Atmospheric CO₂ is the main source of C for the plants metabolism. Furthermore, anthropogenic fossil fuel emissions are increasing CO₂ levels in the atmosphere, constituting an important alteration of the environment for plants and a major source of global climate change. Since iLand is particularly designed for applications under climate change, growth response to CO₂ is explicitly included in the model.

growth response to atmospheric CO₂

Many models applying the radiation use efficiency concept do not explicitly consider changing atmospheric CO₂ levels, assuming a constant photosynthetic efficiency per unit utilizable radiation (e.g, Landsberg and Waring 1997, Mäkelä et al. 2008). However, recent empirical evidence (e.g., Norby et al. 2005) corroborates the effect of raised CO₂ levels on primary production of forests.

In line with the environmental modifier approach applied in iLand we calculate a monthly CO₂ response (using annual CO₂ concentrations and monthly environmental interactions, see below), f_CO2, directly applied to modify radiation use efficiency (cf. the findings of Norby et al. 2005, see Kicklighter et al. 1999 for a variety of other approaches at the approximately same process resolution). Following the analysis of Friedlingstein et al. (1995) we adopt a Michaelis-Menten formulation for CO₂ response (Eq. 1)

\[\begin{aligned} f_{CO2}=\frac{K_{1}(C_{t}-C_{b})}{1+K_{2}(C_{t}-C_{b})} \end{aligned} \]

Eq. 1

where \[\begin{aligned} K_{1}=\frac{1+K_{2}(C_{0}-C_{b})}{C_{0}-C_{b}} \end{aligned} \] \[\begin{aligned} K_{2}=\frac{(2C_{0}-C_{b})-r(C_{0}-C_{b})}{(r-1)(C_{0}-C_{b})(2C_{0}-C_{b})} \end{aligned} \] \[\begin{aligned} r=\frac{NPP_{2\times CO2}}{NPP_{1\times CO2}} \end{aligned} \] with C_t and C₀ the current and reference CO₂ concentrations respectively (reference means the average concentration for which the base quantum use efficiency ε₀ was estimate), and r the productivity increase for a doubling of the CO₂ concentration. The Michaelis-Menten formulation is realistic in that it yields zero productivity for C_t=C_b the CO₂ compensation point, and it saturates to K₁/K₂ when C_t tends to infinity. Also r can be related to the frequently reported factor β by applying Eq. (2) (cf. Friedlingstein et al. 1995)

\[\begin{aligned} r=1+ln(2)\beta \end{aligned} \]

Eq. 2

β was found to be consistent over a variety of temperate forest ecosystems in the empirical data of Norby et al. (2005) at β=0.6

---

environmental sensitivity of CO₂ response

Following Friedlingstein et al. (1995) the actual response to CO₂ is modified by environmental factors. The response of CO₂ fertilization to water stress increases with increasing water stress. In other words, water use efficiency (ratio of carbon gain per unit water used) is expected to increase under elevated CO₂, and water stressed plants are expected to benefit more from this effect. Furthermore, any potential CO₂ fertilization effect depends on the availability of nutrients to be realized. Consequently β₀ is modified by environmental modifiers accounting for these two effects (Eq. 3)

\[\begin{aligned} \beta=\beta _{0}\cdot f_{N}\cdot (2-f_{sw}) \end{aligned} \]

Eq. 3

where β₀ is the baseline response and f_N and f_sw are the nitrogen and soil water response functions also applied in modeling primary production. f_CO2 is calculated monthly to capture the monthly variation in the environment, however, intra-annual variations in C_t are neglected.

---

discussion

This method of an environmentally sensitive CO₂ response has been successfully applied in global vegetation modeling (e.g., Friedlingstein et al. 1995, Berthelot et al. 2005). Kicklighter et al. (1999) present a study comparing different approach at the approximate same level of process resolution, including the one adapted for iLand. Since this effect is not included in models such as PICUS or 3-PG, some more discussion points:

• It is important to note that the CO₂ fertilization effect influences GPP and not timber growth. Environment-related changes in allocation (e.g. increasing harshness increases the allocation to root compartments) might strongly influence the simulated response of stem growth (cf. the findings of Körner et al. 2005).
• Furthermore, absolute effects are related to GPP, which in turn is also environmentally sensitive, i.e. although the CO₂ fertilization effect is highest under strong water stress GPP is simultaneously reduced by water stress. The absolute effect is thus highly dynamic and will vary both spatially and temporally.
• We apply a generic CO₂ sensitivity, as supported by Norby et al. (2005). However, since environmental modifiers are species-specific also the CO₂ response will differ across species.
• We currently only model the effect of atmospheric CO₂ on productivity. Other aspects, such as adaptations in tree ontogeny or leaf physiology are currently neglected in iLand. Also allocation changes in response to elevated CO₂ are not taken into account explicitly. For future modifications in this regard Friedlingstein et al. (1999) present an approach that would be structurally suitable for implementation in iLand. However, within the scope of the current project, this effect is excluded, supported by the analysis of Friedlingstein et al. (1999), who found very little allocation change for temperate forests both in the Pacific Northwest and Central Europe.