Flood did not cause the taxonomic and functional homogenization of Chironomidae (Diptera) community in a Neotropical floodplain

the taxonomic and functional homogenization of Chironomidae (Diptera) community in a Neotropical floodplain

Acta Limnologica Brasiliensia, 2024, vol. 36, e6 Among the aquatic communities of the Upper Paraná River floodplain, Chironomidae larvae (Diptera) correspond to the most frequent, abundant, and diverse benthic invertebrates (Takeda et al., 2004;Petsch et al., 2015Petsch et al., , 2017;;Bertoncin et al., 2019).Chironomidae is divided into 11 subfamilies and 22 tribes (Ferrington Junior, 2008).The composition and distribution of Chironomidae larvae in the Upper Paraná River floodplain are strongly related to environmental filters, such as the high heterogeneity of habitats, which are influenced by the environmental changes promoted by the hydrological regime (Takeda et al., 2004).
We investigated the effect of flooding on the taxonomic and functional beta diversity of Chironomidae larvae in 18 lakes in the Upper Paraná River floodplain (Brazil).We investigated the hypothesis that the flood promotes taxonomic and functional biotic homogenization of Chironomidae among floodplain lakes.If this hypothesis is confirmed, we expect to find a lower taxonomic and functional beta diversity of Chironomidae among lakes during the flood than during the drought period.

Study area
Our study was part of the Long-Term Ecological Research program (PELD/CNPq).The sampled area comprises the last dams free stretch on the Paraná River in Brazil (Agostinho et al., 2008).We sampled 18 lakes in two periods (Figure 1): flood (March 2011) and drought (September 2011).

Sampling period
In 2011, large fluctuations were observed in the hydrometric level of the Paraná River.In March 2011, during the first sampling, the water level was higher than 6m, when all lakes were connected by the flood (Souza Filho, 2009).In September 2011, during the second sampling, the water level was below 3m, characterizing a drought period (Figure 2).

Data collection
We sampled five bottom samples in each lake with a modified Petersen grab (0.0345 m -2 ).The sedimentological content was fractioned with the aid of a set of 2mm, 1mm and 0.2mm sieves.The biological material retained on the last sieve was fixed in commercial alcohol (92.6°) for subsequent screening under a stereoscopic microscope in the laboratory.

Introduction
Floodplains are areas periodically flooded by the lateral overflow of rivers (Junk et al., 1989).Usually, during the flood, aquatic environments tend to be biologically less distinct since flood pulse tends to increase connectivity among habitats and homogenize the environmental characteristics (Thomaz et al., 2004(Thomaz et al., , 2007)).On the other hand, during the drought, many lakes are isolated again and become more distinct from each other due to chemical and physical changes in their habitats (Thomaz et al., 2004;Bozelli et al., 2015).
The flood pulse can not only promote higher environmental similarity (i.e., related to limnological features such as nutrients, pH, dissolved oxygen and turbidity) but also contribute to an increase in biological similarity, resulting in a decrease in beta diversity (Thomaz et al., 2007;Petsch, 2016).Beta diversity can also be defined as the variability in species composition between sampling units (Anderson et al., 2006), generated both by species differences and replacement (Baselga, 2010).A reduction in beta diversity may indicate processes related to taxonomic biotic homogenization, characterized by increasing species similarity between space or time (McKinney & Lockwood, 1999;Olden et al., 2004).
One of the community responses to hydrological cycles is related to species traits once a functional approach can improve our understanding of how communities are shaped by ecological processes such as flooding (e.g., attributes of higher dispersion or tolerance to low dissolved oxygen conditions).Functional attributes represent any measurable morphological, physiological or phenological characteristic at the individual level that indirectly affects biological efficacy through detected effects on its growth, reproduction or survival (Violle et al., 2007).Also, communities can be more similar regarding their response functional traits due to some anthropogenic or natural cause, driving functional homogenization (e.g., Gentilin-Avanci et al., 2021;Feng et al., 2023;Rivera et al., 2023).One of the main anthropogenic actions that decrease functional diversity in freshwater ecosystems are deforestation and agriculture (Colares et al., 2022).These interferences promote habitat loss and chemical changes in freshwater ecosystems, causing some rare species to disappear, thus promoting the functional redundancy of the remaining species (Violle et al., 2017) and functional homogenization (Petsch, 2016).On the other hand, the flood pulse is the main example of a natural cause of biotic homogenization in freshwater ecosystems (Thomaz et al., 2007;Petsch, 2016).
Chironomidae larvae were desiccated and mounted on slides with Hoyer solution according to the methodology proposed by Trivinho-Strixino & Strixino (1995) and later identified in an optical microscope using Epler (2001) and Trivinho-Strixino (2011) keys.The identified organisms are deposited in the Zoobenthos laboratory (NUPELIA/UEM).
We also measure environmental features during sampling, such as water temperature (°C) and dissolved oxygen (mg/L) near the bottom, electrical conductivity (µS/cm), depth (m), pH, turbidity (NTU), chlorophyll (µg/L), total suspended material (µg/L), total nitrogen (µg/L) and total phosphorus (µg/L).Fluviometric level data of the Paraná River were obtained through daily measurements with a ruler installed in Porto Rico city (Paraná, Brazil).

Data analysis
To test our hypothesis of biotic homogenization during the flood period, we investigated whether Chironomidae beta diversity among lakes is lower in flood than drought.Taxonomic beta diversity was based on a Jaccard dissimilarity calculated from Chironomidae incidence data.We also calculated Bray-Curtis dissimilarity based on abundance data, but the results were similar compared to Jaccard dissimilarity (mean distance to centroid in flood: 0.593; drought: 0.587; P = 0.862).In contrast, functional beta diversity was calculated from a dendrogram performed using Chironomidae traits, and then the Gower distance was applied.We calculated the taxonomic and functional beta diversity of Chironomidae separately for each hydrological period as the mean distance of each lake to the centroid in a multivariate space (PCoA) using a multivariate dispersion analysis (PERMDISP; Anderson et al., 2006).This way, higher average distances to centroids indicate higher beta diversity values, i.e., higher dissimilarity between species and traits.To test if beta diversity values differ between hydrological periods, we used permutational null models with 999 randomizations.We performed all analyses in the R program version 3.4 (R Core Team, 2022) using the vegan (Oksanen et al., 2017), FD (Laliberté et al., 2014) and adespatial (Dray et al., 2018) packages.We constructed the figures using the ggplot2 package (Wickham et al., 2016).

Results
We found a total of 56 morphospecies of Chironomidae represented by three subfamilies: Chironominae (46 morphospecies), Tanypodinae (9 morphospecies) and Orthocladiinae (1 morphospecies).We found 51 morphospecies in the drought and 26 in the flood period (Table 1).The most abundant morphospecies in the flood period were those of the genus Chironomus, while in the drought period, they were those of the genera Tanypus and Polypedilum.We observed that 65% of the morphospecies found in the flood period had adaptations for oxygen uptake (e.g., abdominal tubules and/or hemoglobin), and 63% of the morphospecies found in the drought period had the same adaptations.
We observed the highest dissolved oxygen values, pH, turbidity, chlorophyll and nutrients in the drought period (September 2011).On the other hand, the highest temperature and conductivity values were recorded during the flood period (March 2011) (Table 2).Despite the lower number of Chironomidae morphospecies in the flood period, the hypothesis that the flood reduced taxonomic beta diversity was not supported, as the mean distance to the centroid was similar between the flood (0.619) and the drought period (0.624) (F (1,16) = 0.261; P = 0.878; Figure 3A) using Jaccard dissimilarity (based on incidence).We found similar results estimating taxonomic beta diversity with Bray-Curtis dissimilarity based on abundance (mean distance to the centroid in flood = 0.593; mean distance to the centroid in drought = 0.587; F (1,16) = 0.031; P = 0.862).We also found a similar functional beta diversity of Chironomidae among lakes during the drought period (mean distance to the centroid = 0.583) and flood period (mean distance to the centroid = 0.543) (F (1,16) = 1.313;P = 0.267; Figure 3B).

Discussion
Contrary to our hypothesis, we did not find taxonomic or functional homogenization of Chironomidae larvae due to flood pulse in the Upper Paraná River floodplain.Despite large variations in dissolved oxygen, pH, turbidity, chlorophyll, and nutrient levels between the flood and drought, the Chironomidae larvae variability among lakes was similar between hydrological periods.Our findings suggest that while Chironomidae morphospecies abundance and richness may decrease due to the flood effect (Petsch et al., 2015), taxonomic and functional beta diversity among lakes remained similar between different hydrological conditions.We added evidence for the controversy of flood homogenization effect because some studies found biotic homogenization due to flood pulse (e.g., Lopes et al., 2014 andBozelli et al., 2015, investigating zooplankton in Brazilian floodplains) while other studies did not (e.g., Petsch et al., 2021a, investigating zooplankton and aquatic macrophytes in the Upper Paraná River floodplain).
The identity of tolerant species due to a disturbance as flood can vary among sites (such as found for land use disturbance; Johnson & Angeler, 2014;Petsch et al., 2021b).Chironomidae morphospecies tolerant to flood modifications may differ across lakes, creating some variability during the flood period.For example, changes in the substrate, influenced by the hydrological regime of the Paraná River, did not drastically change the composition of Chironomidae but decreased their density (Anjos et al., 2011).The oligotrophication promoted by dams upstream of the Upper Paraná River floodplain also changed the dominance of morphospecies persisting under such stressful conditions (Rosin et al., 2009).TN TP Flood 27.8 (±0.8) 5.05 (±2.8) 6.3 (±0.5) 50.0 (±5.4) 2.72 (± 4.2) 1.37 (±0.9) 3.3 (±2.3) 776.9 (±95.1)20.2 (±5.2) Drought 23.7 (±1.9) 6.5 (±1.8) 6.9 (±0.6) 35.8 (±18.5)15.8 (±8.9) 1.5 (±0.8) 15.7 (±17.5)1251.8 (±436.8)60.5 (±59.The drought period is supposed to have higher beta diversity due to higher lakes isolation, which could lead these habitats to a heterogeneous state (Thomaz et al., 2007;Bozelli et al., 2015; but see Bertoncin et al., 2019 for extreme drought effects).However, flood conditions can also create some taxonomic and functional beta diversity of Chironomidae among lakes throught species replacement.For example, Xestochironomus is the unique morphospecies recorded with the wood mining trait and is favored by flooding because more branches and wood are carried to the pond due to flood pulse (Junk et al., 1989;Petsch et al., 2015).Indeed, Xestochironomus only occurred during the flood in a small pond with high riparian vegetation, contributing to generating beta diversity during the flood period.
Chironomidae larvae are among the most resistant macroinvertebrates in freshwater habitats (Molineri et al., 2020).They are widely distributed and functionally diverse (Martins et al., 2021;Arpellino et al., 2023).Many Chironomidae morphospecies show functional traits related to high resistance and resilience to flood effects (Serra et al., 2017).For example, we found a high proportion of Chironomidae larvae with adaptations for oxygen uptake in both periods, indicating their ability to support varying oxygen levels in the lakes.This finding corroborates the idea that oxygen availability is a crucial factor influencing the distribution and survival of many Chironomidae larvae (Rosin & Takeda, 2007) and pupae (Arpellino et al., 2023).During the flood, dissolved oxygen tends to be lower due to the accumulation of organic matter and greater decomposition (Anjos et al., 2011).When comparing adaptations for oxygen uptake of Chironomidae in drought and flood periods, we found no differences in the percentage of morphospecies with such adaptations, indicating the similarity of some response traits of Chironomidae between flood and drought periods.
We did not find evidence of biotic homogenization of Chironomidae due to flood pulse using taxonomic or functional facets.We highlight the complexity of river-floodplain systems because other mechanisms besides increasing connectivity and environmental similarity due to flood pulse may structure the Chironomidae community during the inundation period, such as variation in the identity of flood-tolerant morphospecies across habitats.Our findings contribute to understanding the ecological success of this family of aquatic insects in the Upper Paraná River floodplain and have implications for the conservation and management of river-floodplain ecosystems.

Figure 2 .
Figure 2. Fluviometric level of Paraná River in 2011.When the Paraná River level reaches 6m, a total flood occurs (March 2011, blue), while the Paraná River level below 3.5m indicates a drought period (September 2011, orange).

Figure 3 .
Figure 3. Taxonomic (A) and functional (B) beta diversity of Chironomidae among flood (blue) and drought (red) floodplain lakes in the upper Paraná River floodplain.

Table 2 .
Environmental variables (mean ± standard deviation) of 18 lakes in the Upper Paraná River floodplain during the flood (March 2011) and drought period (September 2011).