Summary• Data from 13 long-term (> 1 yr), field-based studies of the effects of elevated CO 2 concentration ([CO 2 ]) on European forest tree species were analysed using meta-analysis and modelling. Meta-analysis was used to determine mean responses across the data sets, and data were fitted to two commonly used models of stomatal conductance in order to explore response to environmental conditions and the relationship with assimilation.• Meta-analysis indicated a significant decrease (21%) in stomatal conductance in response to growth in elevated [CO 2 ] across all studies. The response to [CO 2 ] was significantly stronger in young trees than old trees, in deciduous compared to coniferous trees, and in water stressed compared to nutrient stressed trees. No evidence of acclimation of stomatal conductance to elevated [CO 2 ] was found.• Fits of data to the first model showed that growth in elevated [CO 2 ] did not alter the response of stomatal conductance to vapour pressure deficit, soil water content or atmospheric [CO 2 ]. Fits of data to the second model indicated that conductance and assimilation responded in parallel to elevated [CO 2 ] except when water was limiting.• Data were compared to a previous meta-analysis and it was found that the response of g s to elevated [CO 2 ] was much more consistent in long-term (> 1 yr) studies, emphasising the need for long-term elevated [CO 2 ] studies. By interpreting data in terms of models, the synthesis will aid future modelling studies of responses of forest trees to elevated [CO 2 ].

The effects of elevated atmospheric CO 2 concentration on growth of forest tree species are difficult to predict because practical limitations restrict experiments to much shorter than the average life-span of a tree. Long-term, processbased computer models must be used to extrapolate from shorter-term experiments. A key problem is to ensure a strong flow of information between experiments and models. In this study, meta-analysis techniques were used to summarize a suite of photosynthetic model parameters obtained from 15 field-based elevated [CO 2 ] experiments on European forest tree species. The parameters studied are commonly used in modelling photosynthesis, and include observed light-saturated photosynthetic rates (A max ), the potential electron transport rate (J max ), the maximum Rubisco activity (V cmax ) and leaf nitrogen concentration on mass (N m ) and area (N a ) bases. Across all experiments, light-saturated photosynthesis was strongly stimulated by growth in elevated [CO 2 ]. However, significant down-regulation of photosynthesis was also observed; when measured at the same CO 2 concentration, photosynthesis was reduced by 10-20%. Wullschleger (1993) and Ryan et al. (1994), are therefore invaluable for improving model predictions.These challenges were faced by the ECOCRAFT network, a group of laboratories conducting field-based experiments on the effects of elevated [CO 2 ] on European forest tree species. The network has existed since 1991, and the experimental results up to 1995 have been compiled by Jarvis (1998). Besford, Mousseau & Matteucci (1998) reviewed observations of photosynthetic rates in the ECOCRAFT experiments, and concluded that 'both up and down-regulation of photosynthesis has been found. Downregulation appears to be associated with either poor nutrient status or accumulation of starch, occurs more often late in the growing season and in the older needles of conifers.'The group then faced the problem of quantifying these results in such a way that they could be included in models. This problem was addressed by establishing a central relational database of model parameters (Medlyn & Jarvis 1999). The parameters required, and the methods of deriving them from experimental data, were agreed upon by project working groups comprising both experimentalists and modellers. This paper reports on the photosynthesis parameters from the database, presenting them in formats useful for modelling. First, lists of photosynthesis parameters for different species, extracted from the database, are presented. This catalogue should provide a useful resource for modellers. Second, quantitative methods (i.e. meta-analysis) are used to estimate the effects of elevated [CO 2 ] on the parameters across experiments. Different hypotheses for the effects of long-term elevated [CO 2 ] on photosynthesis are examined. This analysis aids our understanding of photosynthetic responses to elevated [CO 2 ] and suggests ways in which we can improve model formulation.Photosynthesis is a key process when stu...

No abstract

The net ecosystem exchange (NEE) of CO 2 between temperate forests and the atmosphere governs both carbon removal from the atmosphere and forest growth. In recent years, many experiments have been conducted to determine temperate forest NEE. These data have been used by forest modellers to better understand the processes that govern CO 2 fluxes, and estimate the evolution of these fluxes under changing environmental conditions. Nevertheless, it is not clear whether models capable of handling short-term processes, which are mostly source-driven, can provide an accurate estimate of long-term forest growth, which is potentially more influenced by sink-and phenology-related processes. To analyse the interactions between short-and long-term processes, we developed the ASPECTS model, which predicts long-term forest growth by integrating, over time, hourly NEE estimates. Validation data consisting of measurements of NEE by eddy-covariance and forest carbon reservoir estimates were obtained from mixed deciduous and evergreen experimental forests located in Belgium. ASPECTS accurately estimated both: (1) the NEE fluxes for several years of data; and (2) the amount of carbon contained in stems, branches, leaves, fine and coarse roots. Our simulations demonstrated that: (1) NEE measurements in Belgian forests are compatible with forest growth over the course of the 20th century; and (2) that forest history and long-term processes need to be considered for accurate simulation of short-term CO 2 fluxes.

-Belgian forests covered 693 181 ha in 2000, representing 22.7% of the total land area. As no biomass or carbon stock data are included in the Flemish and Walloon regional forest inventories, species-specific wood densities, biomass expansion factors s.s. and carbon content values were critically selected from the literature. Based on these conversion and expansion factors, and on data from the forest inventories, the total C-stock in the living biomass of productive Belgian forests was assessed at 60.9 Mt C in the year 2000. The overall mean C-stock amounted to 101.0 t C ha -1 . This value was in the higher range of values reported for the neighbouring countries, mainly due to a high mean growing stock in the Belgian forests (261.9 m 3 ha -1 ). The conversion from wood volume to wood biomass based on wood density values reported in the literature appeared to introduce the largest variability in the assessment of the carbon stocks. Additional measurements of wood densities in Belgian forests could help to reduce the uncertainty related to this factor. Because of the time-consuming and destructive character of the determination of biomass expansion factors s.s. (BEFs), the establishment of new BEFs does not have the highest priority in the framework of improving the assessment of the biomass carbon stock in the Belgian forests. As the median C-content value for all species except beech was equal to the default IPCC-value of 50% carbon in dry matter, it seems appropriate to use this value for future calculations.forest inventory / carbon stock / biomass expansion factor / wood density / aboveground and belowground biomass / carbon content Résumé -Estimation des stocks de carbone dans la biomasse des forêts en Belgique. La forêt belge couvrait 693 181 ha en 2000, ce qui représente 22,7 % de la surface totale du pays. Les inventaires forestiers flamands et wallons n'incluent pas de mesures directes de biomasse ou de stock de carbone. Pour calculer les stocks de carbone dans les arbres forestiers à partir des volumes de bois fort, les infradensités du bois, les facteurs d'expansion de la biomasse et les teneurs en carbone ont été sélectionnés dans la littérature. En 2000, les stocks de carbone dans la biomasse des forêts belges productives étaient de 60,9 Mt C, soit 101,0 t C ha -1 . Cette dernière valeur est relativement haute comparée avec celles observées dans les forêts des pays avoisinants, à cause d'un grand volume de bois fort dans les forêts belges (261,9 m 3 ha -1 ). L'infradensité du bois semble être le facteur qui introduit la plus importante variabilité dans le calcul des stocks de carbone dans la biomasse. Des mesures complémentaires des infradensités du bois des essences les plus communes en Belgique pourraient aider à réduire significativement les incertitudes sur les mesures des stocks de carbone dans les forêts. La détermination expérimentale des facteurs d'expansion est coûteuse en temps et se base sur des analyses destructives. Ainsi, l'obtention de facteurs d'expansion propre à la situati...