1. INTRODUCTION

Karst areas in southwest China, a humid monsoon region with distinct dry and rainy seasons, are characterised by typical regional soil and vegetation that distinguish them from other regions at the same latitude (Peng et al., 2008). Carbonate-derived laterite is widely distributed in karst areas with high clay contents, low permeabilities, and high water-holding capacities (Waller and Wallender, 1993; Liu, et al., 2016; 2020). These soil features determine the calciphile and drought endurance of local vegetation (Peng et al., 2008). Moreover, the complex and unique topographical and geological conditions of karst areas lead to water and nutrient loss (Wang et al., 2019). Given these circumstances, the importance and specificity of these regions have alarmed people about their ecological conditions.
Cracks are physical phenomena that exist in clayey and expansive soils (Sima et al., 2014; Acharya et al., 2015). In karsts, the alternating wet and dry conditions and the special dilation-shrinkage characteristics of carbonate-derived laterite allow for soil cracks to form with ease (Liao et al., 2000; Zhang et al., 2021). Additionally, shallow subsurface cracks may be influenced by plant roots (Shem et al., 2009; Zhu et al., 2020), burrowing fauna (Li et al., 2019), and artificial factors (Mossadeghi-Björklund et al., 2016; Tang et al., 2016). For example, Eucalyptus robusta can absorb approximately 9 t hm−2 year−1 of carbon dioxide while releasing oxygen (Ghannoum et al., 2010; Teixeira et al., 2020); further, rapid water absorption through its root systems result in soil crevices and cracks (Shem et al., 2009).
Cracks can enhance water infiltration and induce preferential flows (Zhang et al., 2014), making water one of the most important factors limiting karst forest development (Zhang et al., 2018; Liu et al., 2020). Preferential flow, generated by the rapid penetration of moisture through soil pore channels, is a nonuniform and unstable water flow (Bouma and De Laat, 1981; Sheng et al., 2014). Crack properties including configuration, orientation, inclusion, angle, depth, and volume can affect water distribution and preferential flow (Zhao et al., 2014). For example, the configuration of a linear crack has a high saturated hydraulic conductivity compared with other shapes and thus develops preferential flow (Liu and She, 2020), whereas regular rectangular crack networks made of steel sheets generate preferential flow only under high-intensity precipitation or irrigation conditions (Zhu et al., 2018). Inclusions within soil cracks are normally coarse particles and rock fragments in karsts (Chen et al., 2011). Although cracks and fissures in nature exist with complex combinations of properties (Zhang et al., 2014; Zhu et al., 2020), the small-scale control of the preferential flow of isolated cracks is critical for water loss in soil-epikarst zones (Liu and She, 2020).
Preferential flow visibility is usually achieved by Brilliant Blue FCF (C.I. Food Blue 2), which is a tracer that marks the area where stained water flows (Bouma and De Laat, 1981; Hagedorn and Bundt, 2002) and traces the flow paths to reveal soil profile characteristics (Hagedorn and Bundt, 2002; Sheng et al., 2009). Recently, ground-penetrating radar (GPR) has been successfully applied to karsts to reveal structural information about the subsurface and transform invisible geological features into trended signals (Estrada-Medina et al., 2010; Fernandes et al., 2015), thereby providing more detailed information about soil profiles. GPR can accurately locate soil cracks and identify crack widths as low as 1–2 mm (Levatti et al., 2017). The combination of‘ time-lapse GPR surveys and infiltration experiments has revealed channelling soil water characteristics (Di Prima et al., 2022). However, the combination of field-based GPR data and lab-based infiltration experiments on crack properties has not been tested to generalise the flow infiltration process.
To clarify whether soil cracks enhanced preferential flow and effected infiltration processes, a case study approach was chosen. GPR was applied to physically identify the natural soil cracks and place “manmade cracks” in transparent Plexiglas columns to observe infiltration over time. This study aimed to i) identify the soil crack properties through GPR envelopes with excavated pedons; ii) evaluate the effects of crack inclusion, width, and configuration on infiltration processes and preferential flow; and iii) discuss the preferential flow characteristics caused by different cracks to further explore the infiltration mechanism in karsts.