Drought Effects on Benthic Macroinvertebrate Community Resilience and Functional Diversity in Wetland Mesocosms

doi:10.21203/rs.3.rs-5303802/v1

Download PDF

Research Article

Drought Effects on Benthic Macroinvertebrate Community Resilience and Functional Diversity in Wetland Mesocosms

https://doi.org/10.21203/rs.3.rs-5303802/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

This study investigated the impacts of short-term (4 months) and long-term (8 months) drought on benthic macroinvertebrate (BMI) communities in wetland mesocosms, focusing on community composition, functional diversity, and resilience mechanisms. Using controlled experiments, we compared BMI communities in three treatments: control (C), short-term drought (S), and long-term drought (L). The results showed that drought duration significantly influenced on some community and functional diversity indices. While the C wetland maintained stable communities, S and L wetlands exhibited distinct recovery patterns post-drought. The S wetland demonstrated higher community similarity to pre-drought conditions compared to the L wetland, suggesting greater resilience to short-term disturbances. Functional diversity indices revealed complex responses, with drought altering feeding habits, living types, and voltinism traits. Long-term drought led to dominance by multivoltine species and reduced functional evenness. This study highlights the importance of drought duration in shaping BMI communities and emphasizes the need for long-term monitoring to understand wetland ecosystem recovery dynamics under climate change scenarios.

Wetland mesocosm

Benthic macroinvertebrates

Functional diversity

Drought

Community resilience

Wetlands are vital ecosystems that provide critical habitat for a wide range of species and play key roles in biogeochemical cycling, energy flow, and maintaining ecosystem stability (Cherry 2011). Within these ecosystems, benthic macroinvertebrates (BMIs) serve essential ecological functions, such as facilitating organic matter decomposition, enhancing water quality through filtration, and acting as a primary food source for higher trophic organisms, including amphibians, fish, and birds (Oertli 1993; Cayrou and Céréghino 2005). BMIs are highly responsive to variations in aquatic environmental conditions, including water flow, oxygen levels, and habitat structure, making them reliable bioindicators for assessing wetland health and functional diversity (Stewart et al. 2013).

Climate change has increased the frequency and intensity of droughts, reducing water retention in wetland ecosystems and threatening their ecological balance (Burkett and Kusler 2000). Drought induces physical and chemical changes, such as lower water levels, higher temperatures, reduced dissolved oxygen, and altered habitat structure, which significantly impact BMI communities (Diaz et al. 2019). For instance, rising temperatures reduce dissolved oxygen levels, negatively affecting oxygen-dependent species and favoring thermally tolerant taxa, leading to shifts in community structure and decreased biodiversity (Stewart et al. 2013; Dallas and Rivers-Moore 2014). As water levels decline, some species can no longer inhabit previously suitable areas, resulting in changes to wetland channels and aquatic vegetation distribution (Zambrano et al. 2019). In response to these stressors, some BMI species exhibit behavioral adaptations, such as migrating to more favorable habitats, while others undergo physiological adaptations, such as entering dormancy or developing desiccation resistance (Southwood 1977; Vineetha and Nandan 2021). However, these strategies are often insufficient to fully mitigate the long-term effects of drought.

According to Southwood (1977) habitat templet theory and Poff' (1997) environmental filtering concept, drought acts as a selective filter, favoring species with functional traits that enhance survival under altered conditions. This filtering process reshapes community structure and functional diversity, with species better adapted to drought becoming dominant (Loreau et al. 2001). These environmental changes are closely linked to the life history, behavioral traits, and biological characteristics of BMIs, which are critical factors in determining their resistance and resilience to drought (Poff 1997; Sweeney and Vannote 1978; Vannote and Sweeney 1980; Lake 2003; Bonacina et al. 2023). The functional traits of BMIs are increasingly recognized as vital indicators for understanding their responses to various environmental stressors, particularly hydrological disturbances (Bêche et al. 2006; Schriever et al. 2015; Leigh et al. 2016). In this context, research on functional diversity using biological traits has become an increasingly important tool in ecological studies, as it evaluates the roles species play within ecosystems based on their functional characteristics (Tierno de Figueroa et al. 2010; Domisch et al. 2017; Pacifici et al. 2017). Functional diversity is typically assessed using four key indices: functional richness (FRic), functional evenness (FEve), functional divergence (FDiv), and functional dispersion (FDis) (Díaz and Cabido, 2001; Mason et al. 2005; Petchey and Gaston 2006; Villéger et al. 2008). These indices evaluate the distribution of species' functional traits within a community and the diversity of their ecological roles (Richards et al. 1997; Gagic et al. 2015). Functional richness (FRic) represents the range of functional space occupied by species within a community, indicating how diversely ecological functions are performed (Mason et al. 2005). Functional evenness (FEve) reflects how evenly functional traits are distributed within a community and is associated with efficient resource use (Mason et al. 2005; Goswami et al. 2017). Functional divergence (FDiv) assesses the extent to which species are concentrated at the edges of the functional space, reflecting adaptability to extreme treatments (Mouchet et al. 2010), while functional dispersion (FDis) indicates how widely species' functional traits are spread, showing occupation of diverse ecological niches (Laliberté and Legendre 2010).

Studies examining the effects of drought on BMI communities from a functional trait perspective have shown that certain functional groups become dominant during dry periods, leading to changes in the community's functional diversity (Ledger et al. 2011; Stubbington et al. 2016; Aspin et al. 2018). For instance, during droughts, BMI species tend to exhibit rapid growth or metamorphosis, resulting in declines in functional richness and evenness (Bogan and Lytle 2011; Kim et al., 2014b; Bogan et al 2015; Aspin et al. 2018; Stuart et al. 2020). Additionally, species with specific functional traits either survive in desiccation-resistant forms or utilize their dispersal abilities, such as flight, to relocate to new habitats in response to drought (Vineetha and Nandan 2021). However, most studies rely on observational data from natural environments, which limits the ability to accurately control for the complexity of environmental variability. Additionally, these studies often fall short in providing a deep understanding of the long-term effects of drought on BMI communities and changes in their functional traits. To address this research gap, the present study explores the effects of drought on BMI functional traits within a wetland mesocosm. Using Southwood (1977) habitat templet theory and Poff (1997) environmental filtering concept, we examine how drought selectively influences the functional traits of BMIs. According to these theories, the physical characteristics of the habitat and hydrological disturbances act as filters, selecting species with specific functional traits. This study analyzes the impact of drought on BMI community structure and functional diversity, while also investigating how BMI communities exhibit resilience and adaptability under varying drought conditions. Through this approach, we aim to provide deeper insights into how climate change-induced drought affects wetland ecosystems.

Understanding the effects of drought on BMI communities is essential for predicting how wetland ecosystems will respond to the increasing frequency and intensity of hydrological disturbances driven by climate change (Aspin et al. 2018). To elucidate the specific mechanisms driving these changes, controlled experimental systems that allow for precise manipulation and monitoring of environmental variables are necessary. Wetland mesocosms provide an ideal platform for studying the impacts of environmental disturbances, particularly drought, on BMI communities. As artificial ecosystems that closely replicate the ecological and environmental conditions of natural wetlands, mesocosms enable controlled experiments where variables such as water level, temperature, and nutrient availability can be precisely adjusted (Odum 1984; Ding et al. 2017). Such experimental control offers a unique opportunity to isolate the drivers of changes in community composition and functional diversity—factors that are often difficult to disentangle in natural wetland systems due to their inherent complexity and variability (Wong et al. 2004; Kim et al. 2014a). For instance, Kim et al. (2014b) demonstrated that BMI communities could rapidly recolonize following drought, with populations showing swift recovery in mesocosm experiments. Furthermore, mesocosm studies have shown that the abundance of certain species, such as mosquito larvae and snails, can fluctuate significantly depending on the duration of drought, providing valuable insights into species-specific responses to varying drought conditions (Hockaday et al. 2024). Despite the numerous advantages of mesocosms, they have inherent limitations when it comes to fully replicating the complex interactions that occur in natural wetlands, such as interspecific competition, predation, and seasonal physical changes. While mesocosms allow for controlled experiments and the manipulation of well-defined environmental stressors, they do not completely capture the complexity of natural ecosystems (Buikema and Voshell 1993; De Szalay 1996). Nonetheless, mesocosms remain a powerful tool for investigating the functional roles and adaptive strategies of species under controlled environmental conditions. Through this approach, we can gain valuable insights into the resilience mechanisms of BMI communities and their adaptations to drought, which are essential for predicting how wetland ecosystems will respond to future climate change scenarios and for developing conservation and management strategies (Odum 1984). It is important to acknowledge that although mesocosm experiments provide controlled conditions for studying ecological processes, these artificial systems may not fully replicate the intricacies of natural wetlands. Key factors such as species interactions, seasonal variability, and the influence of large, mobile species may be underrepresented in mesocosms. Additionally, the limited size of mesocosms can constrain the expression of certain trophic interactions, potentially influencing the overall outcomes of the experiments. We have carefully considered these limitations in interpreting our results and have discussed the potential implications of these constraints in detail in the discussion section.

In this study, we utilize a wetland mesocosm approach to evaluate the differential responses of BMI communities to three drought treatments. The treatments include a control wetland (C) with sustained water levels, a short-term drought wetland (S) lasting four months, and a long-term drought wetland (L) lasting eight months. By analyzing the impact of drought duration on BMI community composition and functional diversity, we aim to provide a more detailed understanding of how wetland ecosystems respond to prolonged or extreme drought conditions. Specifically, we seek to test the following hypotheses: (1) the duration of drought will have a proportional effect on the structure and functional diversity of BMI communities, with longer droughts leading to more significant changes; (2) BMI communities will recover rapidly following short-term drought, while long-term drought will result in slower recovery and more substantial shifts in community structure; and (3) during the post-drought recolonization process, species with specific biological traits, such as multivoltinism and strong dispersal ability, will become dominant. By testing these hypotheses, this study aims to provide critical baseline data for the conservation and management of wetland ecosystems in the context of increasing hydrological variability due to climate change.

2.1 Study site and experimental set-up

To investigate the effects of drought on BMI communities, we conducted controlled experiments using wetland mesocosms at the Korea University Experimental Field Station, located in Deokso-eup, Namyangju-si, Gyeonggi-do, South Korea (37°35'1.75"N, 127°13'52.81"E). (Fig. 1a). To maintain a consistent water depth of 30–40 cm throughout the study period, each mesocosm received a continuous supply of groundwater, and water level control pipes were installed to prevent overflow and ensure the stability of hydrological conditions (Fig. 1b). Additionally, drainage pipes were installed at the bottom of each mesocosm to facilitate water level regulation and the removal of excess water. The mesocosms were positioned 1 meter apart to minimize edge effects and provide uniform exposure to environmental conditions such as sunlight and wind. On December 16, 2019, the existing soil and vegetation were removed from each mesocosm, and new soil was added to ensure consistency and prevent contamination from previous biota. On May 13, 2020, we planted four aquatic plant species, including Nymphoides peltata, Equisetum hyemale, Lythrum anceps, and Iris ensata, based on their ability to enhance BMI habitat complexity (Keast 1984; Papas 2007). A groundwater supply system was installed to maintain continuous water flow, with water pipes centrally positioned in each mesocosm to control the water level at approximately 30 cm.

Three experimental treatments were established: (1) Control (C) wetland, with a constant water level, (2) Short-term drought (S) wetland, with drought conditions applied from April to July 2021, and (3) Long-term drought (L) wetland, with drought conditions applied from December 2020 to July 2021. Each treatment had three replicates, totaling nine mesocosms. We used metal ladders positioned above each mesocosm to conduct BMI surveys without disturbing the substrate or introducing foreign material.

2.2 Sampling and environmental data

2.2.1. BMI Sampling

BMI communities were sampled monthly over 18 occasions from July 2020 to June 2022, with no surveys conducted during the winter months of December, January, and February. BMI sampling was performed quantitatively at three randomly selected points within each wetland mesocosm using a dredge net (20 × 40 cm, 0.1-mm mesh). Collected samples were immediately preserved in 80% ethanol in the field and transported to the laboratory. In the lab, specimens were sorted using a stereomicroscope at 10x magnification, and individuals were identified to the lowest possible taxonomic level using regional taxonomic keys. Any organisms not identifiable to species level were categorized as "sp." All specimens were counted, and the data from the three points within each mesocosm were combined for analysis.

2.2.2. Plant Sampling

Plant species richness and cover were assessed in three wetland treatments. Surveys were conducted at three time points: July 2021 (during drought for the short-term and long-term drought wetlands), November 2021 and April 2022 (after the cessation of drought treatments).

The data from the vegetation surveys were organized in a matrix format (relevés × species), where each cell represents the abundance of species within each relevé. To facilitate numerical and statistical analysis, the abundance/dominance indexes for each species were converted into percentage cover values, following the method proposed by Canullo et al. (2012) (Table 1). The scientific names of species were assigned in accordance with the National List of Species of Korea 2023. The classification of plant species into aquatic and terrestrial types was determined according to their wetland preference and life form. This classification was based on the criteria outlined in “Wetland preference and life form of the vascular plants in the Korean peninsula” (Choung et al. 2021).

Table 1

Meaning of the abundance/dominance indexes of Braun-Blanquet (Braun-Blanquet 1964) and transformation values proposed by Canullo et al. (2012).
Abundance/dominance Indexes of Braun-Blanquet	Plant coverage	Transformation values of abundance/dominance indexes
r	Rare species in the relevés	0.01%
+	< 1%	0.05%
1	1% − 5%	3.00%
2	6% − 25%	15.00%
3	26% − 50%	37.50%
4	51% − 75%	62.50%
5	76% − 100%	87.50%

Vegetation plots were georeferenced and established in permanent locations, with each plot measuring 25 m² (5 × 5 m). To ensure accuracy in monitoring, a wooden frame was constructed around each plot, and all plant species within the frame were surveyed. Plant cover was recorded using the Braun-Blanquet scale, which was later converted to percentage values for analysis (Braun-Blanquet 1964). This method provided a consistent framework for data collection and facilitated comparative analysis across the surveyed plots.

2.2.3. Environmental data

During the study period, we used air temperature and precipitation data for the Deokso area provided by the Korea Meteorological Administration (http://www.kma.go.kr). Additionally, during field surveys, we measured water temperature (°C), pH, dissolved oxygen (DO, mg/L), electrical conductivity (µS/cm), and salinity (psu) at each wetland using a multi-parameter water quality meter (Professional Plus, YSI, Ohio, USA).

2.3 Data analysis

2.3.1 Biological traits

We classified all species according to their biological traits to compare the functional diversity of BMIs (Table 2). biological traits were selected based on relevant literature, considering the environmental treatments at the sampling sites (Poff et al. 2006; Apsin et al. 2018; Vineetha and Nandan 2021). The five selected biological traits were further subdivided into 18 categories, based on ecological, habitat, and life history characteristics, with each trait category abbreviated and expressed as a code.

Table 2

Biological traits classification of benthic macroinvertebrates
Category	Trait	Trait state	Code
Mobility	Swimming ability	none	SA1
		able (weak and strong)	SA2
	Dispersal	active	D1
		passive	D2
Ecology	Feeding habit	collectors	FH1
		piercers	FH2
		shredders	FH3
		predators	FH4
	Living type	burrowers	LT1
		climbers	LT2
		sprawlers	LT3
		clingers	LT4
		swimmers	LT5
Life history	Voltinism	semivoltine (< 1 generation/y)	V1
		univoltine (1 generation/y)	V2
		multivoltine (> 1 generation/y)	V3

2.3.2 Environmental factors and species richness

We compared the environmental factors, species richness, and abundance among the three wetlands during the study period. The comparison of environmental factors, as well as the average species richness and abundance for each wetland, it was conducted using one-way analysis of variance (ANOVA). All comparisons and statistical analyses were performed using SPSS v.25 (IBM Corp., Armonk, New York, USA).

2.3.3. Analysis of BMI community changes among different wetland treatments and stages

To compare changes in BMI communities, we established three different wetland treatments and divided them into three stages, resulting in a total of nine groups for data analysis (Table 3). Phase 1 represents the initial colonization stage for BMIs among all wetland treatments. Phase 2 represents the mid-colonization stage for the control wetlands, while the short-term and long-term drought treatments entered a second colonization stage due to the effects of drought, designed to observe the recolonization process. Phase 3 represents the late colonization stage for the control wetlands. In contrast, the short-term and long-term drought treatments were categorized as being in a second mid-colonization stage. Phase 1 to 3 were divided to compare and analyze the effects of drought on the formation and changes in BMI communities at different stages of colonization.

Table 3

Colonization stages of benthic macroinvertebrates among different wetland treatments
Wetland treatments	Colonization stage
Wetland treatments	Phase 1	Phase 2	Phase 3
Control	Initial-colonization	Mid-colonization	Late-colonization
Short-term drought (spring drought)		Second colonization	Second mid-colonization
Long-term drought (winter-spring drought)		Second colonization	Second mid-colonization

To compare the similarity in species composition and abundance among the nine groups, the Bray-Curtis similarity was calculated using Primer 7.0 (PRIMER-E Ltd., UK). This analysis utilized the data on species richness and abundance counts observed in each group. We calculated the Jaccard similarity index to compare the species composition of benthic macroinvertebrate communities among the three wetland treatments during the three phases of colonization. The index was calculated for each pair of phases (Phase 1 vs. Phase 2, and Phase 2 vs. Phase 3) within each wetland treatments. The Jaccard similarity index is defined as the ratio of the number of species common to both communities to the total number of species observed and was used to assess the similarity in species composition (Jaccard 1900; 1901).

$$\:Jaccard\:index=\frac{{S}_{12}}{{S}_{1}+{S}_{2}-{S}_{12}}$$

Symbols: S₁ the number of species in Assemblage 1; S₂ the number of species in Assemblage 2; S₁₂ the number of shared species;

The spatial variability of BMI communities was analyzed using non-metric multidimensional scaling (NMS). Prior to the analysis, BMI abundance data were normalized using log₁₀ (x + 1) transformation. Rare species, defined as those with a frequency of one or fewer occurrences during the study period at each wetland, were excluded from the analysis. To test the statistical significance of the NMS results, a multi-response permutation procedure (MRPP) was performed, and indicator species analysis (ISPAN) was conducted to identify indicator species within each community. NMS, MRPP, and ISPAN analyses were conducted using PC-ORD v.7.0 (MjM Software, OR, USA) (McCune et al. 2002).

2.3.4. Community and functional diversity indices

We compared the BMI communities and functional diversity among nine groups using four community indices and FD indices. The community indices were calculated based on species richness and abundance data, using the Dominance Index (DI) proposed by McNaughton (1967), the Shannon Diversity Index (Hʹ) by Shannon (1948), the Richness Index (R1) by Margalef (1958), and the Evenness Index (Jʹ) by Pielou (1966). The FD indices were calculated based on species richness, abundance, and biological trait data among the nine groups, using four FD indices: FRic, FEve, FDiv, and FDis. The FD indices were calculated using the "dbFD" function in the FD package of the R software (Laliberté et al. 2015). FRic, FEve, and FDiv were calculated using the formulas proposed by Villéger et al. (2008), while FDis was calculated using the formula proposed by Laliberté and Legendre (2010). The descriptions, calculations, and references for the four selected indices (FRic, FEve, FDiv, FDis) are provided in Table 4. The community and functional diversity indices calculated for the nine groups were statistically compared using one-way analysis of variance (ANOVA) in SPSS 25.0 (IBM Corp., Armonk, New York, USA).

Table 4

Functional diversity measures S, the total species richness; PEW, partial weighted evenness; dG, mean distance to the center of gravity; Δd, sum of abundance weighted deviances; Δ|d|, absolute abundance-weighted deviances from the center of gravity; a_j, the abundance of species _j and z_j is the distance of species j from the weighted centroid c.
Indices		Formula	Authors
Functional richness	FRic	Quickhull algorithm	Villéger et al. 2008
Functional evenness	FEve	$\:\frac{{\sum\:}_{l=1}^{S-1}\text{m}\text{i}\text{n}\left({\text{P}\text{E}\text{W}}_{l,}\frac{1}{S-1}\right)-\frac{1}{S-1}}{1-\frac{1}{S-1}}$	Villéger et al. 2008
Functional divergence	FDiv	$\:\frac{\varDelta\:d\:+\:\stackrel{-}{dG}}{\varDelta\:\left\|d\right\|\:+\:\stackrel{-}{dG}}$	Villéger et al. 2008
Functional dispersion	FDis	$\:\frac{\sum\:{a}_{j}{z}_{j}}{\sum\:{a}_{j}}$	Lalibert and Legendre 2010

3.1. Environmental factors

From July 2020 to June 2022, the mean air temperature recorded was 12.5 ± 10.5°C, with a maximum of 36.8°C (July 27, 2021) and a minimum of -18.6°C (January 8, 2020) (Fig. 2). Total precipitation over this period was 3,015 mm, with the highest rainfall occurring between June and August, consistent with the monsoon climate characteristic of South Korea. There were no significant differences in the average temperature and daily precipitation from 2020 to 2022 (p > 0.05). During the study period, there were no significant differences in dissolved oxygen (DO), water temperature, pH, or electrical conductivity (EC) among the three wetlands (p > 0.05). However, in Phase 3, the water depth in the C wetland (37.5 ± 2.0 cm) was significantly higher than that in the L wetland (33.0 ± 3.0 cm) (p < 0.05).

3.2. Species richness and abundance

3.2.1. BMIs

A total of 66 species and 92,911 individuals of BMIs were identified during the study. The C and S wetlands each supported 48 species, while 41 species were found in the L wetland. The total abundance recorded was 27,669 individuals in the C wetland, 36,963 in the S wetland, and 28,279 in the L wetland. Average species richness was 14.5 (± 3.9) species in the C wetland, 15.4 (± 3.9) species in the S wetland, and 13.8 (± 3.6) species in the L wetland, while average abundance was 1,571.1 (± 762.3) in the C wetland, 2,640.2 (± 1,620.4) in the S wetland, and 2,128.4 (± 1,438.1) in the L wetland (Fig. 4).

3.2.2. Plant species richness and cover

A total of 39 plant species, representing 37 genera across 19 families, were recorded in the surveyed wetlands. Species richness and coverage varied between the wetlands depending on the duration of the drought (Fig. 5). In the July 2021 survey, species richness was highest in the C wetland, with a total of 15 species (12 aquatic, 3 terrestrial), followed by the S wetland with 22 species (15 aquatic, 7 terrestrial), and the L wetland with 28 species (15 aquatic, 13 terrestrial). By November 2021, species richness had decreased across all wetlands, with 11 species (10 aquatic, 1 terrestrial) in both the C and S wetlands, and 9 species (7 aquatic, 2 terrestrial) in the L wetland. The April 2022 survey recorded a further reduction in species richness, with 8 aquatic species in the C wetland, 10 aquatic species in the S wetland, and 8 aquatic species in the L wetland. Species richness in the S and L wetlands was significantly higher than in the C wetland during the July 2021 survey (p < 0.05).

Total plant cover also showed changes over time. In July 2021, total plant cover was 214.5% (210.0% aquatic, 4.5% terrestrial) in the C wetland, 176.5% (158.5% aquatic, 18.0% terrestrial) in the S wetland, and 321.7% (213.0% aquatic, 108.7% terrestrial) in the L wetland. By November 2021, plant cover had decreased to 103.5% (103.0% aquatic, 0.5% terrestrial) in the C wetland, 147.0% (143.5% aquatic, 3.5% terrestrial) in the S wetland, and 140.0% (139.0% aquatic, 1.0% terrestrial) in the L wetland. In April 2022, total plant cover remained relatively stable, with 124.0% (125.0% aquatic) in the C wetland, 107.0% (107.0% aquatic) in the S wetland, and 122.0% (122.0% aquatic) in the L wetland. The plant cover across the three wetlands exhibited a general decline after July 2021, with a convergence towards similar values by November 2022.

The S and L wetlands, which experienced drought, initially exhibited high species diversity and cover due to the emergence of terrestrial plants. However, both species diversity and cover declined over time as the drought persisted. In contrast, the C wetland, which did not experience drought, maintained relatively consistent species diversity and cover throughout the study period.

3.3. BMIs community

Cluster analysis using the Bray-Curtis similarity index revealed three distinct clusters at a 70% similarity threshold (Fig. 6). “Cluster A” included BMI communities from Phase 1 across all wetland conditions (C, S, and L). “Cluster B” comprised communities from Phases 2 and 3 in the C wetland, while “Cluster C” contained communities from Phases 2 and 3 in the S and L wetlands. These results suggest different trajectories of community change in the S and L wetlands compared to the control. When comparing species richness and Jaccard index among the three phases of colonization in the three wetland conditions (C, S, and L), notable differences in species similarity were observed (Table 5). In the C wetland, species richness increased from Phase 1 to Phase 2 due to the introduction of new taxa, despite some species loss. From Phase 2 to Phase 3, species richness slightly decreased (JI = 0.58), indicating a relatively stable community composition over time. In contrast, the S wetland showed a significant decline in species richness from Phase 1 to Phase 2 (JI = 0.59), with substantial changes in community composition from Phase 2 to Phase 3. The L wetland exhibited a marked decrease in species richness from Phase 1 to Phase 2 (JI = 0.33), reflecting a low similarity in community composition before and after the drought period. However, from Phase 2 to Phase 3, there was a slight increase in species richness, accompanied by an increase in similarity to the previous community composition.

Table 5

Jaccard similarity index for benthic macroinvertebrate communities among three wetland treatments at three colonization stages.
Wetland treatments		Phase 1	Phase 2	Phase 3
Control wetland	SR	24	27	25
	CSR		16	19
	JI		0.46	0.58
Short-term drought wetland	SR	33	26	26
	CSR		22	15
	JI		0.59	0.41
Long-term drought wetland	SR	25	23	26
	CSR		12	16
	JI		0.33	0.48
SR (species richness): the number of species in each wetland during each phase; CSR (Common species richness): the number of shared species between the two compared phases; JI (Jaccard index): the similarity index representing the compositional similarity between two phases (calculated as index = CSR/SR₁ + SR₂-CSR)

Non-metric multidimensional scaling (NMS) analysis indicated a pattern of variation in BMIs community composition based on wetland condition and drought duration (Fig. 7). Axis 1 explained 67.3% of the variation, while Axis 2 represented 16.3% of the variation. Electrical conductivity (EC), pH, abundance (A), species richness (R1), and functional evenness (FEve) were key factors distinguishing the three groups: Group 1 (initial-stage communities, Phase 1), Group 2 (Phases 2 and 3 in the C wetland), and Group 3 (Phases 2 and 3 in the S and L wetlands). The multi-response permutation procedure (MRPP) revealed significant differences in BMI communities among the three wetland treatments (C, S, and L) (p < 0.05) (Table 6). Differences were significant between Phases 1 and 2 among all wetlands (p < 0.05). Between Phases 2 and 3, no significant differences were found in the C and L wetlands (p > 0.05), while significant differences were observed in the S wetland (p < 0.05).

Table 6

Differences in species composition among the study sites, as revealed by multi-response permutation procedures analysis.
Groups compared	Test statistic (T)	Chance-corrected within-group agreement (A)	p-value
C1×C2	-3.884	0.240	0.006*
C1×C3	-4.145	0.281	0.006*
C2×C3	-0.110	0.004	0.373
S1×S2	-3.973	0.242	0.005*
S1×S3	-4.180	0.254	0.006*
S2×S3	-2.508	0.088	0.013
L1×L2	-2.936	0.148	0.010*
L1×L3	-3.680	0.223	0.006*
L2×L3	-1.301	0.041	0.103
C2×S2	-2.059	0.083	0.027*
C2×L2	-2.695	0.107	0.014*
S2×L2	-1.375	0.058	0.094
C3×S3	-1.010	0.055	0.151
C3×L3	-1.693	0.087	0.061
S3×L3	0.998	-0.047	0.861
C: control, S: short-term drought treatments, L: long-term drought treatment; 1: initial colonization, 2: mid-colonization, 3: late-colonization.

The analysis of indicator species for each wetland treatment and colonization stage revealed distinct patterns of species presence among the different treatments (Table 7). In the C wetland, Paracercion calamorum was identified as the indicator species in Phase 1, while Anax nigrofasciatus emerged as the indicator species in Phase 2. For the S wetland, four indicator species were identified in Phase 1: Anax parthenope julius, Hydaticus (Prodaticus) grammicus, Molanna moesta, and Sigara (Tropocorixa) substriata. In Phase 2, Physa acuta was identified as the indicator species, and in Phase 3, a species of Dixidae was identified. Meanwhile, in the L wetland, Lepidostoma sinuatum was identified as the indicator species in Phase 3.

Table 7

Indicator values of benthic macroinvertebrates at site during the study period. Indicator values (IVs) are based on the percentage of perfect indication, combining the above values for relative abundance and relative frequency.
Site	Species name	IV	P-value
C-phase 2	Anax nigrofasciatus	47.3	0.004
S-phase 1	Anax parthenope julius	27.2	0.027
	Hydaticus (Prodaticus) grammicus	38.1	0.034
	Molanna moesta	36.6	0.015
	Sigara (Tropocorixa) substriata	43.9	0.019
S-phase 2	Physa acuta	17.4	0.050
S-phase 3	Dixidae sp.	26.5	0.021
L-phase 1	Paracercion calamorum	48.3	0.015
L-phase 3	Lepidostoma sinuatum	37.1	0.015

3.4. Biological traits of BMIs

The functional diversity analysis based on the biological traits of BMIs identified four distinct clusters among the nine groups: “Cluster A” (C1), “Cluster B” (C2 and S1), “Cluster C” (C3 and L1), and “Cluster D” (S2, L2, S3, and L3) (Fig. 8). The patterns of biological traits among these clusters revealed certain differences. Notably, “Cluster D”, which included groups S2, S3, L2, and L3, exhibited a pattern similar to the BMI community clustering results (Fig. 6). The frequency of clingers (LT4) for ‘living type’ (LT) was higher in S2, S3, and L3 within Cluster D compared to the other clusters. The ‘feeding habit’ (FH) showed a higher frequency of piercers (FH2) in Cluster D compared to other clusters. The frequency of ‘voltinism’ was not similar within Clusters A, B, C, and D. When comparing the three wetland treatments, the control wetland showed similar patterns among phases, while S and L drought wetlands exhibited differences between phases 1, 2 and 3. The frequencies of ‘swimming ability’ (SA) and ‘dispersal’ (D) showed similar patterns among all clusters. The frequency of ‘food type’ (FT) revealed a relatively high occurrence of carnivores (FT2) and omnivores (FT3) across all clusters, while herbivores (FT1) were observed at a lower frequency. In Phase 2 of the S and L wetlands, the frequency of herbivores (FT1) declined, followed by a recovery in Phase 3. However, in the L wetland, the recovery of herbivores (FT1) was limited, as certain species did not reappear, resulting in a lower recovery rate compared to the S wetland.

3.5. Functional diversity

We compared species richness, abundance, community indices, and functional diversity (FD) indices among the three wetland treatments at each stage (Table 8). Species richness (SR) showed no significant differences among the C, S, and L wetlands and exhibited a general decreasing trend over time (p > 0.05). The abundance (A) in the C wetland showed no significant differences among stages (p > 0.05), while the S and L wetlands exhibited significant differences. In Phase 3 of the S wetland, the abundance was more than four times higher than in Phase 1 (p < 0.05). Additionally, in Phase 3, the abundance in the S and L wetlands was significantly higher than in the C wetland (p < 0.05).

Table 8

Comparison of average species richness, abundance, community (DI, H’, J’, and R1) and functional diversity (FRic, FEve, FDiv, and FDis) indices (one-way ANOVA, *p < 0.05*; multiple pairwise comparison, Tukey HSD).
	Phase 1			Phase 2			Phase 3
	C	S	L	C	S	L	C	S	L
SR	10.6 ^a (± 1.2)	10.3 ^a (± 1.9)	8.8 ^a (± 1.9)	9.6 ^a (± 2.5)	8.8 ^a (± 2.3)	8.6 ^a (± 2.2)	8.5 ^a (± 1.9)	8.9 ^a (± 2.2)	8.9 ^a (± 2.7)
A	401.9 ^ab (± 362.1)	350.8 ^abc (± 303.4)	252.7 ^a (± 106.3)	471.4 ^ab (± 306.9)	895.2 ^bcd (± 664.7)	748.0 ^abc (± 445.1)	482.3 ^ab (± 408.9)	1330.0 ^d (± 535.1)	1149.9 ^d (± 635.8)
DI	0.73 ^a (± 0.10)	0.74 ^a (± 0.12)	0.61 ^a (± 0.10)	0.70 ^a (± 0.08)	0.75 ^a (± 0.14)	0.70 ^a (± 0.08)	0.66 ^a (± 0.14)	0.70 ^a (± 0.13)	0.78 ^a (± 0.07)
H’	1.38 ^a (± 0.24)	1.32 ^a (± 0.25)	1.54 ^a (± 0.18)	1.36 ^a (± 0.24)	1.25 ^a (± 0.41)	1.40 ^a (± 0.16)	1.49 ^a (± 0.31)	1.32 ^a (± 0.27)	1.24 ^a (± 0.15)
J’	0.60 ^a (± 0.13)	0.60 ^a (± 0.11)	0.73 ^a (± 0.13)	0.62 ^a (± 0.10)	0.58 ^a (± 0.19)	0.67 ^a (± 0.09)	0.69 ^a (± 0.16)	0.60 ^a (± 0.16)	0.57 ^a (± 0.07)
R1	1.52 ^b (± 0.40)	1.34 ^b (± 0.44)	1.29 ^ab (± 0.40)	1.36 ^ab (± 0.29)	1.29 ^a (± 0.35)	1.24 ^a (± 0.47)	1.37 ^ab (± 0.42)	1.12 ^a (± 0.25)	1.29 ^a (± 0.37)
FRic	0.046 ^b (± 0.010)	0.047 ^b (± 0.022)	0.037 ^ab (± 0.018)	0.027 ^ab (± 0.017)	0.045 ^ab (± 0.018)	0.037 ^ab (± 0.021)	0.022 ^a (± 0.017)	0.033 ^ab (± 0.016)	0.037 ^ab (± 0.019)
FEve	0.371 ^a (± 0.088)	0.427 ^a (± 0.065)	0.452 ^a (± 0.156)	0.381 ^a (± 0.127)	0.444 ^a (± 0.257)	0.327 ^a (± 0.170)	0.429 ^a (± 0.177)	0.387 ^a (± 0.156)	0.463 ^a (± 0.139)
FDiv	0.919 ^c (± 0.032)	0.907 ^bc (± 0.081)	0.886 ^abc (± 0.048)	0.860 ^abc (± 0.069)	0.864 ^abc (± 0.090)	0.788 ^a (± 0.127)	0.818 ^abc (± 0.071)	0.803 ^abc (± 0.102)	0.778 ^a (± 0.125)
FDis	0.241 ^a (± 0.072)	0.242 ^a (± 0.081)	0.288 ^a (± 0.045)	0.275 ^a (± 0.082)	0.275 ^a (± 0.040)	0.297 ^a (± 0.026)	0.281 ^a (± 0.054)	0.254 ^a (± 0.052)	0.298 ^a (± 0.043)
Different letters indicate statistically significant differences (one-way ANOVA test). SR, species richness; A, abundance; DI, dominance index; H, diversity index; J’, evenness index; R1, richness index; FRic, functional richness; FEve, functional evenness; FDiv, functional divergence; FDis, functional dispersion.

The indices DI, H, and J showed no significant differences among the C, S, and L wetlands (p > 0.05). The R1 of C, S and L wetlands exhibited a decreasing trend, and in Phase 1 of the S wetland, the R1 index was significantly higher than in Phases 2 and 3 (p < 0.05). FRic of the C, S and L wetlands tended to decrease or remain stable as the stages progressed. However, there was a significant difference between Phase 1 and 3 in the C wetland (p < 0.05). The FDiv, FEve, and FDis showed no significant differences across the C, S, and L wetlands (p > 0.05).

4.1 Environmental factors and their influence on wetland conditions

The temperature and precipitation during the study period were consistent with Korea’s typical monsoon climate, with no significant differences among the three wetland conditions, indicating that climatic variability was unlikely to have directly impacted the wetland ecosystems (Kim et al. 2014b). Water level fluctuations in the L wetland during the drought period may be closely associated with the proliferation of terrestrial plants and their subsequent influence on soil characteristics (Xu et al. 2015).

As drought conditions persisted, terrestrial plants likely adapted by extending their root systems deeper and wider in search of water, absorbing more moisture from the soil and releasing it into the atmosphere through transpiration (Zambrano et al. 2019; Mohammadi Alagoz et al. 2023). For example, Echinochloa crus-galli, which showed high coverage in the L wetland, may have contributed to the reduction in water levels during the prolonged drought through its well-developed root structure and effective water absorption (Hu et al. 2023). Additionally, changes in soil structure due to root growth may have created a more porous environment, facilitating water infiltration into the ground and further reducing surface water (Liu et al. 2020; Wang et al. 2022). In contrast, in the C and S wetlands, aquatic plants such as Iris ensata, Veronica anagallis-aquatica, Nymphoides peltata, and Equisetum hyemale, which were identified as dominant species, likely contributed more effectively to maintaining water levels by reducing evaporation through their water retention capabilities (Ahmad et al. 2016; Gaballah et al. 2020; Choung et al. 2021). In conclusion, despite consistent climatic conditions across the sites, differences in water levels between the wetlands were likely influenced by the spread of terrestrial plants, resulting changes in soil structure, and the role of aquatic plants. However, as environmental variables such as soil moisture and transpiration were not directly measured, further detailed analysis is necessary.

4.2 Differences in species richness and community composition among the three wetland treatments

This study analyzed the differences in community structure and recovery processes of BMIs in among three wetlands, to assess the impact of drought intensity and duration on wetland ecosystems. The results indicate that BMI communities remained stable in the C wetland, while certain changes in community structure were observed in the S and L wetlands.

The BMI community structure in the C wetland remained stable throughout the study period. During the transition from Phase 1 to Phase 2, new species were introduced, leading to an increase in species diversity despite the loss of some species. In the transition to Phase 3, there was a slight loss of species, but both the Bray-Curtis similarity and Jaccard index indicated high similarity, suggesting minimal community variation within the C wetland. These results suggest that aquatic ecosystems, in the absence of external disturbances, can maintain long-term stability (Bogan et al. 2015). This finding aligns with previous studies that have demonstrated community structure in stable habitats remains relatively unchanged over time (Hillebrand et al. 2008).

The NMS analysis revealed that the C wetland exhibited relatively higher functional evenness compared to the S and L wetlands. This finding aligns with previous research indicating that functional diversity tends to be maintained in stable environments (Loreau et al. 2001). Studies emphasizing the importance of functional evenness in stable ecosystems further support these results. Adler et al. (2007) highlighted that functional diversity contributes to ecosystem stability and resilience, playing a key role in regulating species populations through interspecific interactions. For example, indicator species analysis identified Anax nigrofasciatus as a key indicator species in Phase 2. As a top predator, this species is sensitive to specific habitat conditions, and its presence as an indicator suggests that the C wetland provided suitable conditions with abundant prey resources for sustaining higher trophic levels. While species such as chironomids, Radix auricularia, Cloeon dipterum, and Ischnura asiatica were abundant in the C wetland, their population levels were relatively lower compared to those in the S and L wetlands. his suggests that the presence of top predators, such as A. nigrofasciatus, facilitated stronger interspecific interactions, preventing excessive population growth of certain species and contributing to the overall stability of the community (Adler et al. 2007).

The S wetland experienced changes in BMIs communities due to drought, with a decline in species richness after the drought. However, the similarity between pre- and post-drought communities was higher compared to the L wetland. This can be attributed to the survival and rapid recolonization of drought-tolerant species, such as Physa acuta, which remained resilient after the drought (Gulanicz et al. 2018). As P. acuta lacks flight capabilities, it does not migrate to other habitats during drought but instead survives by burrowing into the mud or reducing its metabolic activity (Aspin et al. 2018). This response aligns with the pulse disturbance theory, which explains how certain species rapidly recolonize habitats following disturbances (Bender et al. 1984; Lake 2003). Additionally, species that have previously experienced drought are generally regarded as relatively resistant or resilient to predictable drying disturbances, a pattern similar to post-drought recovery observed in intermittent streams (Bogan et al. 2015). This similarity suggests that the mechanisms of resilience in wetland ecosystems may be comparable to those observed in stream ecosystems.

As the community progressed into Phase 3, the BMIs of abundance increased more than fourfold compared to Phase 1, indicating rapid recovery following the short-term drought. During this recovery, Dixidae sp. was identified as an indicator species, suggesting that the introduction of new species played a significant role in community recovery. Dixidae sp. likely functioned as an early colonizer and decomposer, rapidly establishing itself after the disturbance, in line with the pulse disturbance theory (Bender et al. 1984). By decomposing organic material, Dixidae sp. likely contributed to nutrient cycling and supported the recovery of ecosystem functions (Gullan and Cranston 2014). This process had a significant impact on lower trophic levels, leading to increased reproduction of aquatic insects and creating favorable conditions for higher-level predators, such as dragonflies, to recolonize the habitat (Boulton and Lake 1992). As a result, a total of 10 species of Odonata were identified. This suggests that decomposers like Dixidae sp. played a key role in restructuring ecosystem functions during early habitat recovery, providing the foundation for top predators to reoccupy the ecosystem (Gullan and Cranston 2014). Furthermore, Adámek et al. (2022) demonstrated that Diptera larvae maintained high species richness and abundance in semi-aquatic habitats due to their ability to breathe air directly through spiracles, independent of dissolved oxygen levels. This adaptability aligns with our findings, where Dixidae sp. played a pivotal role in early habitat recovery, indicating a similar pattern between the two studies. Additionally, MRPP analysis revealed significant differences in community structure between Phase 2 and Phase 3 in the S wetland, indicating ongoing changes as the community continued to recover from the drought.

The L wetland experienced the most severe community changes and recovery processes. Although species loss was lower than in the S wetland, the L wetland exhibited the greatest dissimilarity in community structure compared to its pre-drought state. This reflects the extreme environmental conditions that long-term drought imposed on sensitive species. Many aquatic macroinvertebrates struggled to survive under prolonged drought, leading to a sharp decline in community similarity. These findings align with studies by Dallas and Rivers-Moore (2014) and Stuart et al. (2020), both of which emphasize the significant impact of long-term drought on aquatic species survival and the resulting changes in community structure. In Phase 2, Chironomidae emerged as the dominant taxon. Chironomidae are well-adapted to low-oxygen and stagnant water environments, and their rapid reproductive capacity allows them to quickly colonize new habitats (Kim et al. 2014b; Martel-Cea et al. 2021; Saffarinia et al. 2022). These traits likely contributed to the high dominance of Chironomidae in the L wetland following the drought. As the system progressed into Phase 3, some species diversity was restored; however, the effects of the prolonged drought remained evident. Lepidostoma sinuatum was identified as a key indicator species in Phase 3, suggesting that it is a drought-tolerant species capable of adapting to the altered environmental conditions. Similarly, Whiles et al. (1993) study found that Lepidostoma rapidly recolonized disturbed habitats as an early colonizer, playing a pivotal role in the reassembly of communities following disturbance. This observation supports the conclusion that L. sinuatum was able to adapt to the post-drought habitat changes, contributing significantly to the recovery of the community structure (Ehlers et al. 2020).

Overall, the varying drought conditions demonstrated that species with different levels of resilience and adaptability within the BMI community were selected, ultimately shaping the community structure (Boulton and Lake 1992; Bogan and Lytle 2011). These findings highlight the importance of understanding the impacts of drought on ecosystems and underscore the need to consider species-specific adaptive strategies in ecosystem management (Lytle and Poff 2004).

4.3 Functional Responses of BMIs to Drought Disturbance

Drought serves as an environmental filter, selecting for biological traits suited to specific conditions while excluding others (Southwood 1977; Poff 1997; Vineetha and Nandan 2021). This study compared the functional traits of BMIs among the three wetlands, revealing notable changes in functional traits in response to drought, particularly in ‘feeding habits’, ‘living types’, ‘food types’, and ‘voltinism’. These results are consistent with previous studies, emphasizing that drought creates favorable conditions for species with specific biological traits, thereby altering both functional diversity and community composition in ecosystems (Bogan et al. 2015; Apsin et al. 2018; Vineetha and Nandan, 2021)

The ‘feeding habits’ of BMIs showed some differences in specific traits across the three wetlands. In the C wetland, functional traits remained relatively stable, with an increase in predator (FH4) frequency as food resources became more abundant. In environments with abundant food resources, predators tend to dominate within the community (Ledger et al. 2008). In contrast, the S and L wetlands, both affected by drought, exhibited an increase in the frequency of piercers (FH2) and a decline in predators (FH4). The rise in piercers (FH2) suggests their ability to efficiently utilize resources in the resource-limited conditions caused by drought (Chase 2007; Chester and Robson, 2011). This demonstrates that in habitats with intensified resource limitations, certain functional groups, such as piercers, are able to occupy more favorable conditions for survival.

The ‘living types’ of BMIs responded differently to changes in habitat conditions caused by drought. In the C wetland, living types remained generally stable, while in the S and L wetlands, drought-induced changes in the habitat led to distinct responses based on species traits. Species highly dependent on the physical structure of the habitat, such as climbers (LT2), experienced a decline in frequency during the drought but showed a tendency to increase as the habitat recovered. The clingers (LT4), which depend on stable substrates, exhibited higher frequencies in the S and L wetlands. This increase could be attributed to the presence of terrestrial plants in these wetlands, which likely enhanced the structural complexity of the habitat, creating more surfaces for clingers to attach to. The additional physical habitat structure provided by terrestrial plants likely facilitated improved habitat conditions for these species, promoting their recolonization after the drought. This added complexity in the habitat may have created more favorable environments for clingers, enabling them to re-establish more effectively in the post-drought period. The swimmers (LT5), primarily represented by Coleoptera, were less sensitive to physical habitat changes due to their strong mobility and aerial dispersal abilities (Vineetha and Nandan 2021). These observations highlight how species’ responses varied based on their living types, with ecological traits and habitat dependency playing key roles in determining their resilience or vulnerability to drought-induced environmental changes.

The changes in resource availability among three wetlands were reflected in the ‘food type’ of BMIs. In the C wetland, where resource availability remained stable, no significant changes in feeding habits were observed. The herbivorous species (FT1), which rely on specific plant resources, are more sensitive to environmental stress compared to carnivorous and omnivorous species. In environments experiencing stressors such as drought, plant resources can rapidly decline, leaving herbivores with limited or no alternative food sources, making them more vulnerable to these stress factors (Cummins 1973). In contrast, carnivores and omnivores can exploit a broader range of food resources, enabling them to survive more consistently in resource-limited environments factors (Cummins 1973).

In the C wetland, multivoltine (V3) were relatively dominant, while semivoltine (V1) maintained a stable presence. However, in the S and L wetlands, multivoltine (V3) rapidly occupied the habitat following the drought, indicating their ability to reproduce quickly under stress conditions due to their shorter life cycles (Williams 2006; Herbst et al. 2019). In the S wetland, semivoltine (V1) also managed to recolonize as the habitat recovered. In contrast, in the L wetland, semivoltine (V1) nearly disappeared even during the recovery phase, with multivoltine (V3) dominating the community. This pattern aligns with the findings of Kim et al. (2014b) and Vineetha and Nandan (2021) which suggest that long-term drought creates conditions favorable for species with shorter life cycles, allowing them to reproduce more rapidly. As a result, these conditions can reduce ecological diversity by favoring species that are able to exploit such environments, ultimately leading to a decline in the overall biodiversity of the habitat.

4.4 Impact of drought on BMI community and functional diversity indices

Species richness showed a declining trend in all wetlands throughout the study period, whereas species evenness exhibited different patterns among three wetlands (Table 8). In the C wetland, evenness gradually increased as the BMIs community stabilized. The increase in species evenness observed in stabilized habitats can be attributed to the decline or disappearance of dominant species, as confirmed in our study within the C wetland (Brown et al. 2018).

In contrast, the S wetland exhibited a temporary decline in species evenness following the drought, followed by an increase during the community recovery phases. This pattern aligns with the findings of Ledger et al. (2013), which suggest that, in the early stages of post-drought recovery, certain functional groups may dominate the community. However, as time passes and the community continues to recover, species evenness gradually increases as more species reestablish and diversify within the habitat. Meanwhile, in the L wetland, species evenness continued to decline even after the drought. This may be the result of prolonged drought conditions, which allowed certain species to maintain dominance, limiting the recolonization of other species and leading to a gradual simplification of the community structure.

In terms of trait-based indices (Table 8), both the C and S wetlands showed a gradual decline in FRic as the community stabilized. This reflects a reduction in competition within stable habitats, allowing certain functional groups to become dominant and leading to a decrease in functional richness. In the L wetland, community instability may have led to the emergence of new ecological niches, potentially facilitating species turnover. As a result of the drought, functional diversity declined, with certain species becoming dominant and creating opportunities for new species to occupy available space. However, this turnover process is more likely to contribute to a simplification of the community rather than a recovery of functional diversity. Herbst et al. (2019) also noted that such conditions can lead to the rapid colonization of short-generation species, further destabilizing the community structure.

FEve in the L wetland initially declined following the drought but showed an increasing trend during the community recovery phase. This may be due to species from specific functional groups (e.g., semivoltine) dominating the community by exploiting available resources. Over time, however, other functional groups (e.g., predators) gradually returned, regulating excessive population growth and promoting more balanced species interactions within the community. This pattern is consistent with the findings of Ledger et al. (2011).

Roscher et al. (2014) demonstrated that in disturbed functional spaces, species that benefit from niche availability rapidly occupy those spaces, leading to an initial increase in functional evenness (FEve). However, as dominance by certain functional groups resumes, functional evenness can subsequently decrease. Similarly, in this study, the S wetland exhibited a temporary increase in FEve following the drought, likely due to the rapid colonization of species with diverse functional traits. Over time, however, specific species, such as Radix auricularia, became dominant, leading to a decline in evenness.

Environmental disturbances such as prolonged droughts can significantly limit functional diversity within ecosystems by allowing specific functional groups to dominate. According to the theory of FDiv explained by Villéger et al. (2008), FDiv measures how species are distributed in functional trait space, with higher values indicating a more even distribution of species across diverse functional roles. In the L wetland, the prolonged drought (8 months) created conditions that favored a few species, which monopolized resources and dominated the community. As a result, other species were unable to thrive or perform their functional roles, leading to a reduction in overall functional diversity. Consequently, species became more concentrated around the functional center, decreasing FDiv.

In both the C and L wetlands, FDis showed a gradual increase, whereas in the S wetland, FDis increased during the early post-drought recovery phase but declined in Phase 3. This trend suggests that in the S wetland, species with varying functional traits initially colonized the habitat during recovery, but over time, a few dominant species stabilized the community, reducing functional trait diversity and FDis. According to Laliberté and Legendre (2010), FDis measures species' relative abundances and functional distances. As certain species dominate, functional trait diversity decreases, resulting in lower FDis. In contrast, while L wetland exhibited a rising trend in FDis, this does not necessarily imply greater community resilience or stability. The increase may reflect the dispersion of specific functional traits rather than community recovery. Long-term ecosystem recovery involves complex factors, including species interactions, survival rates, and functional role diversity. Therefore, the increasing FDis in L wetland alone cannot conclude that the community has achieved stability or resilience. Instead, it emphasizes the need to interpret functional diversity indices in the context of broader ecological factors. Moreover, despite the L wetland showing higher FRic, FEve, and FDis than the S wetland, it is difficult to conclude that L wetland is more stable or biodiverse. The Jaccard similarity index revealed that L wetland had the lowest similarity to its pre-drought community, despite showing species composition similarities with S wetland. Differences in functional trait groups further suggest distinct recovery trajectories between the wetlands.

The findings of this study have important implications for wetland management and conservation in the context of climate change. The observed changes in functional diversity and community structure, particularly after long-term drought, suggest that increasing drought frequency could lead to significant long-term alterations in wetland ecosystems. Conservation strategies should focus on protecting drought-vulnerable functional groups and enhancing ecosystem resilience. However, it's crucial to note that while our mesocosm experiment provides valuable insights, it may not capture all the complexities of natural wetland systems. Field studies are necessary to validate these findings and to understand how factors not represented in our mesocosms, such as larger spatial scales and more diverse species interactions, might influence drought responses. Future research should focus on long-term field studies of wetland responses to drought, incorporating a wider range of environmental variables and species interactions. Additionally, investigating the genetic and physiological mechanisms underlying species' drought tolerance could provide further insights into ecosystem resilience.

In conclusion, this study underscores the complex dynamics of wetland ecosystem recovery following drought disturbances. Our findings highlight the need for adaptive management strategies that consider both short-term resilience and long-term shifts in community structure and function. As climate change continues to alter hydrological patterns, understanding these ecological responses will be crucial for effective wetland conservation and the preservation of their vital ecosystem services.

This study investigated the responses of BMI communities to short- and long-term drought conditions in wetland mesocosms, focusing on the dynamics of functional diversity and species recovery. Although the L wetland exhibited higher values for functional diversity indices such as FRic, FEve, and FDis compared to the S wetland, these indices alone do not confirm greater ecosystem stability or resilience. The Jaccard similarity index revealed that the L wetland had the lowest similarity to its pre-drought community, despite sharing species composition similarities with the S wetland. This suggests that the prolonged drought led to significant changes in species composition in the L wetland, affecting its ability to recover fully. Differences in functional trait groups between the wetlands highlight the distinct recovery trajectories, with some species adapting more successfully to post-drought conditions. The findings underscore the complexity of wetland ecosystem recovery processes, where functional diversity indices must be interpreted in the context of species interactions and environmental conditions. Given the variations observed in species recovery and functional diversity, long-term monitoring of wetlands affected by drought is essential. Such studies should focus on changes in species composition, functional traits, and biodiversity to better understand the intricate dynamics of wetland recovery and inform future conservation and management strategies.

Conflict of interest

The authors declare that they have no conflicts of interest.

Ethical approval

No ethical approval was required for this study as experimental work was conducted with an unregulated invertebrate species.

Funding

This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (2019R1I1A2A01042109).

Author Contribution

Data curation, Formal analysis and Investigation, Min Kyung Kim and Dong Gun Kim; Project administration and Supervision, Dong Gun Kim; Writing – original draft, Min Kyung Kim; Writing – review and editing, Tae Joong Yoon and Dong Gun Kim.

Acknowledgement

We thank Prof. Sun Hee Hong (Hankyong National University) for assistance in data analysis.

Data Availability

Data supporting the findings of this study are available from the corresponding author upon reasonable request.

Adámek Z, Konečná J, Karásek P, Všetičková L, Podhrázská J, Zajíček A (2022) From perennial to intermittent headwater stream: macroinvertebrate community response to climate-induced summer water scarcity. Aquatic Ecol 56(4):937–950.
Adler PB, HilleRisLambers J, Levine JM (2007) A niche for neutrality. Ecol Lett 10(2):95–104.
Ahmad A, Hadi F, Ali N, Jan AU (2016) Enhanced phytoremediation of cadmium polluted water through two aquatic plants Veronica anagallis-aquatica and Epilobium laxum. Environ Sci Pollut Res 23:17715–17729.
Aspin TW, Khamis K, Matthews TJ, Milner AM, O’Callaghan MJ, Trimmer M, Woodward G, Ledger ME (2019) Extreme drought pushes stream invertebrate communities over functional thresholds. Glob Chang Biol 25(1):230–244.
Bêche LA, Mcelravy EP, Resh VH (2006) Long-term seasonal variation in the biological traits of benthic-macroinvertebrates in two Mediterranean-climate streams in California, USA. Freshw Biol 51(1):56–75.
Bender A, Case TJ, Gilpin ME (1984) Perturbation experiments in community ecology: theory and practice. Ecology 65:1–13.
Bogan MT, Lytle DA (2011) Severe drought drives novel community trajectories in desert stream pools. Freshw Biol 56(10):2070–2081.
Bogan MT, Boersma KS, Lytle DA (2015) Resistance and resilience of invertebrate communities to seasonal and supraseasonal drought in arid-land headwater streams. Freshw Biol 60(12):2547–2558.
Bonacina L, Fasano F, Mezzanotte V, Fornaroli R (2023) Effects of water temperature on freshwater macroinvertebrates: A systematic review. Biol Rev 98(1):191–221.
Boulton AJ, Lake PS (1992) The ecology of two intermittent streams in Victoria, Australia: II. Comparisons of faunal composition between habitats, rivers, and years. Freshw Biol 27(1):99–121.
Braun-Blanquet J (1964) Pflanzensoziologie: Grundzüge der Vegetationskunde. Springer, Berlin.
Brown LE, Khamis K, Wilkes M, Blaen P, Brittain JE, Carrivick JL, Fell S, Friberg N, Füreder L, Gislason GM, Hainie S, Hannah DM, James WH, Lencioni V, Olafsson JS, Robinson CT, Saltveit SJ, Thompson C, Milner AM (2018) Functional diversity and community assembly of river invertebrates show globally consistent responses to decreasing glacier cover. Nat Ecol Evol 2(2):325–333.
Buikema JAL, Voshell JJR (1993) Toxicity studies using freshwater benthic macroinvertebrates. In: Rosenberg, DM, Hesh, VH (eds.), Freshwater biomonitoring and benthic macroinvertebrates. Routledge, Chapman & Hall, New York, pp. 344–398.
Burkett V, Kusler J (2000) Climate change: potential impacts and interactions in wetlands of the United States. JAWRA J Am Water Resour Assoc 36(2):313–320.
Canullo R, Allegrini MC, Campetella G (2012) Reference field manual for vegetation surveys on the Conecofor Lii network, Italy (National Programme of Forest Ecosystems Control—Unece, ICP Forests). Braun-Blanquetia 48:5–65.
Cayrou J, Céréghino R (2005) Life-cycle phenology of some aquatic insects: Implications for pond conservation. Aquat Conserv -Mar Freshw Ecosyst 15(6):559–571.
Chase JM (2007) Drought mediates the importance of stochastic community assembly. Proc Natl Acad Sci 104(44):17430–17434.
Cherry JA (2011) Ecology of wetland ecosystems: water, substrate, and life. Nat Educ Knowl 3(10):16.
Chester ET, Robson BJ (2011) Drought refuges, spatial scale and recolonisation by invertebrates in non-perennial streams. Freshw Biol 56(10):2094–2104.
Choung Y, Min BM, Lee KS, Cho KH, Joo KY, Hyun JO, Nam GH, Kim JS, Cho SY, Lee JS, Jung SY, Lee J (2021) Categorized wetland preference and life forms of the vascular plants in the Korean Peninsula. J Ecol Environ 45:1–6.
Cummins KW (1973) Trophic relations of aquatic insects. Annu Rev Entomol 18(1):183–206.
Dallas HF, Rivers-Moore N (2014) Ecological consequences of global climate change for freshwater ecosystems in South Africa. S Afr J Sci 110:1–11.
De Szalay, FA, Batzer, DP, Resh, VH (1996) Mesocosm and macrocosm experiments to examine effects of mowing emergent vegetation on wetland invertebrates. Environ Entomol 25(2):303–309.
Díaz SM, Settele J, Brondízio E, Ngo H, Guèze M, Agard J, Arneth A, Balvanera P, Brauman K, Butchart S, Chan KM, Garibaldi LA, Ichii K, Liu J, Subramanian S, Midgley G, Miloslavich P, Molnár Z, Obura D, Pfaff A, Polasky S, Purvis A, Razzaque J, Reyers B, Roy CR, Shin Y, Visseren-Hamakers I, Willis K, Zayas C (2019) The global assessment report on biodiversity and ecosystem services: Summary for policy makers IPBES, Bonn.
Ding N, Yang W, Zhou Y, Gonzalez-Bergonzoni I, Zhang J, Chen K, Vidal N, Jeppesen E, Liu Z, Wang B (2017) Different responses of functional traits and diversity of stream macroinvertebrates to environmental and spatial factors in the Xishuangbanna watershed of the upper Mekong River Basin, China. Sci Total Environ 574:288–299.
Domisch S, Portmann FT, Kuemmerlen M, O'Hara RB, Johnson RK, Davy-Bowker J, Baekken T, Zamora-Muñoz C, Sáinz-Bariáin M, Bonada N, Haase P, Döll P, Jähnig SC (2017) Using streamflow observations to estimate the impact of hydrological regimes and anthropogenic water use on European stream macroinvertebrate occurrences. Ecohydrology 10(8).
Ehlers SM, Al Najjar T, Taupp T, Koop JH (2020) PVC and PET microplastics in caddisfly (Lepidostoma basale) cases reduce case stability. Environ Sci Pollut Res 27:22380–22389.
Gaballah MS, Abdelwahab O, Barakat KM, Aboagye D (2020) A novel horizontal subsurface flow constructed wetland planted with Typha angustifolia for treatment of polluted water. Environ Sci Pollut Res 27:28449–28462.
Gagic V, Bartomeus I, Jonsson T, Taylor A, Winqvist C, Fischer C, Slade EM, Steffan-Dewenter I, Emmerson M, Potts SG, Tscharntke T, Weisser W, Bommarco R (2015) Functional identity and diversity of animals predict ecosystem functioning better than species-based indices. Proc Royal Soc B 282:20142620.
Goswami M, Bhattacharyya P, Mukherjee I, Tribedi P (2017) Functional diversity: an important measure of ecosystem functioning. Adv Microbiol 7:82–93.
Gulanicz T, Kobak J, Poznańska-Kakareko M (2018) Effects of water level fluctuations and substratum drying on the survival and behaviour of the invasive freshwater snail Physa acuta Draparnaud, 1805. Mar Freshw Res 69(9):1389–1396.
Gullan PJ, Cranston PS (2014) The insects: an outline of entomology. John Wiley & Sons, Hoboken.
Herbst DB, Cooper SD, Medhurst RB, Wiseman SW, Hunsaker CT (2019) Drought ecohydrology alters the structure and function of benthic invertebrate communities in mountain streams. Freshw Biol 64(4):567–580.
Hillebrand H, Bennett DM, Cadotte MW (2008) Consequences of dominance: a review of evenness effects on local and regional ecosystem. Ecology 89(6):1510–1520.
Hockaday AC, Leon AS, Patterson K, Pennings SC (2024) Freshwater wetlands for flood control: How manipulating the hydroperiod affects plant and invertebrate communities. PLoS One 19(7).
Hu X, Liu R, Mao H, Xu Y, Chen B, Li Y, Yang X (2023) Inter-Species Investigation of Biological Traits among Eight Echinochloa Species. Plants 12(17):3085.
Jaccard P (1900) Contribution an Problème de l’immigration post-glaciaire de la flore alpine. Bull Soc Vaudoise Sci Nat 36:87–130.
Jaccard P (1901) Etude compararative de la distribution florale dans une portion des Alpes et du Jura. Bull Soc Vaudoise Sci Nat 37:647–579.
Keast A (1984) The introduced aquatic Myriophyllum spicatum as habitat for fish and their invertebrate prey. Can J Zool 62:1289–1303.
Kim DG, Kang HJ, Baek MJ, Lee CY, Kim JG, Bae YJ (2014a) Analyses of benthic macroinvertebrate colonization during the early successional phases of created wetlands in temperate Asia. Fundam Appl Limnol 184:35–49.
Kim DG, Lee CY, Choi LJ, Kang HJ, Baek MJ, Kim JG, Bae YJ (2014b) Drought effects on the colonization of benthic macroinvertebrate communities in the early successional phases in experimental mesocosm wetlands. J Freshw Ecol 29(4):507–524.
Laliberté E, Legendre P (2010) A distance-based framework for measuring functional diversity from multiple traits. Ecology 91:299–305.
Laliberté E, Legendre P, Bill S (2015) Measuring functional diversity (FD) from multiple traits, and other tools for functional ecology. R package.
Lake PS (2003) Ecological effects of perturbation by drought in flowing waters. Freshw Biol 48:1161–1172.
Ledger ME, Edwards FK, Brown LE, Milner AM, Woodward G (2011) Impact of simulated drought on ecosystem biomass production: an experimental test in stream mesocosms. Glob Chang Biol 17(7):2288–2297.
Leigh C, Bonada N, Boulton AJ, Hugueny B, Larned ST, Vander Vorste R, Datry T (2016) Invertebrate assemblage responses and the dual roles of resistance and resilience to drying in intermittent rivers. Aquat Sci 78:291–301.
Liu Q, Liu J, Liu H, Liang L, Cai Y, Wang X, Li C (2020) Vegetation dynamics under water-level fluctuations: Implications for wetland restoration. J Hydrol 581:124418.
Loreau M, Naeem S, Inchausti P, Bengtsson J, Grime JP, Hector A, Raffaelli D, Schmid B, Tilman D, Wardle DA (2001) Biodiversity and ecosystem functioning: current knowledge and future challenges. Science 294(5543):804–808.
Lytle DA, Poff NL (2004) Adaptation to natural flow regimes. Trends Ecol Evol 19:94–100.
National Institute of Biological Resources (2023) National list of species of Korea. Retrieved Oct. 10, 2024, available from http://species.nibr.go.kr/index.do.
Martel-Cea A, Astorga GA, Hernández M, Caputo L, Abarzúa AM (2021) Modern chironomids (Diptera: Chironomidae) and the environmental variables that influence their distribution in the Araucanian lakes, south-central Chile. Hydrobiologia 848:2551–2568.
Mason NWH, Mouillot D, Lee WG, Wilson JB (2005) Functional richness, functional evenness and functional divergence: the primary components of functional diversity. Oikos, Copenhagen, 111:112–118.
Margalef R (1958) Information theory in Ecology. Gen Syst 3:36–71.
McCune B, Grace JB, Urban DL (2002) Analysis of ecological communities. MJM Software Design, Gleneden Beach, Oregon, USA.
McNaughton SJ (1967) Relationships among functional properties of Californian grassland. Nature 216:168–169.
Mohammadi Alagoz S, Zahra N, Hajiaghaei Kamrani M, Asgari Lajayer B, Nobaharan K, Astatkie T, Siddique KHM, Farooq M (2023) Role of root hydraulics in plant drought tolerance. J Plant Growth Regul 42(10):6228–6243.
Mouchet MA, Villéger S, Mason NWH, Mouillot D (2010) Functional diversity measures: an overview of their redundancy and their ability to discriminate community assembly rules. Funct Ecol 24:867–876.
Pacifici M, Visconti P, Butchart SH, Watson JE, Cassola FM, Rondinini C (2017) Species’ traits influenced their response to recent climate change. Nat Clim Chang 7(3):205–208.
Papas P (2007) Effect of macrophytes on aquatic invertebrates: A literature review. In: Technical Report Series, Arthur Rylah Institute for Environmental Research, Heidelberg, VIC, Australia, vol 158, p 29.
Petchey OL, Gaston KJ (2006) Functional diversity: back to basics and looking forward. Ecol Lett 9:741–758.
Pielou EC (1966) Shannon’s formula as a measure of specific diversity: its use and misuse. Am Nat 100:463–465.
Poff NL (1997) Landscape filters and species traits: Towards mechanistic understanding and prediction in stream ecology. J N Am Benthol Soc 16:391–409.
Poff NL, Olden JD, Vieira NKM, Finn DS, Simmons MP, Kondratieff BC (2006) Functional trait niches of North American lotic insects: traits-based ecological applications in light of phylogenetic relationships. J N Am Benthol Soc 25:730–755.
Saffarinia P, Anderson KE, Herbst DB (2022) Effects of experimental multi-season drought on abundance, richness, and beta diversity patterns in perennially flowing stream insect communities. Hydrobiologia 849(4):879–897.
Schriever TA, Bogan MT, Boersma KS, Cañedo-Argüelles M, Jaeger KL, Olden JD, Lytle DA (2015) Hydrology shapes taxonomic and functional structure of desert stream invertebrate communities. Freshw Sci 34(2):399–409.
Shannon CE (1948) A mathematical theory of communication. Bell Syst Tech J 27:379–423.
Southwood TRE (1977) Habitat, the templet for ecological strategies? J Anim Ecol 46:337–365.
Stewart BA, Close PG, Cook PA, Davies PM (2013) Upper thermal tolerances of key taxonomic groups of stream invertebrates. Hydrobiologia 718:131–140.
Stuart RE, Ingram T, Closs GP (2021) Recolonisation of a fish and invertebrate community in a wetland following a drought: the importance of deep water refugia. N Z J Mar Freshwater Res 55(3):431–445.
Sweeney BW, Vannote RL (1978) Size variation and the distribution of hemimetabolous aquatic insects: two thermal equilibrium hypotheses. Science 200:444–446.
Odum EP (1984) The mesocosm. Bioscience 34(9):558–562.
Oertli B (1993) Leaf litter processing and energy flow through macroinvertebrates in a woodland pond (Switzerland). Oecologia 96:466–477.
Richards C, Haro R, Johnson L, Host G (1997) Catchment and reach-scale properties as indicators of macroinvertebrate species traits. Freshw Biol 37(1):219–230.
Tierno de Figueroa JM, López-Rodríguez MJ, Lorenz A, Graf W, Schmidt-Kloiber A, Hering D (2010) Vulnerable taxa of European Plecoptera (Insecta) in the context of climate change. Biodivers Conserv 19:1269–1277.
Vannote RL, Sweeney BW (1980) Geographic analysis of thermal equilibria: a conceptual model for evaluating the effect of natural and modified thermal regimes on aquatic insect communities. Am Nat 115:667–695.
Villéger S, Mason NW, Mouillot D (2008) New multidimensional functional diversity indices for a multifaceted framework in functional ecology. Ecology 89:2290–2301.
Vineetha S, Nandan SB (2021) Biological traits and trait combinations of benthic macroinvertebrates in a wetland under hydrological disturbance. In: Proceedings of the Zoological Society, vol 74, no 3, Springer, New Delhi, pp 339–356.
Wang H, Zhu X, Zakari S, Chen C, Liu W, Jiang XJ (2022) Assessing the effects of plant roots on soil water infiltration using dyes and hydrus-1D. Forests 13(7):1095.
Whiles MR, Wallace JB, Chung K (1993) The influence of Lepidostoma (Trichoptera: Lepidostomatidae) on recovery of leaf-litter processing in disturbed headwater streams. Am Midl Nat 356–363.
Williams DD (2006) The biology of temporary waters. Oxford: Oxford University Press.
Wong DC, Maltby L, Whittle D, Warren P, Dorn PB (2004) Spatial and temporal variability in the structure of invertebrate assemblages in control stream mesocosms. Water Res 38(1):128–138.
Xu X, Zhang Q, Tan Z, Li Y, Wang X (2015) Effects of water-table depth and soil moisture on plant biomass, diversity, and distribution at a seasonally flooded wetland of Poyang Lake, China. Chin Geogr Sci 25:739–756.
Zambrano J, Garzon-Lopez CX, Yeager L, Fortunel C, Cordeiro NJ, Beckman NG (2019) The effects of habitat loss and fragmentation on plant functional traits and functional diversity: what do we know so far? Oecologia 191(3):505–518.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Drought Effects on Benthic Macroinvertebrate Community Resilience and Functional Diversity in Wetland Mesocosms

Status:

Version 1

Abstract

Figures

1. Introduction

2. Methods

2.1 Study site and experimental set-up

2.2 Sampling and environmental data

2.2.1. BMI Sampling

2.2.2. Plant Sampling

2.2.3. Environmental data

2.3 Data analysis

2.3.1 Biological traits

2.3.2 Environmental factors and species richness

2.3.3. Analysis of BMI community changes among different wetland treatments and stages

2.3.4. Community and functional diversity indices

3. Result

3.1. Environmental factors

3.2. Species richness and abundance

3.2.1. BMIs

3.2.2. Plant species richness and cover

3.3. BMIs community

3.4. Biological traits of BMIs

3.5. Functional diversity

4. Discussion

4.1 Environmental factors and their influence on wetland conditions

4.2 Differences in species richness and community composition among the three wetland treatments

4.3 Functional Responses of BMIs to Drought Disturbance

4.4 Impact of drought on BMI community and functional diversity indices

5. Conclusions

Declarations

Funding

Author Contribution

Acknowledgement

Data Availability

References

Additional Declarations

Status:

Version 1