Are Soils Forming on Buildings Indicators of Pollution in The City? A Case Study from Lublin (Eastern Poland)

The study was conducted between 2020 and 2023 in Lublin. It is the largest city in eastern Poland (147.5 km²), with a population of approximately 332,000 (status for 2022). The average population density for the city is 2252 people per square kilometre, which is several times higher than the Polish average of 122 people per square kilometre. Built-up areas: residential, commercial, industrial, as well as roads and tracks cover 34.8% of the city’s area. Urban green areas account for 30.8%, of which forests account for 14.2% and water accounts for 2.4%, making Lublin one of the greenest cities in Poland. The remaining 32.0% is developed for arable land. The varied relief of the western part of Lublin is related to the occurrence of loess cover (there are numerous gullies and dry valleys), while the more monotonous eastern part is made up of gneisses, marls and gaizes, covered by a thin layer of eluvial and deluvial sediments (Superson et al. 2018, Dobrowolski, Chabudziński 2021). Lublin is located in a temperate climate zone, with an average multi-year air temperature of 8.1°C and an average multi-year precipitation of 551 mm (Kaszewski 2008). The western part of the city is dominated by loess loam soils in a complex with brown soils, while on the eastern side there are podzols and rusty soils formed mainly from sandy formations, brown soils and, in places, black earths. In the river valleys, chernozemic muds, proper muds and peat silt soils have developed (Plak 2007, 2018).

The first stage of the work was to carry out a field reconnaissance with the aim of finding abandoned or neglected buildings on which the soils under study could potentially occur. A useful indicator of such sites was the presence of vegetation overgrowing fragments of roofs and gutters. Subsequently, eight soil profiles, of which 14 soil samples were taken, were examined at the selected sites (Fig. 1). The following building elements were sampled: brick walls (five samples), cracks and fissures present in concrete structural elements (two samples), gutters on buildings (five samples) and bridge foundations (two samples). The dominant vegetation at the sampling locations is also described (Table 1). The location of the points was determined using a GPS receiver.

Fig. 1.

Location of study area.

All samples were air-dried and sieved through a 2-mm sieve. The soil material was then subjected to standard physical and chemical analyses as follows: –

Grain size composition was determined using the aerometric-sieve method (Sheldrick and Wang 1993).

Analyses of HMs content were carried out in triplicate. Arithmetic averages of the results are presented in the paper. Correctness of HMs determination were carried out based on reference samples of soils SO-2 and SO-4 from Canada Centre for Mineral and Energy Technology. Precision of analyses was within the range from about 1.9% to about 8% (e.g. 1.92% for Cu, 2.02% for Cd, 2.43% for Pb, 3.04% for Zn, 7.4% for Cr and 8.1% for Ni) Detection limits for F-AAS are 1.5·g·dm⁻³ for Cd, 3·g·dm⁻³ for Cu, 6·g·dm⁻³ for Cr, 10·g·dm⁻³ for Ni and Pb and 1·g·dm⁻³ for Zn.

The results obtained were statistically analysed. Minimum and maximum values, arithmetic means, standard deviation (SD) and coefficient of variation (CV) were determined. The Pearson correlation coefficient (p < 0.05) was used to determine the relationship between the analysed HMs and selected properties of the studied soils.

Cluster analysis (CA) is a method for geometrically multidimensional clustering of data. Depending on the method used, data can be clustered based on, among other things, geometric shortest distance, variance, mean, centre of gravity, etc. We used Warde’s method, which considers the variance within a group. This method gives the clearest results. The data for clustering have been normalised. This is a statistical method that allows the significance of each type of data to be equalised (e.g. metals whose natural content is very high with those whose content is low).

Principal component analysis (PCA) allows the number of variables describing an object (in our case, a sampling site) to be reduced to a smaller number of variables: principal components. Each new principal component depends on one or more core variables. The highest values are taken by those variables that influence the principal component to the highest level. At the same time, these principal variables are most strongly correlated with each other. This means that PCA can be used as a method to group data according to their variability. It also means that the variables most strongly influencing a given principal component can have similar origins.

The degree of HMs contamination of the investigated soils was described using the following geochemical indices: geoaccumulation index (I_geo), EF, PLI and potential ecological risk index (RI).

The I_geo, developed by Muller (1969), is used to assess HMs contamination based on its content in the soil material under study in relation to a specific GB. The I_geo is calculated according to the formula: Igeo =log2[ Cn1.5· GB ] where: –

C_n – concentration of individual HMs in the sample,

The I_geo covers seven environmental quality classes: –

I_geo ≤0: unpolluted

The EF is an indicator of the possible influence of anthropopressure on the concentration of HMs in the soil. It is based on a comparison of the content of the element analysed against the concentration of a reference element, that is, one that is particularly stable in the soil and does not actively participate in biogeochemical cycles (Sutherland 2000, Plak 2018). Reference elements are most commonly Fe, Al, Mn and Rb. EF is determined according to the formula: EF=Cn/GBnCref /GBref where: –

C_n – concentration of the analysed element in the sample,

Based on the EF, the following categories of enrichment can be distinguished (Acosta et al. 2011): –

EF <2 – minimal enrichment,

Soil contamination can also be assessed using the pollution load index (PLI). It allows the magnitude of contamination in soils to be expressed as an integrated load of all HMs or other toxic elements analysed. PLI values below 1 indicate the absence of contamination, while values above or equal to 1 indicate the presence of contamination (Tomlinson et al. 1980). PLI is calculated as PLI=CF2·CF2·CF3·…·CFn′n where: –

n – amount of HMs and

PLI values below 1 indicate the absence of contamination, while values above or equal to 1 indicate the presence of contamination.

The study also uses an ecological RI to assess the degree of ecological risk caused by HMs concentrations in water, air and soil (Håkanson 1980). The index is calculated according to the following formula: RI=∑i=1nEri where: –

n – amount of HMs,

Potential ecological RI is defined as the sum of the ecological RI for metals in a given sample (Soliman et al. 2015). There are four categories of index (Håkanson 1980): RI <150 – low, RI 150– 300 – moderate, RI 300–600 – considerable and RI >600 – high.

Based on the determined contents of the pollutants studied, indicators commonly used in the literature were calculated to determine the degree of environmental pollution and the associated risks. At the same time, an extremely important element when considering the level of soil contamination in urbanised areas is the determination of GB. The primary purpose of determining this parameter is to distinguish concentrations that are so-called natural from concentrations that deviate from natural, that is, anthropogenic. Most studies of soils located in areas with different functions and anthropopressures use the GB values determined by Pasieczna (2003) to assess the degree of their HMs contamination. The variability of GB of Polish soils is mainly due to the variable chemical composition of the parent rocks on which these soils were formed. In the case of the studied soils, these regularities do not occur due to the specific soil-forming process on anthropogenic soil. For edifisols, the local GB values for Lublin determined by Plak (2018) were used (Table 4). To assess the degree of HMs contamination, the local GB was determined based on the analysis of the average HMs content from the C horizons of the bedrock in the replicate profiles of soils located in Lublin. The methodology used for soil extraction and determination of HMs for edifisols and for GB were the same. Against this background, a number of regional or local anomalies of geological or anthropogenic origin may be identified (Gałuszka 2007).

In the case of the studied soils, these regularities do not occur due to the specific soil-forming process on anthropogenic ground, hence this approach was followed by the authors for the issue. As the investigated soils are characterised by a small thickness (maximum 12 cm), pollution indices were calculated for the whole profiles from the average HMs content in the individual levels.

The investigated soils were formed on buildings with different histories and purposes (Fig. 2). They were characterised by low thickness (up to a maximum of 12 cm) and the presence of a relatively large number of artefacts in the profile. The morphology was mostly characterised by one or two genetic horizons, with only the 6LU profile delineating three genetic horizons. The typical sequence of soil horizons is: Au(h)-Cu(Au2). Regardless of the location and type of technogenic parent material, the dominant soil species is sandy loam or loamy sand (USDA 2002). This material is derived from weathering mortar that is technogenic parent material for edifisols, or it is the material transported by wind and rainwater.

Fig. 2.

Analysed soils: 1LU – watermill, 2LU – brick wall, 3LU – industrial plant (dyehouse), 4LU – railway outbuilding, 5LU – industrial plant, 6LU – car garage, 7LU – railway ramp, 8LU – concrete bridge pillar.

A characteristic feature of soils of this type is the content of artefacts, which in the case of ed-ifisols in Lublin reached up to 20% (vol.). These are mainly rubbish and various types of waste delivered to the soil by man and avifauna (Fig. 3).

Fig. 3.

A) Type of artefacts and the percentage of edifisols in them (% vol.). B) Examples of artefacts in edifisols: 1 – metal components, 2 – roofing felt, 3 – integrated circuit, 4 – trash (bottle stoper), 5 – slag, 6 – glass shards, 7 – plastic elements.

Among the artefacts present in the edifisols formed in the gutters, bituminous substances (roofing felt) and glass shards predominated. A large group of artefacts consisted of elements of building materials such as polystyrene and building foil, as well as fragmented technogenic parent material (Fig. 3).

The texture determines the course of various physicochemical and biological processes and also influences soil structure and the development of the plant cover (Bednarek et al. 2004). The soils studied were mainly characterised by the sand grain size. The average sand content was 64.1%, while the silt content was 34.8% and the clay fraction averaged 1.1% (Table 2). Edifisols formed on masonry and concrete elements were characterised by a high content of skeletal parts (>2 mm). The pH (in H₂O) value of the tested soils ranged from 7.1 to 8.2 (Table 3). The reaction of soil samples taken from masonry and concrete structural elements was mainly alkaline, which is closely related to the presence of mortar, of which lime is a component. The reaction of the soils formed in the gutters was mostly neutral. The percentage of CaCO₃ reached up to 12.4, and the percentage of OC varied considerably among the samples. The highest percentage of OC (8.0%) was found in the soil formed on the water mill ruins (profile 1LU), while the lowest (0.3%) was recorded in profile 3LU.

ECEC is a very important and frequently used indicator in soil quality assessment. It depends on the reaction, humus content and type, grain size and bedrock, as well as on the use. The soils studied were characterised by a relatively high cation exchange capacity: from 5.41 to 109.20 cmol·kg⁻¹, with the proportion of base cations in the sorption complex (BS) ranging between 65.36% and 98.93% (Table 3). In general, higher sorption capacity values were recorded in the subsurface horizons and in soils formed on walls.

Significantly lower sorption capacity (ECEC) values occurred in profiles located in roof gutters (profile 5LU and 6LU). The saturation of the sorption complex with alkaline cations in these profiles was also the lowest, ranging from 65.63% to 87.88%. The sorption complex of the analysed soils is saturated mainly with Ca²⁺ cations, which represent up to 95.54% of the sorption complex (profile 8LU).

In the soil profiles studied, there was a large variation in the content of HMs (Table 4), which is mainly related to their properties, specific location and the impact of the city. The calculated CVs for individual metals were: Cd – 62%, Ni – 86%, Cu – 56%, Zn – 82%, Cr – 45%, Mn – 65%, Pb – 191% and Fe – 65%. The higher the CV value, the greater the anthropogenic influence (Martin et al. 2006).

Analysis of HMs content following aqua regia extraction indicated that profile 8LU was most contaminated with Pb and Cd, with average lead concentrations measuring 617.7 mg·kg⁻¹ (range: 475.8–759 mg·kg⁻¹) and cadmium measuring 6.0 mg·kg⁻¹ (range: 6.0–6.1 mg·kg⁻¹) (Table 4). The concentrations of other analysed HMs were also increased compared to other profiles. The contamination of this profile with HMs may be attributed to the use of corrosion protection preparations in viaduct structure maintenance. Profiles 1LU, 2LU and 3LU exhibited the lowest HMs contamination following aqua regia extraction and developed on brick walls. For instance, the average Cu content in these three profiles was 9.7 mg·kg⁻¹ (range: 2.9–16.2 mg·kg⁻¹) and the Zn content was 121.1 mg·kg⁻¹ (range: 24.5–257.3 mg·kg⁻¹), with lead content found to be below the detection limit. Another distinct group of edaphic profiles comprised those formed in gutters. In profiles 5LU and 6LU, the average Zn concentrations were 553.3 mg·kg⁻¹ (range: 436.6–670.0 mg·kg⁻¹) and 738.5 mg·kg⁻¹ (range: 340.4–1177.0 mg·kg⁻¹), respectively, while the average Cr concentrations were 88.0 mg·kg⁻¹ (range: 80.2–95.8 mg·kg⁻¹) and 45.5 mg·kg⁻¹ (range: 43.9–47.4 mg·kg⁻¹), respectively. Elevated concentrations of these two elements may be linked to their release due to substrate degradation.

The Pb, Zn and Cu contents of the soils studied were similar to the values reported by Charzyński et al. (2015) in soils of this type in different parts of Europe and Africa. However, the mean HMs contents of the edifisols were significantly higher than the mean HMs contents indicated by Plak (2018) for the soils of Lublin; the mean Zn content was five times higher, while Ni content was more than two and a half times higher. The average Cu and Zn contents were also higher than the average values for Polish soils reported by Kabata-Pendias and Pendias (1993).

The HMs content of edifisols was compared with the HMs content of street dust in Lublin, as reported by Zgłobicki et al. (2018). The results showed no correlation, indicating that road dust is not a source of HMs enrichment in edifisols.

The HMs content determined in the tested samples did not exceed the permissible levels specified in the regulations of the Ministry of the Environment in most cases (Journal of Laws No. 2016, item 1395). The permissible limits were slightly exceeded in some levels of the studied soils for Cd. In the 6LU profile produced in the trough, the permissible contents of Zn were exceeded, and for Cd, the level exceeded even twice the permissible content.

The analysis of correlation coefficients calculated for the contents of artefacts, ECEC, HMs and OC revealed significant correlations between Fe and OC, as well as between Ni and Zn. In addition, a strong positive correlation was observed among Ni, Cu and Mn. The highest correlation was found between Pb and Cd, with a correlation coefficient of r = 0.92. Many authors of studies on urban soil contamination indicate that the distribution of HMs depends primarily on the type of anthropogenic activity, which determines both the quantitative and qualitative aspects of the pollutants present in soils. These regularities are mainly marked in industrial cities and agglomerations, where the different functional zones are clearly separated (Horváth et al. 2015, Li et al. 2013, Mireles et al. 2013, Wang et al. 2007).

Based on PCA, two main groups of elemental origin can be distinguished: natural and anthropogenic (Fig. 4). The group of elements with a natural origin can include Cr, Mn, Ni and Zn. The second group consists of Pb and Cd, whose supply to the soil is of anthropogenic origin. Cu can be included in a separate group, and the genesis of this element in edifisols can be mixed. These are surprising results, especially since in other soil types (e.g. Szuszkiewicz et al. 2016), Pb was always a metal indicative of anthropopressure. Pb in the investigated soils significantly exceeded background, especially in profiles 7LU and 8LU.

Fig. 4.

A) Results of PCA; B) results of cluster analysis.

CA analysis yielded a dendrogram grouping variables with similar multidimensional characteristics into the same cluster. According to the diagram, two distinctive groups can be distinguished. The first includes the edifices taken from profiles 1LU, 2LU and 3LU. These are profiles developed on the masonry because of brick and mortar weathering. The second group includes edifisols from the other locations, formed by the accumulation of material in gutters or fissures. This is confirmed, for example, by the study of Szuszkiewicz et al. (2016), who showed that soil composition is more influenced by the bedrock than by the type of pedogenesis. This is confirmed by CA in the case of edifisols collected in the Lublin area, which divides these soils according to the type of ‘bedrock’.

Analysis of the I_geo magnitude for the soil profiles analysed showed the lowest values in profiles 1LU, 2LU, and 3LU by far, which are characterised by no or moderate pollution for most elements (Fig. 5). Only Cr exceeded values > 2, which indicates the highly contaminated category. I_geo values in profiles 4LU, 5LU, 6LU, 7LU and 8LU were characterised by contamination at moderate to extremely contaminated levels. Particularly high values were recorded for Ni, Cu, Zn and Cr. Zn in profiles 5LU and 6LU and lead in profile 8LU reached the pollution categories as extremely polluted. The largest oscillations occurred for Ni and Pb, where no enrichment was found in the profiles 1LU, 2LU and 3LU, while strong contamination was found in the other profiles.

Fig. 5.

Heavy metal content as determined by the indicator I_geo.

Overall, the studied soils were most contaminated with Cr, Zn, Cu and Ni, and the number of profiles classified as grade 3 (moderately and strongly contaminated) or higher was eight profiles for Cr and five profiles each for Zn, Cu and Ni. The average I_geo values for the studied HMs in the edifisols were in the following order: Mn < Pb < Cd < Ni < Cu < Zn < Cr (Table 5).

The average EF values for the analysed HMs were characterised by a wide variation (Fig. 6). In most of the profiles studied, the EF index for Mn and Pb was in the range 0 < EF > 2, indicating no enrichment of these elements. Only in profile 8LU, a very high enrichment of Pb was observed (37.60), which may be due to the immediate vicinity of the profile next to a frequently used road. Natural contents of the analysed elements were also recorded for Cd in profiles 4LU, 5LU, 6LU and for Ni in profiles 1LU, 2LU, 3LU. In the remaining profiles, the EF values calculated for Cd, Ni, Cu and Zn oscillated between 2.67 and 9.04, corresponding to enrichment categories ranging from moderate to significant enrichment. The highest EF values were recorded for Cr, which shows significant enrichment in almost all profiles. Particularly, high values for Cr were determined for profiles 1LU and 2LU, and profile 3LU was characterised by a very high enrichment of this element. Six profiles also showed EF values above 5, that is, significant enrichment for Zn. The average EF values for the studied HMs in the edifisols were in the following order: Mn < Cu < Zn < Ni < Cd < Cr < Pb (Table 5).

Fig. 6.

Heavy metal content determined by the factor EF.

The determined integrated PLI contamination load made it possible to determine the amount of contamination, taking into account all the elements analysed. On its basis, it was found that all profiles were characterised by the presence of a contaminant (PLI ≥ 1).

The lowest PLI values (2.94-4.69) were recorded for profiles 1LU, 2LU, 3LU, 4LU, and thus in soils that formed from weathering mortar and brick and which are located in residential districts. In contrast, the highest PLI values (9.40-12.19) were found in profiles 5LU, 6LU, 7LU and 8LU located in industrial districts (Fig. 7). At the same time, large fluctuations of PLI were observed in the profiles. This may indicate a lack of uniformity in the distribution of elements and may also be due to the different origins of the strata building up the profiles (Arenas-Lago et al. 2014, Islam et al. 2015).

Fig. 7.

PLI for the soil profiles studied.

A low level (< 40) of potential ecological RI for Zn was found in all profiles (Table 5). For Cr, only one profile showed moderate risk, while moderate risk was found for five profiles for Cu and for four profiles for Ni. The index reached the highest values for Cd. In all profiles, it exceeded a value of 40, and in seven levels, the values fell into the category of significant or high and very high risk. The highest level (364) was found for Cd and (309) for Pb in profile 8LU, located close to the road.

The graph of potential ecological RI (total index for all metals) is similar to the graph of PLI (Fig. 8). Profiles 1LU, 2LU and 3LU showed the lowest indicator category (129-141). Profiles 5LU, 6LU and 7LU showed significant ecological risk (360-422). The highest level of this indicator was 840 (very high potential ecological RI) for profile 8LU.

Fig. 8.

Potential ecological risks for the soil profiles studied.

The determined geochemical indices indicated, in addition to the criteria contained in the Regulation of the Minister of the Environment of 1 September 2016 on the manner of conducting assessment of the pollution of the earth surface (Journal of Laws, item 1395), an elevated degree of pollution or enrichment with particular HMs.