Revealing a conservation challenge towards floodplain disconnection: decreasing turnover and increasing nestedness of mollusc metacommunities

Floodplain disconnection (FDC) has become a profound threat to many large river floodplains. Despite current research examining the impacts of FDC on aquatic biodiversity, little is known about how FDC affected freshwater molluscs from multifaceted biodiversity perspectives, especially at long-term temporal scales. We selected 7 lakes in the Yangtze River floodplain to evaluate how river-lake disconnection impacted taxonomic and functional alpha and beta diversity of molluscan assemblages during the past 60 years. Our results showed that five disconnected lakes experienced dramatic decline in species richness (decreased by 62.4%) and functional richness (54.6%) but increases in functional evenness (20.0%) after disconnection. Whereas such temporal changes in the two connected lakes (CLs) were much slighter. Somewhat unexpectedly, the overall taxonomic (increased by 7.0%) dissimilarities slightly increased through time, due to opposite changes in the turnover (ca. 40% decrease) and nestedness-resultant (ca.110% increase) components through time. The functional nestedness experienced more significant increase (162.0%), leading to relatively considerable increase in functional overall dissimilarity (64.0%). Such opposite changes in turnover and nestedness distinctly modified the structure of both taxonomic and functional beta diversity. The increase in nestedness-resultant component highlights the conservation priority of CLs keeping relative high levels of biodiversity. The overall correlations between changes in both alpha and beta diversity among two diversity facets were generally weak, indicating the different diversity facets would provide complementary information for biodiversity conservation and restoration. Thus, we recommended incorporating functional information of species into traditional taxonomy-based approaches in large river floodplains studies.


Introduction
Floodplains are one of the most important ecological landscapes and ecosystems in the world, which supported high levels of biodiversity and provided a set of ecological services (e.g., water supply and storage, food production, entertainment, aesthetic enjoyment and hydrologic regulation) to human society (Arantes et al. 2018;Tane et al. 2014;Wang et al. 2019). However, due to increasing human disturbances, most river floodplains in the world experienced dramatic and rapid biodiversity decline and ecological degradation during recent decades, making them become one of the most threatened ecosystems on Earth (Revenga et al. 2005;Tockner and Stanford 2002;Wang et al. 2019). Among the multiple anthropogenic disturbances (e.g., hydrological alterations, overexploitation of biological resource, water pollution) which impacting floodplain ecosystems, the loss of riverlake connection was usually regarded as one of the most severe and widespread threats to aquatic biodiversity, especially in those large river floodplains (Revenga et al. 2005;Saunders et al. 2002). This is because river-lake connection played a pivotal role in the material, energetic and biological exchanges between the river mainstem and its surrounding waterbodies (Arthington et al. 2006).
Recently, assessing the impact of hydrological disconnection on the floodplain ecosystems has become a hot topic concerned by ecologists and environmental managers (Jiang et al. 2022;Li et al. 2022). The existing research indicated floodplains disconnection would significantly change the original hydrological rhythms and lead to a series of adverse impacts (e.g., habitat loss and degradation, weakening water self-purification), and finally caused dramatic losses of biodiversity and ecosystem balance. However, most of the studies examined the effects of hydrological disconnection on aquatic biodiversity were conducted at the spatial scale, only few studies evaluated the effects at the temporal scales (i.e., the comparisons for the same lake groups before and after the disconnection). Such before-after comparisons in aquatic biodiversity is a more intuitive way to evaluate how local aquatic assemblages responded to the hydrological disconnection in river floodplains, and then provide suggestions for ecological managements and restoration (Steinfeld and Kingsford 2013;Wang et al. 2019;Wegener et al. 2017).
According to previous studies, the hydrological connectivity played an important role in sustaining the local aquatic biodiversity (Bracken et al. 2013). Typically, a high level of hydrological connectivity would support a high level of local richness (i.e., alpha diversity), and it is convenient to the migration, diffusion and gene exchanges between river and lakes, allowing some species flow from optimal habitats to suboptimal habitats (i.e., mass effects) (Shmida and Wilson 1985). Therefore, the highly connected floodplains usually appeared a high level of regional homogenization (i.e., a low level of beta diversity) (Liu and Wang 2010;Wang et al. 2016). After the loss of hydrological connectivity, there would be a decrease in the alpha diversity but increase in beta diversity in these disconnected lakes, due to the decreased mass effect and increased environmental heterogeneity between disconnected lakes (Carrara et al. 2012;Liu and Wang 2010;Paillex et al. 2013;Penha et al. 2017).
Many previous studies traditionally focused on the changes of taxonomic diversity (i.e., TD) to floodplain disconnection, whereas basically ignored the response of other diversity (e.g., phylogenetic and functional) facets diversity (i.e., PD and FD) (Cardoso et al. 2017;Lepš et al. 2001). The TD was an effective indicator of ecological status, but TD had some limitations, as it regarded different species as an identical unit and ignored the considerable differences in functional traits and phylogeny among different species (DıÁz and Cabido, 2001). Thus, ecologists have realized the need of incorporating functional information of species into the biodiversity studies (D'agata et al. 2014). Comparing to TD indices, functional diversity (FD)indices usually respond more strongly and reliably to environment changes, due to their closer associations with environmental gradient (Mccarthy 2004;Villéger et al. 2014). Furthermore, FD can directly reflect the influence of species to the ecosystem function (Lei et al. 2016;Mcgill et al. 2006). Furthermore, integrating FD to biodiversity studies can provide important complements to TD, and guidance to environmental managers and ecologists for effective biodiversity conservation (Alahuhta et al. 2019;Dolédec et al. 2017;Liu and Wang 2018).
Freshwater molluscs (i.e., gastropods and bivalves) are one of the most diverse and important organism groups in river floodplains, playing an important role in material and energy circulation of floodplain ecosystems (Christian and Harris 2008;Gössling et al. 2004). Whereas floodplain disconnection has led to dramatic biodiversity decline of freshwater molluscs, making them become one of the most threatened groups (Obolewski et al. 2009). Up to now, few studies evaluated the impacts of hydrological disconnection on the freshwater molluscs from multiple diversity dimensions (Jiang et al. 2021). Thus, there is an urgent need to evaluate how different facets of alpha and beta diversity of molluscan assemblages respond to river-lake disconnection, especially through a comparison before and after disconnection.
In this study, we selected 7 lakes in the Yangtze River floodplain to evaluate how riverlake disconnection affected taxonomic and functional alpha and beta diversity of freshwater molluscs during the past 60 years. Historically, the 7 lakes were all connected with the Yangtze River stem, whereas only 2 lakes (Lakes Dongting and Poyang) remained connecting in the current period. Firstly, we tested the changes of alpha and beta diversity and compared the changed between historical and current periods. Secondly, we examined the correlations between changes in diversity between different facets. Specifically, we predicted the two facets of alpha diversity would decrease through time, whereas beta diversity would increase after disconnection.

Study area and data collection
The Yangtze River is the third largest river in the world, with a mainstream length of 6397 km, covering a drainage area of 180,000 km 2 . The floodplain in the middle and lower reaches of the Yangtze River is China's highest concentrations of freshwater lakes (Yin et al. 2020). This region's ecology health had a great benefit on human beings, as it provided bioproduction, water resource, energy etc. for human beings (Wang et al. 2016;Zhang et al. 2018). However, over the past 50 years, the Yangtze floodplain was seriously impacted by human activities, such as hydrological construction projects, sand dredging, pollutant emission, overfishing (Wang et al. 2016. Among these threats, the hydrological disconnection was recognized as the most threaten factor (Jiang et al. 2021;Liu and Wang 2010) .
We selected 7 lakes (Lakes Poyang, Dongting, Taihu, Chaohu, Honghu, Huama and Donghu) with molluscan investigation data both in the historical and current periods for subsequent analysis. The 7 studied lakes included the five largest freshwater lakes in China, and their lake area ranges from 27.5 to 3210.0 km 2 , water depth varies from 2.2 to 6.4 m and the distance to the Yangtze Rivers is between 4.9 and 73 km (Table S2). Of the 7 lakes, only Lakes Poyang and Dongting kept connectivity in the current period, whereas other 5 lakes were disconnected from the Yangtze River stem. The molluscan presence/absence data were collected from published surveys (including scientific reports, books, online data and grey literature) since 1950s (Supplementary file: Table S1-S4). We updated the mollusc taxonomy according to He and Zhuang (2013), Bouchet et al. (2017) and Mollus-caBase (2021).
Then given that the variable disconnection time (the hydraulic projects which caused river-lake disconnection continued up to 1980s in this floodplain) and data availability of the 7 lakes, we divided the data into two periods: historical period  and current period (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018). For each period, all available data were gathered to build up exhaustive species lists. For each lake, multiple sampling gears were used in the molluscan survey, generally including modified trawl net, mussel rakes, hand-rakes and modified Petersen grab. Although this mixture of sampling methods does not yield strictly comparable abundance data, we deem this approach viable as our aim was to compile records of all molluscan species occurring in the studied lakes.

Mollusc trait information
In our study, we selected 2 continuous traits (maximal body size, shell height/width) and 4 categorical traits (respiration, feeding habit, reproduction mode and locomotion and substrate relation) to calculate functional diversity (Supplementary file: Table S5). These biological traits have been widely used in recent mollusc and macroinvertebrate studies, as they were considered as key traits reflecting the response and adaptation of molluscs to environment (Cai et al. 2019;Usseglio-Polatera et al. 2000). All information of the molluscan biological traits was obtained by Liu et al. (1979) and He and Zhuang (2013).

Calculation of taxonomic and functional alpha and beta diversities
We calculated two facets (i.e., taxonomic and functional) of alpha and beta diversity of molluscan assemblages for 7 lakes between historical and current periods. We used the species richness (SR) to represents the taxonomic alpha diversity. Function alpha diversity was quantified with functional richness (FRic), functional evenness (FEve) and functional divergence (FDiv). The three indices represent complementary components of functional diversity, and measure the total spread of traits, evenness and divergence of spread in trait space, respectively (Mason et al. 2005;Villéger et al. 2008). In the computation process of functional alpha diversity, we first used the Gower distance to calculate trait distances between species (Gower 1971) using the function "gowdis" in package FD (Laliberté et al. 2014), and then applied the principal coordinates analysis (PCoA) trait vectors to generate the functional space composed of multiple dimensions using the function "cmdscale" from the package stats. In this study, the first 4 PCoA vectors were applied, as these vectors usually considered as the minimum required dimension number (i.e., PCoA vectors) below which the quality of functional space was insufficient (Maire et al. 2015). All the three functional indices were calculated using "dbFD" function in the R packages FD (Laliberté et al. 2014) and vegan (Oksanen et al. 2012).
We also calculated taxonomic and functional beta diversity of molluscan assemblages and partition each facet of beta diversity into turnover and nestedness-resultant components, following the framework proposed by Baselga (2010;2012). We firstly measured 1 3 overall taxonomic beta diversity (Tβ sor ) and its turnover (Tβ sim ) and nestedness-resultant components (Tβ sne ) based on the partitioning of Sorensen dissimilarity using the function "beta.pair" from the R package betapart (Baselga et al. 2018). Then, we used the derived Sorensen index to calculated overall functional beta diversity (Fβ sor ) based on the above constructed functional space ) and also partition it into functional turnover (Fβ sim ) and nestedness-resultant (Fβ sne ) components as taxonomic dissimilarities. The three functional dissimilarity matrices were obtained using the function "functional. beta. pair" from the R package betapart. Furthermore, the percentage of turnover (p turn ) and nestedness (p nest ) of the overall dissimilarity for each beta diversity facet were also calculated.

Statistical analysis
Firstly, for the alpha diversity, we applied the paired t test to compare whether the two facets of alpha diversity indices (i.e., Taxonomic: SR; Functional: FRic, FDiv, FEve) differed between historical and current periods, using pairwise.t.test functions in R. Then we used the Spearman correlation analysis to examine the correlation among the changes (the current value minus the historical value) in four diversity indices.
For the beta diversity, we used principal co-ordinates analysis (PCoA) which was separately performed based on taxonomic and functional dissimilarity matrices (β sor , β sim and β sne ) to visualize patterns of mollusc assemblages during the two different periods. Meanwhile, we applied the permutational analysis of multivariate dispersions (PERMDISP, Anderson, 2006) with 999 permutations to test whether the multifaceted community compositional heterogeneity varied among two periods. In addition, we explored Pearson coefficient-based Mantel tests (999 permutations) to assess the relationships between temporal changes in pairwise dissimilarity among different diversity facets.
We ran all analyses (Spearman correlation analysis, PCoA, PERMDISP and Mantel tests) in the package vegan (Oksanen et al. 2019) in R version 4.0.1 (R Development Core team, 2020).

Result
In the historical period, before the loss of river-lake connectivity, 151 species belonging to 48 genera, 18 families, 10 superfamilies, 10 orders and 2 classes were recorded across the 7 studied floodplain lakes (Table 1). In the current period, there were 83 species, belonging to 37 genera, 15 families, 10 superfamilies, 8 orders and 2 classes. Overall, 69 species, 12 genera, three families, one superfamily and one order were extirpated from the species pool through time, whereas one species (Ampullaria gigas) which belongs to a new order appeared in the current period.
The correlations between the temporal changes in alpha diversity were insignificant among different diversity indices (Table 2).

Temporal change in multifaceted beta diversity of mollusc assemblages
In the historical period, the mean value of taxonomic dissimilarity (Tβ sor ) among the 7 studied lakes was 0.533 (range 0.200 to 0.671), mainly contributed by the turnover component (Tβ sim , mean value = 0.369, p turn = 69.1%) ( Table 3). The mean value of functional dissimilarity (Fβ sor ) was 0.369, with balanced contribution of turnover (Fβ sim , mean value = 0.190, p turn = 51.1%) and nestedness-resultant (Fβ sne , mean value = 0.179, p nest = 48.9%) components (Table 3). In the current period, both taxonomic and Table 1 List of families, superfamily, class and order, and number of genera and species within each family in the 7 studied lakes during the historical     components dramatically changed (in opposite direction), resulting in the significant modification of multifaceted beta diversity structure (from the domination of turnover component to the domination of nestedness-resultant component). Furthermore, by comparing the different pairs of lakes, we found such decreases in turnover and increases in nestedness-resultant were both more remarkable in pairs of connected and disconnected lakes (C-Dis) than in pairs of disconnected lakes (Dis-Dis) ( Table 4).
The PCoA ordination plots based on taxonomic and functional overall dissimilarity showed the heterogeneity of the mollusc communities in the current period were slightly higher (showing larger dispersion) than that in the historical period (Fig. 2a, d). As for their turnover components, the Tβ sim appeared smaller dispersion in the current period, whereas the dispersion of Fβ sim were similar among two period (Fig. 2b, e). In addition, the two facets of nestedness components were both showed obviously larger dispersion in the current period (Fig. 2c, f). Similarly, the PERMDISP analyses showed that there are only slight differences in the average distance to the group centroid for the taxonomic (Tβ sor ) between 2 periods, whereas the overall dissimilarity of functional (Fβ sor ) showed a marginally significant higher (P = 0.111) average distance to the group centroid in the current periods (Fig. 3). As for their turnover components, such distances of Tβ sim insignificantly declined through times, whereas the distance of Fβ sim kept stable between 2 periods (Fig. 3). Moreover, the two facets of nestedness-resultant components all showed a marginally significant or significant increase in such distance through time (the P value of Tβ sne ,and Fβ sne is 0.123 and 0.009, respectively) (Fig. 3).

Relationships between changes in different facets of beta diversity
The correlations between changes in taxonomic beta diversity and its two components with their corresponding functional dissimilarities were weak or moderate (r of mantel tests ranges from 0.339 to 0.727). Despite the major trends towards for multifaceted heterogenization for the changes in overall dissimilarities and nestedness-resultant components, the taxonomic and functional facets of turnover component both appeared opposite homogenization tendency (Fig. 4).   Fig. 3 Mean distances ± SE to the group centroid based on overalldissimilarity (β sor ), turnover (β sim ) andnestedness-resultant (β sne ) components of molluscanassemblages at taxonomic and functional diversity facets among historical (1956-1982, green) and current (2003-2018,red)periods Table 3 Taxonomic and functional overall dissimilarity (β sor ), turnover (β sim ), nestedness-resultant (β sne ), and contribution of turnover (p turn ) and nestedness-resultant (p nest ) to dissimilarity for historical period and current period.
Values are mean ± standard deviation with range in brackets

Impacts of river-lake disconnection on alpha diversity
The lakes on the Yangtze floodplain were historically connected with the mainstream and constituted a highly diverse and complex floodplain ecosystem. This naturally formed environment often provided suitable and stable habitats, abundant foods, and resources for aquatic organisms, and thus supported high level of molluscan alpha diversity in the historical period (Lydeard et al. 2004;Obolewski et al. 2009). However, after river-lake disconnection, we found considerable decreases both in taxonomic (decreased by 54% for the 7 lakes and ca. 60% if only considering the five disconnected lakes) and functional richness (ca. 40%) through time. Such dramatic loss of molluscan species richness after river-lake disconnection were also reported in previous studies (Obolewski 2011;Pringle et al. 2000) and can be attributed to following reasons. The river-lake disconnection usually caused a  . 4 Bivariaterelationship between temporal changes in taxonomic and functional betadiversity and their turnover and nestedness components among the 7 floodplainlakes. The Pearson correlation (r) and significance (P value) of Mantel testsare also provided in each panel. The one red circle represents pair of 2riverconnected lakes, the blue circles represent pairs of 5 dis-connectedlakes, and the black circles represent pairs of river-connected anddis-connected lakes. Grey areas indicate the major trend. TH taxonomichomogenization, TD taxonomic differentiation, FH functional homogenization,FD functional differentiation 1 3 series of environmental changes, including the alterations of hydrological rhythm, decrease in the spatial and temporal heterogeneity of habitats, and simplification and degradation of aquatic macrophytes . These changes were usually reported as important threats to the survival, growth and reproduction of many mollusc species and the causes of molluscan biodiversity decline (Lopes-Lima et al. 2018;Reckendorfer et al. 2006;Zieritz et al. 2018). Moreover, the losses of the exchange pathways also negatively affected mollusc diversity. For example, the glochidium of some mussel species (e.g., Anodonta woodiana, Lamprotula leai, Hyriopsis cumingii) usually parasitized on migratory fishes to finish their life history and dispersion between lakes and the mainstream (Gillis et al. 2008;Rogers-Lowery and Dimock Jr 2006). In addition, other human disturbances, e.g., water eutrophication, sand dredging and overfishing, were also responsible for the decline in molluscan species richness in the studied lakes, especially the five disconnected lakes (Jiang et al. 2022;Wang et al. 2019). The reduction in FRic (ca. 40%) is also considerable (especially in disconnected lakes) and only slightly lower than reduction in species richness, indicating a low level of functional redundancy of molluscan assemblages. Similar to significant decline in taxonomic richness, the molluscan assemblages in disconnected lakes experienced dramatic loss of functional traits and strategies. Several important traits occurred in historical periods (e.g., large body size, filter-feeding habits, burrower and ovoviviparity) showed substantial reduction (species number of these traits in the current periods were only ca. 1/3 ~ 1/4 of that in the historical period) after river-lake disconnection. Such dramatic losses of functional traits would inevitably lead to the decrease in functional richness. The considerable losses of functional traits and strategies of molluscan assemblages in the studied lakes deserves close attention by conservationists and environment managers, as this may impact overall lake ecosystem functioning and services (Biggs et al. 2020;Delgado et al. 2016). As one of the most important groups in freshwater lakes, mollusc assemblages are via their various traits of vital importance in maintaining lake ecosystem functions, such as promoting material circulation, restoring the aquatic environment, and stabilizing the environment ecosystem (Lydeard et al. 2004;Ruehl and Trexler 2011). Moreover, the increases in FEve through time implied that dramatic losses of unique functional traits made the overall functional traits of mollusc assemblages become more similar.
Comparing with the 5 disconnected lakes, the two connected lakes (Lakes Poyang and Dongting) showed slight changes in the two facets of alpha diversity, especially the functional diversity. Lakes Poyang and Dongting kept hydrological connectivity and relatively well ecological status, and thus experienced slighter loss of mollusc species (ca. 33%) than the five disconnected lakes. The two lakes are the two largest freshwater lakes in China, supporting the species-richest mollusc assemblages (Shu et al. 2014) and higher level of functional redundancy (several species have the same or similar traits). Thus, their mollusc functional diversity showed a higher resistance (the ~ 33% decrease in species richness only resulted in ~ 3% decrease in functional richness) to the loss of species than speciespoor lakes (Jiang et al. 2021;Mouchet et al. 2010;Mouillot et al. 2013).

Impacts of river-lake disconnection on beta diversity
For beta diversity, we found the overall taxonomic dissimilarities only slightly increased (ca. 7%) through time,. Such results were somewhat surprising, as previous studies showed the river-lake disconnection would block the species exchange pathway between river and lakes and increase the between-lake environmental heterogenization, and led aquatic 1 3 assemblages tend to be heterogeneous (Ding et al. 2017;Zhang et al. 2019). In the present study, we found opposite change directions in both two facets of turnover (decrease) and nestedness-resultant (increase) components, which masking the changes in overall dissimilarities. For the taxonomic dimension, the decrease in turnover but increase in nestedness were induced by the imbalanced extirpation of species between different lakes, especially between connected (fewer extirpated species) and disconnected ones. Specifically, the percentage of common species (a) were nearly unchanged (on average decreased from 31.3 to 28.4%) and maximum non-shared species (b) slightly increased (49.4-61.4%) after riverlake disconnection, whereas the percentage of minimum non-shared species (c) decreased (19.3-10.1%) over time. The unchanged a but decreasing c indicated the decrease in taxonomic turnover, while the increasing difference in species richness (b minus c) certainly led to the increase in taxonomic nestedness. After disconnection, several species-poor lakes were nearly completely nested by species-rich large lakes, e.g., Lake Huama was totally nested to Lakes Poyang and Dongting.
Somewhat different from the slightly changed taxonomic overall dissimilarities, the functional beta diversity experienced relatively large increase (64.0%), due to the considerably high increase in functional nestedness. Those disconnected lakes experienced significant loss of functional traits through time, whereas the trait losses in connected lakes were quite slight because of high functional redundancy. Thus, the increase in functional nestedness exceeds the increase in taxonomic nestedness and we observed a relatively large increase (increase in functional nestedness prevails over the decrease in functional turnover) in functional beta diversity.
The opposite changes in turnover and nestedness-resultant components highlight the necessity of partitioning overall beta diversity into independent components in biodiversity studies: the two components indeed represent complementary processes in community assembly (Angeler 2013;Dapporto et al. 2014). In the studied lakes, despite the slight changes in overall dissimilarity, the structures of multifaceted beta diversity of mollusc assemblages were dramatic altered from the dominance of turnover to the dominance of nestedness. Such modification had strong implications for biodiversity conservation and restoration in the studies lakes. According to previous studies, the dominance of turnover component indicated that all the sites or most of sites in overall region should be considered and conservated, while the dominating nestedness implied a requirement to preferentially focus on the species-rich sites (Angeler 2013;Baselga 2012). Therefore, the temporal shift from the dominance of turnover to the dominance to nestedness-resultant components at two diversity facets means conservation of the high-diversity lakes (e.g., Lake Poyang and Doting) become more urgent and preferential.

Correlations between changes in different indices diversity facets
Although the two facets of mollusc biodiversity showed the same direction of changes after disconnection, the correlations between changes in both alpha and beta diversity among different facets were generally weak. Such weak correlations indicated that evaluating functional diversity dimension indeed provided complementary information that cannot be portrayed by examining only taxonomic diversity. Thus, we recommend incorporating functional information of mollusc species into traditional monitoring and assessment approaches in evaluating the impacts of FDC and other anthropogenic disturbances in the large river floodplains, for more effective conservation and restoration practices in the future.
Our study showed significant impacts of river-lake disconnection on mollusc assemblages in the Yangtze River floodplain lakes, but there still have some limitations and caveats. Firstly, the number of studied lakes were quite limited (five disconnected lakes and only two lakes kept connectivity through time), though we exhausted our ability to collect data. While there were still more available lake mollusc investigation data in the studied region, but unfortunately, the available historical mollusc data were quite limited before disconnection. For the connected lakes, Lakes Poyang and Dongting are the two remaining lakes which are currently connected to the Yangtze River stem. Therefore, it is impossible to compile dataset including more lakes for examining the temporal changes before and after disconnection.

Conclusion
In this study, we examined how long-term river-lake disconnection drove taxonomic and functional alpha and beta diversity of molluscan assemblages in the Yangtze floodplain lakes. We found significant decrease in taxonomic and functional alpha diversity through time. Moreover, although the two facets of overall beta diversity change relatively slightly, their turnover and nestedness-resultant component dramatically changes in opposite direction, indicating a significant modification of beta diversity structure. The dominance shifted from turnover component to nestedness-resultant component provided useful implication for biodiversity conservation in this region: the high-diversity lakes (i.e., Lake Poyang and Doting) should be protected and concerned preferentially. Moreover, the weak correlation between changes in different facets signified that embracing functional information into taxonomy-based approaches would provide important and complementary information and enhance the ability to examine and evaluate aquatic biodiversity in the large river floodplains. Overall, our study showed that floodplain disconnection caused profound and severe impacts on the mollusc fauna in Yangtze River floodplain. We appeal to promote the conservation of freshwater molluscs and other organism groups in this floodplain, especially the high conservation priority for two connected lakes (i.e., Lake Poyang and Doting) with high diversity.
Ethical approval All applicable international, national and/or institutional guidelines for sampling, care and experimental use of organisms for the study were followed and all necessary permissions and approvals were obtained in respect of the original collections of the data. Consent to participate Informed consent was obtained from all observers participating in the study.