A review of advancements in the theory and characterization of soil macropore structure

Definition and classification of macropores

The concept of soil macropores is subjective. The most straightforward and commonly used method for defining macropores is based on the equivalent diameter (ED). However, there is still no consensus in the academic community regarding the classification of macropore diameters. Consequently, the pore definitions provided in different studies vary significantly, as summarized in Table 1. Yang et al. (2018) used CT to investigate the impact of different soil amendment methods on soil pore properties and pore distribution. They categorized soil pores into two categories: macropores (>1 mm in diameter) and smaller pores (0.13–1.0 mm). Kong et al. (2020) employed NMR to ascertain the pore size of soil subjected to freeze-thaw conditions and classified the pores into tiny pores (0−0.2 µm), medium pores (0.2–10 µm), and large pores (>10 µm). Gao et al. (2018) observed loess pores with SEM and subsequently classified the pores into four categories: micropores (ED <2 µm), minipores (2 µm <ED <8 µm), mesopores (8 µm <ED <32 µm), and macropores (D >32 µm). In order to examine the influence of salt content and salt type on pore structure, Wang et al. (2020a) and Wang et al. (2020b) used MIP to determine the pore size distribution, which was classified into micropores (<0.04 µm), small pores (0.04−0.4 µm), and mesopores (0.4-4 µm). Overall, pore classification thresholds and terminology vary among researchers owing to differences in the research objectives, methods, and equipment used. Nevertheless, it is common to classify soil pores into three categories: large, medium, and small.

Although the definition of macropores in terms of pore size remains elusive, their classification based on origin and morphology has gained widespread acceptance within the academic community (Liu, Qu & Ying, 2001). Soil macropores can be classified based on their origin, with biological pores referring to plant root pores and soil animal burrows, abiotic pores resulting from freeze-thaw and wet-dry cycles, and human activities such as transportation and cultivation, and natural macropores forming based on soil textural factors. Macropores can also be categorized by their morphology into holes, fissures, cracks, tubes, and irregular macropores (Zhang, Xu & Pei, 2012). Holes have a spherical structure, cracks have a ratio of long to short axes greater than 10, fissures are smaller than cracks, and tubes have an approximate cylindrical shape. Studies have shown that the shape of macropores is correlated with their origin. For example, macropores formed by plants and animals are usually tubular, whereas cracks or fissures are often formed by physical processes, such as collapse, freeze-thaw, and wet-dry cycles. The macropores formed by the soil aggregates are primarily irregular. Zong et al. (2015) discovered that macropores in soil formed by decaying roots often assume tubular or elongated shapes. However, pores created by repeated wet-dry cycles are smaller, more random, and less continuously distributed. Other researchers have proposed alternative nomenclatures for the classification of macropore morphologies. For example, Zong et al. (2015) classified pores into regular (0.5–1.0), irregular (0.2–0.5), and elongated (0.0–0.2) based on the range of the shape factor. Dhaliwal & Kumar (2022) categorized pores into equant, oblate, triaxial, prolate, planar, acicular-planar, and acicular based on the ratio of the intermediate to large axis and the short to intermediate axis. Although the shape factor is a useful parameter for pore classification, there is no fixed set of standard thresholds for classification; therefore, thresholds need to be tailored to the experience and actual pore distribution patterns of each research object (Li et al., 2019a; Li et al., 2019b). The classification and definition of macropores are crucial for better understanding the characteristics of different types of macropores and are essential prerequisites for the study of soil macropores.

Factors affecting macropore structure

The structure of soil macropores is affected by various factors. Anthropogenic factors, such as coal mining, mechanical compaction, and tillage management, play a significant role, though non-anthropogenic factors, including animals, plants, soil types, and wet-dry cycles, also exert an influence, as shown in Table 2. These factors can lead to significant differences in the macropore structure of soil. In addition to the description in the previous section that the shape of macropores can be affected by their origin, Dhaliwal & Kumar (2022) found that the emergence of distinct macropore shapes is linked to the soil type and management practices. Zhao, Hu & Li (2020) determined that several factors, including plant roots, freeze-thaw cycles, temperature fluctuations, and precipitation patterns, can affect pore morphology. Moreover, they observed that the higher the organic matter content of the soil, the more irregular the pore shape. The functioning of macropores can also be affected by many factors, Coal mining and mechanical compaction have destructive effects on the macropores, diminishing their multifunctionality. Conversely, plant roots typically enhance soil pore structure and increase porosity and pore connectivity. The behavior of different soil animals also vary in their effects on macropores. Furthermore, freeze-thaw and wet-dry cycles contribute to the increase in macropore number and complexity.The researchers can choose the most relevant macropore properties for their study purposes. For example, if the objective is to investigate the impact of biological factors on soil macropores, the focus should be on macropore connectivity. Instead, if anthropogenic factors, such as mechanical compaction, are the subject of study, attention should be given to the macropore shape and diameter.

Indicators of macropores and effects of macropores on soil ecological functions

The properties of macropores can be described using various indicators (Schlüter & Vogel, 2022). These indicators can be divided into two categories: those that describe macropore morphology and those that characterize macropore functions. The indicators of macropore morphology include parameters such as pore equivalent diameter, Ferret diameter, fractal dimension, roundness, sphericity, shape factor, and tortuosity. These indicators provide information on the shape and structure of the macropores. Indicators that characterize the macropore function include parameters such as pore number, porosity, pore size distribution, maximum pore area, pore-throat ratio, Euler number, inclination angle, surface area density, network density, and junction density. Table 3 presents the definitions and calculations of some commonly used macropore indicators.

Soil macropores play a pivotal role in ecosystems and exert profound effects on the ecological functions of soil (Keller et al., 2013). These effects are primarily observed in the hydrological cycle, soil quality, and biodiversity (Li et al., 2023). Macropores, the main space within the soil that retains water and gases, facilitate the diffusion and transportation of water, gases, and solutes. Despite comprising only 0.1–5% of the soil volume, macropores serve as the preferred channels for soil water infiltration and evaporation. They can conduct approximately 90% of water flux, significantly influencing the rate, amount, and path of soil water transport (Hayashi, Ken’Ichirou & Mizuyama, 2006). Furthermore, soil solutes migrate within the preferential channels formed by macropores along with the movement of water, influencing the content and distribution of nutrients in the soil (Zhang et al., 2022b). More importantly, soil macropores provide essential living spaces and oxygen for plants, animals, and microorganisms within the soil. This not only impacts the reproduction of biological populations and the differentiation of ecological niches but also plays a significant role in maintaining biodiversity.

The macropore indicators is closely associated with the functional behavior of macropore (Alessandrino et al., 2023). For instance, an elevated porosity level is associated with enhanced aeration and drainage, although it may also result in a diminished water-holding capacity. The functional roles of pores of different sizes in soil are distinct. Macropores of a larger size are primarily responsible for facilitating the transportation of air and water, whereas those of a medium and smaller size are more involved in the processes of water retention and nutrient cycling (Ajayi et al., 2021). The functional behavior of macropores is also influenced by their connectivity, which can be quantified using Euler number and pore-throat ratio. Well-connected macropores can form effective pathways for gas and water, enhancing the overall performance of the soil. In contrast, poorly connected macropores may become isolated “dead pores”, contributing minimally to soil function. Scholars often select specific indicators to describe the properties of macropores quantitatively based on their research objectives. Feng et al. (2020) investigated the effect of mechanical compaction on the structural properties of soil macropores by characterizing the macroporosity, macropore number, area, volume, and connectivity. In another study, Kan et al. (2021) characterized the diameter, volume, surface area, and distribution of macropores in different types of soils and then calculated the macropore angle, tortuosity, actual length, and curvature to understand their relationship with water content. Li et al. (2019a) and Li et al. (2019b) analyzed loess permeability by considering factors such as porosity, pore diameter, tortuosity, and fractal dimensions. By selecting appropriate indicators, researchers can effectively describe the morphology and properties of macropores and understand their behavior in different environments.

Staining observation

SO involves directly dyeing soil in the field or laboratory, taking photographs, processing, and analyzing digital images using image-processing techniques. This method is intuitive and is suitable for large-scale studies. The solution flows through the pore space in the soil, and the stained area represents the existence of the pore space. The more connected the pore space in depth, the deeper the staining. Liu et al. (2012) stained and observed the pore structure of soil profiles using brilliant blue. They characterized the volume of the macropores by calculating the stained area ratio. This allowed them to compare and analyze the soil macropore structure and its origin under different vegetation cover types. Additionally, SO allows for the observation of soil macropore flow paths and morphological features, making it a useful tool for characterizing and discriminating preferential flows. For example, Wang & Zhang (2011) used a combined tracer of dye and iodine to characterize macropore structure and heterogeneous soil water flow patterns. Dye staining was used to visualize the macropore structure, whereas iodine staining was used to visualize the non-uniform flow patterns.

Although convenient and efficient, use of SO for determining soil macropores is limited to soils with good permeability. This is because soils with good permeability facilitate the diffusion of the dye, producing easily observable staining patterns that provide information on the number of macropores and the type of pore flow. However, dyes can only diffuse through the soil matrix to a certain extent, making it challenging to accurately measure the number and area of macropores in soils with high pore density and poor permeability. Moreover, SO can cause significant damage and disturbance to soil macropores, preventing replication of the experiment at the same site. Furthermore, soil particles may adsorb dye, leading to an incomplete observation of preferential flow paths. Additionally, certain dyes may be unsafe and could potentially contaminate the soil (Filipovic et al., 2020). More importantly, macropore information obtained through SO has limited accuracy. Although it provides an approximate spatial distribution of macropores, it cannot quantitatively characterize morphological parameters, such as pore diameter, volume, and roundness, or functional parameters, such as pore connectivity and pore number.

Soil water retention curve, mercury intrusion porosimetry, and gas absorption

In addition to the macropore parameters derived from images, further characteristics, such as the distribution and quantity of macropores, can be determined by analyzing the curves. Soil water retention curve (SWRC), mercury intrusion porosimetry (MIP), and gas absorption (GA) are techniques commonly used for this purpose.

The use of measured SWRCs to obtain the soil pore distribution status is one of the most commonly employed and well-established methods in the study of soil pore distribution. A SWRC is a curve that indicates the capacity of a soil to hold and supply water. These curves can be divided into two categories: moisture absorption curves and dehumidification curves. By determining the SWRC, a series of values pertaining to soil water content and soil suction can be obtained. In the event that the pore structure is deemed to be regular, the soil pore size and pore size distribution can be indirectly inferred from the empirical formula between pore suction and effective pore radius. The water suction in large pores is relatively low, whereas that in small pores is high. Consequently, if there are a greater number of large pores in the soil, the SWRC will be relatively flat with a small slope. Conversely, if there are a greater number of small pores in the soil, the curve will be relatively steep with a large slope. Iiyama (2016) investigated the impact of drying conditions on the pore structure of soil through an in situ SWRC test. De Oliveira, Cássaro & Pires (2021) employed the traditional water use capacity function (first order derivative of SWRC) to estimate soil PSD.

The measurement of SWRC is frequently conducted in a laboratory setting rather than directly in situ. This is due to the inherent limitations of in situ measurements of SWRC, which are constrained by the natural wet/dry cycle. This cycle prevents soils from reaching a state of complete saturation or extreme dryness, thereby narrowing the scope of in situ SWRC measurements. Furthermore, there is frequently a discrepancy between laboratory and in situ measurements, and the application of laboratory-derived SWRC data to in situ conditions may result in errors. The process of laboratory measurement of SWRC typically necessitates the pre-treatment of soil samples, including water saturation and drying. These operations may induce alterations to the natural pore structure of the soil, thereby impeding the non-destructive characterization of the pores. The SWRC primarily reflects the water retention capacity of soil pores, simplifying the spatial distribution of the pore structure and the morphological characteristics. However, it is not capable of quantifying all the characteristics of soil. This is because it does not account for the characteristics of the pore space, such as the shape of the pore space, connectivity, and so on.

MIP is a precise and effective quantitative analysis technique of the size and distribution of pores in rock and soil masses. This relies primarily on the applied pressure to overcome the surface tension of the soil samples, allowing mercury to enter the macropore structure of the soil. Subsequently, the size and spatial distribution of macropores can be inferred indirectly based on the applied external pressure; therefore, MIP can be used to investigate the relationship between the macropore characteristics and macroscopic mechanical properties of soil samples. Compared to SO, MIP is simpler to operate. During mercury compression tests, only the inlet pressure and volume were measured and recorded. The soil pore parameters can then be calculated and analyzed using the appropriate formulas. MIP is frequently used to assess the porosity and pore size distribution of materials that feature open and interconnected pores (Wang et al., 2020a; Wang et al., 2020b). By meticulously analyzing the mercury intrusion and extrusion curves (Fig. 1C), comprehensive insights can be gained into the distribution patterns, connectivity, and permeability of macropores across the entire soil sample. Key pore characteristics such as total pore volume, average porosity, fractal dimension, and specific surface area can be quantitatively determined. MIP offers advantages in terms of wide measurement range and good repeatability of measurement results (Giesche, 2006). Unlike other pore characterization techniques, which have limited scale ranges, MIP can effectively detect pores across the entire range, particularly small pores. Consequently, MIP is indispensable for nanoscale pore studies.

For technical reasons, the MIP should be performed on completely dried soil samples. Therefore, sample pretreatment is essential for removing pore water and ensuring a smooth, bubble-free surface, minimizing the impact of mercury pressure on the soil structure (Stoltz, Cuisinier & Masrouri, 2012; Ma et al., 2020). However, the pretreatment process can alter the original macropore structure of the soil sample (Yang, Li & Nian, 2023). The primary limitation of MIP is its inability to identify closed macropore structure and accurately determine the geometric characteristics of isolated macropores. During the experimental process, the forced intrusion of non-wetting liquid mercury at high pressure can disrupt the macropore structure of the soil, leading to the deformation of the macropores and influencing the cumulative intrusion volume of mercury (Zhang et al., 2020). This can distort the detected macropore structure and result in inaccurate measurements of parameters, such as macroporosity. Additionally, pores located behind the neck of the bottle are prone to the ink-bottle effect, where mercury remains trapped in these pores upon withdrawal. Consequently, these pores may be misclassified as smaller, underestimating the total porosity (Zhang & Kong, 2013; Hashemi et al., 2015; Wang et al., 2020a; Wang et al., 2020b). Moreover, the larger the pores are, the greater the amount of mercury trapped inside them (Wang et al., 2022). Mercury compression testing involves expensive instrumentation and is labor-intensive and time-consuming. Furthermore, mercury is highly volatile and toxic, posing a potential risk in contaminated laboratories. After mercury pressure testing, mercury-contaminated samples require centralized disposal.

Similar to MIPs, GA operations rely on the principle that physical adsorption occurs at low temperatures on the surface of a sample placed in a gas system. Thus, the adsorption and desorption isotherms of the gases can be obtained. Through subsequent analysis, parameters such as pore morphology, pore-specific surface area, pore size distribution, and pore volume can be determined (Zhang & Kong, 2013). Nitrogen is commonly used as the adsorbate. However, owing to limitations in the sample size and manometer accuracy, nitrogen adsorption (NA) may overlook the presence of macropores. Unlike MIP, which often targets connected macropores, NA primarily characterizes micropores and mesopores. Consequently, it is often used alongside MIP to provide a more comprehensive qualitative and quantitative characterization of soil pore space (Zhu et al., 2019). MIP and GA exhibit similar limitations. Their pore analyses are based on simplified assumptions: pores are treated as cylindrical, and the measured pore radius represents the equivalent radius. This does not accurately or intuitively reflect the spatial structure of soil pores (Zhou, Lv & Li, 2009; Yang et al., 2013a; Yang et al., 2013b).

Nuclear magnetic resonance

Nuclear magnetic resonance (NMR) involves the resonance jump of atomic nuclei between energy levels under the influence of an external magnetic field. The distribution of soil pore throats is quantitatively characterized by acquiring the NMR relaxation signals of fluid hydrogen in the pore spaces of the samples. Subsequently, the parameters related to the pore structure and pore size are calculated and determined. One key constant in NMR analysis is transverse relaxation time T2. This time constant describes the recovery process of the transverse component of the nuclear magnetization intensity. The T2 distribution curve is particularly important for analyzing the macropore structure of the soil. By inverting the T2 distribution curve, the size and distribution of macropores within the soil can be determined (Gao et al., 2022). Kong et al. (2020) utilized NMR to obtain pore-size data from soil samples, as displayed in Fig. 2B. Pores with equivalent diameters ranging from 0.01–1 µm are primarily responsible for the porosity of the sample. Furthermore, Chu et al. (2022) applied NMR to investigate the pore damage law and generated T2 distribution curves and NMR pseudo-color images of the samples (Fig. 2C). NMR has a broad testing range and rapid speed, making it a common tool for studying the pore size distribution of soil under freeze-thaw conditions (Li et al., 2016). When compared over other techniques, such as SO, MIP, and GA, NMR offers several advantages. Not only can it determine the pore size and pore size distribution through the T2 distribution curve but it can also visualize the soil pore space in three dimensions using pseudo-color images. Furthermore, NMR allows for the quantification and estimation of water content in the soil as well as its spatial variability (Tian et al., 2014; Ma et al., 2020). This technique also permits observation of the infiltration and distribution of different solutes within the soil (Mantle, Sederman & Gladden, 2001).

In the NMR studies of soil samples, the transverse surface relaxation rate and pore shape factor are essential parameters for obtaining pore diameter variations from the NMR T2 spectrum. However, these parameters cannot be directly measured in practical applications, which is a significant limitation of NMR. To address this issue, Yang, Li & Nian (2023) developed a method that relied on scaled NMR T2 spectral curves obtained from MIP data. This approach enables the accurate determination of the NMR T2 spectra of soil samples, along with the pore radius conversion coefficient. Novotny et al. (2023) introduced diffusion in the internal field method, which eliminates the need to know the transverse surface relaxation rate. This method can successfully determine the pore size distribution of undisturbed soil samples and allow the estimation of water retention curves based on the pore size distribution. In addition, NMR spectroscopy is expensive. Owing to its fundamental principle, NMR is restricted to detecting the pore space occupied by a saturated medium. Prior to testing, samples require extensive pretreatment, including size adjustment, drying, vacuuming, and water saturation. This arduous process can potentially cause damage and alter the soil pore structure, preventing accurate characterization of the in situ pore structure. Consequently, NMR is not widely used for routine soil pore characterization.

Scanning electron microscopy

Scanning electron microscopy (SEM), a form of observation between transmission electron and optical microscopy, utilizes a focused high-energy electron beam to scan and image the surface of a specimen point-by-point. The interaction between the beam and material stimulates various physical information that is then collected, amplified, and re-imaged to characterize the microscopic morphology of the material. The primary imaging signals of SEM are secondary electrons. SEM offers several advantages, including a high resolution of 3–6 nm, a large depth of field, a strong sense of image stereoscopy, and a large adjustable range of magnification (100–200,000). It can also be used for various other analytical functions. In practical applications, it is often used to observe the microscopic morphologies of materials. Using SEM, researchers can observe the microparticle morphology and pore structure of soil samples on the micron or even submicron scale (Al-Mukhtar, Khattab & Alcover, 2012). SEM images are analyzed using image-processing software such as Image-Pro Plus (IPP). The software facilitates the easy acquisition of several two-dimensional pore parameters, including porosity, number, area, perimeter, equivalent diameter, shape coefficient, and anisotropy (Tang et al., 2012; Li & Li, 2017; Li et al., 2019a; Li et al., 2019b). Furthermore, the fractal dimensions could be derived using additional calculations (Liu et al., 2023). To investigate the impact of depth on the microstructure of loess, Li & Li (2017) utilized IPP to postprocess SEM images and segmented the pores based on their diameter. The results indicated that as the soil depth increased, pore connectivity became weaker, the areas of macropores and mesopores decreased, and the areas of small pores and micropores increased.

SEM is in exceptionally high demand for such samples. Before testing, the samples must undergo a series of preprocessing steps, including drying and gold spraying. During the dewatering and drying of the soil samples, there is a risk of shrinkage and deformation, making it challenging to maintain the original structural form. Additionally, owing to the random nature of SEM imaging, the sample must be scanned from left to right and from top to bottom, requiring meticulous operational skills. Furthermore, the sample size for SEM is small, typically less than 10 mm ×10 mm ×10 mm, necessitating the selection of appropriately magnified images for qualitative or quantitative analyses. If the magnification is too high, the number of observable pores diminishes, reducing the representative value and significance of the microscale guidance (Zhang & Kong, 2013). Post-processing using image processing software, such as IPP, cannot completely eliminate human error and image noise, which can also impact pore identification. The key limitation of SEM is its inability to probe and visualize the characteristics beneath the surface or within a three-dimensional body, as it only allows the observation of the surface morphology of the sample.

Computed tomography

CT, which is a widely used diagnostic tool in the medical field, has recently been adopted in soil science applications (Taina, Heck & Elliot, 2008). This advancement allows for the nondestructive, quantitative characterization of the pore structure of in situ soils. The underlying principle involves placing soil samples in a scanning chamber and utilizing X-rays to scan each layer to obtain cross-sectional samples. Macropore extraction relies on the distinct X-ray absorption capacity of media of various densities within the soil (Figs. 3A, 3B). Compared with other techniques, the primary advantage of CT is its ability to generate continuous images reflecting the pore structure of the soil without any preprocessing. This capability offers a true representation of the three-dimensional structure of soil pores, ensuring nondestructive and accurate characterization (Cheng, Liu & Zhang, 2012; Wang, Qian & Gao, 2021). By combining multiple two-dimensional slices obtained from scanning with image processing software such as Avizo, VG Studio MAX, MATLAB, ImageJ, MAVI, Blob3D, and Fiji, the three-dimensional spatial structure of a sample can be reconstructed (Wen, Chen & Shao, 2022). This approach reveals a wealth of information regarding the three-dimensional characteristics of the soil structure, including the pore connectivity, sphericity, and fractal dimensions (Figs. 3C, 3D, 3E). Currently, CT is utilized to investigate how various land-use types (Wang, Qin & Bai, 2018), vegetation types (Kan et al., 2023), tillage practices (Wang, Qian & Gao, 2021), soil improvement measures (Yang et al., 2013a; Yang et al., 2013b), and reclamation age (Cai et al., 2018) impact soil pore characteristics.

Currently, CT is only suitable for high-clarity testing in a controlled laboratory environment and cannot be used to conduct nondestructive in situ scanning of soil directly in the field. When performing CT, it is necessary to select an appropriate sample size and resolution for pore identification. The scanning resolution of CT is inversely proportional to the maximum size of the sample, and image resolution plays a decisive role in pore quantification (Li et al., 2019a; Li et al., 2019b). Therefore, the selection of the sample size is crucial. A low resolution may result in small pores and inactive voids being overlooked, leading to inaccurate calculations of parameters such as total porosity and connectivity. Although reducing the sample size can improve the scanning resolution, the results obtained may be less representative because of the strong non-homogeneity of the soil space and the inability of small samples to contain large pores (Cheng, Liu & Zhang, 2012). In the image-processing stage, CT has the clear disadvantage of involving a large number of cumbersome and time-consuming image analyses. Improving the scanning accuracy blindly can increase the amount of data, leading to greater difficulties in postprocessing. The most significant challenge in analyzing CT images is the effective segmentation and extraction of pores. A low scanning resolution can result in poor contrast at the solid/porous interface of CT images, potentially leading to errors in contact position judgments. Although medical and industrial CT share the same underlying physical principles, their distinct applications require distinct system layouts and designs, resulting in different resolutions and sample size requirements. Compared to industrial CT, medical CT is less expensive and allows larger sample sizes to be analyzed; however, it exhibits relatively low resolution. Currently, medical CT has a resolution of approximately 600 µm, whereas industrial CT ranges from 0.5–100 µm (Liu et al., 2016; Yi et al., 2020). Researchers can select the appropriate CT based on the sample size, pore characterization requirements, and financial budgets.

Research hotspots

In Fig. 5, VOSviewer was used to visualize the co-occurrence density of keywords at this stage. Table 5 lists the top 30 high-frequency keywords from the articles at this stage. Combining Fig. 5 and Table 5, the main research hotspots in this field are centered around “quantification”, “organic-matter”, “soil structure”, “physical-properties”, “tillage”, “image analysis”, “preferential flow”, “transport”, “solute transport”, “management”, “pore structure”, and other related topics. Among the 279 articles, “quantification” appears most frequently, with a total of 64 occurrences, followed by “organic-matter” with 49 occurrences.

Figure 6 shows the clustering density of keywords in the research articles on the CT characterization of soil macropores using VOSviewer. The tool clusters related topics within the research domain and indicates their relevance among these topics. In the mapping, distinct colors represent the categories assigned to each topic, whereas the size of the topic labels reflects their frequency and occurrence. Larger labels indicate higher frequencies. The distance between topics signifies their similarity; topics that are closer share more similarities. Notably, 92 keywords appeared more than six times. VOSviewer categorizes these keywords into four clusters, purple, blue, green, and yellow, revealing four distinct research fields based on their relevance. For instance, the purple cluster (Group 1) focuses on analyzing the impact of soil pore space on “water” through “image analysis.” This cluster highlights processes such as water flow and pollutant transport, which can be influenced by various macropore properties. The keywords in this group include “preferential flow”, “solute transport”, “hydraulic conductivity”, “water-flow”, and “hydraulic conductivity”. The blue cluster (Group 2) emphasizes CT-characterizable pore properties such as “physical-properties”, “macroporosity”, “fractal dimension”, “pore size contribution”, “bulk density”, and “macropore networks”. Groups 3 and 4—the green and yellow clusters—primarily explore factors that impact soil pore structure, including “soil structure”, “tillage” practices, “vegetation” and “root” systems, land use, and “management” practices.