Title: Nonlinear and threshold responses of grassland productivity and species composition to increased CO2 vary with soil type Authors
Submitted to: American Geophysical Union
Publication Type: Abstract Only
Publication Acceptance Date: October 1, 2011
Publication Date: December 31, 2011
Citation: Fay, P.A., Jin, V.L., Jackson, R.B., Gill, R.A., Way, D., Polley, H.W. 2011. Nonlinear and threshold responses of grassland productivity and species composition to increased CO2 vary with soil type. In: Proceedings of the EOS Trans. American Geophysical Union, San Francisco, California. Paper No. B41A-0197. Technical Abstract: Climate change is likely to cause non-linear responses in ecosystem function and threshold changes in species composition. Here we report aboveground net primary productivity (ANPP) responses to a continuous CO2 concentration gradient (250 to 500 µL L-1) in experimental grassland communities on three soils differing in water holding capacity and other properties. Communities consisting of four C4 grasses, two C3 forbs, and one legume were established on a lowland clay (vertisol, n=32), an upland clay (mollisol, n=32), and an alluvial sand (alfisol, n=16). The communities were positioned in a stratified random design in the CO2 gradient for five growing seasons, and were irrigated to mimic the average growing season rainfall regime for the study site in Central Texas. ANPP increased with CO2 almost two-fold more on the upland clay and alluvial sand than on the lowland clay (p < 0.0001), because of strong linear responses to CO2 on these soils (R2 = 0.50 to 0.59, p = 0.002) compared to a saturating response to CO2 on the lowland clay (R2 = 0.48, p= 0.01). On the two more responsive soils, the mesic tallgrass Sorghastrum nutans replaced the more drought adapted mid-grass Bouteloua curtipendula at elevated CO2, while B. curtipendula largely replaced S. nutans at low CO2, especially on the upland clay. Evidence for a similar composition change was not found on the lowland clay. Thus, two soils displayed a threshold change in community composition that accounted for up to 57% of variation in ANPP for those soils. Variation in ANPP and species composition with CO2 were accompanied by linear increases in soil water content (SWC, 0 - 20 cm, volumetric), most strongly on the alluvial sand (R2 = 0.39, p < 0.009) and by weak decreases with CO2 in soil N. Structural equation models (SEMs) explained 34 to 52% of the variation in ANPP, and indicated that CO2 effects on ANPP on the upland clay were primarily explained by CO2 effects on species composition, and on the alluvial sand by CO2 effects on SWC. Responses to elevated CO2 in SWC, ANPP, and species composition were explained by reduced stomatal conductance and increased photosynthetic water use efficiency in both grasses. In addition, S. nutans gained more in WUE at elevated CO2 than B. curtipendula, while B. curtipendula at elevated CO2 had lower light-saturated photosynthetic capacity, quantum use efficiency, and dark respiration than S. nutans. Thus, at elevated CO2, shading by the taller S. nutans likely lowered B. curtipendula carbon assimilation and growth. We conclude that elevated CO2 strongly increased ANPP on upland clay and alluvial sand soils where there were also gains in soil moisture and threshold changes in species composition driven by physiological differences in the two dominant grass species. As a result, CO2 effects on ANPP will likely differ with soil type across the landscape.