the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Apr 2026
Soil contamination and soil-mediated human health risks associated with household coal combustion in residential areas of Zavkhan Province, Mongolia
Abstract. Soil contamination by heavy metals represents a growing environmental and public health concern in cold–dry rural settlements where coal-based household heating remains dominant. This study investigates how coal combustion alters soil element dynamics and associated human health risks by applying a process-oriented, integrated soil system assessment in a residential area of Uliastai city, western Mongolia.
Surface soils (0–10 cm) from 38 sites were analyzed using ICP-OES and ICP-MS to determine major and trace element concentrations. Multivariate statistical analysis (principal component analysis, PCA) was combined with contamination indices (enrichment factor and geo-accumulation index) and human health risk assessment to explicitly link contamination sources, transport pathways, soil retention processes, and potential human exposure.
Results reveal a clear separation between anthropogenically influenced metals (As, Pb, Cd, Zn, and Cu) and elements predominantly controlled by geogenic background conditions (Cr, Co, and Ni). Very high to extreme enrichment and geo-accumulation levels for As, Pb, Zn, Cd, and Cu indicate substantial anthropogenic alteration of surface soil metal pools. Comparison of soil, coal, and ash compositions identifies coal combustion ash as the primary source of metal enrichment, acting as a concentrated reservoir that is redistributed to soils via atmospheric deposition and surface processes. Human health risk assessment shows that the most enriched metals, particularly As and Pb, dominate both non-carcinogenic and carcinogenic risks, with inhalation and ingestion pathways contributing most strongly to potential exposure.
The findings demonstrate that soil contamination in Uliastai reflects systemic changes in soil functioning driven by household energy practices rather than isolated concentration exceedances. By integrating source identification, contamination intensity, and health risk within a unified soil system framework, this study provides mechanistic insight into soil–human interactions and offers a transferable approach for assessing soil impacts in coal-dependent rural environments.
- Preprint
(682 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-703', Anonymous Referee #1, 15 Apr 2026
-
RC2: 'Comment on egusphere-2026-703', Anonymous Referee #2, 16 Jun 2026
The study by Battsengel et al. addresses a scientifically relevant question by investigating the effects of coal combustion on soil contamination and potential health risks. This topic is particularly pertinent in regions such as Mongolia, where coal remains a necessity for heating and cooking due to the lack of accessible alternatives. However, I have identified several methodological limitations in the study that warrant attention.
1. Issues with PCA
The authors present Principal Component Analysis (PCA) as a method to distinguish the origin of soil contamination by clustering metals in soil, coal ash, and coal samples (lines 195–196). While PCA can theoretically achieve this, the current implementation raises some concerns:
The conclusions drawn from the PCA (e.g., lines 295–297) are not supported by the results. For PCA to effectively assess coal-contamination source, coal or ashes samples should be included in the PCA (as presented l.195-196) and be used to analyse co-occurrence of metals. However, Figure 2 and its accompanying text do not attribute sample origins (soil, coal, or ash) to the data points. At this stage of the paper, the identification of two metal groups ({As, Pb, Cd, Zn, Cu} and {Cr, Co, Ni}) does not demonstrate that these groupings are relevant to determining contamination origin. The interpretation in lines 295–297 thus relies on literature (not cited) rather than the study’s own results. Hence, the objectives 3 ("Differentiate geogenic and anthropogenic…") and 4 ("Reveal dominant sources of contamination…") are not fulfilled by the current PCA analysis. Similarly, l. 291-297, the authors argued that “sample score with high PC1 Scores correspond to locations with elevated concentrations of anthropogenically enriched metals”. However, at the stage and without comparing with coal and ashes, the conclusion should be “sample score with high PC1 Scores correspond to locations with elevated concentrations of metals X, Y, Z” without findings of their origins.
Besides, I wonder if PCA is not driven by few samples with loadings >4 and up to 8 (top right and bottom right of the quadrants) against (-1) to 2 for the others .
There is no way to associate PCA data points with their geographic locations (e.g., Figure 1). Adding sample numbers on the PCA would be useful to identify samples with fig. 2 and 3-4.
2. Comments on risk analysis
Secondly, the main conclusion of their risk analysis is that children are more sensitive than adults, and are especially exposed through inhalation (l. 350-351). However, this is not a result as the difference in sensitivity is directly linked to the coefficients used in eq. 3 to 5 and parameters of table 5 that are fixed, and determined to capture the highest sensitivity of children. In my opinion, if authors would have done a risk analysis, they should assess how often the risk limits is reach, or what should be the maximum concentration of single or mixed metals in soils
3. General Remarks on Soil Contamination Analysis Results
The results and conclusions from the Igeo and EF analyses are unclear. Please clarify the importance, use, and specific conclusions of each analysis. Additionally, Figures 4 and 5 are difficult to read.
I recommend that the authors construct a PCA including coal and ash samples to identify metal clustering and cross-reference the EF/Igeo results (e.g., identifying metals associated with soils heavily contaminated by coal) with the PCA results to better assess Objective 3.4. Minor Concerns
- Paragraph 2.3:
The enrichment factor (EF) is defined twice (lines 164 and 158).
The geo-accumulation index (Igeo) is inconsistently formatted as “Igeo” or “I_geo”. Please standardize.
The computation of EF, particularly the definition or calculation of (M/Fe)background (line 177), is imprecise.- Paragraph 2.4:
Justify why PCA is an effective method for determining relationships among heavy metals (lines 186–190).
The objectives and methods to achieve them are unclear. For example, it is not specified whether soils or metals will be clustered (lines 198–199) and “etc.” is not a valid objective, please be more specific.
It is unclear which elements are analyzed: line 233 mentions 26 elements, while line 245 mentions 11.
Provide LOQ/LOD values for all analyzed elements.Paragraph 3.2:
In lines 279–280 and 282–283, the anthropogenic vs. geogenic origin of elements is assumed without Coal/ash measurements, EF/Igeo calculations, or literature references.Table 4:
Units should be written in lowercase.
The inclusion of median standard deviation (or median absolute deviation?) is unclear without the median value.
Why are some values in bold? Please explain in the legend or remove the bold formatting.Other Concerns:
Line 250: What does “ger” refer to?
PCA results should include cos² values (quality of variable representation on each axis).
The legend in Figure 2 is imprecise.
Long names for metals are used in lines 375 and 380 instead of elemental abbreviations. Please use the long name the first time an element appears to introduce its abbreviation, then consistently use the abbreviation thereafter.Citation: https://doi.org/10.5194/egusphere-2026-703-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 211 | 47 | 17 | 275 | 15 | 18 |
- HTML: 211
- PDF: 47
- XML: 17
- Total: 275
- BibTeX: 15
- EndNote: 18
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Thank you for the opportunity to review this paper. As my expertise is primarily in human health risk assessment, my comments will primarily focus on those aspects of the paper. However, as a general comment, I feel the paper would be much improved by the inclusion of analytical results for representative samples of coal ash and other potential contaminating material from Uliastai. The paper infers that metals observed at higher relative levels were due to anthropogenic sources - this conclusion would be reinforced if those sources had also been analysed.
The human health risk assessment component of the manuscript has several serious issues and I consider that the authors should consider redrafting the manuscript without this section and focussing solely on the soil contamination issues. Some (but not all) of the issues with the risk assessment are:
- The authors have used outdated versions of several key references. Specifically, the 1997 version of the USEPA exposure factors handbook has been used, rather than the 2011 edition, and the 2007 USEPA draft risk assessment of coal combustion wastes has been used, rather than the final risk assessment published in 2014.
- The authors elected to determine exposure through three exposure routes: ingestion, dermal and inhalation. The equations for these exposure routes are presented in equations (3), (4) and (5). Equation (4) dermal exposure contains a term FE, which the authors describe as the dermal exposure ratio, with a value of 0.61. However, in the USEPA equation this term is PC, the chemical-specific dermal permeability constant, which for metals has a value of 0.001. This different will obviously have a large impact on estimates of dermal exposure.
- Equation (5) for inhalation exposure, contains a term PEF, standing for particle emission factor. This factor does not appear in the USEPA RAGS part F guidance on inhalation exposure and it is not further explained by the authors.
- The authors list reference doses in Table 3, mainly references to USEPA 2007. As noted previously, USEPA 2007 has been replaced by USEPA 2014. Neither reference contains dermal reference doses and it uncertain where those in Table 3 have been derived from. The inhalation reference doses are actually reference concentrations (RFCs) and consequently have different units to those identified in the table. Finally, some of the values presented in Table 3 differ from those in the source reference.
- Equations (6) and (7) show the derivation of HIs and HQs. HIs are conventional used to combine estimates of risk across different exposure routes. Table 5 (line 330) states that the contents of the table are HIs but provides separate estimates for each exposure route, presumably HQs.
- In comparing Tables 5 and 6, ingestion is shown as the major route of exposure for non-cancer risks and inhalation as the major route for cancer risk. This is not possible, as the same exposure estimates should be used for both cancer and non-cancer risk assessments.
- Lines 350-351. The authors determine cancer risk separately for children and adults. This approach is incorrect, as the cancer potency factors refer to exposure over a lifetime. A lifestage weighted exposure could have been used or the adult exposure, as a satisfactory surrogate for lifetime exposure.
- The authors make no mention of the exposure of the population to metals through other sources of exposure. In most cases, dietary exposure will be the major source of exposure, with drinking-water also contributing in some cases. Without an indication of the magnitude of exposure from these sources, the risk estimates from soil only will underestimate the human health risk. These sources of exposure are likely to also be impact by the contamination sources mentioned in the paper.