Response of the winter soil moisture of different vegetation types to rainfall events in karst slope land

Soil moisture, rainfall, and vegetation are interactive processes (Andrew et al. 2003; Daniele et al. 2008; Eunhyung & Sanghyun 2019). Soil moisture affects the distribution of vegetation types (Le et al. 2013; Zhao et al. 2019). Moreover, the vegetation type is considered the main factor controlling soil moisture (Chen et al. 2007). Among them, vegetation characteristics, such as bush layers, affect the redistribution of rainfall, thereby affecting the replenishment and change of soil moisture (Duan et al. 2016). Therefore, if the response of soil moisture to rainfall for different vegetation types can be clearly understood, it would help understand the interactions between soil moisture, vegetation, and precipitation and may have scientific significance for guiding the implementation of various ecological restoration projects (Hou et al. 2018).

In recent years, increasing attention has been paid to the response of soil moisture to rainfall for different vegetation types. The results of their studies showed that in different periods, the response of soil moisture of different vegetation types to rainfall is different. For example, Li et al. (2015) found that there are significant differences in the response patterns of soil moisture to rainfall in vegetable and forestland soils throughout the year. Also, Tian et al. (2019) found that in the vegetation growing season, the profile distribution patterns of soil moisture response of shrubs and bare land vegetation are similar, while the profile distribution patterns of meadows, high-coverage grasslands, and medium-coverage grasslands are significantly different.

The above research results provide effective method references and valuable information for further research in non-karst areas; however, the research results may not be applicable to karst areas because the properties of bedrock and soil are also important factors that affect the dynamic changes of soil moisture (Zhao et al. 2020). Compared with other areas, in karst areas, the bedrock is mostly carbonate rock and karstification is strong, which makes the soil layer thin in such areas and leads to broken ground surfaces (Deng et al. 2020) and strong soil permeability. These characteristics may cause the magnitude, speed, and time of soil moisture response to rainfall to be different from those in other regions. Thus, the response of soil moisture to rainfall in karst regions differs from that in other regions.

Previous research has focused on the response of soil moisture to rainfall for different vegetation types in karst regions (Zhang et al. 2011; Fu et al. 2016; Yolanda et al. 2016; Yang et al. 2019). For example, Yang et al. (2021) found that under the same rainfall conditions, the response time of grassland soil moisture to rainfall is shorter than that of soil moisture of farmland and bare land and that the sensitivity of soil moisture in grassland is higher than that in farmland and bare land. Zhou et al. (2019) found that with the increase in rainfall, the increase in the soil water content in forestland and shrubland is greater than that in bare land and grassland. These studies have also explained the response of soil moisture of different vegetation types to rainfall in terms of response magnitude and response time. However, the current studies on the response of soil moisture to rainfall mainly focus on non-winter seasons (Chen et al. 2010; Guo et al. 2016), and there is not enough research conducted during the winter season. Also, the changes in soil moisture in winter still form an important part of hydrological processes and affect vegetation restoration and ecological reconstruction. On the one hand, because the climate in the research region is a humid mid-subtropical monsoon climate and the temperature in winter is not too cold, some vegetations continue to grow (Wang et al. 2020). However, vegetation growth inevitably requires a certain amount of moisture to support it. In winter, when rainfall is not abundant, soil moisture conditions still threaten vegetation growth, which makes it particularly important to understand the response of soil moisture to rainfall in winter (Zhang et al. 2012; Li et al. 2019; Ding et al. 2020). On the other hand, although winter precipitation is relatively low in humid karst areas (Xiao et al. 2021), soil moisture changes are still an important part of the hydrological process. Moisture stored in shallow soil layers in karst areas (Chen et al. 2021) becomes more important for moisture storage during winter drought conditions (Fu et al. 2016; Yang et al. 2019). Therefore, the response of soil moisture of different vegetation types to rainfall in winter is still an important topic that needs to be studied accurately.

Hence, to investigate the dynamic response of soil moisture of different vegetation types to rainfall in winter, the response magnitude, response time, and response rate of soil moisture of different vegetation types to rainfall in different soil layers were calculated and the dynamic response of soil moisture to rainfall was analyzed. Thus, the soil hydrological processes related to rainfall for different vegetation types in karst areas were revealed, and important reference values for the key management parameters of water resources were provided.

The experimental site is located in Guanling County, southwest of Guizhou Province, China. The geographic coordinates are 105°43′E and 25°44′N, and the elevation is 367–1,831 m. It belongs to the Wumeng Mountains and has large undulations. The types of landforms in this territory are complex and diverse, the surface is broken, and limestone is widely distributed, which makes it a typical karst landform. The climate is humid mid-subtropical monsoon with an average annual temperature of 16.2 °C. The annual rainfall ranges between 1,205.1 and 1,656.8 mm and is concentrated between June and August, comprising ∼52–55% of the annual rainfall, while the rainfall in winter accounts for ∼3–5% of the annual rainfall. The main soil type in the area is Rendzic Leptosols. Natural vegetation here mainly consists of forests, shrubs, grass, and other vegetation. The forestland is mostly evergreen broad-leaved forest mixed with a small number of deciduous trees and cypress forests. Artificial vegetation mainly consists of corn, plantain, emperor bamboo grass, and sugar cane.

For each vegetation type, five monitoring points with a size of 5 m × 5 m, an interval of ∼20 m, and four sample plots were randomly selected (a total of 20 monitoring points). The HOBO H21_USB soil moisture monitoring system was used at each monitoring point, and the ECH2O-5 probe (The METER Group, Inc., USA) was horizontally inserted in the soil layers at different depths (0–5 cm, 5–10 cm, 10–15 cm, and 15–20 cm). Real-time monitoring of soil moisture changes during rainfall was performed, the data collection frequency was 10 min, and the monitoring time was from July 5, 2019 to January 6, 2020. Considering the actual start and end dates of winter in the study area, soil moisture data for this study were selected from November 5, 2019 to January 6, 2020.

The HOBO RG3_M automatic recording rainfall thermometer observation map was used for capturing the rainfall data. Concerning the monitoring points of each vegetation type, two monitoring points were selected to place this type of rainfall thermometer on the forest surface to observe the penetrating rainfall in the forest and the near-surface temperature under the vegetation. The data collection frequency and the observation period were synchronized with the soil moisture data.

In each plot, soil samples were collected from four soil layers (5, 10, 15, and 20 cm) to determine the soil properties. A 100-cm³ ring knife was used (bottom surface: 20 cm², height: 5 cm) to collect undisturbed soil from the four soil layers of each sample plot to determine the bulk density, total porosity, capillary porosity, noncapillary porosity, and saturated hydraulic conductivity of the soil. The sampling time was July 5–21, 2019. The organic matter content of the soil was determined using the potassium dichromate titration method. Moreover, the mechanical composition of the soil was measured using a laser particle size analyzer.

The soil bulk density, soil porosity, saturated hydraulic conductivity, soil particle size, and organic matter content are important factors that affect the dynamic changes in soil moisture. In this study, they were used as auxiliary factors to explain the dynamic changes in soil moisture for different vegetation types.

There was a total of 20 monitoring points for the 4 sample plots, out of which there were 5 monitoring points for each sample plot. Each monitoring point was divided into four soil layers for monitoring. A total of 80 soil moisture probes were used to obtain time-series soil moisture data through probe monitoring. The average soil moisture time series for each soil layer of each vegetation type at the five monitoring points were calculated as the soil moisture time series of the soil layer of the vegetation type. The average value of the measurement results of the 20 soil samples at the 5 monitoring points and the 4 soil layers of each vegetation type was calculated as the soil properties.

Usually, after rainfall, it can be assumed that the rainfall infiltration has reached a certain soil profile layer when its soil moisture starts to increase (Laio et al. 2001; Wang et al. 2008; Lozano-Parra et al. 2015). Therefore, the increase of soil moisture at 5, 10, 15, and 20 cm after rainfall was used to define and describe the soil response process. Hence, the rainfall impact period data was recorded 10 min before the beginning of rainfall to 12 h after the end of rainfall. Considering the accuracy of the soil moisture probe, a rainfall event with an increase in soil moisture of more than 0.2% (Lin & Zhou 2008) was considered an effective rainfall event. We then calculated the change in the soil moisture during rainfall. We used the time of the first soil moisture increase, which was greater than 0.2%, as the start of the soil response and the time of the soil moisture increase, which was less than 0.2%, as the end of the soil response.

There were 14 rainfall events during the observation period. According to the standards of the National Meteorological Administration of China, the rainfall events were then divided into light rain events and moderate rain events. Rainfall of less than 10 mm within 24 h was defined as light rainfall, and rainfall between 10 and 25 mm within 24 h was defined as moderate rainfall. Of the total rainfall events, 12 were light rainfall events and 2 were moderate rainfall events.

Quantification of soil moisture response can provide reference indicators for process-based soil hydrological modeling (Green & Erskine 2011; Clark et al. 2017). Therefore, in this study, we used a series of quantitative indicators to characterize the response of soil moisture to quantitatively describe the dynamic response of soil moisture to rainfall events under different vegetation types.

Among them, i represents the soil layer (i = 2, 3, 4) and ASWI^j_i−1 and ASWI^J_i are the soil moisture increments of the i − 1 and i layers in the jth soil response event, respectively.

Among them, ET_j and ST_j are the start and stop times of the jth soil response event, respectively.

As shown in Figure 5(b), during a light rainfall event, the soil moisture response duration (Duration) of the grassland and arable land changed slightly within the same soil layer, and the maximum Duration was observed in the third grassland layer. As shown in Figure 6(b), during a moderate rainfall event, the Duration of each vegetation type varied significantly between the different soil layers, from 0.23 to 54 h. Compared with the other vegetation types, the Duration of each grassland soil layer was the highest, and the Duration in the case of moderate rainfall events was significantly greater than that in the case of light rainfall events.

This indicated that in the case of light rainfall, the grassland soil moisture response lasted longer than that of the arable land. Compared with the other vegetation types, in the case of moderate rainfall, the grassland soil moisture response was the longest, and the soil moisture response time difference between the adjacent soil layers of the shrubland was minimum.

As shown in Figure 7, in the case of light rain, the S_max and S_mean of the grassland and arable land decreased with the increase in the soil layer, and S_max was the highest (1.62 vol.%/min) in the first layer. As shown in Figure 8, in the case of moderate rain, the S_max and S_mean of the first grassland layer and the farmland soil moisture were much higher than those of the other soil layers, indicating that the soil moisture response speed was fastest in this layer. In the case of moderate rain, the S_max and S_mean of the shrubs decreased with the increase in the soil layer. The S_max of the forestland increased and the S_mean decreased with the increase in the soil layer. In general, the response speed of soil moisture of all vegetation types was fastest in the first layer. The response speed of the soil moisture in the case of moderate rain was significantly greater than that of light rain, and the response speed of the grassland soil moisture to rainfall was significantly greater than those of the other vegetation types.

Through this study, it was found that compared with other vegetation types, karst slopes in grassland have the largest response to rainfall, longest response duration, and fastest response speed in winter. These results differ from the non-winter research results of Yang et al. (2021) in karst areas. Thus, the responses of soil moisture to rainfall for different vegetation types in karst areas in winter and non-winter seasons are different. The reason may be the seasonal difference in different vegetation types (Duan et al. 2016). For example, compared with forestland and shrubland, grassland in winter has less litter and low vegetation coverage. Compared with winter grassland, arable land has relatively high vegetation coverage and relatively large rainfall interception due to its dry winter crops. Therefore, compared with other vegetation types, the indicators of the grassland soil moisture response to rainfall in winter are larger than those of other vegetation types. In addition, in winter karst environments, compared with other vegetation types, the grassland vegetation coverage is low and the interception is relatively small. In addition, karst rock fissures are currently being developed (Li et al. 2020). After rainfall, most of the surface runoff quickly enters the ground along with rock fissures, resulting in faster and longer responses to rainfall in winter grassland.

The geological background of karst areas is complex. Even if they are all carbonate rocks, the parent rock formed by the soil is different, and the soil bulk density, soil pore water, capillary water holding capacity, and other soil properties are different (Sheng et al. 2018). For example, the Rendzic Leptosols formed by calcareous dolomite were compared with the Rendzic Leptosols formed by pure limestone, which has strong water permeability, small water holding capacity, and low natural water content (Huang 1988). These differences in soil properties lead to differences in the response of soil moisture of different vegetation types during rainfall. The sample plots used in this study are located on a pure limestone slope, which only represents the response of the soil moisture of different vegetation types to rainfall on a pure limestone slope.

Second, karst areas have great spatial heterogeneity. Even if the bedrock is the same, the vegetation and seasonal characteristics of different climate types are different, which may cause differences in the response of soil moisture to rainfall (Chen et al. 2009). For example, in the same limestone karst area, in the case of a humid mid-subtropical climate, the seasonal vegetation differences are more obvious. In summer, the vegetation is dark green, and the canopy closure is relatively large. In autumn, yellowish-brown patches appear, and in winter, most leaves fall off, leaving only dry branches. In the semi-humid climate of South Asia, although vegetation significantly differs from summer to winter, only some dominant species are deciduous in winter and the forest layer is complex, where there are abundant plants in the middle forest layers, and the winding density of rattan (Huang 1988). These different vegetation characteristics may cause the interception of winter vegetation of this climate type to be greater than that of the mid-subtropical humid climate type, which may lead to differences in the response of winter vegetation soil moisture to rainfall (Chen et al. 2010). This study area belongs to the humid mid-subtropical climate, which only represents the response of different vegetation soil moistures to rainfall in this climate type, while the response of vegetation soil moisture to rainfall in other climate types needs further investigation.

In conclusion, this study can better explain the response of soil moisture to rainfall in winter for different vegetation types in the humid mid-subtropical monsoon climate of pure limestone slope land. The results of this study have enriched the understanding of the response of soil moisture of different vegetation types to rainfall in karst slopes and supplemented this knowledge system. This study also showed certain reference values for better understanding the interaction between the soil moisture, vegetation, and precipitation.

A limitation of this study is that it did not conduct enough replicate experiments, because our study area was not very large; therefore, we were unable to include several replicas of the sample plots. Additionally, to minimize the uncertainty caused by variations in lithology, soil, and climate, we wanted to ensure that the sample plots for all four vegetation types were located on a single slope. However, such locations were difficult to find. Even if a plot had been established in the vicinity of the study area, the topography or vegetation type could have varied, and topographic variation is another factor that affects soil water content (Gao et al. 2020). If the same plots had been established away from the study area, the climate, lithology, and soil conditions could have changed. This is particularly important in mountainous environments where there can be significant differences in local climates. In addition, the vegetation type at a sample site can change over years of monitoring. The study area has a humid subtropical monsoon climate with rapid vegetation growth that results in abandoned grassland usually being replaced by shrubs within a few years. For these reasons, it was difficult to prepare sufficient replication within the study. Therefore, the study did not use data from multiple winters.

To improve the ecological environment in the karst region, since the 1990s, China has approved and implemented a large number of ecological restoration projects, for example, the Grain for Green Program and the Rocky Desertification Control Project (Li et al. 2022). Most ecological restoration projects not only aim to recover ecosystem structures and functions but also focus on improving human wellbeing (Qiu et al. 2022). Our results have two implications for vegetation restoration projects in karst regions. The first is that in humid karst areas, where vegetation can still grow during winter, winter soil moisture is an important part of the hydrological process. Therefore, the relationship between winter soil moisture and vegetation growth should not be ignored in the implementation of ecological restoration projects. The second is that in humid karst areas, the response pattern of soil moisture to rainfall is different for different vegetation types during the winter and non-winter seasons. Therefore, when implementing ecological restoration projects, different ecological restoration modes should be selected according to local conditions and more attention should be paid to soil moisture storage in grasslands during winter.

In this study, to explain the response of soil moisture of four vegetation types to rainfall in karst areas in winter, the response magnitude, response time, and response speed of the soil moisture of different vegetation types to rainfall were calculated using the time-series data of soil moisture of different vegetation types.

The results showed that the response of soil moisture differed between different vegetation types in winter and non-winter seasons in karst areas. Among the four vegetation types, soil moisture response magnitude to rainfall in grassland and arable land had a similar distribution pattern along different soil depths, while in scrubland and forestland, it decreased gradually along the soil depth. In addition, compared with other vegetation types, for grassland soil moisture, the response magnitude, response duration, and response speed to rainfall were the largest, longest, and fastest, respectively. Overall, the findings of this study complement the understanding of the soil moisture of different vegetation types in karst slope land with regard to the rainfall response and enrich the related knowledge system. However, whether the response of soil water to rainfall in winter under other conditions, such as different bedrocks, climate types, vegetation types, and karst areas, will have new characteristics is a further question that needs further investigation.

E.Y. and Q.Z. conceived and wrote the manuscript; W.Y. and D.P. were involved in the experimental assays; and Y.W. polished the language of the manuscript.

This work was supported by the Joint Fund of the National Natural Science Foundation of China and the Karst Science Research Center of Guizhou province [grant number U1812401]; the National Science Foundation of China [grant number 41761003]; the Scientific and Technological Research Project of Guizhou Province [grant number Qiankehe Jichu [2019]1433 and Qiankehe Pingtai Rencai [2017]5726]; and the Foundation Programme for Outstanding Talents in Higher Education Institutions of Guizhou Province [grant number [2018]042].

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.