4.2 Influence of different life cycles on species and OTU detection
We hypothesized that species richness will differ depending on life cycle differences between mero- and hololimnic taxa. Accordingly, we found a strong effect of sampling season on all tested groups, except Coleoptera. Species richness of merolimnic and hololimnic taxa were both affected by sampling day but in a different way. In contrast to merolimnic taxa, hololimnic taxa showed a more stable number of species between autumn 2017 and spring 2018 likely due to the absence of emergence events, as hypothesized. For the most species-rich hololimnic group, Annelida, most species were detected in autumn, likely due to an increase of annelid populations and hence stronger eDNA signal in the water. Highest densities of aquatic annelid species are known for summer and autumn (Cook, 1969; Learner et al. 1978; Smith, 1986) with densities for the species Dero digita peaking in autumn (Smith, 1986) and in our data set the species was only detected once in October. Contrary to our second hypothesis, we did detect a lower species richness in summer also for the hololimnic group Annelida, but the decrease was more steady instead of an abrupt drop in species richness after peaking as for merolimnic species.
In general, merolimnic species detection in the water depends on several factors, for example time and duration of the flight period and emergence, abundance of the species in the water, number of generations per year, larval development and occurrence of dormancy/diapause. Diptera were the most species-rich merolimnic order with most taxa belonging to family Chironomidae. We detected a high diversity of Diptera in spring and early summer which is consistent with the results of other studies, implying a reduced diversity in winter and the highest diversity in early summer before emergence, as emergence of chironomids is correlated with high temperatures (Armitage et al., 2012; Bista et al., 2017). The high number of species in spring is likely due to a high growth rate preparing the larvae for emergence in summer, also explaining the drop in detected species richness for the later summer month. For the species‑rich mero- and hololimnic groups, except Annelida, no effect of sampling day on OTU richness was detected, possibly due the high number of OTUs per group which can show distinct responses to seasonal changes even within the same species. In contrast, for Ephemeroptera, Plecoptera and Trichoptera, only OTU richness was affected by sampling day, likely due to the lower numbers of species and therefore smaller shifts in species richness over time. Plecoptera and Trichoptera OTU richness showed one peak each, in winter and early spring, which is likely linked to an increase in biomass that will probably reach its peak right before emergence coinciding with our second hypothesis expecting a low richness in summer. Plecoptera OTU richness peaked a bit earlier than Trichoptera OTU richness which is congruent with an on average earlier emergence of many Plecoptera species compared to most Trichoptera species (Graf et al., 2008; Graf et al., 2009; Graf et al., 2022a,b; Schmidt-Kloiber and Hering, 2015). By comparison, Ephemeroptera OTU richness had two peaks, one in autumn and one in late spring. Species details revealed that from 21 Ephemeroptera species, ten have a bivoltine (mainly genus Baetis ), one a flexible (Caenis beskidensis ) and two a semivoltine (genusEphemera ) lifecycle (Buffagni et al., 2009; Buffagni et al., 2022; Schmidt-Kloiber and Hering, 2015). In contrast, the detected Plecoptera and Trichoptera were mostly univoltine species.
Occurrence even between species within a merolimnic order differed between seasons. The indicator species analysis revealed thatGlyphotaelius pellucidus is detectable in all winter and spring months and absent in summer which coincides with the species having a long flight period after emergence and a known dormancy (Graf et al., 2008; Graf et al., 2022; Schmidt-Kloiber, A. and Hering D., 2015). In contrast Philopotamus ludificatushas also a long flight period and dormancy but was detected less frequently but had similar read numbers as G. pellucidus (Graf et al., 2008; Graf et al., 2022; Schmidt-Kloiber and Hering, 2015). As neither differences between life cycle characteristics, nor read numbers were present, the less frequent detection of P. ludificatus was likely because the species was rarer at the sampling location or because of other unknown factors influencing the detectability of the species. Other species like the Ephemeroptera Baetis rhodani andBaetis vernus are detectable throughout the year with B. vernus being less frequently detectable than B. rhodani . The more frequent detection of B. rhodani can be explained due to the species being also sometimes trivoltine and B. vernus being only known for a bivoltine life cycle (Buffagni et al., 2009; Buffagni et al., 2022; Schmidt-Kloiber and Hering, 2015) and therefore, larvae ofB. rhodani are probably more frequent. Additionally the OTUs detected belonging to species B. vernus differed in their seasonal occurrence with one OTU only occurring in autumn. It is known that different OTUs can show distinct responses to environmental changes (Beermann et al., 2018), and for the B. vernus group cryptic diversity has been recorded (Ståhls & Savolainen, 2008). This strengthens the assumption that the differences in detection between the OTUs is based on different responses to environmental changes and therefore an underestimation of the diversity within B. vernus . A limitation in using eDNA to assess patterns of seasonality is the persistence of eDNA in the environment for up to several days or weeks, and apart from that, the uncertainty of capturing living or dead cells (Dejean et al., 2011; Pilliod et al., 2014; Thomsen et al., 2012a,b). Nonetheless, our results demonstrate that the patterns we found are consistent with the phenology of the different taxa thus further encouraging that the DNA we detected mostly originated from living organisms.