Search

Climate warming is releasing carbon from soils around the world, constituting a positive climate feedback. Warming is also causing species to expand their ranges into new ecosystems Yet, in most ecosystems, whether range expanding species will amplify or buffer expected soil carbon loss is unknown. Here we used alpine grasslands as a model system to determine whether the establishment of herbaceous lowland plants in alpine ecosystems influences short-term soil carbon storage under warming. We found that warming (

International audience

International audience; Predictions of soil organic carbon (SOC) losses in the northern circumpolar permafrost area converge around 15% (± 3% standard error) of the initial C pool by 2100 under the RCP 8.5 warming scenario. Yet, none of these estimates consider plant-soil interactions such as the rhizosphere priming effect (RPE). While laboratory experiments have shown that the input of plant-derived compounds can stimulate SOC losses by up to 1200%, the magnitude of RPE in natural ecosystems is unknown and no methods for upscaling exist so far. We here present the first spatial and depth explicit RPE model that allows estimates of RPE on a large scale (PrimeSCale). We combine available spatial data (SOC, C/N, GPP, ALT and ecosystem type) and new ecological insights to assess the importance of the RPE at the circumpolar scale. We use a positive saturating relationship between the RPE and belowground C allocation and two ALT-dependent rooting-depth distribution functions (for tundra and boreal forest) to proportionally assign belowground C allocation and RPE to individual soil depth increments. The model permits to take into account reasonable limiting factors on additional SOC losses by RPE including interactions between spatial and/or depth variation in GPP, plant root density, SOC stocks and ALT. We estimate potential RPE-induced SOC losses at 9.7 Pg C (5 - 95% CI: 1.5 - 23.2 Pg C) by 2100 (RCP 8.5). This corresponds to an increase of the current permafrost SOC-loss estimate from 15% of the initial C pool to about 16%. If we apply an additional molar C/N threshold of 20 to account for microbial C limitation as a requirement for the RPE, SOC losses by RPE are further reduced to 6.5 Pg C (5 - 95% CI: 1.0 - 16.8 Pg C) by 2100 (RCP 8.5). Although our results show that current estimates of permafrost soil C losses are robust without taking into account the RPE, our model also highlights high-RPE risk in Siberian lowland areas and Alaska north of the Brooks Range. The small overall impact of the RPE is largely explained by the interaction between belowground plant C allocation and SOC depth distribution. Our findings thus highlight the importance of fine scale interactions between plant and soil properties for large scale carbon fluxes and we provide a first model that bridges this gap and permits the quantification of RPE across a large area.

Climate change in temperate mountain systems and associated increase in temperature and decrease in precipitation are expected to have strong implications for vegetation productivity, species diversity and carbon turnover in subalpine grasslands. Little is known, however, about the interaction between the effects of climate change and those of local land use management and possible changes in landscape structure. Pasture woodlands in the Swiss Jura Mountains are a traditional landscape, resulting from a long-lived sustainable use of grasslands and woodlands, and as such provide a suite of important ecosystem services to human society. These range from carbon sequestration and biodiversity preservation, to provision of timber and forage for livestock, and last but not least an aesthetic value, much appreciated by tourism. In this thesis various aspects of ecosystem functioning have been studied, investigating the combined effects of experimental climate change and land use on structurally different wooded pastures. An altitudinal gradient method has been used to simulate future climate change conditions, by imposing warmer and drier climate on subalpine turfs transplanted at lower elevation. The resulting gradient in mean annual temperature and precipitation – ranging from cold and wet in the subalpine zone, to warm and dry in the colline zone – has allowed for the detection of tipping points and altered states of ecosystem functioning in response to the treatments. The method employed provided also the possibility for a direct comparison of three land use types: unwooded pastures, sparsely wooded pastures, and densely wooded pastures (the result of pasture management intensity), in their response to climate perturbation. During the four years of experimental work, a series of observations have been made at the plot scale (square metre) in terms of plant performance and biogeochemical cycles, as well as at the landscape scale (hectare) in terms of forage production. A general threshold level for ecosystem resistance to experimental climate change was detected between the moderate IPCC scenario (+2 K mean annual temperature; -20 % annual precipitation) and the intensive IPCC scenario (+4 K mean annual temperature; -40 % annual precipitation). A concomitant gradient in ecosystem response to climate change was observed across the three land use types. The intensively managed unwooded pasture type was consistently more affected by the experimental treatment and rarely exhibited signs of resistance, especially under the intense climate change scenario. A drastic loss of plant species diversity, reduction of herbaceous biomass, impaired litter decomposition and soil microbial metabolic activity have all contributed to the altered state of ecosystem functioning. In contrast, the two extensively managed wooded pasture types showed considerable resistance to climate perturbation in terms of both above and belowground ecosystem processes. The reported inter-annual variation in herbaceous diversity and biomass production within these land use types demonstrated their resilience (recovery) potential too. Using a modelling approach for upscaling these results to the heterogeneous landscape of pasture woodlands in the Swiss Jura Mountains, has proven that extensively used wooded pastures could grant sustainable ecosystem services in terms of forage provision for cattle under climate change. Considering that the two experimental climate change intensities implemented this study are the projected ‘best’ and ‘worst’ case scenarios for the coming decades, the reported resistance of wooded pastures to climate change has to be embraced, and sustainable land use set as a goal in high altitude mountain pastures.