Conservation tillage regulates aggregate organic carbon and microbial communities in aeolian sandy soil

Aeolian sandy farmland in arid and semi-arid regions is often constrained by loose texture, weak aggregate stability, and low soil organic carbon (SOC), which limit soil quality improvement under intensive cultivation. Identifying management practices that improve aggregation and carbon retention while altering soil microbial communities is important for sustainable production in these systems. A three-year field experiment was carried out on aeolian sandy farmland in Hotan, Xinjiang, China. Four tillage management treatments were set up, including conventional tillage (CT), sole no-tillage (NT), no-tillage with sheep manure organic fertilization (NTOF), and no-tillage with straw mulching (NTSM). The NT and NTSM treatments showed greater mega-aggregate formation and higher mean weight diameter (MWD) than CT, and NTSM gave the most evident increase in aggregate-associated organic carbon, particularly in the mega-aggregate fraction. The implementation of all three conservation tillage practices had no significant effect on microbial alpha diversity; however, NT, NTOF and NTSM significantly influenced bacterial community composition, whilst NTSM significantly altered fungal community composition and caused the bacterial network to shift from being predominantly characterized by negative correlations to positive correlations. Relationships between mega-aggregate microbial community characteristics and SOC or mean weight diameter (MWD) were treatment dependent. Structural equation modelling suggested potential linkage pathways among soil properties, microbial community structure, SOC accumulation, and aggregate stability. Overall, NTSM showed the strongest overall response, as reflected by improved soil conditions, greater aggregate-associated carbon accumulation, enhanced aggregate stability, and distinct changes in mega-aggregate microbial communities. These findings suggest that no-tillage with straw mulching may be a promising management option for improving soil structure and carbon retention in aeolian sandy farmland.

Aeolian sandy soils are widely distributed in arid and semi-arid regions and are often characterized by coarse texture, weak aggregation, and low soil organic carbon (SOC) content (Dai et al., 2019). These properties make it difficult to maintain stable soil structure and to accumulate organic carbon under intensive agricultural management. In such soils, aggregate formation is constrained by the limited abundance of fine particles and inorganic binding agents, including clay minerals and iron and aluminum (hydro)oxides (Denef and Six, 2006). As a result, aggregation in aeolian sandy soils may depend more strongly on organic binding agents and biological processes than in clay-rich soils (Y. Zhang et al., 2021). Soil aggregates provide physical habitats for organic matter and microorganisms and are closely associated with water retention, nutrient cycling, erosion resistance, and SOC protection (Oades, 1984; Six et al., 2000). However, because aggregates in aeolian sandy soils are weakly developed, they are particularly susceptible to disruption by mechanical tillage, which may further reduce structural stability and limit SOC accumulation (Nielsen and Ball, 2015; Weidhuner et al., 2021).

Tillage practices can modify aggregate dynamics and soil carbon distribution in cropland systems. Conventional tillage repeatedly disturbs the soil matrix, which can break down larger aggregates, expose previously protected organic matter, and increase the loss of SOC through mineralization and erosion (Sae-Tun et al., 2022). In contrast, conservation tillage practices, especially those involving reduced soil disturbance and residue retention, have been associated with improved soil structure, greater aggregate stability, and higher aggregate-associated SOC in many agroecosystems (Chavarria et al., 2018; Pearsons et al., 2023; M. Zhang et al., 2024). No-tillage can improve soil structure and reduce organic matter mineralization by minimizing soil disturbance and enhancing the physical protection of organic carbon within soil aggregates (Samantha L. Colunga et al., 2025). However, in arid and semi-arid sandy soils, no-tillage alone may be insufficient to comprehensively improve soil quality because, in the absence of surface straw cover, its capacity to reduce wind erosion, suppress soil evaporation, regulate soil hydrothermal conditions, and provide organic carbon substrates for microorganisms is limited (Du et al., 2022). As an important complementary practice to no-tillage, straw mulching can enhance surface protection, increase soil moisture, and supply residue-derived carbon for microbial growth and activity (F. Jiang et al., 2024). Nevertheless, in crop–livestock systems of arid northwestern China, crop straw is not only a soil amendment but also an important feed resource for cattle and sheep. Therefore, whether straw should be retained on the soil surface as mulch or removed for livestock feeding and subsequently returned to the field as manure represents a practical trade-off in residue utilization (Valbuena et al., 2012). Accordingly, it is necessary to evaluate whether no-tillage alone is feasible in arid sandy farmland. If no-tillage without straw cover is insufficient, further comparison between direct straw mulching and manure return after livestock feeding is needed to clarify their distinct ecological functions.

Soil microorganisms are closely linked to aggregate formation and SOC transformation, but their responses to tillage are not uniform across aggregate size fractions. Bacteria and fungi may differ in their associations with soil particles, organic substrates, and aggregate stability (Z. X. Han et al., 2025; Reed Close et al., 2026). For example, bacterial groups have often been related to early-stage aggregation and micro-aggregate processes, whereas fungal hyphae and fungal-derived organic compounds are frequently associated with larger aggregates and aggregate stabilization (Six et al., 2004; Totsche et al., 2018). Tillage-induced changes in soil microhabitats may therefore be accompanied by shifts in microbial community composition and co-occurrence patterns (C. L. Han et al., 2024; Zheng et al., 2023). Previous studies have also reported changes in specific microbial groups, such as Desulfovibrio, Gram-negative bacteria, and arbuscular mycorrhizal fungi, under different tillage systems (Cao et al., 2022; Li et al., 2024; M. N. Zhang et al., 2024). However, these findings have mainly been derived from bulk soil or general aggregate fractions, and microbial community responses within mega-aggregates of aeolian sandy soils remain insufficiently characterized (W. S. Liu et al., 2024; H. T. Wang et al., 2025).

Although conservation tillage has been widely studied as a strategy for improving soil structure and SOC storage, less attention has been paid to how different conservation tillage practices simultaneously affect aggregate stability, aggregate-associated organic carbon, and microbial communities within mega-aggregates in aeolian sandy soils. This knowledge gap is important because mega-aggregates represent a sensitive structural fraction in sandy soils and may serve as an important site for recent carbon accumulation and microbial colonization. Moreover, the relationships among soil environmental changes, aggregate-associated SOC, and mega-aggregate microbial communities under different tillage practices remain unclear (Crawford et al., 2012).

To address these questions, four tillage regimes were evaluated over three years in aeolian sandy farmland in Hotan, Xinjiang, China: conventional tillage (CT), no-tillage (NT), no-tillage with organic fertilizer application (NTOF), and no-tillage with straw mulching (NTSM). The objectives of this study were: (i) to assess the effects of different tillage practices on aggregate distribution, aggregate stability, and aggregate-associated organic carbon; (ii) to characterize the diversity, taxonomic composition, and association patterns of bacteria and fungi inhabiting mega-aggregates under different tillage practices; and (iii) to examine the potential relationships among soil properties, aggregate-associated organic carbon, aggregate stability, and mega-aggregate microbial community characteristics. We hypothesized that conservation tillage would improve aggregate stability and increase aggregate-associated SOC in aeolian sandy soils, with the strongest effects occurring under straw mulching. We further hypothesized that these soil structural and carbon responses would be accompanied by shifts in microbial community composition and network characteristics within mega-aggregates.

Experimental site and design

The field experiment was initiated before the 2023 maize growing season, and 2023 was the first complete experimental year. The treatments were continuously maintained during the 2023, 2024, and 2025 maize growing seasons. The experimental site lies at 37°42′ N, 79°08′ E and 1,243.7 m above sea level. Climatically, the area belongs to the warm-temperate arid desert zone, with an annual mean temperature of 11.3 °C and a frost-free period of about 177 days. Annual evaporation reaches about 2,239 mm, whereas annual precipitation is only 36–37 mm and occurs mainly from May to August. Soil texture analysis showed that the experimental soil belonged to the sandy loam class, with sand, silt, and clay contents of 65.87%, 23.09%, and 1.52%, respectively; the soil type was identified as aeolian sandy soil. Spring maize cultivar Dingyu 678 was grown as a single-crop system. Before the experiment was initiated, the initial soil physicochemical properties were measured and are provided in Supplementary Table 1. Four tillage regimes were compared in this study: conventional tillage (CT), no-tillage (NT), no-tillage with organic fertilizer application (NTOF), and no-tillage with straw mulching (NTSM). A randomized complete block design was used, and the four treatments were each represented by six replicate plots, yielding 24 plots in total. Individual plots covered 119.7 m². Appendix S1 provides the detailed tillage procedures and field management practices used during the experiment.

Soil sampling

Soil sampling was performed on 7 July 2025, when maize was at the tasseling stage, a period generally associated with relatively high microbial activity (Ayiti et al., 2022; Jia et al., 2021). This date is consistent with the typical phenological timeline for spring maize sown in mid-April in the Hotan Oasis, as under local plastic mulch drip irrigation conditions, maize typically begins to tassel 50–55 days after sowing in mid-April. Each plot was considered one independent replicate. For each plot, five points between maize rows were sampled within the 0–25 cm cultivated layer, and the collected soil was pooled to form a composite sample for physicochemical analysis. The composite samples were kept in a cool and shaded place and air-dried naturally before physicochemical analysis. For aggregate fractionation and microbial community analysis, a separate fresh soil sample was collected from each experimental plot. This sample was not mixed with the composite soil sample used for physicochemical analyses. During sampling, standard soil sampling procedures were followed, and disturbance to the original soil structure was minimized. The collected soil samples were then placed in an ice box, surrounded by shock-absorbing foam, and immediately transported to the laboratory by vehicle. In the laboratory, the soil blocks were gently broken apart along their natural planes of weakness. Stones, crop residues, and visible roots were carefully removed from the soil samples using tweezers, after which the samples were used for aggregate fractionation.

Test parameters

Soil physicochemical analysis and aggregate fractionation

Standard soil agrochemical methods were used to analyze soil physicochemical properties (Lu, 2000). For aggregate separation, fresh soil samples were fractionated under field-moist conditions using a modified dry-sieving procedure. Approximately 300 g of fresh soil at field-moist condition was placed on a two-sieve assembly with mesh sizes of 2.0 mm and 0.25 mm. The sieve set was moved vertically with an oscillation amplitude of approximately 3 cm and was oscillated 60 times within 3 min. Based on the retained material, soil aggregates were divided into three size classes: mega-aggregates (>2 mm), macro-aggregates (0.25–2 mm), and micro-aggregates (<0.25 mm) (Nie et al., 2014). After wet sieving, aggregates from each size fraction were collected and weighed. The mean weight diameter (MWD) was then calculated following the method described previously (N. Li et al., 2015). The separated aggregate samples were further divided according to subsequent analyses. A portion of the mega-aggregate fraction was immediately stored at −80 °C for microbial community analysis, whereas the remaining aggregate samples from all three size fractions were air-dried for SOC determination. MWD was calculated as follows:

M W D = ∑ i = 1 n X i W i In this equation, X i is the mean diameter of aggregate class i, W i is its proportion of the total dry aggregate mass, and n refers to the number of aggregate classes.

Soil microbial analysis

High-throughput amplicon sequencing was conducted on mega-aggregate samples (>2 mm) by Shenzhen Micromicro Technology Group Co., Ltd. (Shenzhen, China). Raw paired-end reads were imported into QIIME 2 for sequence processing. After quality filtering, the DADA2 plugin was applied for denoising, paired-end read merging, and chimera removal. The resulting amplicon sequence variant (ASV) feature tables and representative sequences were used for downstream microbial community analyses. Bacterial taxonomy was assigned using the QIIME 2 taxonomy plugin against the SILVA 138.2 database. Reference sequences were trimmed to the target amplicon region, and the confidence threshold was set to 0.70. Fungal taxonomic classification was performed with the vsearch method using the UNITE ITS reference database with INSD annotations, also with a confidence threshold of 0.70. Mitochondrial and chloroplast sequences were removed before subsequent analyses. Microbial co-occurrence networks were constructed using SparCC correlations. Bootstrap resampling was used to estimate the significance of correlations. Before network construction, taxonomic units occurring in fewer than 20% of samples were removed. Only correlations with |r| > 0.30 and P < 0.05 were used to define edges in the co-occurrence networks. Network construction and topological parameter calculation were performed in R 4.4.1. We used the Zi–Pi framework to describe the ecological roles of network nodes. With Zi = 2.5 and Pi = 0.62 as cut-off values, nodes were categorized as peripherals, connectors, module hubs, or network hubs. Network files were then exported in GraphML format for subsequent visualization in Gephi.

Statistical analysis

Statistical processing was carried out using R 4.4.1. For soil physicochemical and aggregate-related measurements, treatment differences were analyzed with one-way ANOVA. The suitability of ANOVA was checked by testing normality with the Shapiro–Wilk test and variance homogeneity with Levene’s test. When the data satisfied these requirements, Fisher’s LSD procedure was used for pairwise comparisons among treatments, with significance defined at P < 0.05. For microbial community analyses, bacterial and fungal relative abundances within mega-aggregates were summarized at the phylum level. The top 12 phyla were displayed using stacked bar plots generated with ggplot2, and the remaining phyla were grouped as “Others”. To reduce the influence of differences in sequencing depth, microbial diversity was analyzed using normalized data. Alpha diversity was characterized using observed richness, Chao1, Faith’s phylogenetic diversity, and Shannon indices. Alpha-diversity indices were compared among treatments using the Kruskal–Wallis test, with Dunn’s test applied for subsequent pairwise comparisons. Community dissimilarity was quantified using the Bray–Curtis metric and displayed by principal coordinate analysis (PCoA). We used Spearman correlation and Mantel tests to quantify how microbial community attributes were related to key soil variables. Spearman’s rank correlation analysis was used to assess relationships between dominant taxa and key soil variables, and the results were visualized using linkET. Mantel tests were further used to evaluate relationships between SOC, MWD, microbial diversity, community composition, and key microbial genera under different tillage treatments.

Potential pathways linking conservation tillage, soil properties, mega-aggregate microbial community characteristics, SOC content, and aggregate stability were evaluated with SEM using the R package piecewiseSEM. A PCA-based soil property index was derived from standardized AN, AP, and MC values, allowing these correlated soil variables to be represented by a single composite variable in the SEM. The first principal component (PC1) was then used to represent overall soil properties in the SEM. Variance inflation factor diagnostics indicated no serious multicollinearity among the candidate soil variables (all VIFs < 5; Table 1).

Effects of conservation tillage on soil physicochemical properties

The four tillage regimes produced distinct changes in selected soil physicochemical properties (Table 2). Relative to CT, SOC content increased by 17.87% under NT, 11.63% under NTOF, and 22.56% under NTSM. Alkali-hydrolysable nitrogen also varied significantly among treatments, following the sequence NTSM > NT > NTOF > CT. Available phosphorus content was higher in both the NT and NTSM treatments than in the CT treatment, with the highest content observed in the NTSM treatment. Soil moisture content under NTSM was significantly greater than that under CT. In contrast, available potassium, pH, electrical conductivity, and bulk density did not vary significantly with tillage regime.

Aggregate distribution and stability

Under all treatments, micro-aggregates were the predominant fraction, accounting for more than 77.81% of the total aggregate mass (Figure 1A). Tillage treatments significantly affected the proportions of mega- and micro-aggregates. The mega-aggregate fraction was significantly higher under NT and NTSM than under CT. In absolute terms, it increased from 11.43% under CT to 16.37% under NT and 17.80% under NTSM. Expressed as relative changes, these increases were 43.18% and 46.94% under NT and NTSM, respectively, compared with CT. No significant treatment-related change was detected for the macro-aggregate fraction. In contrast, the micro-aggregate fraction decreased by 6.10 percentage points under NTSM relative to CT, equivalent to a relative decrease of 7.67%. Significant treatment differences were also observed for MWD (Figure 1B). Compared with CT, MWD increased by 26.10% under NT and by 31.36% under NTSM.

Aggregate-associated organic carbon across aggregate size fractions

Aggregate-associated organic carbon differed significantly across tillage treatments and aggregate size fractions (Figure 2). Compared with the CT treatment, the NT, NTOF, and NTSM treatments significantly increased the soil organic carbon content in mega-aggregates, with increases of 4.95%, 10.21%, and 12.35%, respectively. Furthermore, NTOF and NTSM produced higher carbon levels in macro-aggregates, while NT and NTSM increased carbon levels in micro-aggregates.

Diversity and composition of mega-aggregate microbial communities

For mega-aggregate bacterial and fungal communities, alpha diversity was described using four indices: observed species, Chao1, Faith’s phylogenetic diversity, and Shannon index (Tables 3, 4). Among bacterial alpha-diversity metrics, only Chao1 differed significantly, decreasing in NTOF relative to CT, whereas observed species, Faith’s phylogenetic diversity, and Shannon index remained statistically unchanged across treatments. There were no significant differences in fungal alpha-diversity indices among the four tillage treatments.

Principal Coordinate Analysis (PCoA) revealed significant differences in the microbial community composition of bacteria (P = 0.01) and fungi (P = 0.005) among the different tillage treatments. The NT, NTOF, and NTSM treatments significantly influenced bacterial community structure, whereas fungal community structure was significantly affected only by the NTSM treatment. The PCoA results showed that axes 1 and 2 together represented 24.84% of the bacterial community variation and 31.30% of the fungal community variation (Figures 3B, D).

At the phylum level, NTOF and NTSM were characterized by greater representation of Acidobacteriota and Gemmatimonadota but lower representation of Planctomycota relative to CT. Relative to CT, the three conservation tillage treatments showed a common shift toward higher Myxococcota and lower Chloroflexota abundances (Figure 3A; Supplementary Figure 1). In the fungal community, Ascomycota showed greater representation under NTOF and NTSM than under CT (Figure 3C; Supplementary Figure 2).

Network-level responses of mega-aggregate microbial communities

The balance between positive and negative bacterial associations varied across treatments: CT, NT, and NTOF networks were dominated by negative associations, while the NTSM network contained a higher proportion of positive associations (Figure 4; Supplementary Table 2). The NT and NTOF networks were characterized by reduced node numbers, fewer links, and lower average degree compared with CT. The NTSM network, however, retained a node and link structure close to that of CT, together with a modest increase in average degree and a decrease in average path length (Supplementary Tables 3, 4). According to Zi–Pi analysis, under the CT, NT, NTOF, and NTSM treatments, 4, 1, 2, and 3 bacterial nodes were classified as module hubs, and 4, 2, 3, and 5 nodes were classified as connectors (Figure 5A). Acidobacteriota, Verrucomicrobiota, Pseudomonadota, and Gemmatimonadota were identified as key bacterial groups (Supplementary Table 5).

In the fungal co-occurrence network, negative correlations predominated across all tillage treatments (Figure 6; Supplementary Table 2). The NT and NTOF treatments produced simpler network structures than CT, as indicated by fewer nodes and links, lower average degree, and greater average path length. Under NTSM treatment, the number of nodes and links increased, the average degree remained largely unchanged, and the average path length increased (Supplementary Tables 3, 4). Based on Zi–Pi classification, one module hub and four connectors were detected in the CT fungal network, while two module hubs and three connectors were detected in the NTSM fungal network (Figure 5B). Most of these key fungal nodes belonged to Ascomycota (Supplementary Table 6).

Relationships between mega-aggregate microbial community characteristics and SOC and MWD

The Mantel test indicated that, under conservation tillage conditions, bacterial community composition and key genera were significantly correlated with soil organic carbon (SOC) (Supplementary Table 7). Genus-level correlations showed that Hirschia was inversely associated with SOC, and Rubellimicrobium was inversely associated with MWD, with both relationships reaching statistical significance (Figure 7A).Furthermore, under NTOF and NTSM treatments, fungal community composition and key genera were significantly correlated with MWD (Supplementary Table 8). Specifically, the genus Beauveria was positively correlated with SOC, while the genera Arthroderma and Penicillium were positively correlated with MWD (Figure 7B).

Potential linkage pathways between conservation tillage and aggregate stability

A piecewise structural equation model (SEM) was applied to evaluate the potential associations among conservation tillage, soil properties, microbial communities within mega-aggregates, SOC content, and aggregate stability (Figure 8). Conservation tillage had a positive path coefficient for soil properties, which in turn showed positive links with fungal and bacterial microbial communities. SOC content was linked mainly to the fungal microbial community and then to aggregate stability through significant positive paths. In contrast, the direct paths from the bacterial microbial community to SOC content and from both microbial communities to aggregate stability were not significant. These results suggest a potential pathway in which conservation tillage is linked to improved soil properties, shifts in the fungal microbial community within mega-aggregates, higher SOC content, and greater aggregate stability.

Effects of conservation tillage on soil fertility and aggregate stability in aeolian sandy soils

The results of this study indicate that, relative to CT, conservation tillage treatments increased several soil fertility indicators in aeolian sandy soils, with the most evident response observed under NTSM. Specifically, NT, NTOF, and NTSM all increased soil organic carbon content, whilst NT and NTSM also significantly increased alkali-hydrolysable nitrogen and available phosphorus content; furthermore, NTSM significantly increased soil moisture content (Table 2). These findings are consistent with existing research, which suggests that no-tillage farming combined with straw mulching can enhance SOC accumulation and soil nutrient retention capacity by reducing soil disturbance, increasing the input of surface organic matter, and minimising losses from organic carbon mineralisation (Gómez-Muñoz et al., 2021). In arid and semi-arid regions, straw mulching can also improve soil water and thermal conditions by reducing soil water evaporation and buffering surface temperature fluctuations, thereby enhancing nutrient availability (Cong et al., 2024). Furthermore, continuous organic matter input is also considered to contribute to increased alkali-hydrolysable nitrogen content and improved soil fertility potential (Cao et al., 2022). Therefore, from a soil environmental perspective, compared with CT, NTSM is more conducive to enhancing the organic carbon content, moisture content, and nutrient levels of aeolian sandy soils, thereby providing more favourable environmental conditions for subsequent aggregate formation and stability.

Improvements in the soil environment are further reflected in changes to aggregate structure and stability indices. In the present study, higher mega-aggregate proportions and MWD were observed under NT and NTSM than under CT, suggesting that reduced soil disturbance, particularly when combined with straw mulching, was associated with improved aggregate stability in aeolian sandy soils (Figure 1). Aggregates are considered key indicators of soil structural quality and function (Gupta and Germida, 2015), and in aeolian sandy soils—which are characterized by low clay content and fragile structure—aggregate formation relies more heavily on organic matter input and the maintenance of biological processes. Under conservation tillage management, greater amounts of crop straw are incorporated into the soil. This straw first decomposes into fine-particle organic carbon, which serves as the nucleation centre for macro-aggregates. Furthermore, Straw-derived organic inputs may support microbial activity and the formation of organic binding materials, thereby contributing to the development of larger aggregate fractions (Hati et al., 2021). Conservation tillage has also been reported to improve the physical protection of SOC within larger aggregates and to increase soil structural stability (Li et al., 2025; Zhang et al., 2024). In this study, the NTSM treatment not only improved the soil environment but also exhibited a higher proportion of mega-aggregates and a higher MWD, indicating that, under the experimental conditions, the NTSM treatment was more effective in enhancing aggregate stability in aeolian sandy soils compared to the NT and NTOF treatments.

However, although the NTOF treatment increased SOC and certain nutrient contents in aeolian sandy soils, its effects on aggregate proportion and MWD were not significant. These results indicate that conservation tillage practices do not produce uniform changes in aggregate formation and stability in aeolian sandy soils. In addition to the organic matter input itself, aggregate stability may also be influenced by the method of application, decomposition rate, and the intensity of soil disturbance (Abdelhafez et al., 2018; Gil et al., 2024). Studies have shown that the relatively rapid organic matter turnover and mineralisation processes following the application of organic fertiliser may also weaken its short-term beneficial effect on aggregate stability (Comin et al., 2013). Therefore, under the conditions of this field experiment, the combination of reduced disturbance and straw retention appeared to be more favourable for improving soil properties and aggregate stability than organic fertiliser input alone.

Aggregate-associated carbon responses across aggregate size fractions

This study indicates that different conservation tillage practices exert distinct effects on aggregate-associated organic carbon across particle size fractions. Relative to CT, mega-aggregate SOC was higher under all three conservation tillage treatments, whereas macro-aggregate SOC increased mainly under NTOF and NTSM, and micro-aggregate SOC increased under NT and NTSM (Figure 2). Overall, NTSM increased the organic carbon content in aggregates of all particle sizes, suggesting that straw mulching not only promotes overall soil organic carbon accumulation but also enhances carbon sequestration within aggregates of different particle sizes. This pattern agrees with previous findings that residue-retaining no-tillage systems often favor SOC accumulation within aggregate fractions (S. L. Colunga et al., 2025). This pattern may reflect increased residue-derived carbon inputs, slower organic carbon turnover, and greater retention of carbon in particulate and mineral-associated forms.

Higher carbon levels in mega- and macro-aggregates may reflect reduced mechanical disturbance, greater surface straw retention, and enhanced retention of newly added carbon within larger aggregate fractions. Previous studies have indicated that larger-sized aggregates are not only the primary sites for the enrichment of newly incorporated organic carbon, but also serve as important structural units for the physical protection of soil organic carbon (Qi et al., 2022). In conservation tillage systems, straw-derived residues can be transformed into particulate organic carbon, providing substrates that may favor macro-aggregate development and organic carbon retention (Li et al., 2024). At the same time, long-term conservation tillage may also reduce the intensity of carbon turnover between aggregates, thereby enhancing the carbon retention efficiency within larger-sized aggregates (Li et al., 2025). Consequently, the increase in organic carbon within mega-aggregates observed in the NT and, in particular, the NTSM treatments in this study is consistent with their performance in improving aggregate structure, and also suggests that larger-sized aggregates in aeolian sandy soils may be one of the primary vehicles through which conservation tillage promotes carbon accumulation.

It is worth noting that although the NTSM treatment reduced the proportion of micro-aggregates, the organic carbon content within these micro-aggregates still increased significantly. This suggests that changes in aggregate size distribution are not entirely synchronous with the carbon enrichment process within aggregates. In other words, under conservation tillage conditions, a decrease in the proportion of micro-aggregates does not necessarily imply a reduction in their carbon sequestration capacity. This phenomenon may be related to conservation tillage’s improvement of soil temperature and moisture conditions, reduction in the rate of organic carbon mineralization, and alteration of carbon transformation pathways between aggregates of different sizes (X. T. Liu et al., 2021). Therefore, from the perspective of aggregate-associated organic carbon, conservation tillage not only alters the particle size distribution of aggregates in aeolian sandy soils but may also alter the manner in which organic carbon is retained by aggregates of different particle sizes.

Mega-aggregate microbial community responses to tillage and residue management

Microbial alpha-diversity indices within mega-aggregates showed limited variation among tillage treatments after three years of conservation tillage implementation. With the exception of a slight decrease in the bacterial Chao1 index under the NTOF treatment, most alpha diversity indices for bacteria and fungi did not show significant differences between treatments (Tables 3, 4). Previous studies have also reported that microbial richness and evenness may remain relatively stable during the early to intermediate stages of conservation tillage adoption (Ibáñez et al., 2024; Y. Z. Li et al., 2020; Navarro-Noya et al., 2013). For soil types such as aeolian sandy soils, which are characterized by low organic matter content and fragile structure, the richness profile of microbial communities within mega-aggregates may remain relatively stable when conservation tillage is implemented for a short duration; consequently, alpha diversity indices may not be the most prominent indicators of change.

In contrast, different conservation tillage practices exerted more pronounced effects on microbial community composition. PCoA results revealed significant separation in bacterial and fungal community composition across different treatments, with the NT, NTOF and NTSM treatments exerting significant effects on bacterial community structure, whilst marked changes in fungal community structure were primarily observed under the NTSM treatment (Figure 3). This suggests that, at the mega-aggregate scale, changes in community composition better reflect the ecological differences resulting from tillage management than alpha diversity. Earlier studies on aggregate fractions also found that microbial community composition was more sensitive to conservation tillage than alpha-diversity indices (C. L. Han et al., 2024; Li et al., 2024). Furthermore, this may be related to differences in the response strategies of bacteria and fungi to environmental changes and carbon source quality. Bacteria typically respond more rapidly to changes in soil disturbance, nutrient availability, and readily degradable organic carbon; consequently, NT, NTOF and NTSM were all able to induce significant changes in bacterial community structure(C. Wang and Kuzyakov, 2024). Interestingly, whilst both NTOF and NTSM introduce organic matter into the soil, only NTSM alters the fungal community structure. This is primarily because the organic matter in NTOF tends to consist of nutrient-rich and more readily degradable organic matter, whereas the NTSM treatment continuously introduces straw residues into the topsoil. Straw is rich in lignocellulosic structural carbon components, which can alter fungal community structure by providing more suitable substrates and microhabitats for fungal colonization, hyphal expansion and residue decomposition (Ma et al., 2013). Consequently, NTSM is more likely to induce fungal community reorganization within mega-aggregates.

Co-occurrence network analysis further indicates that different conservation tillage practices not only influence the composition of mega-aggregate microorganisms but also alter the interaction patterns among them. Relative to CT, NT and NTOF simplified the bacterial co-occurrence networks, as indicated by reductions in node number, link number, and average degree. The NTSM network retained a similar overall size to CT, while positive associations became more prevalent (Figure 4; Supplementary Tables 2-4). For fungal communities, NT and NTOF produced simpler networks, whereas NTSM generated a larger network with more nodes and links (Figure 6; Supplementary Tables 2-4). These results indicate that the effects of different management practices on bacterial and fungal networks within mega-aggregates are inconsistent, with the NTSM treatment having a more pronounced effect on the fungal co-occurrence network. This response may be related to the ability of filamentous fungi to colonize aggregate surfaces and contribute to particle binding. Long-term conservation tillage can reduce soil disturbance and may create conditions that are more favorable for fungal hyphal development (Zhong et al., 2019). Straw mulching under conservation tillage may improve soil pore structure and create conditions more suitable for fungal colonization (J. J. Wang et al., 2014). Furthermore, NTSM treatment caused the bacterial network to shift from being predominantly negatively correlated to predominantly positively correlated; this may be related to the increased levels of organic carbon, readily available nutrients, and soil moisture under NTSM treatment, alongside the continuous input of straw-derived organic substrate (Su et al., 2020). Higher organic carbon inputs and more favorable moisture conditions may have alleviated resource constraints among certain bacterial groups within mega-aggregates, thereby increasing the proportion of positive correlations.

Potential links between changes in the soil environment and microbial communities and aggregate stability and organic carbon accumulation

A synthesis of the above results shows that conservation tillage impacts aeolian sandy soils not only through changes in individual indicators but also through combined effects on the soil environment, aggregate structure, aggregate-associated organic carbon, and microbial communities in mega-aggregates. Compared with CT, conservation tillage—especially NTSM—improved soil organic carbon, nutrients, and moisture. It increased the proportion of mega-aggregates and MWD, and promoted organic carbon accumulation in aggregates of different sizes. Mega-aggregate bacterial and fungal community profiles varied among the four tillage treatments. These results suggest a close link between improvements in the soil environment, aggregate structure, and microbial communities, which may jointly influence aggregate stability and organic carbon accumulation in aeolian sandy soils.

In microbial association networks, highly connected taxa are often considered important indicators of community organization (Berry and Widder, 2014; Zhan et al., 2018). Mantel tests and correlation analyses further support this potential link. Within the bacterial community, Hirschia showed a significant negative correlation with SOC, whilst Rubellimicrobium exhibited a significant negative correlation with MWD; within the fungal community, Beauveria showed a significant positive correlation with SOC, whilst Arthroderma and Penicillium exhibited significant positive correlations with MWD (Figure 7). Among the key bacterial genera, Hirschia was significantly and inversely associated with mega-aggregate SOC. This genus exhibits typical strong surface-adhering characteristics, which facilitate the formation of biofilms on particle surfaces, making it better suited to growth under oligotrophic conditions (Chepkwony et al., 2019; Chepkwony and Brun, 2021; Schlesner et al., 1990). However, the implementation of conservation tillage measures increased the organic carbon content within mega-aggregates. Higher SOC in mega-aggregates may favor taxa with greater carbon-resource demand, which could partly explain the lower relative abundance of Hirschia under carbon-enriched conditions. Rubellimicrobium was also significantly and inversely associated with aggregate MWD. This genus comprises aerobic Gram-negative bacteria commonly found in ecologically stressed desert sandy soils; they are better adapted to colonizing the surfaces of soil particles with weaker physical protection (Dastager et al., 2008; J. R. Han et al., 2023; Riedel et al., 2014). As the implementation of conservation tillage enhanced aggregate stability, their relative abundance conversely decreased.

Among the key fungal genera, Beauveria showed a significant positive correlation with SOC.Beauveria is widely known as an entomopathogenic fungus, and some species can also survive in soil, colonize the rhizosphere, and establish endophytic associations with plants. These ecological traits may allow Beauveria to persist in relatively resource-rich soil microhabitats. Therefore, the positive association between Beauveria and SOC in this study more likely indicates that this genus responded to SOC-rich mega-aggregate environments, particularly under treatments with greater residue-derived C inputs(B. Wang et al., 2021; Q. Zhang et al., 2023). Although endophytic or rhizosphere-colonizing Beauveria may indirectly affect plant growth and plant-associated microbial communities, we did not measure root-derived C input, fungal necromass, or C transformation processes (Mantzoukas et al., 2021; Sui et al., 2023; Y. F. Zhang et al., 2025). Thus, whether Beauveria actively contributes to SOC sequestration remains uncertain and requires further functional verification. In contrast, both Arthroderma and Penicillium showed a significant positive correlation with MWD, but their roles in the soil ecosystem differ. Previous studies have linked Arthroderma to keratin degradation, as members of this genus can produce keratinases that participate in the decomposition of nitrogen-rich recalcitrant substrates (Caretta et al., 1990; Kandemir et al., 2022; Mouliková et al., 2023). This may be because the implementation of conservation tillage measures has enhanced aggregate stability; aggregates encapsulate and protect more organic matter, thereby providing Arthroderma with a rich food source and suitable habitat, which in turn increases its abundance. As a typical filamentous saprotrophic fungus, Penicillium plays a significant role in promoting aggregate formation. Previous studies have shown that the entanglement of fungal hyphae and the production of extracellular polymers can enhance particle aggregation, and it has been confirmed that Penicillium fungi promote aggregate formation in sandy soils (Jiang et al., 2023; Lehmann et al., 2020; Tisdall et al., 1997).

Furthermore, piecewise SEM results further indicate the existence of potential coupling pathways between conservation tillage, soil environmental factors, the proportion of mega-aggregates, changes in the microbial community within mega-aggregates, and SOC and MWD (Figure 8). Specifically, the implementation of conservation tillage measures altered the nutrient and moisture conditions of aeolian sandy soils; these improvements reshaped the microbial community structure and altered key fungal and bacterial taxa, with changes in the fungal community structure within mega-aggregates promoting the sequestration of organic carbon in aeolian sandy soils, ultimately enhancing aggregate stability. However, the nonsignificant direct path from conservation tillage to aggregate stability should not be interpreted as evidence that tillage practices had no effect on aggregate stability. Rather, this path represents the remaining direct effect after accounting for soil properties, mega-aggregate microbial community characteristics, and SOC content in the SEM. Therefore, the treatment-level increases in aggregate stability observed under NT and NTSM were not inconsistent with the SEM results..

The present results suggest that conservation tillage was associated with coordinated changes in soil properties, aggregate size distribution, aggregate-associated carbon storage, and mega-aggregate microbial community characteristics in aeolian sandy farmland. Among the conservation tillage practices, NT and NTSM were associated with higher mega-aggregate proportions, greater MWD, and increased aggregate-associated organic carbon relative to CT. In terms of overall effects, NTSM demonstrated the most significant performance in improving the structural condition of aeolian sandy soils and promoting the accumulation of aggregate-associated organic carbon.

In mega-aggregates, microbial alpha diversity showed limited variation among treatments, whereas bacterial and fungal community composition and co-occurrence network structure displayed clear treatment-related shifts, especially under NTSM. Correlation analyses and SEM further indicated that conservation tillage was associated with coordinated changes in soil properties, mega-aggregate microbial community structure, SOC accumulation, and aggregate stability in aeolian sandy soils.

Overall, this study provides field evidence for understanding soil structure improvement, aggregate-associated organic carbon accumulation, and corresponding changes in the microbial community of mega-aggregates under conservation tillage on aeolian sandy soils, and offers a reference for optimizing soil management in aeolian sandy soil farmlands in arid and semi-arid regions.