Stoichiometric linkages between plant litter, trophic interactions and nitrogen mineralization across the litter–soil interface

Soil Biology & Biochemistry 92 (2016) 102e110

Contents lists available at ScienceDirect

Soil Biology & Biochemistry journal homepage: www.elsevier.com/locate/soilbio

Stoichiometric linkages between plant litter, trophic interactions and nitrogen mineralization across the litteresoil interface Yolima Carrillo a, *, Becky A. Ball b, Marirosa Molina c a

Hawkesbury Institute for the Environment, University of Western Sydney, Penrith, NSW 2751, Australia School of Mathematical and Natural Sciences, Arizona State University at the West Campus, Glendale, AZ 85306, USA c US Environmental Protection Agency, 960 College Station Rd, Athens, GA 30605, USA b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 14 April 2015 Received in revised form 24 September 2015 Accepted 4 October 2015 Available online 16 October 2015

The common notion for describing N mineralization in models is that it results from decomposer organisms trying to meet their stoichiometric demands based on their own elemental composition and that of the resource. However, in addition to influencing C and nutrient availability, plant litter also influences the composition of both the litter and mineral soil community e importantly not in the same manner e resulting in altered trophic interactions. Since decomposer groups and their consumers vary in their elemental composition and demands, a change in composition and abundance of soil functional groups may result in a change in the stoichiometry of the whole soil food web, thus altering their stoichiometric relations with the available resource with potential functional consequences. We use experimental data and quantitative food web modeling to investigate the impact of the changes in the litter and soil food webs brought about by the differing stoichiometry of plant litter on (a) N mineralization, (b) the contribution of different functional groups to mineralization, and (c) the stoichiometric flexibility of the system, assessed as the ability to mineralize materials with different stoichiometry. Our simulations suggested that the effects of litter stoichiometry on trophic interactions, their impacts on N mineralization and the relative contribution of functional groups may not behave as a continuum across the litter and soil interface. Further, changes in food webs associated with variation in plant stoichiometric traits can influence the relative importance of functional groups, which given their particular stoichiometric demands may affect ecosystem-level N cycling. Our results suggested that litter materials of intermediate N contents, or litter mixtures encompassing materials with different nutrient contents and thus resulting in mixtures of intermediate stoichiometry, may promote food webs that are better suited to deal with changing substrate stoichiometry. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Soil food web model Litter quality Soil community Nitrogen Mineralization Stoichiometry Trophic interactions

1. Introduction As temporal and spatial scales decrease, the importance of resource chemical composition and the decomposer community as regulators of decomposition processes increases (Lavelle et al., 1993; Adair et al., 2008; Wall et al., 2008). The effects of the interactions of these two factors are not well understood. Strong relationships between the stoichiometric composition of substrates and their decomposition rates have been demonstrated (Meentemeyer, 1978; Aerts, 1997; Manzoni et al., 2008) and result from the degree of stoichiometric imbalance between decomposers

* Corresponding author. Tel.: þ61 0468 329 048. E-mail address: [email protected] (Y. Carrillo). http://dx.doi.org/10.1016/j.soilbio.2015.10.001 0038-0717/© 2015 Elsevier Ltd. All rights reserved.

and their resources (Hessen et al., 2004; Manzoni et al., 2008). The impact of the soil community composition on decomposition and €la € et al., 1991; mineralization rates is also amply supported (Seta Bradford et al., 2002; Heemsbergen et al., 2004; Carrillo et al., 2011). Experimental evidence of interactive effects between plant litter elemental composition and the structure of the soil community is scarce, however, these interactions may be responsible for unexplained variability observed when exploring the general patterns of decomposition and elemental composition (Parton et al., 2007; Manzoni et al., 2008; Ågren et al., 2013). For example, Carrillo et al. (2011) experimentally demonstrated that the relationship between litter nitrogen (N) and phosphorus content with N mineralization was dependent on the structure of the soil community. Similarly, Buchkowski et al. (2014) demonstrated that the role of the microbial biomass in controlling decomposition was

Y. Carrillo et al. / Soil Biology & Biochemistry 92 (2016) 102e110

determined by the relation of microbial and available resource stoichiometries. Understanding these types of interactions is key to increasing our ability to predict nutrient dynamics and decomposition in the face of changes in plant and soil community composition. In order to understand substrate qualityecommunity interactions, however, separately considering the litter and the soil layer is important, yet very rarely done. It is becoming increasingly clear that a disconnect between the regulation of processes in the litter layer and in the underlying soil exists (Parton et al., 2007; Adair et al., 2008). Recent studies are challenging the notion of the litteresoil continuum, and suggesting that how carbon (C) and nutrient availability regulate decomposer function and vice versa may vary for the litter and soil layer (Fanin et al., 2012; Ball et al., 2014; Mooshammer et al., 2014a). The common notion used for describing N mineralization in C and N models is that it results from decomposer organisms trying to meet their stoichiometric demands based on their own elemental composition and that of the resource (Manzoni and Porporato, 2009). More recent advances consider various microbial physiological mechanisms to deal with the stoichiometric imbalances (Mooshammer et al., 2014b). For the most part the decomposer organisms are described as acting in isolation or as a single group, but not functioning as a community comprised of various functional groups. However, in addition to determining C and nutrient availability, plant litter elemental composition also influences the composition of both the litter and mineral soil community, including microbes and fauna. Such changes in composition result in altered community interactions, including competitive, facilitative, and trophic interactions, which in turn, could affect the outcome of the decomposition process (Wardle, 2002; Kaiser et al., 2014). In this study, we focus on trophic interactions in the soil and litter food webs as major drivers of C and nutrient mineralization (Moore et al., 2003; Bardgett and Wardle, 2010). C mineralization is affected by the grazing-driven turnover activity and respiration of consumed populations (Bardgett et al., 1993; Mikola and Setala, 1998), while N mineralization occurs mainly due to excretion of excess N (Bardgett and Chan, 1999; Bonkowski, 2004). In addition, top-down control of microbial decomposer populations by members of the soil fauna regulates decomposer demand for C and nutrients, and subsequently, their role as consumers and prey (Hedlund and Ohrn, 2000). While trophic interactions are an important component of the role of the soil community in the decomposition process, they have been given little attention when trying to understand the regulation of decay by resource chemical composition. Soil and litter communities and thus their food web structures are very responsive to changes in resource chemistry (Carrillo et al., 2012) with potential impacts on function. The functional consequences of the changes in food web structure associated with changes in plant litter chemistry may be mediated by its impacts on the stoichiometric relations among food web members and the detritus available. These relations may vary for litter and soil, given the dramatic difference in C and nutrient stoichiometry in these two environments (Cleveland and Liptzin, 2007). It has been suggested that the result of trophic transfers between the members of the soil food web can depend on the quality of resources (Herlitzius, 1983; Wardle, 2002; Bardgett, 2005). In one of the few studies to address this question, Hanlon (1981) found that the influence on fungal respiration exerted by a collembolan grazer depended on the nutrient concentration of the growth medium. Such dependency may result from the fact that decomposer groups and their consumers vary in their elemental composition and demands and thus their stoichiometric relations depend on the trophic groups involved and the particular C and nutrient sources available (Hessen et al., 2004; Fanin et al., 2013; Kaiser et al., 2014). Thus, a

103

change in composition and abundance of soil functional groups as a response to variation in substrate may result in a change in the stoichiometry of the whole soil food web with potential functional consequences. For instance, the resulting food web could vary in its ability to process substrates of differing stoichiometry thus altering the system's stoichiometric flexibility, which is defined as the system's ability to modify the balance of elements while maintaining function (Sistla and Schimel, 2012). The inherently complex nature of these interactions makes them difficult to assess via experimental means as this would require isolating the effect of the soil populations from that of litter quality, making modeling approaches not just desirable but necessary. Organism-oriented models, which explicitly incorporate soil organisms and their interactions with the biophysical environment, have a high explanatory value and permit the evaluation of the effects of intervention and management (Paustian, 1994; Smith et al., 1998). The quantitative food-web modeling approach initiated by Hunt et al. (1987) has been applied to several natural and agricultural systems and has proven useful in simulating C and N mineralization rates and in explaining rates in terms of the relative contribution of groups of organisms and particular trophic interactions (Hassink et al., 1994; de Ruiter et al., 1994; Berg et al., 2001; Schroter et al., 2003; Bezemer et al., 2010). In this approach food webs are constructed by aggregating species into functional groups and N and C cycling are analyzed in relation to the structure and functioning of the food webs. We used this approach to simulate N mineralization from surface applied litter of differing N contents and from mineral soil based on observed population sizes and the trophic interactions among the members of the soil and litter food webs. With simple modifications our model could further describe mineralization as being regulated not only by trophic interactions or litter elemental composition, but by the interplay of both, in both the litter and soil. In order to do this, we added simple features to describe differential composition of the litter and differential abilities of fungi and bacteria to degrade organic matter fractions. This approach allowed us to isolate the effects of the changes in soil food web structure prompted by the added litter from the effects of the quality of litter on trophic transfers. In a field experiment we assessed the abundances of soil and litter microbial detritivores, protozoa, nematodes and microarthropods and measured N mineralization over the course of six months after the surface application of six plant materials of contrasting chemical compositions (Carrillo et al., 2011; Ball et al., 2014). We calibrated the model using the measured soil populations and N mineralization after the application of one substrate. We then used the model to investigate the impact of the changes in the litter and soil food webs brought about by the differing stoichiometry of plant litter on: (a) N mineralization, (b) the contribution of different functional groups to mineralization, and (c) the stoichiometric flexibility of the system, assessed as the ability to mineralize materials with different stoichiometry. 2. Methods Net N mineralization rates and abundances of trophic groups in soil and litter from a field study in the Piedmont region of Georgia USA and previously published in Carrillo et al. (2011) and Ball et al. (2014) were used to calibrate and validate the trophic transfer N mineralization model. Briefly, the field approach involved using a 100-m2 site divided into 24 2  2 m aluminum flashing enclosed plots arranged in four blocks and cleared of vegetation and litter. Mineral soil from the top 5 cm of 6 separate 25  50 cm areas within these plots was removed, sieved, frozen (to kill fauna) and placed back in the field enclosed in 5 mm wire mesh. Before application of treatments, soils were left bare for 6 weeks to allow

104

Y. Carrillo et al. / Soil Biology & Biochemistry 92 (2016) 102e110

recolonization and dissipation of the flush of nutrients associated with disturbance. Sieving and freezing produced homogenized soil where we could then observed the impacts of litter types, however it should be taken into account that these procedures likely excluded some of the least disturbance/stress resistant organisms. Litter of five different types and their mixture were randomly assigned to areas within each block and surface applied to plots at a rate of 327 g m2 (6 litter types  4 plots ¼ 24 plots). Materials applied were air dried green litter of cereal rye (Secale cereale L.), air dried green litter of crimson clover (Trifolium incarnatum L.), air dried green litter of false indigo (Amorpha fruticosa L.), wheat straw (Titricum aestivum L.), pine needles (Pinus taeda L.), and an even mixture of these by weight. Initial C, N, phosphorus, lignin, cellulose and hemicellulose contents of the litter were assessed (Carrillo et al., 2011). Lignin, cellulose and hemicellulose were measured using sequential neutral detergent/acid detergent digestion (Van Soest, 1994). There was a wide range in the N contents of initial plant litter, so that C to N ratios ranged from 20 in the clover to 128 in the pine. Substrates were allowed to decompose for six months. Net N mineralization was determined via incubation of nylon bags containing exchange resins. Details and N mineralization rates are reported in Carrillo et al. (2011). Biomass of bacteria, fungi, protozoa, nematodes and microarthropods in soil and litter were estimated from abundances reported in Carrillo et al. (2011) and Ball et al. (2014). The concentration of total and some individual phospholipid fatty acids (PLFA; as mmoles per gram of soil or litter) were used to obtain estimates of biomass of microbial groups. Total microbial biomass in soil and litter was estimated using the conversion factor in Bailey et al. (2002) and a Kc factor of 0.32. The fatty acid 18:2u6 was used to estimate fungal biomass in soil using the conversion factor provided by Klamer and Baath (2004). The biomass of bacteria was estimated as the difference between total microbial biomass and fungal biomass. Fungal biomass in litter was estimated using the relative differences observed in the concentration of the fatty acid 18:2u6 in relation to its overall average concentration in the six litter materials studied. The relative deviation from the average (as percentage of the average) was applied to an assumed average ratio of fungi to bacteria in plant residues of 2 (Beare et al., 1990). Litter fungal and bacterial biomasses were estimated from total microbial biomass using the calculated ratio. While the abundance of 18:2u6 can be a good indicator of fungi, it should be treated with caution as it may be present in plant cells (Frostegard et al., 2011). A laboratory assay was conducted to obtain a conversion factor to estimate biomass of protozoa from the concentration of the phospholipid fatty acid biomarkers in soil and litter. Individual PLFA can be used as markers for protists as their cell membranes contain specific long-chain polyunsaturated fatty acids (White et al., 1996; Desvilettes et al., 1997). The fatty acids 20:2u6,9c; 20:3u6,9,12c; 20:4u6,9,12,15c have been used as indicators of protozoa abundance in soils (Cavigelli et al., 1995; Mauclaire et al., 2003). Protozoa from a composite soil sample from the field site were cultured for 2 days at 30  C in phosphate agar as substrate and Escherichia coli as a food source. Immediately after, the abundance of ciliates, amoebae, and flagellates in this suspension (as numbers per ml) was estimated in five replicates using a dilution method (most probable number, MPN) based on Singh (1946). Immediately after the subsampling for enumeration, known volumes of the original suspension (10, 20 and 40 ml in triplicate) were placed in Teflon bottles, centrifuged (24,000 RCF, 30 min) and the pellet frozen at 80  C. PLFAs were extracted from thawed samples using the same extraction procedure as for soil and litter samples. From the concentrations of cells in the initial suspension obtained with MPN, biomass C of each protozoan group was estimated using the conversion factors in Beare et al. (1992) and total protozoan biomass

per milliliter of suspension was calculated and extrapolated for the volumes used for PLFA extraction. A linear relationship was found between the concentration (mmoles/ml of suspension) of the fatty acid 20:4u6,9,12,15c and the calculated protozoan biomass in the three analyzed volumes of culture suspension (as mg C) (R2 ¼ 0.70; mg C protozoa ¼ 3.5525 þ 54,860  mmoles of 20:4u6,9,12,15c). This relationship was used to estimate protozoan biomass in soil and litter samples from the field experiment 20:4u6,9,12,15c fatty acid values. Estimates of protozoan biomass ranged between 0.5% and 1.9% of total microbial biomass estimated with PLFA. Numbers of nematodes and microarthropods per g of soil or litter were converted to mg of biomass C using the conversion factors in Beare et al. (1992). Obtained values of biomass for all microbial and faunal groups were converted to mg per square meter of soil down to 5 cm using a bulk density value of 1.1 g cm3. Litter biomasses per gram for litter were converted to biomass per m2 based on the initial rate of litter application and considering the amount of litter remaining in litter bags at the time of sampling. The trophic transfer model to obtain N mineralization rates from feeding interactions between functional groups in litter and soil was written using STELLA® 8.1 (iSee Systems). The field observed average biomasses of the major soil groups in mineral soil and litter were used as inputs to the model (average of three measurements for soil and one measurement for litter). Consequently, the simulated daily rate was considered the average rate for the study period. C flows between trophic groups are derived from feeding rates which are in turn split into an excretion rate, a biomass production rate and a mineralization rate. Feeding rates are calculated assuming that the biomass production rate of a group balances the rate at which material is being lost through natural death and predation. Feeding rate of a group on a prey or on a substrate and N mineralization rates are calculated as in de Ruiter et al. (1993). N flows occur in parallel and in proportion to C flows through the use of the C/N ratios of organisms and organic fractions. Specific death n rates were made temperature dependent using a Q10 ¼ 3 (Andre et al., 1990). Temperature was measured in the field every two hours throughout the sampling period and the overall average was used for simulations. Physiological parameters (C/N, assimilation, production and death rates) for the functional groups in the litter and soil food webs were taken from the literature and assumed to be the same for all simulations (Table 1). Two separate food webs were modeled, one for the litter and one for the soil. Organisms found in litter were assumed to only consume litter material and prey inhabiting the litter; organisms found in soil were assumed to only consume material and prey in the mineral soil. Nitrogen mineralized by the litter community

Table 1 Model parameters used for simulation of trophic transfers and resulting C and N mineralization. Parameters

Bacteria Fungi

Death rate (day1)a C/N Production efficiencyb Assimilation efficiencyb Fraction to labile poola Fraction to cellulose poola

0.0033

a b c

5c 0.3

Protozoa Nematodes Microarthropods

0.0033 0.02

0.006

0.005

10c 0.3

7a 0.4

10a 0.37

8a 0.4

1

1

0.95

0.43

0.5

0.8

0.8

0.8

0.8

0.8

0.2

0.2

0.2

0.2

0.2

From Hunt et al. (1987). From de Ruiter et al. (1993). Obtained through calibration.

Y. Carrillo et al. / Soil Biology & Biochemistry 92 (2016) 102e110

joins the mineral nitrogen pool in soil from where both litter and soil microbial groups can uptake it according to their demand. The soil community was modeled as a simplified food web composed of five functional groups: bacteria, fungi, protozoa, nematodes and microarthropods (Fig. 1). The differences in trophic transfers considered in litter and soil corresponded to observed differences in the composition of the major functional groups assessed. For example, soil microarthropods were mostly prostigmatids followed by oribatids. As oribatids are considered fungivores and small soil prostigmatids are known to be important in regulating populations of nematodes (Coleman et al., 2004), the microarthropods in the model consumed fungi and nematodes. The specific proportion found for each group was included in the model as a factor modifying the total consumption demand. Litter to be mineralized was divided into cellulosic/hemicellulosic material (measured), lignin (measured) and labile material (estimated by default). Soil organic matter was split into labile and resistant materials 1 and 2 (1 easier to degrade than 2). The proportions of labile and resistant fractions 1 and 2 in soil were defined to be 5%, 25% and 70% respectively. These fractions were based the composition of soil organic matter parameters in Paustian et al. (1992) application of the Century model (~3% labile and remaining 97% slow and passive pools) and further optimized via calibration for this model. The three fractions are processed by bacteria and fungi. It was assumed that for each unit of mass demanded and consumed by bacteria 94% was labile, 5.5% was cellulose (or Resistant 1) and 0.5% lignin (or Resistant 2). These percentages are based on the relative proportions among the decay constants of these C pools used to model mineralization in the

Labile

Lignin

CERES-N sub-model (Schomberg and Cabrera, 2001). The percentages were then modified to reflect known differences in substrate utilization of bacteria and fungi, specifically the greater ability of fungi to degrade resistant fractions (Romani et al., 2006). Thus, in the case of fungi, 89% of consumed material was labile, 9.9% was cellulosic and 1.1% was lignin. Each functional group contributes to the residue pools through death and waste (Table 1). For determining the C/N of the litter fractions it was defined that the C/ N of the lignin fraction is 2 times that of the labile fraction, the C/N of the cellulose fraction is 3 times that of the carbohydrate fraction, and the C/N of the total litter is the weighted average of its three fractions and was the average between values at the beginning and end of incubations. C/N ratio of the organic fractions in soil was calculated in the same manner. The C/N ratio of soil organic materials that are available to decomposers was assumed to be 30. N mineralized from litter is assumed to be leached into the soil and join the mineral N pool in soil, of which N mineralized in soil is also part. When the C/N of the prey or substrate is higher than that of the predator, then immobilization occurs and N is taken from the soil mineral N pool (Fig. 1). The model was calibrated using the observed population biomasses, C/N and average net mineralization rate for rye litter, because its C/N ratio falls in the middle of the C/N range. To evaluate the performance of the model, we compared observed average net mineralization rates with modeled rates obtained with the observed population biomasses and average C/N ratios of the other five litter materials. Model sensitivity of the newly added features was assessed by running the model with 10% increased/decreased parameter values.

Nematodes

Bacteria

Cellulose

105

Protozoa Fungi Microarthropods

Litter Mineral N in Soil

Soil

Labile Resistant 1 Resistant 2

Nematodes

Bacteria Protozoa Fungi

Microarthropods

Fig. 1. Functional groups and trophic interactions considered in the soilelitter food web model. In the interest of clarity, the arrows representing nitrogen mineralization are drawn for the whole soil and litter food webs although they are modeled for each individual feeding interaction. Nitrogen immobilization is also represented as a flow to the microbial biomass but it is modeled separately for bacteria and fungi. Nitrogen mineralized by the litter community joins the mineral nitrogen pool in soil from where both litter and soil microbial groups can uptake it according to their demand.

106

Y. Carrillo et al. / Soil Biology & Biochemistry 92 (2016) 102e110

3. Results

3.2. Model performance

3.1. Biomasses of soil and litter populations

The sensitivity of this food web modeling approach to variation in physiological parameters (production efficiency, assimilation efficiency, death rates) has been determined previously (Hunt et al., 1987; de Ruiter et al., 1994; Berg et al., 2001). We evaluated the impact of the two features we added to the model. One of them was the differential preference of fungi and bacteria for degrading the fractions of organic material. We defined that fungi consumed a greater proportion of the cellulose and lignin in litter and of the resistant fractions in soil than bacteria. To test for the impact of this assumption we ran simulations in which the opposite was defined. We found that assuming that fungi had a greater preference for hard-to-degrade materials led to greater mineralization rates. This assumption thus appeared reasonable can be explained if the differences in the C/N of fungi and bacteria are considered. Bacteria have a lower C/N ratio than fungi, which implies that their demand for nitrogen is higher. Thus when bacteria consume a greater proportion of resistant material, which has higher C/N, the net release of nitrogen will be lower. The second feature added was that the C/N ratio of the fractions was determined by the C/N ratio of the litter and the proportions of labile, cellulose and lignin. While these proportions were known for the litter, that was not the case for the mineral soil e as the true or even approximate composition of the functional fractions in mineral soil is not known. While, this may limit the power of the predictions in the mineral soil, the model was not very sensitive to changes in the proportions and it was the overall C/N ratio that mostly affected the predictions, as the greater the C/N of the overall material the greater the C/N of all fractions. An increase or decrease in the percentage of labile material of 20% determined a change of 11% in the mineralization rate while a 20% increase of decrease in C/ N lead to a change in 50% of the predicted rate. To assess the model performance, simulations were run using the observed population biomasses, measured C/N and C composition of the litter fractions as inputs. The simulated net mineralization rates were then compared with the average net values measured in the field (Fig. 2). The model was able to predict the general trend in observed net average mineralization rates with regards to C/N of substrates, indicating higher mineralization rates with lower C/N. Simulated values were well within the observed range of values. No consistent pattern of overestimation or underestimation was detected. The model correctly predicted the relative

Statistical comparisons of measured population abundances over the course of the field experiment have been previously reported (Carrillo et al., 2011; Ball et al., 2014). Here we present averages and estimations of mass (g C m2) for simulations (Table 2). Although the concentration of biomass in litter per unit of mass is much greater than in mineral soil, when considering the amount of soil down to 5 cm in depth, the majority of the living mass per unit of area (between 83 and 92%) was found in soil. Bacteria represented between 63% and 75% of all biomass (soil and litter included). Fungi comprised between 23% and 35% of all biomass. The biomass of litter populations varied with litter type. Total microbial biomass was greater in clover and Amorpha litter and was lowest in pine, wheat straw and rye. Fungi were more abundant than bacteria. Fungal biomass was negatively related to litter C/N (p ¼ 0.02, R2 ¼ 0.86) and litter C/P (p ¼ 0.045, R2 ¼ 0.82); the fungito-bacteria ratio was highest in rye, wheat straw and the mixture and lowest in clover, Amorpha and pine. Protozoan biomass was largest in clover and straw and lowest in pine, rye and Amorpha. The percentage of nematode biomass in litter was never above 0.01%. Nematodes were mostly bacterial feeders. The biomass of nematodes in pine litter was substantially lower than in other materials and the bulk of the mass corresponded to fungal feeders. The proportion of bacterial feeding nematodes was lowest in pine, rye and wheat straw. Microarthopod biomass in litter reached 1% of total biomass in the case of clover and rye and had the lowest values under pine, Amorpha and the mixture. Soil populations varied with litter type but differences were less pronounced than in litter. No correlations were detected between soil groups biomass and litter stoichiometry. Soil total microbial biomass was highest in soils under Amorpha, rye and the mixture and lowest under pine. In contrast to litter, bacteria were more abundant than fungi in soil. Fungi-to-bacteria ratios also varied with litter type. The lowest fungi-to-bacteria ratio occurred under pine and wheat straw. Protozoan biomass in soil was not very responsive to litter type. The biomasses of nematodes and microarthropods in soil were comparable and represented up to 0.1% of total soil biomass. Soil nematodes were mostly omnivorous and bacterivorous. Microarthropods were mostly prostigmatids and oribatids.

Table 2 Biomass of functional groups in litter and soil (mg C m2) used for simulation of trophic transfers and resulting C and N mineralization. Functional group

Litter substrate Amorpha

Clover

Mixture

Wheat straw

Rye

Pine

Litter bacteria Litter fungi Litter total microbial biomass Litter protozoa Litter nematodes Bacterial feeding litter nematodesa Litter microarthropods Soil bacteria Soil fungi Soil total microbial biomass Soil protozoa Soil nematodes Soil microarthropods Fungal feeding soil microarthropodsa

1489 2273 3771 (262) 53 (10) 0.172 (0.097) 0.1634 10.3 (1.6) 22,420 6822 (1620) 29,242 (5406) 282 (22) 23 (16) 20 (6) 6

2227 2061 4288 (402) 238 (29) 0.077 (0.009) 0.07161 46.3 (6.9) 16,897 5474 (1865) 22,371 (1739) 250 (14) 19 (6) 14 (2) 4.2

931 2356 3282 (468) 162 (28) 0.170 (0.071) 0.1598 13.4 (1.6) 18,288 5491 (1832) 23,779 (2314) 259 (19) 22 (8) 16 (3) 4.8

606 1843 2449 (371) 284 (53) 0.337 (0.195) 0.19209 20.1 (4.1) 18,080 4240 (1667) 22,320 (1893) 245 (9) 19 (4) 14 (7) 4.2

352 1850 2202 (476) 98 (23) 0.181 (0.03) 0.13575 27.5 (2.6) 17,026 7812 (1015) 24,838 (28) 277 (7) 29 (12) 22 (9) 6.6

1170 1532 2702 (56) 91 (5) 0.028 (0.004) 0.00448 3.1 (0.7) 16,701 3973 (1423) 20,674 (2212) 242 (2) 16 (5) 15 (2) 4.5

Standard errors in parentheses. Standard errors in litter from four samples taken at the end of the sampling period. Standard errors in soil correspond to three sampling dates averages obtained from four replicates for each date. Values without a standard error are estimates (see food web modeling methods). a Percentages of bacterial feeding nematodes in litter and fungal feeding microarthropods in soil were calculated from the average abundances of feeding groups and mites orders respectively.

Y. Carrillo et al. / Soil Biology & Biochemistry 92 (2016) 102e110

14 9 C/N=80.9

4 -1

Amorpha Clover

Mixture

Rye

Wheat straw

C/N=114.1

Pine*

Fig. 2. Observed and simulated net average nitrogen mineralization rates over the course of six months of decomposition. Bars indicate standard error. *Observed value of pine (0.01 mg m2 day1, i.e. net immobilization) too small to appear in figure.

magnitudes of mineralization rates of Amorpha and clover (clover higher than Amorpha, despite the lower C/N of Amorpha). The greater mineralization under clover reflected the effect of the larger population of bacteria in clover than in Amorpha litter, which would have led to greater consumption of substrate and as a result, greater release of mineral N. Clover litter also had a considerably higher population of protozoa which, by preying on bacteria, further contributed to more net release. Consistent with observed values, the model produced low and similar values for pine and wheat straw, even though the C/N of pine is substantially higher than that of straw. Again, simulated rates reflect the effect of the composition of the communities, as the greater populations of bacteria observed in pine litter would have generated a greater release of N. The prediction of the mineralization rate of the mixture was very close to the observed rate. The prediction of mineralization rate under rye was also very close to the observed but this was expected as this was the substrate used for model calibration. In summary, the model behaved reasonably well in predicting net mineralization rates and it was able to describe the effects of populations' sizes and predation pressures on the N release from litter and soil. 3.3. Impact of changes in the soil and litter food webs and contribution of functional groups to mineralization To isolate the effect of the structure of the food webs generated by different litter types, we ran model simulations in which the C/N

10

4

5

2

0

0

Pine

6

Rye

15

Wheat straw

8

Mix

20

Clover

a

We assessed the ability of the various food webs (soil and litter combined) generated under different litter chemistry to mineralize materials with different stoichiometry, as an indication of changes in the system's stoichiometric flexibility. For this, we ran simulations in which the field communities observed under each litter type were set to degrade substrates of varying hypothetical C/N in addition to the C/N of the litter actually applied in the field. This was done using standard model settings and field functional group biomasses in combination with C/N ratios ranging from 10 to 100. To visualize how the different communities were able to mineralize the different hypothetical substrates we plotted observed field mineralization rates for each litter type against simulated mineralization rates under the varying C/N values. This indicated that not all communities were equally suited to mineralize all substrates. Close matching (positive relationships) between observed and simulated rates only existed when the available litter's C/N was low (40 or below) (Fig. 4). For the litters with high C/N, modeled

10

Soil N mineralization

Amorpha

mg mineral N m -2

25

3.4. Impact of food web changes on stoichiometric flexibility

b

Litter N mineralization FAUNA

FUNGI BACTERIA

Pine

C/N=61.1

19

Rye

mg inorganic N m -2

simulated

C/N=40.0

24

Wheat straw

observed

C/N=18.7

Amorpha

29

and C compositions of litter and soil organic matter were held fixed at an intermediate value (C/N ¼ 30 as the sole input for litter quality and different food web structures observed in the field under the different litter treatments as input for the soil and litter communities). In this way mineralization was a function solely of the different community structures in litter and soil associated with each litter type. Simulated total N mineralization showed stronger responses to different treatments in litter than in soil (Fig. 3). The maximum difference between litter treatments in soil was 30%, while in litter it was 108%. The contribution of microbial and fauna groups varied between litter and soil. There was greater variation in the contribution of bacteria, fungi and particularly fauna across litter types in litter than in soil. This is consistent with the greater response of litter populations to litter type. Overall, the contribution of faunal groups to mineralization, relative to microbes, was greater in litter than in soil (with the exception of Amorpha). The contribution of groups also varied with litter type. To evaluate whether the contribution of individual groups was related to the stoichiometry of the litter in which their populations were generated, we assessed the correlation between litter C/N and the contribution to mineralization of each group to litter and soil mineralization. In most cases, the contribution of functional groups was not related to the C/N ratio of the litter material (data not shown). However, the contribution of fungi to mineralization in litter decreased strongly as the C/N of the litter increased (p ¼ 0.007; R2 ¼ 0.87).

Mix

C/N=20.5

Clover

34

107

Fig. 3. Simulated contribution to total net nitrogen mineralization by bacteria, fungi and fauna in mineral soil (a) and litter (b). Simulation used a fixed litter C/N and C composition and the different soil and litter food webs for different litter types. Note different y-axis scales.

108

Y. Carrillo et al. / Soil Biology & Biochemistry 92 (2016) 102e110 55

C/N= 10; R2= 0.28 C/N= 20; R2=0.58

A

35

Actual C/N; R2=0.92

A A

25

A

C/N= 40; R2=0.47

A

C/N= 80; R2=0.01

15 A

5

C/N= 100; R2=0.00

-5

0

5

10

15

20

Clover

Mixture Amorpha

-25

Rye

-15

Wheat straw

-5

Pine

Modeled N mineralization (mg m-2 )

45

Food web 25

30

Observed N mineralization (mg m-2 ) Fig. 4. Modeled N mineralization rates of hypothetical substrates with different C/N ratios by field assessed soil and litter food webs generated under six plant litter materials. Values are plotted against the observed mineralization rates of that species in the field. Actual C/N is the value of the material under which soil food webs were generated. Only lines for significant relationships are plotted. Letter “A” indicates values produced by the communities generated under Amorpha.

mineralization by the communities generated under the high C/N litters was substantially lower than actual mineralization. 4. Discussion 4.1. Impact of changes in the soil and litter food webs and contribution of functional groups to mineralization The greater variation in mineralization rates simulated in litter compared to soil is consistent with the response of whole food web biomass, indicating the overall differences in the abundance of functional groups in mineral soil and in litter translated into differences in rates. Moreover, the contrast in responses of litter and soil did not only manifest in the magnitude of the variation among litter types, as the patterns of response of N mineralization rates were also not correlated between litter and soil. This reflects the fact that litter type did not affect the litter and soil communities in the same manner and thus the effect of populations and their trophic interactions on mineralization should not be expected to be correlated. While litter communities are directly affected by the attributes of litter, as it constitutes their source of energy and nutrients as well as their habitat, the impact on the soil community is less direct and is mediated by processes of redistribution of the resource by fragmentation, leaching or translocation into the mineral soil (Heal et al., 1997; Coleman et al., 2004). This was evidenced in the lack of relationships between litter stoichiometry and functional groups biomasses in soil contrasting the significant relationships of litter C/N and C/P with litter fungi. We speculate that the stoichiometry of the litter-derived materials, once processed and incorporated into soil, would show relationships with the soil population's biomasses. Thus stoichiometric driving in the soil still occurs but would be preceded by the processes involved in litter mixing and incorporation into the soil. However, it is also plausible that because of the smaller stoichiometric imbalance between resource and microbes in soil than in litter, C limitation and access to it would be more important than nutrient stoichiometry in controlling the responses of soil groups. Our modeling results reinforce the notion, as suggested by previous experimental data (Ball et al., 2014), that litter and soil do not behave as a continuum in regards to the impacts of litter on communities and mineralization processes.

The greater proportional contribution of faunal groups to mineralization in litter than in soil, was partly a result of a greater representation of fauna in the litter food webs compared to those in soil, but importantly it was also due to its effect on predation interactions. The influence of predation pressure on mineralization activity can be evidenced in the total mineralization rates in litter which were closely related to the biomass of protozoa, the consumers of bacteria (R2 ¼ 0.89, p ¼ 0.004) and microarthropods, consumers of fungi (R2 ¼ 0.46, p ¼ 0.017), but not significantly related to the biomass of bacteria, or total food web biomass. The strong influence of the variation in protozoan biomass can also be evidenced in the larger variation of the contribution of bacteria (Fig. 3) relative to that of the fungi. Because of bacterial higher N demands, due to their stoichiometry, enhanced predation pressure result in larger impacts per unit of biomass relative to fungi. Thus, the functioning of the litter food web indicated that the responses of the fauna groups and their top-down control of decomposer N demand were strong drivers of the release of N by the whole food web. In soil, in contrast, total mineralization was positively related to the biomass of both decomposers and fauna groups (p < 0.05), indicating a smaller degree of top-down control. These results are influenced by the steady-state assumption of the model, that groups consume as much as is needed to compensate for losses due to death and predation. This assumption determines that a larger biomass of the predator will result in greater demand by the prey and thus prompt more mineralization. Although this assumption may not always hold true, there is evidence of compensatory responses to predation in soil food webs (Hedlund and Augustsson, 1995), which together with the population sizes observed under different litter types and the predictions of the model, suggests topdown control of decomposer activity and impact on N mineralization is considerable and may be more important in the litter than in the soil food web. That the contribution of fungi to mineralization in litter was negatively related to litter C/N suggest that fungi were less important drivers of mineralization in food webs generated under recalcitrant materials. This was associated with a negative relationship between fungi populations and C/N and C/P suggesting that their decreased role was due to the negative impact of low nutrient contents on their populations. Thus these results suggest that changes in food webs invoked by plant stoichiometric traits

Y. Carrillo et al. / Soil Biology & Biochemistry 92 (2016) 102e110

can shape the relative importance of functional groups, which given their particular stoichiometric traits may affect ecosystemlevel N cycling. 4.2. Impact of food web changes on stoichiometric flexibility Assessing the ability of the various food webs generated under different litter chemistry to mineralize materials with different stoichiometry indicated that the food webs generated under litter types of intermediate quality were better suited to handle changes in substrate chemistry. When degrading litter with high C/N (80 and 100), the food webs generated in the low C/N litters (clover, Amorpha and the mixture), produced substantially underestimated rates, suggesting that these food webs lacked the ability to maintain function when processing low N materials. On the other hand, the food webs generated under pine and wheat straw, the materials with highest C/N, produced the largest overestimations of most materials, particularly those with low C/N. Together, these results suggest that when processing litters varying in C/N the food webs produced by substrates of intermediate C/N were better able to generate rates that were similar to those produced under their original substrate. In other words, the attributes of the food webs brought about by intermediate litter C and N stoichiometry, i.e. the specific representation of groups with their specific stoichiometric demands, granted more flexibility to mineralize N from new substrates of contrasting stoichiometry. While we focused on N mineralization this greater flexibility would be expected to impact the mineralization of C, with further impacts on ecosystem C cycling. Litter stoichiometry dissimilarity has been recently suggested as a main driver of the widely observed but not fully understood ‘home field advantage’ phenomenon (Veen et al., 2015), whereby plant litter decomposition occurs faster when taking place near the source of the litter. Our findings may further suggest that the role of stoichiometry may be mediated by its impacts on and interaction with food web structure, including added or reduced stoichiometric flexibility. Additional responses, not involving changes in food web structure, are likely involved in the functional response of soil communities to litter stoichiometry. For instance, changes in microbial exoenzyme production (Mooshammer et al., 2012) or in C or nutrient use efficiency (Mooshammer et al., 2014a) are suspected to enable microbes to cope with the large variability in litter stoichiometry and nutrient availability. Our success at simulating N mineralization from multiple substrates based on whole food web trophic structure and interactions suggests they are an important factor that likely acts in combination with enzymatic and metabolic shifts. This poses the question of the relative importance of these mechanisms in shaping how stoichiometrically flexible soil communities can be. Our simulations combining observed litter and mineral soil food webs, and variable litter composition suggested that: (a) the effects of litter stoichiometry on trophic interactions, their impacts on N mineralization and the relative contribution of functional groups may not behave as a continuum across the litter and soil interface and thus both should be taken into account, (b) changes in food webs associated with variation in plant stoichiometric traits can influence the relative importance of functional groups, which given the way they vary in their particular stoichiometric demands may affect ecosystem-level N cycling, and (c) the responses of the soil and litter food webs invoked by changes in litter stoichiometry may constitute a mechanism controlling the stoichiometric flexibility of a system. In particular, our results suggested that litter materials of intermediate N contents, or litter mixtures encompassing materials with different nutrient contents and thus resulting in mixtures of intermediate stoichiometry, may host food webs that are better

109

suited to deal with changing substrate stoichiometry. A related implication is that changes in the structure of the soil and litter food web, for example due to management or environmental shifts, could be more critical for altering N mineralization when the available litter is of particularly high or low C/N.

Acknowledgments This research was supported by USDA-Southern Region SARE (Project number GS05-044) and National Science Foundation grants to the Coweeta LTER Program (DEB-9632854 and DEB0218001). We thank Kathryn Mitchell, Sunny Shanks, Ashley Johnson, Jimmy Blackmon, Ryan Malloy and Seth Wenger for lab and field assistance. We also thank Carl Jordan, Mark Hunter, David Coleman, Miguel Cabrera and Mark Bradford for their input. This paper has been reviewed in accordance with the USEPA's peer and administrative review policies and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use by the USEPA.

References Adair, E.C., Parton, W.J., Del Grosso, S.J., Silver, W.L., Harmon, M.E., Hall, S.A., Burke, I.C., Hart, S.C., 2008. Simple three-pool model accurately describes patterns of long-term litter decomposition in diverse climates. Global Change Biology 14, 2636e2660. Aerts, R., 1997. Climate, leaf litter chemistry and leaf litter decomposition in terrestrial ecosystems: a triangular relationship. Oikos 79, 439e449. Ågren, G.I., Hyvonen, R., Berglund, S.L., Hobbie, S.E., 2013. Estimating the critical N:C from litter decomposition data and its relation to soil organic matter stoichiometry. Soil Biology & Biochemistry 67, 312e318. n, O., Lindberg, T., Bostro € m, U., Clarholm, M., Hansson, A.-C., Johansson, G., Andre Lagerlof, J., Paustian, K., Persson, J., Pettersson, R., Schnürer, J., Sohlenius, B., n, O., Wivstad, M., 1990. Organic carbon and nitrogen inflows. In: Andre Lindberg, T., Paustian, K., Rosswall, T. (Eds.), Ecology of Arable Land-organisms, Carbon and Nitrogen Cycling. Munsksgaard, Conpenhagen, pp. 85e126. Bailey, V.L., Peacock, A.D., Smith, J.L., Bolton, H., 2002. Relationships between soil microbial biomass determined by chloroform fumigation-extraction, substrateinduced respiration, and phospholipid fatty acid analysis. Soil Biology & Biochemistry 34, 1385e1389. Ball, B.A., Carrillo, Y., Molina, M., 2014. The influence of litter composition across the litter-soil interface on mass loss, nitrogen dynamics and the decomposer community. Soil Biology & Biochemistry 69, 71e82. Bardgett, R.D., 2005. The Biology of Soil: a Community and Ecosystem Approach. Oxford University Press, New York, 242 pp. Bardgett, R.D., Chan, K.F., 1999. Experimental evidence that soil fauna enhance nutrient mineralization and plant nutrient uptake in montane grassland ecosystems. Soil Biology & Biochemistry 31, 1007e1014. Bardgett, R.D., Wardle, D.A., 2010. Aboveground-belowground Linkages. Biotic Interactions, Ecosystem Processes, and Global Change. Oxford University Press, New York, 301 pp. Bardgett, R.D., Whittaker, J.B., Frankland, J.C., 1993. The effect of collembolan grazing on fungal activity in differently managed upland pastures e a microcosm study. Biology and Fertility of Soils 16, 255e262. Beare, M.H., Neely, C.L., Coleman, D.C., Hargrove, W.L., 1990. A substrate-induced respiration (Sir) method for measurement of fungal and bacterial biomass on plant residues. Soil Biology & Biochemistry 22, 585e594. Beare, M.H., Parmelee, R.W., Hendrix, P.F., Cheng, W.X., Coleman, D.C., Crossley, D.A., 1992. Microbial and faunal interactions and effects on litter nitrogen and decomposition in agroecosystems. Ecological Monographs 62, 569e591. Berg, M., de Ruiter, P., Didden, W., Janssen, M., Schouten, T., Verhoef, H., 2001. Community food web, decomposition and nitrogen mineralisation in a stratified Scots pine forest soil. Oikos 94, 130e142. Bezemer, T.M., Fountain, M.T., Barea, J.M., Christensen, S., Dekker, S.C., Duyts, H., van Hal, R., Harvey, J.A., Hedlund, K., Maraun, M., Mikola, J., Mladenov, A.G., Robin, C., de Ruiter, P.C., Scheu, S., Setala, H., Smilauer, P., van der Putten, W.H., 2010. Divergent composition but similar function of soil food webs of individual plants: plant species and community effects. Ecology 91, 3027e3036. Bonkowski, M., 2004. Protozoa and plant growth: the microbial loop in soil revisited. New Phytologist 162, 617e631. Bradford, M.A., Tordoff, G.M., Eggers, T., Jones, T.H., Newington, J.E., 2002. Microbiota, fauna, and mesh size interactions in litter decomposition. Oikos 99, 317e323. Buchkowski, R.W., Schmitz, O.J., Bradford, M.A., 2014. Microbial stoichiometry overrides biomass as a regulator of soil carbon and nitrogen cycling. Ecology 96, 1139e1149.

110

Y. Carrillo et al. / Soil Biology & Biochemistry 92 (2016) 102e110

Carrillo, Y., Ball, B.A., Bradford, M.A., Jordan, C.F., Molina, M., 2011. Soil fauna alter the effects of litter composition on nitrogen cycling in a mineral soil. Soil Biology & Biochemistry 43, 1440e1449. Carrillo, Y., Ball, B.A., Strickland, M.S., Bradford, M.A., 2012. Legacies of plant litter on carbon and nitrogen dynamics and the role of the soil community. Pedobiologia 55, 185e192. Cavigelli, M.A., Robertson, G.P., Klug, M.L., 1995. Fatty acid methyl ester (FAME) profiles as measures of soil microbial community structure. Plant and Soil 170, 99e113. Cleveland, C.C., Liptzin, D., 2007. C:N:P stoichiometry in soil: is there a “Redfield ratio” for the microbial biomass? Biogeochemistry 85, 235e252. Coleman, D.C., Crossley Jr., D.A., Hendrix, P.F., 2004. Fundamentals of Soil Ecology, second ed. Elsevier Publishing, San Diego, CA. de Ruiter, P.C., Moore, J.C., Zwart, K.B., Bouwman, L.A., Hassink, J., Bloem, J., Devos, J.A., Marinissen, J.C.Y., Didden, W.A.M., Lebbink, G., Brussaard, L., 1993. Simulation of nitrogen mineralization in the belowground food webs of 2 winter-wheat fields. Journal of Applied Ecology 30, 95e106. de Ruiter, P.C., Neutel, A., Moore, J.C., 1994. Modelling food webs and nutrient cycling in agroecosystems. Trends in Ecology and Evolution 9, 378e383. Desvilettes, C., Bourdier, G., Amblard, C., Barth, B., 1997. Use of fatty acids for the assessment of zooplankton grazing on bacteria, protozoans and microalgae. Freshwater Biology 38, 629e637. Fanin, N., Barantal, S., Fromin, N., Schimann, H., Schevin, P., H€ aettenschwiler, S., 2012. Distinct microbial limitations in litter and underlying soil revealed by carbon and nutrient fertilization in a tropical rainforest. PLoS One 7 (12), e49990. Fanin, N., Fromin, N., Buatois, B., H€ attenschwiler, S., 2013. An experimental test of the hypothesis of non-homeostatic consumer stoichiometry in a plant litter microbe system. Ecology Letters 16, 764e772. Frostegard, A., Tunlid, A., Baath, E., 2011. Use and misuse of PLFA measurements in soils. Soil Biology & Biochemistry 43, 1621e1625. Hanlon, R.D.G., 1981. Influence of grazing by Collembola on the activity of senescent fungal colonies grown on media of different nutrient concentration. Oikos 36, 362e367. Hassink, J., Neutel, A.M., Deruiter, P.C., 1994. C and N mineralization in sandy and loamy grassland soils e the role of microbes and microfauna. Soil Biology & Biochemistry 26, 1565e1571. Heal, O.W., Anderson, J.M., Swift, M.J., 1997. Plant litter quality and decomposition: an historical overview. In: Cadisch, G., Giller, K.E. (Eds.), Driven by Nature: Plant Litter Quality and Decomposition. CAB International, Wallingford, pp. 3e30. Hedlund, K., Augustsson, A., 1995. Effects of enchytraeid grazing on fungal growth and respiration. Soil Biology & Biochemistry 27, 905e909. Hedlund, K., Ohrn, M.S., 2000. Tritrophic interactions in a soil community enhance decomposition rates. Oikos 88, 585e591. Heemsbergen, D.A., Berg, M.P., Loreau, M., van Hal, J.R., Faber, J.H., Verhoef, H.A., 2004. Biodiversity effects on soil processes explained by interspecific functional dissimilarity. Science 306, 1019e1020. Herlitzius, H., 1983. Biological decomposition efficiency in different woodland soils. Oecologia 57, 78e97. Hessen, D.O., Ågren, G.I., Anderson, T.R., Elser, J.J., de Ruiter, P.C., 2004. Carbo sequestration in ecosystems: the role of stoichiometry. Ecology 85, 1179e1192. Hunt, H.W., Coleman, D.C., Ingham, E.R., Ingham, R.E., Elliott, E.T., Moore, J.C., Rose, S.L., Reid, C.P.P., Morley, C.R., 1987. The detrital food web in a shortgrass prairie. Biology and Fertility of Soils 3, 57e68. Kaiser, C., Franklin, O., Dieckmann, U., Richter, A., 2014. Microbial community dynamics alleviate stoichiometric constraints during litter decay. Ecology Letters 17, 680e690. Klamer, M., Baath, E., 2004. Estimation of conversion factors for fungal biomass determination in compost using ergosterol and PLFA 18:2 omega 6,9. Soil Biology & Biochemistry 36, 57e65. Lavelle, P., Blanchart, E., Martin, A., Martin, S., Spain, A., Toutain, F., Barois, I., Schaefer, R., 1993. A hierarchical model for decomposition in terrestrial ecosystems e application to soils of the humid tropics. Biotropica 25, 130e150. Manzoni, S., Jackson, R.B., Trofymow, J.A., Porporato, A., 2008. The global stoichiometry of litter nitrogen mineralization. Science 321, 684e686. Manzoni, S., Porporato, A., 2009. Soil carbon and nitrogen mineralization: theory and models across scales. Soil Biology & Biochemistry 41, 1355e1379. Mauclaire, L., Pelz, O., Thullner, M., Abraham, W., Zeyer, J., 2003. Assimilation of toluene carbon along a bacteria-protist food chain determined by 13C-enrichment of biomarker fatty acids. Journal of Microbiological Methods 55, 635e649.

Meentemeyer, V., 1978. Macroclimate and lignin control of litter decomposition rates. Ecology 59, 465e472. Mikola, J., Setala, H., 1998. No evidence of trophic cascades in an experimental microbial-based soil food web. Ecology 79, 153e164. Moore, J.C., McCann, K., Setala, H., De Ruiter, P.C., 2003. Top-down is bottom-up: does predation in the rhizosphere regulate aboveground dynamics? Ecology 84, 846e857. Mooshammer, M., Wanek, W., Schnecker, J., Wild, B., Leitner, S., Hofhansl, F., Blochl, A., Hammerle, I., Frank, A.H., Fuchslueger, L., Keiblinger, K.M., Zechmeister-Boltenstern, S., Richter, A., 2012. Stoichiometric controls of nitrogen and phosphorus cycling in decomposing beech leaf litter. Ecology 93, 770e782. Mooshammer, M., Wanek, W., Hammerle, I., Fuchslueger, L., Hofhansl, F., Knoltsch, A., Schnecker, J., Takriti, M., Watzka, M., Wild, B., Keiblinger, K.M., Zechmeister-Boltenstern, S., Richter, A., 2014a. Adjustment of microbial nitrogen use efficiency to carbon:nitrogen imbalances regulates soil nitrogen cycling. Nature Communications 5, 3694e3700. Mooshammer, M., Wanek, W., Zechmeister-Boltenstern, S., Richter, A., 2014b. Stoichiometric imbalances between terrestrial decomposer communities and their resources: mechanisms and implications of microbial adaptations to their resources. Frontiers in Microbiology 5. Parton, W., Silver, W.L., Burke, I.C., Grassens, L., Harmon, M.E., Currie, W.S., King, J.Y., Adair, E.C., Brandt, L.A., Hart, S.C., Fasth, B., 2007. Global-scale similarities in nitrogen release patterns during long-term decomposition. Science 315, 361e364. Paustian, K., Parton, W.J., Persson, J., 1992. Modeling soil organic matter in organicamended and nitrogen-fertilized long-term plots. Soil Science Society of America Journal 56, 476e488. Paustian, K., 1994. Modelling soil biology and biochemical processes for sustainable agriculture research. In: Pankhurst, C., Doube, B., Gupta, V., PR, G. (Eds.), Soil Biota. Management in Sustainable Farming Systems. CSIRO, Melbourne, pp. 182e193. Romani, A.M., Fischer, H., Mille-Lindblom, C., Tranvik, L.J., 2006. Interactions of bacteria and fungi on decomposing litter: differential extracellular enzyme activities. Ecology 87, 2559e2569. Schomberg, H.H., Cabrera, M.L., 2001. Modeling in situ N mineralization in conservation tillage fields: comparison of two versions of the CERES nitrogen submodel. Ecological Modelling 145, 1e15. Schroter, D., Wolters, V., De Ruiter, P.C., 2003. C and N mineralisation in the decomposer food webs of a European forest transect. Oikos 102, 294e308. €l€ Seta a, H., Tyynismaa, M., Martikainen, E., Huhta, V., 1991. Mineralization of C, N and P in relation to decomposer community structure in coniferous forest soil. Pedobiologia 35, 285e296. Singh, B.N., 1946. A method of estimating the numbers of soil protozoa, especially amoebae, based on their differential feeding on bacteria. Annals of Applied Biology 33, 112e119. Sistla, S.A., Schimel, J.P., 2012. Stoichiometric flexibility as a regulator of carbon and nutrient cycling in terrestrial ecosystems under change. New Phytologist 196, 68e78. n, O., Brussaard, L., Dangerfield, M., Ekschmitt, K., Lavelle, P., Tate, K., Smith, P., Andre 1998. Soil biota and global change at the ecosystem level: describing soil biota in mathematical models. Global Change Biology 4, 773e784. Van Soest, J., 1994. Nutritional Ecology of the Ruminant, second ed. Cornell University Press, Ithaca, NY. Veen, G.F., Freschet, G.T., Ordonez, A., Wardle, D.A., 2015. Litter quality and environmental controls of home-field advantage effects on litter decomposition. Oikos 124, 187e195. Wardle, D.A., 2002. Communities and Ecosystems: Linking the Aboveground and Belowground Components. Princeton University Press, Princeton. Wall, D.H., Bradford, M.A., John, M.G.S., Trofymow, J.A., Behan-Pelletier, V., Bignell, D.D.E., Dangerfield, J.M., Parton, W.J., Rusek, J., Voigt, W., Wolters, V., Gardel, H.Z., Ayuke, F.O., Bashford, R., Beljakova, O.I., Bohlen, P.J., Brauman, A., Flemming, S., Henschel, J.R., Johnson, D.L., Jones, T.H., Kovarova, M., Kranabetter, J.M., Kutny, L., Lin, K.C., Maryati, M., Masse, D., Pokarzhevskii, A., Rahman, H., Sabara, M.G., Salamon, J.A., Swift, M.J., Varela, A., Vasconcelos, H.L., White, D., Zou, X.M., 2008. Global decomposition experiment shows soil animal impacts on decomposition are climate-dependent. Global Change Biology 14, 2661e2677. White, D.C., Stair, J.O., Ringelberg, D.B., 1996. Quantitative comparisons of in situ microbial biodiversity by signature biomarker analysis. Journal of Industrial Microbiology 17, 185e196.