Compositional changes in natural communities associated with anthropogenic influence often lead to localised extinctions and biodiversity loss. Soil invertebrates are also threatened by urbanisation due to habitat fragmentation, vegetation changes and management, soil alteration, degradation, and disappearing shelter sites. The aim was to assess terrestrial isopod (Oniscidea) assemblages in differently degraded urban forest patches of a metropolitan area (Budapest, Hungary). Study sites were compared by their species richness, composition and the relevant background factors (soil properties, dead wood, litter characteristics, and canopy closure). The degree of urban disturbance was expressed using an urbanisation index (UI) based on built-up density and vegetation cover. The isopods were identified to species level, and were qualified by their habitat preference and naturalness index (TINI). Average Rarity Index (ARI), derived from TINIs provided information on the degree of naturalness/disturbance of each habitat. Altogether 14 isopod species were collected from 23 sample sites. Urbanisation indirectly affected on the composition of isopod assemblages through the quantity of dead wood and soil plasticity. ARIs and UIs of sample sites were negatively correlated. Urban patches harboured habitat generalist, synanthropic and established introduced species with low naturalness value of assemblages. Areas with no or low anthropogenic disturbance maintained stable native, autochthonous assemblages that were characteristic of rural sites in the region. Transitional zones between rural and urban habitats usually maintained a mixed isopod fauna consisting of both urban and rural elements.

biotic homogenisation, disturbance tolerance, ecological character, habitat specialist, urbanisation index, woodlice

Currently increasing number of studies explore the effects of urbanisation on biological communities at a global level (Niemelä et al. 2000, McKinney 2008, Richter and Weiland 2012, Wang et al. 2012a). The alteration and fragmentation of natural habitats generally leads to a shift in species composition, resulting in biotic homogenisation and changes in ecosystem services as well (McPherson 1998, Whitford et al. 2001, McKinney 2006, Tratalos et al. 2007). Human activity, such as construction industry, air pollution and pollutant emissions of vehicles and the use of chemicals, contributes to urban soil degradation (Pouyat et al. 2008).

The majority of soil invertebrates are highly sensitive to disturbances (Barbercheck et al. 2009) and environmental changes (Santorufo et al. 2012). This includes the macrodetritivore fauna, which has an important role in the ecosystems’ nutrient cycling. These invertebrates fragment dead plant material through their feeding activity increasing its surface area and promoting microbial decomposition (Hanlon and Anderson 1980, Bardgett 2005). Woodlice (Isopoda: Oniscidea) are one of the major invertebrate group contributing to these processes (Anderson 1988, Paoletti and Hassall 1999).

Terrestrial isopods can be used as ecological indicators of habitat qualification. They are widespread, limited in their dispersal abilities, and relatively easy to collect and identify. Based on a single species’ ecological needs and tolerances, species composition informs us about habitat characteristics including habitat disturbance/naturalness (Vilisics et al. 2007, Tuf and Tufová 2008, Hornung et al. 2008, 2009). Urban areas are hot spots for species introduction (Vilisics and Hornung 2009) threatening natural communities. Successfully established introduced species are usually eurytopic and/or cosmopolitan ones leading to global urban soil fauna homogenisation and convergence (McKinney 2006, Pouyat et al. 2015).

To study the effects of urbanisation on oniscidean fauna differently urbanised woodland habitat patches were compared. We expected that our data would be in accordance with the following hypotheses:

(1) the ’Intermediate Disturbance Hypothesis’ (IDH; Connell 1978) that predicts diversity being the highest in habitat with moderate levels of disturbance;

(2) the ’Habitat specialist hypothesis’ that indicates that the abundance and species richness of forest specialist species will decline along a rural–suburban–urban gradient (Magura et al. 2008) and

(3) the ’Synanthropic species hypothesis’ that predicts that abundance and species richness of synanthropic species will increase along a rural–suburban–urban gradient (Magura et al. 2008).

Urban soils are overwhelmed by strong human physical effects (e.g. grading and irrigation) and tend to lack the effects of native factors (e.g. topography and drainage) that formed development of soil characteristics during a long time period. These soils vary widely in their characteristics and are dependent on both direct and indirect effects resulting from urban land use and cover change (Pouyat et al. 2010). The various plots differed slightly in soil parameters, mainly in CaCO₃ content. In accordance with results of previous urban studies (Craul and Klein 1980, Short et al. 1986) higher CaCO₃ content, higher pH, but lower humus content and soil plasticity (K_A) was experienced in habitats with high urbanisation index compared to rural sites. Building materials, for example concrete, asphalt and bricks can be the sources of increased CaCO₃ level (Alexandrovskaya and Alexandrovskiy 2000), which can lead to higher soil pH (Pouyat et al. 2015).

The 14 isopod species found in our survey represent 50% and 25% of the recorded fauna of Budapest and Hungary, respectively (Hornung et al. 2007a, 2008, Vilisics and Hornung 2008, 2009). Korsós et al. (2002) and later Vilisics and Hornung (2009) summarised data on the oniscidean fauna of the entire metropolitan area of Budapest, including different kinds of man-made habitats, e.g. gardens, courtyards, densely built-up areas and botanical gardens. The present study focused only on a subset of the diverse urban habitat types on the Buda side of the city and thus fewer species were detected.

Although average species richness (α diversity) is usually low in Hungary, three species per location on average (Hornung et al. 2008), in the present study 6 locations out of the 23 (26%) resulted 4–6 species. These habitats could be referred to fringe areas. The species number and assemblage composition of some other European cities are also known. In an urban study of Olomouc, Czech Republic, sampling city parks, built up areas, gardens, ruderal and natural habitats altogether 17 woodlice species were collected (Riedel et al. 2007). In relation to habitat naturalness the species spectrum was dominated by adaptable and eurytopic species (sensu Tuf and Tufova 2008) meaning habitat generalist species. Ferenți et al. (2015) published eleven species from 19 localities within Salonta town (western Romania) with the highest (6) species richness from a wetland habitat and most sampling sites harbouring only one or two species. In Bucharest (Romania) 17 species were sampled by Giurginca et al. (2017). In three Swiss cities 17 species were mentioned by Vilisics et al. (2012). The highest species richness was found in Lucerne (13), while Zurich had eleven and Lugano nine species. The data mentioned above are influenced by difference in sampling methods, but give a tentative picture of the species richness of different geographic regions. The natural species pools surrounding urban areas in the different (bio)geographical areas may also differ within Europe. In addition to the basic European fauna they may contain different regional biogeographical elements.

The results of the present study failed to fulfil the requirements of the Intermediate Disturbance Hypothesis. There was no consistent pattern in species richness distribution. However, urban habitats harboured more species on average, but without any statistically significant correlation. In a rural – urban gradient study (Debrecen, Eastern Hungary) IDH was also not proved for woodlice, while it was valid for millipedes (Hornung et al. 2007b, Bogyó et al. 2015). In another extensive urban study (Pécs/Hungary; Farkas and Vilisics 2006) either habitat generalists (A. vulgare, Po. collicola, Trachelipus rathkii) or synanthropic species (P. scaber, C. convexus, Ps. pruinosus) dominated densely built-in areas while in the city edge ecotone zones the more habitat specialists also joined the assemblages (T. nodulosus and Pr. politus). Rural – urban transition zones and ecotones also had a mixed isopod fauna in the present study.

The Habitat specialist and the Synanthropic species hypotheses (Magura et al. 2008) were confirmed by the present study. We experienced a species exchange between habitat specialists and generalists and/or synanthropic species along the rural – urban gradient. Typical forest species (Pr. politus, O. planum) were constant elements of rural habitats while habitat generalist and/or synanthropic species (porcellionids, A vulgare, C. convexus) constituted the species assemblages of urbanised forest patches.

The apparent negative association of P. scaber, a typical urban faunal element, with dead wood can be explained by its high tendency for aggregation (Broly et al. 2012). Hornung et al. (2008) found that species in the family Porcellionidae show a clear preference for human settlements in Hungary. Porcellio scaber, C. convexus and Ps. pruinosus are typical elements of urban ecosystems and farmlands. They are common also in many other parts of the world (Hornung et al. 2008). The unusual appearance of species with a low desiccation tolerance (Trichoniscidae species with hygrophilic and/or endogeic nature) was always connected with some kind of constant water supply, e.g. public artesian wells (sites 4, 18, 16). The significant correlation of O. planum and Pr. politus with the amount of wood debris can be attributed to their typical sylvicolic nature (Hornung et al. 2008, Tomescu et al. 2008). Geographical distance may also be important. The inner city is fairly isolated from rural areas preventing species dispersal and colonisation.

Isopods are indicators of the naturalness of vegetation, and the quality and quantity of dead wood and litter in their habitats, which are used by them both as food and shelter (Rushton and Hassall 1983). Habitats with a lot of dead wood were favourable for many natural species. Dead wood was the main driver of species composition in isopod assemblages, probably because they largely contribute to the supply of organic matter and affect microclimatic variables, e.g. humidity. Preventing removal of dead wood from urban green spaces could enhance the survival of the epigeic invertebrate fauna (Bang and Faeth 2011). This indicates the importance of small-scale management decisions on local biodiversity. Urban planning and changes in the management of public areas, e.g. retaining leaf litter, might be advantageous for soil fauna survival which in turn provide food for various vertebrates and may increase biodiversity as a whole (Stagoll et al. 2010, Threlfall et al. 2016).

However, the 1 × 1 km area units used for determination of UIs seems to be a too broad a scale compared to the small-scale heteromorphic sensibility of the investigated flightless epigeic macrodetritivore fauna. Their habitat preference may be controlled on a much finer scale (see also McCary et al. 2017). UI might be also misleading in some cases as e.g. forests with high hiking traffic but without any buildings and sealed surfaces (sites 11 and 23 in our case). These showed low urbanisation (negative UI scores) and naturalness values (ARI) but impact of anthropogenic disturbance is hidden. Similarly, sharp habitat boundaries with mixtures of urban and rural species explains the exceptional position of sample site 14 on Figs 4 and 5. The UIs and the total naturalness index of isopod species (∑TINI) at the sample sites did not show significant relatedness, but the habitats’ ARIs did, which confirms previous statements (Hornung et al. 2009): ARI is a good tool in estimating habitats’ naturalness as it involves the species’ ecological features and enables a more refined evaluation. High species number does not always mean high naturalness from a conservation biological point of view.

Urbanisation often leads to changes in species richness and community composition. New landscapes and habitats are formed that do not occur elsewhere (Niemelä 1999, Mabelis 2005). Species have different responses to anthropogenic habitat modification, depending on their ecological needs and tolerance. Urbanisation worldwide is accompanied by the occurrence and dominance of habitat generalist species with broad tolerances and the establishment of introduced, mainly synanthropic species. These changes lead to homogenisation and convergence of urban faunas on both local and global scales.

Woody habitats within the urban matrix can still support biodiversity to varying degrees. As species have different responses to anthropogenic impacts, the species composition of urban areas can depend greatly on the habitat characteristics of the local and surrounding areas and their distances from natural species pools. Urban patches harbour assemblages that are relatively modest in species richness and have low naturalness values. The composition usually consists of typical homogenising urban species. Transitional zones (fringe areas) between rural and urban habitats might maintain an assemblage of rural, habitat specialist elements with high naturalness value mixed with urban ones. Areas with no or low disturbance maintain species poor but stable native, autochthonous assemblages, with high naturalness value, characteristic for rural sites in the region.

The Terrestrial Isopod Naturalness Index (TINI) and the Average Rarity Index (ARI) give good possibilities to assess urban effects on habitats and serve as potential tools for habitat qualification. Our study demonstrates that maintaining litter layer with dead wood in urban habitats is an essential factor for favouring natural/unique oniscidean assemblages and we suggest that remnants of natural habitats within cities receive further attention in urban planning.

We are grateful to both reviewers for their useful suggestions and to Drs Katalin Szlavecz (JHU/Baltimore, US) and Helen Read (London/UK) for their linguistic corrections. We also thank the Department of Ecology for financial support and facilities necessary to complete this work. This publication was supported by the 12190-4/2017/FEKUTSTRAT grant of the Hungarian Ministry of Human Capacities.