Text

theses

Three concepts form the foundation of the hypotheses of this dissertation. First, that the partnership between switchgrass (Panicum virgatum) and arbuscular mycorrhizal fungi (AMF) allows the plant to grow with very limited agricultural inputs, which makes it an excellent candidate for a biofuel crop. Second, that the soil community including soil arthropods and nematodes also functions in conjunction with AMF to support the resilience of biomass production in switchgrass. Third, that annual N fertilization will disrupt the switchgrass-AMF relationship and cause changes in the soil community that might affect the biomass production in the long run. Commercial production is projected to harvest fields for up to 20 years.

The experiments in this dissertation measured multiple measures across soil community to capture any effects of the experimental treatments that may have cascaded across trophic levels. Soil community measures included AMF colonization, the communities of soil microarthropods, nematode abundance, and microbial community function (enzyme profiles) through BIOLOG ecoplates. In addition, soil measurements of plant available N and P were taken to measure impact of experimental treatments on fundamental soil properties and measurements of biomass yield were taken to capture the impact of experimental manipulations on plant growth. The three experiments were conducted in field and greenhouse settings. Chapter 2 addresses a study of the shift in the soil community of an established switchgrass field that received annual N fertilization. Chapter 3 describes a greenhouse experiment where sterilized field soil was manipulated with additions of AMF, fungal-feeding nematodes, and N fertilizer. Chapter 4 describes a greenhouse experiment where homogenized field soils from " prime ", "marginal", and "poor" soils (rated for farming purposes by the Natural Resources Conservation Service) were manipulated with additions of AMF and N fertilizer.

For the first experiment (Chapter 2), a switchgrass field that was established in 2008 was studied for three years from 2013 - 2015. It was investigated whether yearly additions of nitrogen (N) fertilizer at 100 lb/ac (112.1 kg/ha) changed the soil community. This level of N fertilization fell within the range of best-management practices. Measurements from the established switchgrass field were compared to adjacent unplanted farmland as representative of the soil community prior to switchgrass establishment. The results showed statistically significant impact of N fertilizer on the soil community, but more occurrences of statistically significant differences between the reference area and the established switchgrass field. Soil arthropod communities were statistically different between fertilized and unfertilized plots on 3 of 6 sampling dates. In comparison, the planted areas combined differed from control areas on 4 of 6 dates. Similarly, mycorrhizal structures were statistically different between planted and unplanted plots on 2 of 9 dates, whereas reference area was different from planted areas on 5 of 9 dates. There were statistical differences in nematode abundance and microbial community function as measured by BIOLOG ecoplates. However, further testing showed the differences were in the comparison of the reference area to the planted plots rather than between the planted areas and due to N fertilization. These results suggest that the impact of fertilization is less than that of the change due only to plant establishment. Soil extractable nutrients showed significantly higher amounts of NH4 and NOx and significantly lower amounts of PO4 in fertilized plots. There was evidence that NOx was increasing over time in the soil. Above-ground biomass yields as well as N content of stem biomass were different at the P = 0.1 level, with 75% higher biomass yields as well as 20% higher N content in the fertilized areas. Therefore, since the soil extractable nutrients were statistically different, and the plant measurements between fertilized and unfertilized areas are less statistically supported, soil changes do not directly translate into plant growth. These results suggest that switchgrass is insensitive to manipulations of N levels, as the factors related to the response of the plant to fertilization were not significant at the P = 0.05 level. The results also show that the soil community is similarly insensitive to the changes due to fertilization because the soil community responded less often to fertilization than to the difference between planted and reference areas.

The second set of experiments (Chapter 3) tested the hypothesis that plant-soil interactions exist in agriculturally-produced switchgrass with two related greenhouse experiments. The soil community and edaphic conditions were manipulated in a fully factorial design to see how commercial mycorrhizal inoculum, fungal feeding nematodes, and N fertilizer impacted biomass yield of switchgrass during one growing season. A subset of the samples was overwintered and grown for a second season to measure any lag effects from the initial treatment, since switchgrass is a perennial crop. Plant biomass and plant N content showed no statistically significant differences due to the experimental manipulations in the one-season or two-season experiment. There were no clear trends in the response of mycorrhizal colonization and structure formation based on experimental treatments in either experiment. The soil arthropod community changed due to N fertilization in the one-season experiment, but not in the two-season experiment. Nematode abundance showed no differences in response to experimental treatments in both experiments. When all soil response variables were combined in an NMDS analysis, none of the experimental treatments consistently affected the distribution of results for either experiment. These results suggest that there are not any strong plant-soil feedbacks in this system. Although there was a limited response of the soil community to experimental treatments, there was no statistically significant response in the plants that could be attributed only to the experimental manipulations.

The third set of experiments (Chapter 4) investigated whether P. virgatum growth responded differently to soil manipulations (adding N fertilizer and commercial mycorrhizal inoculum) in three soils: “prime farmland”, “farmland of local importance”, and “not prime farmland” as rated by the Natural Resources Conservation Service farmland classification system. Additionally, the soil was used for two consecutive years with no additional manipulation to test for soil exhaustion. A 3x2x2 factorial design greenhouse experiment was conducted to test the hypothesis that switchgrass biomass yields and soil community would respond with different effect size or direction to edaphic manipulations in three soils. In the first year, measurements of soil factors were different with statistical significance due to soil type. Higher extractable NO3 in fertilized treatments additionally was statistically significant. Statistically significant differences in measurements of plant factors (biomass and N content) were primarily due to soil type. While both stem and root N content were higher in fertilized treatments, the trend was not statistically significant. Soil community factors (mycorrhizal structures, soil arthropod morphospecies, carbon utilization of the microbial community, and nematode abundance) which were primarily multivariate data, were only different with statistical significance due to soil type, not inoculation or fertilization. In the second year, when no additional manipulation occurred, soil extractable nutrients were lower than in the first year (P < 0.05). Within the second-year-only analyses of soil factors, statistically significant differences were primarily due to soil type. However, lower amounts of extractable PO4 were statistically significant in inoculated treatments. Plant factors were different with statistical significance only for soil type. Soil community factors showed stronger response to inoculation than in the first year, in addition to the statistically significant differences attributable to soil type. These results showed the resilience of switchgrass and the soil community to perturbation across all soil types. While there were signs that the soil was approaching a state of exhaustion, this was not reflected in plant biomass yields. The lag response that was found suggests that treatment effects build up over time; since inoculation was not statistically significant in the first year but was in the second year for BIOLOG ecoplate results and nematode abundance. This building effect is worthy of further investigation.

In general, the manipulations to the soil resulted in changes to the soil extractable nutrients. However, the soil community did not consistently change in response to experimental manipulations. Similarly, plant biomass and plant N content generally did not respond to these experimental manipulations. These results speak to the resilience of the soil community to perturbation, and the physiological resilience of switchgrass to a variety of growth conditions. These results show that switchgrass is an excellent candidate for biofuel production. Best-management practice of annual N fertilization should be reconsidered to reduce the energetic cost of fertilizer to ensure net energy gain in the production of biofuel. However, the results also indicate that the current best-management practice of applying low levels of N fertilizer is not outweighing the benefit to the soil and soil community of the extensive root system of switchgrass.

ETD_11240

Ph.D.

Includes bibliographical references

External ETD doctoral

doi:10.7282/t3-k593-kc08

The author owns the copyright to this work.