Our results showed that seedbanks and HZs are important refuges for some invertebrate taxa during drying, but their use could not be related to drying duration. We observed that the taxonomic and functional diversities of the studied aquatic refuges constituted a substantial portion of the benthic assemblages. Although the relative abundance of resistance traits was not higher in aquatic refuges than in benthos and the resistance forms were the less abundant resistance trait, we found that the capacity to move to the interstitial zone and small body size were the most important resistance traits in the HZ and benthos, while for the seedbank, the most important resistance trait was aerial respiration. Finally, we found a negative relationship between resilience traits and the duration of drying.
Taxonomic and functional diversities and contribution from aquatic refuges
Significant parts of the taxa and traits of the whole community were present in refuges. We found that diversity metrics were higher for benthos than for HZ and seedbank because the former includes the whole invertebrate community. The taxonomic and functional overlap between the aquatic refuges and the benthic assemblages (between 5 and 40%) indicated that some taxa with resistance traits used these refuges to avoid drying in situ. Conversely, the whole functionality of the benthic community could not be recovered after drying because the functional richness of the aquatic refuges did not represent all the functionality found in the benthos (only between 7 and 16%), although it is essential to the contribution of different resilience strategies (Chester and Robson 2011).
We found that Diptera, Oligochaeta, Coleoptera and Ephemeroptera were present in both aquatic refuges. The order Diptera contains a high number of species and variability in anatomical designs, ecological specialisations and life history strategies (Yeates and Wiegmann 2005; Wiegmann et al. 2011). For instance, the family Chironomidae – generalist taxa – was always one of the most abundant taxa in all zones jointly with Oligochaeta. Both taxa present desiccation resistance forms and the capacity to move into sediment, and some Oligochaeta species may have aerial respiration (Tachet et al., 2010). These results are in accordance with other studies that found that some taxa of both groups could be terrestrial and/or semiaquatic species, explaining why they resist drying conditions better, and they may be the only taxa able to survive long dry periods on seedbanks (6.8–8 months) (Tronstad et al. 2005; Chester and Robson 2011; Datry et al. 2012; Stubbington et al. 2016). Mayflies found in seedbanks, Procloeon spp. and Caenis spp., have eggs that are resistant to desiccation. Caenis spp., Ecdyonurus spp. and Baetis spp., found in the HZ, show the capacity of mayflies to move to the sediment (Tachet et al. 2010). All Coleoptera found in the HZ present the capacity to move to this refuge, whereas any coleopteran taxa of the seedbank (e.g., Nebrioporus spp.) present resistance forms (Tachet et al. 2010). These Coleoptera could survive in the dried streambed because they have aerial respiration and some desiccation tolerance due to their impermeable protective cuticle (Holdgate 1956; Pallarés et al. 2017).
We found some “exclusive” taxa from aquatic refuges. However, this exclusivity should be considered with caution because we only sampled the benthos once and some taxa could have hatched before or after the sampling time (e.g., Ecdyonurus has at least two generations per year and the adults emerge between April and June and between July and October; Rawlinson, 1939). In the same way, different flood events when the flow resumes – typical after summer drying under the Mediterranean climate – could limit the presence of some species or their abundances (Pires et al. 2000; Muñoz 2003).
Differences in resistance trait abundances
The HZ showed the lowest resistance trait abundances, but the most abundant resistance trait categories in HZ were body size and the capacity to move to the sediment, as we expected. Few taxa of the HZ (e.g., Ecdyonurus spp. or Hydropsyche spp.) were not burrowers or lived in the interstitial space, but all of them presented small sizes and were also crawlers. These trait combinations allow them to move to the aquatic refuge in the damp sediment and/or in the HZ following the moisture when the water flow decreases (Maazouzi et al. 2017; Vadher et al. 2017), depending on the sediment characteristics, the local hydrology (i.e., the location of downwelling and upwelling zones) and the abilities of the taxa (Stubbington 2012; Dole-Olivier et al. 2022).
Furthermore, seedbank and benthos presented similar resistance trait abundances, but the most abundant trait in seedbank was aerial respiration, partially in agreement with our initial hypothesis, as we expected the resistance form trait (resistance eggs or diapause) to show the highest abundance. Aerial respiration is favoured when flow contraction starts and the water conditions become harsher when the oxygen concentration in water decreases (Bonada, Rieradevall & Prat, 2007; Mellado-Díaz, Alonso & Gutiérrez, 2008).. Taxa with this characteristic are the last to disappear from the surface flow (Stubbington et al. 2017).
Although our second hypothesis was partially supported, some taxa found in aquatic refuges not only have the resistance traits we considered in this study but also others, and combinations of these traits might be associated with the use of aquatic refuges; as a consequence, we could have underestimated the relative abundance of total resistance traits. Moreover, traits are not independent, which is a known limitation of trait-based approaches (Resh et al. 1994; Martini et al. 2021), and may cause the appearance of an individual trait decoupled from the environmental conditions (trait syndromes; Poff et al., 2006), explaining the occurrence of unexpected trait categories.
Drying intensity and resistance strategies
Contrary to our expectations, resistance strategies were not related to drying intensity. We only found a negative relationship between resilience trait categories and the duration of drying, suggesting the loss of exclusively resilient taxa with an increase in flow intermittence. This means that drying “filtered” taxa without resistance traits (e.g., Tinodes spp. or Ancylus spp.), so these taxa recolonise the intermittent reaches only from upstream perennial reaches when flow resumes, implying that the abundance of resilience traits in perennial reaches must be higher than in intermittent reaches (Leigh et al. 2016) and explaining our result. Additionally, the positive relationship between the drying duration and the centroid positions in FS-Axis 1 (values between -0.10 and 0.11) implied a reduction in small taxa (such as Ephemeroptera, Trichoptera, Plecoptera and some Diptera), increasing the presence of larger Diptera, Oligochaeta and Coleoptera with drying (Figure 2), which show resistance strategies (discussed before).
In contrast, two of the most intermittent reaches of this study (in the Llobregat and Fluvià rivers) did not have water in the HZ, making it impossible for aquatic invertebrates to find refuge. The success of most resistance strategies decreases with an increase in the drying harshness conditioned by the presence of water (HZ or isolated pools) or moisture in the sediment (seedbank) (Bogan et al. 2017). The complete disappearance of the water may be part of the natural range of drying intensity in the reach, but almost all invertebrates in the reach are not adapted to these harsh conditions (Fritz and Dodds 2004). For instance, in seedbanks, a large number of taxa can show generalist resistance forms at moderate drying, but only a few specific taxa present suitable strategies at very high drying intensities, such as the use of cocoons and housing against desiccation (Crabot et al. 2021a). Although invertebrates in IRES are well adapted to flow intermittence, severe periods of drying strongly alter the composition and structure of invertebrate communities (Doretto et al. 2018; Chanut et al. 2022).
Since our hypotheses were not completely supported, future research is needed to extend our knowledge about the traits involved in resistance mechanisms and to evaluate the effect of drying intensity on the use of aquatic refuges. The current context of global change, increasing drying severity (extent and duration of drying; Datry et al., 2017), compromises the success of resistance strategies (Stubbington and Datry 2013; Delvecchia et al. 2022). Our results showed that the taxa recovered from aquatic refuges by resistance mechanisms could represent up to 40% of benthic invertebrate assemblages in some reaches, but the expected increase in drought severity and water abstraction activities compromise the water needed to maintain the HZ as well as sufficient moisture in streambed sediments. The importance of protecting the integrity of biodiversity recovery mechanisms in IRES is a priority. Management strategies should therefore aim to ensure the required amount of water to maintain these aquatic refuges during dry periods to preserve aquatic biodiversity.