Long-term responses of a grassland ecosystem to CO2, nitrogen, diversity, and composition

Background/Question/Methods Predicting terrestrial ecosystem change in the face of global change remains an enormous challenge for many reasons, including the difficulty of testing effects of global change drivers and their interactions on community and ecosystem processes. A long-running perennial grassland experiment, BioCON, can provide some advances in this respect. Beginning in 1998, 361 plots differing in the number and identity of species (from 1 to 16 species per plot, drawn from four plant functional groups) were grown under all combinations of ambient and elevated CO2 and N at the Cedar Creek site in Minnesota. Results/Conclusions Over a ten-year period, CO2 and N have influenced many processes, including plant composition, diversity, productivity, and reproductive output, as well as soil communities and nitrogen cycling. Some responses have been consistent over time, while others varied in either idiosyncratic or patterned ways. For example, CO2 and N effects on photosynthesis (a species level trait) or species diversity (a community level trait) have been relatively stable over time. Other responses vary with time, as two examples show. First, changes in plant composition reflect different species responses to CO2 and N superimposed on dynamics driven by ontogenetic and successional characteristics of the species pool. These changes are marginally well explained by differences in functional traits or plant functional groups. Second, the CO2 stimulation of productivity (a species/ecosystem trait) has been consistently greater (since 2000) under N-fertilized than ambient conditions, although the 10-year data suggest a long-term pattern that may take 15-20 years to characterize. Although largely conjecture (based on circumstantial evidence), it appears that complex feedbacks between CO2 and N availability, soil microbial communities, and C and N biogeochemistry may be responsible for the changing nature of the response to elevated CO2 at different N supply rates. Finally, some questions can only be addressed over time. For instance, given the reduced leaf conductance under elevated CO2 that occurs in most plants (including in BioCON), an enduring hypothesis is that water savings will lead to a greater CO2 fertilization effect in drier years, climates or sites than in moister ones. However, data from BioCON surprisingly show neither a direct response to interannual variation in rainfall nor an alteration of the CO2 effect on biomass production. Results from BioCON highlight the need for long-term studies in perennial ecosystems, as the time needed for plant compositional shifts and/or biogeochemical “waves” to occur is typically long by typical research standards.