Changes in soil faunal density and microbial community under altered litter input in forests and grasslands

Abstract

Root and foliar litter inputs are the primary sources of carbon and nutrients for soil fauna and microorganisms, yet we still lack a quantitative assessment to evaluate the effects of root and foliar litter on various groups of soil organisms across terrestrial ecosystems. Here, we compiled 978 paired observations from 68 experimental sites to assess the directions and magnitudes of adding and removing foliar and root litter on the soil faunal density and microbial biomass that was evaluated by phospholipid fatty acids (PLFAs) across forests and grasslands worldwide. We found that litter addition had only a marginal effect on soil faunal density but significantly increased the soil total microbial-, fungal- and bacterial-PLFAs by 13%, 14%, and 10%, respectively, across ecosystems, suggesting that the soil microbial community is more sensitive to carbon source addition than soil fauna, particularly in soils with low carbon to nitrogen ratios. In contrast, removing litter significantly decreased the soil faunal density by 17% but had few effects on soil microorganisms. Compared with foliar litter, root litter input had a more positive effect on the development of soil fungal taxa. The effect of both litter addition and removal on soil faunal density and microbial biomass did not differ between humid and arid regions, but a greater influence was observed in grasslands than in forests for soil microbial community. Our results highlight that the increasing litter production under a global greening scenario would stimulate microbial activity in grasslands more than in forests, and this stimulation would be greater for soil microbes than soil fauna.

Graphical abstract

1. Introduction

More than 50% of the global terrestrial plant production returns to soil surface as plant litter [1]. Plant litter provides organic carbon and nutrients that are required for soil biota [2], [3], which directly accelerate organic matter mineralization and nutrient cycling by digesting and fragmenting plant litter [4], [5]. It is well documented that global lands have been greening since the 1980s [6]. Greening is an encouraged strategy to combat the consequences of climate change, such as drought and increasing CO₂ emissions, and to protect biodiversity [7]. Litter production will change with increasing net productivity due to global greening [8], [9]. However, 45.6% of the global vegetated area, especially that in evergreen broadleaved forests and grasslands, has experienced inconsistent trends in vegetation cover and aboveground net productivity [10], further increasing the uncertainty of litter production and its couples with soil biota involved in biogeochemical cycles.

Detritus input and removal treatments (DIRTs) have been widely used as an effective experimental method to assess the effect of plant litter on soil microbial biomass and community structure by manipulating litter inputs in terrestrial ecosystems [11], [12]. Litter incorporates in a substantial part of soil food web and provides carbon source for soil organisms [13], [14], and there is a conventional view that increasing litter inputs increase the biomass of soil organisms that dominate detrital food webs [15], [16]. However, inconsistent results have been observed in many studies. For example, Nadelhoffer et al [17]. did not find any changes in microbial biomass in a deciduous forest after adding aboveground litter, and several other studies have also reported that litter removal had little impact on microbial biomass and community [18], [19]. Litter input not only provides carbon and nutrients for soil organisms, but also affects the habitat of soil biological community by regulating soil surface microclimate, such as light, temperature and moisture [20]. The inconsistent results observed in different studies are unlikely to provide an integrated understanding of how litter input affects soil fauna and microbial communities, and this knowledge gap greatly limits our understanding of the flows of carbon and nutrients. A recent meta-analysis found that double litter inputs increased the total PLFAs by 19.7% in various forests, but whether this influence varies between litter types (i.e., foliar and root litter) and microbial communities (i.e., bacteria and fungi) remains unclear.

Plant litter contains foliar, twig, root, flower, and other components in terrestrial ecosystems [21]. Generally, foliar litter accounts for more than 60% of the aboveground litter, representing the largest substrate input to the soil [22]. Moreover, root litter represents approximately 33% and 48% of the annual litter input in grasslands and forests, respectively [23], [24]. Different biotic taxa have distinct feeding preferences such that the microbial community structure may change with different resource availability [25], [26]. It is widely accepted that foliar litter with low carbon to nitrogen ratio (C/N) supports bacterial energy channels, and the abundance of bacterial taxa increases with large amounts of labile C input from high-quality foliar litter with lower C/N ratios [27], [28], [29] and thus facilitates fast nutrient turnover. In contrast, root litter (low-quality with a higher C/N ratio than foliar litter) usually stimulates oligotrophic microbial communities (i.e., fungi) that degrade recalcitrant polymers (such as lignin and cellulose) through a large C investment to produce extracellular enzymes [30], [31], [32] with slower nutrient turnover [33]. Therefore, different resources input may have a profound influence on the microbial community due to the divergent metabolic strategies of the bacterial and fungal communities in response to substrate quality [34]. Although the coupling between litter inputs and soil biotic community has important implications for terrestrial biogeochemical cycles, our current understanding is not fully understood.

In this study, we hypothesize that litter addition increases soil faunal density and microbial biomass, with a greater effect of foliar litter than root litter. Here, a meta-analysis was conducted by collecting 978 paired observations (with and without litter input) from 68 sites to assess the influence of root and foliar litter addition and removal on soil faunal density and microbial biomass (evaluated by PLFAs) across forests and grasslands. Our objectives were 1) to quantify the effect of litter addition/removal on soil faunal density and PLFAs and 2) to distinguish the potential factors driving the responses of soil organisms to litter inputs.

2. Materials and methods

2.1. Data collection

We conducted a comprehensive literature search for peer-reviewed articles published before December 2020 through Web of Science (http://apps.webofknowledge.com/) and China National Knowledge Infrastructure (http://www.cnki.net/). We compiled studies that reported the impacts of litter addition and removal on soil faunal density and microbial biomass that were evaluated by PLFAs. The search terms used in this meta-analysis were “detritus/litter exclusion” OR “detritus/litter removal” OR “detritus/litter input” OR “detritus/litter addition” OR “detritus/litter manipulation” AND “soil fauna” OR “soil animal” OR “microbial biomass” OR “phospholipid fatty acid” OR “PLFA”.

To avoid publication bias from different studies, the following criteria were applied during data compilation.

First, only foliar and root litter under increased and/or reduced litter inputs were considered in this study because studies reporting other litter types (i.e., twig and wood) were limited.

Second, studies that applied multiple magnitudes of litter addition and/or removal were considered independent. In this study, the control plots received natural litter input at the study sites, while the litter addition plots received twice as much litter as the control plots.

Third, the microbial biomass was evaluated by the phospholipid fatty acid (PLFAs) analysis. The PLFAs were categorized as total-, fungal-, bacterial-, Gram-positive bacterial-, and Gram-negative bacterial-PLFAs in this study.

Fourth, all data included in this study were retrieved from field measurements, and data from laboratory incubations and model predictions were omitted. Averaged values from different sampling times were excluded to ensure that all data were independent. All data were compiled from only forests and grasslands and not from other ecosystems.

Fifth, no other factors (i.e., increasing temperature, nitrogen addition, etc.) except for litter addition or removal were considered in this study. For example, one study [35] considered four levels of N addition (no N addition, low N, medium N, and high N) nested in three litter input treatments. Here, only the data collected from the ambient plots without N addition were included in our dataset. According to this principle, the different sampling sites and litter types were considered to be independent and appropriate for inclusion in this meta-analysis.

Sixth, the means, standard deviations (or standard errors), and sample sizes of the variables should be provided directly or could be calculated. If the data were presented in figures, we used Engauge Digitizer 4.1 (Free Software Foundation Inc., Boston, MA) to extract these data. Data contained in Supporting Information were also included in our dataset.

Seventh, other information, such as the coordinates, altitude, and mean annual temperature (MAT) and precipitation (MAP) of the experimental sites was also included in our dataset. If the coordinates of the experimental sites were not provided in a given article, we extracted the coordinates from other studies with experiments conducted at the same site. The missing elevation data were retrieved by Google Earth (Google Inc., Santa Clara, CA, USA) according to the given coordinates. The missing MAT and MAP data were retrieved from the National Oceanic and Atmospheric Administration (http://www.noaa.gov) at the meteorological station closest to the experimental site. We also included the experimental duration, soil chemical properties (pH and C/N ratio), and litter C/N ratio.

A total of 978 paired observations were retrieved from 68 articles that reported soil fauna density and microbial PLFAs under litter addition or removal in forests and grasslands (Table S1). The latitude ranged from 41.00 °S to 56.01 °N (Fig. 1), the altitude ranged from 4 m to 3300 m, the MAT ranged from −2.0 °C to 27.6 °C, and the MAP ranged from 331 mm to 3720 mm.

Fig. 1.

Global distribution of the experimental sites included in this meta-analysis. Red circles represent forests, and blue circles represent grasslands. The symbol sizes are proportional to the sample size.

2.2. Data analysis

We calculated the individual response ratio ($\ln R$) to assess the effects of litter addition/removal on soil faunal density and microbial PLFAs. $R$ is defined as the ratio of the mean value of a given variable in the treatment group (${\overline{X}}_t$) to that in the control group (${\overline{X}}_c$) (Eq. (1)), which follows a normal distribution with a Gaussian function (Fig. S1). The response ratio was calculated as follows:

$$\ln R=\ln \left(\frac{{\overline{X}}_t}{{\overline{X}}_c}\right)$$ (1)

For each study, the standard deviations ($S$) for both the control and treatment groups were extracted to calculate the variance (v; Eq. (2)) and the weighting factor for each entry (w_ij; Eq. (3)). Then, weighted response ratios ($\ln RR$; Eq. (4)) were calculated by random mixed effect model that fitted $\ln R$ as the response variable and included study identity as a random effect factor to include both the within-study variance and an extra variance component to account for between-study variance [36]. Given the limited number of paired observations, we used a resampling test with 999 interactions to assess the 95% bootstrap confidence intervals ($\mathrm{C}{\mathrm{I}}_S$; Eqs. (5,6)):

$$v=\frac{S_t^2}{n_t\overline{X_t^2}}+\frac{S_c^2}{n_c\overline{X_c^2}}$$ (2)

$$w_{ij}=\frac{1}{v}$$ (3)

$$\ln RR=\frac{{\sum }_{i=1}^n{w}_i\ln R_i}{{\sum }_{i=1}^n{w}_i}$$ (4)

$$S_{\ln RR}=\sqrt{\frac{1}{{\sum }_{i=1}^n{w}_i}}$$ (5)

$$95\%\mathrm{C}{\mathrm{I}}_S=\ln RR\pm 1.96S_{\ln RR}$$ (6)

where ${\overline{X}}_t$, $S_t$, and $n_t$ represent the means, standard deviations and sample sizes of treatment group (t), respectively, and ${\overline{X}}_c$, $S_c$, and $n_c$ are the means, standard deviations and sample sizes of control group (c), respectively.

We also calculated the percentage change (P, Eq. (7)) in each category based on the weighted response ratio as follows:

$$P(\%)=\left(e^{\ln RR}-1\right)\times 100$$ (7)

The response of a variable to litter addition or removal was considered to be significant if the 95% CI did not overlap zero. If the percentage change and 95% CI of a variable in response to litter addition/removal were greater than zero, it was considered a "significant positive effect". If the percentage change and 95% CI of a variable in response to litter addition/removal were less than zero, it was considered a "significant negative effect".

Furthermore, we used the aridity index ($I_a$) defined by De Martonne [37] to classify the aridity degree of the study sites as shown in Eq. (8):

$$I_a=\frac{\text{MAP}}{\text{MAT}+10}$$ (8)

where MAP and MAT represent the mean annual precipitation (mm) and mean annual temperature (°C), respectively. The study sites where $I_a$ was less than or equal to 30 were considered to be arid zones, and sites with $I_a$ values greater than 30 were considered to be humid zones. In this study, 10.6% and 89.4% of the sites were located in arid and humid zones, respectively (Table S1).

We used a Mann–Whitney test to examine whether $\ln RR$ of variables differed significantly among ecosystem types, litter types, experimental durations, and climatic zones. We also conducted linear regression analyses to test the correlations of soil biotic community (soil faunal density, and total-, fungal-, bacterial-, Gram-positive bacterial-, and Gram-negative bacterial-PLFAs) with climatic factors (MAT and MAP), edaphic factors (soil C/N ratio and pH), and litter C/N ratio. These analyses were performed in SPSS 13.0 (SPSS Ins., Chicago, IL, USA).

3. Results

3.1. Overall responses of soil organisms to litter addition/removal

Litter addition did not significantly affect soil faunal density and Gram-positive bacterial-PLFAs, but significantly increased total-, fungal-, bacterial-, and Gram-negative bacterial-PLFAs, by an average of 13%, 14%, 10%, and 14%, respectively. Litter removal significantly decreased the soil faunal density by 17%, while the total-, fungal-, bacterial-, Gram-positive bacterial-, and Gram-negative bacterial-PLFAs changed marginally (Fig. 2).

Fig. 2.

Percentage changes in soil faunal density and PLFAs in response to litter (a) addition and (b) removal. Error bars represent the 95% confidence intervals. The sample sizes are shown next to the bars. PLFAs: phospholipid fatty acids.

3.2. Factors influencing the response of soil biota to litter addition/removal

Foliar litter addition significantly stimulated soil total-, fungal-, bacterial-, and Gram-negative bacterial-PLFAs, whereas soil faunal density and Gram-positive bacterial-PLFAs showed little change (Fig. 3). Soil faunal density, and total-, Gram-positive bacterial-, and Gram-negative bacterial-PLFAs decreased after litter removal, regardless of foliar and root litter removal. Removing root litter resulted in a greater decrease in soil fungal-PLFAs than removing foliar litter, but the decrease in soil bacterial-PLFAs after foliar litter removal was greater than that after root litter removal (Fig. 3).

Fig. 3.

Percentage changes in soil faunal density and PLFAs in response to litter addition (green) and removal (orange) with different litter types. (a) Faunal density, (b) total-PLFAs, (c) fungal-PLFAs, (d) bacterial-PLFAs, (e) Gram-positive bacterial-PLFAs, and (f) Gram-negative bacterial-PLFAs. Error bars represent the 95% confidence intervals. The sample sizes are shown next to the bar. Lowercase letters indicate significant differences among different litter types (P<0.05). PLFAs: phospholipid fatty acids.

Litter addition had greater positive effects on soil fungal-, Gram-positive bacterial-, and Gram-negative bacterial-PLFAs in grasslands than in forests, while the effect of litter removal did not differ between forests and grasslands (Fig. 4). Moreover, the effects of litter addition and removal on soil fauna density, and total-, fungal-, bacterial-, Gram-positive bacterial-, and Gram-negative bacterial-PLFAs were marginally different between humid and arid zones (Fig. 5).

Fig. 4.

Percentage changes in soil faunal density and PLFAs in response to litter addition (green) and removal (orange) in different ecosystems. (a) Faunal density, (b) total-PLFAs, (c) fungal-PLFAs, (d) bacterial-PLFAs, (e) Gram-positive bacterial-PLFAs, and (f) Gram-negative bacterial-PLFAs. Error bars represent the 95% confidence intervals. The sample sizes are shown next to the bar. Lowercase letters indicate significant differences between forests and grasslands (P<0.05). PLFAs: phospholipid fatty acids.

Fig. 5.

Percentage changes in soil faunal density and PLFAs in response to litter addition (green) and litter removal (orange) in different climatic zones. (a) Faunal density, (b) total-PLFAs, (c) fungal-PLFAs, (d) bacterial-PLFAs, (e) Gram-positive bacterial-PLFAs, and (f) Gram-negative bacterial-PLFAs. Error bars represent the 95% confidence intervals. The sample sizes are shown next to the bar. Lowercase letters indicate significant differences between humid and arid regions (P<0.05). PLFAs: phospholipid fatty acids.

A greater increase in soil total-PLFAs under litter addition was observed in studies with a duration longer than 1 year than those with periods shorter than 1 year. However, this effect on soil faunal density tended to decrease more under long-term (>1 yr) than short-term (≤1 yr) manipulations. Duration-independent was found for both soil faunal density and total-PLFAs after litter removal (Fig. 6).

Fig. 6.

Percentage changes in soil fauna density and PLFAs in response to litter addition (green) and removal (orange) for experiments with different durations. (a) Faunal density, (b) total-PLFAs, (c) fungal-PLFAs, (d) bacterial-PLFAs, (e) Gram-positive bacterial-PLFAs, and (f) Gram-negative bacterial-PLFAs. Error bars represent the 95% confidence intervals. The sample sizes are shown next to the bar. Lowercase letters indicate significant differences between experimental durations of longer and shorter than 1 year (P<0.05). PLFAs: phospholipid fatty acids

3.3. Faunal and microbial responses versus climatic and edaphic factors

The response of soil faunal density to litter addition decreased with increasing elevation (Fig. 7a), while the response of bacterial-PLFAs increased linearly with increasing MAT and MAP (Fig. 7k and l). The response of Gram-negative bacterial-PLFAs to litter addition showed a significant negative correlation with MAT and MAP (Fig. 7q and r).

Fig. 7.

Response ratios of soil faunal density and PLFAs related to elevation, MAT and MAP. (a-c) Faunal density, (d-f) total-PLFAs, (g-i) fungal-PLFAs, (j-l) bacterial-PLFAs, (m-o) Gram-positive bacterial-PLFAs, and (p-r) Gram-negative bacterial-PLFAs. The fitted lines represent slopes from linear mixed effects, with 95% confidence intervals (shaded) shown in each panel. The sizes of the circles represent the relative weights by the weighting factors for each paired observation. The sample sizes, adjusted R², and P values are shown in each panel for litter addition and removal separately. The dotted lines are shown at effect sizes of zero. PLFAs: phospholipid fatty acids, lnR: response ratio.

The response of soil faunal density and total-, fungal-, and Gram-negative bacterial-PLFAs to foliar litter addition decreased significantly with increasing soil C/N ratio (Fig. 8). However, the responses of soil faunal density and fungal-PLFAs increased with increasing litter C/N ratio in forests but not in grasslands (Fig. 9a and b), while the response of Gram-negative bacterial-PLFAs increased in both forests and grasslands (Fig. 9f). The response of soil faunal density was negatively correlated with soil pH (Fig. 8b), while an opposite response was observed for Gram-negative bacterial-PLFAs (Fig. 8l).

Fig. 8.

Response ratios of soil faunal density and PLFAs related to the initial soil C/N ratio and pH. (a, b) Faunal density, (c, d) total-PLFAs, (e, f) fungal-PLFAs, (g, h) bacterial-PLFAs, (i, j) Gram-positive bacterial-PLFAs, and (k, l) Gram-negative bacterial-PLFAs. The fitted lines represent slopes from linear mixed effects, with 95% confidence intervals (shaded) shown in each panel. The sizes of the circles represent the relative weights by the weighting factors for each paired observation. The sample sizes, adjusted R², and P values are shown in each panel for litter addition and removal separately. The dotted lines are shown at effect sizes of zero. PLFAs: phospholipid fatty acids, lnR: response ratio.

Fig. 9.

Response ratios of soil faunal density and PLFAs related to the litter C/N ratio. (a) Faunal density, (b) total-PLFAs, (c) fungal-PLFAs, (d) bacterial-PLFAs, (e) Gram-positive bacterial-PLFAs, and (f) Gram-negative bacterial-PLFAs. The fitted lines represent slopes from linear mixed effects, with 95% confidence intervals (shaded) shown in each panel. The sizes of the circles represent the relative weights by the weighting factors for each paired observation. The sample sizes, adjusted R², and P values are shown in each panel for litter addition and removal separately. The dotted lines are shown at effect sizes of zero. PLFAs: phospholipid fatty acids, lnR: response ratio.

4. Discussion

Plant litter input provides direct and original carbon and nutrient resources for soil fauna and microorganisms through a complex soil food web. Our global synthesis found that soil faunal density changed marginally after litter addition but decreased significantly after litter removal (Fig. 10). The nonsignificant influence of litter addition on soil faunal density suggested that natural litter input is not a limiting resource for soil fauna in terrestrial ecosystems [38], [39], while this response decreased with increasing elevation, suggesting that a warmer environment at lower elevations supports higher density and activity of soil fauna [40], [41]. We also found that soil pH, soil C/N ratio, and litter C/N ratio represented the primary effects on the response of soil faunal density to litter addition, with negative correlations observed between the response of soil faunal density and both the soil pH and C/N ratio (Fig. 8a and b). Moreover, soil fauna density responded more to litter with higher C/N ratios, which was consistent with a previous study [42]. One possible explanation is that soil fauna is incorporated in the early stage of decomposition of low-quality litter with a high C/N ratio and low initial nitrogen availability. However, the decrease in nutrient release from plant litter could be responsible for the reduction in soil faunal density after litter removal.

Fig. 10.

Schematic diagram illustrating soil organisms in response to litter addition (left) and removal (right) over forests and grasslands. The numbers represent the below and above confidence intervals for weighted percentage changes in soil faunal density, and total-, fungal-, bacterial-, Gram-positive bacterial-, and Gram-negative bacterial-PLFAs under litter addition and removal. The blue numbers indicate significant positive effects, the red numbers indicate significant negative effects, and the black numbers indicate non-significant changes after litter addition or removal. PLFAs: phospholipid fatty acids.

Microbial activity benefits from increasing litter inputs [43], [44], resulting in a significant increase in soil microbial-PLFAs after litter addition. This growth is due to the external organic substrates that provide energy-rich, labile C sources for the microbial community [45]. Consistent with previous studies [46], [47], our study demonstrated that the response of soil total-PLFAs was more sensitive to the input of root litter with higher C/N ratios (Figs. 9b and S2), but this effect varied greatly between different microbial communities. For example, litter addition significantly increased fungal-, bacterial-, and Gram-negative bacterial-PLFAs, but had little effect on Gram-positive bacterial-PLFAs (Fig. 10), suggesting a diverse preference of microorganisms for labile C availability [48], [49], [50]. In most terrestrial ecosystems, fungal taxa and Gram-negative bacteria are more inclined to organic matter with low nutrition and high C/N ratio [51], [52], showing rapid growth rates and higher biomass C/N ratios when liable carbon source decreased [53]. Increases in C and N availability due to litter input significantly increased the fungal- and Gram-negative bacterial-PLFAs, and this effect was stronger with the decreasing initial soil C/N ratio (Fig. 8e and k), as fungi and Gram-negative bacteria are considered to be N-limited with an increasing initial soil C/N ratio [54]. Different preferences among different litter types also support this idea [55], [56]. The soil bacterial-PLFAs likely preferred foliar litter, which has a lower C/N ratio than root litter (Fig. S2). In contrast, lower-quality litter with more recalcitrant components contributed more to the growth of soil fungal-PLFAs. The response of fungal-PLFAs was positively associated with the litter C/N ratio (Fig. 9c), supporting the notion that fungi often have lower nutrient requirements than bacteria and have higher carbon use efficiency on low-quality substrates [57], [58].

Environmental factors (i.e., elevation, MAT, and MAP) and other driving factors (i.e., ecosystem types and experimental duration) may influence the responses of soil organisms to litter addition and removal. The different correlations between bacterial- and Gram-negative bacterial-PLFAs and litter addition with climate (MAT and MAP) may be partially explained by the differential frequency of studies in humid and arid regions (Fig. 7k, l, q, and r) because most experiments were conducted in humid areas (Table S1). Adding litter under warmer and wetter conditions may facilitate the growth of thermophilic and moisture-resistant bacteria [59], while Gram-negative bacteria are more abundant in soils with lower temperatures and moisture [60]. Moreover, the microbial response to altered litter input was greater in grasslands than in forests (Fig. 4). One possible reason is that grasslands are subjected to substantial human management (i.e., grazing or mowing), which resultes in greater carbon inputs from roots than aboveground litter [22]. Another reason is that the decomposition rate of litter was generally higher in grasslands than in forests [61]. A higher decomposition rate results in a faster rate of nutrient release, which can attract a large number of soil microorganisms to settle and flourish on substrate materials. Furthermore, soil microorganisms increased more significantly under long-term litter addition (Fig. 6), resulting from continuous carbon and nutrient release from decomposing litter to microorganisms over time.

Although meta-analysis provides a statistical method to quantify the mean effect sizes across different categories, the overall effect of litter addition or removal on soil faunal density and microbial biomass is synthesized with some uncertainties. First, the database is mainly compiled from temperate forests and grasslands in the Northern Hemisphere, and data collected from other regions have not been reported. Therefore, the lack of sufficient results outside of the studied areas (particularly in some arid regions) may produce bias in this assessment. Second, root litter addition experiments received less attention due to the difficulty and high cost of performing root litter experiments. Thus, few observations related to root litter addition were available. Third, the effect of litter removal on soil microbial community was not determined to be sufficient in this study due to the short duration of most experiments (Table S1). Previous studies [18], [62] have found that the soil microbial biomass was reduced significantly only after 7 years of litter removal, so the lack of sufficient studies from long-term experiments may affect our evaluation. Therefore, future studies should further consider the long-term effects of root litter inputs on soil biota, and comprehensive analyses involving multiple cascading factors should be incorporated to provide a full understanding of the fate of soil organisms under litter input and to provide a reference for the prediction of belowground community dynamics in response to climate change.

5. Conclusion

Litter addition significantly increased soil PLFAs, but this effect decreased with increasing initial soil C/N ratio. However, litter removal resulted in decreases in soil faunal density and microbial biomass. The development of fungal-, Gram-positive bacterial-, and Gram-negative bacterial-PLFAs were dependent on root litter input, whereas bacterial-PLFAs were more dependent on foliar litter input, resulting from the difference in the initial litter C/N ratio. The responses of soil faunal density and microbial biomass to altered litter input were greater in grasslands than in forests, and this effect was independent of site aridity. Long-term litter addition had a more significant effect on soil microbial biomass, but the effect of litter removal was independent of experimental duration. These results suggest that the increased labile C availability associated with litter inputs is beneficial to microbial activity under global greening, which is thought to be conducive to nutrient cycling in terrestrial ecosystems. Our study provides evidence that changes in foliar and root litter input have various effects on the structure and stability of soil microbial communities.

Declaration of competing interests

The authors declare that they have no conflicts of interests in this work.

Biographies

Xinyu Wei a Ph.D. candidate in School of Geographical Sciences, Fujian Normal University. She obtained her degree of Master of Science of Silviculture from Sichuan Agricultural University in 2020. Her research interests focus on the formation and stabilization of soil organic matter in the forest ecosystem.

Xiangyin Ni received his Ph.D. in Sichuan Agricultural University in 2017, and currently is a Professor at School of Geographical Sciences, Fujian Normal University. His-research interests include soil organic matter formation and biogeochemical cycling in forests, focusing on the formation processes and stabilization mechanisms of soil organic matter from plant residues and microbial residues. His academic research has resulted in publications in PNAS, Global Change Biology, Soil Biology and Biochemistry, etc. His researches have been included in the Sixth Assessment Report released by the IPCC.