Ground beetles in city forests: does urbanization predict a personality trait?

Introduction

Urbanization leads to substantial changes in natural habitats (McKinney, 2006) with profound effects on wildlife, from community level (e.g., Blair, 2001; McKinney, 2008; Rodewald et al., 2014) to within-species level. Within species, differences along rural–urban gradients have been detected in morphology (e.g., Weller & Ganzhorn, 2004; Giraudeau et al., 2014), physiology (e.g., Partecke, Schwabl & Gwinner, 2006), life history (e.g., Beck & Heinsohn, 2006; Rodewald & Shustack, 2008), and behaviour (e.g., Martin & Réale, 2008; Lehrer & Schooley, 2010; Sol, Lapiedra & Gonzalez-Lagos, 2013; Dahirel et al., 2017).

Behaviour is often the first response to altered conditions (Sih, Ferrari & Harris, 2011; Tuomainen & Candolin, 2011) with potentially large consequences on individuals (Sol, Lapiedra & Gonzalez-Lagos, 2013), population dynamics and biodiversity (Tuomainen & Candolin, 2011). Therefore, understanding individuals’ behavioural responses to urbanization is important for identifying which individuals or species may adapt to changing conditions or may disappear from altered habitats. Some behavioural traits, such as high exploration, high boldness and high aggressiveness, might be especially suited for coping with novel conditions (Sih, Ferrari & Harris, 2011; Phillips & Suarez, 2012; Lowry, Lill & Wong, 2013) and/or for colonizing urbanized habitats (Edelaar & Bolnick, 2012). Indeed, vertebrate urban populations are on average bolder (e.g., Prosser, Hudson & Thompson, 2006; Evans, Boudreau & Hyman, 2010; Uchida et al., 2016), less vigilant (Chapman, Rymer & Pillay, 2012), more aggressive (Evans, Boudreau & Hyman, 2010; Scales, Hyman & Hughes, 2011) and more exploratory (Martin & Réale, 2008) than their rural conspecifics.

In contrast to vertebrates, behavioural responses of invertebrates to urbanization are largely unknown. In general, work on urbanization is biased towards birds and mammals (Magle et al., 2012; Beninde, Veith & Hochkirch, 2015) and there was a call for considering other taxa (Beninde, Veith & Hochkirch, 2015), especially arthropods (McIntyre, 2000). Arthropods, and insects in particular, are important contributors to ecosystem functions (e.g., Mulder et al., 1999; Yang & Gratton, 2014; Griffiths et al., 2016). Thus, impacts of anthropogenic change on arthropods might have substantial consequences for whole ecosystems. In addition, some work suggests that birds and insects might respond differently to urbanization (Beninde, Veith & Hochkirch, 2015) which may be due to different ecological requirements and life histories. The vertebrate species (e.g., birds, squirrels, lizards) tested for behavioural differences across urbanization gradients are noticeable due to their size and/or behaviour, for instance to people walking in parks. Consequently, these species are likely not only indirectly (e.g., by habitat fragmentation or light pollution) but also directly (e.g., by supplemental feeding, harassing, unintentional disturbances by dogs or walking humans) influenced by human activities (following categorisation for (in)direct human effects by Clucas & Marzluff, 2011). In contrast, most invertebrate species (and some vertebrates) have a concealed lifestyle (e.g., they are small and/or hidden in vegetation or litter layer) which reduces direct contact with humans. As these less visible species should be mostly indirectly influenced by human activities, behavioural responses to urbanization might be different compared to openly visible mammals or birds. For instance, species with a concealed lifestyle may be influenced by urbanization if environmental conditions that change with urbanization (and that are not direct human interference) trigger behavioural responses.

Potential environmental drivers facilitating population differences in behaviour along urbanization gradients have been largely neglected. Abiotic drivers could include, amongst others, temperature, exposure to chemicals or humidity conditions. For instance, it is often warmer in more urbanized areas (Pickett et al., 2001) and temperature influences behaviour, particularly of ectothermic organisms. Biotic drivers may include population density and intraspecific competition, altered prey items, pathogens or unfamiliar predators. For example, urbanization often influences abundance (e.g., Shochat et al., 2006). Population density, in turn, affects behaviours such as exploration-activity (Le Galliard, Paquet & Mugabo, 2015), aggression (Cubaynes et al., 2014), sociability (Le Galliard, Paquet & Mugabo, 2015), foraging (Mobæk et al., 2012) and dispersal (Tuda & Shima, 2002).

Here, we studied effects of urbanization on ground beetle behaviour. Ground beetles are suitable invertebrates for urbanization studies as they depend heavily on environmental factors, such as temperature, humidity or food availability (Lövei & Sunderland, 1996; Rainio & Niemelä, 2003). Consequently, they are known to react quickly to environmental change, making behavioural responses to urbanization likely. Urbanization has already been shown to affect carabids, such as their species composition and abundance (Weller & Ganzhorn, 2004; Sadler et al., 2006; Niemelä & Kotze, 2009; Magura, Lovei & Tothmeresz, 2010; Davies, Bennie & Gaston, 2012; Kotze et al., 2012).

We expected more exploratory and high-risk taking ground beetle individuals to cope better with new challenges and new niches and/or to be more prone to colonize urbanized areas. We predicted this to result in a higher proportion of consistently more exploratory or risk-prone individuals in urbanized compared to rural areas, leading to increased population averages in these behaviours in more urbanized areas. Consistent individual differences in behaviour (’animal personality differences’ sensu Dall, Houston & McNamara, 2004) have been recorded in many species (Gosling, 2001), recently also including invertebrate taxa (Kralj-Fišer & Schuett, 2014).

In a two-year field study we tested ground beetles from differently urbanized forests for differences in their exploratory and risk-taking behaviour. We considered temperature and population density as potential environmental drivers facilitating behavioural differentiation between populations living under different levels of urbanization. We studied individuals of four forest-dwelling ground beetle species from eight within-city forest patches of two different urbanization levels (four populations per level). The four ground beetle species differ in their dispersal abilities and niche widths (cf. Lindroth, 1985/6; Turin, 2000). In the second year, we additionally tested individuals of two species repeatedly for their exploratory and risk-taking behaviour in the laboratory to assess consistent personality differences.

Material & Methods

Results

Predictors of behaviour in the field

In three species the number of square visits in the field was linked to the urbanization level or the interaction between urbanization and year; a non-significant trend for an effect of urbanization was also found in the fourth species (A. parallelepipedus) (Table 2; Fig. 1). In line with prediction, individuals mostly showed more square visits in more urbanized than in less urbanized sites: this was generally true in 2015 but in 2016 patterns were less clear (significant interaction between urbanization and year in N. brevicollis and P. oblongopunctatus). The number of square visits was temperature-dependent for all species (Table 2) and either increased with rising temperatures (A. parallelepipedus and C. nemoralis) or showed a temperature optimum (N. brevicollis and P. oblongopunctatus). Sexes differed in their behaviour: in all species, except for N. brevicollis, females had less square visits than males (Table 2). Population densities did not predict the number of square visits in any species (Table 2).

Discussion

In line with prediction, ground beetles were more exploratory (i.e., showed more square visits) in highly urbanized forests compared to less urbanized forests (three out of four species) but for those two species studied over two years, effects were year-dependent. Exploratory behaviour was not influenced by population density but temperature-dependent in all and sex-dependent in three species. Individuals consistently differed in their exploratory behaviour over time, i.e., personality differences existed among individuals (tested in two species). Individuals were generally consistent in their exploratory behaviour across tests in the field and in the laboratory, suggesting behaviour tests in the laboratory can be used to predict behaviour when tested in the field. Exploratory individuals were also bolder (less likely to show thanatosis; Supplemental Information) in three species.

Our results that ground beetles vary in their behaviour between differently urbanized areas (at least under certain conditions) are in line with the few other existing studies that tested effects of anthropogenic change on invertebrate behaviour. Insects varied in their behaviour as response to traffic noise (Lampe et al., 2012; Lampe, Reinhold & Schmoll, 2014; Orci, Petroczki & Barta, 2016) or reacted differently towards chemicals when originating from urbanized as compared to rural sites (Tuzun et al., 2015). However, there are still too few studies to draw general conclusions about behavioural responses of invertebrates to urbanization. It is possible that such responses are generally less pronounced in invertebrates and therefore less often published compared to responses in vertebrates. Moreover, year-dependent effects of urbanization on behaviour—as revealed in our study—point towards more complex patterns where other environmental factors interact with urbanization. For a deeper understanding of behavioural responses of invertebrates to urbanization more long-term studies are necessary. Nevertheless, results reported here were quite similar across species even though our study species differed in their niche width, dispersal ability and reproductive season. This indicates that the patterns we found might be more general, potentially even covering many other forest-dwelling arthropods.

The factors that could contribute to behavioural differences in less and highly urbanized areas are manifold (e.g., Tuomainen & Candolin, 2011). We here assessed the possibilities that differences in temperature or population densities could lead to behavioural differences in differently urbanized areas. While we reported temperature effects on behaviour, temperature during tests did not differ between less and highly urbanized forests. Contrary to our findings, temperatures are often higher in urban than rural areas (Pickett et al., 2001) but this might be less pronounced in city forest patches of different urbanization or during certain times of the year. Furthermore, in only one species the population density was higher in highly urbanized areas and we generally detected no effect of density on behaviour. Other factors, that could explain differences in behaviour of ground beetles between the differently urbanized sites may, for instance, include litter layer depth (Marcus et al., 2015) or food availability. The litter layer is the primary habitat for ground beetles and their prey; and litter decomposition rate increases along urbanization gradients (Pouyat, McDonnell & Pickett, 1997), resulting in reduced litter depth (Van Nuland & Whitlow, 2014). Moreover, activity of ground beetles is reduced with higher food supply (Lenski, 1984). Thus, if reduced litter layers (or other environmental factors) in more urbanized areas lead to reduced availability of prey, increased locomotory activity and exploratory behaviour could be beneficial for the ground beetles for locating food. Exploratory or bold behaviour might also be generally favourable in these forests: urbanization often leads to changes in the structure of (invertebrate) assemblages, including prey items of ground beetles (Steinberg et al., 1997; Van Nuland & Whitlow, 2014; Bogyo et al., 2015), such that beetles may encounter novel prey items. Indeed, abundance of non-native earthworms increased with urbanization (Steinberg et al., 1997), with earthworms being common prey items of carabids (cf Turin, 2000).

Year-dependent effects of urbanization on exploratory behaviour suggest that other environmental conditions, which varied between years, interacted with the effect of urbanization on beetle behaviour. Our results are corroborated by one of the few studies that tested temporal effects of urbanization: arthropod biodiversity changed with urbanization but differently through time (Van Nuland & Whitlow, 2014). Environmental conditions that interact with effects of urbanization could be changes in food availability, soil moisture or depth of the litter layer. If, for instance, litter layers had been thicker in 2016, this could have led to a reduced effect of high urbanization on behaviour (e.g., because prey may be more abundant in thicker layers, potentially reducing the need to explore). Future studies should consider potential environmental variables that might interact with effects of urbanization on studied traits. For testing such interactions, more long-term studies are urgently needed.

Exploration was temperature-dependent. While this result is not surprising given that poikilotherms crucially depend on surrounding temperatures, it highlights how important it is to consider ambient temperature in behavioural and urbanization studies. This is particularly true for field studies, where temperature conditions are not standardized. Temperatures are often higher in urban than rural areas (Pickett et al., 2001). Given that the investigated beetles reacted to increased temperature with more exploratory behaviour, it is possible that different temperatures across the differently urbanized forests caused different behavioural levels among populations. Yet at least during behavioural testing we did not find temperature differences between less and more urbanized forests. Nevertheless, future studies should further investigate whether temperature differences between urban and rural areas can drive behavioural responses to urbanization in poikilotherms.

Reported behavioural differences at population level could arise if individuals (a) react flexibly to altered environmental conditions and/or (b) if individuals consistently differ in their behaviour (i.e., show animal personality differences, sensu Dall, Houston & McNamara, 2004) (Sih, Ferrari & Harris, 2011; Brown, 2012; Lowry, Lill & Wong, 2013; Sol, Lapiedra & Gonzalez-Lagos, 2013). In our study we found consistent personality differences in exploratory behaviour in the two species that were also tested in the laboratory (N. brevicollis and P. oblongopunctatus). Even though we cannot conclusively differentiate between the above mechanisms (and both mechanisms might work at the same time), our data provide some valuable first insights. Behaviour was repeatable and/or (rank-) consistent for at least a few days after individuals had been taken out of their natural environment and kept under standardized conditions in the laboratory. These results may suggest that personality compositions varied between differently urbanized areas, or that individuals reacted differently to the environments (I × E), or that influences of environmental conditions on behaviour still carried-over when conditions changed. However, in the latter case we would not expect to find repeatable behavioural differences among individuals within populations in sites of similar urbanization (assuming homogenous environmental conditions). Rather, individuals should be all similar in their response, leading to low repeatability, despite high individual consistency. Tentatively, our results seem to suggest that different personality types are adapted to different environmental conditions with consistently more exploratory individuals being favoured in highly urbanized sites. Such behavioural changes in urbanized areas might lead to altered population dynamics and potentially even modified community structure. For instance, more exploratory individuals are often also higher risk takers (e.g., Van Oers et al., 2004; this study), more aggressive (e.g., Verbeek, Boon & Drent, 1996; Schuett, Dall & Royle, 2011) and may often follow different life-history strategies than less exploratory individuals (Réale et al., 2010); this likely influences population dynamics.

We showed that ground beetles exhibit consistent personality differences in both, the reproductive (P. oblongopunctatus) and non-reproductive season (N. brevicollis). Males and females of both species showed personality differences in exploratory behaviour and higher exploration was associated with higher risk-taking in both sexes for three species. In contrast, sexes differed in their mean levels of exploration: with the exception of N. brevicollis, males of all species were more exploratory than females. This may suggest that sexes vary in their reaction towards urbanization. In addition, all species, except N. brevicollis, were measured during their reproductive season. Thus, males may generally be more exploratory and/or active during the reproductive phase in order to locate mates for reproduction. Higher trapping rates of males compared to females during the reproductive phase support this idea (Drees & Huk, 2000; Weber & Heimbach, 2001).

Supplemental Information