Inter-annual variability of carbon exchange in Mediterranean shrubland ecosystem

Only a few long-term studies on inter-annual variability in energy and mass exchanges of Mediterranean shrubland ecosystems have been recently published. Since Mediterranean maquis ecosystems experience a wide variation in rainfall and temperature, inter-annual differences in turbulent fluxes are expected. Mediterranean-type ecosystems normally show two main peaks of growth (spring and fall) and are exposed to pronounced summer drought periods. Consequently, their behavior is even more complex and responds more dramatically to perturbations in water conditions. The main objectives of this research are (i) to characterize dynamics of ecosystem carbon cycling, (ii) to identify the driving factors affecting ecosystem exchanges, and (iii) to investigate intra- and inter-annual variability of ecosystem fluxes. Seven completed years of energy and mass fluxes measured using Eddy Covariance (EC) technique over a secondary succession shrubland ecosystem (maquis) located in Sardinia, Italy are here reported. Inter- and intra-annual flux variations were investigated by computing the correlation between climatic variables and ecosystem carbon fluxes at different time lags. For this purpose, climatic drivers were shifted one month at a time by using a moving time window of twelve months for all the measurement period and year by year. Most of the observed available energy resulted in positive Bowen ratio. Maquis ecosystem showed two distinct intense physiological activity periods (spring and fall), while during the summer, vegetation continued to photosynthesize even under extreme drought conditions, although at a reduced rate. The estimate of net ecosystem exchange (NEE) showed differences among years depending on drought and temperature conditions, although the ecosystem acted as a net C sink every year. A positive relationship was found between seasonal gross primary production (GPP) and latent heat flux (LE), while a negative relationship was found between annual NEE and air temperature. The ecosystem was a C source in summer and winter and C sink in spring and fall. The total C sequestration varies from 90 to 466 g C m-2 y-1. For all climatic drivers, NEE and GPP showed similar correlograms. A shift of 2-3 months in ecosystem response was observed depending on air temperature, soil temperature and PAR (positive correlations), while a negative correlations is showed for ecosystem respiration (Reco). At annual scale, the ecosystem response was mainly driven by air temperature and PAR where a more pronounced lag effect was observed in cooler and drier years. Soil temperature showed similar lag effect as for air temperature.