Urban areas are deeply modified spaces for human residence, covering almost 600 000 km² and housing around 40% of the world population (Ellis & Ramankutty, 2008). Urban development involves a syndrome of abiotic and biotic changes at different spatial and temporal scales, including fragmentation, pollution, heath island effect, changing land use and introduction of exotic species (Martinson & Raupp, 2013; Fenoglio et al., 2020). Worldwide, species richness of plants and animals decreases in central urban areas, but sometimes it increases in suburban areas due to intermediate levels of human disturbance (McKinney, 2008).

For terrestrial arthropods, urbanization have overall detrimental effects on abundance and diversity, directly related to age of cities and more severe for species with high dispersal abilities belonging to orders Coleoptera and Lepidoptera (Fenoglio et al., 2020). Particular attention has been devoted to studying the urbanization effects on epigeic carabid assemblages inhabiting fragmented forests within urban-rural gradients in the Northern Hemisphere (Niemela & Kotze, 2009). In most cases, abundance and species richness of Carabidae are reduced from rural surroundings to city cores, notoriously so for brachypterous, larger-sized and stenotopic species (Niemela & Kotze, 2009; Martinson & Raupp, 2013). These general tendencies could be distorted by environmental conditions at plot scale and persistence of local species pool, in fact urbanization doesn't seem to drive to a single synanthropic assemblage (Magura et al., 2010) and management of urban forest fragments could raise abundance and richness (Belskaya et al., 2020).

Urbanization effects on epigeic arthropod assemblages inhabiting sandy soils have been documented for several arid zones and spatial scales. In San Diego County (USA), area and age of native scrub fragments influenced abundance and richness of arthropods, both being lesser in aged and small-sized fragments (Bolger et al., 2000). In Phoenix area (USA), changes of land use from desert environments to agricultural, industrial or residential zones promoted higher abundances of detritivores (springtails) and predators (spiders) sustained by soil nutrient input and supplemental watering (McIntyre et al., 2001; Shochat et al., 2004). Similar outcomes were registered for beetle assemblages in New Damietta city (Egypt), where higher values of abundance, richness and equitability were registered in coastal sandy environments compared to compacted and fertilized soils in urbanized sites (El Surtasi et al., 2012). In fragmented European dunes, psammophilous butterflies, grasshoppers and spiders were on higher risk of local extinction due to their low dispersal abilities and scarce probability to survive in trampled patches (Bonte et al., 2003; Bonte & Maes, 2008), while psammophilous beetles appear to be more resilient in similar scenarios (Comor et al., 2008). Urban development promoted introduction of exotic plant species that alters habitat structure and hence composition of arthropod assemblages as were evidenced in cases of marram grass (Webb et al., 2000) and weed bitou bush (Wilkie et al., 2007) in Australia. Also, urbanization promoted overwhelming abundances of exotic arthropod species that displace native species belonging to same functional guilds, for instance cosmopolitan species of ants, cockroaches, isopods and spiders in coastal areas of California (Suarez et al., 1998; Bolger et al., 2000; Longcore, 2003).

In Alicante province, beaches and dunes have been disturbed by a poorly planned urban development since the 1960s to date. Particularly, El Saladar and Arenales del Sol beaches have been rated as very unattractive urban sites because their intensive development and low landscape values (Asensio-Montesinos et al., 2017). Regarding conservation status of vegetation at El Saladar dunes, it was classified as a clearly degraded site, with a mixed coverage of native and non-native species, also an impoverished coast to inland structure (Albertos et al., 2010).

The purpose of this preliminary work is to characterize diversity and structure of epigeic arthropod assemblages inhabiting urban-influenced coastal environments, based on a short sampling made near to El Altet (Alicante, Spain) between November 2004 and March 2005. To our knowledge, this is the first study on epigeic arthropod assemblages influenced by urbanization in coastal environments of Southeastern Spain.

The field work was done between November 2004 and March 2005, autumn and winter of Mediterranean Region, in a mosaic of dune and anthropogenic habitats at El Saladar dunes and Arenales del Sol urbanization, located near to El Altet, Elche district, Alicante province, southeastern Spain (Fig. 1a).

The sampling area included two urban vegetation fragments covered with a mixture of yellow dune and ruderal vegetation, a small fragment (0.25 ha, 38°15'15” N, 00°31'01” W) and a medium fragment (1.08 ha, 38°15'02” N, 00°30'59” W). Also relatively undisturbed dune vegetation (84 ha) including three dune types, grey (38°15'55” N, 00°31'19” W), heath (38°16'16” N, 00°31'27” W) and yellow (38°15'59” N, 00°31'14” W) (Fig. 1b).

A total of 80 pitfall traps arranged in eight sampling transects were deployed in the following habitat types (Fig. 1c–d):

In each transect, 10 pitfall traps were installed to collect epigeic arthropods. Traps were disposable plastic glasses (10 cm diameter), buried at ground level, separated each other by 5m distance and halffilled with a preservative solution (70% ethylene glycol, 30% water). Trapped specimens were weekly extracted and preserved in ethanol (70%).

In the laboratory, macroarthropods were sorted to order and families under a stereoscope (20×–40×) and using appropriate taxonomic keys (Barrientos, 1988). Determination level had higher resolution only for 22 abundant Coleoptera species of Carabidae, Chrysomelidae, Scarabaeidae, Staphylinidae and Tenebrionidae families. Ants were excluded because their bulky aggregations, as well as epiphytic and flying insects incidentally collected by pitfall traps.

Statistical analysis were conducted with accumulated data of entire sampling period for each transect and transects were grouped according to their habitat types, yellow dune (YD), ruderal vegetation (RV) and, grey dune and heath dune joined because both are natural habitats with semi-consolidated to consolidate sandy substrates (GD+HD). In following lines, the term abundance is used for the sake of simplicity, but it is clear that the number of individuals caught in pitfall traps reflects “activity-density” and not the real taxa density (Southwood & Henderson, 2009).

Diversity of epigeic arthropod assemblages was assessed for mean abundance of arthropod families in each group of habitat types (YD, RV, GD+HD), with two diversity metrics (Krebs, 1989; Hammer et al., 2001): 1) individual rarefaction, to estimate the number of expected taxa for any sample size smaller than the largest sample, using a log Gamma function for computing combinatorial terms and giving as a result linear plots of sample size vs. number of taxa with 95 percent confidence intervals based on standard errors (square roots of variances); 2) diversity profiles, for to calculate a family of diversity indices dependent upon a single continuous parameter a, derived from the exponential of the so-called Renyi index, obtaining total species number for α = 0, Shannon index for α = 1 and Simpson index for α = 2 and giving as a result a plot of diversity profiles with 95 percent confidence intervals based on 2000 replicates (bootstrapping).

Structure of epigeic arthropod assemblages was assessed for log-transformed abundance of 46 arthropod families and 22 beetle species in each sampling transect, with two multivariate analysis (Legendre & Legendre, 1998; Hammer et al., 2001): 1) Non-metric multidimensional scaling (NMDS) to place sampling transects in a bi-dimensional coordinate system according a their similarities measured with Bray-Curtis distance; 2) Analysis of Similarities (ANOSIM) to test the statistical significance of pre-established groups of habitat types (YD, RV, HD+GD) (p < 0.05). Also relationships between scores of NMDS axes with taxa abundances were measured with Spearman rank correlations (p < 0.05).

Diversity metrics and multivariate analysis were performed with PAST statistical software (Hammer et al., 2001).

Overall, 9463 specimens of epigeic arthropods were collected between November 2004 to March 2005 comprising 15 orders and 46 families (Appendix 1). Most abundant orders were Coleoptera, Dermaptera, Isopoda, Araneae and Amphipoda (Fig. 2a). Most abundant families were Forficulidae, Tenebrionidae, Carabidae, Porcellionidae, Curculionidae and Talitridae (Fig. 2b). The subset of 22 Coleoptera species comprised 3041 specimens belonging to following families: Carabidae (seven species), Chrysomelidae (two species), Curculionidae (one species), Scarabaeidae (one species), Staphylinidae (one species) and Tenebrionidae (ten species) (Appendix 2).

Linear plots of sample size (specimens) vs. number of taxa (families) in groups of habitat types (HD+GD, YD, RV) are presented in Fig. 3a. According to these plots, number of families of epigeic arthropods was higher in HD+GD habitat type and was lower in YD and RV habitat types, whose confidence intervals are widely overlapping. Diversity profiles for groups of habitat types are presented in Fig. 3b. Comparison of these profiles shows higher diversity in HD+GD habitat type and lower diversity in YD and RV habitat types, both with confidence intervals entirely overlapped.

NMDS plot for the eight sampling transects based on epigeic arthropod families data is presented in Fig. 4a. Sampling transects were arranged from relatively undisturbed dunes (GD, YD1) to ruderal vegetation (RV1, RV2, RV3) along the first axis, and also from sandy soil sites (YD1, YD2, YD3) to compacted soil sites (HD) along the second axis. Also, transects were grouped in NMDS plot according to their habitat type, ruderal vegetation (RV1, RV2, RV3), yellow dunes (YD1, YD2, YD3) and semi-consolidated to consolidate dunes (GD, HD). Results of ANOSIM test are depicted in Fig. 4a (global) and Appendix 3 (global, mean ranks and pairwise comparisons). Correlations between log-transformed abundances of epigeic arthropod families with NMDS axes are presented in Table 1. Significant correlations with first axis were positive for Dysderiidae, Curculionidae, Staphylinidae, Forficulidae, Pyrrhocoridae and Porcellionidae, and were negative only for Myrmeleontidae and Phalangiidae. While with second axis, positive significant correlations were established for Lycosidae, Julidae and Cydnidae, and negative significant correlations for Carabidae, Scarabaeidae and Tenebrionidae.

NMDS plot for the eight sampling transects based on Coleoptera species data is presented in Fig. 4b. Sampling transects were arranged from sandy soil sites (YD1, YD2, YD3) to compacted soil sites (HD) along the first axis, and also from undisturbed dunes (GD, YD) to ruderal vegetation (RV1, RV2, RV3) along the second axis. Also, transects were grouped in NMDS plot according to their habitat type, ruderal vegetation (RV1, RV2, RV3), yellow and grey dunes (YD1, YD2, YD3, GD) and heath dune (HD). Results of ANOSIM test are depicted in Fig. 4b (global) and Appendix 3 (global, mean ranks and pairwise comparisons). Correlations between log-transformed abundances of Coleoptera species with NMDS axes are presented in Table 2. Significant correlations with first axis were negative for Harpalus fulvus Dejean, 1829, Masoreus wetterhalli Gyllenhal, 1813, Psammodius porcicollis (Illiger, 1803), Erodius carinatus Solier, 1834, Halammobia pellucida (Herbst, 1799), Pachychila frioli Solier, 1835, Pseudosericius pruinosus (Dufour, 1820) and Xanthomus pellucidus Mulsant & Rey, 1856, and positive only for Ocypus olens (Müller, 1764). While with second axis, significant correlations were positive for Macrothorax morbillosus (Fabricius, 1792), Ocypus olens (Müller, 1764), Leichenum pulchellum (Lucas, 1849) and Tentyria laevis Solier, 1835, and negative for Timarcha espagnoli Bechyné, 1948 and Tentyria elongata Waltl, 1835.

Percentage composition of orders and families shows in general terms a resource apportionment between a native epigeic fauna typical of arid ecosystems composed mainly by psammophilous Coleoptera and instead, another promoted by human disturbances composed mainly by exotic or invasive Dermaptera and Isopoda. This general appreciation about the epigeic fauna and its relationships with human disturbances will be confirmed in diversity metrics and multivariate analysis discussed in following lines.

The synthetic interpretation of individual rarefaction and diversity profiles shows three different scenarios: 1) high abundance and low diversity of epigeic arthropods in sites disturbed by fragmentation and land use change (RV), 2) intermediate abundance and low diversity of epigeic arthropods in relative undisturbed sites with mobile sand substrates (YD), 3) low abundance and high diversity of epigeic arthropods in relative undisturbed sites with semi-consolidated to consolidated sand substrates (HD+GD). These findings coincide with known effects of land use change on epigeic arthropods from other arid environments, where increased productivity is an extraordinary amount of energy that supports a greater abundance of arthropods, particularly some detritivores (Collembola, Dermaptera) and predators (Araneae) (McIntyre et al., 2001; Shochat et al., 2004). In other scenarios, abundance of exotic or invasive arthropods is directly related to area and age of native vegetation fragments surrounded by an urban matrix (Bolger et al., 2000). In the present study, detritivores particularly abundant in ruderal vegetation were Talitridae, Forficulidae, Armadillidae and Porcellionidae, which would be displacing native detritivores of family Tenebrionidae. In turn, phytophagous Apionidae, Curculionidae and Pyrrhocoridae would be taking advantage of exotic plants, and predators Dysderiidae and Staphylinidae would be feeding on new available prey.

The significant correlations between family abundances and scores of NMDS axes suggest associations between these taxa and disturbance, substrate and vegetation observed in each habitat type. If non-significant correlations but greater than 0.55 and previous ecological studies for Mediterranean island entomofauna (Español, 1958, 1965) are taken into account, four family-groups could be distinguished: 1) psammophilous, associated with mobile sands, including Carabidae, Myrmelontidae, Phalangiidae, Scarabaeidae, Sparassidae and Tenebrionidae, 2) synanthropic or eurytopic, associated with ruderal vegetation, including Apionidae, Armadillidae, Curculionidae, Dysderidae, Forficulidae, Porcellionidae, Pyrrhocoridae, Staphylinidae and Talitridae, 3) associated with semi-consolidated and consolidated sand, including Chrysomelidae, Cydnidae, Julidae, Lycosidae and Pyrgomorphidae, 4) generalists, not clearly associated with any particular condition, including Catantopidae, Gnaphosidae, Oligotomidae, and Thomisidae.

About family-groups, psammophilous would be potential indicators of dune conservation status, being taxa with a high degree of endemism and specialization in all desert ecosystems (Crawford, 1981). Families associated with ruderal vegetation included detritivores, phytophages and predators that took advantage of the increased productivity and with regard to Dysderidae, Forficulidae and Porcellionidae are the same ones recorded in urban-influenced North American desert ecosystems (Bolger et al., 2000; Longcore, 2003). Apionidae and Curculionidae associations with ruderal vegetation can be explained by greater availability and nutritional value of host plants, however these taxa have several dune-associated species that deserve further research in Alicante province (Velázquez de Castro & Martín-Cantarino, 1992). The classification of Catantopidae and Pyrgomorphidae as generalists coincides with spatial patterns observed in orthopteran assemblages from other southeastern Spain dunes (Clemente et al., 1985; Badih et al., 1997).

The significant correlations between species abundances and scores of NMDS axes suggest associations between these taxa and disturbance, substrate and vegetation observed in each habitat type. If non-significant correlations but greater than 0.55 and previous ecological studies for Carabidae (Sauleda-Pares, 1985a), Curculionoidea (Velázquez de Castro & Martín-Cantarino, 1992) and Tenebrionidae (Sauleda-Pares, 1985b; Martín-Cantarino, 1994) in Alicante province are taken into account, six species-groups could be distinguished: 1) psammophilous, associated with mobile sands, including Erodius carinatus Solier, 1834, Halammobia pellucida (Herbst, 1799), Harpalus fulvus Dejean, 1829, Masoreus wetterhalli Gyllenhal, 1813, Pachychila frioli Solier, 1835, Psammodius porcicollis (Illiger, 1803), Pseudosericius pruinosus (Dufour, 1820), Tentyria elongata Waltl, 1835 and Xanthomus pellucidus Mulsant & Rey, 1856; 2) synanthropic or eurytopic, associated with ruderal vegetation, including Leichenum pulchellum (Lucas, 1849), Macrothorax morbillosus (Fabricius, 1792), Ocypus olens (Müller, 1764) and Tentyria laevis Solier, 1835; 3) associated with semi-consolidated sand, including Cymindis lineola Dufour, 1820, Timarcha espagnoli Bechyné, 1948 and Timarcha intermedia Herrich-Schaeffer, 1838; 4) associated with heath dune, including Asida (Granulasida) ricoi Martinez, 1873 and Pimelia baetica Solier, 1836; 5) halophiles, associated with wet sands represented by Cicindela lunulata Fabricius, 1781; 6) generalists, not clearly associated with any particular condition, including Coniocleonus excoriatus (Gyllenhal, 1834), Licinus punctatulus (Fabricius, 1792) and Scarites buparius (Forster, 1771).

At species level, it should be noted that psammophilous darkling beetles are taxa particularly vulnerable to habitat loss, as has been observed in other urban influenced environments on Mediterranean coast (Fattorini, 2011; El Surtasi et al., 2012). However, some species with good dispersal abilities could survive in dune patches with relatively undisturbed soil and vegetation conditions (Comor et al., 2008). A similar ecological pattern can be expected for P. porcicollis, a small, detritivore, sand-digger and flying dispersal aphodine inhabiting dunes along Mediterranean coast (Kim & Lumaret, 1981). Regarding species of Timarcha genus, both were associated with grey and heath dunes, perhaps due to their strict food preference on some native host plant, as has been observed in other Iberian species (González-Megías & Gómez, 2001). Regarding species associated with patches of ruderal vegetation, M. morbillosus and O. olens were opportunistic predators favored by prey availability at disturbed sites. Likewise, O. olens coincides with the predicted traits for ruderal rove beetles in European urban-rural gradients, namely large body size, eurytopic and good migratory ability (Bohac, 1999).