Diversity and Metacommunity Structure of Aquatic Macrophytes: A Study in Mediterranean Mountain Wetlands

Abstract

This study investigated the mechanisms determining macrophyte species composition in 23 Andalusian Mediterranean mountain wetlands (southern Spain). We employed a methodology combining two approaches: a pattern-based approach utilizing Elements of Metacommunity Structure (EMS) and a mechanistic approach involving Redundancy Analysis (RDA) and variance partitioning. This allowed us to identify the relevance of interactions between environmental and spatial factors. Data collection in these wetlands included macrophyte samples and physicochemical variables, alongside spatial variables generated using Moran’s Eigenvector Maps (MEMs). To refine the analysis of metacommunity structuring, the species matrix was partitioned based on macrophyte dispersal strategy (charophytes by spores and macrophyte vascular plants by seeds). Our results reveal that the macrophyte metacommunity in these wetlands exhibits quasi-clumped species loss for the total community, while charophytes and vascular plants showed quasi-random species loss. In conclusion, this study demonstrates that macrophyte communities in Mediterranean mountain wetlands do not follow a simple species replacement pattern. Instead, they are organized in a quasi-nested pattern, strongly shaped by environmental filters and, to a lesser extent, by spatial connectivity, with a prominent role for random processes. Understanding these mechanisms is crucial for predicting species responses to environmental changes and for designing effective conservation strategies within these vulnerable ecosystems.

1. Introduction

Aquatic macrophytes or hydrophytes (encompassing green macroalgae, charophytes, vascular plants and bryophytes) are keystone elements in the functioning, production and biogeochemical cycles of aquatic ecosystems [1,2,3]. Their presence has been documented across a multitude of ecosystems, from lotic to lentic environments, where they are essential indicators of good ecological status [4,5,6]. Macrophytes also provide a favorable habitat for the occurrence, refuge and feeding of other taxonomic groups, including zooplankton [7], macroinvertebrates [8,9] and fish [10]. Their role extends to supporting the stability, resilience and ecosystem services of these ecosystems [11,12,13].

Mountain wetlands, recognized as biodiversity hotspots [14], provide suitable habitats for the presence of many taxa, including aquatic macrophytes. The distribution of these organisms is influenced by a wide range of factors, such as altitude, water chemistry, climate, habitat connectivity and anthropogenic activities [15,16,17,18,19,20].

The fragmented structure of mountain wetlands, interconnected within broader landscapes, can lead to the formation of macrophyte metacommunities. A metacommunity is defined as a set of interacting local communities linked by the dispersal of multiple potentially interacting species [21]. This concept offers a powerful framework for understanding how local communities are shaped by both local processes (i.e., environmental variables or species interactions) and regional processes (i.e., dispersal or colonization). In essence, this conceptualization emphasizes the interconnectedness among patches (wetlands in this case) and the role of species movement (aquatic macrophytes) between them. To address this, the Elements of Metacommunity Structure (EMS) framework [21] offers a classic and widely adopted approach. It categorizes patterns observed within metacommunities, providing insights into the dominant ecological processes at play. The EMS framework evaluates three key elements: coherence, turnover and boundary clumping. These elements enable the differentiation of various metacommunity patterns, which were later refined by Presley and collaborators to include new metacommunity structures [22]. Analyzing these patterns allows for inferences regarding the relative importance of environmental variables, species interactions, species dispersal capacities and stochastic variability in shaping the organization of diversity within the studied landscapes. A valuable framework for understanding the dynamics and persistence of aquatic macrophytes in this wetland metaecosystem is thus provided. Therefore, understanding the structure and connectivity of these metacommunities is crucial for predicting species responses to environmental change and for designing effective conservation strategies.

Mediterranean mountain wetlands are exceptionally vulnerable ecosystems that serve as critical sentinels of global and regional environmental change. Due to their unique environmental conditions and often isolated nature, they are highly susceptible to the impacts of climate and global change, as well as other anthropogenic pressures [23,24]. They also represent ideal ecosystems to explore these dynamics and determine the contribution of environmental factors and spatial connectivity in explaining the regional patterns of aquatic macrophytes. In this context, this study aims to investigate the metacommunity structure and dynamics of aquatic macrophytes (charophytes and vascular macrophytes) in a selection of Mediterranean mountain wetlands in Andalusia, Spain. The analysis began by examining the total macrophyte community structure. Then, recognizing the distinct dispersal strategies of charophytes (via spores) and vascular macrophytes (via seeds), we performed subsequent separate analyses for each group. This tiered approach enabled us to thoroughly investigate how these differences in dispersal modes influence their respective community patterns, allowing for a more fitting ecological interpretation of our findings. Specifically, we will address the following questions: (i) What are the spatial patterns of occurrence and co-occurrence of different macrophyte species across the wetland network, considering both the total community and differentiating by dispersal strategies (spores and seeds)? (ii) What environmental (e.g., water chemistry, habitat size) and spatial factors (e.g., distance to other wetlands) at both the local and landscape levels are significant predictors of metapopulation parameters? By addressing these questions, this research will contribute to a better understanding of the ecological processes governing macrophyte communities in these vulnerable ecosystems and provide insights for their effective management and conservation.

2. Materials and Methods

2.1. Study Area

This study was conducted in 23 Mediterranean mountain wetlands located in Andalusia, southern Spain (Figure 1). The wetlands span an elevational range from 387 to 2068 m above sea level (m a.s.l.) and are distributed across the Betic mountains, specifically within the provinces of Cádiz, Seville, Málaga, Granada, Almería and Jaén. The study area is characterized by a Mediterranean climate, which features significant intra- and inter-annual fluctuations in temperature and precipitation. Consequently, the investigated wetlands exhibit variable hydroperiods, ranging from temporary to permanent.

2.2. Environmental and Spatial Variables

Macrophyte samples were collected during spring (May–June) of 2017 along transects, with sampling points randomly selected within each transect. Specimens were then transported to the laboratory for species-level identification using a Leica MZ12.5 stereomicroscope and appropriate taxonomic keys [25,26,27].

Concurrent with macrophyte sampling, several physical and chemical properties of the pond water (environmental variables, E) were measured. Electrical conductivity (EC, mS cm⁻¹) was measured using a portable conductivity probe (Hanna HI 98312—Hanna Instruments, Woonsocket, RI, USA), and turbidity (NTU) was assessed using a portable turbidity meter (Hanna HI 93703—Hanna Instruments, Woonsocket, RI, USA). Pond depth (D, m), surface area (WS, m²) and elevation (Alt, m) were measured in situ. Pond isolation was calculated using a Gaussian kernel density estimator [28]. This method assesses isolation by considering the weighted density of neighboring wetlands, where the influence of a neighbor decreases with increasing Euclidean distance according to a Gaussian function. The specific formula used for the pond isolation index of each wetland (i) was

$$Pond Isolation=1-\left(\pi r^{-2}\sum _i{10}^{{\left({-d}_i\right)}^2{2r}^{-2}}\right)$$

where d_i is the Euclidean distance between a pair of wetlands, calculated directly from their geographical coordinates (latitude and longitude). The kernel scale (r) was set to 0.5 degrees. This value was selected to represent an ecologically relevant spatial scale for interactions within the wetland complex, approximating an influence radius of approximately 45–55 km in the study region. Given that the presence or absence of vascular macrophytes and charophytes defined distinct subsets of the wetland complex, three distinct sets of pond isolation values were computed. These sets correspond to the specific wetland subsets included in each analysis: total (the complete dataset), spore (wetlands where charophytes were present) and seed (wetlands where vascular macrophytes were present). The substrate type (TS) and hydroperiod for each pond were obtained from Blanco et al. [29].

Water samples for nutrient analysis were collected in sterile polyethylene vials and stored in dark, refrigerated conditions (4 °C). In the laboratory, total phosphorus (TP, µg L⁻¹) was measured after digesting unfiltered water samples with potassium persulfate [30]. Total nitrogen (TN, mg L⁻¹) was analyzed using the ultraviolet method on digested, unfiltered water samples [30].

To account for the spatial component in shaping wetland macrophyte communities, we generated a set of spatial variables using Moran’s Eigenvector Maps (MEMs) analysis. This method produces orthogonal spatial variables directly derived from the geographical coordinates (latitude and longitude) of the wetlands [31]. The inclusion of a spatial component, acting as a surrogate for unmeasured ecological processes, is crucial in metacommunity analyses. It uncovers underlying mechanisms that are often difficult to measure directly in field studies [32,33]. MEMs are particularly advantageous as they capture various spatial patterns in communities [34,35] through their associated eigenvalues. The resulting MEMs represent different spatial scales, allowing for a nuanced exploration of spatial effects. Specifically, MEMs associated with positive eigenvalues represent broad-scale (B) spatial patterns, often related to processes such as dispersal and colonization limitations. Conversely, MEMs with negative eigenvalues indicate fine-scale (F) spatial patterns, which are typically associated with local community dynamics and biotic interactions [31,33,36,37]. This distinction between broad and fine spatial scales is paramount, as it directly relates to the interplay of different ecological processes in structuring communities [33].

MEMs were calculated using a distance-based approach implemented via the adespatial package [38] in R version 4.2.2 (R Core Team). Due to variations in the number of wetlands included in each specific dataset (i.e., total, spore and seed), the resulting MEM eigenvectors and their associated eigenvalues differed among these matrices. For the total dataset, a total of 22 MEMs were generated. MEM1 through MEM5 exhibited positive eigenvalues, thus representing broad-scale spatial patterns. The spore dataset yielded 14 MEMs, with MEM1 uniquely representing the broad-scale pattern. Finally, for the seed dataset, 21 MEMs were produced. MEM1 through MEM6 were characterized by positive eigenvalues, signifying broad-scale spatial patterns.

2.3. Data Analysis

Two common approaches in the study of metacommunities were used to assess non-random distribution patterns in macrophyte metacommunity structure (Figure 2). To achieve a more precise view of metacommunity structuring, and as previously noted, our analyses were performed on the complete species matrix as well as on subsets defined by dispersal strategy (charophytes via spores and vascular macrophytes via seeds).

One approach was the pattern-based approach that used the Elements of Metacommunity Structure (EMS—Figure 2) framework [21,39]. This analysis examines species distributions that may arise from underlying environmental gradients and identifies specific metacommunity structures (Figure 3). We employed a hierarchical variation partitioning method, a departure from the classical approach. This choice was driven by its superior suitability for analyzing environmental variability across both broad and fine spatial scales [40]. This approach is highly recommended for use with the Elements of Metacommunity Structure (EMS) framework, as it significantly enhances the understanding and identification of relationships between environmental and spatial filters [41]. This analysis used an interaction presence–absence matrix (sites x taxa) ordinated by reciprocal averaging (RA; [42]), grouping species with similar ranges and wetlands with similar species composition [43]. An α-level of 0.05 was used to determine the significance of the three metrics of EMS.

The second, a mechanistic approach, involved performing a Redundancy Analysis (RDA) and hierarchical variation partitioning (Figure 2). This allowed for the identification of the primary environmental and spatial variables explaining the variation in macrophyte species distribution [40,41]. Mechanistic methods often distinguish between different metacommunity models by partitioning the variance explained by spatial and environmental variables. For example, the significance of these variance partitions can be indicative of particular metacommunity types [44,45,46]. The statistical procedure identifies the key environmental (E) and spatial (S) variables (broad as B and fine as F in Figure 2) explaining the variation in the overall macrophyte metacommunity and within the two dispersal strategy groups (spores and seeds). Explanatory variable selection for the RDA employed a forward selection process with a double-stopping criterion, as described by Blanchet et al. [47]. Following variable selection, variation partitioning was used to assess the proportion of community composition variation explained by environmental and spatial variables (pure and shared fractions—R²) [48]. Prior to the mechanistic analysis, the species presence–absence data matrix was Hellinger-transformed to minimize the influence of rare species [49]. Additionally, environmental variables were transformed using log (x + 1) to reduce the effect of differing measurement scales. This analysis was conducted using the vegan package [50] in R version 4.2.2 (R Core Team).

Figure 3. A schematic representation of the method used to examine the Elements of Metacommunity Structure (EMS). I represents Morisita’s index [51]. The dashed black line represents a nested subset according to Leibold and Mikkelson [39]. The dashed blue and green lines represent the subdivisions of the nested subset category according to Presley and collaborators [22].

Figure 3. A schematic representation of the method used to examine the Elements of Metacommunity Structure (EMS). I represents Morisita’s index [51]. The dashed black line represents a nested subset according to Leibold and Mikkelson [39]. The dashed blue and green lines represent the subdivisions of the nested subset category according to Presley and collaborators [22].

3. Results

The environmental variables obtained from the studied wetlands are shown in

Table 1. The environmental descriptors of the studied wetlands. The acronyms for the environmental variables are reflected in the text. The numbers of ponds are coincident with those shown in Figure 1. Type of substrate (TS): S—siliceous; L—lime; T—Trias. Hydroperiod: T—temporary; P—permanent. Three pond isolation values have been calculated based on the geographic wetland coordinates (latitude and longitude). The first values shown in the table correspond to all wetlands, the second to wetlands where vascular macrophytes are present and the third to wetlands with charophytes present.

Ponds	EC (mS cm−1)	Turbidity (NTU)	D (m)	WS (m2)	Alt (m)	TS	Hydroperiod	TP (µg L−1)	TN (mg L−1)	Pond Isolation
1	0.21	6.94	1.5	2026.4	538	S	T	147.23	2.01	0.75/0.90/1.00
2	0.58	3.58	3	380.65	970	L	P	62.73	0.32	0.49/0.64/--
3	0.62	45.33	2	17,908.9	406	T	T	391.87	4.85	0.36/0.47/--
4	0.86	137	3	59,161.22	729	T	T	263.55	2.31	0.67/0.72/0.33
5	0.33	3.90	0.75	913.68	1162	L	T	19.95	0.57	1.00/1.00/--
6	0.94	3.76	2.5	551.17	387	T	T	16.30	0.37	0.87/--/0.38
7	10.4	1.93	5	67,113.49	794	T	P	25.69	1.82	0.91/0.99/0.38
8	0.15	2.36	1	2336.88	1347	L	T	22.56	0.84	0.85/0.89/--
9	0.34	8.17	2.5	5752.28	896	L	P	2428.83	3.08	0.78/0.78/0.27
10	0.29	3.51	0.4	17.34	1317	L	P	80.46	2.57	0.23/0.27/--
11	0.11	5.46	1.5	1197.69	1775	S	P	53.34	0.48	0.13/0.15/--
12	0.03	1.63	0.2	22.74	2068	S	T	77.59	0.45	0.15/0.19/--
13	0.12	23.27	2.5	4787.22	1830	L	T	110.71	1.01	0.58/0.68/0.25
14	0.13	61	1.5	4084.82	1713	L	T	175.40	1.55	0.68/0.79/0.30
15	0.28	67	2.5	1513.04	1340	L	P	45.51	0.52	0.67/0.78/0.29
16	0.18	72	0.5	180.06	1050	S	T	712.67	8.82	0.42/0.48/0.16
17	0.07	3.73	0.6	51.81	1892	S	P	37.17	1.2	0.43/0.49/0.17
18	1.56	2.60	1.25	283.71	1275	L	P	13.17	0.73	0.30/0.34/0.10
19	0.31	1.54	2	265.15	1340	L	P	12.13	0.26	0.00/0.00/0.00
20	0.28	>1000	0.5	959.03	1967	L	T	3310.38	28.54	0.30/0.33/--
21	0.42	41.27	2	229.71	1661	L	P	104.46	1.94	0.47/0.54/0.09
22	0.35	8.09	2	4669.39	1264	L	T	110.45	4.02	0.83/0.95/0.28
23	0.25	6.50	2.75	4406.9	1288	L	T	14.21	0.54	0.81/0.93/0.30

. The majority of the studied wetlands are freshwater or subsaline, with only two wetlands (Laguna Chica and Balsa Pocico) falling outside these ranges. Similarly, most wetlands exhibited clear water, with a notable exception in Laguna de Castril, presenting highly turbid water (NTU > 1000). All wetlands were shallow, with depths equal to or less than 3 m. Furthermore, most were small in size, with surface areas less than 1 ha in 48% of the studied wetlands. Regarding their location, 87% of the wetlands were situated above 700 m a.s.l., with 61% located on a silty substrate, followed by those on a siliceous substrate (22%). Concerning the hydroperiod, the distribution between the two observed typologies was very equitable, with 52% being temporary and 48% permanent. The TP and TN values indicated a high content of both nutrients in most of the studied wetlands. However, some significant exceptions were observed, such as Charca de la Franciscuela, which showed very low values for both nutrients. Finally, the isolation of the wetlands was generally high, with 83% having values greater than 0.5.

Table 2. Species richness of charophytes and vascular macrophytes in studied wetlands. Numbers of wetlands coincident with those shown in Figure 1.

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22	23
Charophytes
Chara connivens				x			x
Chara curta																						x
Chara globularis				x		x	x		x						x							x	x
Chara vulgaris						x	x											x	x		x	x
Nitella flexilis																							x
Nitella syncarpa																	x
Nitella translucens	x
Sphaerochara prolifera													x	x		x
Tolypella glomerata															x
Tolypella hispanica																					x		x
Vascular plants
Baldellia ranunculoides																				x			x
Callitriche lusitanica	x
Callitriche stagnalis											x	x
Dasmasomium polispermum			x	x
Myriophylum alterniflorum	x
Myriophylum spicatum				x			x
Polypogum amphibium				x
Potamogeton berchtoldii																							x
Potamogeton pectinatus							x
Potamogeton polygonifolius																	x
Potamogeton pusillus									x
Rannunculus hederaceus												x
Rannunculus trychophylus		x			x														x		x	x	x
Rannunculus peltatus peltatus	x		x					x					x	x	x	x				x
Zannichellia contorta																		x
Zanichellia pedunculata																						x
Zanichellia peltata																			x
Zannichellia palustris		x								x

presents the diversity of aquatic macrophytes across the studied mountain wetlands, listing ten charophyte species and eighteen vascular macrophyte species. These species are distributed throughout each wetland, with species richness ranging from one to six. Laguna de Siles exhibits the highest species richness, whereas Laguna del Hondonero, Charco del Nevazo Largo, Charca de Juan Ramos and Balsa del Almiar display the lowest. Additionally, a significant group of eleven other ponds harbors only two species. Regarding their geographical distribution, a strong disparity is observed between charophytes and vascular macrophytes. The latter, dispersed by seeds, are present in all sampled wetlands, while charophytes, dispersed by spores, are found in only 65% of them.

Species richness per wetland ranged from one to eight, with Chara globularis Thuill. being the most prevalent among the charophytes and Ranunculus peltatus Schrank subsp. peltatus the most prevalent among the vascular macrophytes. The least prevalent charophyte species, each found in only a single wetland, include Chara curta Nolte ex Kütz, Nitella flexilis (L.) C. Agardh, Nitella syncarpa (Thuill.) Chev., Nitella translucens (Pers.) C. Agardh and Tolypella glomerata (Desv. In Liosel) Leonh. A similar pattern was observed for vascular macrophytes, with Callitriche lusitanica Schotsman, Myriophyllum alterniflorum DC., Potamogeton berchtoldii Fieber, Potamogeton pectinatus L., Potamogeton polygonifolius Pourr., Potamogeton pusillus L., Ranunculus hederaceus L., Zannichellia contorta (Desf.) Cham. & Schltdl., Zannichellia pedunculata Rchb. and Zannichellia peltata Bertol. each occurring in only one wetland.

Table 3. Results of Elements of Metacommunity Structure (EMS) analysis for total, seed and spore matrices. Interpretations of metacommunity type following Leibold and Mikkelson [39] and Presley et al. [22]. Abbreviations: embAbs—embedded absences; Coh Z—Z value of coherence; Tur Z—Z value of turnover; p—p values; df—degrees of freedom; Sim mean—mean value for the null model distribution; Sim sd—standard deviation value for the null model; p—the probability from a z-test to assess the significance of the observed index to the null matrices.

		Total	Seeds	Spores
Coherence	embAbs	57	6	18
	Coh Z	−9.84	−5.83	−3.72
	p	<0.001	<0.001	0.002
	Sim mean	275	112	45
	Sim sd	22	18	8
Turnover	Turnover	1801	467	203
	Tur Z	−0.52	−0.65	−0.92
	p	0.610	0.517	0.367
	Sim mean	1845	480	236
	Sim sd	85	21	37
Boundary clumping	Index	1.50	1.17	0.87
	p	0.002	0.175	0.384
	df	20	19	12

presents the results of the Elements of Metacommunity Structure (EMS) analysis for the three matrices investigated in this study: total community, vascular macrophytes and charophytes (see Supplementary Material S1 for the original data used in this study). Interpretations of these results, aimed at identifying the metacommunity structure obtained in each case, were conducted following Leibold and Mikkelson [39] and Presley and collaborators [23]. The results indicate a clear distinction in metacommunity structure. The overall macrophyte community exhibited a quasi-clumped species loss metacommunity pattern. In contrast, both charophytes and vascular macrophytes individually showed a quasi-random species loss metacommunity structure.

The results of RDA and variation partitioning, illustrating the contributions of environmental and spatial variables (fine- and broad-scale) for the total, charophytes and vascular macrophytes matrices, are displayed in

Table 4. Results of RDA and variation partitioning showing contributions of environmental and spatial variables (fine- and broad-scale) for total, spore and seed matrices.

	Fraction	df	Adj. R2	Variables
Total	E	2	0.13	Hydroperiod and size
	B	1	0.08	MEM1 and MEM2
	F	1	0.06	MEM20
	E ∩ B	4	0.05	Hydroperiod, size, MEM1 and ME2
	E ∩ F Residuals	3 --	0.01 0.67	Hydroperiod, size and MEM20 --
Spore	E	3	0.32	Hydroperiod, isolation and size
	F	2	0.18	MEM3 and MEM12
	E ∩ F Residuals	5 --	0.12 0.38	Hydroperiod, isolation, size, MEM3 and MEM12 --
Seed	E Residuals	-- --	-- --	Hydroperiod, substrate and elevation --

and Supplementary Material S2. For the overall macrophyte community matrix (total), two environmental variables (hydroperiod and size) and three spatial variables (MEM1, MEM2 and MEM20), related to the broad and fine scales, significantly explained macrophyte distribution (Figure 4). Variation partitioning indicated that environmental and spatial variables accounted for similar proportions of the total variation. For the charophyte matrix (spore), environmental variables (hydroperiod, isolation and size), along with two broad-scale spatial variables (MEM3 and MEM12), explained species distribution, which was mainly driven by environmental filters (Figure 4). Finally, the vascular macrophyte matrix (seeds) showed a greater influence from environmental variables only (hydroperiod, substrate and elevation), with no significant influence from spatial variables (Figure 4).

4. Discussion

A fundamental inquiry in community ecology centers on the mechanisms that dictate species composition [52]. Historically, research predominantly suggested that local environmental factors governed community structure. However, this exclusive focus frequently obscured the significant role of processes acting at broader regional scales. A pivotal shift in this perspective occurred with the introduction of the metacommunity concept by Leibold and collaborators [21]. This concept profoundly underscored that identifying the primary mechanisms and their scales of operation is crucial for predicting final species composition and distribution [22]. Our present study, through its application of two robust analytical methodologies—EMS and RDA with variation partitioning—has enabled a more precise discernment of the principal drivers shaping macrophyte metacommunity organization in mountain Mediterranean ponds. This integrated approach, previously employed in metacommunity studies by other authors [41,53,54,55], has proven instrumental in revealing the critical interplay between environmental and spatial factors.

Our results reveal a complex picture of species distribution and turnover, characterized by a modest richness of ten charophyte and eighteen vascular macrophyte species. The species richness per wetland, ranging from one to eight, could highlight the variability in ecosystem complexity and environmental conditions among the studied sites. In this sense, the numerous charophyte and vascular macrophyte species found in only a single wetland highlight the potential for high beta diversity and the presence of rare or habitat-specialist species within this network [56]. Our results with the overall matrix indicate that the macrophyte metacommunity in Mediterranean mountain wetlands exhibits a quasi-structure composition, with clumped species loss. This nested structure indicates that species turnover is not significantly different from random, suggesting a greater influence of stochastic factors or mass effects that dilute the signal of clear and ordered species replacement [57]. Unlike the nested clumped species loss pattern, which has been widely reported in studies on aquatic metacommunities, e.g., [53,58,59,60,61], the specific pattern observed here has been documented in fewer studies [57,61,62]. Nevertheless, both patterns indicate that the metacommunity is primarily composed of species-poor sites [57,63]. This metacommunity typology, common in fragmented and isolated habitats, indicates low species replacement and the presence of several generalist species across most wetlands [62]. This is clearly observed in our study by the wide distribution of some taxa, such as Chara vulgaris L., Ch. globularis, R. peltatus and Rannunculus trycophyllus Chaix, which showed a wide distribution across all studied mountain ponds, while a large number of species occurred in only one or two ponds. This contributes to the appearance of nested subsets according to [39], meaning that ponds closer to others tend to be richer (i.e., Siles), while more isolated ponds have poorer subsets of the richer wetlands (i.e., Charca de Loma León, Laguna del Rico, Laguna de Castril, etc.). This pattern might be driven by environmental gradients, habitat fragmentation or dispersal limitations [41,64]. This community structuring pattern is common in seasonal rivers [65] and has also been observed in tropical reefs [66].

A division of the data into spore and seed matrices reveals contrasting metacommunity structures, both indicating a quasi-random species loss metacommunity structure [54]. This divergence from the overall pattern could be attributed to differences in dispersal mechanisms and life history traits between these groups [64]. Regarding the former, angiosperms showed higher species-specific variability of niche breadths compared to charophytes [67]. Furthermore, dispersal constitutes a crucial aspect of plant life, particularly in the context of contemporary challenges posed by anthropogenic habitat fragmentation and climate change [68]. Moreover, macrophytes exhibit diverse seed and spore dispersal mechanisms, such as anemochory, hydrochory and zoochory. Spore dispersal, in particular, might lead to more stochastic colonization and extinction events compared to seed dispersal, which can be more directed or influenced by specific environmental cues [69].

The Redundancy Analysis (RDA) of the entire macrophyte community revealed that both environmental and spatial variables (representing broad and fine scales) significantly influenced species distribution. The subsequent variation partitioning revealed a similar proportion of explained variance attributed to these two sets of factors. This suggests that local environmental conditions and the spatial arrangement of the ponds, including the associated dispersal capacity of species, are crucial in shaping the overall macrophyte metacommunity structure. This finding aligns with the understanding that species distributions in fragmented landscapes are often governed by a combination of habitat suitability and dispersal-related processes [70]. Considering environmental variables, hydroperiod plays a key role in aquatic diversity within Mediterranean ecosystems [71]. Similarly, ecological theory suggests that larger areas tend to exhibit greater species richness, often due to increased habitat availability and a wider range of environmental conditions [72]. In the same way, species richness should increase with island size and decrease with distance from the source pool [73]. Therefore, an increase in island size primarily influences species richness by enhancing niche variety and population size, while isolation reduces species richness by limiting the number of potential colonists dispersing to the island [73,74].

The observed difference in the relative importance of environmental and spatial factors between spore and seed matrices is particularly interesting. These results reveal differences in the types of environmental conditions experienced by the two macrophyte groups: those dispersed by seeds and those dispersed by spores. While seed-dispersed macrophytes are influenced by hydroperiod, substrate and elevation, spore-dispersed macrophytes are affected by hydroperiod, in addition to isolation and size. The greater influence of environmental variables on macrophyte distribution, as opposed to spatial variables, suggests that this distribution is primarily driven by habitat suitability at the local scale. Conversely, spatial variables exclusively influence the species distribution of spore-dispersed macrophytes, characterized by their small size and potential for long-distance dispersal [75]. The enhanced dispersal capability of charophytes would be expected to result in a broader distribution within the surveyed pond pool. This lack of a discernible pattern suggests the substantial influence of environmental conditions on charophyte establishment [76], particularly since these organisms are highly vulnerable to environmental alterations such as eutrophication [77].

The present research highlights a critical insight into metacommunity organization. Their structure can vary depending on the biological traits considered. Initially, the overall macrophyte community displayed a quasi-clumped species loss pattern. However, upon analyzing the data by dispersal mechanism (seeds and spores), a hidden complexity emerged. Each group revealed its own distinct, underlying quasi-random species loss structure. Furthermore, even with this shared underlying structure, the two dispersal-based typologies were driven by different sets of environmental variables. This demonstrates that evaluating communities through the lens of specific biological traits is not merely a refinement. It constitutes fundamental shift, essential for unlocking the true complexity of metacommunity organization and advancing predictive ecology.

Our findings advance the understanding of aquatic plant metacommunity structure in the Mediterranean region by further substantiating the complex interplay between environmental filtering and spatial processes. Understanding how these factors interact to shape metacommunity structure is crucial for effective conservation and management strategies aimed at preserving the biodiversity of these valuable ecosystems. Effective conservation strategies must therefore address both the environmental integrity of individual water bodies and the spatial configuration of the landscape in order to safeguard the unique biodiversity and ecological processes that underpin these valuable macrophyte metacommunities. Consequently, prioritizing the conservation of a diverse range of mountain ponds—characterized by high environmental variability and maximized connectivity—is crucial. This, coupled with the continuous monitoring of their macrophyte diversity, is essential for safeguarding this metacommunity in mountain wetlands, particularly given the growing threats from anthropogenic activities [14,78] and global change. In particular, a suitable conservation measure for this metacommunity would involve preserving wetlands with the highest macrophyte diversity, as well as those serving as refugia for species found in only a single wetland. In this context, it has recently been established that the planetary-scale importance of conserving core areas that sustain a large part of regional biodiversity is crucial [79]. In this regard, at the landscape scale, it is crucial to conserve not only the aquatic ecosystems included in this study but also the human activities that have historically facilitated connectivity among wetlands. A prime example is transhumance, a practice carried out since ancient times [80] that still persists in some areas of Andalusia. This practice allows livestock to disperse seeds among these ecosystems. Wetlands associated with transhumance have been observed to be of high importance for amphibian conservation [81], an aspect that could be equally significant for aquatic macrophytes. Furthermore, in line with ongoing research by the authors regarding an open inventory of mountain wetlands in Andalusia, future studies that include additional wetlands would be beneficial. This would allow for a more thorough analysis of the aquatic macrophyte metacommunity of mountain ponds in Andalusia.