4.2 Effect of land cover on the soil wetting process
In the case of a certain rainfall amount, land-cover type was the key factor that determined the SM response. Different land-cover patterns have different vegetation structures, coverage (canopy, herb, litter), root depths and densities, and soil textures, which influence the soil wetting process after precipitation (Chen et al., 2007a; Jia et al., 2017; Li and Shao, 2006; Zhang et al., 2021).
First, the rainfall-SM average response depth increased by 15.7%, 1.7%, and 26.7% when the crop was transformed to forest, shrub, and grass sites, respectively (Supplemental Fig. S1). Vegetation recovery altered the plant coverage and structure radically, resulting in changes in the way and process of soil acceptance from precipitation (Lozano-Parra et al., 2015; Wang et al., 2013). Specifically, planted forests, with a higher canopy and greater litter coverage, intercepted the large rainfall and weakened the dynamic potential energy of heavy rains in large amounts. One part of the rainfall is intercepted by the canopy and flows down the trunk to the soil directly, and the part of rainfall may reach the litter layer as throughfall. The thick and rough litter hampered the formation of surface runoff and percolated rainwater infiltration to the deep soil layer. This vegetation structure promoted an SM response deeper in planted forests after heavy rains (Fig. 3a). Abandoned grass improved soil texture, decreased soil bulk density, and increased soil porosity, which was also beneficial for rainfall-SM infiltration (Sun et al., 2018). Especially in Fig. 3c, permitting percolation to the 100 cm depth occurred in heavy rains. Planted shrubs with a high coverage could not reach the deepest monitoring layer, probably due to SM consumption by root absorption and vegetation evapotranspiration (Yu et al., 2017). Additionally, the shrub site showed a shallower average wetting depth than that of bare land across the 13 rainfall events (Supplemental Fig. S1). This is because smaller rainfall amounts with short durations have difficulty penetrating dense canopies and thick litter layers of shrubs to percolate into subsurface soil (Wang et al., 2013). Compared to afforestation covers, bare land had a deeper average soil wetting depth than shrub and crop sites after precipitation. Because no vegetation coverage results in direct rainfall and surface soil cracks that increased soil water percolation access together promoted rainwater infiltration into a deeper layer in bare land. These results suggested that land cover strongly influenced the rainfall-SM response depth, while revegetation promoted rainwater percolation to deeper depths across the profile.
Second, the rainfall-surface SM response lag time (RT) and lasting time changed significantly due to land-cover change. The results showed that shrub has the smallest RT and lasting time at the 10-cm depth (Table 3 and Fig. 4). This case was correlated to the initial SM that influenced the speed of rainwater infiltration (Liu et al., 2019). Due to excessive soil water consumption by vast evapotranspiration (including soil and plants) and deeper root water absorption, the shrubland showed the smallest SM across the entire profile, especially in the surface layer. The extremely dry situation accelerated soil water permeation, which was demonstrated by Li et al. (2015) and Sun et al. (2018), who illustrated that a lower initial soil water content accelerated the topsoil permeability rate. Thus, planted shrubs have the shortest response and duration. Compared to shrubs, planted forests and abandoned grass show longer RT and shorter durations. Due to the existence of a dense canopy and a thick litter layer, rainfall is blocked to a large extent and postpones the contact time of rainfall-SM, resulting in a longer response. Meanwhile, much infiltrated rainwater percolates into deep soil or counterbalanced with plant consumption, also decreasing the SM duration at the 10-cm depth. Thus, the effect of afforestation on the surface SM response time or duration after rainfall was significantly different. However, with no vegetation coverage, bare land showed a smaller RT and the longest duration. Because bare land does not have vegetation shade or litter cover, rainwater directly contacts the surface soil and decreases the SM feedback time. Therefore, vegetation recovery delays the surface SM response time but shortens the duration after precipitation, except for shrubs.
Finally, the change in land-cover type influenced the SM response time and wetting front velocity (WFV). For example, the crop site required more time to respond to rainfall, resulting in the longest ART and the slowest WFV across the entire profile. Vegetation restoration altered this situation. Vegetation restoration no only shortened the ART by approximately 30.8%, 33.7%, and 17.4% but also accelerated the WFV by approximately 77.9%, 36.7%, and 87.0% for planted forest, shrub, and grass, respectively, over the 1-m depth (Table 4 and Supplemental Fig. S3). This difference in soil permeability was related to soil texture. Afforestation improved soil texture, decreased soil bulk density and increased soil porosity, which facilitated rainwater percolation across the profile (Sun et al., 2018). Thus, revegetation had the shorter RT and higher WFV, which benifited for alleviating the soil water deficit caused by plant water consumption.
In conclusion, vegetation restoration has a significant impact on ecohydrological processes, including rainfall-SM response depth, RT, and WFV. Despite afforestation consuming more soil water and causing soil desiccation to a large extent. However, the rainfall-SM response process showed that revegetation (mean value of forest, shrub, and grass) increased the soil wetting depth by 14.7%, shortened the ART by 27.3%, and accelerated the WFV by 67.2% compared to cropland over the entire profile. Revegetation contributed to rainwater infiltration and deeper SM replenishment in the rainy season.