The Arid Zone Pollinator Paradox: hidden resilience erosion and structural-functional decoupling in dryland pollination systems

1 Introduction

Insect pollinators sustain the reproduction of roughly 90% of wild flowering plants and nearly 75% of global food crops, making them central to ecosystem functioning, agricultural productivity, and food security (Klein et al., 2007; IPBES, 2016; Ollerton, 2017; Tong et al., 2023). Their economic contribution is especially large in low- and middle-income regions, where crop yields and rural livelihoods depend heavily on animal-mediated pollination (Gallai et al., 2009; Porto et al., 2020). At the same time, pollinator communities face a stack of interacting pressures—habitat loss, land-use intensification, pesticide exposure, climate change, invasive species, emerging pathogens—with knock-on effects for biodiversity, ecosystem functioning, and resilience (Potts et al., 2010; Vanbergen and the Insect Pollinators Initiative, 2013; Goulson et al., 2015; Ganuza et al., 2022).

One complication is that losing species or interactions does not necessarily translate into an immediate drop in ecosystem functioning. Functional redundancy, response diversity, and compensatory interactions can keep ecosystem functions running despite that loss (Yachi and Loreau, 1999; Winfree, 2013). In plant–pollinator networks specifically, broad-interaction generalists can temporarily stand in for lost specialized interactions, holding visitation or pollination function roughly steady even as the underlying network changes shape (Burkle et al., 2013; Popic et al., 2013; Straka and Starzomski, 2014). Functional stability, in other words, can mask substantial changes occurring at the interaction level.

This paper proposes the Arid Zone Pollinator Paradox (AZPP) as a specific hypothesis about that masked change in dryland pollination systems. Its central claim is that, under chronic environmental stress, interaction-network structure can deteriorate before pollination function shows any measurable decline. Specialist-mediated and other functionally constrained interactions are lost disproportionately while the remaining broad-interaction generalists temporarily hold function steady through compensatory use of surviving partners. Extending beyond standard compensatory dynamics, AZPP proposes that structural erosion itself may constrain subsequent recovery: two systems can share similar current pollination function yet differ substantially in their recovery following a subsequent disturbance, depending on the extent of prior structural erosion.

That temporary persistence of function during biodiversity loss is not, on its own, paradoxical, and it is not the contribution we are claiming here. The actual hypothesis is narrower: that the compensatory phase itself progressively depletes the pool of functionally compatible interaction partners and rewiring options, making future structural recovery dependent on history rather than on current conditions alone. Three linked processes follow from this: asymmetric interaction loss, constrained compensatory rewiring, and history-dependent recovery. Here, the term Paradox refers to the apparent mismatch between comparatively stable short-term pollination function and deteriorating interaction-network structure, rather than to a logical contradiction with established ecological theory or compensatory dynamics.

Dryland ecosystems are a natural place to test this. Arid and semi-arid landscapes combine chronic water limitation, spatially patchy floral resources, compressed flowering windows, and substantial year-to-year climatic variability—conditions that plausibly tighten dependence on temporally and functionally compatible plant–pollinator interactions (Maestre et al., 2012; Gérard et al., 2020). None of the mechanisms AZPP proposes are assumed to be exclusive to drylands, though: specialist loss, compensation, and rewiring happen across ecosystems generally. Dryland specificity is treated here as a claim about boundary conditions—where these processes are expected to become more pronounced or persistent—rather than as a claim of ecological uniqueness. Testing that claim means comparing with less seasonally constrained agricultural systems to see whether the predicted pattern is genuinely stronger under arid and semi-arid conditions.

The framework draws primarily on ecological observations from Moroccan drylands, where pollinator communities support substantial agricultural production even as the landscape experiences climatic aridification, habitat homogenization, agrochemical pressure, and declining ecological connectivity (El Abdouni et al., 2022; Sabbahi, 2022). That evidence should be read cautiously, however. Existing Moroccan studies document pollinator decline, interaction loss, generalist dominance, or persistent visitation—all compatible with pieces of the proposed framework—but none have combined long-term measurements of network structure and pollination function in a way that directly demonstrates the predicted sequence. This gap is not unique to Morocco; longitudinal interaction-level monitoring remains scarce almost everywhere. AZPP is therefore offered as a testable hypothesis, not an established ecological trajectory.

AZPP differs from functional redundancy, response diversity, ecological insurance, and adaptive rewiring in one specific respect: it predicts that recovery depends on prior structural history. Those established concepts explain how compensation and resilience work, but none of them specify whether temporary functional compensation will be followed by reduced recovery capacity once structural erosion has occurred. AZPP’s first prediction is that structural indicators should decline before functional impairment becomes evident. Its second is that, among systems with comparable current pollination function, those with greater prior structural erosion should show slower or less complete recovery after comparable stress is lifted. Its third is that interaction loss should be asymmetric and compensatory replacement increasingly constrained, with functionally compatible alternatives contracting rather than remaining freely substitutable. Together these predictions provide empirical footing for distinguishing AZPP from simpler accounts of stress-driven species loss and generalist compensation.

The framework is developed primarily for bee-dominated dryland pollination systems, where the relevant interaction, phenological, and trait-matching processes can be evaluated most directly; whether it extends to other pollinator guilds is an open empirical question, since taxa differ in thermal tolerance, phenology, mobility, and interaction constraints. A rigidity-trap state is treated only as a possible endpoint—one that may occur if structural recovery stays constrained after stress alleviation—and not as an inevitable stage. In the end, AZPP’s value rests on whether longitudinal and comparative studies show that structural-functional decoupling, asymmetric interaction loss, and history-dependent recovery add explanatory and predictive power beyond what existing accounts of compensation and rewiring already provide.

2 Conceptual framework of the Arid Zone Pollinator Paradox

AZPP is formulated here as a hypothesis about the temporal relationship between interaction-network structure, functional performance, compensatory rewiring, and recovery capacity under chronic environmental stress. Its central proposition is not that pollination function necessarily declines slowly, nor that generalist pollinators invariably replace specialists. Rather, AZPP predicts that, under some chronic stress regimes, interaction-network structure can deteriorate before selected pollination functions show measurable decline, creating a period of structural-functional decoupling. During this period, compensatory interactions may preserve current function while progressively reducing the pool of compatible interactions available for future compensation and recovery.

This framework therefore distinguishes four related but non-equivalent processes: structural erosion, functional buffering, interaction rewiring, and post-stress recovery. The key prediction is that their temporal coupling may be altered by chronic stress, such that a system can retain present function while becoming progressively less capable of restoring its previous interaction architecture after disturbance. Figure 1 summarizes the proposed states and alternative trajectories, while the following sections define the mechanisms and empirical expectations underlying the framework.

2.1 Interaction-network structure and ecosystem function are related but non-equivalent dimensions

Pollination systems are not determined solely by the number or abundance of plant and pollinator species. They are organized through interactions among particular species, and the resulting network contains information about which partners interact, how frequently they interact, and how interactions are distributed across the community. Consequently, two communities with similar species richness and abundance can have substantially different interaction structures.

Relevant structural properties include interaction diversity, degree distribution, specialization, nestedness, modularity, interaction evenness, persistence of particular plant–pollinator links, and the diversity of functional traits represented within the interaction network (Olesen et al., 2007; Tylianakis et al., 2010; Bascompte and Scheffer, 2023). These properties should not, however, be treated as interchangeable measures of resilience. Different structural metrics capture different aspects of network organization, and their responses to disturbance may differ among systems. AZPP therefore does not predict that every network metric must decline monotonically. Instead, it predicts that selected indicators of interaction structure or interaction-option diversity may deteriorate before selected measures of pollination function do so. Functional performance represents a different level of organization. Depending on the system and research question, it may be quantified using visitation rates, pollen deposition, pollen receipt or quality, fruit or seed set, or crop-specific pollination efficiency. These measures are influenced not only by network structure but also by abundance, environmental conditions, plant reproductive biology, and the efficiency of individual interactions. Functional persistence should therefore not be interpreted as evidence that network structure remains intact.

This distinction provides the basis for the proposed structural-functional decoupling. A system may maintain a particular function because remaining interactions are sufficiently frequent or effective even though the number, distribution, or diversity of alternative interactions has already declined. Conversely, structural changes may occur without immediate functional consequences when ecological compensation is available. AZPP focuses specifically on the latter situation when such compensation masks a progressive reduction in future interaction options.

2.2 From species loss to asymmetric interaction loss

A central mechanistic assumption of AZPP is that environmental filtering does not necessarily remove interactions in proportion to species loss. A pollinator species may persist locally while losing access to some plant partners, and a plant may persist while losing particular pollinator partners. Consequently, interaction loss can precede or exceed species loss when ecological compatibility becomes more restrictive. This asymmetry can arise because interactions depend on multiple dimensions of compatibility, including phenological overlap, floral morphology, accessibility of floral resources, spatial co-occurrence, resource abundance, behavioral preferences, and other species-specific traits. Narrowly interacting species are potentially exposed to a larger proportional loss of their interaction opportunities when one or more of these dimensions is disrupted. By contrast, broad-interaction pollinators may retain a larger fraction of their potential partners under the same environmental change. The relevant quantity is therefore not simply whether a species survives, but which interactions remain viable after environmental filtering. A pollinator can remain abundant while losing several specialist or temporally restricted plant partners. Likewise, a plant can continue receiving visits while losing particular pollinators that contribute distinctive temporal, morphological, or behavioral functions. Such changes can simplify the interaction architecture without producing an immediate equivalent decline in aggregate visitation or reproductive output.

This distinction also explains why the four mismatch pathways proposed in this framework—phenological, spatial, trait/morphological, and resource/cue mismatches—are best regarded as alternative proximate routes to the same central mechanism: asymmetric interaction loss. They are not presented as four independent theoretical contributions. Their relative importance is expected to vary among systems, and AZPP does not require any particular pathway to dominate. The empirical observations summarized in Table 1 illustrate ecological conditions across different dryland regions that are relevant to components of the proposed mechanism, including changes in pollinator composition, resource availability, interaction opportunities, and environmental constraints. These conditions may provide proximate routes to asymmetric interaction loss. Importantly, the cited studies do not establish that structural erosion preceded functional decline or that recovery became history-dependent. The table therefore serves as empirical motivation and boundary-condition evidence, while the proposed temporal sequence and recovery dependence remain hypotheses requiring comparative longitudinal testing.

Table 1. Observations across global drylands relevant to components of the Arid Zone Pollinator Paradox.

The studies summarized here were not designed to test AZPP or its proposed mismatch pathways and therefore do not constitute empirical observations of structural–functional decoupling or rigidity-trap dynamics. Cross-regional comparisons are presented to illustrate ecological plausibility and motivate testable predictions. Longitudinal interaction-network studies remain particularly limited in Sahelian and Australian dryland systems; available studies from Morocco likewise provide informative snapshots rather than multi-year network-level tests. AZPP is therefore presented as a Morocco-grounded ecological hypothesis requiring comparative longitudinal evaluation across dryland. systems.

2.3 Compensatory interactions and the distinction between current function and future interaction capacity

Interaction loss does not necessarily produce immediate functional failure because surviving species can modify their interaction patterns. Pollination networks are dynamic, and the identity and frequency of interactions can change through behavioral flexibility, changes in abundance, resource availability, and species turnover (Olesen et al., 2007; Tylianakis et al., 2010; Bascompte and Scheffer, 2023). Such changes can buffer the effects of partner loss and help maintain pollination function. However, compensation is not equivalent to unlimited rewiring capacity. For an alternative interaction to provide effective compensation, the interacting species must co-occur and possess sufficient phenological overlap, compatible morphology and floral accessibility, appropriate behavioral or sensory compatibility, and sufficient abundance or encounter probability. Consequently, the number of theoretically possible partners can be much greater than the number of ecologically viable alternatives. This distinction is central to AZPP. During an initial phase of stress, broad-interaction pollinators may maintain function by increasing their use of surviving resources or by establishing interactions with additional available partners. Such changes can make the system appear functionally stable even while the network loses specialized links. If the set of compatible alternative partners is simultaneously contracting, however, the same compensatory process may leave fewer options for subsequent reorganization.

The framework therefore distinguishes present compensatory capacity from future interaction-option capacity. Present compensation concerns whether surviving interactions are sufficient to maintain a measured function under current conditions. Future interaction-option capacity concerns whether sufficient viable alternative interactions remain available to compensate for additional partner loss or to reconstruct interactions following stress alleviation. This distinction also prevents a potentially misleading interpretation of generalism. A broad-interaction niche does not automatically imply greater resilience. A generalist may be highly exposed to a particular pesticide, pathogen, climatic extreme, or habitat change, whereas a more specialized species may possess physiological or behavioral traits that confer greater tolerance to that stressor. AZPP therefore makes no assumption that generalists are intrinsically more resilient than specialists. Its prediction concerns the distribution and replacement of interactions, not an a priori ranking of species by resilience.

2.4 Operationalizing rewiring

For empirical testing, rewiring should be defined as the establishment or persistence of alternative interactions following changes in partner availability, rather than simply any difference between two observed networks. This distinction is important because apparent interaction turnover can result from changes in species abundance or species composition without any behavioral or ecological reorganization of surviving species (Lázaro and Gómez-Martínez, 2022). Accordingly, AZPP predicts that genuine compensatory rewiring should be evaluated against appropriate abundance- and co-occurrence-constrained expectations. If alternative interactions arise largely as a stochastic consequence of increased abundance or random encounter opportunities, observed replacement should resemble such null expectations. By contrast, AZPP predicts that under structural erosion, interaction replacement becomes increasingly constrained by compatibility, resulting in fewer viable alternatives than would be expected from species abundance and co-occurrence alone. This provides a direct empirical route for distinguishing AZPP from a simpler explanation in which environmental stress removes sensitive species and surviving generalists passively fill the remaining interaction space.

2.5 Structural-functional decoupling as the empirical signature of AZPP

The preceding mechanisms lead to the central observable phenomenon of AZPP: structural-functional decoupling. The hypothesis predicts that, during at least part of the stress trajectory, selected structural indicators may decline while selected functional indicators remain comparatively stable. Importantly, this is not equivalent to claiming that ecosystem function is globally stable. A network may maintain crop-specific fruit set while losing interaction diversity, or maintain visitation rates while losing specialist interactions. The predicted decoupling is therefore metric- and function-specific, and its detection requires independent measurement of structural and functional dimensions.

A suitable empirical design should consequently track both dimensions through time. Structural measurements may include specialist interaction persistence, interaction diversity, network specialization, modularity, nestedness, partner diversity, and functional trait-space diversity. Functional measurements may include visitation stability, pollen deposition, pollen receipt, fruit or seed set, and crop-specific pollination efficiency. No single structural metric is expected to be universally diagnostic.

The strongest evidence for AZPP would occur when:

• structural indicators decline before selected functional indicators;
• the persistence of function can be attributed to surviving or newly established compensatory interactions;
• the pool of compatible alternative interactions contracts during the period of functional stability;
• subsequent recovery is slower or less complete in systems that experienced greater prior structural erosion.

Thus, the hypothesis is not simply that structure changes before function. The more distinctive proposition is that structural erosion occurring during the compensatory phase may alter the subsequent trajectory of recovery.

2.6 History-dependent recovery and ecological memory

Recovery following environmental stress can be influenced by ecological legacies of previous disturbance, as prior conditions may alter species composition, resource availability, interaction opportunities, and the capacity of ecological systems to reorganize following subsequent environmental change. Such effects are commonly described in terms of ecological memory or historical legacies (Johnstone et al., 2016). AZPP does not propose ecological memory or history-dependent recovery as new concepts. Rather, it specifies a plant–pollinator network mechanism through which such historical dependence may arise and generates a corresponding testable prediction. During chronic stress, asymmetric loss of functionally constrained interactions may occur while surviving interactions temporarily maintain selected pollination functions. If this structural erosion progressively reduces the pool of viable alternative interactions, it may leave an interaction-network legacy that persists into subsequent recovery. Thus, systems with similar current species abundances, environmental conditions, and pollination function may nevertheless differ in their capacity for recovery because of differences in prior interaction-network erosion. In particular, a system in which specialist interactions and alternative partner diversity have been largely retained may recover more rapidly following stress alleviation than a system in which comparable current function was maintained despite progressive loss of compatible interactions and increasing dependence on a smaller set of broad interactions. AZPP therefore predicts that greater prior structural erosion will be associated with slower or less complete recovery of interaction-network structure following comparable stress alleviation.

This prediction distinguishes AZPP from a purely state-based view of resilience, in which recovery is determined primarily by current environmental conditions and system state. The distinctive proposition is not simply that ecological history matters, but that prior erosion of plant–pollinator interaction structure—and specifically the associated contraction of viable interaction alternatives—can become an explanatory variable for subsequent recovery. Accordingly, among systems with comparable current pollination function, recovery should be slower or less complete in those that experienced greater prior erosion of interaction structure, after accounting as far as possible for current environmental conditions, species abundance, habitat quality, connectivity, dispersal or immigration, resource availability, and disturbance intensity. The prediction is explicitly falsifiable. If recovery trajectories do not differ systematically with prior interaction-network erosion once relevant current-state, demographic, spatial, and disturbance variables are controlled, the history-dependent component of AZPP would not be supported.

2.7 Conditional rigidity and alternative trajectories

Persistent structural erosion may, under some conditions, produce a state in which the network becomes increasingly dependent on a restricted subset of interactions and has fewer viable routes for reorganization. This possibility is conceptually related to the rigidity-trap literature, in which short-term stability or persistence can coexist with declining adaptive capacity (Walker et al., 2004; Carpenter and Brock, 2008). Within AZPP, however, a rigidity trap is not a defining component or inevitable endpoint. It represents one possible trajectory that may emerge if stress persists, compatible interaction alternatives continue to contract, and the remaining interactions become increasingly dominant. Other systems may instead recover structurally after stress alleviation, particularly where habitat connectivity, resource continuity, behavioral flexibility, or species persistence preserve sufficient interaction options. This distinction is important because it makes AZPP broader and more testable than a deterministic collapse model. The hypothesis does not require threshold behavior, irreversible transitions, or functional collapse. The minimum prediction is structural-functional decoupling. A stronger form of support would additionally demonstrate that structural history predicts recovery. A rigidity-trap trajectory would represent a still stronger and conditional outcome.

Accordingly, AZPP allows for several alternative trajectories rather than a single obligatory sequence. Where interaction structure is largely preserved, functional stability may be accompanied by subsequent structural and functional recovery. Where structural erosion occurs but compensation remains effective, function may persist temporarily despite a progressive reduction in available interaction options, potentially resulting in history-dependent recovery. Under more persistent or severe constraints, continued structural erosion and increasingly constrained rewiring may eventually lead to functional impairment and, in some systems, a possible rigidity-trap trajectory. These outcomes are therefore treated as alternative consequences of the interaction between stress, network structure, compensation, and recovery rather than as obligatory sequential stages.

2.8 Conceptual hierarchy and testable predictions

The framework can be summarized as a causal hierarchy in which chronic environmental stress imposes environmental filtering and mismatch, leading to asymmetric interaction loss and subsequent compensatory changes in interactions or rewiring. As compatible alternative partners become less available, compensation may increasingly depend on a reduced pool of interactions, creating structural-functional decoupling in which selected aspects of network structure deteriorate while current function remains relatively stable. The resulting interaction history may then influence recovery capacity, such that systems experiencing greater prior structural erosion may show slower or less complete recovery after stress alleviation. Under persistent constraints, continued structural erosion and constrained rewiring may contribute to functional impairment in some systems, with a rigidity-trap trajectory representing a possible downstream outcome rather than a defining or inevitable stage of AZPP.

The four mismatch pathways are treated as proximate mechanisms that may contribute to asymmetric interaction loss, whereas multi-stressor exposure represents a potential modifier of the magnitude or persistence of structural erosion. Multi-stressor synergy is therefore not a defining prediction of AZPP. Likewise, nonlinear or threshold-like behavior may occur in some systems when compensatory options become sufficiently constrained, but such behavior is conditional and is not required for the hypothesis to be supported.

AZPP builds on, rather than replaces, established concepts of functional redundancy, response diversity, ecological insurance, adaptive rewiring, and ecological memory (Table 2). These concepts describe important mechanisms through which biodiversity, interaction flexibility, or historical legacies can influence ecosystem stability and recovery. AZPP differs in the specific temporal relationship it proposes among these processes: chronic stress may erode interaction structure and viable interaction options while selected pollination functions remain buffered, creating an apparent decoupling between functional stability and structural erosion. The resulting structural history may subsequently influence recovery. Thus, AZPP is not simply a sequence of functional redundancy followed by loss of response diversity. Its distinguishing proposition is that structural erosion occurring during a period of functional compensation can itself alter the subsequent capacity for recovery, generating a testable prediction that recovery trajectories should depend on the prior degree of interaction-network erosion even when current conditions are comparable.

Table 2. Conceptual comparison of the Arid Zone Pollinator Paradox with related existing concepts.

The comparison is conceptual rather than taxonomic. AZPP does not claim that any individual mechanism represented by these established concepts is novel. Its proposed contribution lies in integrating asymmetric interaction loss, structural-functional decoupling, functional compensation, constrained interaction options, and history-dependent recovery into a single testable hypothesis for dryland pollination networks.

This hierarchy generates three principal empirical predictions, developed fully in Section 7: structural-functional decoupling under chronic stress, history-dependent recovery following stress alleviation, and asymmetric interaction loss accompanied by constrained rewiring. Together, they distinguish AZPP from the simpler expectation that environmental stress removes sensitive species, surviving broad-interaction taxa compensate temporarily, and pollination function eventually declines under continued stress. Under that simpler account, recovery should depend mainly on the system’s current state and prevailing environmental conditions. Under AZPP, the preceding history of interaction-network erosion itself shapes subsequent recovery—a claim longitudinal monitoring, recovery experiments, and interaction-level analyses can directly test. Section 7 specifies the observations that would weaken or strengthen each prediction.

3 Proximate pathways of pollinator–plant mismatch

The four pathways below are treated as alternative proximate routes to asymmetric interaction loss rather than as independent predictions of AZPP. They illustrate how chronic environmental constraints may disrupt plant–pollinator interactions through different ecological dimensions. Here, mismatch refers to a failure in one or more ecological conditions required for the establishment or maintenance of plant–pollinator interactions, rather than to the environmental drivers that generate those failures. Drivers such as climate change, habitat fragmentation, agrochemical exposure, and altered floral signaling can generate mismatch by disrupting temporal synchrony, spatial co-occurrence, functional trait compatibility, or sensory recognition between interacting partners. Accordingly, the four pathways are classified according to the ecological dimension in which interaction failure occurs: temporal, spatial, morphological, and attraction/cue mismatch (Figure 2). These pathways may contribute to asymmetric interaction loss, particularly under chronic environmental stress, while surviving or newly established interactions may temporarily buffer selected pollination functions.

Throughout this framework, specialist pollinator refers broadly to pollinator species exhibiting relatively narrow ecological niches that constrain their capacity to maintain interactions under environmental change (Olesen et al., 2007; Winfree, 2013). Such specialization may involve dietary specialization (e.g., oligolecty), interaction specificity, morphological trait matching, phenological dependence, habitat specialization, or reliance on particular floral sensory cues (Tylianakis et al., 2010; Gérard et al., 2020). These dimensions do not necessarily coincide within individual species, and interaction specialization alone does not imply morphological or sensory specialization (Olesen et al., 2007; Tylianakis et al., 2010; Gérard et al., 2020). Vulnerability may increase when multiple forms of specialization overlap, thereby reducing the number of ecologically compatible interaction alternatives available under environmental stress. The framework is therefore intended primarily to represent bee-dominated dryland pollination systems in which specialist wild bees coexist with more broadly interacting taxa, while recognizing that comparable ecological constraints may occur in other pollinator guilds.

A related terminological distinction is important. In this manuscript, generalist and specialist describe the breadth of a species’ interaction niche within the pollination network and are not measures of resilience to environmental stress. A thermally specialized dryland bee may tolerate heat extremes better than a broadly foraging generalist, while interaction generalists may remain vulnerable to pesticide exposure or pathogen transmission. References to broader-interaction taxa sustaining pollination function therefore concern their capacity to substitute for lost interaction partners under particular conditions, rather than a general claim of ecological robustness.

Temporal mismatch emerges when climatic or environmental change disrupts synchrony between flowering phenology and pollinator activity periods. Altered precipitation regimes, warming, or shifts in flowering onset may reduce interaction opportunities for pollinators with narrow phenological or resource windows, while more flexible taxa may partially compensate through host switching or extended foraging periods (Kudo and Ida, 2013; Gérard et al., 2020). Such compensation is not assumed to be universal. Where specialized interactions are lost and compatible alternatives are limited, temporal mismatch may contribute to structural erosion while selected visitation or pollination functions remain temporarily stable (Straka and Starzomski, 2014; Didevarasl et al., 2025).

Spatial mismatch occurs when environmental change reduces effective spatial co-occurrence between flowering plants and their pollinators, preventing otherwise compatible partners from interacting. Habitat fragmentation, agricultural homogenization, and loss of microclimatic refugia can increase foraging distances and reduce suitable habitat, with potentially greater consequences for taxa with restricted ranges or movement capacities (Moreira et al., 2015; Kremen and Merenlender, 2018). In drylands, spatially patchy and climatically variable resources may further constrain interaction opportunities. Broadly foraging pollinators may nevertheless continue to exploit mass-flowering crops or resource-rich patches, allowing some functions to persist while spatially constrained interactions are lost (Magrach et al., 2017; Gay et al., 2024).

Morphological mismatch occurs when changes in environmental conditions disrupt functional compatibility between floral traits and pollinator morphology. Climatic stress may alter pollinator morphology, activity, or floral traits in ways that reduce previously compatible interactions, particularly where interactions depend on narrow trait matching (Gérard et al., 2020; Michael et al., 2025). Specialized interactions may therefore be lost while other pollinators retain sufficient compatibility to provide partial functional compensation (Loy and Brosi, 2022). Where compatible alternatives are limited, however, morphological mismatch may instead produce more immediate functional impairment.

Attraction/cue mismatch occurs when alterations in floral visual, olfactory, chemical, or reward signals reduce correspondence between plant signals and pollinator sensory responses, thereby impairing partner detection or recognition. Climatic stress, atmospheric pollution, and agrochemical exposure can alter floral volatile compounds, nectar chemistry, coloration, or other cues involved in pollinator attraction and host recognition (Gérard et al., 2020; Peralta et al., 2024). Such changes may disproportionately affect interactions involving pollinators that depend on relatively specific sensory cues, while taxa with broader sensory or foraging responses may continue to exploit available resources. Where alternative compatible interactions remain available, this may contribute to temporary functional buffering; where alternatives are limited, functional effects may emerge more rapidly.

The four mismatch pathways therefore represent alternative proximate routes to asymmetric interaction loss, rather than separate theoretical mechanisms or independent predictions of AZPP. Their relative importance is expected to depend on specialization, the availability of compatible alternative partners, resource continuity, and the temporal and spatial configuration of the network. The central hypothesis concerns the resulting sequence: under some chronic stress regimes, constrained interactions may be lost disproportionately, while selected pollination functions remain temporarily buffered; progressive structural erosion may then reduce the pool of compatible interaction options and influence subsequent recovery. The pathways themselves are not proposed as novel components of AZPP, but as ecological routes through which the core mechanism may arise. The dryland observations summarized in Table 1 provide empirical motivation for these components but do not demonstrate the proposed temporal sequence or history-dependent recovery.

4 Theoretical framework: biodiversity, connectivity, and the structural basis of the paradox

4.1 Functional diversity and rewiring capacity

Rewiring capacity was defined operationally in section 2.4. The discussion below turns to how functional diversity among surviving taxa shapes whether that capacity is actually realized. Functional diversity can contribute to the resilience of dryland pollination systems through response diversity, niche complementarity, and differences in tolerance to environmental stress (Yachi and Loreau, 1999; Winfree, 2013). In arid and semi-arid agroecosystems, interspecific variation in physiological and behavioral traits—including thermal tolerance, desiccation tolerance, phenological sensitivity, and foraging flexibility—can influence which pollinators remain active under drought, heat, and reduced floral continuity (Nacko et al., 2023; Gonzalez et al., 2025). Functional complementarity can likewise stabilize pollination through differences in resource use and interaction efficiency, with traits such as body size, tongue length, and proboscis morphology influencing floral accessibility and pollen transfer (Gérard et al., 2020). These forms of diversity provide potential buffering capacity, but they do not imply that generalists are inherently more resilient than specialists; their contribution to AZPP depends on whether surviving taxa retain sufficient compatibility to compensate for lost interactions.

Environmental stress can alter functional traits through changes in body size, activity timing, morphology, or foraging behavior. Because some specialized plant–pollinator interactions depend on relatively narrow trait matching, such changes may selectively disrupt particular interactions even when other taxa remain functionally compatible (Loy and Brosi, 2022). Rewiring refers to the establishment of alternative interactions when existing partners decline, disappear, or become phenologically mismatched (Lázaro and Gómez-Martínez, 2022; Su et al., 2025). Such flexibility can buffer functional loss following partner turnover, but the availability of alternative interactions is not unlimited. Environmental filtering, habitat fragmentation, and resource simplification may reduce the number of species and interactions that are simultaneously compatible in space, time, and functional traits. AZPP therefore predicts that potential interaction options may contract before selected pollination functions decline, increasing the likelihood that observed compensation depends on a progressively smaller interaction pool.

We hypothesize that this process can create a distinction between current compensatory capacity and future interaction-option capacity. In the short term, alternative interactions may stabilize selected aspects of pollination function despite the loss of specialized interactions. However, if environmental filtering continues to reduce the pool of compatible partners, subsequent compensation may become increasingly constrained. The resulting structural erosion need not immediately impair function, but it may reduce the range of interactions available for future reorganization and recovery.

4.2 Network structure and conditional rigidity

Network structure can influence the stability and flexibility of pollination systems through properties such as modularity, nestedness, and connectance (Olesen et al., 2007; Bascompte and Scheffer, 2023). Modularity may limit the propagation of perturbations among interaction compartments, whereas nestedness reflects asymmetric organization in which specialist interactions are often embedded within the interaction portfolios of more connected taxa. The consequences of these properties are context-dependent: nestedness can contribute to robustness under some forms of species loss while also concentrating ecological dependence around a smaller number of interaction partners (Tylianakis et al., 2010). AZPP does not predict a universal directional change in any single topological metric. Rather, it predicts that loss of interaction diversity and compatible alternatives may alter network structure before selected pollination functions decline, with changes in modularity, nestedness, connectance, or other structural properties interpreted jointly rather than as standalone signatures of a rigidity trap.

In dryland agroecosystems, landscape heterogeneity and semi-natural habitats can support diverse pollinator assemblages and a wider range of plant–pollinator interactions, whereas habitat homogenization and floral-resource simplification may reduce interaction diversity and the availability of compatible partners (Tylianakis et al., 2008; Samih et al., 2024). Fragmentation may be particularly consequential for pollinators with restricted movement or habitat requirements, while mass-flowering crops and managed A. mellifera may alter the distribution of realized interactions within simplified agricultural matrices (Magrach et al., 2017; Gay et al., 2024). Under AZPP, the relevant consequence is not simply a change in network connectivity, but a potential contraction of the pool of compatible interactions available for compensation and subsequent recovery.

To conceptualize a possible downstream trajectory of structural erosion, we draw on the rigidity-trap concept from social–ecological resilience theory, in which reinforcing feedbacks can stabilize systems in simplified configurations and reduce adaptive capacity (Walker et al., 2004; Carpenter and Brock, 2008; Tidball et al., 2016). In AZPP, such a trajectory is hypothesized to arise only under conditions where interaction loss, constrained rewiring, and persistent environmental limitations reinforce one another. Temporary dominance by a limited set of broadly interacting pollinators may maintain selected aspects of pollination function while reducing the diversity of interaction pathways available for future reorganization. Importantly, this represents a possible downstream outcome rather than a defining or inevitable prediction of AZPP.

Importantly, a rigidity-trap interpretation requires history-dependent recovery, rather than simple temporary compensation. Following stress alleviation, systems with greater prior structural erosion may recover more slowly or incompletely because specialist recolonization, demographic rebuilding, and re-establishment of compatible interactions can occur over longer timescales than the initial loss of interactions (Bartomeus et al., 2013; Burkle et al., 2013). Structural simplification may further constrain recovery if the pool of compatible interaction partners has contracted. Thus, two systems experiencing similar current environmental conditions and exhibiting similar levels of pollination function could nevertheless differ in recovery trajectory because of differences in their prior interaction-network history. Under AZPP, history-dependent recovery is the critical test for distinguishing persistent structural consequences from transient functional compensation. A rigidity-trap trajectory should therefore be considered only when structural recovery remains constrained after comparable stress alleviation, rather than simply because a network becomes more generalized during disturbance. Conversely, rapid recovery of specialist taxa, interaction structure, and compatible alternatives following stress reduction would indicate transient compensation rather than a rigidity trap.

The framework generates several empirically testable structural signatures, including loss of interaction diversity, contraction of functional trait space, changes in interaction turnover, and altered network organization (Bartomeus et al., 2013; Spiesman and Inouye, 2013; Bascompte and Scheffer, 2023). However, no single metric is expected to provide a sufficient empirical signature of AZPP. The strongest evidence would instead come from joint temporal analysis showing structural erosion while selected pollination functions remain buffered, followed by slower or less complete recovery in systems with greater prior structural erosion. This distinction allows AZPP to be tested against the simpler expectation that functional compensation merely reflects current species abundance and generalism. Longitudinal network studies that track interaction structure, function, and recovery following stress alleviation will therefore be critical for evaluating the hypothesis.

5 Multi-stressor amplification as a contextual modifier

Multi-stressor exposure is treated here as a context-dependent modifier of the core AZPP sequence, rather than as a necessary condition for its occurrence. Climatic, ecological, toxicological, and, in some contexts, epidemiological stressors may act simultaneously to intensify environmental filtering, reduce interaction opportunities, or constrain compensatory rewiring beyond the effects of individual stressors acting independently (Vanbergen and the Insect Pollinators Initiative, 2013; Goulson et al., 2015; Ganuza et al., 2022). The magnitude and direction of such interactions are expected to depend on stressor identity, timing, exposure, and ecological context. Figure 3 summarizes several illustrative combinations and their potential consequences for interaction structure.

For example, elevated temperature can increase metabolic demand, desiccation stress, or sensitivity to toxicological exposure, while drought-related resource limitation and pathogen exposure may jointly affect pollinator condition and foraging performance (Siviter et al., 2021; Bretzlaff et al., 2024). Interactions involving managed and wild pollinators may likewise modify resource availability, competition, or pathogen exposure under particular conditions (Magrach et al., 2017; Maurer et al., 2024). Within AZPP, the relevance of these combinations lies not in assuming that any individual stress response is maladaptive, but in whether their cumulative effects alter interaction frequencies, selectively increase interaction loss, or reduce the availability of compatible alternatives. These examples therefore illustrate potential amplification pathways rather than additional mechanisms or predictions of AZPP.

Nonlinear or threshold-like responses may arise in some systems when the pool of compatible interaction alternatives becomes sufficiently constrained. Such responses could increase the persistence of structural erosion or contribute to subsequent functional impairment, but threshold behavior is not required for AZPP and is not assumed to occur universally. The core hypothesis remains testable under single- or multiple-stressor conditions through measurement of interaction-network structure, selected pollination functions, and subsequent recovery trajectories.

6 Agroecological implications and recovery

Landscape heterogeneity, floral-resource continuity, nesting opportunities, and spatial refugia may help preserve a broader pool of compatible interaction partners than highly simplified agricultural systems (Tylianakis et al., 2008; Samih et al., 2024). In Moroccan dryland systems, approaches such as Farming with Alternative Pollinators and traditional water-management infrastructure such as khettaras provide potentially informative contexts for testing whether resource continuity and habitat heterogeneity are associated with slower structural erosion, greater availability of compatible interaction alternatives, or more complete recovery after stress alleviation (Houssni et al., 2023; Bencharki et al., 2024, Bencharki et al., 2025). These applications are therefore proposed as testable contexts rather than established management prescriptions.

Because recolonization, demographic recovery, and re-establishment of compatible interactions can require multiple generations (Bartomeus et al., 2013; Burkle et al., 2013), management evaluation should consider interaction-network structure and subsequent recovery rather than immediate changes in visitation or pollination function alone. These implications follow from the history-dependent recovery prediction but are downstream applications rather than components of the central hypothesis.

7 Testable predictions of the Arid Zone Pollinator Paradox

To advance AZPP from a conceptual framework toward a falsifiable ecological hypothesis, we propose three principal predictions that can be evaluated through longitudinal monitoring, experimental manipulation, and comparative analyses across dryland systems. Together, they address three linked components of the framework: asymmetric interaction loss and constrained compensation, structural-functional decoupling under chronic stress, and history-dependent recovery. A separate boundary-condition prediction specifies how the expected magnitude or persistence of these processes should differ between dryland and non-dryland systems. The predictions are deliberately formulated without fixed universal thresholds because dryland pollination systems differ substantially in climatic variability, landscape structure, species composition, and monitoring methodology.

7.1 P1. Structural-functional decoupling

Under chronic dryland stress, selected aspects of network structure are predicted to deteriorate before comparable declines become evident in selected pollination functions. In particular, specialist persistence, interaction diversity, and other structural properties may decline while short-term pollination outputs remain comparatively stable through compensatory interactions. The predicted decoupling is expected to become more pronounced under increasing cumulative stress and, where compensation persists, may be detectable before substantial reductions in crop pollination or plant reproductive success occur. Longitudinal studies can test this prediction by jointly tracking pollinator abundance and composition, plant–pollinator interactions, network structure, visitation, pollen deposition, and reproductive output across stress gradients or through time. Observations showing that functional performance declines concurrently with or before the relevant structural changes would weaken the proposed sequence.

7.2 P2. History-dependent recovery

Among systems exhibiting comparable current pollination function and similar post-stress environmental conditions, those with greater prior structural erosion are predicted to show slower or less complete structural recovery following stress alleviation. This prediction distinguishes AZPP from transient compensatory dynamics: if structural degradation simply reflects current environmental conditions, systems with similar present conditions should exhibit similar recovery trajectories regardless of prior network history. By contrast, persistent differences associated with prior structural erosion would support a history-dependent recovery mechanism. This prediction can be tested through longitudinal restoration or stress-release experiments that compare recovery trajectories among systems differing in the magnitude or duration of prior structural erosion. To distinguish historical network effects from alternative explanations, analyses should account for current habitat quality, pollinator and plant abundance, floral-resource availability, landscape connectivity, dispersal or immigration potential, disturbance intensity, phenological conditions, and relevant sampling differences. Where feasible, the strongest design would compare systems with contrasting disturbance histories that are subsequently brought to comparable environmental conditions and monitored through recovery. Persistent differences in recovery after these current-state and demographic factors are controlled would provide stronger evidence for a history-dependent effect of prior interaction-network erosion. Conversely, rapid and equivalent recovery across contrasting disturbance histories, or disappearance of the effect after accounting for these alternative variables, would weaken the history-dependent component of AZPP.

7.3 P3. Asymmetric interaction loss and constrained rewiring

Under chronic dryland stress, interactions involving specialist or interaction-constrained taxa are predicted to decline disproportionately relative to interactions involving broadly interacting taxa, while the pool of functionally compatible alternative interactions contracts as structural erosion progresses. Consequently, compensatory interaction replacement should become increasingly constrained rather than remaining indefinitely available through generalist replacement. This prediction can be evaluated by tracking plant–pollinator interactions through time and quantifying interaction loss, partner replacement, interaction turnover, and the availability of alternative partners based on phenological overlap, morphology, spatial co-occurrence, and abundance. Abundance- and co-occurrence-constrained null models can be used to distinguish compatibility-based replacement from turnover expected solely from changes in species abundance. Patterns in which interaction losses are largely explained by proportional abundance changes, or in which compatible alternatives remain abundant despite progressive structural erosion, would weaken this prediction.

Dryland boundary-condition prediction. Because AZPP is formulated as a dryland-focused hypothesis, its expression is predicted to differ in magnitude or persistence between dryland and non-dryland pollination systems. Under comparable levels of environmental disturbance, dryland systems are expected to exhibit greater or more persistent structural-functional decoupling and stronger effects of prior structural erosion on subsequent recovery. This expectation reflects the possibility that chronic water limitation, greater variability in floral-resource availability, and spatial or temporal discontinuity in resource supply may constrain the persistence of compatible interaction alternatives. Comparative tests should account for current habitat quality, species abundance, floral-resource availability, landscape connectivity, disturbance intensity, phenology, and sampling effort. The hypothesis does not predict that structural-functional decoupling or history-dependent recovery is absent from non-dryland systems; rather, it predicts that these effects should be more pronounced or persistent under dryland conditions exposed to comparable disturbance.

Together, these predictions define the central empirical logic of AZPP. Chronic stress is expected to generate asymmetric interaction loss and contraction of compatible interaction options (P3), allowing selected pollination functions to remain buffered despite measurable structural erosion (P1). If this structural erosion alters the future interaction-option pool, systems with greater prior erosion should subsequently exhibit slower or less complete recovery after comparable stress alleviation (P2). This sequence is proposed as a falsifiable hypothesis rather than an assumed or inevitable ecological trajectory.

8 Structural-functional decoupling as an empirical framework

The proposed hypothesis does not require a single composite index to identify structural-functional decoupling. Because network structure and ecosystem functioning are multidimensional properties measured using variables with different biological meanings and statistical characteristics, collapsing them into a single scalar may obscure rather than clarify the ecological trajectory. We therefore recommend a two-dimensional state-space approach in which structural and functional indicators are measured and, where appropriate, standardized within each study system and tracked jointly through time. Structural change can be represented by complementary indicators including specialist persistence, interaction diversity, functional trait-space diversity, network organization, and the availability of functionally compatible interaction partners. Functional change can be assessed independently using context-appropriate measures such as visitation stability, pollen deposition, fruit or seed set, or crop-specific pollination efficiency. The specific indicators should be selected according to the focal pollination system and measured consistently across sampling periods.

The key empirical signature is not a particular numerical threshold but a temporal divergence between these two dimensions: structural indicators decline while selected measures of pollination function remain comparatively stable. This pattern should be evaluated longitudinally rather than inferred from a single observation. A subsequent reduction in recovery following stress alleviation would provide stronger evidence for the history-dependent component of AZPP and help distinguish persistent structural consequences from transient compensation. Crucially, this state-space formulation avoids imposing universal weights or normalization procedures on heterogeneous ecological variables while retaining the central empirical predictions of the framework. Future studies may develop composite decoupling metrics if standardized measurement protocols and empirical calibration become available, but no such metric is required to test the core AZPP predictions.

If supported empirically, this framework could inform conservation monitoring by encouraging greater attention to interaction-network structure and recovery history alongside conventional species- and function-based indicators. Such applications remain downstream implications of the hypothesis and would require independent evaluation of indicator performance and management feasibility. Ultimately, the framework would be weakened if longitudinal observations consistently showed tight temporal coupling between structural and functional change, recovery trajectories independent of prior structural history, or compensatory interactions remaining broadly available despite progressive structural erosion. Conversely, support would be strengthened by repeated evidence of temporal structural-functional decoupling, compatibility-constrained interaction replacement, and recovery trajectories that systematically differ according to prior network erosion.

9 Significance and scope of the framework

The scientific value of AZPP lies not in claiming universal applicability, but in providing a mechanistically explicit and empirically testable framework for investigating hidden resilience erosion in dryland pollination systems. Rather than functioning solely as a synthesis of pollinator decline, the framework proposes that chronic environmental stress may produce a temporal divergence between interaction-network structure and selected aspects of pollination function. Apparent short-term stability of pollination services may therefore reflect transient compensation rather than preservation of the underlying diversity of interaction pathways (Bascompte and Scheffer, 2023).

A central contribution of the framework is to distinguish this proposed mechanism from simpler explanations based on neutral interaction turnover or generalized compensation. Apparent persistence of pollination function during biodiversity loss may arise partly from stochastic partner replacement driven by changes in abundance. Under AZPP, however, compensatory replacement is predicted to become progressively constrained as the pool of functionally compatible interaction partners contracts. This distinction can be tested using longitudinal interaction data together with abundance- and co-occurrence-constrained null models that account for covariance between abundance and functional traits (Blüthgen et al., 2008; Vázquez et al., 2009; Bartomeus et al., 2013). Evidence that interaction restructuring is adequately explained by abundance alone would weaken the proposed mechanism, whereas persistent non-random replacement associated with phenological, morphological, spatial, or behavioral compatibility would provide stronger support.

The framework is deliberately narrower than a general theory of pollinator resilience. Its empirical focus is on dryland, predominantly bee-dominated pollination systems in which chronic environmental constraints may reduce the availability of compatible interaction alternatives. Extension to other pollinator guilds, climatic regimes, or ecological contexts should therefore be treated as an empirical question rather than assumed from the present framework. The core AZPP hypothesis is defined by three principal predictions: structural-functional decoupling under chronic stress, history-dependent recovery following stress alleviation, and asymmetric interaction loss accompanied by constrained rewiring. Multi-stressor exposure is treated as a context-dependent modifier that may alter the magnitude or persistence of these processes, but it is neither a defining mechanism nor a required condition of AZPP. Likewise, nonlinear or threshold-like responses and rigidity-trap dynamics are considered possible downstream outcomes under sufficiently persistent structural erosion and constrained rewiring, but neither is required for the hypothesis to be supported. The framework does not require a composite or diagnostic decoupling metric; rather, structural and functional dimensions are intended to be measured independently and evaluated jointly through time. The four mismatch pathways described in the framework represent alternative proximate routes to asymmetric interaction loss rather than additional predictions. This scope limitation is important because pollination-network responses can vary substantially among ecological contexts and taxonomic groups (Ollerton, 2017; Hall and Martins, 2020).

The practical significance of the framework lies in encouraging resilience assessment to move beyond species counts or short-term pollination outputs alone. Monitoring that jointly considers interaction-network structure, selected functional measures, and recovery trajectories may help identify structural erosion that remains partially masked by functional compensation. Such an approach complements broader resilience-oriented monitoring frameworks that emphasize ecological change before severe functional degradation becomes evident (Walker et al., 2004; Scheffer et al., 2009; Bascompte and Scheffer, 2023). However, the value of any individual structural or functional indicator will require empirical validation across contrasting dryland systems.

AZPP should therefore be evaluated as a falsifiable, dryland-focused hypothesis whose explanatory value depends on whether its predicted combination of asymmetric interaction loss, structural-functional decoupling, and history-dependent recovery can be demonstrated longitudinally and distinguished from abundance-driven interaction turnover or transient compensation. Failure to reproduce these relationships across appropriately designed studies would require refinement or rejection of the proposed framework rather than reinterpretation of contradictory evidence as support.

10 Conclusion

Pollinator conservation is increasingly moving from retrospective biodiversity assessment toward approaches that consider ecosystem resilience and recovery. AZPP proposed here reframes a potentially important feature of dryland pollination systems: selected pollination functions may remain apparently stable while interaction-network structure and the pool of compatible interaction alternatives progressively erode. This structural-functional decoupling may arise when functionally constrained or specialist-mediated interactions are lost disproportionately and remaining broad-interaction taxa provide partial compensation. Apparent stability in pollination function should therefore not necessarily be interpreted as evidence that the underlying pollination system remains structurally resilient. Instead, functional persistence may coexist with declining interaction diversity and reduced recovery capacity following subsequent disturbance. The central contribution of the framework is to integrate asymmetric interaction loss, constrained compensatory rewiring, and history-dependent recovery into a single explicitly testable hypothesis. The three predictions developed here provide complementary tests of this framework: structural indicators should decline before selected functional impairment becomes evident; systems with greater prior structural erosion should show slower or less complete recovery under comparable current conditions; and interaction loss should be asymmetric, with the pool of functionally compatible alternatives contracting rather than remaining freely substitutable. Longitudinal network monitoring, comparative analyses, stress-release or restoration experiments, and abundance- and co-occurrence-constrained null models provide complementary approaches for testing these predictions. The four mismatch pathways identified in the framework—temporal, spatial, morphological, and attraction/cue-related—are not proposed as novel mechanisms individually. Rather, they represent alternative proximate routes through which environmental filtering may disrupt interactions and contribute to the structural erosion described by AZPP. Similarly, multi-stressor interactions are treated as potential contextual modifiers rather than necessary components of the hypothesis, while threshold-like responses and rigidity-trap dynamics are considered conditional downstream outcomes rather than inevitable stages. The framework is grounded primarily in ecological observations from Moroccan drylands but is not presented as a Morocco-specific universal pattern. Its broader applicability remains an empirical question. Comparative testing across arid and semi-arid agroecosystems, and eventually across contrasting climatic and agricultural contexts, will be necessary to determine whether structural-functional decoupling, asymmetric interaction loss, and history-dependent recovery are consistently expressed and under what boundary conditions. Such studies should also clarify when compensation remains effective, when compatible interaction alternatives become constrained, and whether structural erosion ultimately translates into measurable functional impairment. Ultimately, AZPP should be regarded as a dryland-focused, falsifiable ecological hypothesis rather than an established universal trajectory. Its value will depend on whether empirical testing demonstrates explanatory and predictive advantages over existing accounts of functional compensation, redundancy, ecological insurance, and rewiring. If supported, the framework would suggest that assessment of pollination resilience should consider not only species richness and current ecosystem function, but also interaction-network structure, functional compatibility, and recovery history. This perspective could help identify forms of structural erosion that remain partially masked by functional compensation and improve understanding of how dryland pollination systems respond to continuing environmental change.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. During the preparation of this work the author used Gemini to improve the readability, language, and writing efficiency. After using this tool, the author reviewed and edited the content as needed and take full responsibility for the content of the published paper.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Muzafar Riyaz, Zoological Survey of India, India

Reviewed by: Juan Carlos López-Acosta, Universidad Veracruzana, Mexico

Patrick Lhomme, University of Mons, Belgium