The research locale was situated in Hubei Province, China, with geographical coordinates at 30°46′N, 110°56′E. It experiences a subtropical continental monsoon climate, characterized by an average annual temperature ranging from 14 to 22°C and an average annual precipitation of 1400 mm. The Quercus variabilis (Q. variabilis) plantation within the study area was established in the 1980s and encompasses an expanse of 11.37 hectares. The trees within the sample plot are uniformly distributed, exhibiting an average slope of 20° and an elevation spanning from 800 to 850 meters. The soil in this region is classified as Luvisols and is characterized by a loamy sandy texture at a depth of approximately 40 cm. The vegetation within the study area is predominantly composed of natural vegetation, with an understory accompanied by shrubs such as Camellia sinensis, Eurya nitida, Viburnum erosum, and others, as well as herbaceous plants like Dryopteris fuscipes, Houttuynia cordata, and Senecio scandens. In this region, the annual background nitrogen (N) deposition rate is approximately 30 kg N ha^-1 yr^-1 (35). Dry deposition, dominated by gaseous N pollutants, represents the dominant pathway and accounts for ~62% of total N deposition. Within this fraction, NH_X-N (comprising NH₃ and NH₄⁺) constitutes more than two-thirds of the total N. In contrast, wet N deposition derives primarily from particulate ammonium salts and nitrates (35, 36).

In total, three replicate plots (20 × 20 m) were established in August 2018. These replicate plots were spaced 20-30 meters apart in a randomized block design. Each replicate plot contained four 3 m × 3 m sample quadrats, totaling 12 across all replicate plots for the N addition treatments, with a 10 m buffer zone between each sample quadrat. Four rates of nitrogen addition were applied to four sample quadrats in each replicate plot: 0 kg N·ha^-1 yr^-1 (N0), 30 kg N·ha^-1 yr^-1 (N30), 60 kg N·ha^-1 yr^-1 (N60), and 90 kg N·ha^-1 yr^-1 (N90). These N addition treatments were designed based on the local atmospheric nitrogen deposition background rate of 30 kg N·ha^-1·yr^-1 (35). The experiment was initiated in August 2018, with NH₄NO₃ solution applied annually at multiples (1×, 2×, 3×, and 4×) of this baseline, corresponding to 0, 30, 60, and 90 kg N·ha^-1·yr^-1, respectively. The annual N input for each treatment was divided into four equal aliquots and applied monthly at the beginning of each quarter. Specifically, NH₄NO₃ was dissolved in 2 L of ultrapure water and uniformly sprayed onto the forest floor within each quadrat (25, 37). In 2018, the respective baseline soil organic carbon (SOC), total nitrogen (N), available phosphorus (P), available potassium (K), pH values, and cation exchange capacity (CEC) at a soil depth of 0−20 cm were 25.83 g kg^-1, 1.29 g kg^-1, 11.08 mg kg^-1, 65.60 mg kg^-1, 7.25, and 27.73 cmol kg^-1. Throughout the experimental period, no logging, understory removal, or any other management practices were conducted (37).

In August 2021, the litter was meticulously removed from the soil surface of each sample quadrat, and then PVC tubes (diameter: 5 cm) were used to extract soil samples from a depth of 0−20 cm at 20−40 locations within each sampling quadrat. The PVC tubes were subsequently sealed with plastic wrap and transported back to the laboratory, where they were stored in a refrigerator at a temperature of 4°C.

Four fractions of aggregate-sized materials were then meticulously separated from the soil samples. Initially, the soil, carefully removed from the PVC tubes to preserve the aggregates, was separated using a series of sieves. The soil was gently broken along natural crevices to ensure that all samples could pass through an 8000 μm sieve. This cut-off level was selected to preserve larger fractions found in natural soil. Soils obtained from the same sample quadrats were mixed. Aggregates were isolated by employing a circular sieve shaker machine (type: AS 200 BASIC, Retsch Germany), following the procedure used by Bach and Hofmockel (38). Approximately 200g of soil was placed on a stack of sieves with 2000, 1000, and 250 μm mesh openings. The stack was shaken at approximately 200-250 rpm for 5 minutes. The soil was gently removed from each sieve and weighed to determine the distribution of aggregates. Aggregates isolated from all methods are referred to by size: large macroaggregates (8000-2000 μm), medium macroaggregates (1000-2000 μm), small macroaggregates (250-1000 μm), and microaggregates (<250 μm). A portion of each aggregate fraction sub-sample was taken out and air-dried in a natural state for the determination of soil organic nitrogen (SOM), total N, available phosphorus (available P), available potassium (available K), and pH. A remainder of each aggregate fraction sub-sample was saved and immediately transported to a refrigerator at 4°C for soil microbial biomass, NH₄⁺-N, and NO₃^–N quantification and rates of soil N mineralization determination. Negligible amounts of roots, stones, and other debris were removed (Supplementary Figure S1).

One-way ANOVA and Duncan’s method were used to analyze the variance and multiple comparisons (P< 0.05) on the proportion of the weight of aggregates in soil, soil properties, the levels of AHN components, and net N transformation (40). The interactions between N addition and aggregate size were studied using two-way ANOVA (42). All results were expressed as mean ± standard deviation. The Pearson correlation test examined the relationship between soil properties and AHN component levels (45). Soil properties were analyzed by principal component analysis to examine the environmental variables comprehensively (45). With component levels as the independent variable, a stepwise multiple regression analysis was carried out to further determine the main AHN components affecting net N transformation (28). All the aforementioned statistical analyses were performed using SPSS 21.0 software (SPSS, Inc., IL, USA). Mapping was performed with Origin 2021 (OriginLab Corporation, MA, USA).

The levels of total N, NH₄⁺-N, and NO₃^–N, as well as the microbial biomass of aggregates, generally exhibited an upward trend under N addition. The 8000-2000 μm fraction exhibited the minimum values for total N, NH₄⁺-N, and NO₃^–N, while the <250 μm fraction showed the maximum values, mirroring the trend observed in microbial biomass changes (Figures 1, 2). The available P level notably decreased with the increase in N addition and diminished with the reduction in aggregate size, displaying an inverse trend to the change in organic matter level with size (Figure 1). The organic matter level significantly declined under the N90 treatment compared to N0 (Figure 1, Supplementary Table S1). The available K level did not exhibit a significant response to the alterations in N addition and particle size (Figure 1). The total N, NH₄⁺-N, NO₃^–N, available P, available K, soil organic matter, and microbial biomass under N addition were analyzed using principal component analysis (PCA, Table 1). PC1, PC2, and PC3 accounted for 52.02%, 16.80%, and 12.59% of the variation in properties under N addition, respectively. The total N level and soil microbial biomass had the greatest influence on the overall changes. Available P had a negative effect on the changes in aggregate properties.

N addition generally increased the levels of acid-hydrolyzable N components in aggregates compared to those in the N0 treatment (Figure 3). The components were ranked in the order of AAN > AN > UAN > ASN. Smaller size fractions exhibited higher AAN levels compared to larger sizes (Figure 3). N addition decreased the proportions of ASN but increased UAN in the aggregates (Table 2). The proportion of aggregates in the soil by weight across different sizes was ranked in descending order: 1000-250 μm fraction, 2000-1000 μm fraction, 8000-2000 μm fraction,<250 μm fraction. The impact of N addition on these proportions was found to be insignificant (Table 2).

With the exception of the 8000-2000 μm fraction, the net ammonification (R_minNH₄⁺-N) decreased slightly under N30 and N60 treatments and increased significantly under N90 treatment compared with that of the control group. In contrast, the net nitrification (R_minNO₃^–N) showed a positive response to N addition. Both R_minNH₄⁺-N and R_minNO₃^–N showed the same overall trend of increasing as aggregate size decreased. Maximum values were observed in the<250 μm fraction, while minimum values were noted in the 8000-2000 μm fraction. More than 80% of the net N mineralization (R_minN) is contributed by R_minNO₃^–N, resulting in a similar overall trend in response to N addition for both processes in aggregates. The net mineral N accumulation was the cumulative sum of R_minN throughout the incubation period (Figure 4).

The levels of total soil N, NH₄⁺-N, NO₃^–N, and organic matter had significant positive correlations with the levels of AHN components and TAN, whereas the available P level had significant negative correlations with AAN, AN, UAN, and TAN levels (Figure 5). Equations were established using stepwise regression analysis with levels of AHN components and TAN as independent variables, and R_minNH₄⁺-N, R_minNO₃^–N, R_minN, and net N mineralization accumulation as four dependent variables. AAN and ASN were the dominant factors affecting R_minNH₄⁺-N, with an R² of 0.252 (P< 0.01). ASN and TAN levels jointly affected R_minNO₃^–N, R_minN, and net N mineralization accumulation, with R² values of 0.527, 0.576, and 0.594 (P< 0.001), respectively (Table 3).

In terrestrial ecosystems, bioavailable N predominantly exists in an inorganic form within the soil matrix (46). However, when the inorganic N pool is insufficient to meet the biological needs of plants and microorganisms, low-molecular-weight organic N becomes a critical and direct N source, readily available for assimilation and utilization (26). Our investigation revealed a pronounced influence of soil aggregate size on the concentration of AHN components (Figure 3). Soil aggregates, fundamental to soil structure and microbial metabolic activity, are distinguished by their unique physical micro-architecture and intrinsic material microcirculation mechanisms (47). An elevated microbial biomass was noted within smaller-sized aggregates (Figure 2), suggesting an intensified biochemical reaction potential due to a greater specific surface area and enhanced adsorptive capacities (19). Consequently, soil organic matter (SOM) and total N exhibited a decreasing trend across aggregate size fractions, sequentially arranged as <250 μm, 250-1000 μm, 1000-2000 μm, and 8000-2000 μm (Figure 1), corroborating with the findings of Polláková et al. (48) and Dal Ferro et al. (49). This gradient elucidates the elevated levels of AHN components observed within the smaller aggregates. Upon N addition, TAN within aggregates experienced a significant rise, with the constituent levels descending in the order: AAN > AN > UAN > ASN, in line with prior empirical evidence (50, 51). A conspicuous positive correlation between AAN and total N was detected (Figure 5), paralleling the associations reported in preceding research (52). This relationship may be ascribed to the N addition augmenting both total N and SOM levels (Figure 1), potentially enhancing microbial biomass and SOM decomposition within the soil matrix (53). The primary AAN contributors were identified as microbial residue proteins and organic matter degradation by-products (54), thereby suggesting that elevated soil total N indirectly amplifies AAN levels. AN, conversely, originated from soil’s freshly incorporated inorganic N, inclusive of immobilized and adsorbed NH₄⁺-N (55), manifesting a significant and positive correlation with both total and inorganic N (Figure 5), as extensively demonstrated in preceding literature (29, 55). ASN was primarily derived from the cellular walls of soil microbiota (56), highlighting its close association with microbial proliferation and activity, alongside its role as a high-quality carbon (C) source for microbial propagation (57). Therefore, increments in ASN levels further validate the surge in microbial biomass within the smaller aggregate fractions (Figure 1). UAN, potentially originating from nucleic acids and soil humification intermediates (58), is also ostensibly linked with microbial processes (Figures 2, 3). AAN and AN serve as soluble organic N (SON) pools that are directly utilized by soil microbes and plants, representing a principal source of N mineralization. Hence, their N level after N addition signifies an enriched bioavailable N reserve for soil organisms and a heightened mineralization potential (59).

N addition led to a significant increase in total N levels (Figure 1), suggesting that the soil’s ability to assimilate N exceeded the combined uptake by plants and microorganisms, as well as losses through leaching and gaseous emissions (60). The fluctuations in soil NH₄⁺-N and NO₃^–N levels in response to nitrogen addition mirrored the trend observed for total N. PCA revealed that microbial biomass and total N exerted a more pronounced influence on the soil properties within aggregates (Table 1). N addition relieved the limitations on soil microorganisms caused by N deficiency, resulting in an increase in soil microbial biomass, which is consistent with the findings of Zeng et al. (61). Research has shown that chronic excess N can reduce the availability of soil P and other salt-based ions, negatively affecting the ecosystem (62). In the present study, the level of available P exhibited a marked decline concomitant with the escalation of N addition (Figure 1). This phenomenon is likely due to the increased utilization of soil-available P by soil microorganisms, resulting from the heightened microbial activity caused by nitrogen supplementation (63) Consequently, available P manifested a negative impact on soil environmental fluctuations as delineated by PC1 (Table 1). The Three Gorges Reservoir area, situated within the northern subtropics, harbors a forest ecosystem acutely sensitive to climatic fluctuations. is an indispensable indigenous silvicultural species in the region. Soil P insufficiency persists as a pivotal factor in circumscribing the productivity of local forests (64). Zhang et al. (65) reported that N addition leads to a decline in soil P. Consistently, we observed that N addition significantly reduced P levels in Q. variabilis plantation soils. Although N deposition increases readily mineralizable SON and enhances short-term N availability, potentially benefiting forest productivity, it accelerates soil mineral N loss and may induce long-term P limitation (66). The decline in soil P likely results from enhanced P cycling under elevated N conditions, which stimulates acid phosphatase activity. Increased microbial biomass further promotes P utilization (42). As an ectomycorrhizal species, Q. variabilis exhibits improved N nutrition and mycorrhizal efficiency under N deposition, leading to greater P uptake by roots and consequently reduced soil P content (67, 68) From a management perspective, P fertilization should be considered when necessary to maintain forest productivity. The greater the proportion of aggregates with larger particle size to soil by weight, the more stable the soil structure (19, 25). Short-term N addition predominantly impacted the levels of chemical elements such as C, N, P, and soil organic matter within aggregates. However, the formation of the physical structure of aggregates constituted a continuous, protracted process, governed by a specific mechanism that regulated the fluctuations in the soil chemical milieu (29). Consequently, the proportion of aggregates in the soil by weight remained relatively unaltered under the with of short-term nitrogen supplementation treatments (Table 2).

In the present investigation, the concentrations of R_minNH₄⁺-N, R_minNO₃^–N, and R_minN within soil aggregates exhibited an upward trajectory as the size of the aggregates diminished, culminating in peak values within the<250 μm fraction (Figure 4). We note that the net N transformation rates demonstrated variability across different aggregate sizes under identical N addition conditions, corroborating our second hypothesis. This indicates that R_minN is profoundly impacted by the availability of mineralizable substrates (69). The multiple stepwise regression analysis revealed that AAN and TAN significantly and positively correlated with the net N transformation rate (Table 3). Noteworthy is the observation that both total N and the mineralizable component AHN were markedly elevated in the<250 μm fraction relative to larger aggregate sizes (Figure 1, Table 1), thereby suggesting a heightened potential for N mineralization within this smaller fraction (70). Additionally, the<250 μm fraction, possessing a more extensive specific surface area, facilitated greater exposure of organic particles to macroaggregates, thereby augmenting the likelihood of N mineralization (71). Furthermore, this fraction had higher levels of AAN and AN compared to larger fractions (Figure 3), suggesting higher concentrations of microbial biomass and hydrolytic enzyme activities within the<250 μm fraction (71). Collectively, these factors contributed to a significantly higher net N transformation rate in the<250 μm fraction.

N addition was observed to significantly modulate the net N transformation within soil aggregates (Figure 4). The aggregates showcased a notable divergence in the behavior of R_minNH₄⁺-N relative to R_minNO₃^–N in response to N addition, with the latter representing more than 80% of R_minN. As a result, the patterns of R_minN paralleled those of R_minNO₃^–N in response to N addition (Figure 4). Specifically, under N30 and N60 treatment conditions, R_minNH₄⁺-N displayed negative values during the mineralization phase, distinctly lower than those observed in the control (Figure 4). This phenomenon occurs because R_minNH₄⁺-N represents a net increase, calculated as the final NH₄⁺-N concentration after incubation minus the initial concentration. NH₄⁺-N serves as a preferential nitrogen source for soil microorganisms, which continuously assimilate both the inherent and newly mineralized NH₄⁺-N throughout the incubation period to energize their biological processes (72, 73). The addition of N was seen to substantially enhance soil microbial biomass (Figure 2), suggesting a faster rate of material cycling in the soil compared to the N0 scenario (74). The heightened biochemical activity within the soil system under N30 and N60 treatments led to a more extensive consumption of NH₄⁺-N than was the case in the control group, culminating in a reduction of R_minNH₄⁺-N under these nitrogen-enriched conditions (Figure 4). Additionally, NH₄⁺-N constitutes the precursor for nitrification within ecosystems, and in nitrogen-saturated environments, the majority of soil NH₄⁺-N undergoes nitrification (75, 76). Correspondingly, the R_minNO₃^–N within aggregates saw a significant uptick under N30 and N60 treatments compared to that in the control group, which further underscored the accelerated utilization of NH₄⁺-N as a nitrification substrate. Conversely, under the N90 treatment, there was an accumulation of mineralized NH₄⁺-N due to the introduction of an abundant exogenous supply of NH₄⁺-N catering to microbial requirements (77). In the acidic soil pH typical of the Q. variabilis plantation in the Three Gorges Reservoir area, a decline was observed under the N90 treatment conditions (Supplementary Table S2). Nitrification reactions typically proceed optimally within a pH spectrum of 5.5 to 10, with a significant diminution of activity below a pH of 5.0, a decline attributed to the sensitivity of the nitrifying microorganisms, as well as the equilibrium between free ammonia and NH₄⁺-N (78). A decline in R_minNO₃^–N was observed within the 8000-2000 μm and 2000-1000 μm fractions under the N90 treatment compared to the N60. This trend may be ascribed to the concomitant reduction in available phosphorus levels (79) and soil pH (46) that ensued from the substantial nitrogen additions.