Abstract: | Long-term eddy covariance measurements over a montado oak woodland in southern Portugal have documented a vulnerability to predicted decreases in springtime rainfall, since water availability during spring limits annual CO2 gain, the growth of fodder for animals, and the production of cork by Quercus suber. The current study examined CO2 exchange of three different herbaceous vegetation components distributed over montado landscapes and within the footprint of long-term landscape eddy covariance monitoring studies. Simultaneous measurements with eddy covariance at two sites and with manually operated chambers at multiple locations revealed that slow drainage of shallow basins, the onset of drying at higher sites and a high release of CO2 below tree canopies significantly influenced the overall course of montado ecosystem gas exchange during the spring.Hyperbolic light response models were employed to up-scale and compare herbaceous gas exchange with landscape net ecosystem CO2 flux. The up-scaling demonstrates the importance of the herbaceous understory in determining annual carbon balance of the montado and suggests a relatively small additional CO2 uptake by the tree canopies and boles, i.e., by the aboveground tree compartment, during springtime. Annual flux totals obtained during the extremely dry year 2005 and a normal precipitation year 2006 for the oak woodland and a nearby grassland were essentially the same, indicating that both ecosystems similarly exploit available resources. Based on comparisons with additional temperate grasslands, we can visualize the montado herbaceous cover as a typical European grassland canopy, but where temperature fluctuations in winter control uptake, and where total production depends on springtime rainfall as it controls phenological events and eventually dieback of the vegetation. On the other hand, tree canopies remain active longer during late spring and early summer, modifying the montado response from that of grassland. Uncertainties in flux estimates via both chamber and eddy covariance methodologies currently prevent a full understanding of vegetation/atmosphere coupling, of the recycling of CO2 between the understory communities and trees, and of relationships between exchange rates of individual components of the vegetation mosaic and overall carbon and water balances in montado landscapes. |