumption-free method to get comprehensive index reflecting overall pollution status of HMs in soil as a composite index, satisfying desired properties.

Materials and Methods

Measures of pollution level of HMs

Single Factor assessment (SF): Concentration of HMs (mg/Kg) in soil are calculated and compared with respective standards. However, such standards vary across countries. Single-factor contaminant index P_i was considered [6] as

Where, C_i (mg/Kg) is the detected concentration of i-th HM and S_i (mg⁄Kg) is the evaluation criterion of i-th HM.

Nemerow pollution index P_N is given by

Where, P_i.max and P_i.ave denote respectively the maximum and arithmetic mean of P_is. The index P_N reflects average pollution level of soil in terms of average of various pollutants and also helps to find more serious pollutants [11]. P_N has a one-to-one correspondence with Synthetic Pollution Index (SPI) computed as

Here, each of P_i and P_N is unit free

Limitations

• Addition of P_i and P_j for I ≠ j is not admissible since evaluation criterion of i-th HM and j-th HM are different, like average speed of two cars. For example, consider A=4/8=0.5 and B=3/12=0.25. Here, (A+B)/2=0.375 which is different from the pooled average of A and B=7/20=0.35.

• P_N is influenced by the maximum score. SPI considers two extreme values of data which could be unrealistic outliers and ignores other observations from the data. Thus, value of P_N or SPI depends heavily on heterogeneity of data. Better could be to find pooled average and pooled Standard Deviation (SD) based on all sample observations.

• Classification of soils based on increasing values of P_i under safe, slightly polluted, low polluted, moderately polluted and highly polluted needs to show efficiency of classification in terms of low value of within-group variance and high value of between-group variance. Same is true for classification on the basis of P_N values [13].

Geological Accumulation Assessment (GEO)

Muller [14], gave the formula for calculating Geo-accumulation index (I_geo) to assess pollution level for a single element by

Where C_Sample and C_Background denote respectively concentration of a HM in soil and the same in the background, taken from secondary data, which can change due to natural diagenesis. Background level concentration distinguishes concentration of the HM occurring by natural process and due to anthropogenic influence [15]. For each HM, separate geochemical baseline is determined for geologically different regions [16].

Considering I_geoMax (maximum I_geo value) and I_geoAv (average of I_geo value) of all HMs in a sample, [17] calculated arithmetic mean, SD, Geometric mean and median for I_geo and Nemerow index separately for eight different HMs without specifying most suitable averaging procedure.

Observations

I_geo is relative to the C_Background and depends heavily on its value. For a given value of . For example, let C_Background for Pb be 36.0 mg·kg⁻¹, C_Sample(1)=37 units and C_Sample(2)=38 units. Then, and

Increase of I_geo per unit increase in C_Sample is different at different range of values. Let values of concentration of Pb in C_Sample(3)= 74 and C_Sample(4)= 75 respectively where C_Background for Pb is 36.0 like the above example.

I_geo could be expressed as when geochemical background value and evaluation criterion of i-th HM are same. Thus, I_geo and P_i are likely to be highly correlated.

C_Sample could be better calculated as pooled mean.

C_Background may not be available for areas considered as free from any HM pollutants where the geochemical baseline map is under progress. Non-availability of C_Background at Brahmaputra valley, North East India was found [18].

Classification based in I_geo may not be beyond questions.

Ecological Risk (ER)

ER reflects pollution level due to accumulation of HMs in soil by the interaction effects of several HMs and environmental associations with HMs. Potential ecological Risk Index (RI) integrates ER value of each HM as follows:

Where, T_r denotes the toxicity coefficient and CF denotes the pollution factor.

Multiplicative aggregation was favored by [19], to obtain Pollution Load Index (PLI) for a station by

For the i-th HM in soil, [20] suggested computation of potential ecological Risk Index (RI) by

Where,

Cⁱ: Total amount of i-th HM,

Cⁱ_n: Corresponding background value of the i-th HM

Tⁱ _r : Coefficient of toxicity of i-th HM, taken as 5,1,5,30, and 5, respectively for Cu, Zn, Pb, Cd, and Ni [21].

Eⁱ _r: Single factor ecological risk index of i-th HM and

Higher value of Eⁱ _r indicates higher risk degree of RI.

Classification of regions with respect to different indices gave different results. Unpolluted regions (with PLI<1) had heavy contaminations in terms of I_geo values [22].

Human Health Risk assessment (HHR)

HHR is a tool to assess and manage adverse health effects on humans beings exposed to HMs in polluted environmental media through ingestion, dermal absorption, and breathing inhalation. It also determines the measures required to comply with the relevant health and safety legislation and, thus, reduce level of occupational injuries and ill-health Potential Non-Carcinogenic health Risk (PNCR) and Potential Carcinogenic health Risk (PCR) for more than one HM can be estimated by sum of HQ for each metal and expressed as unit-free Hazard Index (HI) [23].

HI>1 ⟹ existence of PNCR. Total PNCR>1 × 10^-4, implies high carcinogenic risk to human beings. If total PNCR<1 × 10^-6, PCR due to exposure to HM could be negligible.

Comprehensive assessment

Analytic Hierarchy Process (AHP) was used by [10], to combine four indices (SF, GEO, PER and HHR) involving subjective assessment of relative importance of the indices using Saaty’s scale consisting of 9-point items (1-9) followed by selection of weights and comprehensive assessment scores (T) by weighted sum. Here, higher value of T implies lower pollution level. Major disadvantages of AHP are: Determination of importance through subjective approach using pairwise comparisons; Saaty scale depends on the trial perceptions and is not meaningful [24]; AHP is not a legitimate technique [25]; use of AHP along with others Multi Criteria Decision Making (MCDM) methods was suggested [26]. But, results of Data Envelopment Analysis (DEA) depend significantly on selection of inputs and outputs and number of inputs and output variables [27].

Principal Component Analysis (PCA) was used [12], to determine the sources of soil HMs, without verification of normal distribution of variables under investigations, which is an important assumption of PCA. Non satisfaction of assumptions of techniques used may distort the results. PCA considers the low variance components in the data as noise and thus, those components are thrown away, even if those components play major roles in decision making.

PCA is sensitive to scaling, presence of outliers and may not be suitable for small samples. PCA weights are not constant over time and space and thus, comparisons are difficult. Sava [28], observed PCA weights for data accumulated for 11-years and averaging over each factor for year-wise data were different. Use of PCA was not favoured by Environmental Sustainability Index [29].

Cluster analysis for identification of 10 HM accumulation patterns in 47 soil samples was undertaken [30], and three clusters were found, each with a characteristic composition groups. While Cluster-1 showed high concentrations of Zn and Pb, Cluster-2 had high Cu concentration and major concentrations at Cluster-3 were Cr and Cu implying overlapping of Cluster-2 and Cluster-3. Hierarchical Cluster Analysis (HCA) on normalized datasets was used [31], to avoid problems due to scaling of variables and interconnections among the contaminants. However, non-inferential Cluster analysis assumes that the variables are independent. If variables are not independent, the assumption gets violated requiring further necessary action to deal with the situation.

Comprehensive assessment by weighted sum approach falls on the class of linear model (without testing linearity) with full compensability where poor effect of some pollutants can be well compensated by high effects of other pollutants. Different methods of selection of weights satisfying W_i>0 and ΣW_i=1 give different results and no weighting system is above criticism [32]. The multiplicative aggregation with non-linearity properties by has been proposed [33], which prevents rank reversal phenomenon and practically no substitution problems [34].

Proposed method

Ignoring the stage of selection of indicators, following method by [35] is proposed for Comprehensive Assessment of HMs in Soils as a Composite Index (CI) which avoids selection of weights and scaling of raw data. The method requires ensuring that each indicator/index is positively related to CI.

Set up and methodology

Take pooled average of the sample observations. For a region, consider the vector showing performance of n-indices in the current period and vector showing evaluation criteria/base period values. Instead of arithmetic mean of ratios, overall pollution status of HMs (IHM) as CI for the t-th time period may be defined as the Geometric mean i.e.

Multiplicative aggregation given in (9) and (10) can be applied for all types of indices including those in percentages or having skewed distribution and can be applied even for data with small sample size. In line with general convention of index value=100 in the base year, one can take to reflect changes in percentage.

Properties

For a given region, each of (9) and (10) depicts overall improvement/decline of soil pollution by HMs in the current year with respect to the target values or base year or previous year by a unit free continuous function satisfying the following desired properties.

• Independent of order of the chosen indices and independent of change of scale

• Reduces level of substitutability among the component indices.

• Facilitates formation of chain indices like

• Satisfies time reversal test

• Facilitates drawing of graph over time showing path of improvement/decline for a region across time for comparison of regions with respect to such paths. Replacing the base period vector by the vector for the previous year gives improvement in on year-to-year basis.

• Produces no bias for polluted or non-polluted regions since is least affected by outliers.

• Distribution of GM approaches the lognormal form and thus, computation of mean and SD are admissible [36].

Taking log on both sides of equation (9),

Where, denotes Geometric Standard Deviation (GSD)

Since, log (GSD of X₁, X₂, … …, X_n)=usual SD of logX₁, logX₂, ………, logX_n.

The ratio is analogous to Coefficient of Variation (CV) and indicates consistency of the data.

Help to estimate population parameters. For large data, sample GM is taken as population GM and standard error of GM is estimated by

Thus, statistical tests of hypotheses regarding equality of GM’s can be performed across time and space using t-tests on the logarithms of the observations.

Possible to construct separate index for each subset of indices.

Relative importance and contribution of each index can be easily found.

Critical indices are those for which

Improvement in an index and the corresponding gain in are linearly related since 1% increase in increase in if all others are unchanged

Applications

The proposed method can be applied even for skewed data and irrespective of correlations between a pair of indices and help in meaningful comparison of a set of regions of a country.

For comparison of properly defined sub-areas say agricultural lands, river basins, lands in vicinity of industrial activities like mining, production, highways, solid waste dump sites, etc., the proposed I_HM index can also be applied with pre-selected pollutants and indices.

Can be used for classification of regions with respect to overall pollution status of HMs.

The graph of progress/decline of , using the Chain indices for a region will help to assess impact of various mitigation measures across time

Limitations

• Introduction of a new index requires value of that index in the base period and at different time periods.

• The method fails if an indicator takes negative or zero value.

• Assumes no missing data

Results and Discussion

Limitations of existing methods of overall pollution status of HMs were reviewed. The proposed in terms of GM, avoiding scaling and finding weights considers all chosen indices to depict comprehensive improvement or decline of soil pollution by HMs in a region in the current year with respect to base year or previous year and helps in meaningful comparison of a group of regions/countries. The proposed considers GM of the ratios without normalization of individual indices like geometric aggregation adopted in Human Development Index, 2010 [37].

Important features of the proposed measure include:

• Reduction of substitutability among the component indices.

• Produces no bias for polluted or non-polluted regions since effect of outliers is not much.

• Unit-free monotonic continuous function.

• Facilitates formation of chain indices and satisfies time reversal test.

• Possible to find distribution of and compute mean and variance for a sample of stations.

• Allows ranking and classification of regions in terms of values and assesses paths of improvement registered by a region over time.

• Meaningful comparison of countries can be made in terms of progress made on Year-to-Year basis, despite different base period figures across countries,

GM approach is preferred for its additional properties like formation of chain indices, satisfaction of time reversal test, easy computation of contribution of the indices to , linearity between improvement in a ratio of an index and the corresponding gain in . Distance of a country from the SDG goals at a particular time period can be found by replacing the base period data with the SDG goals adopted by the country. Besides, the measure can be used for statistical testing of for the i-th and j-th region at a particular time or for a particular region.

may be computed separately for biogenic and physicogenic aggregates to see which one is more likely to lead to soil improvement.

Conclusion

The proposed aggregation by GM is an improvement over the existing methods of finding overall pollution status of HMs ( HM I ) in soils and is recommended. Researchers and stakeholders may derive benefits of the proposed HM I to find overall results of soil use and management along with identification of critical areas requiring managerial attentions. Future studies to assess impact of HM I in soils on vegetables, fruits, fish, shellfish, crustaceans, etc. may be undertaken.

DECLARATIONS

Competing interests

No funds, grants, or other support was received.

Funding

No funding was obtained for this study.

Financial interests

The author has no relevant financial or non-financial interests to disclose.

Data availability

Not applicable. No dataset used in the study.

Ethics approval

Not required for this methodological paper.

Ethical responsibilities of authors

All authors have read, understood, and have complied as applicable with the statement on ethical responsibilities of authors.

Informed consent

No data collected from individuals.

Author’s contribution statement

Conceptualization, methodology, writing-original draft preparation, writing-reviewing and editing, by the single author.

References

1. Mattia GD, Bravi MC, Laurenti O, Luca OD, Palmeri A, Sabatucci A, et al. Impairment of cell and plasma redox state in subjects professionally exposed to chromium. Am J Ind Med. 2004;46(2):120-125.

   [Crossref] [Google Scholar] [PubMed]

2. Lee KG, Kweon H, Yeo JH, Woo S, Han S, Kim JH. Characterization of tyrosine-rich Antheraea pernyi silk fibroin hydrolysate. Int J Biol Macromol. 2011;48(1):223-226.

   [Crossref] [Google Scholar] [PubMed]

3. Tchounwou PB, Yedjou CG, Patlolla AK, Sutton DJ. Heavy metal toxicity and the environment. In Molecular, clinical and environmental toxicology: Volume 3: Environmental toxicology. Exp Suppl.  2012:133-164.

   [Crossref] [Google Scholar] [PubMed]

4. Gupta N, Khan DK, Santra SC. Heavy metal accumulation in vegetables grown in a long-term wastewater-irrigated agricultural land of tropical India. Environ Monit Assess. 2012;184:6673-6682.

   [Crossref] [Google Scholar] [PubMed]

5. Singh S, Raju NJ, Nazneen S. Environmental risk of heavy metal pollution and contamination sources using multivariate analysis in the soils of Varanasi environs, India. Environ Monit Assess. 2015;187:1-2.

   [Crossref] [Google Scholar] [PubMed]

6. Yang D, Zhu H, Liu J, Zhang Y, Wu S, Xiong J, et al. Risk assessment of heavy metals in soils from four different industrial plants in a medium-sized city in North China. Toxics. 2023;11(3):217.

   [Crossref] [Google Scholar] [PubMed]

7. United States Environmental Protection Agency (USEPA). Exposure Factors Handbook 2011 Edition (Final); Office of Emergency and Remedial Response. 2011.
8. Yoo JC, Beiyuan J, Wang L, Tsang DC, Baek K, Bolan NS, et al. A combination of ferric nitrate/EDDS-enhanced washing and sludge-derived biochar stabilization of metal-contaminated soils. Sci Total Environ. 2018;616:572-582.

   [Crossref] [Google Scholar] [PubMed]

9. Xie W, Yang YD, Hou JY, Kan GQ, Li GC, Zhao XH. Application of multiple evaluation methods to soil environmental quality evaluation of main walnut producing areas in Tianjin. Geophys Geochem Explor. 2021;45(1):207-214.

   [Google Scholar]

10. Zhao H, Wu Y, Lan X, Yang Y, Wu X, Du L. Comprehensive assessment of harmful heavy metals in contaminated soil in order to score pollution level. Sci Rep. 2022;12(1):3552.

    [Crossref] [Google Scholar] [PubMed]

11. Kowalska J, Mazurek R, Gąsiorek M, Setlak M, Zaleski T, Waroszewski J. Soil pollution indices conditioned by medieval metallurgical activity–A case study from Krakow (Poland). Environ Pollut. 2016;218:1023-1036.

    [Crossref] [Google Scholar] [PubMed]

12. Gong C, Wang S, Wang D, Lu H, Dong H, Liu J, et al. Ecological and human health risk assessment of heavy metal (loid) s in agricultural soil in hotbed chives hometown of Tangchang, Southwest China. Sci Rep. 2022;12(1):8563.

    [Crossref] [Google Scholar] [PubMed]

13. Yu L, Cheng J, Zhan J, Jiang A. Environmental quality and sources of heavy metals in the topsoil based on multivariate statistical analyses: A case study in Laiwu City, Shandong Province, China. Nat Hazards. 2016;81:1435-1445.

    [Crossref] [Google Scholar]

14. Muller GM. Index of geoaccumulation in sediments of the Rhine River. Geophys. J R Astron Soc. 1969;2:108-118.

    [Google Scholar]

15. García-Sánchez A, Alonso-Rojo P, Santos-Francés F. Distribution and mobility of arsenic in soils of a mining area (Western Spain). Sci Total Environ. 2010;408(19):4194-4201.

    [Crossref] [Google Scholar] [PubMed]

16. Rojo PA, Frances FS, Garcia-Sanchez A, Alvarez-Ayuso E. Baseline concentrations of heavy metals in native soils of the Salamanca and Valladolid provinces, Spain. Arid Land Res Manag. 2004;18(3):241-250.

    [Crossref] [Google Scholar]

17. Santos-Francés F, Martínez-Graña A, Alonso Rojo P, García Sánchez A. Geochemical background and baseline values determination and spatial distribution of heavy metal pollution in soils of the Andes mountain range (Cajamarca-Huancavelica, Peru). Int J Environ Res Public Health. 2017;14(8):859.

    [Crossref] [Google Scholar] [PubMed]

18. Dutta N, Dutta S, Bhupenchandra I, Karmakar RM, Das KN, Singh LK, et al. Assessment of heavy metal status and identification of source in soils under intensive vegetable growing areas of Brahmaputra valley, North East India. Environ Monit Assess. 2021;193:1-18.

    [Crossref] [Google Scholar] [PubMed]

19. Chakravarty M, Patgiri AD. Metal pollution assessment in sediments of the Dikrong River, NE India. J Hum Ecol. 2009;27(1):63-67.

    [Crossref] [Google Scholar]

20. Hakanson L. An ecological risk index for aquatic pollution control. A sedimentological approach. Water Res. 1980;14(8):975-1001.

    [Crossref] [Google Scholar]

21. Li YM, Ma JH, Liu DX, Sun YL, Chen YF. Assessment of heavy metal pollution and potential ecological risks of urban soils in Kaifeng City, China. Huan Jing Ke Xue. 2015;36(3):1037-1044.

    [Crossref] [Google Scholar] [PubMed]

22. Andem AB, Okorafor KA, Oku EE, Ugwumba AA. Evaluation and characterization of trace metals contamination in the surface sediment using Pollution Load Index (PLI) and geo-accumulation index (Igeo) of ona River, Western Nigeria. Int J Sci Technol Res. 2015;4(1):29-34.

    [Google Scholar]

23. USEPA. Baseline human health risk assessment. Vasquez Boulevard and I-70 Superfund Site Denver Co. 2001.
24. Holder RD. Some comments on the analytic hierarchy process. J Oper Res Soc. 1990;41(11):1073-1076.

    [Crossref] [Google Scholar]

25. Barzilai J. Notes on the analytic hierarchy process. In Proceedings of the NSF design and manufacturing research conference. 2001:1-6.

    [Google Scholar]

26. Ho W. Integrated analytic hierarchy process and its applications–A literature review. Eur J Oper Res. 2008;186(1):211-228.

    [Crossref] [Google Scholar]

27. Berg S. Water utility benchmarking: Measurement, methodology and performance incentives. International Water Association, IWA Publishing. 2010.

    [Crossref] [Google Scholar]

28. Sava AM. Weighting method for developing composite indices: Application for measuring sectoral specialization. J Appl Quant Method. 2016;11(3):77-84.

    [Google Scholar]

29. World Economic Forum. An initiative of the global leaders of tomorrow environment task force. Pilot Environment Performance Index. 2002.
30. Ikenaka Y, Nakayama SM, Muzandu K, Choongo K, Teraoka H, Mizuno N, et al. Heavy metal contamination of soil and sediment in Zambia. Afr J Environ Sci Technol. 2010;4(11):729-739.

    [Google Scholar]

31. Shirani M, Afzali KN, Jahan S, Strezov V, Soleimani-Sardo M. Pollution and contamination assessment of heavy metals in the sediments of Jazmurian playa in southeast Iran. Sci Rep. 2020;10(1):4775.

    [Crossref] [Google Scholar] [PubMed]

32. Greco S, Ishizaka A, Tasiou M, Torrisi G. On the methodological framework of composite indices: A review of the issues of weighting, aggregation, and robustness. Soc Indic Res. 2019;141:61-94.

    [Crossref] [Google Scholar]

33. Ishizaka A, Labib A. Analytic hierarchy process and expert choice: Benefits and limitations. OR Insight. 2009;22(4):201-220.

    [Crossref] [Google Scholar]

34. Stam A, Silva AP. On multiplicative priority rating methods for the AHP. Eur J Oper Res. 2003;145(1):92-108.

    [Crossref] [Google Scholar]

35. Chakrabartty SN. Better composite environmental performance index. Interdiscip Environ Rev. 2018;19(2):139-152.

    [Crossref] [Google Scholar]

36. Alf EF, Grossberg JM. The geometric mean: Confidence limits and significance tests. Percept Psychophys. 1979;26(5):419-421.

    [Crossref] [Google Scholar]

37. Kovacevic M. Review of HDI critiques and potential improvements. Human development research paper. 2010;33:1-44.

    [Google Scholar]