Friday, August 7, 2009 - 8:20 AM

COS 126-2: Evaluating atmospheric CO2-induced microbial feedbacks on plant productivity in a microcosm experiment

Andrew C. Procter, Alexia M. Kelley, and Robert B. Jackson. Duke University

Background/Question/Methods

As atmospheric CO2 rises, the ability of ecosystems to sequester carbon depends on sustained plant productivity.  This in turn depends on nutrient supply by the soil microbial community.  Since ecosystem responses to elevated CO2 may include shifts in soil community composition and activity, we designed a microcosm experiment to test for CO2-induced microbial feedbacks on plant growth.  The microcosm experiment is linked to a field-scale experiment in which a gradient of atmospheric CO2 (250-500ppm) has been imposed on a prairie ecosystem for 3 growing seasons.  We grew Indiangrass (Sorghastrum nutans) seedlings in pots receiving autoclaved soil plus a soil inoculum taken from either the elevated or the subambient endpoint of the CO2 gradient.  The seedlings were placed in growth chambers under either elevated or subambient CO2 for 2 months, to test for feedback from the inoculum on plant growth and nutrition.

Results/Conclusions

Initial results indicate that inoculum had a stronger effect on plant growth than did growth chamber CO2.  Repeated measures ANOVA indicated a significant inoculum (p = 0.01) and inoculum x time (p < 0.01) effect on plant height measured during the course of the experiment.  Plants receiving inoculum from the elevated endpoint of the CO2 gradient grew faster than those receiving inoculum from the subambient endpoint, indicating a positive elevated-CO2 feedback on plant growth.  However, neither growth chamber CO2 nor CO2 x time had significant effects (p > 0.05).  Plants grew faster when receiving live compared to autoclaved inoculum (p < 0.01), indicating the microbial communities had net positive feedback on plant growth.  The inocula differed in microbial biomass; the inoculum from the elevated endpoint of the CO2 gradient had higher microbial biomass C compared to inoculum from the subambient endpoint (p = 0.07).  These inocula did not differ significantly in microbial biomass N.