Preferential flow characteristics and ecohydrological functions of shallow fissure soils in fragile karst ecosystems

Abstract

Introduction:

Preferential flow is a key hydrological process controlling rainfall redistribu-tion, subsurface recharge, and ecohydrological functioning in karst ecosys-tems. However, the formation and ecological significance of preferential flow in shallow karst fissure (SKF) soils remain insufficiently quantified.

Methods:

Field dye-tracer experiments with Brilliant Blue FCF were conducted in SKF soils in Xixiu District, Guizhou Province, Southwest China, to characterize vertical flow patterns and identify soil-hydraulic and root controls.

Results:

Preferential flow showed strong vertical heterogeneity: D_c was high in the 0–10 cm layer (68.62–98.75%) but decreased rapidly with depth, while maximum stained depth varied from 23.0 to 74.6 cm among fissures, indicating marked spatial variability in subsurface hydraulic con-nectivity. Bulk density (BD) was the strongest negative control on D_c (r = −0.697), whereas total porosity (TP), non-capillary porosity (NCP), saturated hydraulic conductivity (Ks), saturated water content (SWC), capillary water content (CWC), field capacity (FC), root length density (RLD), and root surface area density (RSAD) generally promoted preferential flow development. PLSR further identified BD, SWC, CWC, NCP, TP, FC, and Ks as the main predictors, with VIP values greater than 1. These factors explained approximately 52–55% of the observed variation in D_c. The results indicate that preferential flow in SKF soils is jointly regulated by soil structure, hydraulic properties, and root-induced biopores.

Discussion:

Preferential flow provides a dual ecohydrological function by enhancing deep water redistribution and epikarst recharge while potentially increasing water leakage from the shal-low root zone. These findings improve the mechanistic understanding of water movement in karst fissure soils and provide guidance for vegetation restoration and water conservation in rocky desertification regions.

1 Introduction

Soil water infiltration is a fundamental process linking precipitation, soil water storage, plant water uptake, surface runoff, and groundwater recharge (Williams et al., 1996; You et al., 2022; Li Y. et al., 2023). In present hydrological assessments, infiltration is often simplified as relatively uniform matrix flow. Increasing field evidence, however, shows that water and solutes frequently bypass large portions of the soil matrix and move rapidly through limited but highly conductive pathways, a process commonly referred to as preferential flow (Kan et al., 2024; Rao et al., 2025; Wen et al., 2025). Preferential flow is a non-equilibrium transport process occurring through macropores, fissures, root channels, soil-rock interfaces, cracks, and other structural discontinuities (Sohrt et al., 2014; Wang et al., 2018; Kan et al., 2020; Zhao S. Y. et al., 2020; Yang et al., 2023; Hu et al., 2025). Although these pathways may occupy only a small fraction of the soil volume, they can dominate infiltration, deep percolation, and solute redistribution. Preferential flow may therefore enhance infiltration capacity, reduce surface runoff, and promote groundwater recharge, but it may also accelerate nutrient leaching, pollutant transport, and water loss from the root zone (Sohrt et al., 2014; Zhao and Wang, 2023; Rao et al., 2025). Clarifying the occurrence, morphology, and controls of preferential flow is thus essential for understanding soil hydrological processes and ecological water supply in vulnerable landscapes.

Preferential flow emerges from interactions among pore architecture, hydraulic conditions, vegetation, topography, and land use. Bulk density, total and non-capillary porosity, texture, organic matter, hydraulic conductivity, and water-retention properties influence the abundance, continuity, and activation of conductive pathways (Zhang T. et al., 2021; Li J. et al., 2023; Rao et al., 2025; Wen et al., 2025). Vegetation further modifies preferential flow by altering root architecture, litter input, biological activity, aggregation, and soil pore networks (Kan et al., 2020). Vegetation restoration often increases macroporosity, pore connectivity, soil organic matter, and root density, thereby improving infiltration and strengthening preferential flow development (Zhang et al., 2023; Wen et al., 2025). Because vegetation types differ in canopy interception, root distribution, understory cover, and litter accumulation, they may generate distinct preferential-flow patterns (Kan et al., 2020; Kan et al., 2024; Yang and Guo, 2025). These interactions indicate that preferential flow is not a fixed soil property, but an emergent hydrological response shaped by the coupled soil-vegetation system (Kan et al., 2024; Sharghi et al., 2025).

Karst landscapes are among the most hydrologically heterogeneous and ecologically fragile environments worldwide (Wang et al., 2004; Jiang et al., 2014; Wang et al., 2019; Xiong et al., 2023). Carbonate dissolution creates a dual or multiple hydrological structure consisting of soil matrix, epikarst fractures, rock fissures, and underground conduits, which leads to strong surface-subsurface connectivity and rapid water leakage (Yang et al., 2016; Liu et al., 2019; Zhao and Wang, 2023; Liu X. et al., 2025). Southwest China is one of the largest karst regions globally, where shallow discontinuous soils, exposed bedrock, and rocky desertification strongly constrain vegetation restoration and ecosystem sustainability (Wang et al., 2004; Wang et al., 2019; Xiong et al., 2023). During rocky desertification, soil erosion and bedrock exposure reduce soil thickness and water-holding capacity; remaining soils are frequently stored in fissures, solution grooves, and near-surface epikarst fractures (Yan et al., 2021; He et al., 2023). Previous studies have documented preferential transport in surface stony soils and along rock-soil interfaces, and have described the effects of rock fragments and slope aspect on preferential flow (Sohrt et al., 2014; Hou et al., 2023; Tian et al., 2024). However, the vertical organization of flow within shallow karst fissure (SKF) soils remains poorly quantified. In particular, it is unclear how rapidly broadly distributed near-surface infiltration converges into discrete pathways, which covarying soil and root properties best explain this transition, and whether the resulting connectivity primarily supports deeper water storage or promotes leakage from the shallow root zone.

Field dye tracing is a direct and effective approach for visualizing and quantifying preferential flow pathways in situ. It can reveal pathway morphology, depth distribution, connectivity, and spatial heterogeneity in soil profiles (Kan et al., 2020; Zhao S. Y. et al., 2020). Preferential flow intensity and structure have been quantified using metrics such as dye-stained area ratio, maximum stained depth, preferential flow ratio, length index, coefficient of variation, and fractal dimension (Rao et al., 2025; Yang and Guo, 2025). Thus, combining dye-tracer tests with soil physical, hydraulic, and root properties is key to understanding the hydraulic connectivity, ecological functions, and drought resilience of SKF soils in karst regions. The objectives of this study were to determine: (1) how do preferential-flow pathways and dye coverage change with depth in SKF profiles; (2) which soil structural, hydraulic, and root variables most strongly control their development; and (3) what ecohydrological benefits and risks follow from the observed flow patterns? The results provide scientific evidence for understanding water redistribution, plant water-use efficiency, and soil water conservation in karst rocky desertification regions.

2 Materials and methods

2.1 Study area

The study was conducted in Anshun City, Guizhou Province, Southwest China (105°13′–106°34′E, 25°21′–26°38′N), a representative karst area on the Yunnan-Guizhou Plateau. The area is characterized by peak-cluster and peak-forest karst landforms and lies near the watershed divide between the Yangtze River and Pearl River basins. Carboniferous, Permian, and Triassic carbonate rocks are widely distributed, forming extensive limestone and dolomite bedrock systems. The climate is subtropical humid monsoon with plateau characteristics; the mean annual temperature is approximately 14.0 °C, and the mean annual precipitation is around 1,360 mm, most of which occurs from May to September. Anshun City covers a total area of approximately 9,267 km², with the vast majority of its territory underlain by karst geology. Owing to carbonate lithology and intense soil-rock heterogeneity, hillslope soils are typically shallow, discontinuous, and spatially variable. Limestone soil derived from carbonate weathering is the dominant soil type. Natural vegetation is subtropical evergreen broad-leaved forest, with common tree species including Magnolia liliiflora Desr., Machilus cavaleriei Lévl., and Cupressus funebris Endl., and shrub species including Magnolia (Magnolia liliiflora Desr.), Anshun Runnan (Machilus cavaleriei Levl.), Cypress (Cupressus funebris Endl.) and so on. The main shrubs are Camptotheca acuminata (Camptotheca acuminata), Pyracantha spinosa (Pyracantha fortuneana (Maxim.) Li), Zanthoxylum bungeanum (Zanthoxylum simulans Hance.), Hedgerow firewood (Tirpitzia sinensis) and so on (Figure 1 and Table 1).

Figure 1

2.2 Tracer experiment and dye image analysis

Brilliant Blue FCF solution (4 g·L⁻¹) was used as the tracer because of its high solubility, low toxicity, and clear visibility in soil profiles (Weiler and Flühler, 2004; Wang et al., 2018). At each sampling site, 10 L of Brilliant Blue FCF solution (4 g·L⁻¹) was uniformly applied to an approximately 0.5 m × 0.5 m area (adjusted according to fissure width) using hand-held sprayers. The applied volume was equivalent to approximately 40 mm rainfall (Wang et al., 2018), the shape of the application area was adjusted to accommodate the local fissure geometry, while its total horizontal area was maintained at 0.25 m², calculated from the applied water volume and infiltration area. The solution was applied gradually over approximately 60 min to minimize ponding and prevent surface runoff. After application, the plots were covered with plastic film to reduce evaporation. After 24 h, three vertical soil profiles were sequentially excavated along the dye migration boundary at 10 cm intervals in each experimental plot as replicates. These profiles were treated as spatial subsamples rather than independent experimental replicates. The profile boundaries were measured using a surveyor’s tape, and images were taken using a Canon EOS 80D camera under consistent field conditions. The obtained soil properties, root distribution characteristics, and dye staining patterns were integrated to evaluate the effects of soil structure and vegetation on preferential flow development.

To avoid boundary effects and interference from the rock-soil interface, and the width of the image selected for analysis varied among SKF sampling points (10–40 cm) because of spatial heterogeneity in fissure width. Only the central portion of the 0.25 m² tracer-application area was selected and processed in Adobe Photoshop CS5; therefore, the analyzed profile area varied among sampling points according to fissure width and dye penetration depth. Images were corrected for perspective distortion and rescaled to a spatial resolution of 1 mm × 1 mm per pixel using Photoshop CS5. The same thresholding criteria were applied consistently to all images. The processed images were subsequently converted into binary images, in which dye-stained pixels were assigned a value of 1 and unstained pixels were assigned a value of 0. Stained areas were enhanced and converted into binary images, in which stained pixels represented preferential infiltration zones and unstained pixels represented matrix or non-activated zones. MATLAB was then used to calculate the dye-stained area at each depth and to derive the dye-stained area ratio (D_C) and maximum stained depth. Higher D_C values indicate more developed preferential flow and greater hydraulic connectivity (van Schaik, 2009).

The dye-stained area ratio was calculated using Equation 1:

where D_C is the dye-stained area ratio, D is the total stained area; N_D is the unstained area.

2.3 Measurement of soil properties

After the soil profiles had been photographed, intact soil cores were collected (dye-stained and unstained zones) to avoid disturbing the dye-staining patterns before image acquisition. Soil samples were collected vertically at 10 cm depth intervals from the soil surface to the maximum depth of dye penetration. At least 3 soil cores were collected using cutting ring (100 cm³) in each soil depth. The samples were immediately sealed, labeled, and transported to the laboratory for the determination of soil physicochemical properties and root traits. Saturated hydraulic conductivity (K_S) was measured using the constant-head method based on Darcy’s law (Zhu et al., 2022; Li Y. et al., 2023). Soil bulk density (BD), non-capillary porosity (NCP), capillary porosity (CP), total porosity (TP), saturated water content (SWC), capillary water content (CWC), and field capacity (FC) were determined using standard soil physical methods following Li Y. et al. (2023). Soil organic matter (SOM) was determined using the potassium dichromate oxidation method with external heating. Particle-size distribution was measured and classified according to the international classification standards as described by Yan et al. (2019). After the soil physical measurements, the soil samples were removed from the cutting ring, and the roots were carefully separated from the soil and thoroughly washed with water. The cleaned roots were scanned using WinRHIZO Pro software (Carey et al., 2024; Arslan et al., 2026). Root length density (RLD) and root surface area density (RSAD) were calculated by dividing the total root length and root surface area, respectively, by the volume of the corresponding soil core.

2.4 Data analysis

All statistical analyses were conducted after testing data normality and homogeneity of variance; transformations were applied where necessary. Vertical patterns of dye-stained area were extracted using MATLAB 2015, and figures were prepared using OriginPro 2024. Pearson correlation analysis was used to evaluate bivariate relationships between D_C and soil or root variables. Partial least squares regression (PLSR) was then applied to identify the dominant controls on D_C while reducing the effects of multicollinearity among predictors (Carrascal et al., 2009; Shi et al., 2014). The optimal number of PLSR components was selected according to the maximum cumulative cross-validated goodness of prediction (Q²_cum) and the minimum root mean square error of cross-validation (RMSECV). Variable importance in projection (VIP) values were used to rank the relative contribution of each predictor, with VIP > 1 indicating an important variable. Multiple stepwise regression was further used to construct empirical equations linking D_C with the dominant explanatory variables. PLSR and stepwise regression were performed in R 4.5.3.

3 Results

3.1 Vertical characteristics of preferential flow in SKF soils

D_C generally decreased with increasing soil depth, although maximum stained depth varied markedly among fissures (Figure 2). Across the nine fissures, maximum stained depth ranged from 23.0 to 74.6 cm, indicating large spatial variability in the continuity of preferential pathways. In the surface layer (0–10 cm), mean D_C was high, ranging from 68.62 to 98.75%, suggesting relatively widespread initial infiltration. D_C then declined rapidly with depth and eventually stabilized at approximately 16.95–38.16%. The coefficient of variation showed either a fluctuating increase or an initial increase followed by a decrease along the profile, indicating pronounced vertical heterogeneity and a transition from broad near-surface wetting to more channelized subsurface flow.

Figure 2

3.2 Physicochemical and root properties of SKF soils

The SKF soils showed substantial spatial heterogeneity in soil physical, hydraulic, and root properties (Table 2). K_S ranged from 0.38 to 1.76 mm·min⁻¹, while BD varied from 1.05 to 1.24 g cm⁻³. The soil texture was dominated by clay and silt, with clay content ranging from 33.89 to 56.87% and silt content from 35.14 to 60.84%. SOM varied markedly from 57.21 to 177.95 g kg⁻¹. SWC, CWC, FC, and porosity-related indicators also differed among fissures, reflecting heterogeneous hydraulic storage conditions. RLD and RSAD varied substantially, indicating differences in root development and potential biopore formation. These variations provide a structural and hydraulic basis for the observed spatial variability in preferential flow.

3.3 Factors influencing preferential flow in SKF soils

Correlation analysis showed that BD, NCP, and TP were highly significantly correlated with the D_C (p < 0.01), with correlation coefficients of −0.697, 0.512, and 0.506, respectively. Clay content was significantly and negatively correlated with D_C (−0.357). Ks, SWC, CWC, and FC were all highly positively (p < 0.01) correlated with D_C (0.415–0.589). Among plant-related variables, RLD showed a highly significant positive correlation (0.400), whereas RSAD was significantly positively correlated with the DC (0.372). In contrast, CP, silt content, sand content, and SOM showed no significant relationships with the D_C (Figure 3).

Figure 3

3.4 Dominant controls and predictive models of preferential flow

PLSR was used to identify the dominant factors explaining variation in D_C while accounting for interactions and collinearity among predictors. The one-component model provided the most parsimonious solution because it produced the highest Q²_cum and lowest RMSECV (Table 3). This component explained 45.57% of the variance in D_C, and the cumulative explanation increased to 58.03% with two components. VIP analysis identified BD, SWC, CWC, NCP, TP, FC, and K_S as the dominant predictors, with VIP values of 1.603, 1.354, 1.273, 1.178, 1.164, 1.145, and 1.066, respectively (Figure 4). Stepwise regression further generated three empirical equations for estimating D_C (Equations 2–4; Figure 5). The model based on dominant soil and hydraulic variables explained 51.8% of the variation in D_C, whereas models incorporating vegetation-related variables explained 52.1–54.9%. These results indicate that preferential flow in SKF soils is mainly controlled by soil structural and hydraulic properties, with root traits providing additional explanatory power.

Figure 4

Figure 5

4 Discussion

4.1 Morphological characteristics and development patterns of preferential flow in shallow fissure soil

The dye tracer experiment revealed pronounced preferential flow development in SKF soils. Dye-stained area ratio (D_C) decreased progressively with soil depth, whereas maximum dye penetration depth varied considerably among fissures, ranging from 23.0 to 74.6 cm. Surface soil layers (0–10 cm) exhibited high D_C (68.62–98.75%), indicating relatively uniform infiltration immediately after water entry (Hu et al., 2025; Duan et al., 2026). However, the rapid decline in D_C and increasing coefficient of variation with depth suggest that water movement gradually became concentrated within a limited number of preferential pathways. Similar depth-dependent transitions from matrix infiltration to preferential transport have been reported in karst soils and structured porous media (Sohrt et al., 2014; Wang et al., 2018). The large variation in maximum dye penetration depth among fissures highlights the strong influence of soil–rock heterogeneity on preferential flow development. In karst landscapes, shallow and discontinuous soils overlie fractured bedrock, generating highly heterogeneous hydraulic environments (Kan et al., 2020; Ding et al., 2021; Cen et al., 2023; Jiang et al., 2025). Under these conditions, infiltrating water preferentially converges toward fissures and rock–soil interfaces where hydraulic resistance is relatively low (Sohrt et al., 2014). Previous dye tracing studies demonstrated that rock–soil interfaces, fractures, and macropores constitute the dominant preferential flow pathways in karst ecosystems (Zhao and Wang, 2023).

Despite the overall decline in D_C with depth, preferential pathways remained hydraulically active in deeper soil layers, allowing infiltrating water to rapidly bypass large portions of the soil matrix. Such deep preferential transport is particularly important in karst environments because it establishes hydraulic connectivity between shallow soils and underlying fissure systems (Yang et al., 2016; Zhang et al., 2022; Yang et al., 2023). The results therefore support, rather than prove, a mechanism in which local structural connectivity controls whether infiltration is retained within shallow fissure soil or transmitted rapidly toward deeper epikarst zones.

4.2 Coupled effects of vegetation, root systems, and soil structure on preferential flow

The correlation analysis and PLSR results indicate that preferential flow in SKF soils is jointly controlled by soil structural properties and vegetation-associated factors. Among all measured variables, bulk density exhibited the strongest negative correlation with dye coverage (−0.697), whereas total porosity and non-capillary porosity showed significant positive relationships. Furthermore, bulk density possessed the highest VIP value (1.603) in the PLSR model, confirming its dominant role in regulating preferential flow development. Bulk density is closely linked to pore connectivity and soil compaction status (Wang et al., 2018). Lower bulk density generally indicates greater macropore abundance and improved hydraulic continuity, which facilitate rapid water transmission through preferential pathways. Similar findings have been reported following vegetation restoration, where reductions in soil compaction significantly enhanced hydraulic conductivity and infiltration capacity (Zhu et al., 2022; Zhang et al., 2026). The positive effects of non-capillary porosity and total porosity observed in this study further emphasize the importance of macropore networks in controlling preferential flow development.

Hydraulic properties also exerted strong influences on preferential flow. Saturated hydraulic conductivity, saturated water content, capillary water-holding capacity, and field capacity all showed significant positive relationships with dye coverage and VIP values greater than one. These results suggest that soils with favorable hydraulic conditions are more likely to develop continuous preferential pathways capable of transmitting water rapidly through the profile (Zhao et al., 2014; Zhang and Schaap, 2019; Li Y. et al., 2023). Similar relationships between hydraulic conductivity and preferential flow intensity have been observed in both karst and non-karst ecosystems (Kan et al., 2020; Zhu et al., 2022; Yang et al., 2023; Yang et al., 2026). Vegetation effects were mainly reflected through root development. Root length density exhibited a significant positive correlation with D_C (0.400), indicating that root channels contribute substantially to preferential flow formation. Root growth creates biopores that remain effective water-conducting pathways even after root decay (Li J. et al., 2023; Qu et al., 2025). Hu et al. (2025) demonstrated that root systems and pore structures explained more than 70% of the variation in preferential flow characteristics following vegetation restoration. In SKF soils, the interaction between root channels and fissure networks likely further enhances preferential transport.

Interestingly, soil organic matter did not show a significant correlation with dye coverage. This result suggests that although organic matter contributes to aggregate formation and soil quality improvement, preferential flow development in SKF soils is primarily governed by structural controls associated with fissures, macropores, and hydraulic connectivity (Yang et al., 2017; He et al., 2025; Qi et al., 2025; Tong et al., 2025). Therefore, preferential flow formation in shallow karst soils should be regarded as the outcome of coupled interactions among soil structure, hydraulic properties, and vegetation-induced biopore development.

4.3 Ecohydrological functions of preferential flow in water redistribution within SKF soils

Preferential flow can enhance water redistribution in karst fissure soils by bypassing the soil matrix and transferring water rapidly to deeper soil layers and epikarst storage zones. This process may increase subsurface water availability and provide an important water source for vegetation during seasonal drought. Fissure-controlled subsurface flow has been recognized as a key mechanism of water redistribution in karst hillslopes and can contribute to ecosystem resilience under water-limited conditions (Peng et al., 2019; Wei et al., 2021; Yan et al., 2021; Cen et al., 2023; Liu W. et al., 2025; Xia et al., 2026). By moving infiltrated water from evaporation-prone surface horizons to deeper storage zones, preferential flow may partially compensate for the limited water-holding capacity of shallow karst soils (Yang et al., 2017; Zhang et al., 2018). However, dye depth alone does not quantify recharge volume, residence time, or plant uptake. The present results therefore establish the capacity for deep hydraulic connection, but the magnitude of its contribution to ecological water supply remains to be tested directly.

Nevertheless, preferential flow also represents a potential leakage pathway. Highly connected preferential pathways can accelerate deep percolation and reduce water retention in the root zone. Previous studies have shown that preferential flow along rock-soil interfaces is an important mechanism of soil water leakage in rocky desertification regions (Zhao Z. et al., 2020; Cen et al., 2023; Yang et al., 2023; Zhao and Wang, 2023). Preferential flow therefore has a dual ecohydrological function: it enhances subsurface recharge and vertical water redistribution, but it may also increase water loss from shallow rooting zones. The ecological outcome depends on the balance between beneficial recharge and excessive drainage. Moderate preferential flow may improve rainfall-use efficiency and vegetation water supply, whereas excessive connectivity may intensify leakage and reduce shallow soil water storage. To synthesize the above findings, a conceptual framework was developed to illustrate how soil structure, hydraulic properties, and vegetation–root interactions jointly regulate preferential flow development and its dual ecohydrological functions in shallow karst fissure soils (Figure 6).

Figure 6

4.4 Restoration implications derived from preferential flow mechanisms in SKF soils

The identification of bulk density, porosity characteristics, hydraulic properties, and root development as dominant controls on preferential flow has important implications for ecological restoration in rocky desertification regions (Yang et al., 2017; Li Y. et al., 2023). Traditional restoration strategies often focus primarily on increasing vegetation cover, whereas the present results suggest that improving soil structural quality is equally important for enhancing ecosystem water-use efficiency. The strong influence of bulk density and pore structure indicates that reducing soil compaction should be a key objective of restoration practices. Improved soil structure promotes infiltration, enhances hydraulic connectivity, and increases soil water storage capacity (Zhang Y. et al., 2021; Zhou et al., 2022; Wang et al., 2025; Cao et al., 2026). Similarly, vegetation restoration can contribute to preferential flow development through root growth and macropore formation, thereby improving subsurface water recharge and rainfall utilization efficiency (Pereira et al., 2022; Zhu et al., 2022; Fan et al., 2025; Hu et al., 2025).

Root systems play a particularly important role in regulating preferential flow. The positive relationship between root length density and dye coverage suggests that root-induced biopores function as effective water transport pathways (Mu et al., 2026). Nevertheless, excessive preferential connectivity may increase deep leakage losses. Therefore, vegetation selection and planting density should be optimized to balance water acquisition and water conservation, particularly in shallow-soil karst ecosystems (Yan et al., 2019; Lu et al., 2026; Mu et al., 2026). At the landscape scale, preferential flow should not be viewed solely as a mechanism of water loss. Instead, it constitutes an integral component of karst ecohydrological functioning by linking surface infiltration, subsurface recharge, and vegetation water supply (Wang et al., 2018; Ding et al., 2021; Liu X. et al., 2025). Restoration measures that improve soil structure, enhance water-holding capacity, and regulate preferential flow intensity may maximize the beneficial effects of rainfall redistribution while minimizing deep drainage losses and enhancing the water use efficiency of vegetation (Peng et al., 2019; Zhang et al., 2022; Li et al., 2024). Overall, successful rocky desertification restoration requires integrated management of vegetation, soil properties, and preferential flow processes. Such an approach can improve rainfall-use efficiency, strengthen ecosystem resilience, and promote sustainable water resource management in karst regions.

4.5 Limitations and future research

Several limitations should be considered when interpreting these results. First, the dye-tracer experiments were conducted at selected field profiles, and the number of fissures was limited. Although the profiles captured clear heterogeneity in SKF soils, larger sample sizes across different lithologies, slope positions, and vegetation restoration stages would improve the generality of the conclusions. Second, D_C is an effective indicator of preferential flow morphology, but it does not directly quantify transient water flux or solute transport rates. Combining dye tracing with time-series soil moisture monitoring, electrical resistivity tomography, stable isotopes, or hydrological modeling would help quantify flow velocity and recharge dynamics. Future research should further link preferential-flow morphology with plant water uptake, drought resistance, and nutrient transport. Multi-season observations under natural rainfall events are particularly needed to determine when preferential pathways improve ecological water supply and when they increase leakage losses. Such work would provide stronger process-based guidance for vegetation restoration and water conservation in karst rocky desertification regions.

5 Conclusion

This study characterized preferential flow processes in SKF soils and identified their dominant controls under contrasting vegetation and soil conditions. The main conclusions are as follows: Preferential flow in SKF soils is highly heterogeneous and structurally constrained. D_C decreased with depth, while maximum stained depth varied considerably among fissures (23.0–74.6 cm). Soil structural and hydraulic properties are the primary controls on preferential flow development. BD, TP, NCP, SWC, CWC, FC, and K_S regulated the initiation, continuity, and intensity of preferential pathways. Root traits, especially RLD, further enhanced preferential connectivity through biopore formation. The regression models explained 51.8–54.9% of D_C variation. Preferential flow has a dual ecohydrological function in SKF systems. It promotes rapid vertical water redistribution and epikarst recharge, thereby improving rainfall utilization, but it may also increase deep drainage and reduce shallow root-zone water retention. This trade-off highlights the need to regulate preferential connectivity during ecological restoration. Overall, these findings clarify preferential flow mechanisms in karst fissure soils and support improved vegetation restoration and soil management in rocky desertification regions.

References

• 1

  ArslanM.KemperR.DöringT. F.SchmittmannO.AthmannM. (2026). Effects of biological and technical subsoil amelioration on root growth and crop performance in cereal crops. Soil Tillage Res.256:106907. doi: 10.1016/j.still.2025.106907

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  CaoY.ZhouZ.ChenM.LiuJ. E.WangP.WangN.et al. (2026). Effects of gully topographic vertical zone on the spatial heterogeneity of root-soil complex shear performance in the loess plateau. Catena263:109764. doi: 10.1016/j.catena.2025.109764

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  CareyA.NairA.ThomsA. (2024). Evaluating the soil block method and growing media in organic vegetable transplant production. HortScience59, 542–551. doi: 10.21273/hortsci17566-23

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  CarrascalL. M.GalvánI.GordoO. (2009). Partial least squares regression as an alternative to current regression methods used in ecology. Oikos118, 681–690. doi: 10.1111/j.1600-0706.2008.16881.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  CenL.PengX.DaiQ.LiC.YiX. (2023). Creep leakage process of remaining soils in near-surface fissures in a karst area with bedrock outcrops. Catena221:106802. doi: 10.1016/j.catena.2022.106802

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  DingY.NieY.ChenH.WangK.QuerejetaJ. I. (2021). Water uptake depth is coordinated with leaf water potential, water-use efficiency and drought vulnerability in karst vegetation. New Phytol.229, 1339–1353. doi: 10.1111/nph.16971,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  DuanJ.WangL.TangC.LiuY.ZhengH.YangJ. (2026). Divergent effects of grass cover on soil infiltration patterns and water recharge in orchards: taproot vs. fibrous root systems. Soil Tillage Res.258:107045. doi: 10.1016/j.still.2025.107045

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  FanY.YanY.GanF.DaiQ.ZhangX. (2025). Editorial: understanding the relationship between extreme climate events and forest soil hydrology: implications for ecosystem functions. Front. For. Glob. Change8:1579133. doi: 10.3389/ffgc.2025.1579133

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  HeJ.CaoY.ZhangK.XiaoS.CaoZ. (2023). Soil loss through fissures and its responses to rainfall based on drip water monitoring in karst caves. J. Hydrol.617:129000. doi: 10.1016/j.jhydrol.2022.129000

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  HeZ.HeS.ZhengZ.YiH.LuoZ. (2025). Transforming slope cropland into terraced cropland to enhance the retention of labile SOC fractions by aggregates in southwestern China. Catena259:109374. doi: 10.1016/j.catena.2025.109374

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  HouF.ChengJ.GuanN. (2023). Influence of rock fragments on preferential flow in stony soils of karst graben basin, Southwest China. Catena220:106684. doi: 10.1016/j.catena.2022.106684

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  HuJ.RenY.TangM.ZhangZ.YangK.ZhenQ.et al. (2025). Effects of vegetation restoration on infiltration patterns and preferential flow in semi-arid areas with shallowly buried soft bedrock (Pisha sandstone) in China. J. Hydrol.661:133546. doi: 10.1016/j.jhydrol.2025.133546

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  JiangZ.LianY.QinX. (2014). Rocky desertification in Southwest China: impacts, causes, and restoration. Earth-Sci. Rev.132, 1–12. doi: 10.1016/j.earscirev.2014.01.005

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  JiangW.PengT.ZhangX.WangS.MohammadiZ.TangZ. (2025). High-resolution electrical resistivity tomography for quantitative interpretation of sub-surface karst structures: a case study in Southwest China. Geoderma461:117460. doi: 10.1016/j.geoderma.2025.117460

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  KanX. Q.ChengJ. H.HouF. (2020). Response of preferential soil flow to different infiltration rates and vegetation types in the karst region of Southwest China. Water12:1778. doi: 10.3390/w12061778

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  KanX. Q.ChengJ. H.ZhengW. G.ZhangzhongL. L.LiJ.LiuC. B.et al. (2024). Interpretation of soil characteristics and preferential water flow in different forest covers of karst areas of China. Water16:2319. doi: 10.3390/w16162319

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  LiJ.CuiP.YinY. (2023). Field observation and micro-mechanism of roots-induced preferential flow by infiltration experiment and phase-field method. J. Hydrol.623:129756. doi: 10.1016/j.jhydrol.2023.129756

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  LiY.LiK.ZhouQ.ZhaoY.CaiL.YangZ. (2024). Spatiotemporal dynamics and similarity in soil moisture in shallow soils on karst slopes. J. Hydrol.639:131655. doi: 10.1016/j.jhydrol.2024.131655

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  LiY.WangS.PengT.ZhaoG.DaiB. (2023). Hydrological characteristics and available water storage of typical karst soil in SW China under different soil-rock structures. Geoderma438:116633. doi: 10.1016/j.geoderma.2023.116633

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  LiuW.ChenX.WangL.ZhangZ.LiuX.PengT.et al. (2025). The role of karstic soil-rock structures on subsurface stormflow dynamics in Southwest China. J. Hydrol.657:133126. doi: 10.1016/j.jhydrol.2025.133126

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  LiuX.ChenX.ZhangZ.LiuW.GaoF.ChengQ.et al. (2025). The role of rock fractures as a water source for trees growing in karst. Water Resour. Res.61: e2024WR039588. doi: 10.1029/2024wr039588

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  LiuH. Y.JiangZ. H.DaiJ. X.WuX. C.PengJ.WangH. Y.et al. (2019). Rock crevices determine woody and herbaceous plant cover in the karst critical zone. Sci. China Earth Sci.62, 1756–1763. doi: 10.1007/s11430-018-9328-3

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  LuL.ZengF.ZengZ.DuH.SunW.HeX.et al. (2026). Vegetation restoration patterns and soil properties jointly affected the ecosystem multifunctionality in the karst region. Ecol. Eng.226:107915. doi: 10.1016/j.ecoleng.2026.107915

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  MuY.XiongK.LiuZ.LiuG.CaiL. (2026). Seasonal variations in water use strategies in a karst region. Hydrol. Process.40:e70371. doi: 10.1002/hyp.70371

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  PengX. D.DaiQ. H.DingG. J.ShiD. M.LiC. L. (2019). The role of soil water retention functions of near-surface fissures with different vegetation types in a rocky desertification area. Plant Soil441, 587–599. doi: 10.1007/s11104-019-04147-1

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  PereiraL. C.BalbinotL.NnadiE. O.MoslehM. H.TonelloK. C. (2022). Effects of Cerrado restoration on seasonal soil hydrological properties and insights on impacts of deforestation and climate change scenarios. Front. For. Glob. Change5:882551. doi: 10.3389/ffgc.2022.882551

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  QiJ.WeigtS.WangM.RezanezhadF.QuintonW.ZakD.et al. (2025). Hydro-physical and carbon properties of peat across peatland types and climate zones. Geoderma461:117480. doi: 10.1016/j.geoderma.2025.117480

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  QuY. Y.WuQ. X.KhanF. U.WangJ. F.RenX. Z.ChaiX. H.et al. (2025). The impact of root systems on soil macropore abundance and soil infiltration capacity. Plant Soil513, 1197–1214. doi: 10.1007/s11104-025-07237-5

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  RaoC.DongW.SuX.LvH.ShenX. (2025). Experimental study on the influence mechanism of freeze-thaw action on the preferential flow pattern in vadose zone. J. Hydrol.660:133389. doi: 10.1016/j.jhydrol.2025.133389

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  SharghiS. F.BaukeS. L.RahmatiM.BurgerD. J.VereeckenH.AmelungW. (2025). Soil infiltration variability across diverse soil reference groups, textures, and landuse types. Geoderma463:117550. doi: 10.1016/j.geoderma.2025.117550

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  ShiZ. H.HuangX. D.AiL.FangN. F.WuG. L. (2014). Quantitative analysis of factors controlling sediment yield in mountainous watersheds. Geomorphology226, 193–201. doi: 10.1016/j.geomorph.2014.08.012

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  SohrtJ.RiesF.SauterM.LangeJ. (2014). Significance of preferential flow at the rock soil interface in a semi-arid karst environment. Catena123, 1–10. doi: 10.1016/j.catena.2014.07.003

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  TianX.WuW.ZengS.LiY.JiangY. (2024). Differences in soil water movement between the dip and anti-dip slopes of a karst trough valley. J. Hydrol.636:131246. doi: 10.1016/j.jhydrol.2024.131246

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  TongY.ZhangC.YuY.CaoQ.YangZ.ZhangX.et al. (2025). Effects of different cultivation patterns of grasses on plant community structure, soil physicochemical properties, microbial community structure, and functions. Plant Soil515, 1477–1497. doi: 10.1007/s11104-025-07663-5

  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  van SchaikN. L. M. B. (2009). Spatial variability of infiltration patterns related to site characteristics in a semi-arid watershed. Catena78, 36–47. doi: 10.1016/j.catena.2009.02.017

  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  WangF.ChenH.LianJ.FuZ.NieY. (2018). Preferential flow in different soil architectures of a small karst catchment. Vadose Zone J.17, –10. doi: 10.2136/vzj2018.05.0107

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  WangS. J.LiR. L.SunC. X.ZhangD. F.LiF. Q.ZhouD. Q.et al. (2004). How types of carbonate rock assemblages constrain the distribution of karst rocky desertified land in Guizhou Province, PR China: phenomena and mechanisms. Land Degrad. Dev.15, 123–131. doi: 10.1002/ldr.591

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  WangZ.SunM.WangY.LiuX.FengH.ZhuY.et al. (2025). Long-term application of organic amendments increases soybean yield by enhancing soil quality in aggregate scale. Agronomy15:2801. doi: 10.3390/agronomy15122801

  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  WangK.ZhangC.ChenH.YueY.ZhangW.ZhangM.et al. (2019). Karst landscapes of China: patterns, ecosystem processes and services. Landsc. Ecol.34, 2743–2763. doi: 10.1007/s10980-019-00912-w

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  WeiX.ZhouQ.CaiM.WangY. (2021). Effects of vegetation restoration on regional soil moisture content in the humid karst areas-a case study of Southwest China. Water13:321. doi: 10.3390/w13030321

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  WeilerM.FlühlerH. (2004). Inferring flow types from dye patterns in macroporous soils. Geoderma120, 137–153. doi: 10.1016/j.geoderma.2003.08.014

  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  WenY.LiM.GaoP.ZhouJ.AiX.ZhangY.et al. (2025). Soil physicochemical properties and roots promoted preferential flow development after vegetation restoration. J. Hydrol.659:133349. doi: 10.1016/j.jhydrol.2025.133349

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  WilliamsM.RastetterE. B.FernandesD. N.GouldenM. L.WofsyS. C.ShaverG. R.et al. (1996). Modelling the soil-plant-atmosphere continuum in a Quercus-Acer stand at Harvard forest: the regulation of stomatal conductance by light, nitrogen and soil/plant hydraulic properties. Plant Cell Environ.19, 911–927. doi: 10.1111/j.1365-3040.1996.tb00456.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  XiaR.ZhangX.LiuY.LiT.ZhangJ.WangC.et al. (2026). Effect of bedrock outcrops and underground fissures on soil loss, flow hydraulics and hydrological connectivity characteristics of karst slopes in alpine canyon area. J. Hydrol.674:135529. doi: 10.1016/j.jhydrol.2026.135529

  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  XiongK.KongL.YuY.ZhangS.DengX. (2023). The impact of multiple driving factors on forest ecosystem services in karst desertification control. Front. For. Glob. Change6:1220436. doi: 10.3389/ffgc.2023.1220436

  • CrossRef
  • Google Scholar
  • View reference in article

• 46

  YanY.DaiQ.JinL.WangX. (2019). Geometric morphology and soil properties of shallow karst fissures in an area of karst rocky desertification in SW China. Catena174, 48–58. doi: 10.1016/j.catena.2018.10.042

  • CrossRef
  • Google Scholar
  • View reference in article

• 47

  YanY.DaiQ.YangY.YanL.YiX. (2021). Epikarst shallow fissure soil systems are key to eliminating karst drought limitations in the karst rocky desertification area of SW China. Ecohydrology15:e2372. doi: 10.1002/eco.2372

  • CrossRef
  • Google Scholar
  • View reference in article

• 48

  YangD.FangN.SheD.HuangX.ShiZ. (2026). Predicting saturated hydraulic conductivity on dryland hillslopes with vegetation patches by using interpretable machine learning. Catena269:110086. doi: 10.1016/j.catena.2026.110086

  • CrossRef
  • Google Scholar
  • View reference in article

• 49

  YangY.GuoQ. (2025). Response of preferential flow to initial soil water content in coalmining subsidence zones along the middle reaches of the Yellow River, China. Water17:2606. doi: 10.3390/w17172606

  • CrossRef
  • Google Scholar
  • View reference in article

• 50

  YangJ.NieY. P.ChenH. S.WangS.WangK. L. (2016). Hydraulic properties of karst fractures filled with soils and regolith materials: implication for their ecohydrological functions. Geoderma276, 93–101. doi: 10.1016/j.geoderma.2016.04.024

  • CrossRef
  • Google Scholar
  • View reference in article

• 51

  YangW.PengX.DaiQ.LiC.XuS.LiuT. (2023). Storage infiltration of rock-soil interface soil on rock surface flow in the rocky desertification area. Geoderma435:116512. doi: 10.1016/j.geoderma.2023.116512

  • CrossRef
  • Google Scholar
  • View reference in article

• 52

  YangJ.XuX.LiuM.XuC.ZhangY.LuoW.et al. (2017). Effects of "grain for Green" program on soil hydrologic functions in karst landscapes, southwestern China. Agric. Ecosyst. Environ.247, 120–129. doi: 10.1016/j.agee.2017.06.025

  • CrossRef
  • Google Scholar
  • View reference in article

• 53

  YouJ.WangS. Q.XuD. (2022). Experimental investigations on influence of fracture networks on overland flow and water infiltration in soil. Water14:3483. doi: 10.3390/w14213483

  • CrossRef
  • Google Scholar
  • View reference in article

• 54

  ZhangJ.ChenH.FuZ.LuoZ.WangF.WangK. (2022). Effect of soil thickness on rainfall infiltration and runoff generation from karst hillslopes during rainstorms. Eur. J. Soil Sci.73:e13288. doi: 10.1111/ejss.13288

  • CrossRef
  • Google Scholar
  • View reference in article

• 55

  ZhangT.LiJ.PuJ.HuoW.WangS. (2021). Spatiotemporal variations of soil water stable isotopes in a small karst sinkhole basin. Environ. Earth Sci.80:29. doi: 10.1007/s12665-020-09284-w

  • CrossRef
  • Google Scholar
  • View reference in article

• 56

  ZhangY.SchaapM. G. (2019). Estimation of saturated hydraulic conductivity with pedotransfer functions: a review. J. Hydrol.575, 1011–1030. doi: 10.1016/j.jhydrol.2019.05.058

  • CrossRef
  • Google Scholar
  • View reference in article

• 57

  ZhangY.XuX.LiZ.XuC.LuoW. (2021). Improvements in soil quality with vegetation succession in subtropical China karst. Sci. Total Environ.775:145876. doi: 10.1016/j.scitotenv.2021.145876,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 58

  ZhangR.XuX.LiuM.ZhangY.XuC.YiR.et al. (2018). Comparing ET-VPD hysteresis in three agroforestry ecosystems in a subtropical humid karst area. Agric. Water Manag.208, 454–464. doi: 10.1016/j.agwat.2018.06.007

  • CrossRef
  • Google Scholar
  • View reference in article

• 59

  ZhangB.XuC.ZhangZ.HuC.HeY.HuangK.et al. (2023). Response of soil organic carbon and its fractions to natural vegetation restoration in a tropical karst area, Southwest China. Front. For. Glob. Change6:1172062. doi: 10.3389/ffgc.2023.1172062

  • CrossRef
  • Google Scholar
  • View reference in article

• 60

  ZhangS.ZhaoG.LiuY.TianP.MuX. (2026). Soil infiltration responses to vegetation restoration along a precipitation gradient on the loess plateau. Catena268:110088. doi: 10.1016/j.catena.2026.110088

  • CrossRef
  • Google Scholar
  • View reference in article

• 61

  ZhaoS. Y.JiaY. W.GongJ. G.NiuC. W.SuH. D.GanY. D.et al. (2020). Spatial variability of preferential flow and infiltration redistribution along a rocky-mountain hillslope, northern China. Water12:1102. doi: 10.3390/w12041102

  • CrossRef
  • Google Scholar
  • View reference in article

• 62

  ZhaoZ.ShenY.WangQ.JiangR. (2020). The temporal stability of soil moisture spatial pattern and its influencing factors in rocky environments. Catena187:104418. doi: 10.1016/j.catena.2019.104418

  • CrossRef
  • Google Scholar
  • View reference in article

• 63

  ZhaoZ. M.WangQ. H. (2023). Effect of rock film mulching on preferential flow at rock-soil interfaces in rocky karst areas. Water15:1775. doi: 10.3390/w15091775

  • CrossRef
  • Google Scholar
  • View reference in article

• 64

  ZhaoX.WuP.GaoX.TianL.LiH. (2014). Changes of soil hydraulic properties under early-stage natural vegetation recovering on the loess plateau of China. Catena113, 386–391. doi: 10.1016/j.catena.2013.08.023

  • CrossRef
  • Google Scholar
  • View reference in article

• 65

  ZhouQ.ZhuA. X.YanW.SunZ. (2022). Impacts of forestland vegetation restoration on soil moisture content in humid karst region: a case study on a limestone slope. Ecol. Eng.180:106648. doi: 10.1016/j.ecoleng.2022.106648

  • CrossRef
  • Google Scholar

• 66

  ZhuP.ZhangG.ZhangB. (2022). Soil saturated hydraulic conductivity of typical revegetated plants on steep gully slopes of Chinese loess plateau. Geoderma412:115717. doi: 10.1016/j.geoderma.2022.115717

  • CrossRef
  • Google Scholar