Skip to main content
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Jul 1;55(4):e70207. doi: 10.1002/jeq2.70207

Macroaggregates dominate colloidal phosphorus sequestration in Ultisol paddies: Evidence from size‐fractionation experiments

Jiamin Zhang 1, Zhuoling Liu 1, Yanling Wang 1,✉, Lei Gao 2,✉, Evgenios Agathokleous 1
PMCID: PMC13489021  PMID: 42383584

Abstract

Soil aggregates serve as critical reservoirs for colloidal phosphorus (Pcoll), governing its sequestration and release, and thereby influencing nutrient cycling and mitigating P loss in agricultural systems. However, the mechanisms controlling Pcoll dynamics across aggregate fractions, particularly macroaggregates (MAC, 2–0.25 mm) and microaggregates (MIC, <0.25 mm) remain poorly understood in Ultisols. This study systematically investigated Pcoll release via water‐dispersible colloids (Crq), Pcoll content, colloidal surface morphology, and stability across varying slope positions, pedogenic horizons, and cultivation histories. Key findings reveal that Crq and Pcoll concentrations increased significantly with soil depth, by approximately 200%–1400% and 25%–67%, respectively, with MAC exhibiting consistently higher Pcoll retention than MIC. Scanning electron microscopy highlights distinct colloidal structures: MAC contain larger, loosely bound particles, whereas MIC display compact, layered arrangements. Colloidal aggregation intensified at lower slope position and in deeper soil horizons, forming denser clusters. In dynamic KH2PO4 dispersion experiments, both MAC and MIC demonstrated strong colloidal mobility, but MAC demonstrate a greater capacity to immobilize Pcoll. These findings underscore the dominant role of MAC in Pcoll sequestration. To enhance MAC abundance in practice, management strategies such as applying organic amendments, reducing tillage intensity, and optimizing water management can be employed, thereby providing a practical pathway for sustainable phosphorus management in rice paddies.

Core Ideas

  • Macroaggregates retain large amounts of colloids and colloidal phosphorus.

  • Colloidal particles within aggregates exhibit strong dispersion and mobility.

  • Increasing MAC proportion can effectively inhibit colloidal phosphorus loss.

Plain Language Summary

This study explains how differently sized soil clumps act as storage units for phosphorus fertilizer in rice fields. We found that larger clumps (2–0.25 mm) are much better at holding onto phosphorus and preventing it from being washed away with water than smaller clumps (<0.25 mm), with more of this mobile phosphorus present deeper in the soil. The structure of these clumps plays a key role: larger clumps have a loose, open arrangement while smaller ones are tightly packed. These particles form denser clusters further down a slope and deeper in the soil profile. Importantly, although both sizes of soil clumps can release particles that move with water, the larger clumps are far superior at trapping and locking phosphorus in place. This means that farming practices that promote the formation of these larger soil clumps could be a key strategy to improve fertilizer efficiency and prevent nutrient pollution.


Abbreviations

CFC

critical flocculation concentration

MAC

macroaggregates

MAC‐Crq

colloid release quantity from macroaggregates

MIC

microaggregates

MIC‐Crq

colloid release quantity from microaggregates

MP‐M

middle‐aged paddy field at the middle slope

MP‐T

middle‐aged paddy field at the top slope

MP‐U

middle‐aged paddy field at the upper slope

OP‐B

old paddy field at the bottom slope

RMV

relative mobility value

SOC

soil organic carbon

1. INTRODUCTION

Phosphorus (P) is indispensable for agricultural productivity yet presents significant management challenges due to limited global reserves and high environmental mobility, which can lead to the degradation of aquatic ecosystems (F. Li et al., 2020). As fundamental architectural units of soil systems, aggregates play a pivotal role in regulating ecological functions, including nutrient cycling, carbon dynamics, and environmental stress resilience (Laskar et al., 2021). Importantly, macroaggregates (MAC, 2–0.25 mm) and microaggregates (MIC, <0.25 mm) function as vital P reservoirs within soil systems, with their formation and turnover dynamics fundamentally controlling P distribution patterns, bioavailability, and loss potential (Ahmed et al., 2017). Previous studies have suggested that MAC tend to inhibit P mineralization and promote P accumulation, whereas MIC are more prone to erosion and may enhance P transport (Cui et al., 2019; Ge et al., 2017). However, the sequestration and transport mechanisms of the more mobile Pcoll fraction within and between these aggregate classes remain unclear. Furthermore, the size‐dependent distribution of P is strongly influenced by adsorption–precipitation interactions that vary according to aggregate composition (D. Liu et al., 2021; Wu et al., 2021).

The breakdown of soil aggregates releases colloidal particles (<1 µm) into the soil solution, leading to significant alterations in P dynamics (Cao et al., 2022). Due to their large surface area and high charge density, soil colloids can act as important carriers of P, influencing its transport and environmental redistribution (F. Li et al., 2020; B. Li et al., 2022). Notably, when suspended in aqueous environments, soil colloidal particles may either associate with aggregate surface to form relatively stable aggregates, or alternatively undergo dispersion and subsequent migration following aggregate fragmentation (Henderson et al., 2012; Seta & Karathanasis, 1996). Classical Derjaguin–Landau–Verwey–Overbeek theory attributes colloidal stability to the balance of electrostatic and van der Waals forces, which are highly influenced by the local chemical environment within soil aggregates (Ding et al., 2017). Critically, the distinct physicochemical microenvironments of MAC and MIC, including their organic matter composition, mineralogy, and porosity, likely create divergent conditions for colloidal stabilization or dispersion. Consequently, the mechanism by which these aggregate‐specific microenvironments determine the divergent Pcoll sequestration and release capacities of MAC and MIC remains a key knowledge gap.

Despite growing recognition of the role of soil aggregates in P dynamics, current understanding remains largely limited to bulk P fractions, with insufficient attention to the most mobile and reactive fraction, that is, Pcoll. In particular, how aggregate‐scale microenvironments regulate the sequestration, release, and transport of Pcoll remains poorly constrained. Existing studies have primarily focused on either aggregate size effects or colloid behavior in isolation, lacking an integrated framework that links aggregate structure, pedogenic heterogeneity, and colloidal P dynamics. This knowledge gap is especially critical in paddy Ultisols, where strong redox fluctuations, pedogenic horizon differentiation, long‐term cultivation history, and slope position jointly regulate aggregate stability and colloid mobilization (Xiao et al., 2017; J. Zhang et al., 2025; Zheng et al., 2019). However, it remains unclear how these factors jointly regulate Pcoll distribution and mobility across aggregate fractions and soil horizons, particularly along slope gradient.

To address these gaps, our study investigated Pcoll dynamics in paddy soils across a toposequence within the Red Soil Critical Zone Observatory (Sunjia small watershed in Jiangxi Province, China). We systematically analyzed MAC (2–0.25 mm) and MIC (<0.25 mm) collected from four slope positions (top, upper, middle, and bottom) representing distinct pedogenic horizons and cultivation histories. Three key hypotheses were tested: (1) MAC would exhibit greater colloidal release capacity and P retention than MIC due to their structural and compositional differences, (2) aggregate‐derived colloids would demonstrate high environmental mobility, and (3) slope position and associated pedogenic horizon properties (as influenced by long‐term cultivation history) would significantly influence colloidal stability and P enrichment patterns, particularly in the historically cultivated bottom‐slope paddy. By addressing these hypotheses, this work provides new mechanistic insight into P cycling in paddy soils, offering a scientific basis for sustainable P management and loss mitigation strategies.

2. MATERIALS AND METHODS

2.1. Soil sample collection and preparation

The study was conducted in the Sunjia watershed (28°14′ N, 116°54′ E), Jiangxi province, China, a region characterized by a humid subtropical climate and Ultisols derived from Quaternary Red Clay (Soil Survey Staff, 2010). The mean annual temperature is approximately 18°C, with an annual precipitation of 1795 mm. Land uses include peanut upland, paddy fields, orange orchards, and forested areas. Paddy fields were double‐cropped with rice, planted in early April and late July. Fertilization practices involved compound fertilizer (N:P2O5:K2O = 15:15:15) at 50–75 kg ha−1, urea (46.67% N) at 25–40 kg ha−1, and potassium fertilizer (46.67% K2O) at 5–10 kg ha−1 in April, followed by a second application compound fertilizer (10–25 kg ha−1) and urea (10–35 kg ha−1) in July.

To ensure robust statistical analysis, four paddy fields with varying cultivation durations were selected in November 2020, each representing distinct slope positions: middle‐aged paddy field at the top slope (MP‐T, 50–60 years of cultivation), middle‐aged paddy field at the upper slope (MP‐U, 50–60 years), middle‐aged paddy field at the middle slope (MP‐M, 50–60 years), and old paddy field at the bottom slope (OP‐B, > 500 years) (Figure 1). To account for field variability, three replicate soil profiles (1.5 m × 1 m × 1.0–1.2 m depth) were excavated per field following FAO (2006) guidelines. Soil horizons were delineated based on texture, color, and compactness (Figure 1). Composite samples from each horizon were air‐dried, sieved (<2 mm), and stored. The Ap (0–20 cm) and C (85–110 cm) horizons were selected for detailed analysis because they represent the two most pedogenically distinct layers. Their strong contrasts in key properties (e.g., compactness, organic matter, and Fe/Al oxide content), which are known to govern colloidal behavior and aggregate formation, provided the basis for hypothesizing divergent Pcoll dynamics and aggregate size distributions.

FIGURE 1.

FIGURE 1

Schematic diagram of soil profiles at different slope positions and summary of site characteristics. MP‐M, middle‐aged paddy field at the middle slope; MP‐T, middle‐aged paddy field at the top of slope; MP‐U, middle‐aged paddy field at the upper slope; OP‐B, old paddy field at the bottom slope.

2.2. Extraction of water‐dispersed colloids from soil aggregates

MAC (2–0.25 mm) and MIC (<0.25 mm) were isolated via wet sieving (Fan et al., 2023; Yoder, 1936). The isolated aggregates were dried at 50°C to constant weight (±0.01 g over 24 h) to minimize thermal alternation of colloidal surfaces and organic matter, following protocols for studies focusing on colloidal and aggregate‐associated properties (Elliott, 1986). For colloid extraction, 5.00 g of unground aggregates were dispersed in 100 mL of deionized water using ultrasonication (20 kHz, 300 W, 20 min, and 25°C). The suspension was filtered (<0.053 mm), and colloids (<1 µm) were extracted via pipetting (Stoke's law), dried at 60°C, and weighed. Colloids release from MAC (MAC‐Crq) and MIC (MIC‐Crq) was calculated as (Equation 1):

Crq(gg−1soil−1)=M1M2×WSA% (1)

where M1 = colloid weight, M2 = soil mass (5.00 g), and WSA% = aggregate proportion in soil.

Core Ideas

  • Macroaggregates retain large amounts of colloids and colloidal phosphorus.

  • Colloidal particles within aggregates exhibit strong dispersion and mobility.

  • Increasing MAC proportion can effectively inhibit colloidal phosphorus loss.

2.3. Micromorphological and elemental analysis of colloids

Colloid microstructure and elemental composition were analyzed using SEM‐EDS (scanning electron microscope—energy dispersive spectrometer) (Gemini SEM 300). Before scanning electron microscope imaging, freeze‐dried soil colloids were fixed on aluminum stubs with conductive carbon tape and sputter‐coated with gold. Imaging was performed at 10 kV and 5000x magnification. Energy‐dispersive x‑ray spectroscopy (EDS) line scan analysis was conducted at 10 kV with a live time of 60 s per spectrum. Semiquantitative elemental composition (mass percentage) was obtained using Bruker ESPRIT software with standardless ZAF correction. Reported values are rounded to integer percentages, with absolute error (3σ, 99.7% confidence) given to one decimal place.

2.4. Mobility of soil colloids

Colloid mobility was quantitatively assessed using the methylene blue test paper method (M. Zhang, 1993). Specifically, 5.00 mg of colloidal sample was precisely weighed into a 100‐mL beaker and subsequently mixed with 50 mL of ultrapure water. A pre‐stained filter paper strip (4 × 40 cm) was vertically suspended above the beaker with its lower end immersed in ∼1 cm into the liquid phase. Following a 4‐h equilibration period at ambient temperature (25 ± 1°C) to allow complete capillary migration, the vertical migration heights of both the colloidal suspension (h) and ultrapure water control (H) were measured to the nearest 0.1 mm using a calibrated ruler. The relative mobility value (RMV) was then calculated as the ratio h/H with results classified into four distinct mobility categories: highly mobile (RMV > 0.85), moderately mobile (0.85 > RMV > 0.5), slightly mobile (0.5 > RMV > 0.25), and immobile (RMV < 0.25). All measurements were performed in triplicate to ensure methodological reproducibility.

2.5. Critical flocculation concentration (CFC) determination

The CFC of soil colloids was determined following Shang et al. (2010) with modifications. Six 5.00 mg colloidal aliquots were dispensed into 25‐mL colorimetric tubes, each receiving 10 mL of KH2PO4 solution at concentrations of 0, 20, 40, 60, 80, and 100 mg L−1. After dilution to mark with ultrapure water and 10x inversion mixing, suspensions were equilibrated for 3 h. Absorbance measurements were conducted at 660 nm (Ai ) using a UV‐Vis spectrophotometer, with ultrapure water used as the instrumental blank. A 0 was defined as the absorbance of the colloidal suspension without added KH2PO4, representing the fully dispersed reference state. The relative dispersion concentration (Ai /A 0) was plotted against the corresponding phosphate concentration to generate the flocculation profile. Complete colloidal dispersion was indicated by Ai /A 0 > 1.0, while significant flocculation was defined as Ai /A 0 < 0.2. The CFC was operationally defined as the minimum phosphate concentration required to achieve Ai /A 0 = 0.2, with lower CFC values denoting decreased colloidal stability in accordance with established criteria (Q. Y. Hu et al., 1996; H. M. Li et al., 2017). All experimental procedures were performed in triplicate under controlled laboratory conditions (25 ± 0.5°C).

2.6. Soil physicochemical properties analysis

Soil pH was determined potentiometrically in 1 mol L−1 KCl (2.5:1 water:soil). Soil bulk density was measured by the ring knife method. Soil organic carbon (SOC) was measured by the high‐temperature potassium dichromate volumetric method. Total phosphorus and colloidal phosphorus (Pcoll) were analyzed using HClO4–H2SO4 digestion followed by molybdenum–antimony colorimetry. Free iron—aluminum oxides (Fed, Ald) were extracted by dithionite‐citrate‐bicarbonate, while amorphous iron–aluminum oxides (Feo, Alo) were extracted with acidic ammonium oxalate (pH 3.17). Concentrations were measure by ICP‐OES (Lu, 2000). Key properties of Ap and C horizons are summarized in Table 1.

TABLE 1.

Basic physicochemical properties of Ap and C layer soil in paddy fields at different slope positions.

Profile Horizon

BD

(g cm−3)

pHKCl

SOC

(g kg−1)

TP

(g kg−1)

Feo

(g kg−1)

Alo

(g kg−1)

Fed

(g kg−1)

Ald

(g kg−1)

MAC proportion

(%)

MIC proportion

(%)

MP‐T Ap 0.76Db 3.97Ba 30.3Ba 0.78Ba 4.63Ab 4.64Ab 24.4Ab 10.3Ab 11.2Bb 7.80Bb
C 1.20Ba 3.81Db 6.04Ab 0.57Bb 8.38Aa 7.71Aa 63.5Aa 19.2Ba 70.0Aa 15.4Ca
MP‐U Ap 0.82Cb 4.03Aa 28.1Ca 0.71Ca 2.96Cb 4.5ABb 22.8Bb 10.1Ab 10.7Bb 9.07Bb
C 1.27Aa 4.06Ca 5.44Bb 0.68Ab 7.37Ba 7.36Ba 58.2Ba 18.9Ca 52.1Ca 16.0Ca
MP‐M Ap 0.88Bb 3.91Cb 32.9Aa 0.84Aa 3.68Bb 4.42Bb 18.1Cb 7.68Bb 8.40Cb 3.92Cb
C 1.17Ca 4.10Ba 4.66Cb 0.56Cb 5.98Ca 7.68Aa 60.4Aa 21.0Aa 45.3 Da 22.3Ba
OP‐B Ap 0.90Ab 3.97Bb 30.0Ba 0.56 Da 2.74Cb 3.41Ca 15.0Db 6.23Cb 24.7Ab 31.0Aa
C 1.20Ba 4.34Aa 3.97Db 0.37Db 3.26 Da 3.41Ca 49.6Ca 15.2 Da 62.9Ba 26.8Ab

Note: The same uppercase letter means no significant difference in the same pedogenic horizon of different profiles, and the same lowercase letter means no significant difference between different pedogenic horizons in the same profile.

Abbreviations: Ald, Free alumina oxide; Alo, amorphous alumina oxide; BD, bulk density; Feo, amorphous iron oxide; Fed, Free iron oxide; MAC, macroaggregates; MIC, microaggregates; MP‐M, middle‐aged paddy field at the middle slope; MP‐T, middle‐aged paddy field at the top of slope; MP‐U, middle‐aged paddy field at the upper slope; OP‐B, old paddy field at the bottom slope; SOC, soil organic carbon; TP, total phosphorus.

2.7. Statistical analysis

Data processing utilized Excel 2019, with statistical analyses conducted in SPSS 24 (IBM). Normality (Shapiro–Wilk) and homoscedasticity (Levene's test) were verified prior to analysis. Significant differences among treatments were determined by one‐way analysis of variance with Duncan's post hoc test (α = 0.05). Values represent mean ± SD (n = 3).

3. RESULTS

3.1. Colloid release dynamics in Ultisol

The C horizon exhibited significantly higher colloid release quantities (Crq) from both MAC (MAC‐Crq: 0.14–0.26 g g−1 soil−1) and MIC (MIC‐Crq: 0.06–0.09 g g−1 soil−1) compared to the Ap horizon (MAC‐Crq: 0.01–0.03 g g−1 soil−1 and MIC‐Crq: 0.004–0.03 g g−1 soil−1). Both aggregate fractions displayed a consistent concave trend with slope position, initially decreasing before increasing downslope. Notably, MAC‐Crq predominated across all profiles except OP‐B, while C horizon MIC‐Crq showed significant downslope enrichment (Figure 2).

FIGURE 2.

FIGURE 2

Colloid release quantity from macroaggregates (MAC) and microaggregates (MIC) in the Ap and C horizon soils. The same uppercase letter indicates that the difference between different profiles of the same pedogenic horizon and the same grain grade is not significant, the same lowercase letter indicates that the difference between different grain grades of the same pedogenic horizon and the same profile is not significant (p < 0.05). The data are shown as the means ± SD (n = 3). Crq, colloid release quantity; MP‐M, middle‐aged paddy field at the middle slope; MP‐T, middle‐aged paddy field at the top of slope; MP‐U, middle‐aged paddy field at the upper slope; OP‐B, old paddy field at the bottom slope.

3.2. Colloidal characteristics

SEM (10 kV and 5000x) revealed clear morphological differences between soil horizons and colloidal fractions (Figure 3). In the Ap horizon, MAC colloids were dominated by relatively larger and more compact particles with smoother surfaces and closer inter‐particle contacts, indicating a higher degree of aggregation. In contrast, MIC colloids exhibited more platy and layered structures with increased surface roughness and visible interstitial spaces, suggesting comparatively looser packing.

FIGURE 3.

FIGURE 3

Morphological characteristics of colloidal particles in macroaggregates (MAC) and microaggregates (MIC) of the Ap and C horizons in paddy fields at different slope position. MP‐M, middle‐aged paddy field at the middle slope; MP‐T, middle‐aged paddy field at the top of slope; MP‐U, middle‐aged paddy field at the upper slope; OP‐B, old paddy field at the bottom slope.

In the C horizon, both MAC and MIC fractions were characterized by smaller particle sizes and a higher proportion of fragmented units. The particles appeared more irregular in shape and more densely packed, forming clusters with reduced inter‐aggregate spacing. Compared to the Ap horizon, the C horizon showed a higher degree of particle congregation but a lower extent of structural integrity, as reflected by the increased abundance of fine fragments and reduced occurrence of large, coherent aggregates (Figure 3).

EDS analysis identified oxygen (20%–50%) and silicon (19%–31%) as the dominant elements. P distribution varied with pedogenic horizon and aggregate class: In Ap horizons, MIC colloids generally contained more P than MAC colloids, except for OP‐B. In C horizons, the pattern was less consistent; MAC colloids showed higher P at the MP‐T and MP‐M, similar levels were observed in MP‐U, whereas MIC colloids exhibited substantially higher P in OP‐B (Table 2). Notably, MIC colloids in the C horizon of OP‐B showed reduced oxygen content accompanied by increased P.

TABLE 2.

Elemental composition of colloidal particles in macroaggregates (MAC) and microaggregates (MIC) of the Ap and C horizons in paddy fields at different slope position.

Profile Pedogenic horizon Element composition (%)
O Si P Al Fe K
MP‐T Ap‐MAC 47 ± 6.5 25 ± 1.2 5 ± 0.3 16 ± 0.9 4 ± 0.5 2 ± 0.2
Ap‐MIC 48 ± 5.5 24 ± 0.9 9 ± 0.4 13 ± 0.6 4 ± 0.5 1 ± 0.2
C‐MAC 33 ± 5.7 31 ± 1.4 8 ± 0.5 18 ± 1.0 8 ± 1.0 2 ± 0.3
C‐MIC 46 ± 7.3 23 ± 1.2 5 ± 0.4 15 ± 0.9 7 ± 0.9 2 ± 0.3
MP‐U Ap‐MAC 46 ± 8.2 23 ± 1.3 4 ± 0.3 14 ± 0.9 7 ± 1.0 2 ± 0.2
Ap‐MIC 47 ± 8.5 20 ± 1.1 4 ± 0.3 12 ± 0.8 4 ± 0.8 1 ± 0.2
C‐MAC 50 ± 6.4 21 ± 0.9 6 ± 0.4 14 ± 0.7 6 ± 0.7 2 ± 0.2
C‐MIC 46 ± 7.1 24 ± 1.2 6 ± 0.4 15 ± 0.8 6 ± 0.7 1 ± 0.2
MP‐M Ap‐MAC 45 ± 4.6 22 ± 0.8 8 ± 0.3 14 ± 0.5 6 ± 0.5 2 ± 0.2
Ap‐MIC 43 ± 5.1 22 ± 0.9 9 ± 0.4 14 ± 0.6 3 ± 0.4 1 ± 0.2
C‐MAC 44 ± 3.9 22 ± 0.7 10 ± 0.4 14 ± 0.5 7 ± 0.4 2 ± 0.2
C‐MIC 45 ± 5.8 24 ± 1.0 6 ± 0.3 15 ± 0.7 7 ± 0.6 2 ± 0.2
OP‐B Ap‐MAC 47 ± 7.3 23 ± 1.2 6 ± 0.4 15 ± 0.9 6 ± 0.8 1 ± 0.2
Ap‐MIC 40 ± 5.4 28 ± 1.1 7 ± 0.4 15 ± 0.7 8 ± 0.8 2 ± 0.2
C‐MAC 38 ± 5.0 19 ± 0.8 10 ± 0.5 11 ± 0.5 5 ± 0.6 1 ± 0.2
C‐MIC 20 ± 2.1 23 ± 0.6 22 ± 0.6 11 ± 0.4 11 ± 0.7 2 ± 0.2

Abbreviations: MP‐M, middle‐aged paddy field at the middle slope; MP‐T, middle‐aged paddy field at the top of slope; MP‐U, middle‐aged paddy field at the upper slope; OP‐B, old paddy field at the bottom slope.

3.3. Colloidal stability and mobility characteristics

The stability of soil colloids was quantitatively evaluated through CFC analysis, where higher values indicate greater stability (Q. Y. Hu et al., 1996). In the KH2PO4 dispersion system (0–100 mg L−1), both MAC and MIC colloids exhibited stable dispersion, with Ai /A 0 values consistently ranging from 0.8 to 1.2. No significant differences were observed between aggregate fractions (p > 0.05). Complete dispersion (Ai /A 0 > 1) was observed for MAC colloids in MP‐U and MP‐M C horizons at 20–40 mg L−1 KH2PO4, and for MIC colloids in MP‐T and OP‐B C horizons across 20–100 mg L−1 KH2PO4 (Figure 4). Colloidal mobility, assessed through relative mobility value (RMV), showed a consistent nonlinear response to KH2PO4 concentration (decrease–increase–decrease) in both Ap and C horizons, though no significant MAC–MIC differences were detected (Figure 5).

FIGURE 4.

FIGURE 4

Stability changes of soil colloidal particles in macroaggregates (MAC) and microaggregates (MIC) of Ap and C horizons paddy fields at different slope positions. The data are shown as the means ± SD (n = 3). MP‐M, middle‐aged paddy field at the middle slope; MP‐T, middle‐aged paddy field at the top of slope; MP‐U, middle‐aged paddy field at the upper slope; OP‐B, old paddy field at the bottom slope.

FIGURE 5.

FIGURE 5

Mobility changes of colloidal particles in macroaggregates (MAC) and microaggregates (MIC) of Ap and C horizons paddy fields at different slope positions. The data are shown as the means ± SD (n = 3). MP‐M, middle‐aged paddy field at the middle slope; MP‐T, middle‐aged paddy field at the top of slope; MP‐U, middle‐aged paddy field at the upper slope; OP‐B, old paddy field at the bottom slope; RMV, relative mobility value.

P release dynamics revealed distinct patterns between aggregate fractions. In Ap horizons, Pcoll content ranged 0.15–0.36 g kg−1 (MAC) and 0.07–0.41 g kg−1 (MIC), while C horizons showed 0.25–0.45 g kg−1 (MAC) and 0.10–0.15 g kg−1 (MIC). Both fractions exhibited initial decreases followed by downslope increases in Pcoll content. Notably, MAC generally contained higher Pcoll than MIC, except in OP‐B Ap horizon. The C horizon MIC showed particularly strong downslope accumulation (Figure 6).

FIGURE 6.

FIGURE 6

Changes in colloid phosphorus content in macroaggregates (MAC) and microaggregates (MIC) of Ap and C horizons paddy fields at different slope positions. The same uppercase letter indicates that the difference between different profiles of the same pedogenic horizon and the same grain grade is not significant, the same lowercase letter indicates that the difference between different grain grades of the same pedogenic horizon and the same profile is not significant. The data are shown as the means ± SD (n = 3). MP‐M, middle‐aged paddy field at the middle slope; MP‐T, middle‐aged paddy field at the top of slope; MP‐U, middle‐aged paddy field at the upper slope; OP‐B, old paddy field at the bottom slope.

4. DISCUSSION

4.1. Aggregate size‐dependent colloidal release

The differential colloidal release between MAC and MIC significantly influences P retention dynamics (F. Li, 2021). Our findings demonstrate consistently higher colloidal release in MAC than MIC across both Ap and C horizons (except in the Ap horizon of OP‐B field; Figure 2). This is attributable to MAC's larger pore networks and stronger internal binding forces arising from microaggregate amalgamation (D. Liu et al., 2021; Tiwari et al., 2022). The C horizon exhibited more pronounced MAC–MIC differences (Figure 2), likely due to oxidative conditions promoting colloidal dispersion (P. Hu et al., 2023). Given that colloids (<1 µm) constitute ≥80% of soil surface charge and specific surface area (Tang et al., 2015), their migration critically governs nutrient and pollutant transport (Flury & Qiu, 2008). Enhanced colloidal release at lower slope position (Figure 2) aligns with reduced C horizon stability from elevated Fe/Al oxides and diminished SOC (Table 1; Y. Zhao et al., 2015).

4.2. Redox‐driven colloidal behavior

SEM analysis revealed distinct colloidal morphologies between horizons (Figure 3), reflecting redox‐mediated aggregation processes (G.‐L. Zhang & Gong, 2003). Waterlogging in the Ap horizon induces reducing condition, whereas oxidation in the C horizon promotes tight aggregation (Y. Liu et al., 2023). The downslope aggregation trend (Figure 3) matches the pattern of SOC enrichment on lower slopes, a factor known to stabilize colloids (Barton & Karathanasis, 2003).

4.3. Iron–aluminum oxides mediated P retention

EDS data confirmed higher P, Fe, and Al in MAC colloids than MIC (Table 2). This difference is likely due to the enrichment of low‐crystalline Fe/Al oxides, which possess higher specific surface area and reactivity (H. Zhao et al., 2026). These oxides not only serve as key components of organo‐mineral colloids but also provide abundant reaction sites for phosphate adsorption via ligand exchange and metal bridging (J. Li et al., 2025). In the C horizon of OP‐B, anaerobic conditions are likely to develop under prolonged flooding, as commonly reported for paddy soils (J. Zhang et al., 2014). Such conditions may promote the reductive dissolution of Fe3+, thereby weakening colloid‐aggregate bonds and facilitating P release (Henderson et al., 2012).

4.4. Colloidal stability mechanisms

The stability and dispersion behavior of soil colloidal particle are fundamentally governed by multiple factors, including organic matter content, soil pH, ion types, and iron–aluminum oxides content (Chen et al., 2020; Igwe et al., 2009). In KH2PO4 − dispersion systems, we observed that colloidal particles interact electrostatically with H2PO4 − ions, with increasing phosphate concentration enhancing electric field shielding capacity (Zhu, 2018). Our measurements of relative dispersion concentration (Ai /A 0) revealed stable values ranging from 0.8 to 1.2 across all slope position in both MAC and MIC fractions (Figure 4). The minimal variation in Ai /A 0 values between aggregate sizes at equivalent KH2PO4 concentrations suggests complete saturation of phosphorus adsorption sites on colloidal surfaces (Zhou, 2022), establishing an equilibrium state that prevents further surface charge modification with additional P inputs.

4.5. Ionic strength effects on colloidal behavior

Soil colloids demonstrate remarkable adsorption capacity due to their extensive surface area and unique double electrical layer structure (Sen & Khilar, 2006). In our experimental systems (0–100 mg L−1 KH2PO4), the RMV consistently exceeded 0.85 for both MAC and MIC across soil horizons (Figure 5). However, at the highest concentration (100 mg L−1 H2PO4 −), we observed a significant RMV decline resulting from increased ionic strength. This phenomenon can be attributed to compression of the double electric layer and consequent elevation of surface potential, ultimately reducing colloidal stability and migration capacity (J. Hu et al., 2009).

4.6. Aggregate‐scale phosphorus distribution patterns

Consistent with previous reports (F. Li et al., 2020), our data demonstrate significantly higher Pcoll content MAC compared to MIC across both Ap and C horizons (Figure 6), reflecting the strong binding affinity between Pcoll and Fe/Al oxides in larger aggregates (Cui et al., 2019). Notably, C horizon MAC in MP‐T, MP‐U, and MP‐M soils contained greater Pcoll than corresponding Ap horizon MAC (Figure 6), likely due to differences in P adsorption dynamics, mineral composition, and crystallization processes. The particularly high adsorption affinity of iron and aluminum oxides (Hens & Merckx, 2001; Montalvo et al., 2015; Tavakkoli et al., 2013), combined with their elevated concentration in C horizon (Table 1), explains this distribution pattern. In contrast, long‐term cultivation appears to destabilize Ap horizon MAC while MIC maintain stronger P enrichment through size‐dependent deposition advantages.

4.7. Implications for P management and future research

The results of this study provide important implications for P management in paddy Ultisols. The dominant role of MAC in Pcoll sequestration suggests that enhancing MAC formation can effectively reduce Pcoll loss while improving P retention. Agricultural practices that promote aggregate stability, such as applying organic amendments (e.g., crop residues and compost), reduced tillage intensity (e.g., conservation tillage), and optimized water management (e.g., controlled irrigation), can serve as practical strategies to regulate P mobility. In addition, the sensitivity of Pcoll dynamics to slope position, pedogenic horizon, and long‐term cultivation history highlights the need for site‐specific management approaches. For instance, bottom‐slope paddies with prolonged cultivation may require targeted strategies, such as reducing P fertilizer input or incorporating iron‐rich amendments, to mitigate P accumulation and potential release under fluctuating redox conditions.

Future research should further explore redox‐sensitive P partitioning between aggregate fractions and evaluate the generality of these mechanisms across different paddy systems. Integrating aggregate‐scale processes with field‐scale variability will be essential for developing more effective and sustainable P management frameworks.

5. CONCLUSIONS

This study systematically investigated Pcoll dynamics in paddy soil aggregates and yielded three key findings. First, MAC released more Pcoll than MIC yet exhibited stronger retention capacity via Fe/Al oxide binding, particularly in C horizons. Second, both aggregate fractions showed enhanced colloidal mobility under KH2PO4 dispersion, with stability maintained until reaching critical ionic strength threshold. Third, strategic management to promote MAC formation, such as organic amendments, reduced tillage, and optimized water regimes, can simultaneously reduce P loss and improve fertility, offering a sustainable approach for sloping paddy systems. Collectively, these findings provide mechanistic insights for optimizing P management through aggregate‐based strategies in rice cultivation.

AUTHOR CONTRIBUTIONS

Jiamin Zhang: Conceptualization; data curation; investigation; methodology; visualization; writing—original draft; writing—review and editing. Zhuoling Liu: Conceptualization; investigation; methodology. Yanling Wang: Funding acquisition; writing—review and editing. Lei Gao: Funding acquisition; writing—review and editing. Evgenios Agathokleous: Supervision; writing—review and editing.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China (Nos. 42077087 and 42477323), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA0440202), and the National Natural Science Foundation of China Talent Grant (No. W2532031).

Zhang, J. , Liu, Z. , Wang, Y. , Gao, L. , & Agathokleous, E. (2026). Macroaggregates dominate colloidal phosphorus sequestration in Ultisol paddies: evidence from size‐fractionation experiments. Journal of Environmental Quality, 55, e70207. 10.1002/jeq2.70207

Assigned to Associate Editor Tyler Groh.

Contributor Information

Yanling Wang, Email: ylwang@nuist.edu.cn.

Lei Gao, Email: lgao@issas.ac.cn.

DATA AVAILABILITY STATEMENT

The datasets used during the current study are available from the corresponding author upon reasonable request.

REFERENCES

  1. Ahmed, E. H. , Anjum, S. I. , & Zhang, M. K. (2017). Effects of fertilization on phosphorus distribution in water‐stable aggregates of soils with different properties. Toxicological & Environmental Chemistry, 99(1), 32–47. 10.1080/02772248.2016.1172584 [DOI] [Google Scholar]
  2. Barton, C. D. , & Karathanasis, A. D. (2003). Influence of soil colloids on the migration of atrazine and zinc through large soil monoliths. Water, Air, and Soil Pollution, 143(1), 3–21. 10.1023/A:1022886225564 [DOI] [Google Scholar]
  3. Cao, G. , Chang, B. , Zhou, Z. , Hu, L. , Du, W. , & Lv, J. (2022). Soil Aggregate Breakdown with Colloidal Particles Release and Transport in Soil: A Perspective from Column Experiments. Agriculture, 12(12), 2155. 10.3390/agriculture12122155 [DOI] [Google Scholar]
  4. Chen, Y. , Ma, J. , Wu, X. , Weng, L. , & Li, Y. (2020). Sedimentation and transport of different soil colloids: Effects of goethite and humic acid. Water, 12(4), 980. 10.3390/w12040980 [DOI] [Google Scholar]
  5. Cui, H. , Ou, Y. , Wang, L. , Wu, H. , Yan, B. , & Li, Y. (2019). Distribution and release of phosphorus fractions associated with soil aggregate structure in restored wetlands. Chemosphere, 223, 319–329. 10.1016/j.chemosphere.2019.02.046 [DOI] [PubMed] [Google Scholar]
  6. Ding, W. Q. , He, J. H. , Wang, L. , Liu, X. M. , & Li, H. (2017). Effect of nonclassical polarization of Na+ and K+ on the stability of soil colloidal particles in suspension. Surface Review and Letters, 24(07), 1850019. 10.1142/S0218625x18500191 [DOI] [Google Scholar]
  7. Elliott, E. (1986). Aggregate structure and carbon, nitrogen, and phosphorus in native and cultivated soils. Soil Science Society of America Journal, 50(3), 627–633. 10.2136/sssaj1986.03615995005000030017x [DOI] [Google Scholar]
  8. Fan, H. L. , Zhang, J. M. , Li, H. , & Wang, Y. L. (2023). Study on phosphorus storage pattern and loss risk in paddy section of slope cultivated land. Ecology and Environmental Sciences, 32(02), 283–291. 10.16258/j.cnki.1674-5906.2023.02.008 [DOI] [Google Scholar]
  9. FAO . (2006). Guidelines for soil description (4th ed.). FAO. [Google Scholar]
  10. Flury, M. , & Qiu, H. (2008). Modeling colloid‐facilitated contaminant transport in the vadose zone. Vadose Zone Journal, 7(2), 682–697. 10.2136/vzj2007.0066 [DOI] [Google Scholar]
  11. Ge, N. N. , Shi, Y. , Yang, X. L. , Zhang, Q. Y. , Li, X. Z. , Jia, X. X. , & Wei, X. R. (2017). Distribution characteristics of soil carbon, nitrogen, phosphorus and aggregates in farmland with different soil textures on the Loess Plateau. Chinese Journal of Applied Ecology, 28(5), 1626–1632. 10.13287/j.1001-9332.201705.021 [DOI] [PubMed] [Google Scholar]
  12. Henderson, R. , Kabengi, N. , Mantripragada, N. , Cabrera, M. , Hassan, S. , & Thompson, A. (2012). Anoxia‐induced release of colloid‐and nanoparticle‐bound phosphorus in grassland soils. Environmental Science & Technology, 46(21), 11727–11734. 10.1021/es302395r [DOI] [PubMed] [Google Scholar]
  13. Hens, M. , & Merckx, R. (2001). Functional characterization of colloidal phosphorus species in the soil solution of sandy soils. Environmental Science & Technology, 35(3), 493–500. 10.1021/es0013576 [DOI] [PubMed] [Google Scholar]
  14. Hu, Q. Y. , Lan, Y. Q. , & Xue, J. H. (1996). Factors affecting the stability of soil colloid. Soils, 06, 290–294+315. [Google Scholar]
  15. Hu, J. , Shen, Y. , & Wang, X. (2009). Release and mobilization of soil colloid in the natural soil packed column with various water saturations. Journal of Agro‐Environment Science, 28(9), 1829–1836. 10.3321/j.issn:1672-2043.2009.09.010 [DOI] [Google Scholar]
  16. Hu, P. , Zhang, Y. U. , Wang, J. , Du, Y. , Wang, Z. , Guo, Q. , Pan, Z. , Ma, X. , Planer‐Friedrich, B. , Luo, Y. , & Wu, L. (2023). Mobilization of colloid‐and nanoparticle‐bound arsenic in contaminated paddy soils during reduction and reoxidation. Environmental Science & Technology, 57(26), 9843–9853. 10.1021/acs.est.3c03051 [DOI] [PubMed] [Google Scholar]
  17. Igwe, C. A. , Zarei, M. , & Stahr, K. (2009). Colloidal stability in some tropical soils of southeastern Nigeria as affected by iron and aluminum oxides. Catena, 77(3), 232–237. 10.1016/j.catena.2009.01.003 [DOI] [Google Scholar]
  18. Laskar, S. Y. , Sileshi, G. W. , Pathak, K. , Debnath, N. , Nath, A. J. , Laskar, K. Y. , Singnar, P. , & Das, A. K. (2021). Variations in soil organic carbon content with chronosequence, soil depth and aggregate size under shifting cultivation. Science of the Total Environment, 762, 143114. 10.1016/j.scitotenv.2020.143114 [DOI] [PubMed] [Google Scholar]
  19. Li, F. (2021). Study on influencing factors of colloidal phosphorus production and loss control in soil of typical farmland in Zhejiang Province [Doctoral dissertation, Zhejiang University]. 10.27461/d.cnki.gzjdx.2021.000445 [DOI]
  20. Li, H. M. , Ge, Y. C. , Wang, Y. , Dong, W. Q. , & Li, Y. L. (2017). The study of critical flocculation concentration of ammonia nitrogen of colloids in the groundwater. South‐to‐North Water Transfers and Water Science & Technology, 15(4), 59–64. 10.13476/j.cnki.nsbdqk.2017.04.011 [DOI] [Google Scholar]
  21. Li, J. , Hu, X. , Shi, L. , & Xue, L. (2025). Soil organic matter enhanced the soil colloidal phosphorus via co‐precipitation with Fe/Al in paddy soil. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 717, 136863. 10.1016/j.colsurfa.2025.136863 [DOI] [Google Scholar]
  22. Li, F. , Liang, X. , Li, H. , Jin, Y. , Jin, J. , He, M. , Klumpp, E. , & Bol, R. (2020). Enhanced soil aggregate stability limits colloidal phosphorus loss potentials in agricultural systems. Environmental Sciences Europe, 32(1), 17. 10.21203/rs.2.19712/v1 [DOI] [Google Scholar]
  23. Li, B. , Liao, P. , Liu, P. , Wang, D. , Ye, Z. , Wang, J. , & Liu, C. (2022). Formation, aggregation, and transport of NOM–Cr (III) colloids in aquatic environments. Environmental Science: Nano, 9(3), 1133–1145. 10.1039/D1EN00861G [DOI] [Google Scholar]
  24. Liu, D. , Ju, W. , Jin, X. , Li, M. , Shen, G. , Duan, C. , Guo, L. , Liu, Y. , Zhao, W. , & Fang, L. (2021). Associated soil aggregate nutrients and controlling factors on aggregate stability in semiarid grassland under different grazing prohibition timeframes. Science of the Total Environment, 777, 146104. 10.1016/j.scitotenv.2021.146104 [DOI] [PubMed] [Google Scholar]
  25. Liu, Y. , Wang, P. , & Wang, J. (2023). Mechanism of formation and stabilization of soil aggregates: Research progress and prospects. Acta Pedologica Sinica, 60(3), 627–643. 10.11766/trxb202112180686 [DOI] [Google Scholar]
  26. Lu, R. K. (2000). Soil and agro‐chemistry analytical methods. Chinese Agricultural Science and Technology Press. (In Chinese). [Google Scholar]
  27. Montalvo, D. , Degryse, F. , & McLaughlin, M. J. (2015). Natural colloidal P and its contribution to plant P uptake. Environmental Science & Technology, 49(6), 3427–3434. 10.1021/es504643f [DOI] [PubMed] [Google Scholar]
  28. Sen, T. K. , & Khilar, K. C. (2006). Review on subsurface colloids and colloid‐associated contaminant transport in saturated porous media. Advances in Colloid and Interface Science, 119(2–3), 71–96. 10.1016/j.cis.2005.09.001 [DOI] [PubMed] [Google Scholar]
  29. Seta, A. K. , & Karathanasis, A. D. (1996). Water dispersible colloids and factors influencing their dispersibility from soil aggregates. Geoderma, 74(3–4), 255–266. 10.1016/S0016-7061(96)00066-3 [DOI] [Google Scholar]
  30. Shang, S. B. , Wang, F. Y. , & Li, X. Q. (2010). Effects of physical and chemical properties of soil colloids on their stability and mobility. Soils, 42(06), 1015–1019. 10.13758/j.cnki.tr.2010.06.016 [DOI] [Google Scholar]
  31. Soil Survey Staff . (2010). Keys to soil taxonomy. 11th ed. USDA‐NRCS. GPO, Washington, DC. [Google Scholar]
  32. Tang, Y. , Li, H. , Liu, X. , Zhu, H. , & Tian, R. (2015). Unraveling the size distributions of surface properties for purple soil and yellow soil. Journal of Environmental Sciences, 32, 81–89. 10.1016/j.jes.2014.12.011 [DOI] [PubMed] [Google Scholar]
  33. Tavakkoli, E. , Donner, E. , Juhasz, A. , Naidu, R. , & Lombi, E. (2013). A radio‐isotopic dilution technique for functional characterisation of the associations between inorganic contaminants and water‐dispersible naturally occurring soil colloids. Environmental Chemistry, 10(4), 341–348. 10.1071/EN13020 [DOI] [Google Scholar]
  34. Tiwari, E. , Khandelwal, N. , Singh, N. , Biswas, S. , & Darbha, G. K. (2022). Effect of clay colloid‐CuO nanoparticles interaction on retention of nanoparticles in different types of soils: Role of clay fraction and environmental parameters. Environmental Research, 203, 111885. 10.1016/j.envres.2021.111885 [DOI] [PubMed] [Google Scholar]
  35. Wu, X. , Peng, J. , Liu, P. , Bei, Q. , Rensing, C. , Li, Y. , Yuan, H. , Liesack, W. , Zhang, F. , & Cui, Z. (2021). Metagenomic insights into nitrogen and phosphorus cycling at the soil aggregate scale driven by organic material amendments. Science of the Total Environment, 785, 147329. 10.1016/j.scitotenv.2021.147329 [DOI] [PubMed] [Google Scholar]
  36. Xiao, H. , Liu, G. , Liu, P. , Zheng, F. , Zhang, J. , & Hu, F. (2017). Developing equations to explore relationships between aggregate stability and erodibility in Ultisols of subtropical China. Catena, 157, 279–285. 10.1016/j.catena.2017.05.032 [DOI] [Google Scholar]
  37. Yoder, R. E. (1936). A direct method of aggregate analysis of soils and a study of the physical nature of erosion losses. Journal of the American Society of Agronomy, 28(5), 337–351. 10.2134/agronj1936.00021962002800050001x [DOI] [Google Scholar]
  38. Zhang, M. (1993). Study on the mobility of natural clay particles in soil in China by methylene blue paper method. Bulletin of Science Technology, 03, 204. 10.13774/j.cnki.kjtb.1993.03.016 [DOI] [Google Scholar]
  39. Zhang, G.‐L. , & Gong, Z.‐T. (2003). Pedogenic evolution of paddy soils in different soil landscapes. Geoderma, 115(1–2), 15–29. [Google Scholar]
  40. Zhang, J. , Liu, J. , Chen, H. , Du, Y. X. , Li, J. Z. , Sun, H. Z. , & Zhao, Q. Z. (2014). Change in soil oxygen and water contents under alternate wetting and drying in paddy fields. Chinese Journal of Eco‐Agriculture, 22(4), 408–413. [Google Scholar]
  41. Zhang, J. , Wang, Y. , Gao, L. , Liu, Z. , & Agathokleous, E. (2025). Unraveling the link between soil phosphorus saturation and colloid behavior in ultisol: A case study. Journal of Soils and Sediments, 25(6), 1927–1938. 10.1007/s11368-025-04039-1 [DOI] [Google Scholar]
  42. Zhao, Y. , Liang, X. Q. , Fu, C. D. , Zhu, S. R. , Jin, Y. , & Ye, Y. S. (2015). Effect of phosphorus fertilizer input on colloidal phosphorus content in paddy soil profile. Acta Ecologica Sinica, 35(24), 8251–8257. 10.5846/stxb201406031147 [DOI] [Google Scholar]
  43. Zhao, H. , Tan, Q. , Zhang, S. , Yang, W. , Ma, X. , & Li, C. (2026). Contrasting effects of iron minerals on soil organic carbon stability in mangrove wetlands with different salinities. Ecological Processes, 15(1), Article 1. 10.1186/s13717-025-00649-2 [DOI] [Google Scholar]
  44. Zheng, T. , Yang, J. , Zhang, J. , Tang, C. , Liao, K. , & Liu, Y. (2019). Factors contributing to aggregate stability at different particle sizes in ultisols from Southern China. Journal of Soils and Sediments, 19(3), 1342–1354. 10.1007/s11368-018-2143-8 [DOI] [Google Scholar]
  45. Zhou, Y. J. (2022). Occurrence characteristics of colloidal formation layer in red soil of paddy field and its risk of loss of supported phosphorus . [Master's thesis, Nanjing University of Information Science and Technology].
  46. Zhu, L. H. (2018). Ion‐specific effects of positive and negative charge mixed colloid condensation . [Doctoral dissertation, Southwest University].

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The datasets used during the current study are available from the corresponding author upon reasonable request.


Articles from Journal of Environmental Quality are provided here courtesy of Wiley

RESOURCES