Effects of the Quality of Organic Soil Amendments on the Soil Community and on Nitrogen Mineralization in an Agroecosystem in the Georgia Piedmont

Final Report for GS05-044

Project Type: Graduate Student
Funds awarded in 2005: $8,576.00
Projected End Date: 12/31/2007
Grant Recipient: University of Georgia
Region: Southern
State: Georgia
Graduate Student:
Major Professor:
Carl Jordan
University of Georgia
Expand All

Project Information


We studied how the chemical quality of surface-applied organic amendments and the soil foodweb interact to determine nitrogen mineralization. Quality had substantial effects on the microbial communities in soil. By affecting microbial populations quality also influenced the consumer populations. Modeling suggested that: communities generated by different quality amendments have differential abilities for mineralization; communities may be better at mineralizing substrates similar to those that generated them; community structure is less important in determining mineralization for intermediate quality substrates. In soils whose faunal community has been impoverished mineralization is less responsive to quality than in soils with a more complex community.


The majority of agricultural soils in the Southern Piedmont region of the U.S. are degraded (Lal and Livari, 2004). Intense weathering combined with historic and contemporary agriculture further depleted soils that were naturally low in soil organic matter and bases. Farmers in the Southern Piedmont operate under these conditions, which inherently demand high fertilizer inputs. This demand is customarily met with the addition of synthetic fertilizers and, less commonly, the use of organic amendments imported from beyond the farm. Attaining more ecologically and economically sustainable food production systems demands strategies to build soil organic matter and nutrient availability for plants. Regular application of organic amendments to soil constitutes a central component of such strategies. Soil organic amendments such as green manures and crop residues can have an important role not only in building soil organic matter and in conservation of soil and water, but also in supplying nutrients—nitrogen in particular—to subsequent crops in rotations and to simultaneous crops (Gliessman, 1998).

Nitrogen can be the most limiting factor in low-input and organic systems (Clark et al., 1999). Whenever the use of inorganic nitrogen is replaced with organic materials, the availability of N in soil depends largely on the biological processes of nutrient release, which makes understanding these processes increasingly important. Two important factors governing the rate of nutrient release are: (1) the biochemical quality of organic matter inputs, and (2) the functioning of the community of soil organisms, including microorganisms, protozoa, nematodes and arthropods. Furthermore, organic inputs quality and soil communities can interact in soil to determine the rates of release of nitrogen into the soil. The details of this interaction are for the most part ignored. The purpose of this project was to combine experimental and modeling approaches to study how these two factors interact to determine N release into the soil and to inform practical recommendations for enhancing nitrogen availability in soil.

The role of organic inputs quality on nitrogen mineralization:
Release of nitrogen occurs as a sub-process of the decomposition of organic materials. Biochemical quality determines the susceptibility of a substrate to attack by decomposers. Indices of litter quality such as initial N concentration, C:N ratio, Lignin:N and (Lignin + polyphenol ):N ratios are good predictors of nutrient release rates under varied environmental circumstances. In agroecosystems, the selection of organic amendments and the timing of applications determine the quality and release pattern of the decomposition substrate.

The role of the soil community:
In agricultural systems, soil biota play important roles in regulating organic matter decomposition and nutrient release (Andren et al., 1990). Hendrix et al. (1990) stressed the concept of soil biota as fundamental components of sustainable agroecosystems, particularly as regulators of nutrient release. The decomposition of organic matter is a biological process in which microorganisms (bacteria and fungi) play the most direct role. However, the complete soil community (microbes and fauna) and the direct and indirect interactions between them are involved in the process (Coleman et al., 1994). Members of the soil fauna include protozoa, nematodes, micro and macroarthropods. Interactions among soil community members regulate the availability of the nutrient supply that allows for aboveground growth (Wardle, 2002). Indeed, direct trophic interactions are responsible for a great fraction of nutrient release. For example, de Ruiter et al. (1994) estimated that protozoa feeding on bacteria were responsible for up to 95% of the total N released in two arable farming systems.

Effect of quality on soil communities:
Agricultural practices such as application of organic amendments can affect population size and dynamics of organisms in soil communities (e.g. Bulluck III, 2002, van Vliet et al., 2000). Furthermore, the quality of added organic amendments can differentially affect soil populations structure (Robinson et al., 1994; Yang et al. 2003) and function (Bending et al., 2002). Most studies looking at the effect of the quality of organic amendments on communities have focused on the communities inhabiting the layer of added organic material (e.g. Bjornlund and Christensen, 2005; Wardle et al., 2006) or on soil communities after substrate incorporation (e.g. Bending et al., 2002; Salamanca et al., 2006) but few have looked at the effects of unincorporated substrates. However, in natural or managed systems in which the amendment does not get incorporated into the mineral soil (e.g. no-till systems), it is crucial to distinguish between the communities living in the amendment layer and the mineral soil communities as (a) they utilize different organic matter pools as energy and nutrient resources and thus can play different roles in driving processes and (b) the populations that compose them can establish interactions that may influence carbon and nutrient cycling and long term organic matter accumulation (Fontaine et al., 2004). How the mineral soil communities respond to the chemical quality of organic inputs can have implications on the dynamics of nutrients and indigenous soil organic matter (Waldrop and Firestone, 2004). Understanding the short term effects of amendment quality on soil detrital communities is of special importance in integrated agricultural systems where organic soil amendments such as green manure and crop residues are seasonally applied as a strategy to enhance soil fertility. The first objective of this study was to investigate the short term effect that the chemical composition of one-time surface-applied amendments has on the soil microbial and micro and mesofauna community in the mineral soil. We used five different substrates and a mixture representing a gradient of different quality parameters and attempted to determine (a) the effect of substrate type on soil communities and (b) what biochemical parameters most influence the soil micro food web groups’ abundance during the first six months of decomposition.

Interactive effects of soil faunal community and substrate quality on nitrogen mineralization:
Given the soil community’s role in nutrient cycling, modifications in its structure can potentially affect nutrient release processes and therefore nitrogen availability (de Ruiter et al., 1994). Agricultural management practices such as tillage or pest management can affect the composition of the soil community. Among the members of the soil community, the soil fauna is particularly vulnerable extinction or homogenization due to management. It is well known that the composition of the soil fauna can influence the decomposition and mineralization of plant litter and soil organic matter (Bradford et al., 2002; Brussaard, 1998; Coleman et al., 2004; Edwards, 2000; Lavelle et al., 1993) and thus it would be expected that changes in the faunal community imply changes in the dynamics of decomposition and nutrient mineralization. In general, nutrient mineralization and decomposition increase with increasing body size of the soil animals and with increasing complexity of the faunal community. However, interactive effects of the soil fauna community structure and the quality of available substrates on mineralization and decomposition have been documented. Fauna can increase decomposition and mineralization rates of low quality litter (Couteaux et al., 1991; Tian et al., 1992), but the opposite has been observed as well (Schadler and Brandl, 2005). Gonzalez and Seastedt (2001) on the other hand found no fauna-litter interactions. Some studies have shown that the composition of the faunal community can affect the level of control of organic matter quality on soil processes. In particular, a more complex fauna seems to enhance the degree of control of substrate quality on decomposition (Schadler and Brandl, 2005; Smith and Bradford, 2003). The second objective of this study was to investigate the effect of the structure of the faunal community on nitrogen mineralization and in particular to explore one potential way in which fauna could affect nitrogen mineralization: by mediating the control that the quality of surface added substrates exerts on the structure of the micro-food web (the microbial groups and their direct predators). In a factorial arrangement, we exposed soil to surface-applied plant materials of contrasting chemical compositions and restricted the access of size-classes of fauna to mineral soil. After six months of decomposition, we assessed the effects of substrate type on the structure of microbial community in the mineral soil when fauna had been excluded and when it was present, and measured nitrogen mineralization and decomposition under both scenarios after 21, 91 and 165 days. We also estimated the abundances of microarthropods and nematode trophic groups in soil. We asked (a) does soil fauna mediate the effect of substrate type and quality on nitrogen mineralization, (b) does soil fauna mediate the effect of substrate type on the structure of the microbial community and the micro foodweb in the mineral soil, and (c) is the effect of fauna on mineralization associated with its effect on the microbial community and their predators?

Effect of quality on nitrogen mineralization due to its effect on trophic interactions:
It has been suggested that the effect of substrate quality on soil processes is driven by its effects on the soil organisms responsible for the processes (Wardle, 2002). By affecting the soil community structure, substrate quality can affect the trophic interactions occurring in the organic layer and mineral soil habitat. Trophic interactions in the soil food web have major effects on carbon and nutrient mineralization. Carbon mineralization can increase as a result of higher turnover rate activity and respiration of consumed populations due to grazing (Bardgett et al., 1993) while nitrogen mineralization occurs mainly due to excretion of excess nitrogen by the consumer (Woods et al., 1982). Furthermore, 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 (Bardgett, 2005; Herlitzius, 1983; Wardle, 2002). Thus, the quality of organic amendments has the potential to influence the dynamics of nutrients and carbon due to its inherent chemical degradability but also due to its effects on the trophic interactions among soil populations. These two factors may in turn interact to affect mineralization. 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 and that of substrate quality.

The complexity of the decomposition process and biological interactions in soil makes 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 organism-oriented modeling approach initiated by Hunt et al. (1987) has been applied to several natural and agricultural systems (Berg et al., 2001; de Ruiter et al., 1994a; Hassink et al., 1994; Schroter et al., 2003). In this approach food webs are constructed by aggregating species into functional groups and the structure and functioning of the food webs are analyzed in relation to nutrient cycling. Soil food web models have proven useful in predicting C and N mineralization rates (de Ruiter et al., 1994b), in explaining rates in terms of the relative contribution of groups of organisms and particular trophic interactions (Berg et al., 2001).

For the third objective we use the soil food web model scheme of Hunt et al. (1987) to simulate carbon and nitrogen mineralization from surface applied substrates of differing chemical qualities and from mineral soil based on observed population sizes and the trophic interactions among the members of the soil food web. This model describes mineralization as being regulated not only by trophic interactions or chemical quality limitations, but by the interaction of both. We add simple features to describe differential chemical composition of the substrates and differential abilities of fungi and bacteria to degrade organic matter fractions. This approach allows us to isolate the effects of the changes in soil food web structure prompted by the added substrate from the effects of the quality of substrate on trophic transfers.

We calibrated the model using the measured soil populations and nitrogen mineralization after the application of one substrate and then assessed its performance with the other plant materials. We then used the model to investigate (a) the importance of the soil community changes brought about by the quality of substrate on carbon mineralization and nitrogen mineralization; (b) whether nitrogen mineralization is better predicted by the interaction of soil community and substrate quality than by each factor alone; (c) whether the role of the soil communities and their trophic interactions varies depending on the quality of the degrading substrate; and (d) whether some communities are better suited to degrade substrates of certain quality.

Project Objectives:
  1. Assess the response of the microbial community, nematodes and microarthropods to the following amendments over a growing season: (a) green manure from a leguminous alley cropping species (Amorpha fruticosa L., a native woody shrub.), (b) green manure from a leguminous non-woody species (Trifolium incarnatum L. or crimson clover) used as a winter cover crop, (c) green manure from cereal rye (Secale cereale L.), (d) wheat hay (Triticum aestivum L.), (e) pine needles (Pinus taeda L.), and an even mixture of these by weight.

    Investigate the effect of the complexity of the faunal community on nitrogen release into the soil and the interactive effects of the faunal community structure and the quality of organic amendments on nitrogen mineralization

    Using experimental and modeled data, study the effect of the observed responses of the soil community and their trophic interactions on nitrogen release into the soil.


Click linked name(s) to expand/collapse or show everyone's info
  • Yolima Carrillo


Materials and methods:

Objective 1:
To study the effect of the chemical quality of substrates on the structure of the community in mineral soil a field study was conducted in experimental plots in a previously abandoned conventional farm in the Piedmont region of Georgia USA (33°57’N lat. 83°19’W long). Soil is classified as Pacolet sandy clay loam (kaolinitic, thermic typic hapludult) and had 0.7% C and 0.1%N content. In June of 2005 standing vegetation was removed from a 100 m2 plot. Removed vegetation consisted of grasses and forbs. The plot was kept vegetation-free throughout the duration of the study by periodically pulling out any emerging plants. The plot was divided into 24 2 x 2 m sub-plots and sheets of aluminum flashing were buried down to 4-5 cm and placed as dividers. Soil from the first 5 cm of a randomly located area of 35 x 35 cm within each plot was collected and sieved to 4 mm. This soil was then frozen at -80°C for 72 h to kill fauna. After thawing, soil was placed back in the field in metallic trays (35 cm x 35 cm x 5 cm) made of 5-mm mesh. Soil was left bare and allowed to re-colonize and stabilize for 2 months. Six quality treatments, five substrates and a mixture, were randomly assigned to four of the plots and were surface applied at a rate of 327g/m2 both to the inside and to the outside of the tray on July 22.
Substrates used were air dried cereal rye (Secale cereale L.), air dried crimson clover (Trifolium incarnatum L.) air dried false indigo (Amorpha fruticosa L., hereafter “Amorpha”), wheat hay (Triticum aestivum L.), pine needles (Pinus taeda L.), and an even mixture of these by weight. Dry samples of the initial substrates were ground using a Spex CertiPrep 8000-D Mixer Mill and used for chemical analyses and estimation of ash free dry mass. For C/N samples were analyzed on a Carlo Erba Elemental Analyzer and reported as carbon and nitrogen percent dry weights. Total phosphorus was measured from 0.5 g of sample that was ashed, extracted in aqua regia acid, and analyzed on an automated Alpkem Analyzer. Cellulose, hemicellulose, and lignin concentrations were measured from 0.5g of each sample that underwent sequential neutral detergent/acid detergent digestion on an Ankom A200 fiber analyzer. Percentage N and C were determined for remaining biomass at the end of the incubation.

Sample collection and processing:
Soil samples from each tray were collected 21, 91 and 165 days (August, October and February from herein) after substrate application. Soils were immediately processed for fauna, kept refrigerated for up to a day then freeze-dried for PLFA extraction and mineral nitrogen. Approximately 4 grams of soil were extracted for NO3- and NH4+ by shaking in 20 ml of 2 M KCl for one hour. Extracts were analyzed with an Alpkem Continuous Flow Analyzer. Soil moisture content was determined by drying at 100°C for 24h. pH was measured in water with a 1:1 water to soil ratio.

Microbial community:
Phospholipid fatty acids were extracted from 2-mm sieved and freeze-dried soils. Five grams of soil were extracted with the Bligh-Dyer technique (Bligh and Dyer, 1959) by shaking for 2 h with a single phase mixture of methanol-chloroform-phosphate buffer (2:1:0.8 ratio in volume) in a teflon bottle. The mixture was centrifuged to decant soil. Extra volumes of chloroform and methanol were added and the sample was centrifuged again to separate the organic phase containing the lipids which was then fractionated into neutral lipids, glycolipids and phospholipids (Zelles and Bai, 1993) with an activated silica gel column (BondElut; Varian, Palo Alto, CA). Methylation of phospholipids was done with mild alkaline methanolysis (White et al., 1979) to produce fatty acid methyl esters (FAMEs). FAMEs were purified with NH2 aminopropyl columns (Bond, Elut; Varian, Palo Alto, CA).

Samples were analyzed with a Hewlett-Packard (HP; Palo Alto, CA) 6890 series GC with a flame ionization detector and a 30 meter DB-5 (film thickness = 0.25 microns, internal diameter 0.32 mm; Agilent, Santa Clara, CA). Individual PLFAs were quantified in relation to an external standard (20:0 ethyl ester) that was run in duplicate with every batch of samples. Compounds were identified by comparison of their retention times with those of a prepared mixture standard containing 36 FAMES (Matreya, PA; Sigma-Aldrich, MO; Nu-Chek, MN) that was run with each batch. Analysis by GC-MS was used to identify FAMEs common in the samples that were not contained in the standard mixture. Fatty acid notation used follows that in Frostegard and Baath (1996).

Soil fauna:
Nematodes were extracted from approximately six grams of fresh soil with the Baermann funnel method for 72 hours. Nematodes collected were preserved in 5% formaldehyde. Total nematode counts to the level of trophic group were performed for each sample. Tylenchidae were included in the fungivorous trophic group. Microarthopods were extracted for four days on Tullgreen-type extractors. They were enumerated and sorted into Collembola, Prostigmata, Mesostigmata, and Oribatida.

Statistical Analyses:
Phospholipid fatty acid methyl ester concentrations were converted to mole percentage. PLFA community profiles were then subjected to principal components analysis (PCA). Significance of principal components was assessed with one-way ANOVAs. Correlations between PCs and substrate quality parameters and soil variables were assessed with Pearson correlation analysis. The response of total PLFAs and the ratio of fungal to bacterial PLFA markers to substrates over time were analyzed with repeated measures ANOVA. Nematodes and microarthropods abundance data over time were analyzed by repeated measures ANOVA. Data were log transformed prior to analysis. Pearson’s correlations coefficients were determined for relationships of substrate quality and soil variables with soil community variables. All analyses were performed using JMP®. A probability level of 0.05 was used to determine significant effects.

Objective 2:
Objective 2 was addressed in parallel with Objective 1 using the same field set-up (see Materials and Methods Objective 1) but involved some additional treatments. Soil from the first 5 cm of four areas of 25 x 50 cm within each of the 24 sub-plots was collected, mixed and sieved to 4 mm. Care was taken to allow at least 20 cm of distance between the four areas and between their edges and the edge of the sub-plot. Collected soil was then frozen at -80°C for 72 h to kill existing fauna. After thawing, soil was placed back in the field in metallic frames (25 cm x 50 cm x 5 cm) that fitted the previously dug holes. Soil under and around the sides of the frames was relocated as necessary to make sure that there was contact between soil inside and outside the frame. The bottom and the sides of the metallic frames were lined with mesh of different sizes to restrict access of fauna. Mesh sizes used were 5 cm, 2 mm, 100 μm and 40 μm. Soil was added to the mesh-lined frames (boxes from here on) to reach a depth of approximately 3.5 cm to allow room for the addition of plant residue. The boxes were open on top, so that the mesh restricted access to mineral soil dwellers but organisms living in the organic layer could move in and out by jumping the 1-2 cm side wall of the boxes. Soil was left bare and allowed to re-colonize and equilibrate for 2 months. Substrate treatments consisting of five plant residues and a mixture (as in Objectives 1 and 2) were randomly assigned to four of the plots (4 replicates) and were surface applied at a rate of 327 g/m2 both to the inside and to the outside of the boxes on July 22.

Sampling for mass loss and nitrogen mineralization was carried out 21, 91 and 165 days (August, October and February from hereon) after substrate application for mesh-size treatments 40 μm and 5 mm (the most contrasting ones) and for all size-class treatments in August. Nematodes and microarthropods abundances were assessed on all three dates in mesh sizes 40 μm and 5 mm and microbial community composition only in August and for mesh sizes 40 μm and 5 mm. After collection, soils and amendments were immediately processed for microfauna and then freeze-dried for PLFA extraction and mineral nitrogen.
Mass Loss:
Mass loss was measured using 3g of plant material in 13 mm x 14.5 mm nylon litterbags with a combination of 2 mm mesh (to prevent loss of material by fragmentation) and ten 4mm-diameter holes (to allow macroarthropods in the 5mm mesh size treatment). In the case of the mixed substrate treatment, component species were equally represented in the total of 3g. Three bags, one for each sampling, were placed in box on top of the amendment layer. After collecting the litterbags, half of the material was used for nematodes extraction (see below) and the remainder was used for microarthopods (see below). Ash free dry mass of remaining material was calculated after ignition for 3 hours at 500°C.

Nitrogen Mineralization:
Initial soil concentrations of NO3- and NH4+ were estimated from six soil samples collected from randomly selected boxes immediately before substrate application. Release of NO3- and NH4+ from substrates and soil was determined by incubation of nylon bags containing anion and cation exchange resins which were buried under the soil layer in each box to retain NO3- and NH4+ leached from the amendment and soil (Binkley and Matson, 1983). Each nylon bag measured approximately 4.5 cm in diameter and contained 24 g of an even mixture of Na-saturated cation and Cl-saturated anion exchange resins (Sybron Ionac C-250 and ASB-1P, Sybron Chemicals, Birmingham, NJ). One nylon bag was buried in each box immediately before substrate addition, removed after three weeks of incubation and replaced by a new one for the next incubation period (9 weeks). This procedure was repeated for the final incubation period (10 weeks). At each sampling date a soil sample was collected from each tray. Resins and soils (4 g) were extracted by shaking in 2M KCl for one hour (20 ml for the soils and 60ml for the resin-filled bags). Extracts were analyzed for NO3- and NH4+ using an Alpkem Continuous Flow Analyzer. Nitrate and ammonium from soil extractions were converted to μg per gram of soil. To calculate the amount of nitrate and ammonium derived from soil and plant material that had accumulated in the resins, it was assumed that the majority of the nitrate and ammonium collected had been derived from the soil directly above it (approximately 64 cc3 of soil, since the diameter of the resin bag was approximately 4.5cm and the depth of the soil above resin bag was 4 cm). Using an average value of bulk density for the site, nitrate and ammonium accumulated in the resin over each incubation period was converted into μg per gram of soil. Concentration of mineral nitrogen was calculated using the combined amounts of nitrate and ammonium in both soils and resins and net mineralization per day for each incubation period was calculated as the difference between final and initial total nitrogen (NO3- + NH4+) divided by the number of days in each incubation period (Kolberg et al., 1997). Initial soil inorganic N used for the last two incubation periods was the average N content in soil from the four replicates collected at the end of the previous incubation period.

Microbial community:
Microbial community structured was studied using Phospholipid fatty acids (PLFA) following the methods presented for Objective 1. PLFA were extracted from 5 g of 2 mm sieved and freeze-dried soils. PLFA results are reported as percentage of total concentration. Total PLFA were obtained by adding the concentrations (as ug/g soil) of all detected PLFA.

Nematodes and microarthropods were extracted and processed as explained under Materials and Methods for Objective 2. Approximately six grams of fresh soil were extracted for nematodes and 25 g for microarthropods.

Statistical Analyses:
Mass loss, decomposition and fauna data were analyzed by repeated measures two factor ANOVA with mesh-size and substrate type as factors and time as the repeated measure. Fauna densities were log (n+1) transformed prior to analysis. Two-way ANOVA was used for microbial community characteristics in the last sampling date. Tukey’s test and t ratios were used for means comparisons when appropriate. Simple linear regression was used to assess the control on mineralization rate by quality parameters.

Objective 3:
The results of the field study described under Materials and Methods for Objective 2 were used to estimate populations’ biomasses in mineral soil and amendment layer and to measure nitrogen mineralization from the substrates over the course of the first six months of decomposition.

Nitrogen mineralization:
Details of the processing and calculations to estimate average net mineralization rate are presented under Methods and Materials for Objective 3. Values were converted from μg NO3-+NH4+/g soil /day to mg NO3-+NH4+ /m2/day for the upper 5cm of soil using measured soil bulk density of 1.1g/cc3. For model calibration and validation purposes an overall average mineralization rate for the whole period was calculated from the averages for each period.

Estimation of populations’ biomass:
Soil microbial community composition (using phospholipid fatty acids, PLFA), nematodes and microarthropods abundances were assessed on all three dates. Details of the extraction and processing of soil samples are included in Materials and Methods for Objective 2. Populations in the organic layer were only assessed at the end of the incubation. Phospholipid fatty acids were extracted from 1gram of dry and ground plant material and extracted as detailed in Materials and Methods for Objective 2. Nematodes and microarthropods were extracted from litter and processed as detailed in Materials and Methods for Objective 2.
The concentration of total and some individual phospholipid fatty acids (as μmoles per gram) were used to obtain estimates of biomass of microbial groups. Total microbial biomass was estimated using the conversion factor in Bailey et al. (2002) and a Kc factor of 0.32. The fatty acid 18:2ω6 was used to estimate fungal biomass in soil using the conversion factor provided by Klamer and Baath (2004) for soil fungi. The biomass of bacteria was estimated as the difference between total microbial biomass and fungal biomass. Fungal biomass in the organic layer was estimated using the relative differences observed in the concentration of the fatty acid 18:2ω6 in relation to its overall average concentration in the six substrates 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). Fungal and bacterial biomasses in the organic layer were estimated from total microbial biomass using the calculated ratio.

Existing methods for estimation of protozoa abundance are very time and labor intensive and can make population assessments in experimental contexts infeasible or impractical. 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 mineral soil and in the organic layer. Individual PLFA can be used as markers for protists as their cell membranes contain specific polyunsaturated fatty acids (Desvilettes et al., 1997; White et al., 1996). The fatty acids 20:2ω6,9c; 20:3ω6,9,12c; 20:4ω6,9,12,15c have been used as relative indicators of protozoa abundance in soils (Cavigelli et al., 1995; Mauclaire et al., 2003). This is the first attempt to estimate a conversion factor from PLFA to protozoan biomass of which we are aware.

Protozoa from fresh soil samples from the field site were cultured for 2 days at 30C in phosphate agar as substrate and Escherichia coli as a food source. The cultures were then suspended in phosphate agar and the suspensions were pooled into one large flask. Immediately after, the abundance of ciliates, amoebae, and flagellates in this suspension (as numbers per milliliter) was estimated in five replicates using the dilution method (Most Probable Number) based on Singh (1946). Immediately after the enumeration, known volumes of the original suspension (10, 20 and 40 ml in triplicate) were placed in Teflon bottles and centrifuged at 24000 RCF for 30 minutes, after which the supernatant was poured and the samples were frozen at -80ºC. PLFAs were extracted from thawed samples using the same extraction procedure as for soil and organic layer samples.

From the concentrations of cells in the initial suspension obtained with MPN, biomass carbon 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 strong linear relationship (r2= 0.70) was found between the concentration (μmoles/ml of suspension) of the fatty acid 20:4ω6,9,12,15c and the calculated protozoan biomass in the three analyzed volumes of culture suspension (as μg C). The following relationship was obtained:
ug C protozoa = 54860 (μmoles of 20:4ω6,9,12,15c ) + 3.5525.
This relationship was used to estimate protozoan biomass. Estimates of protozoan biomass ranged between 0.5% and 1.9% of total microbial biomass estimated with PLFA.

Numbers of nematodes and microarthropods per gram of soil or subsbtrate were converted to mg of biomass carbon using the conversion factors in Beare et al. (1992). Obtained values of biomass carbon for all microbial faunal groups were converted to mg per square meter of soil down to 5 cm using a bulk density value of 1.1g/cc3. Biomasses per gram of substrate were converted to biomass per square meter based on the initial rate of application and considering the amount remaining in litter bags at the time of sampling.

Food web modeling:
The model was written using modeling software package STELLA 8.1. The food web model was based on the approach by Hunt (1987) to obtain mineralization rates from feeding interactions between soil functional groups of known biomass. We used this approach to simulate mineralization of carbon and nitrogen per day from the substrates and soil organic matter. Only the average biomasses—not the dynamics— of the major soil groups observed in mineral soil over the course of the field study, and at the end of the study for , were used as inputs to the model and played the role of driving the consumption of organic materials and their mineralization. Consequently, the simulated daily rate was considered the average rate for the study period. Carbon 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 is calculated as in de Ruiter et al. (1993):
F=((Dnat * B)+P)/eass. eprod
where F: feeding rate; Dnat: natural death rate; B: biomass of functional group; P: predation rate; eass: assimilation efficiency and eprod: production efficiency. Nitrogen flows occur in parallel and in proportion to C flows through the use of the C/N ratios of organisms and organic fractions. Carbon and nitrogen mineralization rates are calculated as in de Ruiter et al. (1993):
Cmin= eass-(1- eprod)*F
Nmin= eass((1/CNprey)-(eprod/CNpred))*F
where C or Nmin: mineral carbon or nitrogen released per trophic transfer; eass: assimilation efficiency of the consumer; CNprey: carbon-to-nitrogen ratio of the prey or substrate; eprod: production efficiency of the consumer; C/Npred: carbon-to-nitrogen ratio of the consumer; F: feeding rate of consumer on prey. Specific death rates were made temperature dependent using a Q10=3 (Andren et al., 1990). Temperature was measured in the field every two hours throughout the sampling period. The overall average was used for simulations. Physiological parameters (CN, ea, ep, and death rates) for the functional groups in the organic layer and soil food webs were taken from the literature and are shown in Table 3.1.

Two separate food webs were modeled. Organisms found in the organic layer were assumed to only consume organic layer material and prey inhabiting it; organisms found in soil were assumed to only consume material and prey in the mineral soil. The soil community was modeled as a simplified food web composed of five functional groups: bacteria, fungi, protozoa, nematodes and microarthropods. The specific trophic interactions modeled are shown in Figure 3.1. The differences in trophic transfers considered in the organic layer and mineral soil corresponded to observed differences in the composition of the major functional groups assessed. For instance, in the case of the organic layer, the great majority of nematodes were either bacterial or fungal. The specific proportion found for each group was included in the model as a factor modifying the total consumption demand by nematodes. Microarthropods in the organic layer were either oribatids or collembola which are both considered to be mainly fungivores (Coleman et al., 2004). Thus, in the organic layer food web model microarthropods only consumed fungi. In the case of soil, most nematodes found were omnivores or bacterial feeders. In the model, soil nematodes consumed bacteria, protozoa and fungi and their consumption of one particular prey was proportional to the relative abundance of that prey. 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 of oribatids and prostigmatids was used to split the consumption by microarthopods in soil.

Substrate to be mineralized was divided into cellulosic/hemicellulosic material, lignin 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 assumed to be 25%, 5% and 70% respectively. 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 were calculated to approximately reflect the relative proportions among the decay constants of these carbon pools used to model decomposition and mineralization in the original version of the 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. 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. Material return to the three organic matter pools was due to death and excretion using the fractions in Hunt et al. (1987) (Table 3.1).

C/N of the substrate was defined as the weighed average of the C/N of its three fractions. It was assumed that the C/N of the lignin fraction is 2 times that of the carbohydrate fraction and the C/N of the cellulose fraction is 3 times that of the carbohydrate fraction, so that:
CN substrate: (2*CN lignin*lignin fraction)+(3*CN cellulose*cellulose fraction)+(CN labile fraction*carbohydrate fraction)
The CN of the labile pool was calculated using the known CN of the substrate and used to calculate the CN ratio of the other fractions. CN 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. Initial and final (after six months of decomposition) C/N ratios of the substrates were measured with the micro-Dumas combustion method and the average value of the two was used as their C/N.

Nitrogen mineralized from the added substrate is assumed to be leached into the soil and join the mineral nitrogen pool in soil, of which nitrogen 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 nitrogen pool. The amount of nitrogen immobilized corresponds to all the nitrogen that would be required to meet the consumption demand by the predator according to its C/N.

The model was calibrated using the observed population biomasses, C/N of added substrates and average net mineralization rate for rye, which had an intermediate C/N value. To evaluate the performance of the model, we compared observed average net mineralization rates for the study period with modeled rates obtained with the observed population biomasses and average C/N ratios of the six substrates studied.

Research results and discussion:

Results Objective 1.
Chemical composition of substrates:
Initial chemical composition of applied substrates is shown in Table 1.1. Pine needles had the lowest concentration of nitrogen and phosphorus, the highest C/N ratio and lignin percentage and the lowest percentage of cellulose. Wheat hay had the second lowest values of N and P and followed pine needles in its C/N ratio; it had a low concentration of lignin but the highest of cellulose and hemicellulose. The composition of rye was similar to that of hay except for its considerably lower C/N ratio. Amorpha and clover showed the lowest C/N ratios and the highest concentration of P. Among these two, Amorpha had the highest percentage of N and P and the lowest C/N ratio but the highest percentage of lignin after pine needles. The mixture of the five substrates had the average concentration for all parameters.

Effects of substrate type on soil characteristics:
Soil moisture did not differ significantly between substrate types on any of the sampling dates (p>0.05), data not shown). Soil pH was measured only for the October date. No significant difference between substrates was observed (p>0.05), data not shown). Substrate type had a significant effect on soil NO3- on all sampling dates and on NH4+ concentrations only in August (Table 1.2).

Total Phospholipid Fatty Acids:
All identified and unidentified PLFAs were summed to obtain total phospholipid fatty acid content per gram of soil (as μg of PLFA/g of soil). There was a significant interaction of substrate by time but no overall effect of substrate on total PLFAs (Table 1.3). Total PLFAs decreased between August and October and increased between October and February, except for pine in which it decreased throughout the course of the study (Figure 1.1). A significant difference between substrate treatments was observed only for February (p=0.0014), when pine had the lowest concentration and Amorpha, rye and the mixture the highest.

Microbial community:
Seventeen phospholipid fatty acids that were identified and common to all samples were used for PCA analysis. PLFA profiles were influenced more strongly by sampling date than by substrate (Figure 1.2). The relationships of PC1 and PC2 with time were statistically significant (p<0.001 and p=0.019 respectively). Samples collected in October separated along PC1 from samples collected in August and February. On the basis of eigenvector loadings, communities from October were enriched in i17:0 and 15:0, whereas communities from February contained greater amounts of 18:1ω7, 18:1ω9 and 16:1ω5. Communities from August were located in between but tended to be closer to those from February. PC2 distinguished communities from October and February, which had high loadings of a17:0, 10Me18:0 and cy19:0 from those of August, which were enriched in 14:0, 18:2ω9 and 20:3ω6.

Community PLFA profiles were analyzed by PCA one date at a time to remove sampling date effects and focus on the effect of substrate. For this, identified PLFAs were grouped into 5 categories (Table 1.4) to maintain a high ratio of samples to variables. In August (Figure 1.3a) substrates pine and hay separated from all other substrates along PC1 (P=0.04) due to greater abundances of Gram-positive bacteria and actinomycetes and increased presence of Gram–negative bacteria and fungi under the other substrates. PC1 was highly and positively correlated to soil NO3- (r=0.61, p=0.0024) and NH4+ (r=0.70, p=0.0003) and substrate %N (r=0.49, p=0.02), %P (r=0.51, p=0.02) and negatively related to %C (r= -0.59, p=0.004) and C/N (r= -0.62, p=0.002). PC2 separated the rye community from other substrates. However, the relationship between PC2 and substrate was not significant. Amorpha, clover and the mixture form a cluster along PC2 associated with Gram–negative bacteria. In October, neither PC1 nor PC2 showed a significant relationship with substrate. However, pine, hay and the mixture appeared to be separated from rye clover and Amorpha along PC2 (Figure 1.3b). Eigenvector loadings indicated higher relative abundances of Gram-positive bacteria in pine, mix and hay and greater importance of Gram–negative bacteria and fungi in Amorpha, rye and clover. PC2 was positively correlated to %C (r=0.49, p=0.02), % lignin (r=0.46, p=0.03), and lignin-to-nitrogen ratio (r=0.51, p=0.01). In February (Figure 1.3c), pine, Amorpha and rye showed higher abundances of protozoan and fungal PLFA markers, whereas hay, clover and the mix showed enrichment of Gram-positive bacteria (PC1, p=0.12). Pine, Amorpha and hay communities were distinctive for their high loadings of actinomycetes along PC2 while the mixture, clover and rye soils were enriched in the PLFA fungal marker. PC2 (p=0.0002) was positively correlated to %C (r=0.75, p<0.0001), % lignin (r=0.60, p=0.002) and lignin-to-N ratio (r=0.61, p=0.002).

PLFA fungal-to-bacterial ratios:
No interaction of substrate by time was observed (Table 1.3). Highly significant effects of substrate were detected at all sampling dates (p=0.001, 0.0002, 0.009 for August, October and February respectively). Rye, clover and Amorpha had the highest F/B ratios while pine, hay and the mixture had the lowest. F/B declined under all substrates between August and October after which no general trend was observed (Figure 1.4). Differences in F:B between and within dates were driven by differences in the fungal PLFA marker (data not shown).

Significant time/substrate interactions were found for bacterivorous and omnivorous nematodes (Table 1.3). The highest abundances of nematodes groups were found in August, except for the predatory group, which increased as decomposition progressed (Figure 1.5). The greatest differences among substrates treatments were also observed in August and tended to decrease over time, except for the predatory nematodes. Nematode abundances within dates showed high variability, but trends in the ranges were evident and therefore these are presented descriptively, not statistically. In August bacterivorous nematodes reached higher numbers under clover, Amorpha and the mixture. Abundances remained high under clover by October. In October and February abundances under Amorpha had the lowest values among all substrates. In August and October omnivorous nematodes under rye, clover and Amorpha showed the greatest abundances whereas pine and hay had the lowest. Predatory nematodes showed no clear response to substrate addition in Augsut and their greatest response occurred in February when they were most abundant under clover and the mixture. Overall predatory nematodes increased as decomposition progressed except when hay was the added substrate, where they consistently decreased with time. Fungivorous nematodes were rare and constituted only up to 2% of all nematodes and were absent under the Amorpha treatment.

No differences in the responses to time or substrate of oribatid, prostigmatid and mesostigmatid mites were observed. Pooled numbers of all mites are presented. Mites and collembola significantly responded to time and no overall effect of substrate or substrate/time interactions were found (Table 1.3). The greatest abundances as well as differences among substrates were observed in August for mites and Collembola (Figure 1.6). On this date, although no significant effect was found, the highest numbers of mites occurred under rye and Amorpha and those of Collembola under hay.

Substrate quality parameters and soil variables as drivers of relative abundances:
Relationships of the abundances of PLFA biomarkers with initial plant quality variables as well as soil moisture and soil content of nitrate and ammonium were examined with Pearson correlation analysis (Table 1.5). Percentage of hemicellulose and soil moisture at sampling dates did not show any apparent relationship with group abundances. No plant quality or soil variable was related to Protozoan PLFA markers at any date. Several significant relationships were observed in August and fewer are found thereafter. No microbial group showed any relationship to substrate cellulose, lignin or lignin to N ratio in August, however Gram-negative bacteria were positively affected by %P and %N and negatively affected by %C and C/N. The opposite was observed for Gram-positive bacteria. Fungi and F/B were negatively related to %C. Fungal markers were also negatively correlated with C/N. In October the only microbial group that was related to quality parameters was the Gram-negative bacteria. This group was negatively related to % lignin and lignin/N. In February the actinomycetes were strongly and positively correlated to %C and Lignin/N and % Lignin.

Microbial groups also showed significant strong relationships with soil mineral nitrogen but only in August, when fungal, fungal to bacterial markers ratio and Gram-negative bacteria were strongly and positively related to soil NH4+. Gram-negative bacteria and fungi were also positively related to NO3-. Actinomycetes had a negative correlation with soil NH4+.

Relationships between the abundance of nematode trophic groups and microbial biomarker groups were also assessed for every date. In August bacterivorous nematodes were positively correlated with Gram-negative bacteria (r=0.55, p=0.007) and general bacterial makers (r=0.54, p=0.012). Fungivorous nematodes were negatively related to protozoan markers (r=-0.46, p=0.04). In October and February only omnivorous nematodes showed relationships with microbial markers. They were negatively correlated to Gram-negative bacteria (r=-0.56, p=0.007) and bacteria (r=-0.54, p=0.009) in October and positively related to fungi in February (r=0.44, p=0.044), when they also showed a negative relationship with actinomycetes (r=-0.46, p=0.03).

No significant relationships were found between microbial groups and microarthropod groups.

Discussion Objective 1.
Some studies that have looked at the long term effect of mulches (surface applied plant residues or other materials) on mineral soil communities (Forge et al., 2003; Tiquia et al., 2002; Yang et al., 2003), have found that after long-term exposure (one to several years) soil microbial and micro and mesofauna communities differ only when the mulching materials are considerably contrasting (e.g. compost vs. wood chips). In our six-month study we found that although sampling date had a greater effect on the structure of microbial communities than substrate type, within each sampling date the structure of the microbial communities in mineral soil was different under different plant substrates. This was true after 3 weeks, 12 weeks and 29 weeks of substrate addition. This suggests that different factors might be involved in the short term and long term effects of added substrates on soil communities. In the longer term soil characteristics such as organic matter composition and quality and soil pH can be influenced by the quality of organic matter inputs (Binkley and Giardina, 1998) and in turn modify the soil community (Hackl et al., 2005). In the shorter term the influence is probably due to temporary changes in carbon and nutrients availability brought about by the added substrate. It is in the early stages of organic matter decomposition when the soluble fraction is released and can be leached into the mineral soil and become available for mineral soil populations. McMahon et al. (2005) demonstrated that the soluble fraction of ryegrass litter served as carbon source for the microbial biomass in bulk soil.

We found the separation of the communities within dates obtained with PCA to be driven by different soil, substrate and community variables at different times. By week 3 (August), increased abundances of Gram-positive bacteria and actinomycetes under pine and hay were responsible for the distinction between these communities and the communities under clover, Amorpha, the mixture and rye which were in turn enriched in Gram-negative bacteria and fungi. This ordination was very strongly correlated with the N and P concentrations in substrates and with mineral N in soil, indicating that the Gram-negative bacteria and fungi were favored by high nutrient availability and Gram-positive bacteria and the actinomycetes by low nutrient availability (under pine and hay). By week 12 (October), Gram-negative bacteria and fungi were more abundant under Amorpha, rye and clover. The PCA ordination obtained (PC 2) was correlated with added substrate %C, % lignin and lignin-to-nitrogen ratio indicating that Gram-positive bacteria were favored when there were high carbon and lignin concentrations in substrates and a high lignin-to-nitrogen ratio, while Gram- negative bacteria and fungi responded in the opposite manner. After 22 weeks (February), the ordination that distinguished pine, Amorpha and hay communities from the other substrates was correlated with the same quality variables as in October. In this case, indicating that the actinomycetes were more abundant in soil when the added substrate had high carbon and lignin concentrations and lignin-to-nitrogen ratio while fungi were negatively affected by these variables.
While the distinction between communities in August was associated with substrate and soil nutrient content, in October and February %C, lignin and lignin-to-nitrogen ratio had the greatest influence. This seems to parallel what is known about the control of decomposition by quality variables: that the relative importance of quality parameters in controlling decomposition changes as decomposition progresses (Berg and Staaf, 1987; Heal et al., 1997) and specifically that nutrients are of greater importance as a controlling factor in the early stages and lignin plays a greater role in later stages. The fact that most microbial groups showed correlations with soil mineral N after three weeks of decomposition suggests that their early response to substrate quality was mainly indirect through the modification of soil nutrient content due to the quality of the added substrate. Our findings then suggest that the effect of substrate quality on mineral soil biota is indeed mediated by its effect on decomposition; in other words, that quality affects biota by affecting decomposition and therefore what gets released into the soil or remains a component of the decomposing material.

From PCA and correlation analysis results it was possible to identify some patterns in the response of individual microbial groups to substrate quality parameters. Fungi and Gram-negative bacteria in general responded positively to high nutrient content in the added substrate and mineral nitrogen in soil and negatively to %C, C/N and lignin content. Gram-positive bacteria and actinomycetes were stimulated by high %C, % lignin and the lignin-to-nitrogen ratio. A positive response of the soil microbial biomass to mineral nitrogen has been documented before (Hart and Stark, 1997). Our results suggest that this is due to the response of only some members of the microbial community.

The fact that Gram-negative bacteria showed stronger and longer-lasting relationships with more substrate quality parameters than the Gram-positive groups is consistent with the findings of Waldrop and Firestone (2004) who showed that Gram-negative organisms are more responsive to the addition of substrates and with Kramer and Gleixner (2006) who found that Gram-negative bacteria prefer recent plant material over soil organic matter. T

Participation Summary

Educational & Outreach Activities

Participation Summary:

Education/outreach description:

Ph.D. Dissertation:

Linking litter quality, soil microbial and faunal communities and soil processes, August 2007, University of Georgia

Publications in preparation:

Carrillo, Y. Ball, B, Molina, M. Chemical quality of litter as a driver of detrital community assemblage in mineral soil. To be submitted to Soil Biology and Biochemistry.

Carrillo, Y., Ball, B. Jordan, C. Mediation by soil fauna of the effect of litter quality on nitrogen mineralization. To be submitted to Oikos.

Carrillo, Y., Jordan, C. Ball, B. Modeling the effect of the interaction of soil community structure and plant litter quality on C and N mineralization. To be submitted to Oikos.

Papers presented at conferences:

Carrillo, Y., Ball, B., Jordan, C. Chemical quality of litter as a driver of detrital community assemblage in mineral soil. Ecologial Society of America/Society for Ecological Restoration Joint Meeting, San Jose, California August 5-10, 2007

Carrillo, Y., and Ball, B., Biochemical quality of litter as a driver of detrital community assemblage in mineral soil. 11th International Conference of the Soil Ecology Society. Moah, Utah, May 2007.

Carrillo, Y., Organic matter and soil organisms and why they matter. Georgia Organics Conference, Douglas, Georgia, March 2007

Project Outcomes

Project outcomes:

Overall, these findings demonstrate that the interplay of the quality of organic amendments and the structure of the soil community has the potential of influencing the functioning of an agroecosystem and should be taken into consideration when making management decisions. One of the impacts of this general finding as well as the other specific ones is the increased basic understanding of the interaction of two of the most important factors regulating the process of mineralization of nitrogen from organic materials, which is of crucial importance in sustainable agricultural efforts. Some of the findings of this study have already been presented in scientific meetings (see publications and outreach section) and it is expected that three publications will result from these results. Several working hypotheses were proposed and these might potentially be addressed by future research.

Another impact of this project involves the applied implications of our findings. Some general recommendations for enhancing the role of the soil biota in nitrogen mineralization were presented to growers during the Georgia Organics meeting in 2007. We expect to participate in the coming meeting as well. General recommendations for the selection of organic amendments have also been discussed with the Full Moon Farm growers in Athens, Georgia and it is expected that through them and the Organic Matter Management Working Group (funded by Southern SARE) that they lead these recommendations will reach the community of organic farmers in Northeast Georgia. In addition, there has been some discussion of potentially including some of these results in a brochure published by the Cooperative Extension Service of the College of Agriculture and Environmental Sciences at UGA. In general, we expect that this work will increase awareness among growers and researchers of the importance of considering the interactions between the community of soil organisms and the quality of soil organic amendments for determining the availability of nitrogen for plants in organic production systems.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.