Search

International audience; Intra- and interannual variability of precipitation can lead to major modifications of grassland production and carbon storage capacity. Greater understanding of how climatic variability affects net CO2 exchange [i.e. net ecosystem exchange (NEE)] of grazed grasslands is important to adapt grassland management and reduce risks of carbon losses. Since 2002, we continuously measured NEE (i.e. eddy covariance technique) on an upland grassland site (7 ha), divided in two paddocks grazed by heifers (intensive: 1 LSU ha(-1) yr(-1), 213 kg N ha(-1) yr(-1) and extensive: 0.5 LSU ha(-1) yr(-1), no fertilization). For years with dry and warm growing seasons (i.e. 2003, 2005 and 2008), absolute annual NEE was higher in the intensive paddock compared with the extensive paddock. The opposite was observed during years of ample seasonal rainfall and soil moisture (i.e. 2004, 2006 and 2007). Contrasted management led to two distinct plant communities being different in leaf area index (LAI), soil bulk density and soil water holding capacity. Differences in annual NEEs could thus be assigned to interactions between in carbon and water fluxes during dry and wet growth periods. Dry growth periods led to a reduction in weekly gross primary productivity (GPP) in the extensively managed paddock, whereas the GPP was maintained in the intensive paddock. In turn, during wet growth periods, GPP was similar in both paddocks, whereas N amendment and frequent defoliation significantly increased ecosystem respiration in the intensive paddock, presumably through a higher heterotrophic respiration following on a better C substrate quality and availability (rhizodeposition and senescent fine roots). In the extensive paddock, where plant cover was denser (reducing soil temperature) and less decomposable, C losses through heterotrophic respiration were comparatively smaller under wet conditions. Our results demonstrate that grassland subjected to a moderately intensive management could be more resilient in terms of carbon storage during drought and heat waves, presumably because of a trade-off between heterotrophic and autotrophic respiration.

The role of plant community structure and plant functional traits for above- and belowground carbon (C) fluxes was studied for 2 years in a mesocosm experiment with grassland monoliths, using continuous gas exchange measurements and soil analyses. Here we test the response-and-effect trait hypothesis, by applying a mathematical framework used to predict changes in C fluxes after a change in disturbance through the community response (R) and effect (E) traits. Monoliths were extracted from two contrasted long-term field treatments (high vs. low grazing disturbance) and exposed to both low and high (simulated grazing) disturbance during a 2 years experiment. Carbon dioxide exchanges were measured continuously in an open flow system. Net ecosystem productivity and ecosystem C balance were positively correlated at low disturbance with plant species richness. Aboveground net primary productivity (ANPP) and soil C sequestration were, however, unrelated to these variables. Community aggregated leaf (specific leaf area, leaf dry-matter content) and root and rhizome (specific length, tissue density, diameter) traits responded (R) significantly to changes in disturbance, indicating an increased dominance of conservative plant growth strategies at low compared with high disturbance. Applying the mathematical framework, ANPP was predicted by distribution of leaf traits within the community (functional divergence), while mean root and rhizome traits had significant effects (E) on soil C sequestration, irrespective of the experimental disturbance and of the year. According to highly significant linear regression models, between 6% and 61% of the transient changes in soil C sequestration resulted from community root and rhizome (response-and-effect) traits after a change in disturbance.