The response of the soil microbial food web to extreme rainfall under different plant systems

An agroforestry experiment was conducted that involved four planting systems: monoculture of the focal species Zanthoxylum bungeanum and mixed cultures of Z. bungeanum and Capsicum annuum, Z. bungeanum and Medicago sativa and Z. bungeanum and Glycine max. Soil microbial food web (microorganisms and nematodes) was investigated under manipulated extreme rainfall in the four planting systems to assess whether presence of neighbor species alleviated the magnitude of extreme rainfall on nutrient uptake of the focal species by increasing the stability of soil food web. Our results indicate that in the focal species and G. max mixed culture, leaf nitrogen contents of the focal species were higher than in the monoculture and in the other mixed cultures under extreme rainfall. This result was mainly due to the significant increase under extreme rainfall of G. max species root biomass, resulting in enhanced microbial resistance and subsequent net nitrogen mineralization rate and leaf nitrogen uptake for the focal species. Differences in functional traits of neighbors had additive effects and led to a marked divergence of soil food-web resistance and nutrient uptake of the focal species. Climate change can indirectly alleviate focal species via its influence on their neighbors.

Scientific RepoRts | 6:37662 | DOI: 10.1038/srep37662 belowground functional traits could potentially increase soil food-web complexity and stability 16,17 . However, stronger competition among different species may also decrease the stability of the soil food web under extreme rainfall, even in the presence of a legume 10 . The present study examines these contradictory hypotheses by exploring the stability of soil microbial and nematode community responses to variations in plant community composition and extreme rainfall.
There is a close link between aboveground and belowground sub-systems because plants provide organic matter to decomposing microbes, which are an important part of the soil food web and provide nutrients to the soil by decomposing a plant's dead organic matter 18 . Recent investigations showed that subordinate plant diversity increases nematode diversity, abundance and stability 17 . However, many researchers have revealed that the species identity has a stronger influence on the nematode community than the species richness 19,20 . Thus, it is not clear whether the positive biodiversity effect is due to species richness (true biodiversity effect) or to the individual species in a mixture (identical effect).
Extreme rainfall regimes will quickly change soil water content with consequences for habitat, soil food-web stability and resource availability, all of which may cascade onto plant growth 5,21 . Previous studies suggested that the soil microbial community stability indexes and the nematode stability indexes were not significantly affected by plant diversity against extreme rainfall disturbance, indicating a limited potential from plant diversity to buffer soil food webs under extreme rainfall 4 . However, Mariotte et al. 22 suggested that subordinate species moderated drought effects on earthworm communities due to the better drought resistance of subordinate species. Taylor and Wolters found that oribatid mite communities in beech litter were less sensitive to summer drought than those in spruce litter due to the presence of drought-tolerant oribatid mite species, which were strongly related to the litter types 23 . Thus, the plant-soil system can buffer against drought events. Using soil microbial and nematode communities that also have strong species-specific relationships, we hypothesized that plant community composition might be important in affecting soil food-web stability and the nutrient uptake of the focal species under extreme rainfall.
An agroforestry system is an intercropped growth of forest trees and crop plants. This combination of plants helps to improve soil-available nitrogen, increase carbon sequestration, enhance soil biodiversity, reduce threats and ultimately increase agricultural production. Thus, an agroforestry system is one of the potential tools to mitigate climate change, especially under extreme climate conditions 24 .
Zanthoxylum is a widely growing genus in southwestern China, India, North America and Australia 25 . The fruits are used to make Sichuan pepper and have medicinal uses such as for toothache or rheumatism. However, Zanthoxylum species such as Z. bungeanum easily succumb under continuous heavy rainfall, resulting in decreased yield and quality of fruits and increasing mortality rate 26 . Most previous studies focused on increasing nutrient use in agroforestry systems; however, the effects of increasing soil food-web stability are neglected. These effects relate to the uptake of nutrients indirectly under extreme rainfall conditions. The objective of this study was to assess whether an agroforestry system could increase the available nitrogen uptake of the focal species by stabilizing the soil microbial food web under extreme rainfall. The present study tested four hypotheses:1) specific species affect soil microbial and nematode communities, 2) extreme rainfall affects the soil food web differently under different plant communities, 3) legume species increase soil microbes and nematode stability, and 4) planting systems facilitate the focal species relay on species-specific neighbors under extreme rainfall.

Results
Soil properties. Two-way ANOVA showed that planting systems significantly (P < 0.05) affected soil NH 4 +_ N, NO 3 − -N, dissolved organic carbon (DOC) and dissolved organic nitrogen (DON) ( Table 1), extreme rainfall significantly (P < 0.01) affected soil water content, NH 4 +_ N, NO 3 − -N and DOC (Table 1). There was a significant (P < 0.05) interaction between planting system and extreme rainfall on soil NO 3 − -N (Table 1). In the normal rainfall treatment, the Z-M mixed culture significantly (P < 0.05) increased soil NH 4 +_ N when compared with the monoculture and other mixed cultures (Table 1). Soil NO 3 − -N, DOC and DON were greater in the legume mixed culture (P < 0.05) plots than in the monoculture plots and the Z-C mixed culture. Under extreme rainfall, soil NH 4 +_ N, DOC and DON were greater in the legume mixed culture (P < 0.05) plots than in the monoculture plots (Table 1).
Microbial community. The Z-M and the Z-G mixed cultures significantly (P < 0.05) increased total microbial, bacterial and fungal biomass compared with the Z-C mixed culture in normal rainfall condition (Fig. 1). However, the planting system did not have significant effects on the ratios of fungi to bacteria (Fig. 1d).
Compared with normal rainfall, extreme rainfall significantly (P < 0.05) decreased total microbial and fungal biomass in the Z, the Z-C and the Z-M cultures and significantly (P < 0.05) decreased bacterial biomass in the monoculture and mixed cultures (Fig. 1). Under extreme rainfall, the Z-G mixed culture significantly (P < 0.05) increased total microbial and bacteria biomass compared with the monoculture and the Z-C mixed culture. The Z-C mixed culture had a significantly (P < 0.05) lower fungal biomass than the monoculture and the legume mixed culture (Fig. 1c).
Nematode community. Under normal rainfall, the Z-C mixed culture had significantly (P < 0.05) lower total nematode as well as herbivores and bacterivores densities than the monoculture and the legume mixed culture (Fig. 2). The Z-M mixed culture had significantly (P < 0.05) lower fungivores density than the monoculture and the Z-C and the Z-G mixed culture but had a significantly (P < 0.05) higher omnivore-predators density than the Z-C mixed culture (Fig. 2). The Z-G mixed culture had significantly (P < 0.05) greater bacterivores density than the monoculture and the other mixed cultures (Fig. 2c).

Resistance index.
Under extreme rainfall stress, the Z-G mixed culture significantly (P < 0.05) increased the total microbe resistance index versus monoculture and the Z-C and the Z-M mixed culture but significantly (P < 0.05) decreased the bacterivores resistance index ( Table 3). The Z-C mixed culture significantly (P < 0.05) decreased the herbivores resistance index compared with the monoculture, the Z-M and the Z-G mixed culture but significantly (P < 0.05) increased the omnivores-predators resistance index. The Z-M mixed culture had the lowest bacteria and fungivores resistance index (Table 3).

Net nitrogen mineralization.
Under normal rainfall, Z-M and Z-G significantly (P < 0.05) increased the net N mineralization rate when compared to Z and Z-C (Fig. 3). Extreme rainfall significantly (P < 0.05) decreased the net nitrogen mineralization rate in both the monoculture and the mixed culture. Under extreme rainfall, the Z-G mixed culture had a significantly (P < 0.05) higher net N mineralization rate than either the Z or the Z-C mixed cultures (Fig. 3).

Plant leaf nitrogen.
Under normal rainfall, the focal species of Z in the Z-G mixed culture had higher leaf nitrogen content than the Z-C mixed culture (P < 0.05) (Fig. 4a). The neighbors of M had significantly higher leaf N contents than C and G (Fig. 4b). Extreme rainfall significantly (P < 0.05) decreased Z leaf N contents in the Z-M mixed culture, extreme rainfall also tended to decrease Z leaf N contents in the monoculture (p = 0.077). However, extreme rainfall significantly (P < 0.05) increased M leaf N contents in the Z-M mixed culture. Under extreme rainfall, the focal species Z in the Z-G mixed culture had significantly (P < 0.05) higher leaf N contents  Table 2. Response of nematode density in different plant patterns under control and extreme rainfall (Individuals per 100 g dry soil) (mean ± SE). Mixed cultures of Z. bungeanum and G. max (Z-G), Z. bungeanum and M. sativa (Z-M), and Z. bungeanum and C. annuum (Z-C), and monoculture Z. bungeanum (Z). Different uppercase letters indicate significant differences among control (normal rainfall) treatments; different lowercase letters indicate significant differences among extreme rainfall treatments. ANOVA was used to assess the effects of extreme rainfall on soil properties. *P < 0.05, **P < 0.01, ***P < 0.001. than the monoculture, the Z-C or the Z-M mixed culture (Fig. 4a). The G species had the lowest leaf N contents when compared with C and M (Fig. 4b).
Root biomass. Under normal rainfall, there was no significant difference between the neighboring species root biomass (Fig. 4c). Extreme rainfall significantly decreased the Z-M root biomass but significantly increased the Z-G root biomass. Under extreme rainfall, the Z-G mixed culture had a higher root biomass than the Z-M (Fig. 4c).

Correlation analysis.
Under extreme rainfall, the focal species of Z leaf N contents had a significant positive correlation with the total microbial resistance index, the bacterial resistance index and the net nitrogen mineralization rate but had a significant negative correlation with the total nematode resistance index, the bacterivores resistance index and the omnivores-predators resistance index (Table 4). However, the neighboring species leaf N contents had a significant negative correlation with the fungivores resistance index only (Table 4).

Discussion
Legume addition significantly increased the soil-available N, DOC and DON, which was apparently attributed to the N-fixing ability of G. max and M. sativa. Therefore, the input of high-nitrogen residuals (litter and/or root exudates) to soils is the most likely reason for the enhancement of soil carbon and nitrogen content seen in this study. Furthermore, a previous study also recognized that the presence of legume species increased the soil carbon and nitrogen contents 27 . Extreme rainfall significantly decreased soil NO 3 − -N and DOC due to greater soil leaching.
Previous studies have found that mixed cultures (especially the presence of a legume) significantly increase microbial biomass because of enhanced litter and rhizodeposition input compared with monocultures 28,29 . However, in the present study, no significant differences were found in terms of total microbial, bacterial, fungal biomass and the ratio of fungi to bacteria in soils between the mixed cultures and the monoculture under normal rainfall (Fig. 1). This finding is possibly due to legume fixation of atmospheric N 2 leading to higher levels of nitrogen in and around the plant (Table 1) and resulting in the rapid growth of bacteria. However, legume-induced changes in soil microorganisms might exert bottom-up control on higher trophic level organisms, such as microbivores and omnivore-predators, via the food web 28 . Our results also suggested that the legume mixed culture had higher bacterivores or omnivore-predator nematodes (Fig. 2). The legume mixed culture had a higher microbial biomass than the C. annuum mixed culture, likely because C. annuum produces poor-quality litter and root exudates compared to legume species 30 .
Previous studies have revealed that species identity, rather than species richness, has a stronger influence on nematode communities 19,20 . Given their ability to fixed N, legumes are the most common species that consistently influence soil nematodes 27 . Although a recent study argues that neighboring plant diversity increases nematode diversity, abundance and resistance 17 , the role of legumes as a key species in plant communities is well established. Likewise, our study provides evidence that legume addition significantly enhanced bacterivores in the G. max mixed culture and tended to increase omnivore-predator nematodes in the M. sativa mixed culture (p = 0.064) when compared to the monoculture (Fig. 2c,e). The G. max mixed culture significantly increased bacterivores mainly due to the significant increase of Rhabditis and Mesorhabditis, which both belong to Ba1 bacterivores (Table. 2). Ba1 bacterivores have short life spans and are known to rapidly enrich available resources 31 . NO 3 − -N, DOC and DON contents were higher in the G. max mixed culture soil, confirming its nutrient-rich status (Table 1). Therefore, nematode abundance maybe a useful measure to indicate available soil resources. Consistent with the present study, legumes supported large populations of certain bacterivores, especially Rhabditis and Panagrolaimus, both of which are Ba1 nematodes 32 , and legumes increased bacterivores, most notably r-selected Rhabditida 19 . However, our results showed that another legume, the M. sativa mixed culture, remarkably increased Alaimus (Ba4) and Mesodorylaimus (Op5) ( Table 2). Enhancements of some k-selected nematodes (especially Op5), which occupy high trophic levels in soil food webs, could increase the complexity of the soil food web 27 . The differences in the effects of the two legumes on soil nematodes might be due to plant chemistry. Indeed, plants supply food resources not only to plant-feeding nematodes but also to microbivores by forming litter above and belowground and by releasing large quantities of organic material from their living roots 33 . Different plant species vary greatly in terms of litter quality and root exudates; therefore, their decomposers are varied.
Previous studies had suggested that C. annuum could only inhibit root-knot herbivores by releasing different secondary metabolites 30 . The present study revealed interesting findings that C. annuum not only inhibits herbivores but also inhibits bacterivores and fungivores in the C. annuum mixed culture (Fig. 2c,d). Bacterial and fungal feeding nematodes play crucial roles in decomposing litter and increasing microbial activities in soil. Inhibitory effects on nematode density in the C. annuum mixed culture resulted in a significant decrease in the net nitrogen mineralization rate in comparison with the legume mixed culture (Fig. 3).
The Z. bungeanum monoculture had higher fungivores than the mixed cultures, mainly due to Z. bungeanum's waxy leaf, which is difficult to decompose. Slow decomposition of poor-quality litter (low C/N) favors fungi and fungivores, while rapid decomposition of high-quality litter (high C/N) favors bacteria and bacterivores.
Taken together, our results suggest that the identities of a single plant species and plant functional groups are the most important factors influencing microbial and nematode communities. The above findings support our first hypothesis that specific species affect soil microbial and nematode communities.
Extreme rainfall regimes will increase soil water content rapidly with consequences for both soil food web and habitat 5 . However, very few studies have evaluated the effect of extreme rainfall on soil microbial communities 4,21 . The above two results suggest that extreme soil moisture significantly decreases microbial biomass and alters microbial communities because of low levels of oxygen and excessive amounts of water in the soil. These findings are consistent with our results that extreme rainfall significantly reduces total microbial, bacterial and fungal biomass (Fig. 1). First, an excessive amount of water in the soil causes the microbial cell membrane to rupture due to hypoxic condition 34 . Second, this marked decline in microbial biomass corresponds with a significant increase in nutrient loss in the soil (Table 1). This pattern suggests that extreme rainfall-induced leaching is likely to be linked to the decrease in microbial biomass. However, the F/B ratio did not show any significant change under extreme rainfall because of a simultaneous reduction in bacterial and fungal biomass (Fig. 1d). Therefore, this finding indicated that the fungi also did not have sufficient capacity to resist extreme soil water content stress 21,35 , although fungi have strong cell walls and intrinsic resistance potential against moisture stress events. This condition mainly exists because most fungi are located in/on the outer part of soil aggregates and are susceptible to extreme rainfall. Landesman and Dighton 36 suggested no significant change in microbial biomass and community composition in response to extreme rainfall. This discrepancy may be due to different saturation levels and textures of tested soils. Landesman and Dighton 36 investigated sandy soil with low water-holding capacity that cannot induce large fluctuations in soil moisture content, while in the present study, soil water content increased by up to 30%, and the soil had a higher capacity to hold water than sandy soil.
Nematodes respond rapidly to water stress because they depend upon water for movement and migration toward their prey in soil. Previous studies showed that increased soil moisture markedly decreased bacterivore abundance and increased herbivore density 37,38 . However, a recent investigation suggested that extreme rainfall stress increased total nematode abundance depending on the diversity of the plant species 4 . Our findings suggest that extreme rainfall had a significant effect on the density of bacterivores and herbivores depending on the composition of the plant species (Fig. 2b,c). Extreme rainfall significantly increased herbivore abundance in the C. annuum mixed culture (Fig. 2), possibly because high soil moisture content decreased the level of secondary metabolites released by C. annum to the soil. A higher abundance of herbivores in the soil could have serious negative effects on both soil and plants by feeding the roots of C. annum (Fig. 4c). Extreme rainfall significantly decreased bacterivores in the G. max mixed culture and M. sativa mixed culture (Fig. 2c). In the G. max mixed culture, a decrease in bacterivore abundance was mainly due to the decreased density of Rhabditis and Mesorhabditis, both of which belong to Ba1 nematodes (Table 2). However, leaching induced by extreme rainfall had a limit effects on nematode dissemination 39 . The reason for this might be that extreme rainfall significantly increased G. max's root biomass (Fig. 4c) and perhaps changed the root exudates, which negatively influenced Ba1 nematodes in response to extreme rainfall 28,29 . Another reason might be that extreme rainfall disrupted the aggregate structure of soil where the habitable nematodes occur. In contrast to our study, a recent study suggested that Ba1 nematodes increase after extreme rainfall because of their high acclimatization potential due to their high reproduction rate and impermeable cuticle 4 . However, that study showed nematode densities increasing strongly with increasing plant diversity after the extreme rainfall. Different plants have different strategies for withstanding flooding disturbances, e.g., by aerating the soil and thereby maintaining activity 40 . Thus, in the referenced study, a plant species' existing tolerance to extreme rainfall in a community with high plant diversity was another reason for the increase in Ba1 nematodes. In the M. sativa mixed culture in the present study, extreme rainfall mainly decreased the genera Rhabditis, Acrobeloides and Alaimus (Table 2). Nematodes responded differently in G. max and M. sativa mixed cultures under almost the same soil water content. Thus, the resistance of populations to soil organisms may be partly dependent on the traits of the respective plant species.
Extreme rainfall significantly decreased fungivores in monoculture as well as the C. annuum and G. max mixed cultures (Fig. 2). This pattern primarily occurred because extreme rainfall significantly decreased the genera Aphelenchus and Ditylenchus, which are cp-2 nematodes (Table 2). However, extreme rainfall significantly increased fungivores in the M. sativa mixed culture (Fig. 2). This result may be due to the significant reduction  Table 4. Correlation analysis of soil biota resistance index and leaf NH 4 + -N and NO 3 − -N contents. *P < 0.05, **P < 0.01.
Scientific RepoRts | 6:37662 | DOI: 10.1038/srep37662 in omnivore-predator abundance. Thus, the food source or feeding relationship is also a dominant aspect for structuring the soil food web.
In the present study, extreme rainfall significantly decreased omnivore-predator abundance in G. max and M. sativa mixed cultures and tended to decrease them in the monoculture (p = 0.074). The density of the genus Mesodorylaimus (cp-5) decreased significantly in these cultures (Table 2). Cp-5 nematode species are the type most susceptible to environmental stress because of their long life cycle, permeable cuticle and high sensitivity. However, recent investigations revealed that Cp-5 nematode species have a strong resistance to different environmental stresses because of their varied and unidentified feeding strategies and their complex interactions with higher trophic levels 41 . All of the above results confirm our second hypothesis that extreme rainfall affected the soil food web differently under different planting systems.
Although a recent study argued that greater plant diversity is responsible for decreasing ecosystem stability under flooding stress 5 , ecosystem stability is directly related to plant functional diversity, e.g., nitrogen-fixing legumes 42 . Thus, no consistent conclusion exists about how plant community composition affects ecological stability, especially soil food-web resistance.
In the present study, total microbial resistance was higher in the G. max mixed culture than in the monoculture, the C. annuum mixed culture and the M. sativa mixed culture under extreme rainfall stress (Table 3). This finding was mainly due to the significantly increased root biomass of G. max under extreme rainfall, which greatly stimulated the growth of adventitious roots (F. Sun, personal observation) and increased tolerance to hypoxia, both of which safeguarded the root activity and functionality 43 . These factors also resulted in higher microbial resistance than was found in the monoculture or in the C. annuum and M. sativa mixed cultures. Another reason for this finding was that the reduction of bacterivores grazing on bacteria resulted in a higher microbial biomass than was found in the either the monoculture or the C. annuum mixed culture. Although the M. sativa mixed culture contained a legume species, the culture did not show an increase in microbial resistance (Table 3) because extreme rainfall significantly decreased the root biomass of M. sativa and reduced the input of root exudates to soil microorganisms. Therefore, microbial resistance depends on the composition and traits of the species. Bacterivores' resistance was lower in the legume mixed culture than it was in the monoculture (Table 3) due to the stress extreme rainfall had on bacterivores and their abundance (Fig. 2). The composition of the species, even in the presence of the legume, could not enhance soil microbial and nematode resistance simultaneously. This apparent trade-off in the soil food web is used to constrain the composition and diversity of the entire soil community. Thus, these results only partly support our third hypothesis that legume species could increase soil microbes and nematode stability.
Neighbor plants have direct positive effects on whole plant communities by altering and modifying physical and/or biotic conditions 44 . Although the facilitation process is important for the structuring and functioning of an ecosystem, the mechanism behind this process is still poorly understood, especially in the belowground soil food web. Some studies suggest that leguminous plant presence enhances the N uptake of the focal species in ambient weather 45 and in extreme rainfall 8 . However, our results were inconsistent with previous findings. Specifically, under extreme rainfall, only the G. max mixed culture had higher leaf nitrogen contents in the focal species when compared to the monoculture, the C. annuum mixed culture and the M. sativa mixed culture (Fig. 4a). The reason for this was that extreme rainfall significantly increased the G. max species root biomass, thereby enhancing microbial resistance and the net nitrogen mineralization rate. Thus, this process increased soil available N and enhanced nutrient uptake of the focal species. Our data also suggested that the focal leaf N contents had a significant positive correlation with total microbial resistance, bacterial resistance and the net nitrogen mineralization rate but a negative correlation with total nematode resistance and the bacterivore resistance index (Table 4). Thus, maintaining lower trophic level stability (microbial resistance) is more important than maintaining higher trophic level stability (nematode resistance) under extreme rainfall stress. In the M. sativa mixed culture, although a legume species was also present, the focal species had lower leaf nitrogen contents than the G. max mixed culture (Fig. 4a). The main reason was the strong interspecific competition for nitrogen between M. sativa and the focal species under extreme rainfall (Fig. 4b). The leaf nitrogen contents of M. sativa were significantly higher (+ 64%) than they were under normal rainfall. The deterioration effects for the focal species under extreme rainfall occurred because of increased competition for nutrients. These findings confirmed the study's fourth hypothesis, that planting systems could facilitate the focal species' relay on species-specific neighbor species under extreme rainfall. Legume species could facilitate the focal species, but the effect is highly species specific, related to climate change disturbance, and mostly due to belowground community processes. Thus, different mixed cultures would result in various soil food web structures under extreme rainfall and subsequently different biogeochemical cycling, soil function and nutrient availability. This finding is important in agricultural management that will continue to change to alleviate extreme climate conditions under future climate change. Belowground communities must be integrated into land management and climate change mitigation strategies.

Materials and Methods
Study site. The experiment was carried out in Maoxian, Sichuan Province, southwestern China (103°53′ E, 31°41′ N), at an altitude of 1686 m above the sea level. The mean annual temperature, precipitation and evaporation were 8.9 °C, 920 mm and 796 mm, respectively, according to meteorological monitoring from the Maoxian Ecological Station of Chinese Academy of Science. Most rainfall occurred in August, and most drought occurred in February. In this area, a significant increase in extreme precipitation was found in a previous study 2  Two-year-old uniform seedlings of Z. bungeanum were collected from a local nursery based on plant stem base diameter and height. Z. bungeanum belongs to the family Rutaceae and is a perennial plant reproductively, while M. sativa, G. max, and C. annuum are annual plants. M. sativa and G. max are leguminous plant species, while C. annuum is an herbaceous plant. The focal species Z. bungeanum was grown once, in the center of each plot. The neighboring species G. max, M. sativa and C. annuum were planted at the same density in all plots. Each plot area was 2.6 × 2.6 m 2 , and the distance between each plot was 1 m. Each pattern was replicated six times randomly, for a total of 24 plots. In Z monoculture, all the understory plants were completely removed. All the planting systems had no fertilization from the establishment of experiment, and the understory grasses were removed by hand each week; herbicide was not used to avoid potentially negative effects on soil organisms 29 . In August 2015, we conducted precipitation treatments randomly. Control (normal rainfall) and extreme rainfall each had three replications. Control and extreme rainfall plots were maintained under rainout shelters during the experiment period. The shelters were constructed using steel frames and were covered with transparent plastic 47 to protect from precipitation from August 1 to 30. To minimize greenhouse effects, the rainout shelters were 1.8 m aboveground. Tap water was used to mimic extreme precipitation events, and a watering pot was used to compensate for rain. Rain regimes were designated according to Ng et al. 48 in which the average rainfall in the area during August was 3.0 mm/day (based on the average rainfall data during 1983-2013 from the Maoxian Ecological Station of Chinese Academy of Science), and was designated as the control rain regime, while extreme rainfall was designated according to the abnormally high rainfall in August (9.5 mm/day). All the plots were watered (7-9 am and 18-20 pm) from August 1 to 30. Around all plots, thick PVC panels were inserted at the top of the 0.5 m soil layer to prevent lateral water movement between plots and to prevent interaction between roots from neighboring plots.
Field sampling. Soil sampling was conducted on 31 August 2015. A total of five soil cores (5 cm diameter) were collected using the five-spot method from a 10 cm soil depth and combined to form one composite sample for each plot. Z. bungeanum and neighbor species leaves were sampled on the same day, kept in a liquid nitrogen container, and taken back to the laboratory for determining the concentrations of leaf NH 4 + -N and NO 3 − -N. In total, three individual neighbor species were collected from each plot.
Soil properties analysis. Estimation of soil moisture content was performed gravimetrically by oven drying (105 °C for 24 h) 20 g of the field soil sample. A button thermometer (QT-ST010, Beijing Channel Science Equipment Co., LTD, China) was used for soil temperature measurement, and data were recorded from each plot at 1-h time intervals. Soil pH was estimated in a 1:2.5 soil/CaCl 2 (0.01 mol/L) suspension. Soil NH 4 + -N and NO 3 − -N were determined with the help of a flow injection auto analyzer (AA3, Bran+ Luebbe, Germany).
Dissolved organic carbon (DOC) and nitrogen (DON) were estimated using a TOC/TN analyzer (Multi N/C ® 2100(S), Analytik Jena AG, Germany). Net N mineralization rate was calculated according to the changes in NH 4 + -N and NO 3 − -N before and after the 7-d incubation.
Nematode community analysis. The modified cotton-wool filter method was followed to determine soil (50 g) nematodes 51 . From each sample, the total number was identified, and then 100 nematode specimens were randomly selected and identified with the help of an inverted compound microscope. Nematode populations were expressed as the number of nematodes per 100 g dry soil. Based on their feeding habits, nematodes were categorized into four groups: (1) bacterivores, (2) fungivores, (3) omnivores-predators and (4) herbivores 52 .
Neighboring species root biomass analysis. All harvested neighboring root biomass was dried at 75 °C to constant weight and weighed.
Plant leaf nitrogen analysis. Plants readily take up and directly use two main forms of soil available nitrogen (NH 4 + -N and NO 3 − -N) for growth. For leaf NH 4 + -N and NO 3 − -N measurement, 0.2-g leaf samples were homogenized in 2 mL of 10% HCL solution and 5 mL of deionized water, respectively. The resultant supernatants were analyzed using the quantitative colorimetric method 53 .
Statistical analyses. Statistical analysis was performed using SPSS v. 17.0 (SPSS Inc., Chicago, IL). Two-way ANOVA was used to test the effects of planting systems and extreme rainfall on the soil properties and genera Scientific RepoRts | 6:37662 | DOI: 10.1038/srep37662 of nematode. One-way ANOVA was used to test the effects of planting systems on the soil properties, microbial biomass, nematode abundance, net nitrogen mineralization, microbial and nematode resistance index, neighbor species root biomass and leaf nitrogen content under control and extreme rainfall. One-way ANOVA was used to assess the effect of extreme rainfall on soil properties, microbial biomass, nematode abundance, net nitrogen mineralization, microbial and nematode resistance index, neighbor species root biomass and leaf nitrogen content under each planting system. Correlation analysis was used to assess soil biota resistance index and leaf NH 4 + -N and NO 3 − -N contents. Least significant difference (LSD) was used to test differences among treatment means. Statistical significance was determined at P < 0.05.