Monitoring CO2 fluxes and partitioning of soil respiration in a Mediterranean forest ecosystem: an integrated approach to carbon cycle

The increase of greenhouse gases concentrations in the atmosphere is a major driver for global warming and climate change. Carbon dioxide is the primary anthropogenic greenhouse gas and is cumulatively responsible for approximately 55% of greenhouse-gas-related climate forcing, popularly known as "the greenhouse effect". Forests play an important role in the carbon cycle and carbon sequestration at both local and global scales. Trees remove CO2 from the atmosphere through photosynthesis and store carbon in different tree components and in the soil. Whether a forest acts as a carbon sink or source depends on the difference between photosynthetic uptake and respiratory release of CO2. The total flux of CO2 released from the ecosystem (Reco) originates form a range of sources, which can be broadly divided into those originating from aboveground plant tissues (RAa) and soil (RS). Soil CO2 efflux can furthermore be partitioned into belowground autotrophic respiration (roots plus associated microorganisms, RAb) and the decomposition of dead organic matter (heterotrophic respiration, RH). In order to predict likely changes in ecosystem carbon balance under changed environmental conditions, it is therefore necessary to identify and quantify the different sources of CO2 efflux and their dependence on environmental conditions. For this purpose a combined approach based on simultaneous eddy covariance and soil respiration measurements was applied in a maritime pine (Pinus pinaster) forest in Central Italy within the Regional Park of San Rossore – Migliarino – Massaciuccoli (Tuscany). In 2011 a girdling experiment has been developed to partition RS into RH and RAb. The experiment started in spring 2011 and the response has been followed till the end of 2012. Two weeks after the girdling treatment soil respiration in the girdled plots decreased by 30% and remained stable over the period of analysis with an average RH/RS that was estimated around 0.70, suggesting that at San Rossore site RH dominates RS. The anomalous low rainfall regimen recorded during 2011 growing season and 2012 summer offered the opportunity to investigate the decoupled response of respiration to water availability and temperature. RS and RH responded quite predictably to environmental controls. Nevertheless a dichotomous response was observed during the hot and dry season and during the wetter and colder winter. Soil water availability was the major control of RH and RS during the growing season. Severe drought masked the temperature response of respiration which was restored only during the wettest periods. At the ecosystem scale, it can be estimated that RH and total autotrophic respiration (RAt) contributed 40% and 60% of Reco, respectively. Our data suggested that photosynthesis was the major driver of RAt throughout the whole observation period, thus highlighting the important role exerted by newly assimilated carbon on plant tissue respiration. Large and consistent peaks of CO2 emissions were recorded from the soil after drying – rewetting cycles, as has been observed in many other water limited ecosystems (the so called “Birch effect”). These wet-days warm-soil respiration peaks contributed to 50 – 70% of Reco and released an amount of carbon to the atmosphere that was double the amount emitted during the whole dry season. Our data derived from natural field conditions unequivocally indicate that the soil CO2 pulses following rewetting of the dry soil derived from the rapid microbial oxidation of labile carbon compounds. A delayed effect of the water pulse on Rs was also observed, which was ascribed to the slower mobilization of more recalcitrant soil organic matter and to the time lag requested by recently assimilated photosynthates to be translocated from the canopy to the root system. Even though severe summer drought is common at this experimental site with a typical Mediterranean climate, the anomaly in the precipitation regimen experienced in 2011 and 2012 was unusual and significantly impacted the ecosystem carbon balance. Soil water content was clearly the main environmental driver for both gross primary productivity (GPP) and Reco. Nonetheless, GPP was less affected when trees could still access deep soil water reserves with roots. These observations highlight the vulnerability of the Mediterranean-type ecosystems not only in terms of their response to predicted increasing temperatures but also to changing precipitation patterns. This demands for an incorporation of these variables into models for prediction of ecosystems’ feedbacks to climate change. Since the capability of an ecosystem to be a net sink of CO2 depends on its ability to fix and retain more carbon than that is respired back to the atmosphere, carbon losses associated to an increased variability of rainfall events could dramatically impact the ecosystem carbon balance and determine the fate of this forest of being a net carbon sink or source