In this study, we sought to link short-term cover crop C and N cycling to microbial-mediated turnover and subsequent stabilization of below- and aboveground cover crop inputs across various soil microenvironments, operationally defined as soil organic matter (SOM) fractions. Few differences were found in the 13C-PLFA community profiles and bacterial abundances across the three cropping systems and among the microenvironments. Results from this study corroborate the hypothesis that belowground C input plays a larger role than residue C in SOM stabilization; but the hypothesis that both residue and root C are preferentially stabilized in microaggregates by enhanced microbial processing was not corroborated.
To promote the long-term management of agroecosystems as carbon (C) sinks for anthropogenic emissions of CO2, mechanisms of C sequestration need to be better understood and subsequently managed. For example, definitive estimates of the relative contributions of roots versus residues to soil organic C pools in cropping systems are still lacking. Aboveground and belowground plant decomposition dynamics are fundamentally distinct in terms of soil conditions (e.g., temperature and moisture) and interactions with soil microbial systems (Rasse et al. 2005). The influence of roots on soil organic C pools could be relatively greater than the influence of aboveground C inputs because active roots are continuously releasing a range of organic compounds (i.e., mainly carbohydrates, carboxylic acids, and amino acids) into the rhizosphere (Oades 1978; Michulnas et al., 1985). Yet, belowground inputs are generally assumed to parallel aboveground net primary productivity. Recent papers have shown roots to contribute more to soil C and have suggested that roots decay in soil slower than aboveground plant parts, due to the intrinsic biochemical composition and/or conditions for decomposition for belowground biomass (Balesdent and Balabane 1996; Puget and Drinkwater 2001).
Plant roots also influence soil organic matter (SOM) stabilization via their role in soil aggregate dynamics. In the rhizosphere, roots physically enmesh soil particles, while exudates stimulate microbial biomass that in turn synthesizes polymers that act as binding agents (Tisdall and Oades, 1979; Jastrow et al., 1998). Gale et al. (2000) and Puget and Drinkwater (2001) have reported the potential importance of plant roots and exudates to the stabilization of aggregate-associated SOM. Still, few studies have investigated the mechanisms of residue- versus root-derived C stabilization in situ mainly because of the difficulties involved in quantifying belowground biomass production, turnover, and exudation.
Many studies have demonstrated that the soil microbial community is a key regulator of soil C and N cycling. Plant roots, exudates, and sloughed off root material, which are major sources of C and N input to soil, are readily available C and N sources for microorganisms. Microbial immobilization of rhizodeposit-derived C and N, followed by the release of microbial biomass C and N with cell death, and subsequent release of mineral N is an ongoing process during cover crop growth (Coleman et al. 1978). Hence, microorganisms preferentially colonize the rhizosphere, making the rhizosphere an area of intense activity with specific biological, chemical, and physical characteristics (Lynch 1990).
Recently, soil microaggregates-within-macroaggregates have been shown to be microenvironments within the soil structure, where SOM is preferentially stabilized (Denef et al. 2004; Kong et al. 2005). The physico-chemical characteristics of soil aggregates, especially microaggregates, in conjunction with the high concentration of C and N in the rhizosphere determine the distribution and activity of microbial populations. In turn, as the abundance of microorganisms utilizing rhizodeposits increases during cover crop growth, different microenvironments form as a result of interactions between minerals and microbial-processed rhizodeposits (Harris et al. 1963). This perpetuating cycle is expected to facilitate greater microbial C turnover and stabilization as well as more efficient N cycling (i.e., greater mineral N availability and reduced N loss) in the microaggregate of the rhizosphere. Furthermore, changes in physical and chemical soil properties that occur as a result of agricultural management, and subsequent changes in microbial community composition, in turn, influence soil processes (Schimel 1995).
Few studies have elucidated the underlying mechanisms of the short- and long-term effects of cover crops on the relationship between soil structural properties, nutrient cycling, and microbial community structure and function. Hence, the main objectives of this study were:
Objective1: To elucidate the role of aggregates in stabilizing residue-derived C compared to root-derived C across different cropping systems.
Objective2: To identify and quantify the microbial communities (e.g., nitrifying and denitrifying populations) associated with overall C stabilization and N cycling within soil microenvironments across different cropping systems.
Experimental Site and Design
The field study took place at the Russell Ranch experimental site (Davis, CA, USA; 38°32’24” N 121°52’12” W), which is located in a region characterized by wet winters and hot, dry summers. Two soil types dominate the site: i) Yolo silt loam (fine-silty, mixed, nonacid, thermic Typic Xerothent) and ii) Rincon silty clay loam (fine, montmorillonitic, thermic Mollic Haploxeralf). Our field study was conducted during the 2006 maize growing season in three maize-tomato cropping systems: conventional (synthetic fertilizer only), low-input (alternating synthetic fertilizer and cover crop) and organic (composted manure- and cover crop) (Table 1). These maize-tomato cropping systems were arranged in a complete randomized design with three, 0.4-ha replicates, receiving furrow irrigation, and under conventional tillage. Hairy vetch (Vicia dasycarpa), a leguminous plant commonly used as a winter cover crop in temperate agroecosystems, was broadcast-sown within each of the cropping treatment replicates at the end of October 2005.
Cover Crop 13CO2 Pulse-Labeling and Field Operations
To address Objective1, two microplots (1.0 m x 1.0 m) were established in each treatment replicate, shortly after cover crop germination. At five weekly events between March 26th and April 26th, 2006, one of the two plots was enclosed with a portable chamber and pulse-labeled with 13CO2 (99 atom%), for a cumulative total of 6.5 L per labeled microplot. The portable 13C-labeling chambers consisted of a vinyl sheet (TAP Plastics, Sacramento, CA) fitted around a polyvinyl chloride frame (height adjustable), with excess vinyl that was fixed against the contours of the soil surface to serve as a seal around the edges of the frame. During the labeling procedure, total CO2 concentration within the labeling chamber was monitored using a portable infrared gas analyzer (Qubit CO2 Analyzer, Model S-151, Qubit Systems, Kingston, Ontario, Canada). 13CO2 was injected into the chambers to obtain a theoretical increase of 750 ppm in CO2 concentration and then the chamber was removed when the initial CO2 levels in the chamber dropped below 250 ppm (usually after 30 to 45 minutes). 13C-labeling events usually took place between 1100 and 1300 h. To maximize 13CO2 plant uptake, the labeling chambers were replaced over the experimental plots at sunset after each labeling event (~1700 h) to capture overnight 13CO2 respiration and then removed the following morning (~800 h) after CO2 levels in the chamber dropped below 250 ppm. Air temperatures in chambers during 13C-labeling averaged 25° to 30°C. At the final two labeling events, temperatures within the chambers reached 38°C, which was 8°C above the ambient air temperature.
In the final week of April, the aboveground biomass of both microplots were mowed to approximately 2cm long pieces and roto-tilled into the microplots to a depth of 15cm, which also resulted in the chopping and mixing of the root biomass within the microplots. The 13C-labeled aboveground biomass was incorporated into the unlabeled microplot (AG microplot) and the unlabeled aboveground biomass was incorporated into the microplot with 13C-labeled roots (BG microplot) in the same manner. This exchange led to one microplot with labeled aboveground residue and the other microplot with labeled roots. Prior to incorporation, aboveground biomass was weighed and subsampled for elemental and isotopic C concentrations as well as to determine dry matter.
Maize was direct-seeded into the conventional and then the low-input and organic plots in the second and final weeks of May, respectively. Subsequent simulations of field operations (cultivation, weeding, etc.) were done by hand within the microplots.
Estimation of Belowground Biomass and C Contributions
Before cover crop incorporation, soil cores (4cm diameter; 0-15 cm) were collected from each AG and BG microplot to determine belowground standing biomass and both labeled and unlabeled root delta-13C values (δ13C). Field moist soil (three replicates ber microplot) was extracted from the cores and large visible roots were immediately removed. Approximately 100g soil was subsampled from the soil core and suspended in ~200ml deionized water. The slurry was stirred by hand and the roots that separated from the soil were removed and placed in petri dishes on ice. Visual criteria, such as color and elasticity, were used to distinguish cover crop roots from roots of other plants. The remainder of the core was subsampled and processed for roots in the same manner as above. The collected roots were rinsed under a gentle stream of deionized water to removed soil and then dried at 50°C in paper envelopes. The δ13C values measured for these roots were used as the enrichment values for root-derived C. The roots collected from the cores were used to extrapolate the root biomass for the microplot to a depth of 15cm. We acknowledge that these procedures likely excluded very fine roots and overlooked dynamic C contributions from root exudates, and that total-root derived inputs were likely underestimated.
Soil Sample Collection
Two soil cores (4cm diameter; 0-15 cm) were collected October 27th, 2005 (Time-Zero), and then composited and subsampled for baseline δ13C values and total C levels, before the addition of 13C-enriched material. To address Objective2, four soil samples (4cm dia.; 0-15 cm) were taken from each of the 1-m2 microplots throughout the maize growing season that followed the cover crop growing season. Over the course of the 2006 maize growing season, soil core samples were collected from both AG and BG microplots within the conventional, low-input, and organic cropping systems on May 31st (First), July 10th (Second), August 10th (Third), and at harvest (September 13th; Final). Field-moist soil samples were gently broken apart and stored at -20°C until further processing and analyses. Bulk density was determined on an individual soil core basis.
Soil Aggregate Separation
From each soil sample, 30g subsamples were fractionated into three SOM fractions, coarse particulate organic matter (CPOM; >250µm), microaggregate (53-250µm), and silt-and-clay (<53µm), according to the methodology outlined in Six et al. (2000). Frozen soil samples were thawed for 20 min. and submerged in deionized water, at room temperature for 5 min., to slake the soil. Water-stable SOM fractions were obtained using a microaggregate isolator in which the slaked soil was immersed in deionized water on top of a 250µm mesh screen and gently shaken with 50 stainless steel beads (4mm diameter) until only CPOM) and sand are retained on the 250μm mesh screen. During shaking, a continuous and steady stream of water flows through the device to ensure that microaggregates are immediately flushed onto a 53µm sieve and are not exposed to any further disruption by the beads. The material on the 53µm sieve was manually sieved to separate water-stable microaggregates from silt and clay particles. All fractions were collected as a soil suspension, which were centrifuged at 4°C at 5,000 rpm for 15 min. (Sorvall RC-5C Plus Superspeed centrifuge, Thermo Scientific). The supernatant was discarded and the remaining soil was lyophilized and stored at -20°C until further analysis.
Phospholipid Fatty Acid Assay
13C-Phospholipid fatty acids (13C-PLFAs) were extracted and derivatized to fatty acid methyl esters from the microaggregate and silt-and-clay fractions using modifications of the method according to Bossio and Scow (1995). Identification, quantification, and δ13C signature measurement of the PLFAs were determined using a Thermo gas chromatograph-combustion-isotope ratio mass spectrometer (GC-C-IRMS) system composed of a Trace GC Ultra gas chromatograph (Thermo Electron Corp., Milan, Italy) coupled to a Delta Plus Advantage isotope ratio mass spectrometer through a GC/C-III interface (Thermo Electron Corp., Bremen, Germany), with a J&W DB-5 column.
Real-time PCR Quantification of amo A and nos Z genes
Humic acid, a PCR inhibitor, was removed from the whole soil as well as the SOM fractions as follows: 1g of soil was incubated in 2mL of 0.1% Na4P2O7 in 10mM Tris-HCl buffer (pH 8.0)-1mM ethylenediaminetetraacetic acid (EDTA), at room temperature for 30 min. and then centrifuged at 8,500xg for 10 min. at room temperature. The supernatant was discarded and total DNA was extracted from 0.5g humic-washed soil using a Bio 101 Fast DNA SPIN kit for soil (QBiogene, Irvine, CA). The yield of DNA extracted was quantified with a Qubit flourometer (Invitrogen, Carlsbad, CA). DNA concentrations are expressed on a dry soil basis.
To measure changes in total denitrifying bacteria within the soil microenvironments, we employed the SYBR Green based polymerase chain reaction (PCR) primers targeting the nos Z gene, as described previously in Henry et al. (2004). Thermal cycling, fluorescent data collection, and data analysis were carried out with the ABI Prism 7300 sequence detection system according to the manufacturer’s instructions using SYBR Green based detection. Real-time PCR for nos Z contained 5 uL of 1:100 dilutions of DNA extracts from SOM fractions, as a template, 0.5 µL of 12.5 µM stock of both the forward and reverse primers, 10 µL of 2x ABI Power SYBR green master mix, 3.4 µL of nano-pure water to give a total of a 20 µL reaction. The primers (5’-3’) used to detect the nos Z gene were nosZ2F (CGC RAC GGC AAS AAG GTS MSS GT) and nosZ2R (CAK RTG CAK SGC RTG GCA GAA) (Henry et al. 2004). The conditions for nos Z quantification were as follows: 6 cycles of 15 s at 95˚C for denaturation, 30 s at 65˚C for annealing, 30 s at 72˚C for extension, then 40 cycles consisting of 15 s at 95˚C for denaturation, 15 s at 60˚C for annealing, 30 s at 72˚C for extension and 30 s at 83˚C, which served as the data acquisition step, then finally, a one cycle of 95˚C for 15 s, 60˚C for 30 s and 95˚C for 15 s was added to obtain a specific denaturation curve.
To quantify the ammonia oxidizing bacteria (AOB) in the DNA extracted from whole soil and SOM fractions of the 1st sampling event, we used PCR primers to target the functional gene, ammonia monoxygenase (amo A), which encodes for an enzyme involved in NH4+ oxidation. Thermal cycling, fluorescent data collection, and data analysis are carried out with the ABI Prism 7300 sequence detection system according to the manufacturer’s instructions using SYBR-green based detection. Real-time PCR for amo A contained 5 uL of 1:100 dilutions of DNA extracts from SOM fractions, as a template, 0.3 µL and 0.9µL of 5 µM stock of the forward and reverse primers, respectively, 10 µL of 2x ABI Power SYBR green master mix, 0.2 µL of nano-pure water to give a total of a 20 µL reaction. The primers (5’-3’) used to detect the amo A gene were A189 (GGH GAC TGG GAY TTC TGG) and amoA-2R’ (CCT CKG SAA AGC CTT CTTC) (Okano et al. 2004). The conditions for amo A quantification were as follows: 10 min. at 95˚C for enzyme activation as recommended by the manufacturer (Applied Biosystems), then 40 cycles consisting of 15 s at 95˚C for denaturation, 30 s at 55˚C for annealing, 31 s at 72˚C for extension, which served as the data acquisition step, then finally, a one cycle of 95˚C for 15 s, 60˚C for 30 s and 95˚C for 15 s was added to obtain a specific denaturation curve. Purity of amplified products was checked by the observation of a single melting peak (~85˚C) and the presence of a unique band of the expected size in a 2% agarose gel, stained with ethidium bromide.
For both the amo A and nos Z real-time PCRs, the cycle number (CT) that corresponds with fluorescence readings above a threshold is proportional to the starting amount of template DNA. Copy numbers of the amo A and nos Z genes in SOM fractions were determined with an external standard curve generated by a 10-fold dilution series of either cloned amo A or nos Z genes into plasmids. The following equations were used to convert amo A and nos Z real-time PCR results, respectively:
CT = -3.37 x Log10(amo A copies reaction-1) + 41.47 and R2= 0.98
CT = -3.35 x Log10(nos Z copies reaction-1) + 38.67 and R2= 0.99
The AOB and total bacterial population sizes presented here are corrected for efficiency of extraction from soil.
Elemental and Isotopic C Analyses
Subsamples of cover crop above- and belowground biomass, whole soil samples, and SOM fractions were ground and analyzed for elemental and isotopic C concentrations, using a PDZ Europa Integra C-N isotope ratio mass spectrometer (Integra, Germany). Results were expressed as
where 13R = 13C/12C, and the standard is the Pee Dee Belemnite (PDB). Because the soil samples did not react (i.e., evolve CO2) upon addition of 12M hydrochloric acid, we concluded that the whole soil samples were free of inorganic C; therefore, the total C concentrations that were measured were considered equivalent to organic C concentrations. For whole soil and aggregates, the proportion (f) of soil C derived from either 13C-labeled above- or belowground biomass was calculated using the δ13C of the 13C-enriched residue or roots, respectively, against the 13C natural abundance samples in the equation:
, (Equation 2)
where 13Csample = δ13C for the AG or BG soil sample of interest, 13Clabeled material = δ13C for above- or belowground biomass, 13C natural abundance = δ13C of the equivalent soil sample taken at the Time-Zero sampling event. Total soil-, SOM-fraction, and PLFA-C concentrations for the measured variables were multiplied by f to obtain Cnew, the concentration of C derived from either 13C-labeled residue or root biomass. All elemental and isotopic C measurements for the soil samples were converted to a m2 basis using bulk density measurements. Above- or belowground cover crop biomass-C recoveries in the SOM fractions were calculated as the ratio of the soil fraction-Cnew to the total amount of 13C-labeled applied.
Data and Statistical Analyses
The proportion of whole soil- and SOM fraction-C not derived from either reside- or root-13C [i.e., Cold = 1 – Cnew] was used to calculate the mean residence times of the whole soil and SOM fraction-C pools. Mean residence times were calculated by taking the reciprocals of the estimates of rate constants (k), obtained from the following first-order decay equation:
where At = proportion of Cold at the final sampling event, A0 = proportion of Cold in the sample at the Time-Zero sampling event, k = decay rate constant, and t = time between the final soil sampling event and the incorporation of the 13C-aboveground [t = 134 days] or belowground biomass [t = 171 days]. The differences in SOM fraction-Cold concentration between the Time-Zero soil sampling time and the time the 13C-labeled biomass was incorporated into the soil were assumed to be negligible.
To compare differences in whole soil and SOM fraction C and Cnew as well as AOB and nitrifier abundances measurements among the soil sampling events for each cropping system, repeated measures analyses were performed using the PROC MIXED procedure of the Statistical Analysis System (SAS; SAS Institute, 2002). The data were analyzed as a complete randomized design. In the model, cropping system was the main variable and sampling event (i.e., time) was the repeated variable. Differences between means were calculated based on least significant difference tests, with the PDIFF option of the LSMEANS statement. All differences discussed were significant at the p
Baseline C concentrations
Over the course of the season, total soil- and SOM-fraction C and N did not change, hence, only C and N concentrations for whole soil and for SOM fractions measured at the Time-Zero sampling event are listed in Table 2. Both the total C and N concentrations of the organic system (24.0 Mg SOC ha-1 and 2.15 Mg N ha-1) were greater than soil C and soil N levels in the conventional (19.1 Mg SOC ha-1 and 1.49 Mg N ha-1) and low-input systems (19.7 Mg SOC ha-1 and 1.49 Mg N ha-1) (p
Above- and Belowground 13C and Total C inputs
Standing aboveground biomass averaged 554 g m-2 and did not differ from biomass measured outside the microplots, thereby suggesting that the labeling procedure did not affect aboveground cover crop biomass yield. At the time of cover crop incorporation, belowground cover crop root biomass was estimated to be 153 g m-2. Cover crop residue and roots were incorporated at rates of 2.33 and 0.44 Mg C ha-1, respectively, with a root-to-residue biomass C ratio of 0.19.
Residue- and Root-C Contributions to Whole Soil and SOM Fractions Across the Season
Residue- and root-derived C (Cnew) was detected in all cropping systems, with greater belowground C than aboveground C distribution across the season (Fig. 1). Although similar within AG and BG microplots, Cnew dynamics differed among the cropping systems, in that the greatest Cnew concentrations for conventional, low-input, and organic systems were found at the First, Second, and Third sampling events, respectively. At the end of the season, C contributions from the roots constituted ~59% of the total belowground C contributions (i.e., kg Cnew ha-1 / kg cover crop-C ha-1), whereas 6% of the total aboveground C input was recovered as residue-C. Moreover, we found a cropping system*C input (i.e., below- versus aboveground) effect on Cnew, where the root-derived C in the organic system was highest, root-derived C in the low-input system was intermediate, and the lowest values were observed for the root-derived C in the conventional and residue-derived C for all the systems, at the final sampling event.
No differences due to above- or belowground C were found for neither the CPOM nor silt-and-clay fractions (Fig. 2). The high variability inherent with 13C measurements on small samples such as CPOM may have precluded the ability to detect any differences across the sampling times and among the cropping systems. We found more root-derived than residue-derived microaggregate Cnew across the season, but did not detect a cropping system effect.
The relative retention of root C compared to residue C by whole soil and SOM fractions was evaluated using the ratio of root-derived over residue-C concentrations (Figs. 3a and 3b). By the end of the season, this ratio was highest in the organic system (6.27), intermediate in the low-input system (3.07), and lowest in the conventional system (1.25) (Fig. 3a; p < 0.1). A similar trend was observed in the silt-and-clay fraction; however, the relative retention of root C compared to residue C was highest in the microaggregate fraction of the low-input system and similar between the conventional and organic systems (Fig. 3b). The ratios for the microaggregate and silt-and-clay fractions were not different for the conventional and low-input systems, while more root-derived C dominated silt-and-clay-associated Cnew than microaggregate-associated Cnew of the organic system.
Using whole soil and SOM fraction samples collected at the Final sampling event, we quantified the turnover of both residue- and root-derived C in each of the cropping systems. Mean residence times of root-derived whole-soil C were ~27% longer than that of AG, but did not differ among the cropping systems (Table 3). A significant cropping system*SOM fraction effect was found for the MRT’s calculated for the SOM fractions. Turnover of the CPOM in the conventional system (88.5days) was the slowest while the fastest turnover was associated with the silt-and-clay fractions in the low-input (34.4 days) and organic systems (31.8 days). Root versus residue did not affect SOM fraction-C turnover.
13C-PLFA in Microenvironments
Microbial biomass was estimated using the PLFAs extracted from the microaggregate and silt-and-clay fractions. Few differences were found across the season, thus, microbial biomass for only the First sampling period is reported (Fig. 4). Across all three systems, microbial biomass generally decreased according to the following: organic > conventional > low-input. Due to high variability, no differences were found in total biomass between the soils AG and BG plots (Fig. 4). Nevertheless, total biomass of samples enriched with 13C-labeled belowground biomass was highest in the organic system (46.1 nmoles g-1 dry soil), intermediate in the conventional system (41.6 nmoles g-1 dry soil), and lowest in the low-input system (33.7 nmoles g-1 dry soil).
Principal component analysis (PCA; CANOCO version 4.0, Microcomputer Power, Inc. Ithaca, NY) of PLFA-13C extracted from microaggregate and silt-and-clay samples clearly distinguished microbial communities by cropping system, C input, and microenvironment. Between two axes, 27.2% of the variation was explained and all the samples were separated based on cropping system (e.g., conventional, low-input, and organic; Fig. 5). Upon analysis of samples from only the Final sampling event (which was similar to other sampling events), nearly 55% of the variation was explained by cropping system and C input (Fig. 6). Meanwhile, PCA of 13C-PLFA from only the conventional system, taken from the Final sampling event, indicated that PLFA profiles of AG samples are dissimilar to PLFA profiles from BG samples (Fig. 7). Axes 1 and 2 for the latter explained 56.1 and 40.0% of the variation, respectively, and separated the samples by both SOM fraction (microaggregate versus silt-and-clay) and the origin of cover crop biomass.
Ammonia Oxidizing Bacteria (AOB) and Denitrifier Abundance
The abundance of AOB and denitrifying bacteria in the whole soil neither differed among the cropping systems nor differed as a function of sampling time (data not shown). Morever, no effect of cover crop root or residue C was detected for AOB and denitrifier abundances. Hence, copy numbers for both amo A and nos Z for both AG and BG plots were composited and will be reported as an averaged value. At the First sampling event, mean amo A copy numbers in the CPOM and microaggregate fractions were 2.47 x 103 and 2.62 x 103 copies g dry soil-1 (Fig. 8). No AOB were not detected at the Time-Zero sampling event, nor in the in the silt-and-clay fractions at the First and Final sampling events (Fig. 9). Assuming an average amo A copy number of 2.5 AOB cell-1 (Okano et al. 2004), our results translate to means of 9.87 x 102 and 1.05 x 103 AOB cells g dry soil-1 in the microaggregate and silt-and-clay fraction, respectively.
Measurements of the abundance of denitrifying bacteria at the First sampling event were on the same order of magnitude as the AOB and not different among cropping systems (Fig. 9). The CPOM fraction in both the conventional (3.00 x 103 copies g dry soil-1) and organic (3.20 x 103 copies g dry soil-1) systems at the Time-Zero sampling had the greatest copy numbers of nos Z compared to the microaggregate and silt-and-clay fractions (Fig. 9). Denitrifier abundances in the silt-and-clay fraction (~101 copies g dry soil-1) in the three cropping systems were consistently one and two orders of magnitude lower than that of the microaggregate and CPOM fractions, respectively.
Below- versus Aboveground C inputs
Belowground C inputs have often been assumed to parallel aboveground net primary productivity. However, we found the ratio of standing root biomass to aboveground cover crop biomass, 0.19, was very similar to the 0.20 value reported in Puget and Drinkwater (2001). Although over five times more residue C was incorporated into the three cropping systems, approximately 59% and 6% of root versus residue C, respectively, remained as whole-soil C. The latter was similar to results shown by Gale et al. (2000), who found ~50% of root-derived C in soil versus only 13% of shoot-derived C, also for hairy vetch. Furthermore, Balesdent and Balabane (1996) reported that, although the estimated aboveground (345 g C m-1yr-1) was higher than the belowground (152 g C m-1yr-1) corn residues, the latter contributed more to the SOM pool (57 g C m-1yr-1) than did the aboveground (36 g C m-1yr-1) corn residues. The greater contribution of roots versus residue to SOM may be due to the greater contact between root-derived material and the soil matrix, which Balesdent and Balabane (1996) and Oades (1995) have shown can to lead faster stabilization of root-derived C on clay surfaces and within aggregates. In addition, during cover crop growth, root exudates and sloughed off root material are readily available C sources for microorganisms, whereas, aboveground residue are only in contact with soil particles and microorganisms after incorporation and senescence.
In their review, Rasse et al. (2005) estimate that the relative contribution factor, [expressed as the ratio: (root-derived soil C/total root C input)/(shoot-derived soil C/total shoot C input)] was on average 2.4 for various in situ studies of different root systems. If calculated according to the Rasse et al. (2005) study, the relative contribution factor for our study was ~18. The shorter duration of our experiment and the use of 13C in situ labeling compared to the other studies reported in Rasse et al. (2005) may have led to our higher estimate of the relative contribution factor. However, calculating the ratio of root-derived C over residue-derived C, without normalizing for the respective C inputs to our systems, gave a value 2.3, which is similar to the relative contribution factor given in Rasse et al. (2005).
Despite greater belowground C stabilization compared to aboveground C across the systems, more newly stabilized C originated from root C than aboveground C in the organic system (Fig. 3a). In contrast, root and residue C contributed equally to Cnew in the conventional system. For the conventional and low-input systems, the relative contributions of root and residue C to microaggregate and silt-and-clay C were similar. Contrastingly, the ratio of root-derived over residue-derived C was higher in the silt-and-clay fraction than in the microaggregate fraction of the organic system (Fig. 3b). Hence, our results associate a greater potential for root C than residue C stabilization for the silt-and-clay fraction compared to the other SOM fractions in the organic system.
Long-term Crop Management Effects on SOC and Soil N sequestration
Increased N fertilization has been correlated to increased SOC sequestration (Campbell et al. 1991; Dumanski et al. 1998). Despite receiving a relatively high N fertilization rate and producing the largest maize yields and vegetative biomass (i.e., greater maize stover returned to the system), the conventional system showed neither the greatest SOC nor soil N stocks of the three cropping systems after 13 years of cropping. This supports findings from recent studies showing that, while synthetic fertilizer N may increase crop residue returns, N fertilization has a net negative effect on SOC sequestration (Omay et al. 1997; Halvorson et al. 2002; Russell et al. 2005). Moreover, this data imply that fertilizer (e.g., synthetic versus organic) plays a role in long-term SOC and soil N sequestration. The conventional and organic systems both received high rates of N additions (280 and 473 kg N ha-1yr-1, respectively), yet the organic system, where solely organic amendments were applied, sequestered disproportionately more SOC and soil N, than the amount of C and N input it received compared to the other systems, after 13 years of crop management. The greater long-term protection and stabilization of C and N derived from the cover crop and composted manure within aggregate structures may have fostered the gradual accumulation of a large pool of soil organic matter in the organic system compared to the conventional and low-input cropping systems.
Mechanisms of C stabilization in whole soil and SOM fractions
In a previous study, we observed that the organic cropping system disproportionately accumulated soil C relative to its C input level (Kong et al. 2005). The organic system received 1.7 times more annual C additions in the form of crop residues, cover crop biomass, and composted manure than the conventional cropping system; yet, the organic system had an annual C sequestration rate 14 times greater than the conventional system, where the C inputs consisted only of crop residues. We surmised that, in the organic system, along with other cover crop-based systems, the growth of a cover crop enhanced the rate of C sequestration. Data in this study provided a possible mechanism for the substantially higher total soil C levels of the organic system compared to the conventional and low-input systems. In this study, while above- and belowground standing biomass C were similar among all three cropping systems, more root-derived C was recovered in the organic system, reflecting the long-term differences in soil C among the systems. Newly added C from cover crop roots was highest in the organic system and lowest in the conventional system (Fig. 1). This suggests that the organic system preferentially stabilizes more belowground cover crop C inputs than the other systems, a possible explanation for its disproportionate sequestration of C.
Mean Residence Times
Based on the datasets of several studies, Rasse et al. (2005) report that the mean residence time of root-derived C is 2.4 times longer than that of shoot-derived C in soils. The latter estimate is on the same order of magnitude as that for our study, which was 1.4. Faster turnover of microaggregate and silt-and clay support the greater stabilization of these fractions into their respective systems.
Microbial Community Structure and Activity in Microenvironments
The structural organization of soil particles provides a spatially heterogeneous habitat for microorganisms characterized by different substrate, nutrient, and oxygen concentrations and water contents as well as variable pH values. Several studies have shown that soil aggregates represent an ecological niche whose chemical and physical properties may contribute to the heterogeneous distribution of microorganisms and their activity among aggregates of different sizes. Different population numbers and structure of denitrifiers (Seech and Beauchamp 1988, Philippot et al. 1997), rhizobia (Mendes and Bottomley 1998), and diazotrophic communities (Poly et al. 2001) among aggregate size classes have been reported. Our 13C-PLFA community fingerprints illustrate that different microbial communities process cover crop above- versus belowground biomass (Fig. 6). Our 13C-PLFA analyses also suggest that microaggregates and the silt-and-clay fraction are distinct microenvironments that support microbial communities, which process C dissimilarly (Fig.7). More research is necessary to elucidate whether this discrepancy in microbial communities associated with the processing of above- versus belowground biomass may be related to the differences observed in rate and amount of belowground versus aboveground biomass-C stabilization.
We expected that long-term (>10 years) cropping management would create a distinctive microbial community structure and population size that would lead to a differentiation in N cycling and storage. The annual cover cropping and manure amendments characteristic of organic cropping systems were expected to produce a more abundant, active, compositionally diverse, and resilient community of soil microorganisms (Gunapala et al. 1998). However, not only did the ammonium-oxidizing bacteria population size in the whole soil not increase over the course of this field study, but no differences were detected between the different cropping systems. Measurements at the whole soil level suggest that different types of N amendments (e.g., NH4+, organic N, or NO3-) do not have a long-term effect on AOB population size compared to the differences observed in annually fertilized versus unfertilized soil, even when AOB populations were quantified eight months after fertilization (Okano et al. 2004).
Sessitsch et al. (2001) found that bacterial community structure was affected to a greater extent by the particle size fraction than by the kind of fertilizer applied to the long-term cropping systems in their study. Similarly, the effect of SOM fraction size class was more significant than that of cropping system in our data. Interestingly, AOB were more abundant at the First and Final sampling events, whereas denitrifier populations were largest at Time-Zero and First sampling events. This indicates that the potential for ammonification was higher later in the season whereas the conditions for denitrification are more ideal at the start of the growing season. We failed to corroborate our second hypothesis because abundances of neither AOB nor denitrifiers were highest in the microaggregate structures compared to the CPOM and silt-and-clay fractions.
Educational & Outreach Activities
Kong, A.Y.Y., K.M. Scow, K.R. Hristova, and J. Six. Oral presentation. 2008. Linking C and N Cycling to Microbial Function Within Soil Microenvironments in Cover Crop Systems. Soil Science Society of America Meeting. Houston, TX.
Kong, A.Y.Y., K.M. Scow, K.R. Hristova, and J. Six. Poster presentation. 2007. Linking N Cycling to Microbial Function Within Soil Microenvironments in Cover Crop Systems. American Geophysical Union Meeting. San Francisco, CA.
With mounting scientific evidence linking anthropogenic activity (e.g., combustion of fossil fuels) to the rise in greenhouse gas emissions and climate change, it has become more evident that cropping system management must be improved to mitigate global warming. Therefore, it is more pertinent that the scientific community elucidates the spatial and temporal dimensions of soil C and N cycling in order to develop sustainable crop management practices that improve environmental quality, optimize the C sequestration within agricultural lands, and achieve N synchrony for maximum crop yields. Total soil C and N data collected from this study have been incorporated into regional greenhouse gas modeling exercises as a part of a larger modeling effort by biophysical modelers and economists at the University of California, Davis. The models produced from these collaborations will help both policymakers and land managers to develop best management practices to reduce greenhouse gas emissions from California croplands. At the local management level, our data contribute to a more complete understanding of the mechanisms governing nutrient cycling in cover crop systems as well as the potential role over cover crops in profitable and sustainable farming practices.