Introduction

The establishment of large cities in a relatively short time has promoted some of the largest and fastest environmental impacts (Johnson and Munshi-South 2017). Cities refer to the built infrastructure and land use change associated with increasing human population density (Schmitt and Burghardt 2021). The massive growth of urbanization has driven one of the most substantial transformations in water bodies among other habitats for multiple reasons (Goertzen and Suhling 2019; Johansson et al. 2019; Meland et al. 2019). In general, urban streams are subjected to toxins, temperature change, siltation, organic pollutants, and the replacement of riparian vegetation for concrete, which limits rainfall infiltration, modifying flow regimes and increasing runoff and flooding risk (Bogan et al. 2020; Walteros and Ramírez 2020).

The impact of urbanization on different biological groups vary depending on factors related to each city or particular taxa characteristics. With respect to cities, their geographical location, type of original biome, age of establishment, size, human population density and type of economic activities in each one, can generate different selective pressures on organisms (Szulkin 2020). For example, in temperate zones, the effect of urbanization is the opposite to that of dessert biomes. While in temperate zones, urbanization leads to increases in temperature and generates the urban heat island effect, in dessert areas, cities may be cooler than their natural surroundings (during the day) (Imhoff et al. 2010).

Species can respond to the selective pressures of urbanization at a population or individual level (Harris et al. 2013; Salomão et al. 2020). The evolution of environmental tolerance depends greatly on certain aspects of population structure such as the mating system or dispersal strategy, and is a function of both the between-individual variance in environmental optima and the within-individual breadth of adaptation (Lynch and Gabriel 1987). At the individual level, urban alterations exert strong direct and indirect selection on morphological (e.g. body size and body shape) and behavioral traits (e.g. mating behavior and habitat selection) of aquatic macroinvertebrates (Langerhans and Kern 2020; Resende et al. 2021). Thus, it is important to highlight that species respond differently to the same type of disturbance (Le Gall et al. 2018).Therefore, revealing how each species or group of species responds and what are the traits that allow them to survive in urban environments is essential to be able to understand what are the factors that affect each biological group and what is the best way to preserve the biodiversity that inhabits the city.

Despite all negative impacts that cities impose on aquatic ecosystems, urbanization can also lead to the creation of novel anthropogenic water bodies, such as canals, stormwater runoff basins, and recreational ponds in urban parks (Davies et al. 2008; Simaika et al. 2016; Ngiam et al. 2017). Not surprisingly, these novel habitats have become important reservoirs of biodiversity and, consequently, provide key ecosystem services such as cooling from urban heat, water filtration of pollutants, carbon sequestration, materials for construction and food sources (Ngiam et al. 2017; Bogan et al. 2020; Walteros and Ramírez 2020). Due to the availability of different types of aquatic environments (natural vs artificial; lotic vs lentic) in and around cities, some groups of habitat and feeding generalist aquatic organisms, such as Hemiptera, Diptera or Odonata, could find suitable habitats in cities to survive and reproduce in this new ecosystems (Vermonden et al. 2009; Monteiro-Júnior et al. 2014; Luke et al. 2017).

Odonata (Anisoptera [or dragonflies], Zygoptera [or damselflies], and Anisozygoptera), are typically labelled as forest animals (Cordero-Rivera 2006). However, this tendency to inhabit forests seems to be an ancestral condition: families such as Aeschnidae, Libellulidae, and Coenagrionidae, have basal genera (e. g. Nannophlebia, Boyeria and Erythromma) that appear to be forest-dwellers while derived genera (eg. Anax, Pantala and Ischnura) occur mostly in open habitats. This indicates that in odonates, life in open habitats (like farms or cities), is a recent adaptation (Paulson 2006). Although some studies have detected high richness and abundance of Odonata in forests (Rith-Najarian 1998; von Ellenrieder 2000), open and drastically transformed ecosystems such as cities show surprisingly high diversity (Jeanmougin et al. 2014; Luke et al. 2017). A great diversity of species have been recorded in remnants of lakes or wetlands located within cities around the world (Craves and O’Brien 2013; Goertzen and Suhling 2015). These species, mostly belonging to the families Coenagrionidae and Libellulidae, are capable of rapidly colonizing new or restored environments and can maintain viable populations in relatively small habitat remnants.

In Odonata, body size (Rocha-Ortega et al. 2020), body coloration (Sanmartín-Villar et al. 2017; Leveau 2021), the degree of color sexual dimorphism (Tüzün et al. 2017), territorial or reproductive behaviors and habitat selection (Resende et al. 2021) bear an important role when dealing with challenges imposed by human-modified habitats. Open areas would be filtering species on the basis of their body size, resulting in the loss of larger odonate species (Rocha-Ortega et al. 2020). Likewise, habitat modification could determine fitness alterations in species with a fixed coloration-based strategies, due to the spectrum change of the background (Sanmartín-Villar et al. 2017). Finally, warm adapted and more habitat generalist species deal fairly well with conditions of a modified environment (Ferreras-Romero et al. 2009; Resende et al. 2021).

In order to understand how some traits enable dragonfly or damselfly species to inhabit cities, we have reviewed analytically the different studies on the impact of urbanization on Odonata communities worldwide. Although there are already some reviews on this subject (see: Willigalla and Fartmann 2012; Villalobos-Jiménez et al. 2016, and Deacon and Samways 2021), these have not identified which odonate species are tolerating urban environments and what morphological or behavioral traits are linked to such tolerance. We expect that the species that we identify as sensitive to urban environments have larger sizes, compared to species from environments less altered by human, since small body sizes will facilitate heat exchange between the insect and its environment (Sanborn 2005) avoiding the possible warming generated by heat islands in cities. Regarding coloration, polymorphic species with duller colorations are expected to dominate urban environments (Leveau 2021). Finally, we expect that species that are more sensitive to urbanization tend to exhibit more territorial-based behaviors and live in lotic or phytotelmata environments with greater canopy cover (Resende et al. 2021; Goertzen and Suhling 2019; Rocha-Ortega et al. 2020).

Methods

The data were compiled from a searching in Web of Science and Scopus databases focusing on the studies published between 1992 and 2022. The search criteria were “(urban* OR cit*) AND (Odonata OR dragonfl * OR damselfl *)”. These terms were searched in titles, abstracts, or key words. In addition to this general literature search, we examined the two journals that specialize in this order: Odonatologica and International Journal of Odonatology. The survey was conducted in August 2022. The selection process of scientific papers was performed in four steps: (1) identification: We searched for scientific papers in WoS and Scopus using the aforementioned keywords; (2) first screening: We then read abstracts to select scientific articles that evaluated the impact of urban infrastructure on Odonata Communities. In this step, we also removed duplicated papers; (3) second screening: we read in full all selected papers from the previous step to extract the necessary data to answer the questions of our study; and (4) selection of papers for the scientometric analyses.

Of the almost 400 articles returned by the search, after reading the title and the abstract, and checking the keywords, we removed 330 papers from our dataset as they did not meet the requirements established in our study (first screening). Thus, we ended up with 70 articles that evaluated the impact of urban infrastructure on Odonata communities and changes in odonata communities in an urbanization gradient. For each manuscript we seek to collect the following information: (1) authors and year of publication; (2) city and country where the study was conducted, (3) geographical scope; (4) taxonomic scope; (5) habitat type (lotic or lentic); (6) the species richness registered (7) life cycle phase included in the study and (8) the type of impact that was found (on the abundance of individuals, change in richness or species composition) (Supplementary material 1).

Additionally, we identified species sensitive and tolerant to urbanization based on the results of the reviewed articles and linked some of the species’ adaptations to urbanization. For the latter, we searched some morphological (body size, body color, strength of sexual dimorphism, polymorphisms by sex) and behavioral (territorial behavior, flight mode, type of mate guarding, and habitat preference) traits for identified species (Supplementary material 2). We collected this information from databases such as Odonata phenotypic database (Waller et al. 2019) and Encyclopedia of life (2022), and, when necessary, we looked for the information of each species individually. We used the non-information coding (NI) in cases where information for some of the traits were not found for each species.

Phylogenetic data

A phylogeny for all species in our database was obtained by using the informatics tool Phylomatic (Webb and Donoghue 2005) (Fig. 1). Phylomatic uses the Odonata mega-tree (Waller and Svensson 2017) as a backbone onto which species are added based on their taxonomy. For families missing in Waller and Svensson, (2017), we took into account the fossil dating dates proposed by Suvorov et al. (2022), while nodes with no estimated dates were estimated by bladj algorithm in Phylocom (Webb et al. 2008).

Fig. 1
figure 1

Phylogenetic tree assembled from the phylogeny hypothesis of Waller and Svensson (2017) with Odonata species identified as tolerant (gray) or sensitive (green) to urbanization

Full size image

Analysis

To test how traits influence the ability of odonates to live in urban environments, we used phylogenetic path analysis (Hardenberg and Gonzalez‐Voyer 2013). This analysis allows comparing causal hypotheses of the relationship among traits disentangling direct from indirect effects, while correcting for the non-independence of trait data due to common ancestry. Additionally, this approach deals with multicollinearity better than multivariate linear models, because partition the variance in the response among fewer predictors (Gonzalez-Voyer and von Hardenberg 2014).

In order to include the variables in the path analysis, we encoded each character as indicated in Supplementary material 3, assigning higher values to traits that might make species more sensitive to urbanization. We combine some of the variables to include them in the models. Habitat type and habitat openness together were considered as “Habitat preference”, strength of sexual dimorphism and polymorphisms by sex was considered as “Sexual Dimorphism” and type of mate guarding, flight mode and territorial behavior was considered as “Territoriality”. Then, we identified a taxon-specific model representing the relationships between body length, body colors, sexual dimorphism, territoriality, habitat preference and sensitivity to urbanization. In these models, we only considered significant paths, and we ensured that all conditional independencies (i.e., non-significant relationships between non-linked variables) were met (Gonzalez-Voyer and von Hardenberg 2014). To define the trait-only model for each suborder, we tested specific directional relationships based on a priori knowledge and expectations derived from published information. Phylogenetic path analysis models were built and tested in RStudio v. 1.1.456 (R Core Team 2018) using ‘phylopath’ package (van der Bijl 2018).

Results

A total of 70 papers related with the effect of urbanization on Odonata were considered (see Supplementary material 1): 21 were carried out in Europe, 20 in North America, 10 in Asia, 8 in Africa, 8 in South America, 2 in Oceania and 1 in Central America (Fig. 2). Thirty were carried out in lentic environments, 25 in lotic environments and 15 in lotic and lentic environments; 41 were performed using adults, 3 using adults and exuviae, 12 using adults and larvae, 1 using exuviae, 12 using larvae and 1 using larvae and exuviae (Supplementary material 1).

Fig. 2
figure 2

Number of works on the effect of urbanization in Odonata carried out in each continent

Full size image

We identified 88 urbanization-tolerant species (Table 1) and 87 urbanization-sensitive species (Table 2, Fig. 1). The most common families reported as tolerant to urbanization were Aeshnidae, Libellulidae and Coenagrionidae (see Table 1), whilst the most common families reported as sensitive to urbanization were Lestidae, Calopterygidae, Gomphidae, as well as some specialist aeshnids and libellulids (see Table 2).

Table 1 Odonata species tolerant to urbanization
Full size table
Table 2 Odonata species sensitive to urban stressors
Full size table

The phylogenetic path analyses indicated that some traits influenced the use urban areas for dragonflies and damselflies species (Fig. 3). The model obtained for Anisoptera explained 41.7% of variation (Table 3), while the model obtained for Zygoptera explained 44.8% of variation (Table 4). For both suborders, habitat preference was not linked to the sensitivity to urbanization, whereas larger species with weak sexual dimorphism stood out as consistently associated with sensitivity to urban environments (Fig. 3).

Fig. 3
figure 3

Average causal models obtained for relationship between species traits and sensitivity to urbanization. Blue indicates positive relationships, red indicate negative relationships. Values represent standardized path coefficients reported in Tables 3 and 4. BL body length, HP habitat preference, BC body colors, SD color sexual dimorphism, T territoriality, U sensitivity to urbanization

Full size image
Table 3 Standardized path coefficients (Coeff.) with standard errors (SE) between significant explanatory variables of Anisoptera species estimated through phylogenetic path analysis. Coefficients indicate whether significant relationships between variables have a positive or negative trend
Full size table
Table 4 Standardized path coefficients (Coeff.) with standard errors (SE) between significant explanatory variables of Zygoptera species estimated through phylogenetic path analysis. Coefficients indicate whether significant relationships between variables have a positive or negative trend
Full size table

Furthermore, it was found that different traits indirectly influence the species’ sensitivity to urban areas in each suborder. For dragonflies, body size is negatively related to body color, in turn, body color is positively related to the strength of sexual dimorphism and negatively to sensitivity to urbanization (Fig. 3). That is, although there is no direct relationship between body color and sensitivity to urbanization; smaller, yellow–red colourful, monomorphic species tend to be less sensitive to urban areas. Finally, for dragonflies, the territoriality of the species was not related to urbanization sensitivity.

For damselflies, territoriality is positively related to the strength of sexual dimorphism which is negatively related to sensitivity to urbanization (Fig. 3). Thus, with higher territoriality is greater sexual dimorphism and the species are less sensitive to urbanization. Nevertheless, the relationship between territoriality and sexual dimorphism is indirect to the sensitivity to urbanization. Finally, for damselflies, body color was not related to urbanization sensitivity.

Discussion

We found that some traits prompt species to persist in urban environments. According to our expectations, in the studied odonate species, sensitivity to urbanization is linked to morphology. Sensitive species having both a larger sizes and weaker sexual dimorphism, compared to the more tolerant species. This result coincides with previous studies on Odonata (Brito et al. 2021; Rocha-Ortega et al. 2020) and other insects (Shahabuddin and Ponte 2005; Theodorou et al. 2021), where species with larger individuals are the most vulnerable ones in the face of environmental disturbance. Likewise, species with weak sexual dimorphism were identified as more sensitive to urbanization. This is also consistent with previous studies where changes in the environment (e.g. loss of vegetation) could determine fitness costs in monomorphic species, due to the spectrum change of the background (Sanmartín-Villar et al. 2017).

Moreover, contrary to our expectations, we found that for dragonflies, duller colorations were indirectly associated with sensitivity to urbanization, while yellow and red colors were more associated with tolerance to urban areas. In damselflies, no relationship was found between body color and sensitivity to urbanization. In this latter suborder, we found an indirect relationship between territoriality and sensitivity to urbanization, with less territorial species being more sensitive to urbanization. In dragonflies, no such relationship was found for these last variables. Finally, we did not find any association between habitat preference and sensitivity to urbanization.

Why would larger species be more vulnerable in urbanized environments? Possibly, smaller species can quickly recolonized empty source and sink patches inside cities, whereas that building up a large body requires long developmental periods, more food and, consequently, there are more risks associated to such demands (Suhonen et al. 2014, 2022). Other explanation is that having a greater dispersal capacity, larger species search habitats with better conditions around cities, implying a reduced chance to be recorded within urbanized areas (e.g. Prescott and Eason 2018). Regarding sexual color dimorphism, previous studies in Lepidoptera (Franzén et al. 2020) and Odonata (Sanmartín-Villar et al. 2017) have reported that dimorphic or polymorphic species may have an advantage in urban environments with visual heterogeneity created by light pollution and non-native plants. The presence of more than one color morph in the species could facilitate their adaptation to the city, considering the changing environment in which they will have to hide from potential prey or predators.

Different types of stressors influence different species inside an order (e.g. Luke et al. 2017; Seidu et al. 2018; Huikkonen et al. 2019; Jere et al. 2020). In relation to this, we found that the traits identified in species that are tolerant to urbanization are different for dragonflies and damselflies. In dragonflies, urban species tend to have red and yellow colors. The open habitats that occur in cities could promote these reddish colors in dragonflies. As for the mechanism of color production, ommochromes are responsible for red and yellow coloration, and have been associated to antioxidant production (McGraw 2005), and UV ray protection (Needham 1974; Cooper 2010). For example, in Megalagrion calliphya a higher frequency of males and females with red coloration was found in environments with higher solar radiation due to the fact that both sexes acquire coloration to protect themselves from solar rays (Cooper 2010). Although this is reported in a damselfly species, the mechanism that may be behind our finding for dragonflies may have a similar basis to that which explains the maintenance of the red color in places with high solar radiation. In this way, the red and yellow dragonflies are more likely to endure the higher incidence of solar radiation typical of cities, through the antioxidant and protective function of the cuticle-embedded ommochromes.

Contrary to what we expected, urban damselflies tend to present more territorial behaviors compared to their counterparts in conserved sites. In Odonata, the selection and defense of a territory is determined by the presence of optimal sites for reproduction and ovipositing (Kohli et al. 2020). Our results could indicate that the territorial behaviors in damselflies become more important within cities, where there may be fewer available sites for couples to copulate or for females to oviposit, therefore those few available sites are defended more intensively. Results similar to this have been reported for birds, which defend their territories more vigorously in cities than their rural counterpart (Davies and Sewall 2016).

Although the results obtained in this study allow us to have a first approximation of which traits allow Odonata species to adapt to cities, it is important to consider that the lack of information for some of the evaluated traits is a limitation of our study. In addition, we include information for 175 species of the nearly 6400 Odonata species described so far, and we included traits for adults and not for larvae of the different species. Therefore, continuing to generate information about the response of these species to urbanization is essential to confirm the trends we detected.

There is also a limitation in terms of the regions of the world that have been most studied in this topic. Within the papers we reviewed, most studies were carried out in Europe and North America which are places with a large history of urbanization, yet they are less diverse than the tropical region. This finding implies a bias in our results, since we were not able to control for differences in urbanization times, size, or geographic location of cities. For instance, due to physiological adaptations responding to its evolutionary history, species in subtropical and temperate zones might have stronger responses to urban heat islands than species in tropical zones (Langerhans and Kern 2020). In damselflies species, it has been reported that forests can become even more important at lower latitudes so that species could be more sensitive to urbanization there, compared to high latitude species (Paulson 2006). To correct such bias, we suggest that the impact of urbanization on dragonfly and damselfly diversity be assessed with greater emphasis on species-rich areas such as Asia, Africa, and South America.

Given that urbanization is a phenomenon that grows by leaps and bounds every day, we must understand their ecological implications immediately. It is important to know odonate responses since they are an important link in aquatic food webs (Knight et al. 2005) and export a large part of aquatic productivity to terrestrial environments (Popova et al. 2017), both very important functions for the functioning of ecosystems. To know the dynamics of these aquatic insects in cities (transformed environments with little possibility of returning to their initial state) allow us to predict future distribution trends and consequently design appropriate conservation strategies focused on this group and applicable to other groups of aquatic insects. In this regard, management of urban freshwater habitats should account not only the aesthetic part but ecological needs to benefit people and biodiversity, two fundamental components within cities (Ngiam et al. 2017; Goertzen and Suhling 2019; Huikkonen et al. 2019). Additionally, we should be able to recognize biodiversity hotspots where future anthropic activities should be avoided, and to continue building the necessary theory to propose the most suitable conservation plans for our beloved odonates: creatures of the forest and the city!