Spatial patterns in soil physicochemical and microbiological properties in a grassland adjacent to a newly built lake

Abstract Soil water content (SWC) is an important determinant for nutrient cycling and microorganism activity in the grassland ecosystem. Lakes have a positive effect on the water supply of the neighboring ecosystem. However, information evaluating whether newly built lakes improve the physiochemical properties and microorganism activity of adjacent grassland soil is rare. A 15‐hectare artificial lake with a 2 m depth was built on grazed grassland to determine whether the change of soil physiochemical properties and microorganism activity of the adjacent grassland depended on the distance from the lake. SWC and total nitrogen (TN) were greater within 150 m of the lake than at distances over 150 m from the lake. The total organic carbon (TOC) increased first at 100–150 m from the lake and then decreased. The soil microbial biomass and the bacterial and fungal contents increased with increasing years after the construction of the lake. Gram‐negative bacteria and methanotrophic bacteria were greater within a 30 m distance of the lake. Over 60 m away from the lake, Actinobacteria, gram‐positive bacteria, and anaerobic bacteria showed higher abundances. In the area near the lake (<250 m distance), microorganisms were strongly correlated with SWC, EC, TN, and TOC and greatly correlated with the changes of total phosphorous (TP) and pH when the distance from the lake was over 250 m. The results indicated that the newly built lake could be a driving factor for improving the physiochemical properties and microorganism activity of adjacent grassland soil within a certain range.

In soil ecosystems, more than 95% of soil microorganisms live in the surface soil and play important roles in the cycle of nutrients, soil health and ecosystem productivity (Leeuwen et al., 2017;Moll et al., 2017;Pulleman, Six, Uyl, Marinissen, & Jongmans, 2005;Zelles, 1999). Soil microorganisms are sensitive to soil moisture changes, and changes in soil moisture directly affect the metabolism, quantity, and diversity of soil microorganisms (Brockett, Prescott, & Grayston, 2012;McMahon & Chapelle, 1991;Zhao et al., 2018). Changes in the structure and diversity of the soil microbial community were closely related to hydrological conditions (Foulquier, Volat, Neyra, Bornette, & Montuelle, 2013). In a tall-grass prairie of the American Midwest, Fierer et al. (2013) found that the main driving factors of soil bacterial community diversity were the annual average precipitation and total precipitation. In a simulated rainfall experiment, Ma, Huang, Guo, Wang, and Xiao (2012) found that the carbon and nitrogen biomass and activity of soil microbes increased with increasing available water.
In grassland ecosystems, studies on the relationship among soil physiochemical properties, soil microorganisms, and moisture mostly focused on natural rainfall or simulated rainfall (Bastida et al., 2017;Guo, Li, et al., 2018;Guo, Weise, Fiedler, Lohmann, & Tietjen, 2018;Zhao et al., 2016), but the uncertainty and inhomogeneity of natural rainfall restricts its effectiveness for the evaluation of their dynamics and relationships. The presence of lakes generally has a positive effect on the water supply of neighboring ecosystems (Gong et al., 2017;Zhang, Zhao, & Zhang, 2012;Zhao et al., 2016). Therefore, if a lake were built on a grassland, then a simulated wet ecosystem with adjacent grasslands would be formed.
The water shortage of grasslands in arid and semiarid regions could be mitigated. The soil environment, including soil physiochemical properties and soil microorganism activity, of adjacent grasslands would potentially improve. However, information related to the effect of the presence of a lake on the soil physiochemical traits and microbes of grasslands in the direct vicinity is rare. Therefore, this experiment studied the temporal and spatial responses of the physiochemical properties and microorganisms of grassland soil adjacent to a newly built lake. We hypothesized that (a) the soil water content of grasslands near the lake increased and the spatial distribution of soil physicochemical properties of adjacent grasslands changed after lake construction; and (b) the content of taxonomic groups of soil microbes increased with increasing years after the construction of the lake.

| Study site
The study was conducted in the National Field Station of Grassland Ecosystem at Guyuan County, Hebei Province, China (41°44′N, 115°42′E, 1,460 m a.s.l.). The entire station is located in a typical semiarid continental monsoon climate zone with an annual mean precipitation of 400 mm, approximately 80% of which falls in the growing season from July to September. The mean annual temperature is 1.4 ℃, and there are 80-95 frost free days. There are 2,930 annual hours of sunshine, and the annual evaporation is 1,735 mm.

| Experimental design and soil sampling
A 15-hectare artificial lake with a 2 m depth was built on the grazing grassland in the National Field Station of Grassland Ecosystem at Guyuan County, Hebei Province, China, in June 2013 and completed in June 2014. The site of lake construction and the study grassland belonged to the same grazing grassland before the lake was constructed. The lake was constructed by digging the grazing grassland to a 2-m depth and moving out the soil. Initial lake water mainly came from groundwater below 150 m. After water filling was completed, there was no dramatic fluctuation in the water table of the lake. To study the spatiotemporal responses of the soil physiochemical properties and microorganisms of the adjacent grassland to the newly built lake, eight 2 m wide × 2 m long × 2 m high iron cages made of fine iron bars were set perpendicularly at 10, 30, 60, 100, 150, 250, 400, and 600 m distances from the edge of the newly built lake in June 2014, which were named s1, s2, s3, s4, s5, s6, s7, and s8 in the text, respectively. The areas outside of the cages remained in their original grazing state (continuous grazing, 10 goats/ha), and the cages prevented the disturbance of animals from affecting the accuracy of data collection inside the cages.
Soil samples were collected from within cages during the peak growing period of plants (late July) in 2015, 2016, and 2017. Before we took soil samples, sunny days had generally occurred for more than one week. Three soil samples were randomly taken from the 0-to 10-cm and 10-to 20-cm soil layers at the given distance within one day. The collected soil samples from each layer and sampling site were divided into three parts. One part was placed into an aluminum box for soil water content (SWC) determination, one part was dried at 65°C for soil chemical analysis, and the third part was kept at −20°C for phospholipid fatty acid (PLFA) analysis.

| Soil physicochemical properties
Soil physiochemical properties included SWC, pH, electronic conductivity (EC), total organic carbon (TOC), total nitrogen (TN), and total phosphorus (TP). SWC was calculated as follows: SWC = (W 2 -W 3 )*100/(W 2 −W 1 ), where W 1 was the weight of the aluminum box, W 2 was the weight of the initial soil sample with the aluminum box, and W 3 was the weight of the dried soil sample with the aluminum box, which dried at 65°C to constant weight. Soil pH and electrical conductivity (EC) were measured with an acidity meter and conductivity meter (METTLER TOLEDO), respectively (Guo, Li, et al., 2018;Guo, Weise, et al., 2018). TOC was determined by the TOC determinator (MACRO) (Wang, Qin, Cai, Meng, & Zhang, 2014). TN was determined by the Kelvin determination apparatus (FOSS Kjeltec) (Ge, Li, Fan, Hou, & Liang, 2010). TP was measured by the sodium hydroxide fusion-molybdenum antimony colorimetric method (JingHua) (Bao, 2013).

| Soil microorganisms
Phospholipid fatty acid (PLFA) analysis was used to measure the structure of the sedimentary microorganisms according to Bossio and Scow (Bossio & Scow, 1998). Briefly, 8-g dry soil was extracted using a single-phase mixture of chloroform/methanol/citrate buffer (23 ml at a 1:2:0.8 volume ratio). After phase separation, the CHCL 3 layer (extracted lipids) was collected. The extract was transferred into a silica solid-phase extraction column, and neutral lipids and glycolipids were removed by sequential elution with chloroform (5 ml) and acetone (10 ml). Phospholipids were then collected by elution with methanol (5 ml) and dried under N 2 . Afterward, the phospholipid fraction was methylated with a methanol-toluene solution and a potassium hydroxide methanol solution, and H 2 O (2 ml) and acetic acid (0.3 ml) were added. Fatty acid methyl esters were extracted in hexane (2 × 2 ml) and dried under N 2 . The resulting fatty acid methyl esters (FAME) were analyzed on an Agilent 6850 Gas Chromatograph (GC) with MIDI peak identification software. The column was an Agilent capillary column using H 2 as the carrier gas.
The GC temperature progression was controlled by the MIDI software. The fatty acid 19:0 was added as an internal standard before methylation. Identification and quantification of fatty acid methyl esters were conducted automatically by the MIDI peak identification software. The indicators of bacteria and fungi in PLFAs are shown in Table 1. The ratios of gram-positive/gram-negative bacteria and fungi/bacteria were analyzed.

| Statistical analysis
The spatiotemporal responses of the soil physiochemical properties to the newly built lake were analyzed by analysis of variance (ANOVA) using the general linear model (GLM) procedure, with distance, years, and their interaction in the model. Least square means were used for mean separation. Multiple comparisons were used to compare the differences between distances for each year or among years within each distance. The spatiotemporal responses of soil microorganisms were analyzed using principal component analysis (PCA). The relationships between soil physiochemical properties and microorganisms were analyzed using detrended correspondence analysis (DCA) and redundant analysis (RDA). SWC, pH, EC, TOC, TN, and TP were used as environmental data, and the detected specific microorganisms were used as the species data. After the conversion of species and environmental data, all species data were first analyzed using DCA to select the suitable model for data analysis (Reavie, Hall, & Smol, 1995). To eliminate the arch effect, a single-peak response value with a gradient length of <3 was achieved, which indicated that the species data had an obvious linear response relationship with the previous two environmental axes and were suitable for redundant analysis (RDA) (Jongman, Braak, & Tongeren, 1995). There was no abnormal species data in the RDA test, and the variance expansion factors (VIF) of environmental indicators were all lower than the normal value of 20; that is, the typical correlation coefficient of environmental indicators and the axis was stable, which was suitable for further explanation (Teixeira, Assis, Siqueira, & Casagrande, 2008). The relationship of microorganisms and soil physicochemical properties in different soil layers was analyzed using path analysis, and data from different years and distances were combined in path analysis. PCA, DCA, and RDA analysis was completed by the vegan package in R3.3.3, and the path analysis was completed by the sem package in R3.3.3. Effects were considered to be significant if p ≤ .05.

| The responses of soil physiochemical parameters to the newly built lake
A newly built lake had a spatiotemporal effect on the SWC of adjacent grasslands (Table 2). In the 0-to 10-cm soil layer, SWC decreased significantly as the distance from the lake increased from 10 to 150 m and decreased up to 19.02%, 15.48%, and 13.31% in 2015, 2016, and 2017, respectively. At more than 150 m from the lake, SWC was not different among distances. At most of the distances, except for at 10 and 150 m, SWC had no much difference among years since lake construction. In the 10-to 20-cm soil layer, SWC within a 100 m distance decreased significantly with increasing distance from the lake, falling by 10.96%, 10.94%, and 4.38%, in 2015, 2016, and 2017, respectively. SWC did not change significantly among distances >100 m from the lake. There was no significant interannual difference at most of the distances except 10 and 400 m.
The surface soil EC had a fluctuating change with the distance from the lake and the years since construction of lake (Table 2). In the 0-to 10-cm soil layer, EC first increased and then decreased as the distance from the lake increased. The maximum values were 4,816 μs/cm at 150 m in 2015, 1,536 μs/cm at 100 m in 2016, and 1,435 μs/cm at 150 m in 2017. In the 10-to 20-cm soil layer, the soil EC showed a similar trend as that of the 0-to 10-cm soil layer, which first increased and then decreased with increasing distance from the lake. The maximum values were 2,447 μs/cm at 150 m in 2015, 1,677 μs/cm at 60 m in 2016, and 1,640 μs/cm at 100 m in 2017. As the years since the construction of the lake increased from 2015 to 2017, the soil EC in the 0-to 20-cm soil layer decreased.
The soil pH showed different fluctuations in different years with increasing distance from the lake ( Table 2). In the 0-to 10-cm soil layer, the pH in 2015 first decreased until 150 m and then increased with increasing distance from the lake, reaching a minimum value of 7.96 at 150 m. In 2016, the pH was significantly higher in the area near the lake (10-30 m) than in the area farther from the lake (60-600 m). Within the distance of 150 m, pH values in 2017 were greater than in 2015; however, pH values within 250-600 m of the lake were smaller than those in 2015. In the 10-to 20-cm soil layer, the pH value at a given distance increased significantly with increasing years since the construction of the lake; that is, pH values from 2016 and 2017 were significantly greater than those from 2015.
The newly built lake affected the spatial and temporal distribution of soil TN, TP, and TOC in adjacent grasslands (Table 3). In 2015, in the 0-to 10-cm soil layer, the soil TN first increased until 150m and then decreased with increasing distance from the lake, reaching a maximum of 3.13 g/kg at 150 m (Table 3) (Table 3). In 2016 and 2017, soil organic carbon within 150 m of the lake was generally greater than that at more than 150 m distance.
In the 10-to 20-cm soil layer, soil TN showed a similar trend in 2015, 2016, and 2017, increasing until 150 m and then decreasing with the change of distance (Table 3). The soil TP decreased with increasing distance from the lake each year, but the differences among distances decreased with increasing lake establishment time. The soil TOC first increased and then decreased with increasing distance from the lake, reaching a maximum at 150 m. When the years since the construction of the lake increased from 2015 to 2017, the soil TOC decreased within 60 m distance and increased at 600 m distance; however, there was no difference in the distance from 100 to 400 m. 13.67 ± 0.77aA Note: Data are means ± SD; means in rows within each year with different lowercase letters are significantly different at p ≤ .05; means in columns within each distance from lake with different uppercase letters are significantly different at p ≤ .05.

| Responses of soil microorganisms to the newly built lake
The soil microbial biomass (SMB) changed dynamically in different soil layers and at different distances from the lake (Figure 1) With the increase in years since the construction of the lake, the bacterial content increased at most distances (except 100 and 400 m).
In the 10-to 20-cm soil layer, the bacterial content at 10 m was the lowest. In 2015, the bacterial content first increased and then de-

| Response of soil physiochemical properties to the newly built lake
When the new lake was built on grassland in this study, it changed the soil environment of the adjacent grassland to some extent. The spatial variation of soil physicochemical properties of the grassland F I G U R E 5 RDA two-dimensional sequencing diagram of microbial and soil physicochemical properties in different soil layers. The blue lines represent different soil physicochemical indexes, and the red dashed lines represent different microorganism indicators. SWC represents soil water content, EC represents electrical conductivity, TOC represents soil organic carbon, TP represents soil total phosphorus, and TN represents soil total nitrogen occurred following the appearance of the lake. The SWC decreased gradually with increasing distance from the lake and finally became stable. The SWC of the 0-to 10-cm soil layer became stable when the distance was more than 150 meters, while the SWC of 10-to 20-cm soil layer tended to be stable when the distance was more than 100 m.
At the same distance, the magnitude of change in the SWC in the 0-to 10-cm soil layer was greater than that in the 10-to 20-cm soil layer.
These results indicated that the influence of the newly built lake on SWC of the adjacent grassland had a certain range and was greater in the upper layer than in the lower layer of soil. These results verified the findings of Qiu, Fu, Wang, and Chen (2001), who found that environmental influences on soil became weak when soil depth increased.
In the 0-to 10-cm soil layer, the pH value close to the lake (10-60 m distance) was greater than that at more than 100 m distance.
This was probably because the SWC close to the artificial lake was higher. The higher SWC caused the increase in MnO + 2 , FeOH + , and SO 2− 4 contents in soil and the decrease in H + contents due to consumption and thus increased the soil pH value (Willett, 1989). However, this phenomenon in deep soil (10-20 cm) was not apparent, possibly because secretions from roots in this soil layer balanced the change in pH (Sarabia et al., 2018). The EC of different soil layers first increased and then decreased with increasing distance from the lake, reaching a maximum within 100-150 m distance. This change was probably associated with the change of SWC. Carsel and Parrish (1988) reported that water has a great influence on soil salinity dynamics. In this study, the high SWC at 10-60 m distance could dissolve the salinity brought by evaporation, while the low SWC at 250-600 m could not bring salt to the upper soil through evaporation.
Environmental changes also affected the spatial distribution of soil nutrients. In this study, TN, TP, and TOC in the surface soil (0-10 cm) were higher than that in the deep soil (10-20 cm). This was probably because the surface soil had good air permeability, a large amount of litter and high microbial activity, which promoted the accumulation of soil nutrients (Ma et al., 2015). However, under extreme water conditions (drought or flooding), the decreased microbial activity affected decomposition and then decreased soil nutrient contents (Kucera & Kirkham, 1971). In this study, soil TN and TOC first increased and then decreased with increasing distance from the lake, which was probably because the high SWC in the area near the lake affected the microbial activity and then resulted in relatively low nutrient contents in the same area. The TP of different soil layers at 10 and 30 m was significantly higher than that at other distances farther from the lake. This is probably because the high water content in the area near the lake led to the death of less tolerant microorganisms with a large amount of fatty acids and thus increased the soil phosphorus content (Blackwell et al., 2009).

| Response of soil microorganisms to the newly built lake
Soil organisms play an important role in the decomposition of soil organic matter and mineralization of nutrient elements, but they are very sensitive to the soil environment. Changes in the environment can lead to changes in microbial community structure . In this study, microbial biomass and bacterial and fungal contents at most distances decreased with increasing soil depth. This may be because the upper soil (0-10 cm soil depth) has better exchange with external conditions such as water, gas, and nutrients, which provide better conditions for the growth and metabolism of microorganisms (Stone, Deforest, & Plante, 2014).
In addition, the topsoil contained the most resources (higher TOC and TN contents) that were generally strong drivers of microbial biomass. Therefore, microorganisms in the upper soil were more active than those in the lower soil. In the 10-to 20-cm soil layer, the fluctuation of the biomass of microorganisms, bacteria, and fungi was greater within 250 m distance from lake than at farther distances. This was probably because the impact of the lake on the surrounding environment had a certain range, so it showed low magnitude when the distance from the lake exceeded a certain range. In 2017, the third year after the lake was built, the amounts of microorganisms, bacteria, and fungi at most distances in the surface soil were greater than those in 2015 and 2016. The improved soil water conditions could facilitate the decomposition of litter (Smith, Woodin, Pakeman, Johnson, & Van, 2014).
Among the main soil physical and chemical factors, SWC and pH had the greatest impact on soil microorganisms in the 0-to 10-cm soil layer or the 10-to 20-cm soil layer. Soil water content is an important driving factor influencing soil microbial processes (Foulquier F I G U R E 6 The path analysis of the relationship of microorganisms and soil physicochemical properties in different soil layers. B represents bacteria, F represents fungi, the dashed line represents a negative correlation, the solid represents a positive correlation, and the thickness of the line represents the strength of the correlation et al., 2013). Under water pressure, a series of small molecules will be produced in microbial cells to adapt to the new water environment (Schimel, Scott, & Killham, 1989). Due to the different structures of microorganisms, water conditions of microorganisms needed for survival are different, which leads to different microorganism community structures under varied water contents (Harris, 1981;Wood et al., 2001). Our results verified the relationship between SWC and microorganisms. SWC was positively correlated with gram-negative bacteria and negatively correlated with gram-positive bacteria. Many studies have also noted that soil pH value is an important driver of microbial community structural changes (Abbasi & Adams, 2000;Hassink, 1992). This is because most of the reactions in the growth processes of microorganisms are enzymatic reactions. Enzymatic reactions need to be carried out within a certain pH range.

| CON CLUS ION
Grassland in arid and semiarid region generally faced the problem of water shortage. Constructing a lake on the grassland could mitigate this issue to some degree, especially for grassland area close to the lake. After the lake was constructed, SWC, TN, and TOC generally had greater values within 150 m of the lake. The spatial distributions of microorganisms of adjacent grassland soil were also affected after the construction of the lake. The SMB and the bacterial and fungal contents generally increased with increasing years since the construction of the lake. Gram-negative bacteria and methanotrophic bacteria were greater within a 30 m distance of the lake. Actinobacteria, gram-positive bacteria, and anaerobic bacteria showed higher abundances over 60 m away from the lake. The results indicated that the newly built lake could be a driving factor for improving the physiochemical properties and microorganism activity of adjacent grassland soil within a certain range.

ACK N OWLED G M ENTS
We thank Yangfang Bai, Jianxia Qiao, Liping Wang, and many staff members from NFSGE for installing/maintaining the experimental plots and for sample and data collection. This research was supported by the National Key Research and Development Program (2016YFC0501902, 2018YFF0213405).

CO N FLI C T O F I NTE R E S T S
The authors declare no conflict of interest.

E TH I C S S TATEM ENT
None required.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analyzed during this study are included in this published article.