Organic Management Mediates Multifunctionality Responses to Land Conversion from Longan (Dimocarpus longan) to Tea Plantations at the Aggregate Level

Abstract

Soil aggregates, which are highly influenced by land conversion, play key roles in driving soil nutrient distribution and microbial colonization. However, the role of soil aggregates in shaping the responses of microbial community composition and multiple ecosystem functions, especially ecosystem multifunctionality (EMF), to land conversion remains poorly understood. In this study, we investigated the impact of the conversion of a longan orchard (LO) to a conventional tea plantation (CTP) and organic tea plantation (OTP) on soil EMF at the aggregate level and explored the underlying mechanism. Our results showed that EMF was significantly reduced in the conventional tea plantation, with 3.44, 1.79, and 1.24 times for large macro-, macro-, and micro-aggregates. In contrast, it was relatively preserved in the organic tea plantation. Notably, micro-aggregates with higher microbial biomass supported more EMF than larger aggregates under the land conversion conditions. The EMF associated with soil aggregates was found to be regulated by the differences in nutrient content and microbial community composition. Random forest analysis, redundancy analysis, and Pearson analysis indicated that both soil nutrient and microbial community composition within soil aggregates jointly determined EMF. This study highlights that soil aggregation influences the stratification of nutrients and microbial communities, which leads to the differing response of aggregate-related EMF to land conversion.

1. Introduction

Land conversion has increasingly been adopted as a common practice to meet the growing demand for food production and economic development [1]. Specifically, land conversion can impact a variety of ecosystem functions, including primary productivity, nutrient cycling, and biodiversity [2,3,4]. Soil has the capacity to support multiple ecosystem functions simultaneously, such as soil nutrient storage and cycling, carbon sequestration, resource use efficiency, and organic matter decomposition [5]. Previous studies that focused on individual ecosystem functions have likely underestimated the broader effects of land conversion on the provision of multiple complex and interconnected functional processes [6]. Thus, assessing the response of EMF is crucial for developing effective management strategies to safeguard biogeochemical processes and maintain ecosystem functions after land conversion [7].

Soil consists of various aggregate fractions, which are fundamental components of soil structure [8,9]. These aggregate fractions create heterogeneous micro-habitats for microbial colonization and substrate utilization [10,11,12], which, in turn, influence numerous microbial-driven processes, such as organic matter decomposition and nutrient cycling [13,14,15]. Land conversion has been demonstrated to alter soil aggregate structure and formation by modifying vegetation cover and management practices [16,17,18,19]. Despite the central role that soil aggregates play in maintaining soil functions [20,21], limited information is available on the response of EMF to land conversion at the soil aggregate scale and the mechanisms driving these changes. These knowledge gaps hinder our understanding of how land conversion and management practices influence soil EMF, limiting our ability to predict and manage ecosystem responses effectively.

Recently, studies on EMF have attracted increasing attention, with indicators related to soil nutrient cycling commonly used to evaluate EMF [22]. Many studies have highlighted the critical role of soil microbes in driving EMF [23,24]. For example, a positive linear relationship between microbial diversity and EMF was observed in a 30-year period long-term fertilization experiment, with this relationship being mediated by soil physiochemical properties [25]. Land conversion could directly affect soil microbial community composition and diversity via altering vegetation coverage and shoot biomass and indirectly via influencing soil properties, including soil pH and nutrient availability [4,26,27,28,29]. In addition, soil aggregates are considered key regulators of microbial biodiversity, abundance, and community composition in soils [12,30,31]. Previous studies have shown that land conversion can lead to the intensive turnover of soil aggregates, leading to variations in inorganic C and other nutrients within different soil aggregate fractions due to changes in structural stability [17,32,33,34]. These shifts in the distribution of organic C and other nutrients across aggregate fractions subsequently influence soil microbial communities and, ultimately, EMF [35,36,37]. Therefore, understanding the relationships between EMF and abiotic properties, as well as microbial communities at the soil aggregate level, is vital for uncovering the functional consequences of land conversion.

The conversion of subsistence agriculture crops such as longan, rice, and sugarcane to tea plantations is a common land-use change. It has been widely promoted due to its higher economic value in tropical regions [38]. Tea is an essential cash crop in developing countries and is widely cultivated in China, India, and Kenya. China, being the largest tea producer globally, had a total of 3.4 million hectares of harvested area dedicated to tea plantations by 2022, with the harvested area and yield continuing to expand [39,40]. In addition, organic tea plantations have rapidly increased in China to meet the growing consumer demand for organic products [41]. Organic management practices following land conversion have been increasingly adopted to suppress plant pathogens and improve economic returns. However, there is limited information regarding the response of EMF to these various land conversions. Specifically, how soil aggregate-associated physiochemical properties and microbial communities influence EMF remains unexplored. Understanding these mechanisms is essential for optimizing land-use practices and ensuring sustainable agricultural systems.

Thus, we chose a longan orchard (LO), tea plantations that were converted from longan orchards under different management practices—a conventional tea plantation (CTP) and an organic tea plantation (OTP)—to explore the effects of land conversion with different management regimes on soil EMF at the aggregate level. Furthermore, we aimed to explore the potential associations among soil physiochemical properties, soil microbial communities, and EMF within soil aggregates to uncover the mechanisms driving EMF responses to land conversion. The following hypotheses were proposed: (1) land conversion will exert different influences on soil microbial communities and abiotic factors in soil aggregate fractions; (2) soil EMF within these soil aggregate fractions will shift in response to land conversion, with soil microbial communities and abiotic factors contributing differently to EMF. This study would provide insights into how land conversion influences EMF and offer a foundation for developing management strategies for tea plantations converted from longan orchards.

2. Materials and Methods

2.1. Experimental Site

The experiment site was established in Baisha County, Hainan Province (109°38′ E, 19°18′ N). This area has a tropical monsoon climate. The mean annual temperature and precipitation are 22.7 °C and 1900 mm, respectively. The landscape’s topography is flat, and the soil type was laterites (Oxisols) developed from granite and sandstone.

2.2. Land Uses

The original vegetation in our study area was longan. In recent years, longan orchards have been increasingly converted to tea plantations to meet the demand for economic income. We selected three land uses: original longan orchard (LO), conventional tea plantation (CTP), and organic tea plantation (OTP). The tea plantations were established after clearing the longan orchard in 2014, and they were managed conventionally or organically for 3 years.

In June 2014, three 20 × 20 m squared plots were set up in each land use type with a 35 m interval between them. The three land use type systems were separated by 1 km, and a total of 9 study plots were obtained. In accordance with typical longan cultivation regimes, the longan orchard (LO) was supplied with 360 kg ha⁻¹ as urea, 300 kg P₂O₅ ha^−1, and 240 kg K₂O ha⁻¹ per year, respectively. Chemical fertilizers were applied at rates of 450 kg N ha⁻¹, 225 kg P₂O₅ ha⁻¹, and 225 kg K₂O ha⁻¹ per year by fertigation for the conventional tea plantation (CTP). The pesticides (imidacloprid and deltamethrin) (Bayer CropScience (China) Co., Ltd., Hangzhou, China) and herbicides (glyphosate and paraquat) were applied. For the organic tea plantation, the dry sheep manure was applied at an average rate of 6000 kg ha⁻¹ year⁻¹, and contained 270 g/kg TOC, 8 g/kg N, 6 g/kg P, and 5 g/kg K, respectively. In addition, no pesticides or synthetic fertilizers were adopted from the beginning of the conversion, to comply with the organic farming specifications. According to our investigation, the annual tea yields of CTP and OTP were approximately 108 kg hm⁻² and 92 kg hm⁻², respectively. The organic tea production declined by approximately 7.8% per year.

2.3. Soil Sampling and Aggregate Size Fractionation

On 20 July 2017, five soil cores from each treatment plot were collected from 0–10 cm layers and then mixed into a composite soil sample with four replicates per plot. Samples were sealed in plastic boxes, placed on ice in a constant temperature box, and transported back to the laboratory. The samples were processed by removing the roots and large debris and were then stored. To avoid destroying microbial community structure and functioning, soil aggregate fractions were obtained by the wet sieving method [16,42]. Briefly, a soil aggregate analyzer (DIK-2012, Daiki Rika Kogyo Co., Ltd., Saitama, Japan) with a set of stacking sieves of 2 and 0.25 mm was used, with the fresh soil in the sieves being manually moved up and down by approximately 3 cm in deionized water at a rate of 25 times/min for 5 min. Therefore, three sizes of soil aggregates were manually fractionated into the following size fractions: (1) >2000 μm (large macro-aggregates), (2) 2500–2000 μm (macro-aggregates), and (3) <250 μm fractions (micro-aggregates) as previously described [16,30]. The separated aggregate fractions were divided into three subsamples: the first was air-dried for soil property chemical analysis; the second was stored at 4 °C for extracellular enzyme assays; and the third was immediately stored at −80 °C to determine the microbial phospholipid fatty acids. All analyses were carried out in triplicate.

2.4. Soil Chemical Analysis

The analysis methods of soil chemical properties were described in Lu [43]. The soil pH was determined by a pH meter (PHSJ-3F, Shanghai Inesa Scientific Instruments Co., Ltd., Shanghai, China) at a water-to-soil ratio of 2.5:1. The soil organic carbon (SOC) content was assayed by the oxidation method using K₂Cr₂O₇ andH₂SO₄. The total nitrogen (TN) was measured using the Kjeldahl digestion method with a 2300 Kjeltec Analyzer Unit (Foss Analytical, Hillerød, Denmark). Total phosphorus (TP) was extracted with a H₂SO₄-HClO₄ mixture and quantified by the molybdenum blue method.

2.5. Soil Microbial Community Analysis

The soil microbial community structure was determined by phospholipid fatty acid (PLFA) analysis [44]. Briefly, 3 g of soil aggregate sample was extracted with a chloroform–methanol citrate buffer mixture (volume ratio of 1:2:0.8). The centrifuged supernatant was poured into a separating funnel. The remaining soil in the lower part of the tube was extracted again with the same procedure. The two extracted portions were then mixed together in the separating funnel. The lower layer of the mixing solution was transferred and then dried under N₂ at 30 °C after 24 h. On the solid phase extraction column (silicic acid), neutral lipids, glycolipids, and phospholipids were fractionated by chloroform, acetone, and methanol, respectively. Then, phospholipids were transesterified into fatty-acid methyl esters (FAMEs) with methanolic potassium hydroxide. Methyl nonadecanoate (19:0) was added as the internal standard. The extracted FAMEs were analyzed by an Agilent 7890 Series II gas chromatograph (Agilent Technologies Inc., Santa Clara, CA, USA) with a MIDI Sherlock microbial identification system (MIDI Inc., Newark, DE, USA). Soil microbial community composition was evaluated by PLFA, which indicated different microbial groups. The following biomarkers were used: total PLFA was the sum of all identified PLFAs, fungi (18:1 w9c, 18:2 w6c, and 18:3 w6c), and bacteria (i14:0, i15:0, a15:0, i16:0, a16:0, i17:0, a17:0, i18:0, 16:1 w9c, 16:1 w7c, i17:1 w9c, 17:1 w8c, 18:1 w7c, 18:1 w5c, cy17:0 w7c, cy19:0 w7c, 14:0, 15:0, 15:0 DMA, 16:0, 17:0, 18:0, and 20:0) [45,46,47,48]. PLFA abundance was expressed in nmol g⁻¹ dry soil.

2.6. Soil Extracellular Enzyme Assays

We measured the activities of soil C-(β-1,4-glucosidase, BG), N-(leucine aminopeptidase, LAP, and β-N-acetylglucosaminidase, NAG), and P-cycling enzymes (acid phosphatase, ACP) with standard fluorometric techniques [49]. In brief, soil suspensions were dispensed into 96-well microplates. After incubation in the dark at 20 °C (3 h for BG, NAG, LAP, and 24 h for ACP), the fluorescence was measured using a microplate spectrophotometer (Synergy H1, BioTek Instruments, Inc., Winooski, VT, USA). C-, N-, and P-cycling indices were obtained by an equation to normalize the activity of extracellular enzymes belonging to the same functional group [25]. For example, the N-cycling index was calculated as:

N-cycling index = ²(LAP × NAG)

(1)

A vector analysis (vector length [L, unit less] and vector angle [A, ◦]) of soil EEAs was conducted to evaluate microbial nutrient limitations. Vector length and angle were calculated by the log-transformed ratio, as follows [50]:

Vector length (L) = [Ln(BG)/Ln(NAG + LAP)]² + [Ln(BG)/Ln(ACP)]²

(2)

Vector angle (A) = Degress{ATAN²[(Ln(BG)/Ln(ACP), (Ln(BG)/Ln(NAG + LAP)]}

(3)

A relatively longer vector length indicates a greater C limitation; a vector angle <45° or >45° was considered to be the relative degree of N or P limitation, respectively [51,52].

2.7. Soil Multifunctionality Index

Four function variables related to the C-cycling enzyme (BG), N-cycling enzymes (NAG and LAP), and the P-cycling enzyme (ACP) were selected, which are important indicators of ecosystem functioning. We assessed individual soil functions separately and calculated EMF based on the average approach. Firstly, each of the four soil variables was standardized with z-score transformations. The z-scores were averaged to obtain the EMF index [53].

2.8. Statistical Analysis

The SPSS software (Version 19.0, IBM, Armonk, NY, USA) was adopted to conduct statistical analyses. Significant differences between soil physiochemical properties, microbial biomass, enzyme activities, and EMF among different land use types and aggregate fractions were compared by a one-way analysis of variance (ANOVA) followed by the test of least significant difference (LSD). A difference of p < 0.05 was considered to be significant. Effects of land use type, aggregate size, and their interactions on the enzyme activities and EMF were estimated using a two-way ANOVA, where aggregate size and treatment acted as the factors. A Pearson correlation analysis was conducted to measure the relationship between the changes in the soil EMF, individual nutrient cycling indices, and soil physiochemical and microbial properties caused by land conversion and management practice change within soil aggregates. The CANOCO software (Version 5.0, Biometry, Wageningen, Netherlands) was applied for redundancy analysis (RDA) to determine the relative effects of soil properties and microbial community composition on EMF and individual nutrient cycling indices within soil aggregates. The pivotal and credible predictors of aggregate EMF among different factors were evaluated using a random forest analysis, which was performed using the “Random Forest” package in R.

3. Results

3.1. Soil Aggregate Fractions

In our study, the dominant fraction in the distribution of soil aggregates across the three systems was large macro-aggregates, which accounted for 54.03~56.92%, followed by the macro-aggregates, which accounted for 23.04~35.70%, while micro-aggregates were present in the lowest proportions (Figure 1). There was no significant change in the proportion of large macro-aggregate fractions in both the conventional tea plantation (CTP) and organic tea plantation (OTP) compared to longan orchard (LO). Meanwhile, macro-aggregates had a dramatic increase after the conversion of LO to CTP, by 54.95% (p < 0.05), and there was an opposite trend observed in micro-aggregates. Organic management (OTP), compared to CTP, increased the mass proportion of micro-aggregates, while the mass proportion of macro-aggregates decreased. This suggests that organic management may promote finer soil structure by supporting micro-aggregate formation, whereas conventional practices tend to enhance the proportion of larger aggregates.

3.2. Physicochemical Properties of Soil Aggregates with Different Land Use Types

Land use type, aggregate size, and the interaction between them significantly affected soil physiochemical properties (pL < 0.001; pA < 0.001, and pL × A < 0.001, Table S1). Soil pH in large macro-aggregates and macro-aggregates decreased in the following order: LO > CTP > OTP. Soil pH in micro-aggregates in OTP was higher than that in CTP, although LO had the highest pH among the three systems. In both LO and OTP, soil pH was highest in micro-aggregates, intermediate in large macro-aggregates, and lowest in macro-aggregates (Figure 2a). The conversion from LO to CTP and OTP significantly decreased SOC, TN, and TP across the three aggregate fractions (p < 0.05, Figure 2b–d). Compared with LO, SOC and TP contents within the three aggregate fractions were also decreased in OTP, but the decreased magnitude was much lower in OTP than CTP. In contrast, soil TN contents within the soil aggregate particles were increased with the conversion from LO to OTP across aggregate fractions.

Moreover, in LO and OTP, SOC and TN showed the highest concentrations in macro-aggregates, followed by micro-aggregates, and were the lowest in large macro-aggregates. In CTP, however, SOC and TN decreased with the increasing aggregate size. TP levels were highest in LO and lowest in CTP across three aggregate fractions. These findings suggest that land conversion, especially to CTP, significantly alters the distribution of key nutrients across soil aggregate sizes, while organic management helps to mitigate some of these effects.

3.3. Soil Microbial Communities within Aggregate Fractions

As shown in Figure 3, both CTP and OTP significantly altered the total microbial biomass estimated from total PLFA as well as the biomass of bacteria and fungi within aggregate particles (p < 0.05). Except for the fungi biomass of micro-aggregates, which did not differ significantly from that in LO, the total and bacteria biomass within aggregate fractions tended to decrease significantly in CTP. The reduction in bacteria and fungi biomass in OTP was less than in CTP, especially within micro-aggregates. In addition, both CTP and OTP increased the F/B. The distribution of the total microbial, bacteria, and fungi biomass within aggregate fractions was similar for CTP and OTP, with significantly higher biomass in micro-aggregates than large macro-aggregates, with macro-aggregates being intermediate in value (p < 0.05, Figure 3a–c). Moreover, ANOSIM results showed that the microbial community structure was significantly altered among the three land use types within large macro-aggregates and macro-aggregates (Table S2). However, no remarkable differences in microbial community structure were observed within micro-aggregates.

3.4. Soil Aggregate-Related Nutrient Cycling Indices, Extracellular Enzyme Stoichiometries, and Ecosystem Multifunctionality

Land use type, aggregate size, and their interaction significantly influenced the C-, N-, and P-cycling indices, as well as EMF within soil aggregates (p < 0.001, Table S3). In our study, CTP significantly decreased C-, N-, and P-cycling indices across the three fractions compared to LO (p < 0.05, Figure 4a–c). In contrast, the responses of C-, N-, and P-cycling indices to OTP varied. P-cycling indices were significantly higher in OTP than in LO across all the aggregate fractions, whereas N-cycling indices showed a different trend, with even lower values in OTP compared to CTP (p < 0.05, Figure 4b,c). Specifically, soil C-cycling indices following the conversion from LO to OTP varied with aggregate size, increasing within micro-aggregates, but significantly decreasing in large macro-aggregates and macro-aggregates (Figure 4a). Additionally, the distribution of soil nutrient cycling indices did not change with different land use types and management practices. Soil C- and N-cycling indices increased with decreasing aggregate size across all systems (p < 0.05). Conversely, P-cycling indices were highest within macro-aggregates and lowest within micro-aggregates (p < 0.05).

As a comprehensive indicator of soil functional services, ecosystem multifunctionality (EMF) within soil aggregates was significantly affected by land conversion and management regimes. Except for micro-aggregates in OTP, which did not differ significantly from those in LO, soil EMF across all aggregate fractions was significantly decreased in both CTP and OTP (p < 0.05, Figure 4d). Specifically, the impact of land conversion on EMF was less pronounced under organic management, resulting in relatively higher EMF in soil aggregates from OTP than CTP compared to CTP. CTP decreased the EMF in large macro-aggregates, macro-aggregates, and micro-aggregates by 3.44, 1.79, and 1.24 times, respectively. The EMF of large macro-aggregates and macro-aggregates was decreased by 31.35% and 41.09% by OTP, respectively. Nonetheless, EMF was highest in LO. In both CTP and OTP, EMF reached its peak in micro-aggregates.

In the vector analysis of soil extracellular enzyme stoichiometries, both vector length and angle were significantly affected by land conversion and aggregate size. Compared to LO, the vector length was significantly decreased within large macro-aggregates, but increased within micro-aggregates under CTP (p < 0.05, Figure 5a). Moreover, OTP significantly enhanced the vector length across all aggregate fractions. The vector angles were less than 45° in LO and CTP, showing a significant decreasing trend, which suggests an increase in the microbial N limitation. In contrast, vector angles in OTP were greater than 45°, except in micro aggregates, indicating a shift from microbial N limitation to P limitation (p < 0.05, Figure 5b). Furthermore, vector angle increased with aggregate size, regardless of land use type.

3.5. The Relationship among Soil Ecosystem Multifunctionality, Individual Soil Functions, and Driving Factors within Aggregates

A redundancy analysis (RDA) was conducted to examine the relationships between soil properties, microbial community composition, individual nutrient cycling indices, and EMF within the same aggregate fractions under different land use treatments (Figure 6a–c). The first and second axes of RDA explained 62.46% and 34.51% of the total variation in large macro-aggregates, 66.32% and 32.64% in macro-aggregates, 69.51% and 21.35% in micro-aggregates, respectively. Overall, EMF was positively related to pH, the SOC, TN, TP content, and microbial community, but negatively correlated with the F/B. Moreover, C- and P-cycling indices had a more significant impact on explaining soil EMF compared to N-cycling indices.

The Pearson’s correlation analysis revealed that pH was positively correlated with EMF in macro-aggregates and micro-aggregates (p < 0.05,

Table 1. Pearson correlation analysis between soil EMF, physiochemical properties, microbial community composition, and individual nutrient cycling indices in the soil aggregates. * and ** indicate significant differences at p < 0.05 and p < 0.01, respectively.

Soil Aggregate Size	EMF
	>2000 μm	250–2000 μm	<250 μm
pH	0.459	0.992 **	0.727 **
SOC	0.982 **	0.941 **	0.965 **
TN	0.979 **	0.996 **	0.956 **
TP	0.936 **	0.526	0.938 **
Microbial biomass	0.904 **	0.992 **	0.946 **
F/B	−0.831 **	−0.374	−0.607 *
G+/G−	−0.868 **	−0.606 *	−0.585 *
C-cycling index	0.852 **	0.992 **	0.843 **
N-cycling index	−0.300	0.115	−0.207
P-cycling index	0.741 **	0.765 **	0.847 **
Vector length	0.815 **	0.363	−0.235
Vector angle	0.644 *	0.478	0.779 **

). No significant relationship was found between N-cycling indices and EMF. Additionally, the nutrient content and soil microbial biomass significantly and positively regulated soil EMF, except for TP within macro-aggregates. Meanwhile, in large macro-aggregates, EMF was positively related to both vector length and vector angle, while in micro-aggregates, EMF was positively correlated with vector angle.

To quantitatively evaluate the relative contributions of soil physiochemical properties and microbial community to EMF, we employed a random forest model. In large macro-aggregates, vector length, vector angle, and C-cycling indices were the main drivers for EMF (Figure 7a). In macro-aggregates, SOC, microbial biomass, and C-cycling indices were the main contributors to EMF (Figure 7b). In micro-aggregates, P-cycling indices, SOC, and vector angel were the key drivers of EMF (Figure 7c).

4. Discussion

4.1. Land Conversion Concentrates Microbial Communities within Micro-Aggregates

Considering the key roles of soil microorganisms in biogeochemical processes, exploring the microbial community at the aggregate level is essential for understanding the cycling of microbially affected nutrients in soils [12,54]. In this study, the conversion from a longan orchard to a conventional tea plantation significantly decreased soil microbial biomass, including total microbial, bacterial, and fungal biomass across different aggregate size fractions (Figure 3a–c). This reduction is primarily attributed to physical destruction and soil erosion due to land conversion, which may trigger a loss of soil microbial biomass [55]. Moreover, the shift from high a broad-leaved evergreen longan to a low tea plantation, where leaves are frequently harvested, significantly reduced the litter input, and thus the nutrients available to microorganisms [56].

In contrast, organic management practices enhanced the available nutrient input, benefiting the microbial communities that drive soil functions. We observed that the microbial biomass—total, bacterial, and fungal—was higher in organic tea plantations compared to CTP across all aggregate fractions. Notably, the microbial biomass in micro-aggregates of the organic tea plantation was statistically similar to that in the longan orchard. This can be attributed to the protective role of micro-aggregates, which provide a stable habitat for soil microorganisms [12,57]. The distribution of microbial biomass across aggregate fractions was significantly influenced by land conversion, particularly in large macro-aggregates.

In the longan orchard, microbial biomass across aggregate fractions did not differ statistically, possibly due to the long-term accumulation of organic matter such as plant and litter debris and increased labile SOC, which was evenly distributed across various aggregate sizes [58]. However, soil microorganisms were mainly concentrated within smaller aggregates, with significantly higher microbial biomass within micro-aggregates compared to large macro-aggregates in both conventional and organic tea plantations. This result aligns with previous studies [15,39], suggesting that soil microorganisms within large aggregate fractions are more sensitive to environmental perturbation, likely due to the vulnerability of their microenvironment.

Furthermore, the conversion from longan orchards to tea plantations with different management regimes altered soil structure and the quantity and quality of nutrients within soil aggregates. Land conversion, particularly to CTP, increased fungal biomass within micro-aggregates (Figure 3c). Fungal-dominated communities are known for their superior ability to decompose complex and recalcitrant organic C compared to bacteria-dominated communities [59,60]. Thus, the shift in microbial community composition within micro-aggregates under land conversion conditions accelerated the decomposition of recalcitrant C, thus enhancing microbial biomass associated with soil aggregates. This result is consistent with typical observations showing higher labile C in larger fractions and more recalcitrant C in smaller fractions [59], likely due to differences in soil aggregate structure.

4.2. Land Conversion Changes Nutrient Cycling Indices within Soil Aggregates

In our study, the conversion from longan orchards to both conventional and organic tea plantations significantly altered individual nutrient cycling indices and EMF across different soil aggregates. Among the enzyme activities measured, the conventional tea plantation led to reductions in C-, N-, and P-cycling indices for all soil aggregate fractions. This decrease suggests a lower nutrient turnover rate, likely driven by less metabolically active microbes. The decline in microbial biomass associated with the conversion from longan to a conventional tea plantation supports the observed reductions in C-, N-, and P-cycling indices.

Longan orchards, with their tall evergreen trees, typically have greater litter inputs compared to tea plantations, where leaves are frequently harvested [38]. The removal of residues, combined with the loss of vegetation cover and root biomass caused by intensive land conversion, may also trigger reduced nutrient input to the soil, leading to a decrease in enzyme activities [61].

Conversely, land conversion to an organic tea plantation induced different responses for C-, N-, and P-cycling indices. Organic management alleviated the negative effects of land conversion on aggregate-associated C-cycling indices. Compared to the longan orchard, P-cycling indices were significantly enhanced in the organic tea plantation across all aggregate fractions, whereas N-cycling indices decreased significantly. The decreased N-cycling indices in the organic tea plantation were primarily due to increased N assimilation by heterotrophic microorganisms. As we observed, soil aggregate-associated TN was also enhanced in the organic tea plantation. The alleviation of N limitation in the organic tea plantation may have increased the relative demand for P [62], leading to higher P-cycling indices and P limitation within soil aggregates.

Interestingly, we found that the distribution patterns of C-, N-, and P-cycling indices were not significantly influenced by land conversion. A possible explanation is that aggregate size, with its diverse physiochemical and biological properties, plays a more substantial role in affecting enzyme activities than the land use type or management regime [63].

4.3. Biotic and Abiotic Factors Mediating Soil Aggregate EMF under Land Conversion

As a comprehensive indicator of soil functional services, aggregate-related EMF was also strongly reduced by the conversion from a longan orchard to a conventional tea plantation. As expected, the negative effect of land conversion on EMF was preserved by organic management. The result supports prior findings demonstrating that organic fertilization could improve soil EMF [23].

Ecosystem function is driven by biotic and abiotic factors. Our study found a significant positive relationship between microbial biomass and EMF across all aggregate fractions (Table 1). Soil microorganisms act as decomposers, regulating material cycles and nutrient flows, which in turn affects EMF. The conversion from a longan orchard to a conventional tea plantation led to a great loss of microbial biomass, especially within larger aggregate fractions, leading to a reduction in EMF. This contradicts with previous findings suggesting that macro-aggregates provide more habitats for microbes, facilitating better nutrient transformation. This discrepancy may be due to larger aggregates being more susceptible to disturbance caused by land conversion.

Interestingly, micro-aggregates, which had a higher microbial biomass compared to larger ones in both conventional and organic tea plantations, exhibited the highest EMF. This shift in the highest EMF from macro-aggregates in the longan orchard to micro-aggregates in both types of tea plantations underscores the role of aggregate size in influencing EMF. Our finding that longan orchard soil had the highest EMF within macro-aggregates aligns with the highest distribution of SOC and microbial biomass in LO. This finding is supported by our results showing that macro-aggregate-related EMF was primarily driven by SOC, microbial biomass, and C-cycling indices (Figure 6b).

Under land conversion conditions, micro-aggregates promoted extracellular enzyme production by microorganisms, leading to an enhancement in EMF. In addition, SOC, TN, and TP contents within aggregates were pivotal determinants of EMF. Despite variations in nutrient distribution within aggregate particles due to land conversion, microorganisms adapted to optimize resource utilization. Our observations highlight the crucial role of C-cycling indices in mediating EMF across all aggregate fractions (Figure 6), suggesting the importance of soil C-cycling capabilities for the effective implementation of EMF.

Moreover, the conversion from longan to conventional and organic tea plantations may have changed nutrient availability within different aggregates. The energy (C) and nutrient (N, P) limitations reflected by enzyme stoichiometry vectorization can indicate soil biochemical processes [51]. Our random forest analysis demonstrated that lower microbial P limitation significantly impacted higher EMF in micro-aggregates (Figure 7c). Conversely, severe microbial C limitation contributed to significantly reduced EMF within large macroaggregates (Figure 7a). These results underscore the need for appropriate nutrient supply and agricultural management to support soil carbon sequestration following land conversion from longan orchard to tea plantation.

In this study, we found that converting longan orchard to conventional and organic tea plantations caused shifts in microbial communities and nutrient contents within soil aggregates to varying degrees, which further impacted ecosystem multifunctionality. To better predict EMF under land conversion conditions, other biotic and abiotic factors should also be considered. While we measured four nutrient cycling-related enzymes to characterize EMF, future studies should include additional functions, such as the rate of soil nutrient decomposition, to gain a more comprehensive understanding of EMF.

5. Conclusions

In this study, our results confirmed the negative influence of land conversion from longan orchard to conventional tea plantation on soil aggregate-related EMF. Notably, converting to an organic tea plantation could help mitigate this negative impact. Our findings reveal that EMF within larger aggregates is particularly constrained by land conversion. However, the positive influence of organic management on microbial community composition within micro-aggregate can enhance nutrient cycling and help to maintain EMF.

Furthermore, we observed a significant correlation between EMF and both microbial biomass and nutrient content. In summary, our study suggests that organic management is beneficial for enhancing soil multifunctionality during land conversion. Additionally, we highlight the crucial role of aggregate fraction in regulating soil function.