Freshwater ecosystems are endangered worldwide by various human pressures, resulting in dramatic habitat and species loss. Many aquatic invertebrates respond to disturbances in their habitat, and mayflies are among the most sensitive ones. Therefore, we investigated mayfly response to anthropogenic disturbances at 46 study sites encompassing slightly to heavily modified small and mid-sized lowland streams and rivers. Mayfly nymphs were sampled between April and September 2016 using a benthos hand net. A total of 21 species was recorded, with Cloeon dipterum (Linnaeus, 1761) being the most frequently recorded one. Nevertheless, the taxa richness was rather low per site, i.e., between zero and nine. Assemblage structure had a high share of lower reaches and lentic (potamic and littoral) elements, and detritivores (gatherers/collectors and active filter feeders). This indicates that hydromorphological alterations lead to assemblage “potamisation” in small and mid-sized rivers. More mayfly species were related to higher oxygen concentration and lower water temperature, abundance of aquatic vegetation and total organic carbon. Additionally, the assemblage diversity and abundance were negatively associated with increasing intensive agriculture area at the catchment scale. This study confirms mayfly bio-indicative properties, i.e., their sensitivity to alterations of their habitat and pollution, but also provides new data related to mayfly response to the impacted environment. Those data can be used for management and protection activities of lowland rivers and their biota according to the requirements of the European Water Framework Directive.

Environmental stress, Ephemeroptera, feeding guilds, longitudinal zonal associations, pollution

Freshwater ecosystems represent an indispensable resource of water supplies for humans (Carpenter et al. 2011), but they also have a crucial role in biodiversity maintenance and conservation (Previšić et al. 2009; Ivković and Plant 2015). Therefore, it is essential they remain in good ecological status (Dudgeon et al. 2006; Vörösmarty et al. 2010). Nevertheless, the status of many aquatic systems is far from good worldwide (Carpenter et al. 2011). Various anthropogenic impacts represent major threats to aquatic biodiversity and make lotic habitats among the most endangered ones (Malmqvist and Rundle 2002; Hering et al. 2006; Stoddard et al. 2006). Human population growth, increased urbanisation and industrialisation have led to increased demands for land use for purposes of agriculture, forestry, irrigation activities and wetland drainage, resulting in alterations of habitat morphology, hydrological regime and causing degraded water quality, pollution and increased sediment erosion into lotic systems (Waters 1995; Dudgeon et al. 2006; Woodward et al. 2012). By altering their natural condition, such activities largely downgrade the habitat integrity, which results in reduced ecological function and biodiversity (Steffen et al. 2015), including native species loss (Carpenter et al. 2011). The habitat characteristics change dramatically: formation of macrophyte assemblages is disturbed (Jones et al. 2014; Turunen et al. 2017), habitat heterogeneity and availability for macroinvertebrates is reduced (Jones et al. 2012), while primary production (Louhi et al. 2017) and decomposition of organic matter (Lecerf and Richardson 2010) are highly altered.

As freshwater organisms live almost continuously in the aquatic environment, they clearly respond to all those environmental stresses (Morse et al. 2007; Vilenica et al. 2019; 2020). The aquatic assemblages can respond to alterations of their habitats with their structure differing from a reference state, i.e., they can show characteristics of “rhithralisation” (e.g., caused by channel straightening) or “potamisation” (e.g., caused by the impounding) (Jungwirth et al. 2000; Moog and Chovanec 2000; Kokavec et al. 2018; Vilenica et al. 2016; 2019), or there is a change in the trophic structure (Brasil et al. 2013). By observing the assemblages’ structural alterations, we can conclude that the lotic system has been altered, which in the end indicates a certain level of ecological disturbance (Moog 2002; Vilenica et al. 2016). Mayflies are able to colonise all kinds of freshwater habitats but are found to be the most diverse in lotic ones. They are among particularly sensitive aquatic macroinvertebrates, mainly disappearing when faced even with small-scale disturbance in their habitat (Firmiano et al. 2017; Vilenica et al. 2019). Previous studies demonstrated that the majority of species can tolerate a rather narrow range of environmental factors, being highly sensitive to oxygen depletion, acidification, and various contaminants such as metals, ammonia, nitrogen, phosphorous (Moog et al. 1997; Vilenica et al. 2017, 2019). Therefore, the absence/presence of a particular species can tell us a lot about the quality of the environment it inhabits. Ecological assessments in different regions worldwide, as well as at habitats of various ecological status are necessary for effective conservation and management of freshwater habitats and their biota (Hughes et al. 1986; Stoddard et al. 2008). Therefore, in order to obtain additional data on mayfly response to anthropogenic disturbances in their habitat, we investigated mayfly assemblages and their relationship with environmental factors at 46 slightly to heavily modified lotic habitats.

Our results indicate that a relatively high number of mayfly species can be found in anthropogenically impacted freshwater habitats. Nevertheless, at a large part of the study sites (i.e., 72 %) taxa richness was low, i.e., between zero and four taxa, corroborating previous studies (Vilenica et al. 2016; 2019). Mayflies inhabit both lotic and lentic habitats, although upper and middle reaches of fast-flowing streams, and ecologically intact large rivers harbour the highest mayfly diversity (Bauernfeind and Soldán 2012; Vilenica et al. 2016; 2018). Therefore, such low species richness, not typical for a lotic habitat (Bauernfeind and Moog 2000; Zedková et al. 2014; Vilenica et al. 2018), could be a consequence of various disturbances present at those sites, such as channelling, eutrophication, pollution, and microhabitat homogeneity (Axelsson et al. 2011; Carvalho et al. 2013; Ligeiro et al. 2013). In many cases, we observed shoreline erosion, as the emergent vegetation along the habitat edges, together with surrounding vegetation was mowed. This could have resulted in an increased input of sediments into the habitats, which could have influenced the habitat physico-chemical characteristics and hydrological cycle, resulting in reduced water quality and habitat heterogeneity (Mendes et al. 2017 and references herein). Consequently, these habitats showed to be less favourable for a high number of mayfly species. The majority of study sites were inhabited by widespread and generalist species (Popielarz and Neal 2007; Bauernfeind and Soldán 2012), yet sites with more microhabitat heterogeneity and higher water velocity, had also several microhabitat specialists, such as Baetis lutheri and Ecdyonurus torrentis for mesolithal, and Centroptilum luteolum as specialists for macrophytes (Buffagni et al. 2009; 2020).

The Zelina stream in Božjakovina (site 22) and Toplica River upstream from Daruvar town (site 36) showed somewhat higher species richness, yet their assemblages mainly consisted of species inhabiting a wide range of habitats, such as Baetis rhodani, Centroptilum luteolum, Serratella ignita and Caenis luctuosa (Buffagni et al. 2009; 2020; Bauernfeind and Soldán 2012). The most interesting finding was a record of Oligoneuriella rhenana at Toplica River, which is considered rare in Croatia (Vilenica et al. 2015; 2018). Although the species can tolerate some variations of environmental factors, its presence indicates that the ecological condition of Toplica River upstream from Daruvar town is not as poor as at the majority of other sites (Găldean 1999; Petrovici and Tudorancea 2000). Another interesting species was the rarest in our study, a riverine Heptagenia flava, uncommon in Croatian waters (Vilenica et al. 2015). Although the species was reported to have rather high ecological plasticity, usually it does not inhabit heavily polluted rivers (Vidinova and Russev 1997). Therefore, the species record at Česma River in Narta (site 28) could be considered as an accidental finding, as shown by Vidinova and Russev (1997). On the other hand, two eurytopic and euryvalent species (i.e., with wide tolerance towards the environmental conditions and habitat type), Cloeon dipterum and Serratella ignita, were recorded as the most common and the most numerous, respectively (Buffagni et al. 2009; 2020; Bauernfeind and Soldán 2012; Vilenica et al. 2019). Nevertheless, while discussing the total species richness at a particular site, we need to keep in mind that standardised sampling methods generally do not include sampling of underrepresented microhabitats, which could be important for some rare species (Haase et al. 2008). Therefore, in order to obtain a more complete species list, it might be beneficial to complement standardised quantitative sampling with a qualitative one.

Stream channelling is a widely used engineering practice designed for flood control and wetland draining, which affects the majority of hydrogeomorphological characteristics and processes at the channelled habitat. Due to these changes, the biota is also severely affected (Hupp 1992), i.e., the community structure and composition are changed and poorer (Waters 1995). Our results showed that mayfly assemblages have mainly consisted of taxa of potamic (lower reaches) and lentic preferences (e.g., Baetis buceratus, Caenis horaria) or wide range (e.g., Cloeon dipterum, Centroptilum luteolum, Serratella ignita) habitat type preferences (Buffagni et al. 2009; 2020; Bauernfeind and Soldán 2012). Moreover, Baetis vernus, Caenis luctuosa and Serratella ignita, species with relatively strong rhithral affinity (Biss et al. 2002) were predominantly associated with hydromorphologically less degraded sites, while species with more prominent potamic preference, such as Baetis buceratus and Baetis fuscatus (Schöll et al. 2005) were present at sites both with low and high degree of hydromorphological degradation. Some study sites showed a higher share of rhithral elements, yet that was mainly due to the dominance of eurytopic Cloeon dipterum (Buffagni et al. 2009; 2020). As the majority of sites are characterised by low microhabitat diversity, a high level of sedimentation and nutrients, assemblages were dominated by detritivores (Buffagni et al. 2009, 2020).

Previous researches showed that mayflies are highly dependent on specific environmental cues, and many species rapidly disappear when faced with anthropogenic disturbances in their habitat (Bauernfeind and Moog 2000; Goulart and Callisto 2005; Stepanian et al. 2020). Our results corroborate previous studies that showed negative responses of mayflies to high water temperature (e.g., Chadwick and Feminella 2001; Alhejoj et al. 2014) and low oxygen concentrations (e.g., Nebeker 1972; Lock and Goethals 2011). Sites that were characterised by high water temperatures were also often accompanied by low oxygen content and dense aquatic vegetation. High levels of nutrients in the water support such dense growth of vegetation, leading to a decrease of oxygen level (Boeykens et al. 2017). Moreover, the decay of organic matter (especially aquatic vegetation), together with bacterial growth, animal/human metabolic activity and various synthetic sources (such as pesticides, fertilisers, pharmaceuticals, detergents) lead to elevated concentration of total organic carbon (TOC) in water (e.g., Volk et al. 2002). A part of the TOC can be explained by the increased shoreline erosion due to management and clearing of vegetation in the shoreland zone, which probably also negatively affected mayflies in this study. Riparian buffers, especially undisturbed vegetated riparian zones situated adjacent to river and streams, can greatly mitigate nutrients, sediment from surface and groundwater flow through the processes of deposition, absorption and denitrification (e.g., Peterjohn and Correll 1984). Finally, the strong negative association of mayfly assemblages with intensive agriculture in the catchment area corroborates results of previous studies that showed high mayfly sensitivity to agricultural pollution (Siegloch et al. 2014; Zedková et al. 2015). Here, as especially sensitive showed Alainites muticus, Baetis lutheri and Oligoneuriella rhenana, species with low and moderate tolerance to water pollution (mainly occurring in oligosaphrobic and beta-mesosaphrobic waters) (Bauernfeind et al. 2002; Mihaljević 2011). In addition, another species was distinguished as sensitive to such kind of pollution, Baetis rhodani. Those results could come as a surprise, as this eurytopic mayfly has a wide ecological tolerance, and generally contributes as a major part of the macroinvertebrate biomass in many European streams and rivers (Elliott et al. 1988). Nevertheless, as Baetis rhodani is a species complex (Williams et al. 2006), those results should be inspected in more details, using molecular analyses. Our results confirm that water pollution is one of the largest limitation factors for the majority of mayflies (Van Dijk et al. 2013; Zedková et al. 2015).

This study contributes to our knowledge of mayfly relationship with environmental conditions in heavily modified and anthropogenic habitats. Various anthropogenic pressures resulted in changes in mayfly assemblage composition and structure, whereas species richness decreased. For instance, the assemblages consisted mainly of a relatively low number of widespread generalists and species characteristic for lower reaches and lentic habitats. This indicates that hydromorphological alterations could have resulted in assemblage’s “potamisation”. Moreover, highly polluted sites, with high temperatures and low oxygen content, were inhabited almost exclusively with the euryvalent Cloeon dipterum, or were completely unsuitable for any mayfly species, confirming the high sensitivity of mayflies to disturbances in their habitats. Our results can enable planning of management and conservation activities of lowland rivers and their biota according to the requirements of the European Water Framework Directive.

We would like to thank our colleagues from University of Zagreb (Faculty of Science, Department of Biology, Division of Zoology) for their indispensable assistance during the field investigations and for help with sorting the collected material. Miran Katar is thanked for helping us with the artwork. We thank Maja Kerovec for assisting with GIS analysis and Croatian Waters for providing hydromorphological, land use and water physico-chemical data. Finally, the reviewers are thanked for their useful comments and suggestions.