Habitat Destruction Alters the Mechanisms of Species Coexistence by Modifying Competitive Structure

Abstract

Habitat destruction is a major driver of biodiversity decline, yet how it reshapes multispecies coexistence by altering interaction structure remains unclear. We adopt a spatially explicit metacommunity model framework under a homogeneity assumption and introduce a tunable parameter controlling intransitive competition. Within this framework, we represent the system using a generalized Lotka–Volterra model to examine how coexistence mechanisms respond to habitat destruction. Our findings demonstrate that (1) coexistence is not driven by a single mechanism: under transitive competition, it highly relies on niche differentiation, whereas in intransitive structures, coexistence can be maintained even with low niche differentiation. (2) Habitat destruction compresses the feasible coexistence space, but regions dominated by different mechanisms respond asymmetrically, with niche-difference-driven coexistence shrinking and intransitive-dominated coexistence expanding under certain conditions. (3) The difference stems from habitat destruction, altering the relative proportions of intraspecific and interspecific competition, driving the community beyond the coexistence threshold. This reduces the probability of coexistence and reshapes the relative importance of several coexistence mechanisms. This finding provides a new theoretical perspective for biodiversity in fragmented landscapes.

1. Introduction

Habitat destruction is widely acknowledged as a principal factor in the current decline in biodiversity [1,2,3,4]. The classical “extinction debt” theory posits that species persistence is hindered as the proportion of available patches declines [5]. In real ecosystems, habitat destruction frequently results in alterations to spatial structure and connectivity. The foundational framework for understanding how such ecological disturbances drive biodiversity changes is the Intermediate Disturbance Hypothesis (IDH) [6], which posits that local species diversity is maximized at intermediate levels of disturbance. Theoretical studies further demonstrate that, despite the regulation of total habitat area, landscape configuration and habitat fragmentation can significantly impact population dynamics and community composition by affecting species dispersal, local extinction, and recolonization processes [7,8,9]. However, current research has predominantly concentrated on the “outcomes” of diversity changes, with insufficient investigation into how habitat destruction affects biodiversity by altering the dynamics of species interactions. In reality, community stability and coexistence patterns are influenced not only by species’ abilities but also by the structure of competition among species [10]. Consequently, comprehending the effects of habitat fragmentation on communities through the framework of interaction structure may be essential for elucidating its fundamental mechanisms.

Modern species coexistence theory provides a systematic framework for comprehending community stability [11,12]. This theory posits that species coexist via stabilizing mechanisms arising from niche differentiation and relative advantage that offset fitness difference. In the most basic framework, most studies have analyzed pairwise competitive structures, suggesting that systems can achieve stable coexistence when intraspecific competition is stronger than interspecific competition [13,14,15]. However, natural communities are typically composed of multiple species, and their complex dynamics often cannot be explained solely by pairwise interactions. Expanding the traditional pairwise perspective reveals that the structural characteristics of diverse ecological networks, such as collaborative and competitive networks, offer additional mechanisms for multispecies coexistence. For example, collaborative interactions such as mutualism and facilitation have been shown to enhance community stability and support species coexistence [16,17]. However, competition drives competitive exclusion and species extinction, maintaining biodiversity in natural systems faces a significant theoretical challenge. Therefore, investigating competitive networks is crucial. Within competitive interactions, for instance, an intransitive loop similar to “rock-paper-scissors” creates cyclical relationships by weakening the advantage of dominant species [18,19,20]. Thus, while collaborative interactions play a role in community stability, unraveling how the structural complexities of competitive networks, particularly intransitive structures, function remains the critical step in understanding biodiversity maintenance. The source of community stability may be determined by the relative contributions of niche differentiation and intransitive competition, which may co-act within the same system. Intransitive competition should be incorporated into traditional species coexistence frameworks, according to an increasing amount of research [21,22,23,24]. However, systematic theoretical studies on how these two mechanisms transform under habitat destruction remain scarce.

By impacting dispersal processes, colonization rates, and local population reproduction processes, habitat destruction may also affect the intensity of species interactions in addition to changing population size [5]. The competition matrix is an efficient result controlled by spatial processes rather than a static parameter when population dynamics incorporate spatial structure. For example, local competition and species dispersal patterns could shape intraspecific and interspecific interactions in spatially explicit models, affecting the likelihood of coexistence [25,26,27,28]. Competitive structures might therefore face asymmetric changes as a result of habitat removal or reconfiguration. Such changes could have effects that go beyond the simple enlargement or contraction of coexistence areas; they could also result in changes to coexistence mechanisms. Additionally, species spatial distributions and competitive intensities can be changed by spatial heterogeneity and fragmented network structures themselves, which can reorganize the relative roles of equalization and stabilization mechanisms [29].

A crucial framework for exploring community dynamics is provided by the metacommunity model under habitat destruction. The metacommunity model, as opposed to the conventional mean-field model, explicitly includes habitat patch structure and species dispersion dynamics in their analysis. This makes it possible to characterize local colonization–extinction processes and connectivity that affect the stability of communities [30,31]. Examining how habitat destruction affects community dynamics within this framework shows strong theoretical consistency since habitat destruction significantly modifies spatial structure, and spatial characteristics are precisely the subject of metacommunity theory. According to recent research, a spatially explicit metacommunity model is especially useful for examining the ways in which competitive network structures and spatial processes interact [32,33,34]. In this context, it is necessary to minimize the influence of environmental heterogeneity when considering the effect of competitive structure itself on coexistence mechanisms. Meanwhile, in many regional-scale or functionally homogeneous habitat systems, the homogeneity assumption may effectively capture the dynamic characteristics of the spatial network, allowing spatial processes to unfold under controlled conditions, despite the general heterogeneity of real landscapes.

Here, we adopt a three-species competitive system within a metacommunity framework that integrates spatial characteristics and intransitive interactions. Under the homogeneity assumption—where all habitat patches have identical areas and constant pairwise distances—the model can be reformulated into a generalized Lotka–Volterra form, allowing structural niche difference to be quantified directly from the competition matrix. By parameterizing the competition coefficients, we then continuously transform the system from a transitive to an intransitive structure through an adjustable intransitive competition strength parameter. On this basis, we considered the mechanisms of niche differentiation, intransitive competition, and their combined impacts. By analyzing the coexistence regions under these mechanisms and their responses to varying degrees of habitat destruction, we focused on answering the following questions: (1) How does habitat destruction change the species persistence conditions? (2) What might be the causes of such changes?

2. Materials and Methods

2.1. Multispecies Competition Model

2.1.1. Model Formulation

The model framework developed here builds upon the spatially explicit metacommunity model proposed by Luo et al. [35]. This framework integrates species colonization–extinction dynamics, interspecific competition, and landscape configuration into a unified dynamical system. However, analytical investigation of multispecies coexistence under fully heterogeneous landscapes remains challenging due to the high dimensionality introduced by spatial variability. To facilitate theoretical analysis while retaining essential spatial processes, we adopt a spatial homogeneity assumption, under which all habitat patches are assumed to have identical size and constant pairwise distances (Figure 1a). This simplification removes heterogeneity in landscape configuration while preserving colonization–extinction dynamics and dispersal-mediated interactions. Under this assumption, the spatially explicit system can be reformulated into a tractable generalized Lotka–Volterra form, which serves as the core dynamical model in our analysis:

$${\displaystyle \frac{1}{p_k\left(t\right)}}{\displaystyle \frac{d{p}_k\left(t\right)}{dt}}={\displaystyle \frac{1}{A^{z_{ex,k}}}}\left(r_k-\sum _{l=1}^n{\alpha }_{kl}{p}_l\left(t\right)\right)$$

(1)

where $p_k$ denotes the occupancy probability of species k in any patch. The term $r_k=c_k{\lambda }_k-e_k$ represents the generalized intrinsic growth rate of species k, while the competition coefficient matrix α = (α_kl) is given by ${\alpha }_{kk}=c_k{\lambda }_k$ and for $l\ne k$, ${\alpha }_{kl}=Q_{kl}{e}_k$, where ${\alpha }_{kl}$ represents the effect of species l on species k. A represents patch area, and $z_{ex,k}$ is the scaling exponent of extinction rate with patch size for species k. The parameter ${\lambda }_k$ represents the metapopulation capacity of species k, a synthetic spatial metric that integrates landscape configuration [36,37]. In the generalized Lotka–Volterra model, metapopulation capacity ${\lambda }_k$ is defined as the leading eigenvalue of the landscape matrix $M_k$. The s × s landscape matrix $M_k$ (s denotes the number of patches) has entries ${\left(M_k\right)}_{ij}$ specified by:

$${\left(M_k\right)}_{ij}=\left\{\begin{array}{ll}{A}^{Z_{im,k}+Z_{ex,k}+Z_{em,k}}{e}^{-{\displaystyle \frac{D}{{\xi }_k}}},& \mathrm{i}\mathrm{f} i\ne j\\ 0,& \mathrm{i}\mathrm{f} i=j\end{array}\right.$$

(2)

The construction of the competition-related parameter $Q_{kl}$ is described in Section 2.1.2. Other relevant parameters and their values are detailed in

Table 1. Parameters.

Parameter	Description	Value
θ3	The degree of competitive transitivity	0:0.005:2.5
θ	The degree of competitive symmetry	U(0, 0.5)
λ	The metapopulation capacity of a fragmented landscape	—
ck	The species-specific colonization parameter of species k	U(0.0001, 0.01)
ek	The species-specific extinction parameter of species k	1
zim,k	The scaling exponent of the immigration rate with patch size for species k	U(0, 1)
zem,k	The scaling exponent of the emigration rate with patch size for species k	U(−1, 1)
zex,k	The scaling exponent of the extinction rate with patch size for species k	U(0, 1)
ξk	The characteristic dispersal distance of species k	U(1, 2)
A	The size of the patch	${10}^{N\left({\mathrm{1,0.5}}^2\right)}$
D	The distance between pairs of patches	U(0, 10)

. It should be mentioned that model coefficient $1/A^{Z_{m,k}}$ is a positive proportional factor that does not affect the stability of equilibrium solutions or the coexistence criteria. Therefore, it will not be included in later studies.

2.1.2. Competition Matrix

We apply Equation (1) to parameterize multispecies competitive interactions for the purpose of investigating variations in coexistence mechanisms across multiple species under different competitive structures. Specifically, we adopt Yang and Hui’s [38] competitive structure design approach to establish interspecific interaction of three species (species 1, species 2, species 3) as follows: the interaction strengths of species 1 on species 2, species 1 on species 3, and species 2 on species 3 are all set to 1. The parameter θ controls the strength of interaction about species 2 on species 1, and species 3 on species 2, whereas the parameter θ₃ controls the strength of interaction about species 3 on species 1 (Figure 1b). Consequently, the competition matrix Q is defined as:

$$Q=\left(\begin{array}{ccc}0& \theta & {\theta }_3\\ 1& 0& \theta \\ 1& 1& 0\end{array}\right)$$

In this context, the system’s paired competition symmetry is controlled by parameter θ, whereas the degree of intransitive competition is modulated by parameter θ₃. The system displays transitive competition when θ₃ < 1, yet an intransitive loop is formed via competitive interactions when θ₃ > 1. Table 1 provides specific parameter values. The competition coefficient matrix α in Equation (1) could be explicitly obtained from the metacommunity model’s parameters, taking the form:

$$\mathit{\alpha }=\left(\begin{array}{ccc}{c}_1{\lambda }_1& Q_{12}{e}_1& Q_{13}{e}_1\\ Q_{21}{e}_2& c_2{\lambda }_2& Q_{23}{e}_2\\ Q_{31}{e}_3& Q_{32}{e}_3& c_3{\lambda }_3\end{array}\right)\to \left(\begin{array}{ccc}{c}_1{\lambda }_1& \theta e_1& {\theta }_3{e}_1\\ e_2& c_2{\lambda }_2& \theta e_2\\ e_3& e_3& c_3{\lambda }_3\end{array}\right)$$

2.2. Coexistence Analysis

2.2.1. Conditions for Multispecies Coexistence

Invasion criterion is capable of being utilized to assess if multiple species coexist under Equation (1) [11,39,40]. In particular, when one species is absent and the other species are in stable equilibrium, we investigate if any species may effectively invade the system [41,42,43]. The system meets the requirement for multispecies coexistence if each species has a positive invasion growth rate when acting as an invader, that is, ${\displaystyle \frac{d{p}_k}{dt}}·{\displaystyle \frac{1}{p_k}}>0$ for every species k. Let $R_k$ be the invasion growth rate of species k in equilibrium with the other species. The specific calculation method for the invasion growth rate of species k is detailed in Appendix A. The system fulfills the requirement for multispecies coexistence when every species in the system satisfies $R_k>0$ as invaders ($k = 1, 2, 3$). This criterion has been employed in the studies that follow to figure out whether a system state in the parameter space is conducive to coexistence [44,45,46].

2.2.2. Species Coexistence Space

We further describe the species coexistence space along two dimensions: niche difference and intransitive competition, assessing multispecies coexistence. We use the structural niche difference $\Omega$ proposed by Saavedra et al. [47] to quantify niche difference in multispecies systems. The mathematical expression for n species is $\Omega ={\displaystyle \frac{2^n\left|\det \mathit{\alpha }\right|}{{\pi }^{n/2}}}{\int }_0^{+\infty }{e}^{-x^T{\mathit{\alpha }}^T\mathit{\alpha }x}dx$, where α is the matrix of competition coefficients. However, in a two-species system, regions of coexistence and exclusion based on niche difference have a clear division, whereas in a multispecies system, the division becomes diffuse, making coexistence more difficult to interpret. Luo et al. [35] further extended this idea in multispecies systems that take habitat fragmentation and landscape configuration into account, suggesting an improved structural niche difference Ω₁, defined as ${\Omega }_1={\Omega }^{1/(n-1)}$. Given that our study focuses on multispecies coexistence under varying competitive structures, we adopt the rescaled metric Ω₁ to provide a comparable quantification of niche differentiation.

The presence of an intransitive structure does not necessarily imply that it contributes to coexistence. Thus, intransitive metrics alone are insufficient to assess their actual effect. We use the average change in invasion growth rate $\overline{∆R}$ to reflect the extent to which intransitivity promotes coexistence [38,41]. Specifically, when any other species is removed from the system, the average change in the invasion growth rate of species k is described by $\Delta R_k$, and the average change in the invasion growth rate of three species is described by $\overline{∆R}=(\Delta R_1+\Delta R_2+\Delta R_3)/3$; details are in Appendix B. This measure characterizes the total effect of intransitive competition on the average invasion growth rate. Intransitive structure typically increases species invasion growth rates, which promotes coexistence among multiple species when $\overline{∆R}>0$. Conversely, it suggests that intransitive competition does not facilitate coexistence within that system if $\overline{∆R}<0$.

2.3. Habitat Destruction Patterns

We provide three categories and five typical habitat destruction scenarios inside the previously specified model framework in order to comprehensively investigate the impact of various forms of habitat destruction on the processes of multispecies coexistence. From three angles, these patterns describe the process of habitat destruction: decreasing the area of all patches proportionally, increasing the distance between each pair of patches, and removing a fraction of patches randomly.

According to Luo et al. [35], the first category is habitat deterioration (Category A), which reduces patch area. It is further classified as uniform and gradient habitat deterioration. Under uniform habitat degradation (Pattern A1), the area of each patch shrinks simultaneously by an identical proportion. Gradient habitat deterioration (Pattern A2), on the other hand, decreases patch area in a gradient fashion, resulting in differing levels of area loss in different patches. Nonetheless, the total area reduction across all patches is consistent with that under uniform habitat deterioration, guaranteeing that the two types are comparable in terms of the overall severity of loss.

The second category is habitat fragmentation (Category B), which involves increasing the distance between patches without changing the patch area directly [48]. Additionally, there are two types of habitat fragmentation. With uniform habitat fragmentation (Pattern B1), species dispersion connectivity between patches is evenly reduced when distances between all patches rise synchronously by the same amount. Gradient habitat fragmentation (Pattern B2), on the other hand, results in unequal isolation levels between various patch pairings by increasing patch distances in a gradient fashion. Nonetheless, the overall distance increase stays the same as in uniform fragmentation, guaranteeing that the two types’ overall isolation levels are constant.

The third category is the habitat loss (Category C, Pattern C) [35,48], which uses the direct removal of whole patches to mimic total habitat loss. In this case, certain patches are completely removed from the landscape, which changes the structure and connectivity of the landscape while also decreasing the amount of habitat that is accessible.

2.4. The Implementation of Numerical Simulation

We performed numerical simulations of Equation (1) to evaluate species coexistence across parameter space. In each simulated community, the number of habitat patches was fixed at 100. Model parameters ($c_k, e_k, z_{im,k}, z_{em,k}, z_{ex,k}, {\xi }_k, A, D$) were randomly drawn from predefined distributions (Table 1), generating 5000 independent parameter sets. For each parameter set, the transitivity strength ${\theta }_3$ was varied from 0 to 2.5 in increments of 0.005, reconstructing the corresponding competition matrix and evaluating coexistence conditions at each value using MATLAB R2024a (MathWorks, Natick, MA, USA). Habitat destruction was implemented by proportionally reducing patch size or removing a fraction of patches, and proportionally increasing the distance between patches.

3. Results

3.1. Mechanisms of Species Coexistence

We plotted the coexistence regions of different mechanisms on a parameter space defined by structural niche difference ${\Omega }_1$ and intransitive structural strength as coordinate axes. According to simulation results (Figure 2), when ${\theta }_3<1$, the system displays a typical competitive hierarchy pattern. In order to prevent competitive exclusion and preserve community stability, species with varying degrees of competitive advantage mostly depend on adequate niche differentiation. Intransitive competition is present in the system when ${\theta }_3>1$, but their effects on coexistence vary, resulting in different coexistence mechanisms. Specifically, depending on the direction of the average change in invasion growth rate ($\overline{\Delta R}$), the coexistence area divides into two subregions, Latent Intransitivity and Active Intransitivity, each having unique mechanism features along the structural niche difference. The relatively large range of ${\Omega }_1$ is where the Latent Intransitivity area is mainly found. Even though the system has intransitive competitive structures, their ability to stabilize coexistence is minimal in this area ($\overline{\Delta R}<0$). The basic factor sustaining coexistence is structural niche differences. Conversely, regions with greater ${\theta }_3$ and lower ${\Omega }_1$ are concentrated in the Active intransitivity region. Intransitive competition is the primary mechanism maintaining species coexistence in this area, as evidenced by the corresponding coexistence states, which indicate a considerable contribution of intransitive competition to the system’s stable coexistence ($\overline{\Delta R}>0$). Competitive hierarchy, latent intransitivity, and active intransitivity are the three mechanistic domains into which we classify coexistence areas as a result of the clear structural distribution that appears in this parameter space.

3.2. Effects of Habitat Destruction

In this section, we explore how patterns of species coexistence are modified by habitat destruction. We simulate species coexistence responses under various types and intensities of habitat destruction by manipulating patch area and the distance between patches in this spatial implicit metacommunity model.

3.2.1. Effects on Coexistence Outcomes

We first examine the impact of gradient habitat deterioration (Pattern A2) on species coexistence (Figure 3). Overall, the parameter space enabling species coexistence has a general shrinkage tendency as habitat destruction rises from 10% to 40% and then to 70%. Nonetheless, distinct coexistence mechanisms reflect varying evolutionary trends in reaction to habitat destruction. There is a decreasing tendency in both the coexistence zones for latent intransitivity and competitive hierarchy. Both competitive hierarchy and latent intransitivity must rely on a strong structural niche difference (${\Omega }_1$) to keep species coexistence, since habitat destruction reduces these mechanisms’ ability to counteract competitive exclusion. In stark contrast, habitat destruction did not reduce the area that was dominated by active intransitivity. Instead, its distribution spread over both low ${\Omega }_1$ and high ${\theta }_3$ regions as a result of more severe habitat destruction, and intransitive competition took over as the main factor maintaining system coexistence.

We also explored five distinct types of habitat destruction. The findings demonstrated that the particular pattern of habitat degradation had no discernible impact on the previously described behaviors (Figure 4). The changes in coexistence areas corresponding to the three mechanisms—competitive hierarchy, latent intransitivity, and active intransitivity—showed identical variations as habitat destruction degree increased, regardless of whether patch area was reduced via uniform or gradient (Figure 4a), patch distance was increased via uniform or gradient (Figure 4b), or entire patches were eliminated (Figure 4c). In particular, there was a monotonically declining trend in the coexistence regions for both latent intransitivity and competitive hierarchy. On the other hand, when habitat destruction increased, the coexistence area for active intransitivity exhibited a rising tendency. The main finding that “habitat destruction suppresses coexistence under competitive hierarchy and latent intransitivity while promoting coexistence under active intransitivity” holds across all types of habitat destruction, despite minor variations in particular values and rates of change—for instance, uniform habitat fragmentation considerably increases active intransitivity more than gradient habitat fragmentation.

3.2.2. Effects on Coexistence Mechanisms

To examine how coexistence responds to habitat destruction under different mechanisms, we first analyzed metapopulation capacity λ, a crucial spatial metric. Mean λ declined sharply with increasing habitat destruction across all five destruction scenarios (Figure 5a,b). The competitive structure within communities is directly impacted by the reduction in λ. For the purpose of measuring the structural features, we quantified the ratio of intraspecific to interspecific competition ($\rho ={\overline{\alpha }}_{ii}/{\overline{\alpha }}_{ij}$). Under undisturbed conditions, $\rho$ distributions differed among coexistence mechanisms (Figure 5c): Compared to the other two mechanisms, the active intransitivity region exhibited markedly lower $\rho$ values, suggesting weaker intraspecific competition and coexistence rely on interspecific intransitivity structure. In contrast, latent intransitivity and competitive hierarchy both show wide and high distributions of $\rho$ values, indicating coexistence maintained by strong intraspecific competition, where substantial niche difference masks potential intransitive structures.

To uncover the mechanism driving the transition from latent intransitivity to active intransitivity, we analyzed the relationship between the initial self-regulation strength (${\rho }_{initial}$) and its absolute destruction ($∆\rho$) following habitat destruction. (Figure 5d). The data exhibit a strict linear relationship in log-log space, indicating that habitat destruction uniformly erodes a constant fraction of self-regulation capacity, independent of the community’s initial dynamic state. Our findings reveal that structurally robust communities suffered larger absolute destruction yet maintained their dynamic state, while marginally stable communities transitioned despite minimal destruction. As shown in Figure 5f, the transition to active intransitivity is governed by the proximity to the self-regulation threshold ($\rho =1$, where intraspecific effects equal interspecific effects). Communities that retained their latent intransitivity status (orange points) originated from regions of high structural redundancy (${\rho }_{initial}\gg 1$). Conversely, communities that underwent a regime shift (purple points) were clustered exclusively in the region of minimal stability margins. For these structurally vulnerable communities, the proportional destruction was sufficient to drive the system below the threshold (${\rho }_{HD}<1$), thereby activating underlying intransitive loops to sustain coexistence. In contrast, the community itself lacked intransitive structures in the competitive hierarchy region, even if its $\rho$ value had decreased comparably much. As a result, biodiversity was eventually lost as it was unable to sustain stable coexistence through competitive hierarchy (Figure 5g,h).

3.3. Dynamical Transitions Before and After Habitat Destruction

We tracked the temporal evolutionary trajectories of five different habitat destruction patterns and chose example communities before and after habitat destruction in order to graphically illustrate the real-time influence of habitat destruction on species coexistence dynamics (Figure 6). The point at which habitat destruction took place is displayed by the dashed line (t = 5).

Regardless of the type of habitat destruction experienced, the occupancy probability of species for communities that were initially in latent intransitivity experience significant fluctuations and reorganization. Intransitive competition mechanisms, on the other hand, encourage species to quickly form new equilibrium connections, allowing them to effectively shift to active intransitivity and sustain a new steady state of coexistence (see Figure 6, second column). Communities that rely on the competitive hierarchy, on the other hand, lack an intransitive structure. The occupancy probability of disadvantaged species quickly approaches zero under comparable habitat disturbances because they are unable to endure survival stresses. The system quickly degenerates into systems with one or two species (see Figure 6, third column). The vital function that active intransitivity plays after habitat destruction is well supported by this dynamic process.

4. Discussion

Our study investigates how competitive structures and niche differentiation patterns are impacted by habitat destruction, hence reshaping mechanisms for multispecies coexistence. We adopted a multispecies coexistence space based on the generalized Lotka–Volterra framework, which originated from spatially explicit metacommunity models under the homogeneity assumption. This space simultaneously integrates structural niche difference, intransitive competition, and the actual contribution of intransitive competition through a controllable intransitive parameter. The findings indicate that a single mechanism does not always sustain multispecies coexistence. Niche differentiation, intransitive competition, or their combined impacts could identify coexistence scenarios. Under diverse processes, communities respond differently to habitat destruction; coexistence regions controlled by intransitive competition show a growing tendency. Additionally, habitat destruction weakens species coexistence and various coexistence mechanisms have different reactions.

Only in situations of strong structural niche difference (${\Omega }_1$), which takes the form of the competitive hierarchy structure, can multispecies coexistence take place when the competitive structure is primarily transitive (${\theta }_3<1$). This finding is consistent with the predictions of classical niche theory, which holds that structural niche difference is necessary for multispecies coexistence to balance out fitness difference by reducing interspecific competition in comparison to intraspecific competition [11,49,50,51]. In other words, the main mechanism maintaining coexistence in a system devoid of intransitive structures is structural niche difference. Coexistence is divided into two areas under intransitive competitive structures (${\theta }_3>1$). Intransitive competition only exists as a latent structure in one sort of coexistence area, which nevertheless mostly depends on high ${\Omega }_1$ to sustain coexistence. Under the lower ${\Omega }_1$ criteria, the other kind obtains sustained coexistence, suggesting that intransitive competitive structures directly support coexistence. The ecological significance of intransitive structures depends on whether they change the competitive outcomes between species, as demonstrated by recent research showing that these structures do not necessarily foster coexistence [22,38,41,52,53].

By decreasing the metapopulation capacity λ, habitat destruction has notably varied consequences on coexistence states. The entire coexistence space shrinks as λ decreases, which is in keeping with the traditional metapopulation theory result that species persistence is constrained by metapopulation capacity [36,37]. The coexistence space’s internal structure, however, is systematically reorganized: areas that rely on high ${\Omega }_1$, such as the competitive hierarchy and latent intransitivity zones, decline and are compelled to move in the direction of higher ${\Omega }_1$ regions. This suggests that coexistence mechanisms that just rely on niche differentiation need more structural support to endure in unfavorable situations. This is in line with recent studies that examine the connection between stability mechanisms and competitive structure. According to Ranjan et al. [53], the total competitive structure determines three-species coexistence. The coexistence region significantly reduces when competition is hierarchical and niche differentiation is inadequate. On the other hand, when habitat is being damaged, especially when habitat fragmentation is consistent, the active intransitivity area shows an expanding tendency. This suggests that intransitive competition might distribute competitive pressure to maintain species persistence under particular competitive structures. Intransitivity structures can increase species invasion capacity and persistent coexistence probability under specific conditions [41], which is compatible with recent invasion analysis findings and the cyclic competition maintenance mechanism uncovered by the classic three-species intransitivity model [18]. As a result, habitat destruction not only shrinks the “parameter space for coexistence”, but it also fundamentally alters the relative significance of various coexistence mechanisms.

Further analysis reveals that the previously indicated pattern change is caused by habitat destruction, which alters the system’s overall competitiveness, rather than the intransitive structural strength ${\theta }_3$ itself. According to the empirical results [33], habitat destruction results in a decrease in metapopulation capacity. This study showed that the intensity of intraspecific competition was weakened in comparison to interspecific competition when the area of Peneonanthe pulverulenta pulverulenta was reduced, resulting in a considerable fall in metapopulation capacity. In early communities where intraspecific competition had initially provided little benefit, this weakening impact was particularly noticeable. When habitat destruction decreases the relative advantage of intraspecific competition over interspecific competition, communities dominated by latent intransitivity mechanisms that have intraspecific competition slightly greater than interspecific competition are more likely to fall below the critical threshold for intraspecific competition-dominated coexistence. The transition from latent to active intransitivity mechanisms is brought about by the intransitive structure, which was previously inadequate to maintain coexistence alone. According to the paradigm of contemporary coexistence theory, which holds that stabilizing mechanisms rely on the relative advantage of intraspecific vs. interspecific competition, this change represents the systematic reworking of competitive ratio structures under habitat degradation [11]. The coexistence border varies systematically when habitat changes reduce the relative benefit of intraspecific competition [54]. Therefore, habitat destruction triggered changes in dominance within coexistence mechanisms via altering competitive intensity structures, rather than just decreasing community size. On the other hand, under the same habitat destruction-induced changes, communities that are governed by competitive hierarchy processes and do not have regulation of intransitive competitive structures are more likely to go extinct. As a result, this mechanism’s range of coexistence decreases after habitat destruction. These patterns suggest that, rather than only decreasing community size, habitat destruction causes internal community reorganization by changing competitive ratio structures.

Our model framework is intended as a general theoretical approach to isolate and analyze coexistence mechanisms within spatially structured ecological systems. The results are most applicable to systems satisfying two important conditions: populations are organized across discrete habitat patches, and these patches are connected through dispersal, allowing indirect interactions to emerge across space. For example, plant communities in fragmented landscapes often exhibit such patch-based dynamics via seed dispersal across habitat networks [55], and benthic or intertidal communities are similarly structured by discrete substrate patches with dispersal mediated by larval stages [56]. In order to include the impacts of intricate landscape features into a single metapopulation capacity, we use a spatial homogeneity assumption. This method highlights how competitive structures and dynamic responses caused by habitat destruction alter the processes of multispecies coexistence. Although this method emphasizes how habitat change affects coexistence mechanisms, it ignores local interactions and spatial heterogeneity between patches. According to earlier research, local interactions and spatial heterogeneity may significantly change competitive outcomes [9,57]. Additionally, extending the current model to more realistic multispecies scenarios require considering more complex network structures. Indeed, network topology can strongly influence metapopulation dynamics and persistence [58]. Incorporating such structural complexity may therefore modify the effects of intransitive competition and habitat loss identified in this study. Consequently, one of the most important ways to comprehend the ecological relevance of intransitive competition is to embed it into explicit spatial structure and network topology, and to investigate its scale dependency and topological effects in actual environments.