From nitrogen enrichment to ecosystem recovery: mechanisms driving forest ecosystem reorganization under changing nitrogen regimes

1 Introduction

Human activities have fundamentally altered the global nitrogen cycle (Vitousek et al., 1997). Industrialization, fossil fuel combustion, and agricultural intensification have greatly increased emissions of reactive nitrogen (Nr), resulting in unprecedented atmospheric nitrogen deposition across terrestrial ecosystems (Matson et al., 2002; Stevens et al., 2018). Together with climate change and land-use transformation, nitrogen deposition has become one of the major global change drivers influencing ecosystem structure and functioning (Meng et al., 2025). Forests represent critical components of the global carbon cycle and receive substantial nitrogen inputs through wet and dry deposition (Schulte-Uebbing and de Vries, 2018; Lu et al., 2021). Due to their complex canopy structures and high biological activity, forests can efficiently intercept atmospheric nitrogen, resulting in ecosystem nitrogen inputs that may exceed those measured at open-field sites (Lovett and Lindberg, 1993; Avila et al., 2017). Once deposited, reactive nitrogen enters interconnected plant, microbial, and soil pathways that regulate nutrient retention, carbon storage, greenhouse gas emissions, and ecosystem stability (Lu et al., 2021; Li et al., 2026).

Early ecological theory predicted that increasing nitrogen availability would enhance forest productivity because nitrogen frequently limits plant growth in terrestrial ecosystems (Zhang et al., 2024; Li et al., 2025a). Consistent with this expectation, short-term nitrogen enrichment experiments have demonstrated increased photosynthesis, biomass accumulation, and carbon inputs in many nitrogen-limited forests (Liang et al., 2020; Li et al., 2020; Feng et al., 2023). However, increasing evidence from long-term observations and experiments indicates that nitrogen benefits are often transient and strongly context dependent (Zheng et al., 2026). When nitrogen inputs exceed ecosystem demand, forests may undergo a transition from nitrogen limitation release toward nitrogen saturation, characterized by reduced nutrient-use efficiency, altered species composition, soil acidification, increased nitrogen leakage, and changes in carbon cycling (Aber et al., 1989). The contrasting effects of nitrogen deposition raise a fundamental ecological question: why does nitrogen enrichment initially stimulate ecosystem productivity but eventually reorganize ecosystem processes and potentially reduce ecosystem stability?

Previous studies have extensively documented the responses of vegetation, soils, microorganisms, and carbon cycling to nitrogen deposition (Tian et al., 2019; Ma et al., 2022). However, many syntheses have focused on individual ecosystem components, treating plant responses, soil processes, and microbial changes as separate consequences of nitrogen enrichment (Han X. et al., 2025). Such an approach provides valuable information about ecosystem responses but does not fully explain the mechanisms linking nitrogen inputs with ecosystem-level reorganization. Forest ecosystems function through tightly coupled interactions among plants, soils, and microorganisms (Lladó et al., 2017). Nitrogen deposition can modify plant nutrient acquisition strategies, alter carbon allocation belowground, reshape microbial resource-use strategies, and regulate soil organic matter transformation (Tang et al., 2023; Yan et al., 2025). These interconnected feedbacks determine whether nitrogen inputs enhance ecosystem functioning, induce ecological imbalance, or constrain recovery after nitrogen decline. Therefore, understanding nitrogen deposition effects requires a shift from component-based assessments toward a process-based framework emphasizing mechanisms and feedbacks.

In this review, we make three conceptual contributions. First, we distinguish atmospheric N deposition from experimental N addition and frame ecosystem responses in terms of changing ecosystem N states rather than N input magnitude alone. Second, we integrate plant carbon allocation, root–microbial interactions, microbial resource acquisition, extracellular enzyme allocation, microbial CUE, and soil carbon transformation into an evidence-weighted plant–soil–microbial feedback framework. Third, we distinguish ecological legacy, delayed recovery, and ecological hysteresis and identify ecosystem contexts under which declining N deposition may lead to rapid, delayed, partial, or divergent recovery trajectories.

The resulting framework shifts the analytical focus from a simple input–response relationship to a trajectory–state–feedback relationship: changing N regime interacts with initial nutrient status, retention capacity, climate, soil mineralogy, biological traits, and historical N exposure to determine ecosystem state; that state regulates the strength and direction of plant–soil–microbial feedbacks; and those feedbacks influence both the response to enrichment and the capacity for recovery after N decline. This framing is intended as a synthesis and predictive hypothesis rather than a claim that all forests follow a single universal sequence.

2 Literature search and evidence synthesis

This review was conducted as a structured narrative review with an evidence-oriented synthesis strategy. Relevant literature was identified primarily through searches of the Web of Science Core Collection and Google Scholar, supplemented by citation tracking of relevant review articles and primary studies. No publication-year restriction was imposed, allowing the synthesis to incorporate both foundational studies establishing major ecological concepts and recent research addressing changing nitrogen regimes, plant–soil–microbial interactions, soil carbon dynamics, ecological legacies, and ecosystem recovery.

The literature search combined terms related to nitrogen deposition, atmospheric nitrogen inputs, experimental nitrogen addition, forest ecosystems, nitrogen limitation, nitrogen saturation, phosphorus limitation, nutrient stoichiometry, nitrogen-use efficiency, plant carbon allocation, root–microbial interactions, extracellular enzyme allocation, microbial carbon-use efficiency, soil organic carbon, particulate organic carbon, mineral-associated organic carbon, microbial necromass, mineral stabilization, ecological legacy, ecosystem memory, recovery, path dependence, and ecological hysteresis. Search terms were combined in different configurations to capture both direct studies of nitrogen deposition in forests and mechanistic studies addressing the component processes underlying ecosystem responses and recovery.

Studies were prioritized when they (i) focused on forest ecosystems or directly relevant forest biogeochemical processes, (ii) examined atmospheric nitrogen deposition, deposition gradients, long-term nitrogen enrichment, or experimental nitrogen manipulation, and (iii) provided evidence relevant to ecosystem responses, mechanistic pathways, soil carbon transformation, ecological legacies, or recovery following declining nitrogen inputs. Studies from non-forest ecosystems were included selectively when they provided mechanistic evidence relevant to processes considered in forest ecosystems. Such evidence was used to inform mechanisms but was not treated as equivalent to direct evidence from forest ecosystems, and its transferability was considered when interpreting the strength and generality of the proposed pathways.

The literature was organized into four linked domains: (i) atmospheric nitrogen deposition and experimental nitrogen addition; (ii) nutrient limitation, nitrogen retention, and nitrogen saturation; (iii) plant–soil–microbial interactions and soil carbon transformation; and (iv) ecological legacies and ecosystem recovery following declining nitrogen inputs. Studies were classified according to ecosystem process, study design, direction of response, and the directness of evidence connecting changing nitrogen regimes with the proposed mechanism. Citation chaining was further used to identify influential foundational studies, subsequent mechanistic research, and studies specifically addressing recovery after declining nitrogen deposition.

Atmospheric nitrogen deposition and experimental nitrogen addition were treated as complementary but non-equivalent evidence sources. Atmospheric deposition represents a chronic and spatially heterogeneous environmental exposure that is modified by canopy interception, hydrological transport, chemical form, and ecosystem retention. Experimental nitrogen addition provides controlled manipulation that is particularly useful for resolving mechanisms and causal relationships, but its magnitude, timing, chemical composition, and application regime may differ from ambient deposition. Accordingly, nitrogen-addition studies were used primarily to evaluate mechanistic pathways, whereas deposition gradients, long-term observations, and field studies were considered particularly informative for assessing whether these mechanisms operate under realistic exposure regimes.

For synthesis purposes, evidence strength was evaluated qualitatively according to three complementary dimensions: (i) consistency of reported responses across studies, (ii) directness of evidence linking the nitrogen regime to the proposed mechanism, and (iii) ecological relevance to forest ecosystems. Evidence was categorized as high, moderate, or limited, with intermediate categories of moderate–high and limited–moderate used where appropriate. High-confidence links are supported by multiple independent lines of evidence that are directly relevant to forest nitrogen regimes and show broadly consistent responses. Moderate-confidence links are supported by generally consistent evidence but show substantial dependence on ecosystem context, exposure regime, or methodological approach. Limited-confidence links are mechanistically plausible but rely primarily on indirect, sparse, short-term, or non-forest evidence. These categories describe the strength of empirical support rather than statistical certainty.

This evidence-oriented approach was used throughout the review to distinguish relatively well-supported mechanisms from context-dependent pathways and unresolved hypotheses. In particular, microbial CUE-mediated effects on soil carbon persistence and ecological hysteresis are treated as mechanistically important but not universally established causal outcomes. The resulting synthesis therefore emphasizes not only the direction of ecosystem responses to changing nitrogen regimes, but also the conditions under which those responses are likely to differ among forests and across enrichment and recovery trajectories.

3 Nitrogen deposition reorganizes ecosystem resource limitation

3.1 From atmospheric nitrogen inputs to ecosystem nitrogen availability

Atmospheric nitrogen deposition represents an external subsidy to forest ecosystems, but the ecological significance of this subsidy depends on how deposited nitrogen is retained, transformed, and redistributed within the ecosystem (Gilliam et al., 2019). Reactive nitrogen enters forests through wet and dry deposition and can subsequently be intercepted by vegetation canopies, absorbed by foliage, incorporated into plant and microbial biomass, transformed through soil microbial processes, or exported through gaseous and hydrological pathways (Cui et al., 2024; Guerrieri et al., 2024). Thus, atmospheric deposition should not be interpreted as equivalent to ecosystem nitrogen availability. Instead, the effective nitrogen supply experienced by organisms emerges from interactions among deposition inputs, canopy processing, biological demand, soil retention capacity, and environmental conditions (Lin et al., 2024).

Atmospheric deposition and experimental N addition provide complementary but non-equivalent evidence. Deposition is typically chronic, spatially heterogeneous, and filtered through canopy interception, hydrological transport, and ecosystem retention. Experimental N addition is imposed by researchers and can involve pulses, concentrations, chemical forms, and rates that differ from ambient deposition. We therefore use N-addition studies primarily to resolve mechanisms and use deposition gradients, long-term monitoring, and field observations to evaluate whether those mechanisms operate under realistic exposure regimes.

In nitrogen-limited forests, additional nitrogen is initially retained efficiently because biological demand exceeds external supply (Lin et al., 2024). Deposited nitrogen can therefore be incorporated into plant biomass and microbial pools, increasing nitrogen availability for photosynthesis and growth. This mechanism is particularly important in boreal and temperate forests where relatively slow decomposition can constrain internal nitrogen supply (Zhao et al., 2024). However, as nitrogen inputs persist, the balance between nitrogen supply and ecosystem demand progressively changes (Elrys et al., 2023). Once biological uptake and retention capacity become saturated, a larger fraction of deposited nitrogen remains in inorganic pools and becomes susceptible to leaching and gaseous losses (Du and de Vries, 2025).

This transition suggests that the ecological effects of nitrogen deposition are better understood as a shift in ecosystem nitrogen state rather than as a simple dose–response relationship. The same nitrogen input may alleviate nutrient limitation in one forest but contribute to nutrient imbalance in another, depending on historical nitrogen exposure, initial nutrient status, and the capacity of plants and microorganisms to retain nitrogen (Xu et al., 2021). The transition toward nitrogen saturation is therefore an emergent ecosystem property arising from interactions between external nitrogen supply and internal biological regulation.

3.2 Transition from nitrogen limitation to nitrogen saturation

The initial response to nitrogen enrichment can be interpreted as a release from nitrogen limitation. Additional nitrogen may increase plant nitrogen acquisition, photosynthetic capacity, and biomass production, thereby increasing ecosystem carbon inputs (Feng et al., 2023; Stocker et al., 2025). However, this positive response is not necessarily sustained. As enrichment continues, biological demand may become increasingly decoupled from external supply, and a larger fraction of nitrogen may accumulate in inorganic pools or enter loss pathways.

Nitrogen saturation should therefore be understood as a condition in which ecosystem demand and retention capacity become insufficient relative to nitrogen supply, rather than as a universal deposition threshold. This distinction helps explain why forests can show an initial productivity response followed by increasing nitrogen losses or nutrient imbalance. The transition depends on nitrogen supply, biological demand, retention capacity, and the availability of other limiting nutrients (Lovett and Goodale, 2011; Niu et al., 2023).

3.3 Stoichiometric reorganization and secondary nutrient limitation

Secondary nutrient limitation addresses a different question: which resource becomes most limiting after nitrogen limitation is alleviated? Sustained enrichment can alter C:N:P balance, and in some ecosystems, shift the dominant constraint toward phosphorus or other nutrients. These responses are not universal and should be diagnosed using complementary evidence from nutrient availability, uptake, resorption, enzyme allocation, and nutrient-addition responses rather than stoichiometric ratios alone (Treseder and Vitousek, 2001; Wang et al., 2022b).

3.4 Nitrogen-use efficiency and trajectory-dependent ecosystem state transitions

Nitrogen-use efficiency provides a useful indicator of how effectively forests convert available nitrogen into biological production (Congreves et al., 2021). Under strong N limitation, additional nitrogen is often efficiently incorporated into plant and microbial pools, whereas supply beyond biological demand can reduce the fraction supporting biomass production and increase loss pathways (Band et al., 2022). A decline in nitrogen-use efficiency therefore signals weakening biological retention and increasing biogeochemical processing, while its magnitude remains context dependent.

The mechanisms described above indicate that forest responses to changing nitrogen regimes cannot be understood from nitrogen input alone. Instead, external N forcing interacts with ecosystem nutrient state, retention capacity, biological demand, and environmental context to alter the strength and direction of plant–soil–microbial regulation. These interactions can lead to different ecosystem states and response pathways, including the alleviation of N limitation, secondary nutrient limitation, increasing N saturation and loss, and subsequent changes in carbon and nutrient cycling. Figure 1 summarizes this trajectory–state–feedback framework and distinguishes relatively well-supported mechanisms from context-dependent or more uncertain pathways.

4 Plant–soil–microbial feedbacks mediate ecosystem reorganization

Nitrogen deposition does not affect plants, soils, and microorganisms independently (Frey et al., 2004; Yuan et al., 2020). Instead, changes in nitrogen availability modify plant carbon allocation, microbial resource acquisition, and soil organic matter transformation, generating feedbacks that regulate subsequent nutrient availability (Luo et al., 2023; Tang et al., 2023). These coupled processes provide a mechanistic link between atmospheric nitrogen inputs and ecosystem-level changes in carbon and nutrient cycling (Stocker et al., 2025). Conceptually, nitrogen deposition can propagate through a linked sequence of plant nutrient status, carbon allocation, microbial resource acquisition, soil organic matter transformation, and subsequent changes in nutrient availability, which in turn feedback to plant responses. The strength and direction of these feedbacks depend on the balance between nitrogen supply and the availability of other resources. Thus, nitrogen enrichment can reorganize ecosystem processes even when responses of individual ecosystem components appear relatively small. The major mechanisms, mediators, feedbacks, and ecosystem consequences underlying this process are synthesized in Table 1.

Table 1. Evidence-weighted synthesis of mechanisms linking changing nitrogen regimes to plant–soil–microbial processes and ecosystem carbon dynamics.

4.1 Plant functional strategies and carbon allocation

These mechanisms form a coupled regulatory network. Plant carbon allocation links nitrogen availability to belowground carbon inputs, while microbial metabolism and soil biogeochemical transformations regulate the subsequent fate of carbon and nutrients. The following sections focus on the principal feedback links rather than treating plant, microbial, and soil responses as independent effects.

Nitrogen availability strongly influences how plants allocate assimilated carbon among leaves, stems, roots, and symbiotic partners (Li et al., 2020). Under nitrogen limitation, plants commonly maintain substantial carbon investment belowground to acquire nitrogen through fine roots, root exudates, and mycorrhizal associations (Braghiere et al., 2022). Nitrogen enrichment can reduce the marginal benefit of these acquisition pathways, allowing plants to reallocate carbon toward aboveground growth and photosynthetic tissues (Li et al., 2020; Eastman et al., 2021).

This mechanism creates an important connection between nitrogen deposition and soil carbon cycling. Reduced root production, root turnover, and rhizosphere carbon inputs can decrease the quantity of labile carbon supplied to soil microorganisms (Lv et al., 2023). Consequently, an increase in aboveground carbon assimilation does not necessarily translate into a proportional increase in soil carbon inputs. The effects of nitrogen deposition on carbon allocation therefore depend on how plants balance carbon investment between resource acquisition and biomass production.

However, this response may change when nitrogen enrichment induces secondary nutrient limitation (Liang et al., 2026). When phosphorus or other nutrients become limiting, plants may maintain or increase belowground carbon investment to sustain nutrient acquisition (Wang et al., 2022a). This context dependence indicates that plant carbon allocation is regulated by the relative limitation of multiple nutrients, rather than by nitrogen availability alone (Han Y. et al., 2025). This distinction is critical for explaining why forests exposed to similar nitrogen deposition can exhibit contrasting belowground carbon responses.

4.2 Root–microbial interactions regulate belowground carbon inputs

The rhizosphere provides a major interface through which plant responses to nitrogen deposition propagate into soil processes (Henneron et al., 2020). Changes in root biomass and carbon allocation alter the quantity and composition of root-derived substrates available to microorganisms (Feng et al., 2022). These changes can modify microbial growth, enzyme production, nutrient mineralization, and interactions with soil organic matter (Wen et al., 2022).

The rhizosphere therefore functions as a regulatory interface: plant carbon allocation modifies microbial resource availability, while microbial transformations feed back to plant nutrient supply (Feng et al., 2022). The direction of this feedback depends on whether N remains limiting or another nutrient, particularly P, becomes limiting.

4.3 Microbial CUE mediates carbon processing but does not by itself determine SOC stabilization

Microorganisms represent a central biological interface between nitrogen availability and soil carbon persistence (Mou et al., 2023). Nitrogen deposition can modify microbial metabolism through simultaneous changes in nitrogen availability, carbon substrate supply, soil chemistry, and plant–microbe interactions (Zhang et al., 2018; Ma et al., 2022).

Microbial carbon-use efficiency (CUE), defined as the proportion of assimilated carbon allocated to microbial growth rather than respiration, can mediate the fate of plant-derived carbon. However, CUE should not be treated as a sufficient or universal determinant of soil carbon stabilization. Carbon retained in microbial biomass can be rapidly respired or transformed, whereas persistent SOC formation additionally depends on microbial necromass production, chemical accessibility, aggregation, and association with mineral surfaces. Thus, N effects on CUE are best interpreted as one component of a broader carbon-processing pathway rather than as direct evidence of increased SOC persistence.

Accordingly, ecosystem carbon responses should be evaluated as the balance between carbon input, microbial assimilation and growth, and subsequent stabilization versus respiration. The CUE response is therefore an intermediate process, not direct evidence of persistent SOC storage.

4.4 Extracellular enzyme allocation reflects changing microbial resource constraints

Microbial extracellular enzymes regulate the accessibility of carbon and nutrients contained within organic matter (Abay et al., 2024). Nitrogen enrichment can alter microbial investment in these enzymes by changing the relative demand for nitrogen, carbon, and phosphorus (Li et al., 2025b). In particular, nitrogen enrichment may reduce investment in nitrogen-acquiring enzymes while modifying enzymes involved in carbon degradation.

One consequence is a potential reduction in the decomposition of chemically complex organic matter when nitrogen enrichment suppresses ligninolytic enzyme production (Tan et al., 2020). Reduced fungal activity and altered microbial functional groups may further constrain the decomposition of lignin-rich substrates (Wilhelm et al., 2019).

Overall, nitrogen effects on decomposition reflect microbial resource-economic trade-offs rather than a universal increase or suppression. The response depends on which resource limits microbial growth and which enzymes provide the greatest return on microbial investment.

4.5 Microbial functional restructuring alters nitrogen transformation pathways

Changes in microbial communities are important not simply because community composition changes, but because functional traits and metabolic pathways determine the fate of deposited nitrogen (Sun et al., 2020). Key microbial processes include nitrogen fixation, nitrification, denitrification, and nitrate reduction (Xiang et al., 2020).

Under relatively low nitrogen availability, microbial immobilization can contribute to ecosystem nitrogen retention (Nadelhoffer et al., 1999; Gu et al., 2025). With chronic enrichment, however, increasing nitrogen availability may shift microbial metabolism toward pathways associated with nitrogen transformation and loss. Enhanced nitrification can increase nitrate availability, while subsequent denitrification and nitrate reduction can contribute to gaseous nitrogen losses and N₂O emissions (Aber et al., 1989; Magill et al., 2004).

Across N availability gradients, microbial processing can shift from relatively strong immobilization and retention toward greater inorganic N transformation and loss. Enhanced nitrification can increase nitrate availability, while denitrification and related pathways contribute to gaseous losses and N₂O emissions (Nadelhoffer et al., 1999; Magill et al., 2004).

This transition helps explain why chronic N deposition can weaken ecosystem N retention even when total ecosystem N stocks increase. The balance between biological demand, microbial retention, and transformation pathways is therefore central to interpreting N losses.

4.6 Integrated feedbacks link component responses to ecosystem reorganization

The interactions described above form a dynamic feedback network linking atmospheric N deposition, plant nutrient status and carbon allocation, rhizosphere inputs, microbial resource use, enzyme allocation and CUE, soil C–N transformation, and plant-available nutrients. Feedback strength depends on ecosystem state and can reinforce either nutrient retention and productivity or, under prolonged enrichment, nutrient imbalance and N loss.

5 Nitrogen deposition reshapes carbon–nitrogen coupling and ecosystem functioning

5.1 Plant productivity and carbon acquisition

Nitrogen is a major component of chlorophyll, photosynthetic enzymes, and other cellular machinery involved in carbon assimilation (Liu et al., 2025). Consequently, nitrogen enrichment can directly increase leaf nitrogen concentration and photosynthetic capacity in nitrogen-limited forests. Increased photosynthesis can subsequently increase aboveground biomass production and carbon inputs to ecosystems (Li et al., 2020).

This mechanism represents the most direct pathway through which nitrogen deposition can initially enhance forest carbon uptake: N deposition → plant N availability → photosynthetic capacity → carbon assimilation → biomass production.

However, this pathway is not a persistent fertilization effect. Productivity stimulation weakens as nitrogen ceases to be the dominant limiting resource and other constraints become more important (Ren et al., 2017; Luo et al., 2022). Thus, greater plant carbon assimilation should be interpreted in relation to the nutrient state of the ecosystem rather than deposition magnitude alone.

5.2 Carbon inputs, microbial processing, and carbon sequestration

Nitrogen deposition can enhance ecosystem carbon inputs by stimulating plant productivity, litter production, and root-derived carbon inputs, particularly in nitrogen-limited forests (Lu et al., 2021). However, greater carbon input does not necessarily result in greater long-term soil carbon storage.

The fate of newly fixed carbon depends on carbon input, microbial processing, and subsequent stabilization or loss. Greater plant production can therefore increase carbon supply without necessarily increasing persistent SOC storage.

5.3 Nitrogen-mediated changes in decomposition and carbon pathways

Nitrogen enrichment can modify decomposition through changes in microbial nutrient acquisition, extracellular enzyme production, fungal activity, and carbon availability (Ping et al., 2025). Moderate nitrogen enrichment may suppress decomposition by reducing microbial investment in nitrogen acquisition and ligninolytic enzymes, potentially slowing the breakdown of complex organic matter (Wu et al., 2023).

At the same time, excessive nitrogen inputs may alter microbial metabolism and increase respiratory carbon losses (Kangi et al., 2025). Consequently, nitrogen deposition does not exert a unidirectional effect on decomposition. Instead, its effects depend on the balance between:

Together, these observations indicate that nitrogen can simultaneously increase carbon inputs and generate divergent effects on decomposition and carbon persistence. The outcome depends on the balance among N availability, substrate supply, microbial nutrient demand, enzyme allocation, and CUE.

5.4 From microbial processing to POC, MAOC, necromass, and mineral stabilization

The persistence of soil carbon depends on the interaction between plant inputs, microbial transformation, physical protection, and mineral association. A useful distinction is between particulate organic carbon (POC), which is generally more closely linked to relatively recognizable plant-derived particulate material and can be comparatively dynamic, and mineral-associated organic carbon (MAOC), which is often stabilized through interactions with mineral surfaces and associated compounds. Microbial necromass can contribute substantially to MAOC formation, but the magnitude of this pathway depends on mineralogy, organo-mineral interactions, aggregation, substrate quality, and environmental conditions. Nitrogen deposition may therefore alter SOC through multiple pathways, and increases in CUE, microbial biomass, or total SOC should not be interpreted in isolation as evidence of long-term stabilization.

An increase in litter or root inputs represents greater carbon supply, whereas long-term storage requires transformation into forms that are resistant to decomposition and effectively protected through mineral association or aggregation (Button et al., 2022; Georgiou et al., 2022). Consequently, total SOC or CUE alone cannot establish long-term stabilization.

Accordingly, the effect of nitrogen deposition on soil carbon storage depends on whether increased carbon inputs outweigh changes in microbial decomposition and stabilization efficiency.

5.5 Ecosystem state transitions and functional consequences

At the ecosystem scale, carbon responses are coupled to changes in nitrogen regulation (Niu et al., 2023). Nitrogen limitation release can increase productivity and carbon inputs, whereas chronic enrichment may alter C: N: P balance, nitrogen-use efficiency, microbial processing, and nitrogen losses (Xu et al., 2022). The key consequence is therefore reorganization of carbon uptake, transformation, stabilization, and nitrogen retention rather than a uniform increase or decrease in carbon storage (Mozdzer et al., 2021).

6 Ecosystem context determines the direction and magnitude of responses

The magnitude and direction of ecosystem responses to nitrogen deposition cannot be explained by nitrogen input alone (Borer and Stevens, 2022). Instead, responses depend on the ecological context in which nitrogen is added, including the initial nutrient limitation, soil properties and nitrogen retention capacity, climatic conditions, and the history of nitrogen exposure (Schmitz et al., 2024). These factors determine whether additional nitrogen alleviates an existing resource constraint, induces secondary nutrient limitation, or exceeds ecosystem demand and retention capacity. Consequently, forests receiving similar atmospheric nitrogen inputs may exhibit contrasting responses in productivity, nutrient cycling, nitrogen retention, and carbon dynamics.

6.1 Initial nutrient limitation determines response direction

The initial nutrient state of a forest is a primary determinant of its response to nitrogen deposition (Schmitz et al., 2024). Nitrogen-limited ecosystems are more likely to exhibit positive productivity responses because additional nitrogen directly alleviates a major resource constraint (Zheng et al., 2026). In contrast, ecosystems with relatively high nitrogen availability may show weaker productivity responses because nitrogen is no longer the dominant limiting resource. Under these conditions, continued nitrogen inputs may instead intensify nutrient imbalance, promote secondary phosphorus limitation, or increase nitrogen losses (Mason et al., 2022).

6.2 Soil properties regulate nitrogen retention and ecosystem sensitivity

The response to nitrogen deposition is further regulated by the capacity of soils and vegetation to retain and process incoming nitrogen (Bai et al., 2025). Soil organic matter, mineral composition, pH, moisture, and microbial activity influence whether deposited nitrogen is immobilized, transformed, retained, or exported, thereby determining the amount of atmospheric nitrogen that becomes available to plants and microorganisms and the fraction transferred among ecosystem compartments or lost through leaching and gaseous pathways (Li et al., 2021). Consequently, soil properties regulate both the exposure of plants and microorganisms to deposited nitrogen and the balance between biological retention and nitrogen loss.

6.3 Climate and ecosystem history modify response trajectories

Climate provides an additional context that regulates the response of forest ecosystems to nitrogen deposition. Temperature, precipitation, and growing-season conditions influence decomposition, microbial activity, plant nutrient demand, and nitrogen transformation, thereby modifying the sensitivity of ecosystem processes to additional nitrogen inputs (Vallicrosa et al., 2024). These climatic controls can alter both nitrogen demand and the rates at which nitrogen is immobilized, mineralized, transformed, or lost. Ecosystem history is equally important because previous nitrogen exposure can modify soil nutrient pools, microbial communities, vegetation composition, and the strength of plant–soil–microbial feedbacks (Zhang et al., 2025). Consequently, current ecosystem states and response trajectories reflect both present climatic and nitrogen forcing and the accumulated effects of historical nitrogen exposure.

6.4 Context dependence integrates ecosystem sensitivity

Context dependence emerges from interactions among initial nutrient limitation, soil retention and mineralogy, climate, hydrology, mycorrhizal association, forest functional type, and historical N exposure. These controls can jointly alter the sign, magnitude, and persistence of N responses, so similar atmospheric N inputs can produce contrasting outcomes among forests.

6.5 Forest functional type, mycorrhizal association, soil mineralogy, climate, and historical N exposure

Forest functional type, mycorrhizal association, soil mineralogy, climate, and historical N exposure interact to determine baseline nutrient limitation, nitrogen retention, nutrient-acquisition strategies, microbial activity, and the persistence of ecosystem responses. These moderators can therefore alter both the direction and magnitude of responses to changing nitrogen regimes, contributing to contrasting outcomes among forests exposed to similar atmospheric N inputs. The major ecosystem contexts, mediating processes, expected responses, and associated uncertainties are summarized in Table 2.

Table 2. Context-dependent moderators of forest responses to changing nitrogen regimes and ecosystem recovery.

7 Ecological legacies shape recovery trajectories after nitrogen decline

The ecological consequences of nitrogen deposition may persist after atmospheric inputs decline because forests retain legacies in soil nutrient pools, microbial communities, vegetation composition, and biogeochemical feedbacks (Larsson et al., 2024; Schmitz et al., 2024). Recovery can therefore be slower than the development of enrichment responses, making enrichment and recovery distinct trajectories rather than reversible stages of a simple dose–response relationship.

A useful distinction is that enrichment changes ecosystem state, whereas declining deposition tests whether the reorganized ecosystem can return toward its previous configuration. Recovery rate and completeness depend on the persistence of legacies and the strength of internal feedbacks.

7.1 Soil nutrient legacies constrain recovery

Soils can retain nitrogen deposited over decades through accumulation in organic matter, microbial biomass, and mineral-associated pools (Li et al., 2021). Even after atmospheric nitrogen inputs decline, these internal nitrogen pools can continue to supply mineral nitrogen to plants and microorganisms (Johnson, 1992). The resulting decoupling between declining external nitrogen inputs and persistent internal nitrogen availability can delay the recovery of nitrogen limitation.

Historical N enrichment can also leave longer-lasting changes in soil acidity and nutrient stoichiometry (Yan et al., 2020). These changes may alter base-cation availability, phosphorus chemistry, microbial activity, and root nutrient acquisition, potentially delaying recovery after deposition declines. A reduction in external N input therefore need not immediately restore pre-enrichment nutrient cycling.

Soils can thus function as long-term reservoirs of the consequences of historical N exposure, creating a temporal separation between declining atmospheric deposition and recovery of ecosystem nutrient regulation.

7.2 Microbial ecological memory

Microbial communities are another important source of ecological memory (Canarini et al., 2021). Chronic nitrogen enrichment can alter microbial community composition, functional groups, enzyme allocation, and carbon-use strategies (Wang et al., 2023). These changes may persist after nitrogen inputs decline because microbial communities respond not only to current nutrient availability but also to the environmental conditions established during prolonged enrichment.

Long-term N enrichment may reduce microbial investment in N acquisition while changing carbon-processing strategies and extracellular enzyme allocation. When N availability declines, microbial community composition and functional traits may recover slowly or only partially (Chen et al., 2024).

Microbial legacies can influence decomposition, nutrient mineralization, and N transformation after the external driver weakens. Such persistence provides a mechanism for path dependence and delayed recovery; ecological hysteresis should be considered only when enrichment and recovery trajectories differ systematically at comparable levels of the external driver.

7.3 Vegetation legacies alter the direction and rate of recovery

Vegetation responses can also generate persistent ecological legacies (Islam et al., 2026). Chronic nitrogen enrichment may favor species with high nutrient demand or alter competitive relationships among plant species, potentially changing community composition and functional traits (Liao et al., 2024). Once species composition has shifted, subsequent recovery may be constrained by dispersal limitation, regeneration dynamics, altered soil conditions, and continued competition.

Even if soil N availability returns toward pre-enrichment conditions, vegetation may not immediately return to its former composition. Recovery can be constrained by dispersal, regeneration, altered soil conditions, and competition, while vegetation change can feed back to litter quality, root inputs, and rhizosphere processes.

Vegetation recovery therefore need not retrace enrichment-related change, but such non-retracing trajectories should be interpreted as recovery asymmetry or path dependence unless direct evidence demonstrates hysteresis.

7.4 Interacting legacies reinforce altered ecosystem states

Soil, microbial, and vegetation legacies can interact through coupled feedbacks and reinforce an altered ecosystem state established during chronic N enrichment (Li K. et al., 2025). Historical N enrichment can alter soil chemistry and N pools, which affect microbial processes and plant nutrient availability; vegetation changes then modify litter and root inputs, feeding back to soil processes (Manning et al., 2008). The strength of these interactions determines how strongly internal processes continue to maintain the altered state after N deposition declines.

These interacting legacies can prolong ecosystem reorganization after the external driver weakens (Manninen et al., 2024). Recovery therefore depends on the strength, persistence, and interaction of soil, microbial, and vegetation legacies rather than on the magnitude of N reduction alone.

Recovery is therefore determined not simply by the magnitude of nitrogen reduction, but by the strength, persistence, and interaction of ecological legacies across soil, microbial, and vegetation compartments.

7.5 Enrichment and recovery follow asymmetric trajectories

The interaction among ecological legacies helps explain why enrichment and recovery are unlikely to be mirror processes. During enrichment, increasing N inputs can reorganize nutrient availability and plant–soil–microbial processes; during recovery, declining external N may be followed by persistent internal N pools, altered microbial functioning, and vegetation legacies (Xu et al., 2022). The resulting recovery trajectory therefore depends on the persistence and interaction of these internal changes and may range from relatively rapid recovery to delayed, partial, or divergent recovery. Figure 2 conceptualizes these alternative trajectories and emphasizes that recovery asymmetry can arise from ecological legacies without necessarily constituting ecological hysteresis.

7.6 When can delayed recovery indicate ecological hysteresis?

Ecological legacy, delayed recovery, and hysteresis describe progressively stronger forms of trajectory dependence. A legacy is a persistent effect of past N exposure; delayed recovery is a temporal mismatch between driver decline and ecosystem response; hysteresis requires directional path dependence, demonstrated by systematically different ecosystem states or functions at comparable external N levels during enrichment and recovery (Gilliam et al., 2019). The current synthesis supports the first two concepts more broadly than the third, so hysteresis is treated as a testable hypothesis rather than a universal outcome of N-deposition decline.

Not all forests will exhibit the same degree of ecological memory (Webster et al., 2018). The trajectory following declining nitrogen deposition is likely to depend on the magnitude and duration of historical nitrogen enrichment, the initial nutrient limitation of the ecosystem, soil buffering capacity, the strength of plant–microbial feedbacks, the degree of vegetation compositional change, and the persistence of altered soil nutrient pools.

These factors can produce several broad recovery trajectories:

Rapid recovery——Historical changes are relatively weak, internal nutrient pools are rapidly depleted or transformed, and ecosystem feedbacks remain sufficiently reversible to allow a relatively rapid return toward the pre-enrichment state.

Delayed recovery——Persistent soil or microbial legacies maintain altered nutrient cycling and slow the restoration of ecosystem processes.

Partial recovery——Some ecosystem functions recover more rapidly than others, resulting in a mixture of restored and persistent ecosystem properties.

Alternative trajectory——Persistent feedbacks, soil changes, and vegetation reorganization prevent convergence toward the pre-enrichment state and instead stabilize the ecosystem along a different trajectory.

Ecosystem resilience following declining nitrogen deposition should therefore be evaluated by both the magnitude and completeness of recovery and the persistence of internal legacy constraints. Evidence for true hysteresis requires direct comparison of enrichment and recovery trajectories rather than inference from delayed or incomplete recovery alone.

8 Research gaps and future directions

8.1 From deposition magnitude to ecosystem-state prediction

Predicting forest responses to nitrogen deposition requires moving beyond deposition rate alone. The trajectory–state–feedback framework can be represented as: N deposition trajectory → ecosystem nutrient state → plant–soil–microbial feedbacks → ecosystem functional trajectory. Future prediction should therefore incorporate current forcing together with the internal state from which the ecosystem responds.

8.2 Process-based thresholds and early-warning indicators

Process-based indicators may help identify transitions among regulatory states, including the alleviation of N limitation, emergence of secondary nutrient limitation, declining nitrogen-use efficiency, increasing nitrate mobility, shifts in microbial enzyme allocation, and weakening N retention. Because thresholds depend on ecosystem demand and retention capacity, they should not be treated as universal deposition values.

8.3 Integrating plant–soil–microbial feedbacks into ecosystem models

Many ecosystem models still represent plant productivity, soil carbon cycling, and nitrogen cycling as partially independent processes. Future models should explicitly couple plant nutrient acquisition, root carbon allocation, microbial metabolism, soil C–N transformation, and nutrient availability so that internal feedbacks can generate nonlinear responses and delayed recovery within a common framework.

8.4 Incorporating ecological memory into recovery prediction

Recovery prediction should incorporate ecological memory rather than current environmental conditions alone. A useful formulation is historical N exposure + current N deposition + ecosystem state + feedback strength, particularly in regions where atmospheric deposition has declined but soil, microbial, or vegetation legacies may persist.

8.5 Multifactor global change and trajectory-based prediction

Nitrogen deposition interacts with warming, drought, elevated CO₂, and land-use change, which can modify nutrient demand, microbial activity, decomposition, and water availability (Borer and Stevens, 2022). Future assessments should test how these interacting drivers modify the same trajectory–state–feedback relationships and thereby alter recovery trajectories.

9 Discussion

9.1 A unified framework for forest responses to changing nitrogen deposition

The evidence synthesized here supports a shift from an input–response perspective toward a trajectory–state–feedback framework. Atmospheric nitrogen deposition modifies the external resource environment, but its ecological effects depend on initial nutrient limitation, nitrogen retention capacity, biological demand, ecosystem traits, and historical exposure (Stevens et al., 2018). These conditions regulate the strength and direction of plant–soil–microbial feedbacks and thereby determine whether continued N inputs primarily stimulate productivity and retention or promote nutrient imbalance and N loss. The framework therefore emphasizes ecosystem state and internal regulation rather than deposition magnitude alone.

This perspective also helps reconcile contrasting responses reported among forests. Similar atmospheric N inputs can produce different outcomes when ecosystems differ in nutrient limitation, soil properties, climate, or historical exposure, while a single ecosystem can shift from initial stimulation toward nutrient imbalance as internal resource constraints change (Xu et al., 2022). The framework is therefore intended as a synthesis and predictive hypothesis rather than a universal sequence followed by all forests.

9.2 Why nitrogen enrichment and recovery can be asymmetric

Nitrogen enrichment and ecosystem recovery should not be treated as opposite directions along a reversible dose–response curve. Enrichment can reorganize nutrient availability, plant allocation, microbial functioning, and soil processes, whereas declining deposition removes the external pressure without necessarily reversing the internal changes at the same rate (Li K. et al., 2025). Persistent soil N pools, microbial functional changes, vegetation reorganization, and altered soil chemistry can therefore generate delayed, partial, or alternative recovery trajectories.

These effects provide a basis for path dependence, but they do not by themselves demonstrate ecological hysteresis. Ecological hysteresis should be reserved for cases in which ecosystem functioning or state differs systematically at comparable external N levels between enrichment and recovery trajectories. Thus, the broader evidence supports ecological legacy and recovery asymmetry more strongly than it supports hysteresis as a general outcome of declining N deposition.

9.3 Implications for ecosystem resilience and management

The recovery capacity of forests cannot be evaluated from declining atmospheric N deposition alone (Stevens, 2016; Gilliam et al., 2019). Monitoring should combine external N trajectories with indicators of internal ecosystem state, including nitrogen-use efficiency, N retention and loss, soil nutrient pools and chemistry, microbial functional attributes, and vegetation composition. Such integrated indicators can distinguish relatively rapid recovery from delayed or incomplete recovery and help identify ecosystems in which persistent legacies may constrain restoration.

From a management perspective, reducing chronic N inputs remains fundamental, but recovery may also depend on ecosystem-specific constraints such as nutrient imbalance, soil buffering capacity, vegetation reorganization, and persistent belowground feedbacks. Management and modelling approaches should therefore account for both the magnitude of N forcing and the internal state from which ecosystems respond.

9.4 Conclusion and perspectives

Atmospheric nitrogen deposition is more than an external nutrient subsidy. Its ecological consequences emerge from interconnected changes in nutrient availability, plant carbon allocation, microbial metabolism, soil organic matter transformation, and nitrogen retention. Initial nitrogen enrichment may alleviate nutrient limitation and stimulate ecosystem productivity, but continued inputs can alter C:N balance, resource-acquisition strategies, microbial processing, and nitrogen loss. The direction and magnitude of these responses depend on ecosystem state and context, including initial nutrient limitation, soil properties, climate, mycorrhizal association, forest functional type, and historical nitrogen exposure.

Importantly, declining atmospheric nitrogen deposition does not necessarily reverse enrichment-induced changes along the same trajectory. Persistent soil, microbial, and vegetation legacies can maintain altered ecosystem states and produce delayed, partial, or divergent recovery. These responses indicate that recovery should be evaluated as a history-dependent process rather than inferred from changes in external N deposition alone. The trajectory–state–feedback framework proposed here integrates these processes by linking the history and magnitude of N forcing with ecosystem state, internal feedback strength, and recovery trajectory.

This framework also highlights an important boundary in current understanding. Ecological legacy and delayed recovery are broadly relevant to interpreting ecosystem responses after declining N deposition, whereas direct evidence for ecological hysteresis remains more limited. Hysteresis should therefore be treated as a testable outcome requiring systematic differences between enrichment and recovery trajectories at comparable external N levels. Future long-term experiments, coordinated monitoring, and process-based ecosystem models should explicitly track these trajectories to determine when forest ecosystems recover, reorganize, or persist in alternative states under changing nitrogen regimes.

9.4.1 Evidence grading and uncertainty statement

Evidence grades in this review are qualitative and are intended to communicate the consistency, directness, and ecological relevance of supporting evidence rather than statistical certainty. High confidence indicates multiple independent lines of evidence that are directly relevant to forest nitrogen regimes and show broadly consistent responses. Moderate confidence indicates generally consistent evidence but substantial dependence on ecosystem context, exposure regime, or methodological approach. Limited confidence indicates that evidence remains indirect, sparse, or primarily mechanistic. Where evidence is generally strong but shows important contextual dependence, we use an intermediate category of moderate–high confidence; where a mechanism is plausible and supported by some evidence but direct forest evidence remains limited, we use limited–moderate confidence.

These grades are used to distinguish relatively well-supported mechanisms from context-dependent pathways and unresolved hypotheses in the framework and tables. In particular, microbial CUE-mediated effects on SOC persistence and ecological hysteresis are treated as mechanistically important but not universally established causal outcomes of changing nitrogen regimes.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

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: Charles Driscoll, Syracuse University, United States

Reviewed by: Wenyuan He, South China Agricultural University, China

Xingyun Huang, Ministry of Agriculture and Rural Affairs, China