Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2018 Oct 11.
Published in final edited form as: J Appl Ecol. 2017 Feb 24;54(5):1291–1300. doi: 10.1111/1365-2664.12894

Soil drainage facilitates earthworm invasion and subsequent carbon loss from peatland soil

Xinwei Wu 1, Rui Cao 1, Xue Wei 2, Xinqiang Xi 1, Peili Shi 3, Nico Eisenhauer 4,5, Shucun Sun 1,2,*
PMCID: PMC6181200  EMSID: EMS79941  PMID: 30319142

Summary

1. Human activities have been a significant driver of environmental changes with tremendous consequences for carbon dynamics. Peatlands are critical ecosystems because they store ~30% of the global soil organic carbon pool and are particularly vulnerable to anthropogenic changes. The Zoige peatland on the eastern Tibet Plateau, as the largest alpine peatland in the world, accounts for 1‰ of global peat soil organic carbon storage. However, this peatland has experienced dramatic climate change including increased temperature and reduced precipitation in the past decades, which likely is responsible for a decline of the water table and facilitated earthworm invasion, two major factors reducing soil organic carbon (SOC) storage of peatlands.

2. Because earthworms are often more active in low- than in high- moisture peatlands, we hypothesized that the simultaneous occurrence of water table decline and earthworm invasion would synergistically accelerate the release of SOC from peatland soil. We conducted a field experiment with a paired split-plot design, i.e. presence vs. absence of the invasive earthworms (Pheretima aspergillum) nested in drained vs. undrained plots, respectively, for three years within the homogenous Zoige peatland.

3. Water table decline significantly decreased soil water content and bulk density, resulting in a marked reduction of SOC storage. Moreover, consistent with our hypothesis, earthworm presence dramatically reduced SOC in the drained but not in the undrained peatland through the formation of deep burrows and decreasing bulk density of the lower soil layer over three years. The variation in SOC likely was due to changes in aboveground plant biomass, root growth, and earthworm behavior induced by the experimental treatments.

4. Synthesis and applications. We suggest that incentive measures should be taken to prevent further water table decline and earthworm invasion for maintaining the soil C pool in Zoige peatland. Artificial filling of drainage canals should be implemented to increase the water table level, facilitating the recovery of drained peatlands. Moreover, the dispersal of earthworms and their cocoons attached to the roots of crop plants and tree saplings from low-lying areas to the Zoige region should be controlled and restricted.

Keywords: C dynamics, climate change, earthworm, peatland, Qinghai-Tibet Plateau, soil animal, water table

Introduction

The global soil organic carbon (SOC) pool is estimated to be about 1,395 Pg, constituting approximately two-thirds of the carbon in terrestrial ecosystems (Post et al. 1982) and being about two times the atmospheric pool (Lal 2004). The dynamics of the SOC pool thus is important for the global carbon cycle and climate-carbon cycle feedbacks (Melillo et al. 2002; Lal 2004). Peatland SOC is an important component of the global SOC pool (Sahagian & Melack 1998). Peatlands cover only 3% of the global land area but store about 30% of the global SOC pool (Gorham 1991). The Zoige peatland on the Tibetan Plateau is the largest alpine peatland in the world (Sun 1992; Xiang et al. 2009) and covers 4,605 km2, occupying about 23% of the total area of the Zoige plateau and about 1‰ of the total area of peatlands worldwide (Kivinen & Pakarinen 1981; Chen et al. 2014). The Zoige peatland is estimated to contain 0.48 Pg SOC (Chen et al. 2014), occupying approximately 6.2% and 1‰ of SOC storage in China and in the world, respectively (Cui et al. 2015).

Similar to many other peatlands around the world, Zoige peatland is experiencing two major threats that could facilitate and synergistically enhance the loss of SOC. The first one is water table decline (Dong et al. 2010; Yang et al. 2014). Zoige peatland has experienced significant warming with the temperature increasing by 0.4°C per decade since 1970 (Mu et al. 2015). Alongside, precipitation has decreased by on average 22-28 mm per decade (Dong et al. 2010; Yang et al. 2014), which likely has resulted in a decrease in water input into Zoige peatland. Moreover, this peatland has experienced intensive human activities, especially the artificial drainage for pasture expansion since the 1960s (Xiang et al. 2009; Dong et al. 2010). Roughly 1,000 artificial drainage channels, with a total length of 2,864 km, have drained nearly 41% of the total area of Zoige peatland (Dong et al. 2010). Drainage has transformed large areas of peat soil into grazing meadows or even aeolian sandy soils (Xiang et al. 2009), changing plant community composition and vegetation cover, as well as soil properties.

Climate change together with these direct anthropogenic perturbations have collectively led to a 5-15 cm reduction in the water table from the 1970s to the 2000s in Zoige peatland (Xiang et al. 2009). Consistently, records at the two largest rivers, the White River and the Black River, show that the mean annual runoff from the peatland has decreased by about 28% and 35%, respectively, from the 1957s to the 2010s (Li et al. 2014). Numerous studies have shown that the decline of water table, as a global threat to peatlands, may dramatically increase carbon emissions from peatlands and potentially weaken the ecosystem service of carbon storage in the SOC pool (Aerts & Ludwig 1997; Silvola et al. 1999; Strack, Waddington & Tuittila 2004; Smith et al. 2004; Salm et al. 2012).

The second major threat to the Zoige peatland is the invasion of earthworms from urban residential areas. Soil moisture is considered to be critical for the survival, growth, and reproduction of earthworms (Zorn et al. 2008), one of the most significant soil macrofauna invaders in terrestrial ecosystems worldwide (Hendrix et al. 2008). Many earthworm species are unable to establish stable populations in water-saturated soils due to anaerobic conditions (Mather & Christensen, 1992; Ausden, Sutherland & James 2001) and thus are rare in peatlands. However, it is possible that soil fauna including earthworms may invade and successfully establish populations in low-moisture peatland soils (Ausden, Sutherland & James 2001; Bradford et al. 2002; Silvan, Laiho & Vasander 2010) resulting from drainage and reclamation for agriculture or climate change (Curry & Schmidt 2006).

Indeed, in the past decades, earthworm invasion has been observed in the Zoige peatland, although there are no consistent historical records for earthworm communities in the region. In the 1980s, a field survey by Zoige Natural Reserve (Sichuan Academy of Forestry 2006) showed that the peatland and surrounding pastures only contained one epi-anecic earthworm species, Aporrectodea nocturna (Zhao et al. 2013). However, a field investigation in 2008 also found the large-bodied anecic earthworm species Pheretima aspergillum in pastures and drained peatlands (Xiao et al. 2012). The investigators suggested that the invasion of this earthworm species was due to the introduction of crop plants and tree plantations from lower altitude areas (e.g. Chengdu Plain of southwest China; Wu 2008). This was supported by the fact that the earthworm density was relatively high (20-100 ind.m-2) in local urban vegetable gardens and tree plantations, and that this earthworm species was more abundant in the grassland or peatland close to urban plantations (Hu et al. 2016).

Subsequent field investigations further indicate that the earthworm density has been consistently increasing from 1-2 to 5-50 individuals per m2 in the last few years, particularly in the alpine meadow that is typically characterized by soil moisture ranging from 25% to 30% (Zhao et al. 2013; Wu et al. 2013). A field survey in 2015 in preparation of the present study (1 m2 plots, N = 67) further showed that there were 5-60 individuals per m2 in drained peatlands but none in undrained peatlands (with soil moisture levels of mostly >35%; Fig. S1). Importantly, provided that all the drained peatlands/ pastures with soil moisture being ca. 30% would be invaded by this earthworm species, the estimated invaded area could reach 1,800 km2 in the Zoige area, according to the spatial distribution of soil moisture and vegetation type (Dong et al. 2010; Zhang et al. 2014). Studies have demonstrated that earthworm invasions often cause a significant reduction of the soil carbon pool, particularly in ecosystems that are rich in soil organic matter (Kammann et al. 2009; van Geffen, Berg & Aerts 2011; Lubbers et al. 2013). However, there are no studies explicitly examining the effect of earthworm invasion on the SOC pool of peatlands and how this is affected by water table decline.

Here we test the effects of both water table decline and earthworm invasion on the SOC pool of Zoige peatland in order to assess SOC dynamics under ongoing environmental change in a three-year field experiment in a homogenous peatland. As noted above, we predicted that water table decline and earthworm invasion would decrease SOC. Moreover, because earthworms tend to be more active in drained, low-moisture soil than in undrained soil (Silvan, Laiho & Vasander 2010), we predicted that the simultaneous occurrence of water table decline and earthworm invasion would synergistically decrease SOC in Zoige peatland.

Materials and methods

Study Site

This study was conducted in Hongyuan County (32°48′N, 102°33′E) of Sichuan Province in the Zoige peatland of the eastern Qinghai-Tibetan Plateau. The climate is characterized by short and cool spring, summer, and autumn and a long winter. As recorded by Hongyuan County Climate Station (located 5 km from the study site) for the period of 1961-2015, the annual mean temperature was 0.9°C, with maximum and minimum monthly means of 10.9°C and -10.3°C in July and January, respectively.

The area of the Zoige peatland in Hongyuan County is about 492 km2. The depth of peat layers in this region ranges from 0.3 to 10 m having a mean dry mass accumulation rate of 0.03 g m-2 yr-1, an average pH of 6.6-7.0, and an average total carbon (TC) concentration of 58.6 mg L-1 and total nitrogen (TN) concentration of 1.4 mg L-1 (Yang et al. 2014). The most dominant plant species is Carex muliensis, and the other abundant plant species include Caltha palustris, Gentiana formosa, and Trollius farreri (Yang et al. 2014).

The earthworm species Pheretima aspergillum typically occurs in the middle and the lower reaches of the Yangtze River, and Chengdu Plain of southwest China (Wu 2008). This earthworm species had not been found in the Zoige area until 2008 (Sichuan Academy of Forestry 2006; Xiao et al. 2012) and is considered an invasive species to the Zoige peatland (Hu et al. 2016). This species is often active in the top soil layer (ca. 0-10 cm) in mesophytic habitats. It feeds on soil organic matter and rarely appears at the soil surface unless the soil is saturated with water. This earthworm species prefers soils with a soil water content of 25-30%, a soil temperature of 15-25°C, and a pH of 5.9-6.8 (Wu 2008).

Experimental Set-Ups

Main Experiment

To assess the effects of water table and earthworms on peatland SOC, we conducted a field experiment within the homogeneous Zoige peatland, avoiding possible confounding factors, such as variations in soil and vegetation conditions. The experiment was set up in a paired split-plot design, where water table, with two levels (high and low) as a pair per block, was the main plot factor, and earthworm presence, with two levels (presence and absence), was the subplot factor. This resulted in four treatments: (i) high water table, earthworm absent, (ii) high water table, earthworm present, (iii) low water table, earthworm absent, and (iv) low water table, earthworm present. The main plot treatments were replicated six times, resulting in 12 main plots (six blocks) in total (Fig. S2). In each main plot, the subplot treatments (i.e. earthworm presence and absence) were replicated three times (to allow for repeated, destructive samplings in time), resulting in six subplots in each main plot, ending up with 72 subplots (microcosms) in total.

The water table was manipulated by ditching in the homogeneous peatland. In April of 2013, a 240 m long, 0.5 m wide, and 1 m deep drainage ditch (called the major ditch thereafter) was dug in a fenced, flat area of about 2 ha (Fig. S2). The major ditch was connected to a small river. Twelve 6 × 6 m plots (main plots) were deployed (with regular intervals of 60 m between two adjacent plots) approximately 30 m away from the major ditch, with 6 plots distributed along each side of the major ditch. Six plots were drained by a 50 cm deep ditch, and 6 plots were kept intact, which served as low and high water table treatments, respectively (Fig. S2).

We installed polyvinyl chloride (PVC) pipes (1.5 m in length and 5 cm in diameter respectively) in each main plot to record the water table with a ruler every three to ten days during the experiment except for the freezing period (from November to March). The mean annual water table was about 23 cm lower in drained plots than that in undrained plots in 2014 and 2015 (Fig. S3). We also measured soil moisture and temperature 5 cm below the ground surface in the center of the subplots every 30 minutes over three years using data loggers (Watchdog 2000, Spectrum Technologies, Inc., USA). Average daily soil moisture was 10%, 21%, and 18% lower in the drained than in undrained plots in 2013, 2014, and 2015, respectively (Fig. S3), whereas there was no systematic and significant difference in soil temperature (Fig. S3).

On May 10 and 11 of 2013, the 72 microcosms (subplots) were installed in the 12 main plots. In each main plot, every microcosm was deployed about 1 m from the edge of the plot, and microcosms were spaced at least 0.5 m from each other. Each microcosm consisted of a cylindrical chamber (diameter: 30 cm, depth: 35 cm; with covering made of 2 mm PVC sheet) and an intact peatland monolith (30 cm in diameter and 30 cm in depth) inside the chamber. These chambers were deep enough to prevent the emigration of earthworms. This was tested in a pre-experiment (n=3) where none of five earthworms per replicate was found in soil deeper than 20 cm after 15 days. For each microcosm, we first used spades to split the peat to form a cylindrical peatland monolith (but we did not dig the monolith out) to minimize the impact on soil during the installation process. Then, we inserted the cylindrical PVC chamber into the soil covering the peatland monolith. Subsequently, the chamber was further fitted with an aboveground cylinder net, which was made of thin (0.1 mm) steel screen with a mesh size of 0.5 × 0.5 mm (to prevent the emigration of earthworms). The net was 30 cm high relative to ground surface. Measurements made on sunny days showed that the light intensity under such screens was about 80% of full sunlight.

Earthworms (P. aspergillum) were hand-sorted from nearby pastures within two days, until enough individuals were captured to conduct the experiment. We selected medium-sized and vital adult earthworms (about 10 cm in body length) for the experiment (Fründ et al. 2010). Four worms were added to each replicate of the earthworm-present treatment on June 1 of 2013, 2014, and 2015 respectively. The resulting earthworm density was 56 m-2, which was at the upper end of the natural density range (5- 60 m-2) observed in invaded peatland and pastures in the area (Fig. S1).

The experiment was started on June 1, 2013 and terminated on August 20, 2015, covering three growing seasons. We sampled the microcosms (subplots) three times from each main plot, i.e., at the end of the growing season of each year (August 18, 17, and 20 of 2013, 2014, and 2015, respectively). At each time, we randomly sampled two microcosms (one with earthworms and one without) in each main plot. For every sampled microcosm, we first harvested the aboveground plant parts and then took two soil cores (5 cm in diameter and 20 cm in depth) beneath the harvested plants. One soil core was used to measure soil bulk density and water content (in both the upper and lower soil layers; see below). Another soil core was sieved, and all plant roots were picked out for the measurement of root mass. Subsequently, 500 g soil samples were collected from the soil cores for the measurements of soil organic carbon and nutrient concentrations. In 2013 and 2014, we did not dig out the whole peatland monolith because destructive samplings may have influenced the soil condition and plant growth of the whole plots.

In 2015, in addition to soil and plant samples, we sampled earthworms in order to examine possible earthworm density and behavioral responses to variations in the water table. We separately quantified earthworm densities in the upper (0-10 cm depth) and the lower soil layer (10-20 cm depth). We took out the monolith in each microcosm and placed it on a plastic sheet, where each soil layer was hand-sorted for adult and juvenile earthworms and cocoons.

Plant roots and the aboveground parts of plants were dried at 75°C for 48 h and weighed. The soil samples were weighed for each soil layer immediately after being taken to the lab, dried at 105°C for 48 h, and weighed for the calculation of soil water content. Soil bulk density was calculated as the dry mass divided by the volume of the soil core. Moreover, for each soil sample, total soil organic carbon (SOC) and labile organic carbon (LOC) were determined by the potassium dichromate oxidation-outer heating method and KMnO4-chemical Oxidation method, respectively; total N and P were determined by the Kjeldahl method and spectrophotometric Colorimetry (Unicam-200, Unicam, Cambridge, UK), respectively. Soluble N and P concentrations were determined using the alkaline KMnO4 method and 0.5 M NaHCO3 (pH 8.5) solutions, respectively.

Earthworm Response Experiment

Because the density and vertical distribution of the earthworms could be determined only when the cores were destructively sampled at the end of the main experiment, we performed a complementary earthworm response experiment to examine how the vertical distribution of earthworms varied across the water table treatments. We established 24 microcosms in the same plots (i.e., two microcosms per plot) using the same method as for the main experiment. Four earthworms were added to each of the 24 microcosms. The experiment ran for 60 days from June 1 to July 31, 2015. The microcosms were destructively sampled and earthworms were examined 30 and 60 days after the start of the experiment (N = 12 each time).

Data Analyses

For the earthworm response experiment, a generalized linear mixed model (GLMM with Poisson error structure and log link function), with high/low water table as the fixed factor, and sampling date and soil layer as random factors, was used to determine the effect of water table on earthworm density. For the main experiment, the split-plot analysis (GLMM with water table and earthworm presence as fixed factors, and block as a random factor), was used to determine the effects of water table, earthworm presence, and their interactions on SOC, LOC, total N and P, soluble N and P, soil bulk density, and water content (upper layer and lower layer, respectively), aboveground and belowground plant biomass for each sampling year. Once a significant effect was detected, post hoc Tukey HSD tests were used to further elucidate treatment differences at each soil layer. In addition, we conducted correlation analyses to determine the relationships among bulk density, SOC, soil soluble nitrogen concentrations, and aboveground plant biomass in the main experiment. All the statistical analyses were conducted in R 3.3.1 (R Core Team 2016).

Results

Earthworm Performance

Earthworms in the Main Experiment

In the high water table treatment, no adult earthworms were found three years after the start of the experiment (Fig. 1c), but an average number of 0.5 ± 0.3 juveniles (Fig. S4) was found in the upper soil layer. In the low water table treatment, an average number of 3.8 ± 0.5 adult earthworms were found at the end of the experiment, and 83% (3.2 ± 0.4; Fig. 1c) of the initially added number of adult earthworms was found in the lower soil layer (P < 0.001). In addition, an average number of 3.3 ± 0.3 juveniles (Fig. S4) was also found in the upper but not in the lower soil layer.

Fig. 1.

Fig. 1

Open in a new tab

Variations in the number of earthworms between high and low water table treatments. Earthworm survival in the upper (0-10 cm) and lower (10-20 cm) soil layer after (a) 30 days and (b) 60 days of the earthworm response experiment, and (c) at the end of the main experiment. Means ± standard error. ***, P < 0.001.

Earthworm Response Experiment

In the high water table treatment, an average earthworm number of 1.0 ± 0.3 (mean ± 1SE) (Fig. 1a) and 0.3 ± 0.2 (Fig. 1b) were found after 30 and 60 days, respectively, and no earthworms were found in the lower soil layer after 60 days. In the low water table treatment, all the earthworms remained alive after 60 days, and 71% (Fig. 1a) and 79% (Fig. 1b) of the earthworms were found in the lower soil layer after 30 and 60 days, respectively, with the percentages being significantly higher than in the upper layer (both P < 0.001).

Water Table and Earthworm Effects

Bulk Density and Water Content

Water table and earthworm presence interactively affected soil bulk density: in the high water table treatment, the presence of earthworms reduced soil bulk density in both the upper and lower soil layer in 2015, but not in the other years (Table S1; Fig. 2a, b). In the low water table treatment, earthworms consistently decreased soil bulk density in all three years (Table S1; Fig. 2a, b). In addition, drainage significantly decreased soil water content, but the presence of earthworms did not change water content in both the upper and lower soil layer in all three years (Table S1; Fig. S5).

Fig. 2.

Fig. 2

Open in a new tab

Variations in soil bulk density for (a) the upper (0-10 cm) and (b) the lower (10-20 cm) soil layer among the four experimental treatments ((1) high water table, earthworm absent (HWT, -E), (2) high water table, earthworm present (HWT, +E), (3) low water table, earthworm absent (LWT, -E), and (4) low water table, earthworm present (LWT, +E)) in the main experiment. Different letters above the error bars denote statistically significant differences among means (P=0.05), as revealed by GLMM followed by Tukey's HSD test for multiple comparisons. Sample size was 6 for all the treatments. Means ± standard error. The treatments, error bars and sample size were the same in following figures.

Soil Organic Carbon and Nutrients

Experimental water table decline directly and significantly decreased SOC by 13% and 6% (for 0-10 cm and 10-20 cm layers, respectively) in 2013, 17% and 18% in 2014, and 20% and 17% in 2015 (Table S1; Fig. 3a, b). Moreover, it also indirectly decreased SOC by 1% and 8% in 2013, 10% and 22% in 2014, and 8% and 21% in 2015 (Fig. 3a, b) by increasing earthworm activity, as indicated by the significant interaction effect between water table and earthworm presence (Table S1). Specifically, in the high water table treatment, the presence of earthworms did not change SOC in all three years. However, in the low water table treatment, earthworm addition significantly decreased SOC by 13% and 25% in 2014 and by 16% and 31% in 2015 for 0-10 cm and 10-20 cm layers, respectively, but the earthworm effect was not significant in 2013. In addition, experimental water table decline and earthworm addition significantly decreased LOC in 2015 (Table S1; Fig. S6).

Fig. 3.

Fig. 3

Open in a new tab

Variations in (a, b) total soil organic carbon (SOC), (c, d) total N, and (e, f) total P concentrations for the upper (0-10 cm) and the lower (10-20 cm) soil layer among the four experimental treatments in the main experiment.

The effect of experimental water table decline on total N was non-significant in the upper soil layer in 2013, but was significant in the lower soil layer in 2013, and in both soil layers in 2014 and 2015 (Table S1; Fig. 3c, d). Earthworm addition significantly decreased total N of the lower soil layer in 2014 and 2015 in the high water table treatment but not in the low water table treatment in all three years, as indicated by the significant interaction effect between water table and earthworm presence. Experimental water table decline and earthworm addition had non-significant effects on soil total P of both the upper and the lower soil layer in all three years (Table S1; Fig. 3e, f).

Experimental water table decline and earthworm addition interactively affected soil soluble N in both the upper and the lower soil layer over three years (Table 1). Specifically, in the high water table treatment, earthworm addition did not markedly change soil soluble N, whereas in the low water table treatment, earthworm addition significantly increased soil soluble N in both the upper and the lower soil layer in 2014 and 2015 (Fig. 4a, b). Water table decline and earthworm addition increased soil soluble P in the lower but not upper soil layer in 2015 (Table S1; Fig. 4c, d).

Fig. 4.

Fig. 4

Open in a new tab

Variations in (a, b) soluble N and (c, d) soluble P concentrations for the upper (0-10 cm) and the lower (10-20 cm) soil layer among the four treatments in 2014 and 2015 in the main experiment.

Plant Biomass

Aboveground plant biomass was affected by a significant interaction of water table and earthworm presence (Table S1). Specifically, at high soil water level, earthworms did not have a significant effect on aboveground plant biomass in all three years, whereas earthworms significantly increased aboveground plant biomass in the low soil water level treatment in 2014 and 2015 (Table S1; Fig. 5a). The effects of water table decline and earthworm addition on root mass were non-significant in all three years (Table S1; Fig. 5b).

Fig. 5.

Fig. 5

Open in a new tab

Variations in (a) aboveground plant biomass and (b) root biomass among the four treatments during the main experiment.

Relationships Among Dependent Variables

Significant and positive correlations were found among soil water content, soil bulk density, total N, and SOC in each of the three experimental years, and these variables were significantly negatively correlated with soil soluble N and aboveground plant biomass in 2014 and 2015 (Table S2). Furthermore, soil soluble N was also significantly positively correlated with aboveground plant biomass (Table S2).

In the low water table treatment, there was a positive linear regression relationship between soil bulk density and SOC concentration in both the upper and lower soil layer in 2014 (Fig. S7c, d) and 2015 (Fig. S7e, f), but not in 2013 (Fig. S7a, b). A positive linear relationship between bulk density and N availability was found in the upper but not in the lower soil layer in 2014, and in both the upper and lower soil layer in 2015 (Fig. S8). In addition, there was a positive linear relationship between total aboveground plant biomass and N availability in the lower but not in the upper soil layer (Fig. S9). However, all of these relationships were not found in the high water table treatment.

Discussion

Our experiment reveals that water table decline directly decreased SOC (by 17%- 20%) over the three years in Zoige peatland. This is consistent with numerous studies addressing water table effects on SOC in peatlands (Silvola et al. 1999; Strack, Waddington & Tuittila 2004). The negative effect of earthworm addition on SOC is also consistent with some previous studies addressing earthworm effects on SOC in high organic matter content soils (Whalen, Paustian & Parmelee 1999; Bohlen et al. 2004). Importantly, our results for the first time demonstrate that water table decline facilitates the survival of invasive earthworms and shifts their habitats, which indirectly reduced SOC (by 8% and 21% for the upper and lower soil layer, respectively). Notably, at high water table, invasive earthworm populations were not able to establish, suggesting that anthropogenic soil drainage can not only pave the way for successful earthworm invasion, but the simultaneous occurrence of water table decline and earthworm invasion may potentially be the most severe combination of factors decreasing the SOC pool in Zoige peatland.

Water table level is frequently suggested to be the most important driver influencing peatland soil carbon dynamics (Aerts & Ludwig 1997; Strack, Waddington & Tuittila 2004; Salm et al. 2012). Water table decline often facilitates breakdown of dissolved organic carbon and hence increases carbon dioxide release (Salm et al. 2012; Leiber-Sauheitl et al. 2014), leading to SOC reduction in peatlands (Silvola et al. 1999). Our three-year drainage experiment caused a significant SOC reduction of 20% (0-10 cm soil layer) and 17% (10-20 cm layer). The likely mechanism underlying the SOC reduction is that water table decline may increase soil aeration (Berglund & Berglund 2011), as reflected by the reduced soil moisture and bulk density in the present study (Fig. 2; Figs S3 & S5). Increased soil oxygen may stimulate microbial growth and phenol oxidase activity to degrade phenolic compounds (McLatchey & Reddy 1998). This further may have caused the breakdown of organic matter (especially that of labile organic carbon, Fig. S6; Boyer & Groffman 1996) and the release of carbon dioxide in a biogeochemical cascade (Aerts & Ludwig 1997; Fenner & Freeman 2011).

The invasion of earthworms intensified the soil drainage effect by decreasing SOC in the drained but not in the undrained soils. This can largely be attributed to differences in the density and behavioral responses of earthworms to the water table level. Firstly, earthworm density decreased quickly at a high water table level in both the main and the earthworm response experiment. The high daily average soil moisture (>40%) in the undrained treatment is generally not favorable for earthworms (Lubbers et al. 2013) because high soil water content may be lethal and decreases egg hatching and larval development of earthworms (Fig. S4; Ausden, Sutherland & James 2001) due to low oxygen concentrations affecting the metabolism of earthworms (Mather & Christensen 1992). In contrast, in the drained treatments where soil water content was mostly <30% (Figs S3 & S5), earthworm density increased over the course of the experiment.

Secondly, earthworms moved to the lower soil layer (in both the main and earthworm response experiment) in the drained treatment, likely because water table decline induced more favorable conditions deeper in the soil and allowed for higher earthworm activity. Moreover, sufficiently high soil moisture in the lower soil layer (ca. 30% in Fig. S5) compared to lower moisture of the upper layer soil (e.g. <15% in July and August in Fig. S3), as well as lower root density may be more appropriate for earthworm activity.

These density and behavioral responses of earthworms to water table decline further affected the SOC content in the peatland. Specifically, more earthworms within larger activity space reduced soil bulk density in both the upper and the lower soil layer in the drained treatment, which was likely associated with increased aeration (Zhao et al. 2013). The decomposition activities of the earthworms per se could contribute to the SOC decline, alongside with enhanced soil microbial activity (Lubbers et al. 2013). These earthworm-induced changes in soil physical and biological properties facilitated carbon loss and nitrogen mineralization. This assumption was supported by the results of the correlation and regression analyses (Table S2; Figs S6 & S7) among soil bulk density, soil moisture, soil N, and SOC. In addition, possibly because of the habitat shift of earthworms, SOC decreased more in the lower (by 31%) than in the upper soil layer (by 16%), consistent with the prediction that peatland SOC in deep soils is more sensitive to disturbances that aerate the soil (Fontaine et al. 2007).

In addition, water table decline and earthworm invasion might have affected SOC via other pathways. For instance, water table decline and earthworm invasion changed plant biomass and community composition. Increased aeration and earthworm activity increased plant growth (Fig. S9; van Groenigen et al. 2014) possibly by facilitating microbial growth and N mineralization: The presence of earthworms shifted the plant community from being dominated by monocot to dicot species over the experimental period (X. Wu and R. Cao; personal observation). Nevertheless, root biomass did not show a significant response to improved soil conditions, probably because these peatland plants already have high biomass allocation to roots due to low nutrient availability (Fan et al. 2008). Nevertheless, it is possible that enhanced root activities due to improved soil conditions may have fueled higher levels of carbon turnover by facilitating microbial growth and activity (Mosier 1998). Another possible pathway is that water table decline and earthworm addition could have changed soil temperature that is also critical to peatland SOC dynamics (Melillo et al. 2002), although no systematic and significant difference in soil temperature was found among the experimental treatments of the present study (Fig. S3). Moreover, earthworm-induced improvement in soil aeration might have facilitated the colonization of other invertebrates such as collembolans, nematodes, and mites, which may also have contributed to the reduction of SOC.

In summary, we demonstrate here that water table decline may not only directly reduce peatland SOC, but also indirectly by facilitating the establishment, population growth, and vertical distribution of invasive earthworms. This indirect effect supports the finding that water table decline facilitates the invasion of soil fauna into peatland (Ausden, Sutherland & James 2001; Bradford et al. 2002; Silvan, Laiho & Vasander 2010), which mostly decrease SOC in carbon-rich soils. Land managers and policymakers should therefore consider the potential subsequent carbon loss from soil and initiate activities to conserve and restore drained peatlands. Incentive measures, such as artificial filling and blocking drainage canals, should be implemented to increase the water table level, to rewet and revegetate the drained peatlands. Moreover, measures (e.g. such as root cleaning) should be taken to eliminate earthworms associated with vegetables that are commonly transported from low-lying areas to this alpine region. Alternatively, self-sufficient vegetable production (e.g. in greenhouses) should be encouraged. Such strategies would provide opportunities for the sustainable management of peatlands and would make a cost effective contribution to averting further carbon emission from peatlands.

Supplementary Material

Supplementary information

Acknowledgements

We thank Kai He, Yangheshan Yang, and Chuan Zhao for field assistance. We thank Matthew Carroll and one anonymous reviewer for very helpful comments on the manuscript. This study was supported by 973 Program (2013CB956302), National Science Foundation of China (31470482 and 31325004), Strategic Priority Research Program of the Chinese Academy of Sciences (XDA05050702), and the German Research Foundation (FZT 118) to N. Eisenhauer.

Footnotes

The authors declare no competing financial interests.

Author contributions

SS conceived the project and supervised the research; XW designed and set up the experiments; RC, XW and PS were involved in field sampling and laboratory analyses; XW and XX analyzed data; XW, NE, and SS wrote the manuscript. All authors discussed the results and commented on the manuscript.

References

  1. Aerts R, Ludwig F. Water-table changes and nutritional status affect trace gas emissions from laboratory columns of peatland soils. Soil Biology & Biochemistry. 1997;29:1691–1698. [Google Scholar]
  2. Ausden M, Sutherland WJ, James R. The effects of flooding lowland wet grassland on soil macroinvertebrate prey of breeding wading birds. Journal of Applied Ecology. 2001;38:320–338. [Google Scholar]
  3. Berglund Ö, Berglund K. Influence of water table level and soil properties on emissions of greenhouse gases from cultivated peat soil. Soil Biology & Biochemistry. 2011;43:923–931. [Google Scholar]
  4. Bohlen PJ, Pelletier DM, Groffman PM, Fahey TJ, Fisk MC. Influence of earthworm invasion on redistribution and retention of soil carbon and nitrogen in northern temperate forests. Ecosystems. 2004;7:13–27. [Google Scholar]
  5. Boyer JN, Groffman PM. Bioavailability of water extractable organic carbon fractions in forest and agricultural soil profiles. Soil Biology & Biochemistry. 1996;28:783–790. [Google Scholar]
  6. Bradford MA, Jones TH, Bardgett RD, Black HIJ, Boag B, Bonkowski M, et al. Impacts of soil faunal community composition on model grassland ecosystems. Science. 2002;298:615–618. doi: 10.1126/science.1075805. [DOI] [PubMed] [Google Scholar]
  7. Chen H, Yang G, Peng C, Zhang Y, Zhu D, Zhu Q, et al. The carbon stock of alpine peatlands on the Qinghai-Tibetan Plateau during the Holocene and their future fate. Quaternary Science Reviews. 2014;95:151–158. [Google Scholar]
  8. Cui M, Ma A, Qi H, Zhuang X, Zhuang G, Zhao G. Warmer temperature accelerates methane emissions from the Zoige wetland on the Tibetan Plateau without changing methanogenic community composition. Scientific Reports. 2015;5 doi: 10.1038/srep11616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Curry JP, Schmidt O. Long-term establishment of earthworm populations in grassland on reclaimed industrial cutaway peatland in Ireland. Suo. 2006;57:65–70. [Google Scholar]
  10. Dong Z, Hu G, Yan C, Wang W, Lu J. Aeolian desertification and its causes in the Zoige Plateau of China’s Qinghai–Tibetan Plateau. Environmental Earth Sciences. 2010;59:1731–1740. [Google Scholar]
  11. Eisenhauer N, Fisichelli NA, Frelich LE, Reich PB. Interactive effects of global warming and ‘global worming’ on the initial establishment of native and exotic herbaceous plant species. Oikos. 2012;121:1121–1133. [Google Scholar]
  12. Fan J, Zhong H, Harris W, Yu G, Wang S, Hu Z, et al. Carbon storage in the grasslands of China based on field measurements of above- and below-ground biomass. Climatic Change. 2008;86:375–396. [Google Scholar]
  13. Fenner N, Freeman C. Drought-induced carbon loss in peatlands. Nature Geoscience. 2011;4:895–900. [Google Scholar]
  14. Fontaine S, Barot S, Barré P, Bdioui N, Mary B, Rumpel C. Stability of organic carbon in deep soil layers controlled by fresh carbon supply. Nature. 2007;450:277–281. doi: 10.1038/nature06275. [DOI] [PubMed] [Google Scholar]
  15. Fründ HC, Butt K, Capowiez Y, Eisenhauer N, Emmerling C, Ernst G, et al. Using earthworms as model organisms in the laborytory: Recommendations for experimental implementations. Pedobiologia. 2010;53:119–125. [Google Scholar]
  16. Gorham E. Northern peatlands: role in the carbon cycle and probable responsesto climatic warming. Ecological Application. 1991;1:182–195. doi: 10.2307/1941811. [DOI] [PubMed] [Google Scholar]
  17. Hendrix PF, Callaham MA, Jr, Drake JM, Huang C, James SW, Snyder BA, et al. Pandora’s box contained bait: the global problem of introduced earthworms. Annual Review of Ecology, Evolution, and Systematics. 2008;39:593–613. [Google Scholar]
  18. Hu L, Wu X, Zhou Q, Sun S. The ecosystem service of the Zoige Marsh: current status and research prospect. Journal of Southwest University of Nationalities (Natural Science Edition) 2016;42:246–254. (in Chinese with an English abstract) [Google Scholar]
  19. Kammann C, Hepp S, Lenhart K, Müller C. Stimulation of methane consumption by endogenous CH4 production in aerobic grassland soil. Soil Biology & Biochemistry. 2009;41:622–629. [Google Scholar]
  20. Kivinen E, Pakarinen P. Geographical distribution of peat resources and major peatland complex types in the world. Annales Academiac Scientarum Fennicae Series A. 1981;32:1–28. [Google Scholar]
  21. Lal R. Soil carbon sequestration impacts on global climate change and food security. Science. 2004;304:1623–1627. doi: 10.1126/science.1097396. [DOI] [PubMed] [Google Scholar]
  22. Leiber-Sauheitl K, Fuß R, Voigt C, Freibauer A. High CO2 fluxes from grassland on histic Gleysol along soil carbon and drainage gradients. Biogeosciences. 2014;11:749–761. [Google Scholar]
  23. Li B, Yu Z, Liang Z, Song K, Li H, Wang Y, et al. Effects of climate variations and human activities on runoff in the Zoige alpine wetland in the eastern edge of the Tibetan Plateau. Journal of Hydrologic Engineering. 2014;19:1026–1035. [Google Scholar]
  24. Lubbers IM, van Groenigen KJ, Fonte SJ, Six J, Brussaard L, van Groenigen JW. Greenhouse-gas emissions from soils increased by earthworms. Nature Climate Change. 2013;3:187–194. [Google Scholar]
  25. Mather JG, Christensen O. Surface migration of earthworms in grassland. Pedobiologia. 1992;36:51–57. doi: 10.1016/j.ecoenv.2003.08.007. [DOI] [PubMed] [Google Scholar]
  26. McLatchey GP, Reddy KR. Regulation of organic matter decomposition and nutrient release in a wetland soil. Journal of Environmental Quality. 1998;27:1268–1274. [Google Scholar]
  27. Melillo JM, Steudler PA, Aber JD, et al. Soil warming and carbon-cycle feedbacks to the climate system. Science. 2002;298:2173–2176. doi: 10.1126/science.1074153. [DOI] [PubMed] [Google Scholar]
  28. Mosier AR. Soil processes and global change. Biology and fertility of soils. 1998;27:221–229. [Google Scholar]
  29. Mu J, Peng Y, Xi X, Wu X, Li G, Niklas KJ, et al. Artificial asymmetric warming reduces nectar yield in a Tibetan alpine species of Asteraceae. Annals of Botany. 2015;116:899–906. doi: 10.1093/aob/mcv042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Post WM, Emanuel WR, Zinke PJ, Stangenberger AG. Soil carbon pools and world life zones. Nature. 1982;298:156–159. [Google Scholar]
  31. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing; Vienna, Austria: 2016. URL https://www.R-project.org/ [Google Scholar]
  32. Sahagian D, Melack J. Global wetland distribution and functional characterization: Trace gases and the hydrologic cycle. IGBP report 46. 1998 [Google Scholar]
  33. Salm J, Maddison M, Tammik S, Soosaar K, Truu J, Mander Ü. Emissions of CO2, CH4 and N2O from undisturbed, drained and mined peatlands in Estonia. Hydrobiologia. 2012;692:41–55. [Google Scholar]
  34. Sichuan Academy of Forestry. A comprehensive survey on the Zoige National Natural Reserve. 2006 [Google Scholar]
  35. Silvan N, Laiho R, Vasander H. Changes in mesofauna abundance in peat soils drained for forestry. Forest Ecology and Management. 2010;133:127–133. [Google Scholar]
  36. Silvola J, Alm J, Ahlholm U, Nykanen H, Martikainen PJ. CO2 fluxes from peat in boreal mires under varying temperature and moisture condition. Ecology. 1999;84:219–228. [Google Scholar]
  37. Smith LC, MacDonald GM, Velichko AA, Beilman DW, Borisova OK, Frey KE, et al. Siberian peatlands a net carbon sink and global methane source since the early Holocene. Science. 2004;303:353–356. doi: 10.1126/science.1090553. [DOI] [PubMed] [Google Scholar]
  38. Strack M, Waddington JM, Tuittila ES. Effect of water table drawdown on northern peatland methane dynamics: Implications for climate change. Global Biogeochemical Cycles. 2004;18 doi: 10.1029/2003GB002209. GB4003. [DOI] [Google Scholar]
  39. Sun G. A study on the mineral formation law, classification and reserves of the peat in the Ruoergai Plateau. Journal of Natural Resources. 1992;7:334–345. (in Chinese with an English abstract) [Google Scholar]
  40. van Geffen KG, Berg MP, Aerts R. Potential macro-detritivore range expansion into the subarctic stimulates litter decomposition: a new positive feedback mechanism to climate change? Oecologia. 2011;167:1163–1175. doi: 10.1007/s00442-011-2051-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. van Groenigen JW, Lubbers IM, Vos HMJ, Brown GG, De Deyn GB, van Groenigen KJ. Earthworms increase plant production: a meta-analysis. Scientific Reports. 2014;4 doi: 10.1038/srep06365. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Whalen JK, Paustian KH, Parmelee RW. Simulation of growth and flux of carbon and nitrogen through earthworms. Pedobiologia. 1999;43:537–546. [Google Scholar]
  43. Wu P, Zhang H, Cui L, Zhong H, Wang Y. Response of soil macrofauna community to degradation of alpine meadow. Acta Pedologica Sinica. 2013;50:786–799. (in Chinese with an English abstract) [Google Scholar]
  44. Wu W. PhD thesis; Guangzhou University of Chinese Medicine, China: 2008. Studies on the germplasm resources and quality estimation of Dilong (earthworm) [Google Scholar]
  45. Wu X, Griffin JN, Sun S. Cascading effects of predator–detritivore interactions depend on environmental context in a Tibetan alpine meadow. Journal of Animal Ecology. 2014;83:546–556. doi: 10.1111/1365-2656.12165. [DOI] [PubMed] [Google Scholar]
  46. Xiang S, Guo R, Wu N, Sun S. Current status and future prospects of Zoige Marsh in eastern Qinghai–Tibet plateau. Ecological Engineering. 2009;35:553–562. [Google Scholar]
  47. Xiao H, Liu H, Li B, Yuan X, Sun S, Chen Z. Soil fauna diversity in the rotational grazing pasture in a subalpine meadow. Scientia Agricultura Sinica. 2012;45:292–301. (in Chinese with an English abstract) [Google Scholar]
  48. Yang G, Chen H, Wu N, Tian J, Peng C, Zhu Q, et al. Effects of soil warming, rainfall reduction and water table level on CH4 emissions from the Zoige peatland in China. Soil Biology & Biochemistry. 2014;78:83–89. [Google Scholar]
  49. Zhang W, Lu Q, Song K, Qin G, Wang Y, Wang X, et al. Remotely sensing the ecological influences of ditches in Zoige Peatland, eastern Tibetan Plateau. International Journal of Remote Sensing. 2014;35:5186–5197. [Google Scholar]
  50. Zhao C, Griffin JN, Wu X, Sun S. Predatory beetles facilitate plant growth by driving earthworms to lower soil layers. Journal of Animal Ecology. 2013;82:749–758. doi: 10.1111/1365-2656.12058. [DOI] [PubMed] [Google Scholar]
  51. Zorn MI, Van Gestela CAM, Morriena E, Wagenaara M, Eijsackers H. Flooding responses of three earthworm species, Allolobophora chlorotica, Aporrectodea caliginosa and Lumbricus rubellus, in a laboratory-controlled environment. Soil Biology & Biochemistry. 2008;40:587–593. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary information
NIHMS79941-supplement-Supplementary_information.docx (17.9MB, docx)

RESOURCES