Managing Soil Biota in Low-Input and Organic Farming Systems to Enhance Soil Fertility

Objective 1 The light fraction (LF) of soil organic matter, which is an indicator of available carbon and nitrogen, was measured in organic, low input, and conventional tomato plots. The amount of variability in the LF measurements from within each farming system was higher than any differences between farming systems, largely because the LF pool was so small in all soils. Therefore we decided to discontinue these measurements and focus on other means to estimate available carbon (see below).

Objective 2 We broadened this objective to the study of seasonal fluctuations in microbial communities in different farming systems and of the influence of crop type on microbial communities. We found that phospholipid fatty acids (PLFA), found primarily in cell membranes of living organisms, extracted directly from soil provides information on microbial communities in several ways. The sum of all lipids is an estimate of total biomass, specific lipids are biomarkers for certain microbial groups (e.g., bacteria vs fungi, aerobes vs anaerobes), and all lipids together can be used to provide a unique fingerprint for a given soil. A PLFA fingerprint consists of the percentages of different fatty acids detected as peaks in gas chromatograms from a soil extract. Fatty acid percentage values can be entered as multivariate data in principal component analysis to assess similarities among different soil samples. (The PLFA work was co-funded by a grant from the USDA NRI Soil and Soil Biology program).

We tested whether there was consistency in the microbial community composition (by PLFA fingerprinting) of tomato soils at SAFS throughout the season, following tillage and fertilization, at different spatial locations within the field, and within different farming systems. PLFA fingerprints were consistent among field blocks within the same farming system, thus field variability represented by blocks in this study did not have a significant impact on the differences in the observed fingerprints. Communities in organic (cover crop and manure) and conventional (mineral fertilizer only) managed plots were significantly different on most dates. Microbial communities in low input (cover crop and mineral fertilizer) plots were intermediate between the conventional and organic communities. Fungi were greater in relative abundance in organic than conventional soils. PLFA data were related to conventional measurements characterizing microbial populations. Measurements of microbial biomass carbon and nitrogen, substrate induced respiration, basal respiration, and potentially mineralizable nitrogen were not as sensitive to season, farming system, and management practices as were the PLFA fingerprints. A comparison was made of PLFA fingerprints of tomato, corn, safflower, bean and wheat soils all collected from the SAFS plot on a single date in the latter part of the growing season. Though communities under some crops could be distinguished from the others (e.g., wheat, beans), others were more similar to one another (e.g., tomatoes, safflower and corn). Samples from the 2-yr conventional rotation plots were distinctly different from the three 4-yr rotation plots.

To investigate the importance of soil fungi and fungal-feeding nematodes in nitrogen mineralization, soil fungi and nematodes were isolated from SAFS soils and inoculated into microcosm column systems with different C:N ratios of organic substrates. Nitrogen-free sand was amended with appropriate amounts of alfalfa and cellulose, with total N held constant, to create C:N ratios ranging from 11:1 to 45:1. Nitrogen mineralization by Aphelenchus avenae and Aphelenchoides composticola feeding on Rhizoctonia solani and Trichoderma sp. was determined by measuring ammonium and nitrate concentrations in the leachate from the columns at 3-d intervals. Nematode population levels and the fatty acid 18:2 (a fungal biomass indicator), were monitored over a period of three weeks. There were differences between fungal species in the amount of N mineralized in the presence of either nematode species. Average N mineralized per nematode per day ranged from 0.0021 ug to 0.0034 ug. N mineralized by fungal-feeding nematodes feeding increased as the C:N ratios of organic substrates increased. Initial and average nematode population densities were significantly higher in columns containing Rhizoctonia than on Trichoderma. Both nematodes reduced the fungal fatty acid 18:2, the biomarker for fungi.

The fungi, Rhizoctonia solani and Trichoderma sp., and nematodes, Aphelenchus avenae and Aphelenchoides composticola, isolated from field soils, were introduced into alfalfa/cellulose/sand microcosms with C:N ratios of 11:1, 20:1, 30:1 and 40:1. The six treatments included each fungus alone or with each nematode. Fungal biomass was measured by identification and quantification of phospholipid fatty acids (PLFAs) extracted from living cell membranes through gas chromatography. Principal component analysis was used to identify key PLFA patterns associated with fungal colonizers and their nematode grazers. Fatty acid 18:2-omega-6c, a fungal indicator, consisted of 52.2 % + 3.8 of total PLFAs in pure broth cultures, and 10.2% + 2.3 in alfalfa/cellulose/sand for R. solani in the absence of nematodes. Biomass of R. solani, as indicated by fungal PLFAs, was reduced in the early stage of decomposition of the organic matter in the presence of A. composticola, but was not reduced by day 21. There was more N mineralized in the presence of nematodes grazing on R. solani than in their absence. Fungal biomass of R. solani was lower at C:N ratios of > 30:1 than at < 20:1 on day 21, but C:N effects were inconsistent for Trichoderma sp. Objective 3 Conversion of N from incorporated crop residues and manure to available mineral forms requires decomposition and mineralization processes mediated by bacteria, fungi and other components of the soil food web. When carbon is abundant, increases in microbial biomass may result in immobilization of N. In that case, mineralization may be increased by organisms that graze on the primary decomposers. In laboratory studies, bacterial-feeding nematodes mineralized between 1.2 and 5.8 ng N per individual per day, depending on body size, while fungal-feeding nematodes mineralized up to 3.3 ng N per individual per day, depending on host status of the fungal substrate. Since bacterial-feeding nematodes have a higher C:N ratio (± 6:1) than their substrate (± 4:1), considerable mineralization of N is associated with their metabolic activity. In consuming sufficient bacteria to provide the C necessary for their body structure and respiration, nematodes assimilate excess N. The excess N is excreted as ammonia. Fungal-feeding nematodes have a body C:N ratio closer to that of their substrate, so the N associated with C used in body structure is not in excess. However, the N associated with C that is respired exceeds body needs and is excreted. Field observations in the organic and low-input tomato plots of the UC Davis Sustainable Agriculture Farming Systems (SAFS) project indicated that population levels of microbivorous and fungivorous nematodes were low early in the growing season. In those plots, which either do not receive mineral fertilizers (organic) or receive them only when necessary (low-input), symptoms of N deficiency are sometimes seen during the early vegetative growth phase. The goal of this study was to manage the soil food web to enhance population levels of microbivorous and fungivorous nematodes in the spring so that their feeding would release N immobilized in the primary decomposer biomass. To avoid delaying planting of tomatoes until nematode populations responded to spring soil temperatures, we attempted to increase numbers of microbivorous nematodes and other microbial grazers the previous fall. Soil temperature conditions are adequate for nematode populations to increase in the late summer and early fall in northern California, but moisture is a constraining factor. The moisture constraint to soil biological activity was relieved by irrigation and a carbon source was provided in the form of a late summer cover crop. Abundance of microbivorous nematodes, and presumably other organisms at similar trophic levels, increased with the enhancement of biological activity the previous fall. Soil N level increased in the spring with nematode abundance. Subsequent tomato yields also increased when N levels were limiting. A convenient biomarker system is necessary to provide a basis for management and monitoring of the soil food web. From the standpoint of soil fertility, greatest amounts of mineralization probably occur at lower trophic interchanges in the food web, where biomass of organisms is greatest. Members of functional guilds of soil nematodes, classified at the family level, respond similarly to food web enrichment and to environmental perturbation and recovery. Indices derived through nematode faunal analysis provide bioindicators for condition of the soil food web and disturbance of the soil environment. Resolution is enhanced by a weighting system for the indicator importance of the presence and abundance of each functional guild in relation to food web enrichment and structure. Graphical representations of food web structure, based on nematode faunal analyses, allow diagnostic interpretation of its condition (e.g. Fig. 1). Simple ratios of the weighted abundance of representatives of specific functional guilds provide useful indicators of food web structure and enrichment, and of the nature of decomposition channels. These analyses can be used as a basis for management decisions and as a means of monitoring management success. Fig.1 Progression of changes in nematode faunal analyses from dry soil in late August in plots where the soil was irrigated and summer and winter cover crops were grown (dark symbols) or were not irrigated and had no cover crop (light symbols). The star indicates the time at which organic material was incorporated into both plots (April). The soil food web of the dry plots was unable to respond rapidly to the enrichment, while that in the irrigated plots responded immediately. The differences were reflected in availability of soil nitrogen in May and in yields of the subsequent tomato crop. Our results confirm that an abundance of nematodes representing opportunist bacterial- and fungal-feeding guilds indicates a biologically-active soil. When such guilds predominate at planting time, N provided by incorporation of organic material will not remain bound in the primary decomposer biomass but will be mineralized through the grazing activities of the nematodes. Despite extensive knowledge of the physiological, population, and trophic ecology of microbivorous nematodes and other soil fauna, the impact of these fauna on community and nutrient dynamics remains unpredictable. We used a novel “field incubator” technique to study populations of soil fauna and their effect on decomposing organic matter in situ. This technique combines the experimental control of the litter bag approach with more realistic contact between soil and decaying litter. Microbivorous nematodes were more abundant early in the decomposition process and fungal feeders (e.g., Aphelenchus avenae) were more abundant in the presence of high C/N ratio organic matter. Experimental exclusion of microarthropods (mostly mites and Collembola) led to reduced mass loss of leaf litter; it did not affect fungi, the primary prey of microarthropods, or populations of fungal feeding nematodes, their potential competitors. However, microarthropod exclusion was associated with elevated populations of predator/omnivore members of the Dorylaimida. Furthermore, in incubators from which microarthropods were not excluded, they were more abundant under the high C/N conditions that favor fungal decomposition. This suggests that microarthropods and predatory nematodes may be linked either directly via predation or indirectly via other members of the food web. The increase in the Dorylaimida is also interesting because it represents colonization by these supposedly disturbance-sensitive nematodes relatively early in the decomposition process. 5. Objectives 4 and 5 Our emphasis was on determining causes and impacts of seasonal fluctuations in microbial communities in SAFS soils. Denitrification measurements were not continued after preliminary results indicated that denitrification rates in all farming systems were minimal during the growing season. Differences in microbial communities among farming systems were not as great as differences over the growing season. Microbial communities, regardless of farming system, showed the same general changes in composition over time. For example, in 1995, changes in microbial community composition between March and August were usually greater than any differences among farming system or among blocks within one farming system. We found similar trends in 1996 through 1998, with all years showing the same general progression with time for all farming systems. Thus similarities in early spring for all 3 years were greater than similarities between early spring and summer samples for the same year or similarities between farming systems for the same year. Even for soil samples collected under different crops and at different times at the SAFS plots, the same systematic progression in community composition with time could still be discerned. For example, corn samples collected in 1995 were more similar to 1995 tomatoes than they were to corn samples collected in 1997. In conclusion, the relative importance of environmental variables in governing the composition of microbial communities could be ranked in the order: time > specific farming operation (e.g., cover crop incorporation or side dressing with mineral fertilizer) > management system > spatial variation in the field.

Soil microcosm studies were conducted to identify PLFAs enriched, or decreased, by management practices and environmental perturbations simulated under laboratory conditions. Controlled laboratory experiments were necessary to follow up on hypotheses stemming from results collected in the field. Organic tomato soil from the SAFS Project was incubated with different carbon sources and at several moisture contents to test the following hypotheses:

1. Adding organic matter changes community composition to higher relative proportion of fungal and actinomycete markers.
2. Flooding changes community to greater relative proportion of anaerobic markers and lower relative proportion of fungal marker.
3. Flooding changes community more than organic matter addition.
4. Without addition of organic matter, changing water content has little effect on communities

Soils were incubated at moisture contents of 2%, 11%, field capacity, and flooding. Organic matter additions included cover crops (oats/vetch) only, compost (poultry manure) only, cover crops + compost, or no addition of organic matter. Samples were collected before and at several times (6, 11 and 20 d) during the incubation period. PLFA analysis was used to evaluate these hypotheses.

Both soil moisture content and organic matter addition strongly affected the composition of the soil community. Of all the moisture treatments, flooding caused the most significant change in composition and differences were evident within a week. The flooded treatments amended with organic material were substantially different from the unsaturated treatments (both with and without organic matter). With flooding, there was a high relative abundance of lipids associated with gram-positive anaerobic bacteria and low relative abundance of fungal markers. The PLFA fingerprints of soils incubated at 2, 11 or 22% moisture at all sampling times were relatively similar in the absence of organic matter additions; however in the presence of organic matter, differences were substantial. This suggested that significant changes in microbial communities due to stress are unlikely unless there are carbon sources available for populations to grow on. There was a higher relative abundance of fungal, but not actinomycete, markers in the soils amended with organic matter. In conclusion, both carbon and moisture are likely to have major impacts on the microbial community composition of soils in the field. The next challenge is to link these changes in microbial community composition to processes important to productivity and sustainability. Thus we measured relationships between carbon pools and microbial communities in SAFS soils exposed to wet/dry cycles.

California agricultural soils, particularly the surface layer, are subject to numerous extreme wet/dry cycles within the growing season. To investigate the effect of varying levels of organic amendments and wet/dry cycles on soil microbial communities, we compared soils from conventional, low input, and organic farming systems at SAFS. Differences in C availability to soil microorganisms were apparent among the three farming system soils. Microbial and dissolved organic C (DOC) increased with increasing organic inputs. Respiration following re-wetting was fit to a model describing two carbon pools in soil: one readily available and the other slowly available over a longer period of time. Carbon dioxide released from the more slowly available pool was higher in soils receiving greater long-term organic inputs.

Large differences in microbial process rates and community composition suggested greater adaptation to wet/dry cycles in the surface (0-3 cm) than deeper (3-15 cm) layer in all farming system soils. The increase in microbial biomass carbon (MBC) as a proportion of total MBC after soil re-wetting was greater in the surface than deeper layer. Respiration kinetics of the surface layer showed a more rapid response to soil re-wetting, as indicated by higher rate constants for both the rapidly and slowly used C pools, and by greater CO2 mineralized from the slowly used pool. The surface and deeper layers had distinctly different community compositions with higher ratios of cyclopropyl fatty acids to their precursors suggesting greater bacterial stress in response to wet/dry cycles in the deeper layer. This study suggested that adaptation to wet/dry cycles by surface microorganisms had occurred during the several months between the time of field preparation and tillage in April and the later sampling date in July, leading to changes both in microbial process rates and community composition.

Objective 6 We collaborated with Prof. Bruce Jaffee, Dept. of Nematology, and Cooperative Extension Specialist, Jeff Mitchell, in the Dept. of Vegetable Crops. With Prof. Jaffee, we published a paper linking information on microbial substrate induced respiration and nematode community size and structure to data on the density of nematode-trapping fungi. With Dr. Mitchell, we measured nematode community structure and microbial community composition in tomato and cover cropped plots that are part of the Biologically Integrated Farming Systems (BIFS) study near Fresno, California. Scow was involved in a project evaluating potential indices of soil quality with the USDA Soil Tilth Lab in Ames, Iowa. Scow also collaborated with Louise Jackson and Dennis Rolston on the effects of short term disturbance, such as irrigation and tillage, on N dynamics and microbial communities in agricultural soils. Scow’s lab also ran microbial community analysis for numerous growers, farm advisors, and grower groups. These include Bruce Roberts, Ed Weber, Chuck Ingels, and Roger Duncan (farm advisors), as well as the Sonoma Valley Growers Group and the Prune Board. Making these measurements provides our collaborators with valuable data on microbial communities, possibly of use in decision making. In turn, the collaborations give us on-farm applications of our measurements and opportunities to link our measurements to soil and agronomic properties of concern to growers.

Objective 7 Outreach is a significant part of our project and both PIs have been active participants in outreach activities for the past 3 years. This is despite the fact that the general SAFS project had no outreach coordinator past 1997. Scow and Ferris played a major role in two workshops put on by the BIOS (Biologically Integrated Orchard Study) Project during 1998. This involved running afternoon workshops on the soil food web. Also, Dr. Ferris and Scow were keynote speakers and participated in a panel discussion on how to measure soil organisms in a Soil Biology workshop hosted by John Luna and Mary Stabens at Oregon State University. We are a central resource for the FARMS program, a unique educational outreach program for high school students, which provides exposure to and research opportunities in agricultural science. Jessica Hanson, a graduate student in the Scow lab, was a coordinator for the program during 1997 and continued to be heavily involved in the program in 1998. Specific outreach activities are listed in the following section.