1 Introduction

Human activity increasingly affects the natural environment. In 1900, only 13 % of the world’s population lived in the cities (Saier 2007), but by 2011, the percentage had increased to 52 % (United Nations 2013). Huge changes in the soil cover and soil processes can be observed in different cities around the world (Effland and Pouyat 1997; El Khalil et al. 2013). The urban ecosystem is characterised by new types of human-made systems resulted from degradation, destruction and/or substitution of natural systems (Stroganova et al. 1998). Such ecosystem alterations are also identified as an ‘urban stream syndrome’. This term describes the consistently observed ecological degradation of streams draining the urban land (Walsh et al. 2005).

The research on pedogenesis in the urban environment intensified in the 1990s. The urban soils can be defined as bodies located in urban ecosystems, which include a substrate formed by human activity (Blume 1989; Burghardt 1994). According to Sobocká (2003), this term can be also considered as a general terminological concept for soils occurring in urbanised, industrial, traffic, mining and military areas (SUITMA). The main factor for differentiation of SUITMAs from other soils is their location in the abovementioned areas. Therefore, the grouping of SUITMAs is found upon the management system which is an entirely different concept as opposed to soils used for agricultural purposes. The similar approach was used in the World Reference Base for Soil Resources to distinguish Technosols (IUSS Working Group WRB 2007; 2014). They are defined as soils whose properties and pedogenesis are dominated by their technical origin, having a significant amount of artefacts, or a geomembrane, or are sealed with technic hard material (IUSS Working Group WRB 2014).

A very specific type of urban soils develops on buildings. Already in 1840, Vasily Dokuchaev, the major precursor in the field of soil science, mentioned the soils formed on the walls of an abandoned 12-century fortress (Karpachevskiy 1983). On the other hand, Charles Darwin (1890) was the first scientist who described in details the soils formed on the buried ruins of the ancient Roman buildings. While examining the impact of worms on the formation of soil layers (moulds), the author stated that they play a key role in the partial decomposition of raw organic matter. Their activity accelerates the weathering process of structural elements of the buildings due to the increased amounts of humic acids. In addition, Darwin presented the cross sections of soils covering the Roman villas, basilicas and other buildings. The morphology of these soils in relation to their location was also analysed. Afterwards, such soils have not been studied in details until now. Knowledge about their properties is not well developed. However, there are some results of the botanical and archaeological studies conducted in areas where the development of vegetation and soils on the ruins of buildings led to the total concealment of the latter (Lisci et al. 2003; Ceschin and Caneva 2013). Despite the vast and pervasive human perturbation, urban ecosystems can provide a variety of substrata for colonization by flora and fauna (Forman and Godron 1986; McKinney 2002). The building construction materials can serve as ecological niches for various microbes (Hyvärinen et al. 2002), which support the growth of higher plants. Studies of Millard (2004) and Fornal-Pieniak and Chyliński (2012) showed that the establishment of plant communities on abandoned buildings is entirely spontaneous.

The initial study of the soils forming on buildings was conducted in Toruń, Poland by Charzyński et al. (2011a) and Charzyński and Hulisz (2013). The authors proposed a new term to describe these soils, i.e. Edifisols (Latin aedificium = building). This paper is a continuation and extension of the scope of these studies which aimed to characterize Edifisols occurring in different regions of Europe and northwest Africa. It was assumed that most properties of the soils will not vary much depending on the climatic conditions. However, taking into account that similar building materials are used all over the world, the main differentiating factor should be the location within various types of structural elements of a building. That is why, these still pilot studies, on the international scale, took into account the heterogeneity of Edifisols. They allowed the initial, more qualitative than quantitative assessment of the diversity of morphology and properties of these soils.

Moreover, for easier and better communication in the soil science community, proposals are made for the terminology used and the systematic position of the investigated soils in the in the next edition of World Reference Base for Soil Resources. This classification is subject to a constant process of evaluation and improvement, also in the group of Technosols (Séré et al. 2010; Charzyński et al. 2011b; Uzarowicz and Skiba 2011; Charzyński et al. 2013).

2 Study area

The research was carried out in seven countries and at 15 sites which are examples of areas with diverse, natural and anthropogenic environment (Table 1, Fig. 1). The municipalities of different size and function ranging from villages (Rimetea, Romania), small tourist resorts (Tossa de Mar, Spain and Chełmno, Poland) to medium-sized regional centres, such as Toruń (Poland) and Segovia (Spain), and finally metropolises (Milan, Porto, Barcelona, Marrakech and Dakar) were accidentally selected.

Table 1 General characteristics of the study sites
Full size table
Fig. 1
figure 1

Location of the study sites

Full size image

In Poland (Chełmno and Toruń), the climate is warm temperate, fully humid with warm summer (Cfb according to Köppen). More continental climatic conditions characterize the northwestern part of Romania (Rimetea; Dfb). Milan and Bergamo (Italy) have a relatively cool, mid-latitude variant of the humid subtropical climate (Cfa). Some Portuguese (Porto) and Spanish (Avila, Salamanca, Segovia) municipalities are located in the zone of moderate maritime climate with cooler and dry summers (CsB), while subtropical Mediterranean climate with mild winters and warm summers (Csa) is typical for such Spanish cities and towns like Barcelona, Tossa de Mar, Girona and Lloret de Mar. The climate in African locations (Marrakech, Morocco and Dakar, Senegal) is generally warm, semiarid with a short rainy season and a lengthy dry season (BSh)—Martyn (1992), Kottek et al. (2006).

3 Materials and methods

3.1 Fieldwork and laboratory analysis

The research was conducted in 2009–2012. In the selected locations, the reconnaissance fieldwork was carried out in search of poorly maintained or abandoned buildings. The main criterion for the selection of the study sites was the presence of vegetation as an indicator of soil-forming processes occurring on the buildings. The soil material (23 samples in total) was collected from two types of building components: horizontal (roofs, tops of the buildings, bridge surfaces; 17 samples) and vertical (cracks and gaps in the walls; six samples)—Table 1. Additionally, data from the previous publications of the authors (Charzyński et al. 2011a; Charzyński and Hulisz 2013) were used (5 samples) for comparative purposes, which represent the third type of the building components, i.e. concave forms (rain gutters). Also, the dominant vegetation was described at the sampling sites. The nomenclature of plants follows Flora Europaea (Tutin et al. 1993).

All samples were air-dried, disaggregated, homogenized and sieved through a 2-mm mesh. Due to the nature of the studied soils, i.e. developing on the technogenic rock, the whole fraction >2 mm was described as artefacts (IUSS Working Group WRB 2014). The soil material was submitted to standard physical and chemical analyses (van Reeuwijk 2002; Bednarek et al. 2004): particle size distribution (by the hydrometer method combined with the sieve method); pH in water and in 1 M KCl (soil to water ratio 1:2.5); CaCO3 by the Scheibler method; organic carbon (OC) by sample oxidation in the mixture of K2Cr2O7 and H2SO4 (Tyurin method); total nitrogen (Nt) by the Kjeldahl method; and phosphorus soluble in 1 % citric acid (Pca). The content of lead, zinc and copper was determined using the atomic absorption spectrophotometry (AAS) after extraction of samples with a mixture of acids HF + HClO4.

3.2 Data analysis

Data for the 28 soil samples (including five soils described in the literature) were selected for statistical analysis. Scatter plots were drawn using STATISTICA 9.0 software (Statsoft Inc.) to determine the relationships between some soil parameters. The influence of climatic conditions (i.e. BSh, Cfa, Cfb, Csa, CsB and Dfb climates) and location (i.e. on horizontal and vertical building components or in concave forms) on soil properties was tested by Kruskal-Wallis one-way analysis of variance by ranks with post-hoc mean rank multiple comparison (Dunn test) (Zar 1999). The mean values of each soil parameter and standard deviations were calculated. STATISTICA 9.0 software (Statsoft Inc.) was used.

4 Results

4.1 Vegetation

The vegetation cover on the studied soils varied widely. Three groups of plants were distinguished: ruderal, forest and xerophilous (Table 1). The ruderal species were dominant as typical of the habitats disturbed by humans at all of the study sites (e.g. Pimpinella saxifraga, Elymus repens, Lactuca sp., Rubus sp. and Poa sp.). The second group consisted of forest plants that prefer shaded places and soils rich in nitrogen (Hedera helix, Chelidonium majus, Acer platanoides, Digitaria sp.; sites S-Sa3, S-Av1, Pl-Ch1, S-LM1). The last one was the xerophilous species (i.e. Sedum album, Parietaria sp. and Xeranthemum sp.; sites S-Sa1, S-Sa2, S-Se1, S-Ba2, P-Li1, I-Mi2, I-Be2), whereas the Edifisols in the rain gutters (sites from Pl-T1 to Pl-T5) were overgrown with both ruderal and forest plants, e.g. Galinsoga parviflora, Epilobium adnatum, Acer negundo and Populus alba (Charzyński et al. 2011a; Charzyński and Hulisz 2013).

4.2 Soil characteristics

The studied soils were very shallow (maximum thickness up to 10 cm), without visible horizonation and stagnic features (Fig. 2). Most of them (with the exception of soils occurring in the rain gutters) were developed from mineral materials, from sand to sandy loam (Fig. 3). The specific feature of those soils was a high content of artefacts (over 10 % in 15 samples) consisting mainly of brick fragments and slightly weathered mortar.

Fig. 2
figure 2

Examples of Edifisols: 1 S-Av1, Avila I, Spain; 2 I-Be1, Bergamo I, Italy; 3 Pl-T4, Toruń, Poland; 4 S-Av2, Avila II, Spain; 5 R-Ri2, Rimetea, Romania; 6 Pl-T3, Toruń, Poland; 7 M-Ma1, Marrakech, Morocco; 8 I-Mi2, Milan II, Italy; 9 Pl-T5, Toruń, Poland; 10 P-Li1, Lisbon, Portugal; 11 I-Be2, Bergamo II, Italy; 12 Pl-T4, Toruń, Poland

Full size image
Fig. 3
figure 3

Texture of the studied soils

Full size image

The content of organic carbon (OC) and total nitrogen (Nt) significantly varied in the studied soils (Tables 2 and 3). The lowest content of OC (4.1 g kg−1) was recorded at site Se-Da1 located in the crack of the wall. The highest content (up to 394 g kg−1—Pl-T3) was found in the soils developed from the organic material in the rain gutters. The total content of nitrogen (Nt) ranged from 0.4 (Se-Da1) to 21.1 g kg−1 (S-TM1). The C:N ratio was between 3 and 60. Generally, the soil parameters showed the commonly known linear relationship (r = 0.684, p = 0.00006—Fig. 4), but in some cases, the Nt content was disproportionately high in relation to the OC content. As a result, the C:N ratio was extremely narrow (3–5; S-Sa2, S-Sa-3, S-TM1, P-Li1 and I-Mi-1).

Table 2 Properties of the studied soils
Full size table
Table 3 Properties of the Edifisols forming in concave building components (rain gutters)—literature data
Full size table
Fig. 4
figure 4

Scatter plots of the selected soil parameters (n = 28)

Full size image

The pH values in the studied soils ranged from 6.1 to 8.5 in H2O and from 5.7 to 7.9 in KCl, and CaCO3 content was between 0.0 and 476 g kg−1 (Tables 2 and 3). The weakly acid reaction was mainly measured in the organic soils (rain gutters), poor in carbonates. Neutral and alkaline reaction of the samples was caused by the presence of carbonate-rich binder—a parent substance of masonry mortar.

The content of phosphorus soluble in 1 % citric acid (Pca) ranged from 73 mg kg−1 (S-Av2) to 2650 mg kg−1 (I-Mi2)—Tables 2 and 3. The Pca values were not associated with the OC and Nt content (Fig. 4).

The results of heavy metal analysis are presented in Fig. 5. The studied soils were characterised by a very wide range of lead (16–5944 mg kg−1), zinc (30–687 mg kg−1) and copper (7–654 mg kg−1) content, which was considerably different from the background values for world soils. For comparison, the mean concentration of those elements in different soil units (surface horizons) can be as follows: 3–189 mg kg−1 for Pb, 17–125 mg kg−1 for Zn and 13–24 mg kg−1 for Cu (Kabata-Pendias and Pendias 2001). Extremely high Pb concentration was recorded in sample I-Mi2 (5944 mg kg−1), Zn in samples S-Sa3, S-Ba1, P-Li1, Pl-T1, Pl-T2, Pl-T3 (>500 mg kg−1) and Cu in samples Pl-T4, Pl-T5 (448 and 654 mg kg−1, respectively). The analysed heavy metals demonstrated no statistically significant correlations (Fig. 4).

Fig. 5
figure 5

Heavy metal content in the Edifisols

Full size image

4.3 The impact of climate and location on soil properties

The results of Kruskal-Wallis one-way analysis of variance by ranks demonstrated that no significant difference between BSh, Cfa, Cfb, Csa, CsB and Dfb climates was found. However, the soils on different types of building components differed significantly. The Kruskal-Wallis test with post-hoc mean rank multiple comparison revealed that soils in concave forms were characterised by lower pH (pH-H2O 6.7 and pH-KCl 6.4) in relation to soils on horizontal and vertical building components (Table 4). Moreover, soils in concave forms had significantly higher organic matter contents (207 g kg−1) and C:N ratios (27). Furthermore, they significantly differed in CaCO3 content which is lower (15 g kg−1) than in soils on vertical locations (159 g kg−1) (Table 4). Differences in other properties were not significant. Each of the analysed soil characteristics showed a great variability, as evidenced by SD, often above mean values (Table 4).

Table 4 Comparison of soil properties in different localities (mean ± SD). Kruskal-Wallis test was applied with Dunn test as post-hoc nonparametric multiple comparisons. Treatments with no difference are marked by the same letter (a or b)
Full size table

5 Discussion

5.1 Specificity of Edifisols and their similarity to other soils

In the light of this study, several properties of Edifisols should be regarded as characteristic, i.e. shallowness, richness in artefacts and carbonates, very varying content of organic carbon (OC) and total nitrogen (Nt), elevated phosphorus content (Pca) and heavy metal contamination (Tables 2 and 3, Figs. 3 and 5). This description is in line with typical characteristics of urban soils given by Lehmann and Stahr (2007). The studied soils, excluding soils located in the rain gutters, are developed from technogenic substrates, usually abundant in carbonates. As a result, shallow soils, up to 10 cm thick, are formed with properties similar to weakly developed lithogenic soils (Rendzic Leptosols). On the other hand, the soils located in the rain gutters were characterised by low deposition of mortar and the presence of large amounts of organic matter derived from the decomposition of litter fall (leaves, flowers, small twigs) carried by the wind from trees growing nearby or washed off the roof. Their properties were similar to shallow mountain organic soils with folic horizon—Folic Histosols (Küfmann 2003).

Very specific features of certain soils (5 samples) include a very low C:N ratio (below 5), rarely found in natural or technogenic soils. This can probably be explained by a significantly increased input of nitrogen, mainly through bird droppings, which are also an important source of phosphorus (Crowther 1997). In this regard, the abovementioned Edifisols are similar to ornithogenic soils, common in the polar regions. According to Kim et al. (2012), the C:N ratio in the layers enriched with ornithogenic material may range from 1 to 4.

Soils in the cities are very often contaminated with heavy metals, which are a serious threat to the entire urban ecosystems (Grzebisz et al. 2002; Manta et al. 2002; Uzarowicz and Maciejewska 2012). That is why, some authors suggest to use certain trace elements (Cu, Pb, and Zn) as indicators of technosolisation (El Khalil et al. 2013). Most of the studied Edifisols were enriched with lead, zinc and copper. In relation to the background values given by Kabata-Pendias and Pendias (2001), 36 % of the analysed samples had elevated values for Pb content, 61 % for Zn and 68 % for Cu. These metals could enter Edifisols primarily through industrial and traffic pollution (e.g. S-Sa2, P-Li1, I-Mi2), building materials (e.g. S-Sa3, S-Ba1, Pl-T1-T5), etc. It should be noted that the dominant neutral and alkaline soil reaction resulting in immobilization of some metals and reduction of their bioavailability for plants is an additional factor responsible for their accumulation (Brümmer and Herms 1983).

The comparison of soil properties between the three categories (i.e. soils formed on horizontal, vertical and in concave building components) by the Kruskal-Wallis test with post-hoc mean rank multiple comparison revealed distinct properties of soils formed in gutters. They were characterised by lower reaction, higher organic matter content and C:N ratio and lower calcium carbonate content. What is interesting is that soils formed on horizontal and vertical building components did not differ significantly. Moreover, the three investigated categories did not differ in pollution by heavy metals.

The statistical comparison showed that climate was not a key factor in the differentiation of the studied soil properties contrary to location. Due to the limited dataset, however, it should be assumed that the climate may affect the rate of soil formation processes in Edifisols. Therefore, the presented results should be interpreted with caution because of the pilot character of the study. This issue should be further researched.

5.2 Stages of the soil development on the buildings

Soils forming on buildings in the areas currently inhabited are usually ephemeral and young (Burghardt 2001). This can be associated with a specific character of objects on which they developed. Ruins of no historical value (e.g. S-Av1) might be demolished within a short time as they are a blot on city centres. The rain gutters with soils may break under the load of deposited material or they may be cleaned (e.g. Pl-T2). Buildings of historical importance (e.g. I-Mi1), still used by man, are periodically cleaned of all the soil and vegetation. Buildings are a type of geomembrane which, in the case of Edifisols, blocks the contact with natural or technogenic soils occurring on the Earth’s surface. Lack of soil sealing does not reduce the impact of the external environment on the soil development and properties.

Figure 6 presents the stages of the development of Edifisols taking into account the functioning of the building over time. After the construction of a building, the growth of plants and soil development on its surfaces depend on the building maintenance and the state of conservation (i.e. clearing and repairing), the building material and the climate. According to Lisci et al. (2003), it usually takes more than 10 years for plants to colonize a wall (in the Mediterranean Basin). Following the aforementioned authors, two modes of plant colonization on the buildings can be distinguished. The first one is independent of the position and moisture conditions. The initial deterioration of the building structure caused by bacteria, fungi and lichens can lead to the formation of a substrate suitable for the germination of seeds of hardy pioneer plants. The biological processes favour the intensification of physical and chemical weathering of building elements. Simultaneously, it may be followed by the input of mineral and organic matter (atmospheric dust, bird excrements and human wastes). The detritus of the pioneer plants can also form a soil substrate (Folic layer). The second mode of colonisation occurs mostly on horizontal positions with a sufficient water supply. In such conditions, mosses are the pioneer plants. They accelerate the formation of an initial soil substrate by trapping the atmospheric dust.

Fig. 6
figure 6

Block diagram presenting the development of the Edifisols in relation to the building’s age and fate

Full size image

In areas formerly inhabited and now abandoned, where a long-term soil-forming process occurs and is uninterrupted by human intervention, an increase in the thickness of Edifisols is observed. It can lead to a complete coverage of the buildings, first by plants and then by the soil cover. This process lasts until the transformation of a building into a natural-looking land form, e.g. the Mayan ruins in Mexico and Guatemala (Fig. 7). However, in the case of smaller buildings, e.g. Roman villas, the ruins can totally disappear, covered by well-developed soils (Darwin 1890; Barone et al. 2013; Fig. 7), which originally started as Edifisols. According to WRB (IUSS Working Group WRB 2014), they start as Technosols and when they become thicker, their Reference Soil Group may change.

Fig. 7
figure 7

The example of the extreme Edifisols—an unrecognizable building ruins completely covered by soil and vegetation (Tikal, Guatemala)

Full size image

5.3 Proposal for classification of Edifisols

So far, the studied soils have not been included in a proposal for classification of urban soils, except for a suggestion made by Burghardt (1994) who, however, outlined only one of the possible scenarios for the development of soils on man-made structures, i.e. the formation of so called ‘aerosols’ as a result of aeolian accumulation of the dust produced by vehicular traffic.

The newest developments in the taxonomy of technogenic soils i.e. a proposal to integrate the systematic of urban soils into the new Russian soil classification system (Prokof’eva et al. 2013) or the Greinert concept for Systematics of Polish Soils (Greinert et al. 2013) do not ensure an adequate classification of Edifisols due to the lack of an appropriate unit.

Based on the research on soils developing on buildings, we propose a new taxonomic unit called Edifisols (Latin aedificium = building), which should be introduced in the technogenic soils taxonomies. It is also important to provide a possibility of precise classification of the described soils in the international classification of WRB because they develop worldwide in urbanised areas. In the currently valid edition of WRB (IUSS Working Group WRB 2014), Edifisols may be classified as Isolatic Technosols or Leptosols, Histosols and Regosols with the Isolatic supplementary qualifier. It also seems likely (although further studies are required to confirm it) that some of the well-developed soils covering the ancient ruins (see e.g. Darwin 1890; Barone et al. 2013) could meet the criteria for Pheozems and Umbrisols. Isolatic qualifier comprises two opposite types of soils, i.e. constructed soils of ‘green roofs’, intentionally placed on the top of buildings and Edifisols. These two groups of soils are characterised by different genesis and properties. The above classification does not specify the nature of Edifisols, so the next edition of WRB should include the additional qualifier Edific. The definition of this qualifier should account for location on the buildings and spontaneous development as a result of weathering of technogenic material in situ with the supply of mineral and organic matter. The proposal is given below:

Edific (ef)

having a thickness of ≤100 cm and developing spontaneously on buildings, without intentional human activity from technic hard material as a result of weathering of technogenic substrates in situ and with the supply of mineral and organic matter carried by wind, rainwater or animals.

6 Conclusions

This paper characterises a new unit of technogenic soils, which is proposed to be called Edifisols. These soils can develop on different types of building components: horizontal (e.g. flat roofs, tops of buildings, bridge surfaces), vertical (cracks and gaps in walls) and concave (e.g. rain gutters).

Edifisols are formed as a result of initial, relatively natural soil-forming processes occurring on technogenic substrates. Therefore, they may be considered as certain analogues of natural initial or weakly developed soils (Rendzic Leptosols, Folic Histosols and ornithogenic soils). They are often ephemeral soils susceptible to an influx of any kind of contamination and transformation. The influence of the natural environment on the discussed soils is mainly manifested in the presence of different plant species (e.g. xerophilous, thermophilous).

The results of the statistical analysis showed that climate settings in Poland, Romania, Italy, Spain, Portugal, Morocco and Senegal did not affect the studied properties of soils. The properties of Edifisols are closely linked to their position on buildings. Within the investigated types of Edifisols, only soils formed in the concave structural elements have distinct characteristics. They are characterised by a higher content of organic matter, higher values of the C:N ratio, lower pH and lower content of CaCO3 in comparison with Edifisols developed on horizontal and vertical elements.

According to the World Reference Base for Soil Resources (IUSS Working Group WRB 2014), Edifisols can be classified using the Isolatic qualifier which refers to soils formed both as a consequence of intentional human activity and with no human impact. According to the authors, the next edition of WRB should be supplemented with the qualifier Edific (ef) to be used for the Reference Soil Groups of Technosols, Histosols, Leptosols, Phaeozems, Umbrisols and Regosols.

The authors have undertaken a pioneer challenge to identify the nature of Edifisols. The research on those interesting and worldwide occurring technogenic soils should be continued. It seems that the most important aspect of future research should consist in the determination of the rate of soil formation on the buildings under various climatic conditions and the plant cover.