Introduction

Arbuscular mycorrhizal (AM) fungi, a well-studied group of root-associated microorganisms, can form symbiotic associations with ~80% of terrestrial plant species and exert profound influences on soil carbon (C) cycling (Averill et al. 2014; Cheng et al. 2012; Lin et al. 2017; Smith and Read 2008; Soudzilovskaia et al. 2015; Wurzburger and Brookshire 2017). AM fungi provide plants with nutrients and water, in return for plant carbohydrates, forming abundant hyphae in the soil (up to 81–111 m cm−3) (Bardgett and van der Putten 2014; Jakobsen et al. 1992; Kiers et al. 2011; Smith and Read 2008). AM hyphae provide an important pathway of C flux from plants to soil, contributing to soil C storage (Averill et al. 2014; Johnson et al. 2002; Kaiser et al. 2015; Qin et al. 2017; Wilson et al. 2009). However, soil C is also released through biological respiration (Hughes et al. 2008; Nottingham et al. 2010; Tome et al. 2016).

Efflux of soil CO2 represents a substantial C loss from terrestrial ecosystems to the atmosphere (Raich and Schlesinger 1992). Soil CO2 efflux is a combination of autotrophic and heterotrophic respiration derived from root respiration and decomposition of litter and soil organic matter (Hanson et al. 2000). Root respiration is a major source of carbon dioxide (CO2) efflux from grassland soils and plays an important role in ecosystem-level carbon cycling (Hasibeder et al. 2015). The contribution of root respiration to total soil respiration of grassland ecosystems ranges from one-third to more than one-half, depending on measurement methods, grassland system, or season (Hasibeder et al. 2015). Previous experiments have indicated that increased belowground biomass is associated with increased respiration (Ru et al. 2017), however this relationship is not clear. In previous studies, AM hyphal respiration is typically treated as a component of root respiration (Hanson et al. 2000; Kuzyakov 2006), but studies conducted in the past decade have suggested AM hyphal respiration can make a substantial contribution to total soil respiration. For example, AM hyphae account for ~14% of total soil CO2 efflux in moist tropical forests in the Republic of Panama (Nottingham et al. 2010), ~11% in cultivated apple orchards located at Vadena (Tome et al. 2016), ~25% of soil respiration in croplands in Thuringia, Germany (Moyano et al. 2007), and ~27% in a temperate established grasslands in North Yorkshire, UK (Heinemeyer et al. 2012). Furthermore, significant seasonal changes in AM hyphal respiration have been observed in several of these previous studies (Heinemeyer et al. 2012; Moyano et al. 2007; Nottingham et al. 2010; Tome et al. 2016), indicating sampling time is an important factor affecting hyphal respiration. Although grassland ecosystems are generally recognized as AM-dominated communities (Averill et al. 2014; Smith and Read 2008), no previous studies have investigated seasonal variations in AM hyphal respiration in grassland field studies or assessed the relative influence of plant functional groups on AM hyphal respiration.

AM fungi are completely reliant upon carbohydrates or fatty acids supplied by host plants (Kiers et al. 2011; Luginbuehl et al. 2017; Smith and Read 2008), which indicates the amount of C released via hyphal respiration is driven by the host-plant. C allocation from host-plant to AM fungi is based on soil fertility and plant/fungal species identity, especially plant functional groups (Grman 2012; Hoeksema et al. 2010; Johnson et al. 2015; Kiers et al. 2011; Liu et al. 2012). C4 grasses or non-N2-fixing forbs often show a more positive response to AM fungi than C3 grasses or N2-fixing forbs (Hoeksema et al. 2010), and plants reward beneficial AM fungi with more carbohydrates (Kiers et al. 2011). In addition, Grman (2012) reported relatively high phosphorus (P) amendments significantly reduced AM hyphal length (an indicator of plant C allocation to AM fungi) associated with Bromus inermis and Elymus repens (C3 grasses) while hyphal length was not altered for Andropogon gerardii and Schizachyrium scoparium (C4 grasses). These findings indicate C4 grasses or non-N2-fixing forbs may provide more C to AM fungi than other plant functional groups. However, C efflux from grassland soil following host-plant allocation to AM fungi is less well understood.

Shading host plants significantly decreases AM fungal abundance by reducing host-plant photosynthetic rates (Johnson et al. 2015), showing AM mutualism may be carbon-limited. Defoliation (removal of plant shoot tissue by grazing animals or machinery) can either promote or reduce C allocation belowground (Bardgett et al. 1998; Barto and Rillig 2010; Gehring and Whitham 2003; Ren et al. 2018), potentially regulating the respiration of AM fungi. Defoliation is a key component of livestock grazing, shaping grassland ecosystems (Allen et al. 2011; Liu et al. 2015; Mikola et al. 2009). A better understanding of how defoliation drives AM hyphal respiration will help predict the influence of grazing on grassland soil C dynamics.

AM hyphal length is likely linked with total soil respiration. Previous studies reported AM hyphae are the largest component of total AM fungal biomass and a substantial contributor to soil C, accounting for 20–30% of total soil microbial biomass and ~20% of plant C soil inputs (Grman 2012; Jakobsen and Rosendahl 1990; Leake et al. 2004; Smith and Read 2008). Carbon-labeling experiments demonstrate AM hypha provide a rapid and crucial pathway of C flux from plants to soil and the atmosphere (Hughes et al. 2008; Johnson et al. 2002; Kaiser et al. 2015; Moyano et al. 2007). Moreover, AM hyphal length is tightly correlated with the formation of soil macroaggregate, potentially sequestering soil C (Kohler et al. 2017; Wilson et al. 2009).

We conducted a field experiment in the grasslands of Inner Mongolia, which are part of the Eurasian steppe, in which ~80% of plant species can be colonized by AM fungi (Bao and Yan 2004; Tian et al. 2009), and mycorrhizal plants typically comprise ~95% of the aboveground plant community biomass (Yang et al. 2014). This temperate grassland ecosystem provides an excellent platform to evaluate the influence of AM fungi on soil C dynamics. In our study, mowing was utilized to simulate defoliation by herbivores (Liu et al. 2015; Mikola et al. 2009). We aim to elucidate AM hyphal contribution to total soil CO2 efflux across grassland plant functional groups, and we hypothesized (1) plant functional groups differentially affect soil CO2 efflux of AM hyphae, other soil microbes, or roots; (2) mowing will reduce AM hyphal length and subsequent hyphal respiration due to reduced plant C allocation belowground; (3) soil respiration is associated with AM hyphal length and root biomass production.

Materials and methods

Study site

This experiment was conducted at the Duolun Restoration Ecology Station of the Institute of Botany of the Chinese Academy of Sciences, which is located in the Inner Mongolian steppe (42°02′N, 116°17′E, 1324 m above sea level), China. Monthly mean temperature was 13.4–21.5 °C from May to September, and monthly mean precipitation was 39.6–77.6 mm (90% falling from May to September) in 2017 (Fig. S1). Soil in this region is classified as Haplic Calcisols (FAO classification). Soil properties (0–10 cm) are: sand 62.8%, silt 20.3%, clay 17.0%, total organic C 1.9%, plant-available nitrogen (N) 28.5 mg kg−1, plant-available P 6.2 mg kg−1, and pH 7.2. Livestock grazing is the major land-use practice in the study area. The study site has been fenced for livestock since 2003 (Tian et al. 2016). Prior to initiating this experiment, the study site was dominated by plants that associate with AM fungi, including Stipa krylovii Roshev., Artemisia frigida Willd., Cleistogenes squarrosa (Trin.) Keng, Agropyron cristatum (L.) Gaertn., Leymus chinensis (Trin.) Tzvelev, and Potentilla acaulis L (Table S1) (Bao and Yan 2004; Lu et al. 2017; Tian et al. 2009; Yang et al. 2014).

Experimental design

We utilized a randomized block design consisting of factorial combinations of four plant functional groups that were either mowed or non-mowed. By removing all plants of each non-target functional group, or leaving an intact diverse plant community (Wu et al. 2015), we established C3 grasses, C4 grasses, forbs, or diverse control treatments (Fig. S2; Table S1). N2-fixing forbs account for less than 2% of the total aboveground biomass in this study site (Yang et al. 2014) and were therefore combined with non-N2-fixing forbs in our study. In total, eight treatments were randomly assigned into five blocks, resulting in 40 plots. Each plot was 1.5 × 1.5 m and separated from the others by a 2-m aisle/buffer.

Removal of non-target plant functional groups was initiated at the beginning of the growing season (May, 2015) by clipping aboveground biomass at the soil surface, and continued every two weeks until August. To simulate defoliation from grazing, in half the plots aboveground biomass of the entire plant community was cut at a height of 5 cm above the soil surface on 15 May, 20 June, 10 July, and 03 August in 2015, 2016, and 2017. This stubble height is representative of cattle grazing in our study area (Liu et al. 2015). Aboveground biomass for each mowed plot was added to the final harvest biomass (25 August in 2017) to estimate annual aboveground net primary production (ANPP). The aboveground biomass in non-mowed plots was harvested on 25 August in 2017 to estimate ANPP. All plant materials harvested from each plot were dried (65 °C for 48 h) and weighed. ANPP was expressed by total aboveground biomass per m2.

Soil respiration determination

To restrict roots and AM hyphae, we utilized three types of soil mesocosms constructed from PVC pipes (Fig. 1) in each plot, creating a gradient of root and AM hyphal restriction (RM treatments). In two of the PVC mesocosms (15 cm in height and 10 cm in diameter), we cut two 10.5 × 8 cm rectangular openings into the sides and covered these openings and the circular bottom of the PVC mesocosm with 25 μm nylon mesh (roots excluded but AM hyphae present; −R + M mesocosm) or 0.45 μm nylon mesh (roots and AM hyphae excluded; −R − M mesocosm) (Fig. 1), following the in-growth core method (Johnson et al. 2001; Nottingham et al. 2010; Tome et al. 2016). Additional PVC tubes (5 cm in height and 10 cm in diameter) without mesh to allow access to roots and AM hyphae (+R + M mesocosm) were utilized; all mesocosms were designed so solid PVC extended to the same soil depth of 3 cm (Fig. 1). Soil cores (13 cm in height and 10 cm in diameter) were randomly collected from each plot using a hammer corer, roots and stones were removed from soil by sieving, and all mesocosms were filled with approximately 450 g (dry wt) of this sieved soil. All mesocosms were installed in the same location in which soil was collected, ensuring equal installation disturbance for all mesocosms. Mesocosms were placed vertically, so mesh-covered openings were completely below the soil surface (Fig. 1). We installed mesocosms on 15 May 2015, at the beginning of the growing season. All PVC tubes extended to a height of 2 cm above the soil surface to allow soil respiration measurements.

Fig. 1
figure 1

Conceptual diagram, illustrating three soil respiration mesocosms. +R + M mesocosms: no mesh (in-growth of roots, AM hyphae, and free-living soil microbes); −R + M mesocosms: 25-μm mesh (restriction of roots, in-growth of AM hyphae, and free-living soil microbes); −R − M mesocosms: 0.45-μm mesh (restriction of roots and AM hyphae, in-growth of free-living soil microbes). Dark colored lines indicate roots and light colored lines indicate AM hyphae

Full size image

Because more than one year was needed for soil conditions (e.g. bulk density and hyphal density in −R + M mesocosm) to recover from installation disturbances (Nottingham et al. 2010), we determined soil respiration two years after mesocosms were installed using LI-8100XT (LiCOR Instruments, Lincoln, NE, USA) at 9:00 a.m. - 12:00 p.m. on 16 June, 06 July, 26 July, 05 August, and 26 August 2017. Soil CO2 efflux measured from +R + M mesocosms (+R + Mefflux) was assumed to result from respiration of roots, AM hyphae, and free-living soil microbes, efflux measured from −R + M mesocosms (−R + Mefflux) was assumed to result from respiration of AM hyphae and free-living soil microbes, while efflux measured from −R − M mesocosms (−R − Mefflux) was assumed to result only from respiration of free-living soil microbes, based on previous methods (Neumann and Matzner 2014; Nottingham et al. 2013; Tome et al. 2016). Plant and fresh leaf litter in each PVC mesocosm were removed from the soil surface before efflux measurements were taken, so contribution of CO2 from litter decomposition and green plants was considered negligible.

Quantification of BNPP and AM hyphal length

We collected two soil cores (3 cm in diameter and 10 cm in depth) from each mesocosm at the conclusion of our soil respiration measurements in late August 2017. Sieved moist soil was air-dried for hyphal length determination. In addition, to improve representation of annual belowground net primary productivity (BNPP), five soil cores (3 cm in diameter and 10 cm in depth), with one located in the center and four in the corners of each plot, were collected and combined to form a composite sample. Roots were separated from soil using a 2-mm sieve and washed, dried at 65 °C for 48 h, and weighed.

Quantification of hyphal length followed the modified membrane filter protocol (Jakobsen et al. 1992). A 4-g soil sample was blended with 250 ml water. Hyphae in 5-ml aliquots were collected on 25-mm membrane filters (1.2-μm pore size) and stained with Trypan Blue. Dark-to light-blue stained aseptate hyphae with characteristic unilateral angular projections were counted as hyphal length (Leifheit et al. 2015). Hyphal length was recorded in 25 random fields of view per filter. The lengths of stained hyphae were determined by the grid line intercept method at 200X magnification. Hyphal length of each soil sample was assessed with six replicates.

To estimate the potential role of ANPP in the strength of plant C allocation to AM hyphal length, we defined a new index: hyphal production efficiency (HPE). This index is derived from the concept of root:shoot biomass ratios and reflects host-plant resource allocation to hyphae. HPE for each plant functional group was calculated as the ratio of AM hyphal production to plant ANPP, where hyphal production is the total hyphal length per m2 in 0–10 cm soil depth. Thus, HPE refers to the hyphal length (m) produced per each gram of plant shoot biomass. Higher HPE indicates greater exploration of soil volume by AM hypha and likely reflects stronger plant C allocation to AM fungi.

Calculations and statistical analysis

Two-way ANOVA was used to analyze the effects of blocks, plant functional groups, mowing, and their interactions on HPE and plant above- and belowground biomass, to determine the magnitude of the response. We employed three-way ANOVA to test the effect of block, plant functional groups, mowing, RM treatment, and their interactions on the length of AM hypha and cumulative CO2 efflux. Cumulative measurements of soil CO2 efflux from 16 June to 26 August in 2017 were estimated by successive linear interpolation and we assumed emissions followed a linear trend over the periods when no samples were taken (Lang et al. 2017). Plant functional group, mowing, and RM treatments were considered as fixed factors, while block was treated as a random factor.

The respective soil CO2 efflux of roots, AM hyphae, and free-living microbes to total soil respiration were calculated as follows (Heinemeyer et al. 2006; Nottingham et al. 2010):

$$ \mathrm{Root}\ \mathrm{respiration}\ \left(\%\right)=\left[{\left(+\mathrm{R}+\mathrm{M}\right)}_{\mathrm{efflux}}-{\left(-\mathrm{R}+\mathrm{M}\right)}_{\mathrm{efflux}}/{\left(+\mathrm{R}+\mathrm{M}\right)}_{\mathrm{efflux}}\right)\Big]\times 100; $$
(1)
$$ \mathrm{AM}\ \mathrm{hyphal}\ \mathrm{respiration}\ \left(\%\right)=\left[{\left(-\mathrm{R}+\mathrm{M}\right)}_{\mathrm{efflux}}-{\left(-\mathrm{R}-\mathrm{M}\right)}_{\mathrm{efflux}}/{\left(+\mathrm{R}+\mathrm{M}\right)}_{\mathrm{efflux}}\right)\Big]\times 100; $$
(2)
$$ \mathrm{Free}-\mathrm{living}\ \mathrm{microbial}\ \mathrm{respiration}\ \left(\%\right)=\left[{\left(-\mathrm{R}-\mathrm{M}\right)}_{\mathrm{efflux}}/{\left(+\mathrm{R}+\mathrm{M}\right)}_{\mathrm{efflux}}\right]\times 100 $$
(3)

The +R + Mefflux, −R + Mefflux and − R − Mefflux are CO2 efflux measured for +R + M, −R + M and − R − M mesocosms, respectively. Repeated-measures ANOVA were used to examine variations in seasonal dynamics of soil CO2 efflux of roots, AM hyphae, and free-living microbes for each plant functional group over the growing season.

Percentage contributions of roots, AM hyphae, and free-living microbes to the total soil CO2 efflux were estimated by each cumulative soil CO2 efflux partition divided by cumulative soil CO2 efflux from −R + Mefflux mesocosms in each plot (Heinemeyer et al. 2006; Nottingham et al. 2010). Mowing or the interactions with other factors did not have significant effects on any soil cumulative CO2 efflux measurement (Fig. S3; Table S2), therefore, we combined mowed and non-mowed plots as replicates for soil CO2 efflux analysis. One-way ANOVA was used to test the effect of plant functional group on cumulative soil CO2 efflux and percentage contributions of roots, AM hyphae, and free-living microbes.

Regression analysis was conducted to determine relationships between AM hyphal length and soil respiration in root-restricting mesocosms (−R − M and − R + M). In addition, regression analysis was conducted to determine the relationship between hyphal length (R + M+) and ANPP as well as the relationship between hyphal length (R + M+) and BNPP for each plot. Least significant difference was used to detect differences between mowed and non-mowed plots. Duncan’s multiple range test was used to compare differences among plant functional groups and RM treatments. All ANOVAs were analyzed using SAS version 8.0 (SAS Institute, Cary, NC, USA, 2002) and regression analysis was performed with Sigma plot 12.0 (Systat Software Inc., San Jose, CA, USA).

Results

AM hyphal length and plant productivity

No roots were detected in −R + M or − R − M mesocosms. Root restriction simultaneously decreased AM hyphal length in forbs and the diverse (control) plant communities but did not affect hyphal length in C3 grasses and C4 grasses (Fig. 2). Restriction of AM hyphae and roots (−R − M) significantly decreased hyphal length in C3 grasses with mowing, C4 grasses, forbs and diverse plots compared to restriction of roots but not AM hyphae (−R + M) (Fig. 2). Mowing increased hyphal length in the −R + M and + R + M mesocosms of C3 grasses.

Fig. 2
figure 2

Effect of mowing (M) and root/mycorrhizal restriction (RM) treatments on extra-radical mycorrhizal hyphal length in (a) C3 grasses, (b) C4 grasses, (c) forbs, and (d) diverse (control) plant communities. +R + M mesocosms: no mesh (in-growth of roots, AM hyphae, and free-living soil microbes); −R + M mesocosms: 25-μm mesh (restriction of roots, in-growth of AM hyphae, and free-living soil microbes); −R − M mesocosms: 0.45-μm mesh (restriction of roots and AM hyphae, in-growth of free-living soil microbes). Bar groups with different capital letters indicate a significant difference (p < 0.05) among plant functional groups within mowed or no-mowed treatment, and # indicates a significant difference between mowed and non-mowed plots within a plant functional group in a given year. Bars represent means +SE. * = p < 0.05; ** = p < 0.01; *** = p < 0.001; NS = p > 0.05

Full size image

Mowing significantly increased ANPP for each plant functional group (Fig. S4). In contrast, mowing did not alter BNPP. There were significant biomass production differences between plant functional groups. C3 grasses produced more ANPP than other plant functional groups, while diverse plots had the greatest BNPP.

Soil respiration

Plant functional group and sampling date significantly affected total soil CO2 efflux, (Table 1; Fig. 3a). Greater total soil CO2 efflux was detected in diverse communities, while C4 grasses tended to have lowest total soil CO2 efflux. Total CO2 efflux peaked at the middle of the growing season. Similarly, soil CO2 efflux of free-living soil microbes and AM hyphae also reached a peak at the middle of the growing season (Table 1; Fig. 3b and c). There was lower microbial CO2 efflux but higher AM fungal respiration in C3 grasses plots at mid-season (Fig. 3b and c), while forbs and diverse plots had higher microbial CO2 efflux but lower AM fungal respiration at the middle of growing season (Fig. 3b and c). However, sampling date did not alter soil CO2 efflux of roots (Table 1; Fig. 3d). Overall, the diverse plots tended to have higher root CO2 efflux.

Table 1 F ratios resulting from repeated-measures ANOVA testing the effects of plant functional groups (F) and sampling date (D) on the soil CO2 efflux of free-living soil microbes, AM hyphae, and roots
Full size table
Fig. 3
figure 3

Seasonal dynamics of soil CO2 efflux of (a) total respiration, (b) microbial respiration, (c) AM fungal respiration, and (d) root respiration. Each respiration was calculated using values from CO2 efflux accumulated in +R + M (in-growth of roots, AM hyphae, and free-living soil microbes); −R + M (restriction of roots, in-growth of AM hyphae and free-living soil microbes) and − R − M (restriction of roots and AM hyphae, in-growth of free-living soil microbes) mesocosms, according to Eqs. 1–3 (see methods). Bars represent means ±SE. The X-axis data indicate each respiration measurement date

Full size image

Restricting roots and/or AM hyphae (RM treatments) significantly decreased cumulative soil respiration of each plant functional group, while mowing and its interaction with RM treatment did not affect soil respiration of any plant functional group (Fig. S3). The cumulative soil respiration of C3 grasses was highest in +R + M mesocosm (20.89 ± 1.64 mol m−2), immediate in −R + M (16.75 ± 1.19 mol m−2) and lowest in −R − M (9.97 ± 0.50 mol m−2) mesocosms (Fig. S3a). The cumulative soil respiration of C4 grasses in +R + M and − R + M mesocosms was on average 44% higher than in −R − M mesocosm (Fig. S3b). A similar trend was found in forbs and diverse plots, with at least 27% higher cumulative soil respiration in +R + M mesocosm, relative to −R + M and − R − M mesocosms (Fig. S3c and d).

Plant functional group significantly affected cumulative free-living microbial respiration, as forbs or diverse plots produced at least 25% greater efflux, compared to C3 or C4 grasses plots (Table S3). C3 grasses plots had three times greater cumulative AM hyphal respiration, compared to forb communities or diverse plots. Cumulative root respiration in diverse plots was two times greater than C4 grasses, while there were no significant differences among C3 grasses, C4 grasses, or forbs.

We calculated percentage contributions of roots, AM hyphae, and free-living microbes. Based on soil cumulative CO2 efflux of each plant functional group, AM hyphal respiration constituted 32% (C3 grasses), 23% (C4 grasses), 9% (forbs), and 9% (diverse plots) of total soil respiration (Table 2). Plant functional group had no effect on the relative contribution of roots to total soil respiration. Significantly greater AM hyphal contributions to soil respiration occurred only in C3 grasses plots. The greatest relative contribution of free-living soil microbes to soil respiration was assessed in forb community plots, and the lowest contribution was assessed in C3 grasses plots.

Table 2 The relative contributions (%) of free-living soil microbes, AM hyphae and roots to soil respiration for each functional group
Full size table

Soil respiration in relation to hyphal length and BNPP

AM hyphal respiration was positively correlated with hyphal length in the root-restricting mesocosms (−R + M) (Fig. 4a). Soil CO2 efflux was significantly correlated with BNPP in non-restricting (+R + M) mesocosms (Fig. 4b). But the correlation between the two sets of data was weak (Fig. 4a: R2 = 0.1703, Fig. 4b: R2 = 0.1376).

Fig. 4
figure 4

Soil CO2 emission in relation to (a) AM hyphal length in –R + M mesocosms (restriction of roots, in-growth of AM hyphae, and free-living soil microbes), and (b) belowground net primary productivity (BNPP) in +R + M (in-growth of roots, mycorrhizal hyphae, and free-living soil microbes) mesocosms. Statistics (R2 and p-values) for regressions are indicated, and dotted lines indicate the 95% confidence interval. Symbol reference: “∆” = C3 grasses, “○” = C4 grasses, “□” = forbs, and “◊” = diverse (control) plant communities. Solid black = mowed plots, open symbols = non-mowed plots

Full size image

Hyphal length in relation to ANPP and HPE of plant functional group

ANPP was significantly correlated with hyphal length across all plant functional groups (Fig. 5). HPE was strongly affected by plant functional group and mowing (Fig. 6). C4 grasses and forbs had higher HPE than C3 grasses in mowed or non-mowed treatments. Mowing significantly reduced HPE of C4 grasses, forbs, and the diverse plant communities but did not affect that of C3 grasses.

Fig. 5
figure 5

Relationship between annual aboveground net primary productivity (ANPP) and extra-radical mycorrhizal hyphal length. Regression was estimated using a linear model with ANPP as a continuous predictor. Statistics (R2 and p-values) for regressions are indicated, and dotted lines indicate the 95% confidence interval. Symbol reference: “∆” = C3 grasses, “○” = C4 grasses, “□” = forbs, and “◊” = diverse (control) plant communities. Solid black = mowed plots, open symbols = non-mowed plots

Full size image
Fig. 6
figure 6

The effect of mowing (M) on (a) hyphal production efficiency (HPE) of different plant functional groups (F): C3 grasses, C4 grasses, forbs, and diverse (control) plant communities. Bar groups with different capital letters indicate a significant difference (p < 0.05) among plant functional groups within mowed or no-mowed treatment, and # indicates a significant difference between mowed and non-mowed plots within a plant functional group. Bars represent means +SE. * = p < 0.05; ** = p < 0.01; *** = p < 0.001; NS = p > 0.05

Full size image

Discussion

Our study provides field evidence from temperate grasslands that AM hyphal respiration is an important source of soil CO2 efflux. Consistent with our hypothesis, soil CO2 efflux of AM hyphae, other soil microbes, and roots was significantly different among plant functional groups. Furthermore, AM hyphal respiration was positively correlated with hyphal length in in-growth mesocosms. AM hyphal and microbial respiration peaked at the middle of the May to August growing season, while no significant seasonal variations in root respiration were detected. Contrary to our hypothesis, mowing did not influence hyphal respiration.

Our results suggest the contribution of roots to total soil respiration was >30%, while AM hyphae was <10%, and free-living soil microbes contributed the remaining ~60% in diverse plant communities. In comparison, a previous study in a temperate established grassland showed roots only contributed ~10% of soil respiration, while AM hyphae produced >25%, with free-living microbes contributing the remaining ~60% (Heinemeyer et al. 2012). While free-living soil microbes appear to have similar contributions regardless of study, it is notable that the contributions of AM hyphae vs plant roots were quite different. Our study suggests that these substantial differences in proportional contribution of AM hyphae to soil carbon flux is strongly influenced by host-plant functional group. Previous studies indicate host identity affects the contribution of AM hyphae to soil respiration. For example, contribution of AM hyphae was ~10% in an apple orchard (Tome et al. 2016), and ~15% in a moist tropical forest (Nottingham et al. 2010), but was reported as >25% when associated with a C3 grain-crop (Moyano et al. 2007).

We found that plant functional groups differ in their production of AM hyphae, and hyphal length was positively correlated with soil CO2 efflux in root-restricting mesocosms. These results indicate that plant functional group may influence soil respiration by mediating AM fungal abundance. This confirms previous studies linking AM fungal abundance and soil respiration using mesh in-growth cores and isotope methods (Heinemeyer et al. 2006; Moyano et al. 2007; Tome et al. 2016). In our study, AM hyphal cumulative respiration was greater in C3 grasses compared to all other plant functional groups. C3 grasses produced greater ANPP, which was positively correlated with hyphal length, therefore, we propose the greater aboveground biomass drove soil C inputs and consequently increased hyphal respiration. C3 grasses are always dominant and fast-growing species in our study site, indicating a high photosynthetic capacity, providing abundant C to AM fungal partners. We also predicted the turnover rate of AM hyphae associated with C3 grasses would be more rapid than other plant functional groups.

Previous studies used hyphal length, AM fungal root colonization, or concentration of AM phospholipid biomarkers to represent C allocation from plants to AM fungi (Grman 2012; Vestberg et al. 2012; Walder et al. 2012). These approaches did not consider the potential effects of ANPP, which may be a critical aspect of determining host-plant ability for C allocation to AM fungi. In our study, the significant relationship between ANPP and hyphal length suggests C allocation from plants to AM fungi are indeed linked. Therefore, HPE might be a more accurate indicator for comparing C allocation from plants to AM fungi across plant functional groups or plant communities differing in productivity. We suggest HPE can be widely used for distinguishing shifts in resource allocation with variation in edaphic or environmental conditions and across functional, taxonomic, physiological, or phenological plant functional groups, allowing for large quantities of HPE data to be generated globally.

Our results indicate C4 grasses, forbs, or diverse plant communities were more efficient in hyphal production compared with C3 grasses. These findings are consistent with each functional group’s dependency on AM fungi, as C4 grasses and forbs typically respond more positively to AM symbiosis compared with C3 grasses (Hoeksema et al. 2010; Wilson and Hartnett 1998). Generally, C4 grasses and forbs allocate more carbohydrates or fatty acids to AM fungi, compared to C3 grasses, for enhanced nutrient uptake. As C3 grasses were associated with lower HPE, yet we observed greater hyphal respiration suggests AM symbiosis with C3 grasses may be responsible for greater soil C loss in grasslands. However, forbs and C4 grasses tend to more efficiently produce AM hyphae, yet with lower hyphal respiration, indicating AM symbiosis with these hosts can provide a potential benefit through increased C sequestration in grassland soils.

Plant functional groups may also affect AM hyphal respiration by mediating the composition and diversity of AM fungal communities. It has been shown that plants select for the AM fungi that provide the greatest benefit in local environments (Johnson et al. 2010), indicating the composition of AM fungal communities are strongly influenced by the plant host. Furthermore, AM fungal species differ in their functional traits, including patterns in carbon allocation from host plants, development of AM fungal root colonization, and extra-radical hyphal length (Dai et al. 2013; Engelmoer and Kiers 2015). Therefore, alterations in composition and diversity of AM fungal communities have the potential to regulate AM hyphal respiration, and, while beyond the scope of our current study, should be investigated in future studies.

While mowing increased AM hyphal length, it did not significantly alter hyphal respiration in C3 grasses plots. Plants tend to upregulate C allocation to AM fungi when they obtain greater benefits from the association, such as acquiring limiting soil nutrients (Kiers et al. 2011; Smith and Smith 2011). Our results suggest mowing increased AM hyphal length, presumably because mowing alters plant C resource distribution, promoting greater benefit from mycorrhizal symbioses (Bardgett et al. 1998; Eom et al. 2001; Gehring and Whitham 2003). Previous studies showed AM hyphal CO2 efflux is determined by recent plant C supply (Heinemeyer et al. 2006; Kaiser et al. 2015). In contrast, our results suggest mowing did not alter hyphal respiration in spite of reduced plant C supplies after aboveground biomass removal. Our study indirectly supports the hypothesis suggested by Nottingham et al. (2010) that root carbohydrate reserves maintain AM hyphal C supply following low-intensity defoliation.

ANPP showed a positive linear relationship to AM hyphal length, and mowing increased both ANPP and hyphal length compared to non-mowed plots, indicating hyphal length is promoted by increased ANPP. ANPP is indicative of C assimilation, determining plant C supply to AM fungi (Johnson et al. 2015; Smith and Read 2008). This linear relationship between ANPP and hyphal length can also be interpreted as a positive effect of hyphal length on plant growth. Several studies observed hyphal length was significantly correlated with AM functioning, such as plant N and P uptake, stability of soil macroaggregates, or soil organic C (Moore et al. 2015; van der Heijden 2004; Wilson et al. 2009). Our results provide field evidence that ANPP is linearly and positively associated with hyphal length, emphasizing the essential role of plant C assimilation in AM fungal abundance and illustrating that AM functioning is related to AM abundance.

Sampling date exerted a significant effect on total, hyphal and microbial respiration associated with different plant functional groups. Hyphal and microbial respiration peaked at the middle of the growing season in our study. While there was no significant seasonal variation in hyphal respiration in moist tropical forests (Nottingham et al. 2010), similar seasonal dynamics in total soil respiration and precipitation have been previously reported at our study site (Xia et al. 2009). In addition to demonstrating the seasonal dynamics of AM fungal and microbial soil CO2 efflux, our study indicated that root respiration did not show significant seasonal variations. This indicates that root CO2 efflux was probably not limited by water or temperature during the grown season. Therefore, our study suggests higher soil respiration at the middle of the growing season are due to the contribution of hyphal and microbial respiration, rather than root respiration. Sampling date is a combined factor including variation in temperature, precipitation, and plant phenological stage, which potentially affect plant C flow to AM fungi and other soil microbes (Barrett et al. 2014; Birgander et al. 2017).

Regression models with low R-squared values (0.14–0.29) were observed in our study (Figs. 4 and 5), indicating there is still at least 70% of the variation in the data left unexplained. There is inherent greater amount of unexplainable variation in field study. Furthermore, AM hyphal length has been underestimated because hyphal length was determined at the soil depth of 10 cm and the AM fungal respiration is also underestimation as a result of the drawbacks for the in-growth core method. It is important to note our experimental design, developed following previous studies of Moyano et al. (2007) and Nottingham et al. (2010), may underestimate mycelial and microbial respiration, while overestimating root respiration. The 0.45-μm nylon mesh (−R − M) excluded microorganisms from the rhizosphere, potentially leading to an underestimation of microbial respiration. The 20-μm nylon mesh (−R + M) was designed to restrict roots, however this may have inadvertently reduced hyphal length, as hyphal length generally decreases with distance from plant roots (Thonar et al. 2011). In addition, the incomplete transparency of –R + M mesocosms (i.e. constructed with only two rectangular openings on PVC tube), compared to mesocosms designed without restriction of roots and AM mycelia (+R + M) may have further reduced AM hyphal length, due to lower availability for hyphal access. Therefore, respiration estimates of root restriction mesocosms (−R + M) may underrepresent respiration of AM hyphae. Moreover, interactions between roots, AM fungi, and free-living bacteria all influence soil respiration in combination with both biotic and abiotic factors. For example, transfer of water and nutrients from AM fungi to bateria can stimulate bacterial activity, regulating C flux from soil to the atmosphere (Moore et al. 2015; Worrich et al. 2017). Roots colonized by AM fungi affect root respiration (Valentine and Kleinert 2007). Our results illustrate AM hyphal respiration is an important component of total soil respiration in grassland ecosystems; these ecological interactions need to be further explored and teased apart in future studies.

Conclusions

Our study suggests AM hyphal and microbial respiration are a critical source of soil respiration in temperate grasslands. Plant functional group is the main factor regulating AM hyphal and microbial respiration at different stages of growing season. Plant functional group influences AM hyphal contribution to total soil efflux, with C3 grasses contributing the greatest soil CO2 efflux compared to other plant functional groups. Moreover, C3 grasses have lower hyphal production efficiency than C4 grasses, forbs and diverse plant communities. Anthropogenic changes can drive variation in plant community composition and biodiversity loss, likely affecting soil C sequestration via AM hyphal respiration. For example, plant communities can be altered by grazing or N deposition, with a resultant shift to C3 grasses dominance (Clark and Tilman 2008; Song et al. 2011). Our research suggests this shift in plant community composition will increase AM hyphal respiration and reduce hyphal production efficiency, an important contribution to soil C losses. Diminished soil C has broad implications spanning from declining local grassland productivity to disruption of global carbon cycles.