Introduction
Aquatic and terrestrial ecosystems are connected through fluxes of resources (e.g. carbon and nutrients) and organisms, mediated by physical forces (water, wind) and dispersal, as part of a meta-ecosystem (Loreau et al. 2003). River networks are typical meta-ecosystems, in which resources are exchanged across longitudinal (i.e. from headwaters to downstream sections) and lateral (i.e. between riparian and instream habitats) dimensions (Gounand et al. 2018, Harvey et al. 2020, Cid et al. 2022) and transformed by organisms through ecosystem functions such as primary production or decomposition (Hotchkiss et al. 2015, Tiegs et al. 2019). Ecosystem functioning and community composition naturally change along the river network as the basal trophic resource switches from brown – i.e. based on coarse detrital organic matter riparian input – to green – i.e. based on instream algal production – from upstream to downstream (Vannote et al. 1980, Finlay 2001). Although the structuring of resources, communities and their functioning along the river network have been theorized > 40 years ago (Vannote et al. 1980, Gounand et al. 2018), limited empirical evidence of the mechanisms driving such structuring exist, apart from simulations (Jacquet et al. 2022a).
Leaf-litter is a key source of organic carbon and nutrients in rivers, exchanged in both lateral (through leaf fall and runoff) and longitudinal (up to downstream flow) dimensions (Catalàn et al. 2022, Scherer-lorenzen et al. 2022). The decomposition rate of leaf litter in both aquatic and terrestrial habitats depends on several factors including, leaf chemical quality and palatability (Pastor et al. 2014, del Campo et al. 2021a), environmental factors such as humidity and temperature (Tiegs et al. 2019, Annala et al. 2022) and decomposer community (i.e. microorganisms and invertebrates) composition and activity (Abelho and Descals 2019, Boyero et al. 2021). Although leaf-decomposer (shredder) invertebrate communities typically differ between instream and riparian habitats, terrestrial and aquatic decomposers may colonize riverbed and riparian habitats during streambed drying (Corti and Datry 2016) and overbank flooding events (Hutchens and Wallace 2002, Steward et al. 2022), respectively. Thus, lateral movement of consumers could link aquatic and terrestrial exchanges of organic matter (Abelho and Descals 2019, Barthélémy et al. 2022, Scherer-lorenzen et al. 2022), but little evidence of such links exists at the river network scale (Harvey et al. 2020). As a result of decreasing leaf input from headwaters to mainstem, decomposer abundances and their resource use also typically decrease from up to downstream (Collins et al. 2016, Jacquet et al. 2022a). However, dispersal limitation mechanism preventing species from reaching their favoured habitat or resource because of low connectivity, could lead to sub-optimal functioning in isolated headwaters, despite abundant resources (Leibold et al. 2017, Thompson et al. 2017).
Drying, i.e. the loss of surface water, is one of the most fundamental types of disturbance in river ecosystems, triggering drastic changes in environmental conditions that induce shifts in community composition (Boulton 2003, Datry et al. 2014) and carbon-related biogeochemical processes (Datry et al. 2018). By fragmenting the river continuum, drying can also alter the transport and dispersal of resources and organisms, limiting exchanges between populations (i.e. metapopulations), communities (metacommunities) and ecosystems across the network (Gauthier et al. 2020, Cid et al. 2022). Typically, the diversity and abundance of aquatic organisms such as invertebrates decreases as drying frequency and duration increase (Datry et al. 2014, Soria et al. 2017, Sarremejane et al. 2020) but the colonization of dry riverbeds by terrestrial organisms can counteract those diversity losses (Corti and Datry 2016, Steward et al. 2022). As a result of the negative effect of drying on shredder invertebrates, decomposition rates are usually lower in rivers that experience drying, particularly during their dry phase (Corti et al. 2011) but also during their flowing phase due to legacy effects (Datry et al. 2011, del Campo et al. 2021b). Fungal and bacterial decomposition may however be little affected by drying, if sufficient moisture preserves microbial activity (Riedl et al. 2013, Foulquier et al. 2015). Drying may also affect leaf-litter quality, altering labile carbon contents and nutrient immobilization by microbial communities, which often reduce palatability for invertebrates (Pastor et al. 2014, del Campo et al. 2021a). At least 50% of the world’s river networks dry, on average, for one day per year due to geological and/or climatic features (Messager et al. 2021), and this proportion is likely to increase due to climate change and other anthropogenic alterations (e.g. water abstraction, irrigation; Datry et al. 2023). We urgently need to understand how fragmentation by drying modifies spatial and temporal interactions between organisms and resources at the river meta-ecosystem scale, including across aquatic-terrestrial boundaries (Scherer-lorenzen et al. 2022) to be able to predict when and where adaptive management strategies may be needed to preserve functional river ecosystems under global changes.
Drying and fragmentation may create mismatches between resource and consumer availability if drying and low connectivity prevent efficient aquatic consumers from being present and use resources (Thompson et al. 2017). Such mismatches could result in changes in the biodiversity – ecosystem functioning (BEF) relationship as communities among highly disturbed or isolated environments may not be composed of the most efficient consumers; hence translating into a weak or null BEF relationship across networks fragmented by drying (Brose and Hillebrand 2016, Leibold et al. 2017). However, intermediate levels of disturbances and connectivity may strengthen the BEF by 1) allowing selection of the species best adapted to the environment and resource use and 2) preventing the dominance of functionally inefficient, highly competitive species occurring under low levels of disturbance and high connectivity (Cardinale and Palmer 2002, Leibold et al. 2017).
Here, we aim to examine how resource quality and quantity, the structure of consumer communities and ecosystem functioning respond to drying across lateral and longitudinal dimensions. To do so, we monitored across three seasons leaf resource stocks, invertebrate communities and decomposition rates in the instream and riparian habitats of 20 sites in a river network fragmented by drying. We hypothesize (H1, Fig. 1) that increasing flow intermittence promotes accumulation of leaves of low chemical quality (low conditioning by microorganisms), decreases in invertebrate abundance and diversity and thus lower decomposition rates instream. Instream responses to drying may also change with connectivity and effects may be stronger in less connected headwaters that typically have lower diversity and post-drying recolonization capacity, likely reducing functioning rates there (H2, Fig. 1). Drying may strengthen relationships between instream and riparian resource community and functioning due to homogenization of environmental conditions among these two habitats as the riverbed dries (H3, Fig. 1). Relationship between riparian and instream communities and functioning may also be stronger in isolated headwaters where riparian and instream habitats are more environmentally similar (e.g. shading, moisture) and edge effects more dominant than in more connected downstream sections (H4; e.g. Harvey et al. 2019). We also hypothesised that the relationship between communities and functioning may weaken with increasing flow intermittence and isolation, potentially driven by mismatches (H5).