Coppicing shifts CO2 stimulation of poplar productivity to above-ground pools: a synthesis of leaf to stand level results from the POP/EUROFACE experiment

Authors

  • Marion Liberloo,

    1. University of Antwerp, Research Group of Plant and Vegetation Ecology, Department of Biology, Campus Drie Eiken, Universiteitsplein 1, 2610 Wilrijk, Belgium
    Search for more papers by this author
  • Martin Lukac,

    1. NERC Centre for Population Biology, Division of Biology, Imperial College London, Silwood Park Campus, Ascot SL5 7PY, UK
    Search for more papers by this author
  • Carlo Calfapietra,

    1. University of Tuscia, DISAFRI, Via San Camillo De Lellis, I-01100 Viterbo, Italy
    2. National Research Council (CNR), Institute of Agro-Environmental & Forest Biology, Via Salaria km 29,300, 00015 Monterotondo Scalo (Roma), Italy
    Search for more papers by this author
  • Marcel R. Hoosbeek,

    1. Department of Environmental Sciences, Earth System Science – Climate Change group, Wageningen University, PO Box 47, 6700AA Wageningen, the Netherlands
    Search for more papers by this author
  • Birgit Gielen,

    1. University of Antwerp, Research Group of Plant and Vegetation Ecology, Department of Biology, Campus Drie Eiken, Universiteitsplein 1, 2610 Wilrijk, Belgium
    Search for more papers by this author
  • Franco Miglietta,

    1. Institute of Biometeorology – National Research Council (IBIMET-CNR), Via Caproni 8, 50145 Firenze, Italy
    Search for more papers by this author
  • Giuseppe E. Scarascia-Mugnozza,

    1. University of Tuscia, DISAFRI, Via San Camillo De Lellis, I-01100 Viterbo, Italy
    2. National Research Council (CNR), Institute of Agro-Environmental & Forest Biology, Via Salaria km 29,300, 00015 Monterotondo Scalo (Roma), Italy
    Search for more papers by this author
  • Reinhart Ceulemans

    1. University of Antwerp, Research Group of Plant and Vegetation Ecology, Department of Biology, Campus Drie Eiken, Universiteitsplein 1, 2610 Wilrijk, Belgium
    Search for more papers by this author

Author for correspondence:
Marion Liberloo
Tel: +32 3 820 22 79
Fax:+32 3 820 22 71
Email: marion.liberloo@ua.ac.be

Summary

A poplar short rotation coppice (SRC) grown for the production of bioenergy can combine carbon (C) storage with fossil fuel substitution. Here, we summarize the responses of a poplar (Populus) plantation to 6 yr of free air CO2 enrichment (POP/EUROFACE consisting of two rotation cycles). We show that a poplar plantation growing in nonlimiting light, nutrient and water conditions will significantly increase its productivity in elevated CO2 concentrations ([CO2]). Increased biomass yield resulted from an early growth enhancement and photosynthesis did not acclimate to elevated [CO2]. Sufficient nutrient availability, increased nitrogen use efficiency (NUE) and the large sink capacity of poplars contributed to the sustained increase in C uptake over 6 yr. Additional C taken up in high [CO2] was mainly invested into woody biomass pools. Coppicing increased yield by 66% and partly shifted the extra C uptake in elevated [CO2] to above-ground pools, as fine root biomass declined and its [CO2] stimulation disappeared. Mineral soil C increased equally in ambient and elevated [CO2] during the 6 yr experiment. However, elevated [CO2] increased the stabilization of C in the mineral soil. Increased productivity of a poplar SRC in elevated [CO2] may allow shorter rotation cycles, enhancing the viability of SRC for biofuel production.

Introduction

Forests are a major sink of carbon (C) on a global scale, as they contribute up to 70% of the terrestrial carbon fixation (Waring & Schlesinger, 1985; Melillo et al., 1993). Increased C sequestration in forest biomass or soils would therefore help to mitigate rising atmospheric CO2 concentrations ([CO2]). Tree plantations managed for the production of renewable bioenergy could combine C storage with the production of C neutral energy. Here, we show the results of a 6 yr CO2 study on a poplar short rotation coppice culture (SRC). The poplar trees, planted in typically high densities (Mitchell et al., 1999), were managed for the production of bioenergy from woody biomass and were growing under elevated [CO2] and soil nutrient fertilization treatments.

The CO2 fertilizing effect on forest net primary production (NPP) is well established. In their synthesis, Norby et al. (2005) showed that the NPP of four closed canopy temperate forests (including the one discussed here) increased by 23% on average when growing in free air carbon dioxide enrichment (FACE). The authors showed that in forests with a low leaf area index (LAI), the increased NPP in elevated [CO2] resulted mainly from increased light absorption through enhanced LAI, whereas in high-LAI forests NPP was enhanced through increased light use efficiency (LUE). In addition, a recent study of McCarthy et al. (2006) suggested that enhanced above-ground carbon storage may only occur when resource availability supports an increased LAI. However, there remain some important uncertainties that need to be resolved in order to predict future forest CO2 responses accurately.

Firstly, the capacity of a forest to sequester carbon in the long term depends, apart from its rate of carbon uptake, on the allocation of carbon to various carbon pools which have different turnover rates (DeLucia et al., 2005). Soils comprise up to three times as much carbon as vegetation (IPCC, 2001), and the long-term pools of soil carbon are stored in the most recalcitrant fractions (Six et al., 2002). Therefore, carbon storage in soils appears to be the best option for long-term carbon sequestration (Medlyn et al., 2005). Norby et al. (2005) found that the similar NPP response to CO2 of four FACE forests did not necessarily hold an equal trajectory for long-term C cycling. Across these four FACE sites, the percentage of NPP gain partitioned to wood, for instance, varied between 11 and 93%. At Duke forest, most of the extra carbon assimilated by the pine forest was allocated to wood production (DeLucia et al., 1999; Hamilton et al., 2002), whereas in the sweetgum forest at ORNL, the stimulation in wood production in the first year was replaced in subsequent years by an equal stimulation in fine root production (Norby et al., 2004). Understanding the controls on C partitioning is therefore of major importance.

Secondly, important questions remain about the impact of other environmental variables, such as nutrient availability, on NPP under elevated [CO2]. According to the theory of progressive nitrogen limitation (PNL; Luo et al., 2004), it is predicted that NPP will decline through time in the absence of an external N input. Increased NPP under elevated [CO2] could immobilize a larger amount of N in plant and microbial biomass, speeding up the onset of PNL (Finzi et al., 2006). However, increases in N use efficiency (NUE) decrease the amount of N required to maintain an increase in NPP, and therefore delay the onset of PNL (Finzi et al., 2002). Nutrient availability has, however, proved to be a decisive factor in regulating the impact of elevated [CO2] on NPP (Kinney & Lindroth, 1997; Sigurdsson, 2001; Reich et al., 2006). Given the close linkages between the C and N cycles, responses to N fertilization and CO2 enrichment are not likely to be simple or additive (Norby, 1998).

The potential of managed forests to act as a terrestrial C sink and produce biofuels that replace fossil fuels may depend on their management (Eriksson et al., 2007). Short rotation coppice (SRC) cultures are typically cut back and harvested every 3–5 yr to maximize yield (Sennerby-Forsse, 1995). However, harvesting has been found to both increase (Johnson & Curtis, 2001) and decrease (Yanai et al., 2003) carbon storage in soil organic matter (SOM). Fertilization of SRC can substantially increase productivity (Tamm, 1991) and may also increase C storage by decreasing CO2 efflux from the soil (Nohrstedt et al., 1989). It is therefore essential to assess any possible interactions between elevated [CO2] and different management practices, such as coppice and fertilization, on productivity and C storage.

The POP/EUROFACE experiment, established in 1999, combined a fast-growing agroforestry system capable of a substantial biomass production with a large-scale FACE system. The FACE technique has been shown to be one of the best approaches to expose plants to elevated [CO2] without altering the environmental conditions and to conduct research truly at the ecosystem level (Hendrey & Miglietta, 2006). During the last decade, several other large FACE experiments have been deployed, studying effects of elevated [CO2] on a mix of temperate tree species (sweetgum at ORNL: Norby et al., 2001; pine at Duke Forest: Finzi et al., 2002; five different aspen clones at AspenFACE: Dickson et al., 2000; a deciduous forest at Bangor FACE: A. R. Smith et al., unpublished). For a complete overview of the different FACE research stations, see http://public.ornl.gov/face/global_face.shtml.

In this synthesis, we present the main findings of the POP/EUROFACE experiment over the duration of the 6 yr experiment. We discuss the effects of elevated [CO2] on the productivity and the different above- and below-ground C pools of this intensively managed poplar bioenergy plantation in relation to different management regimes, such as coppicing vs single stems, fertilization and species choice. In particular, we investigate the combined effects of elevated [CO2] and fertilization on physiological processes and plant growth from the leaf level up to the tree and stand levels. Finally, we highlight the perspectives of a SRC for the production of bioenergy and C storage.

The POP/EUROFACE experimental design and plant material

The FACE facility was located in central Italy, near Viterbo, on 9 ha of agricultural land. In late spring 1999, the main plantation was established using hardwood cuttings of Populus × euramericana (Dode) Guinier (clone I-214). Within the main plantation, six experimental plots were planted with hardwood cuttings of three different species: a local selection of P. alba L. (clone 2AS-11), P. nigra L. (clone Jean Pourtet), and P. × euramericana (Dode) Guinier (clone I-214) (Fig. 1a,b). The plantation was designed and managed as a SRC with typically high plant densities (Mitchell et al., 1999). A FACE infrastructure was installed in three of the six experimental plots, while the other three plots were left under natural conditions and represented the ambient plots. Over the 6 yr of CO2 enrichment, the long-term mean [CO2] measured at the centre of the plots, just above the canopy, was equal to 540 ppm, on average, across the three FACE rings (Fig. 2). The elevated [CO2], measured at 1 min intervals was within 20% deviation from the target concentration of 550 ppm for 94% of the time during the first rotation, and for 78% of the time during the second rotation (Fig. 2). The quality of the [CO2] control by FACE (i.e. the variability of measured [CO2] against the target) declined with increasing canopy height. Such a decline was more pronounced in the second growth cycle when the height differences between the three poplar species became larger, thus creating larger air turbulence and faster vertical dispersion of released CO2 (Fig. 2). Daytime CO2 enrichment was provided from bud burst to leaf fall. A detailed description of the setup and of the initial performance of the FACE facility can be found in Miglietta et al. (2001).

Figure 1.

(a) Aerial view of the POP/EUROFACE site. The plantation covered 9 ha and consisted of three elevated [CO2] plots (CO2 enrichment to 550 ppm with FACE infrastructure) and three ambient [CO2] plots (380 ppm). (Source: Scarascia-Mugnozza et al., 2006, with kind permission of Springer Science and Business Media.) (b) Schematic layout of the plantation (black square, elevated [CO2] plot; white square, ambient [CO2] plot) (arrow points to the north) and of an experimental plot, planted with three poplar species –Populus alba (A), P. nigra (B) and P. × euramericana (C) – and consisting of fertilized and unfertilized halves (adapted from Gielen et al., 2003). A permanent growth plot with scaffoldings providing access to the canopy is marked in grey in the centre of each sector. (c) Schematic overview of the management and evolution of the POP/EUROFACE experiment. Poplars were planted as single stem cuttings in the spring of 1999 and grown for 6 yr under elevated [CO2]. Poplars were cut back after 3 yr, creating a multi-stem coppice. During the second rotation, half of each experimental plot was fertilized.

Figure 2.

Performance of the free air CO2 enrichment (FACE) system. Average monthly [CO2] in the three FACE rings during 6 yr of CO2 fumigation (closed circles, n = 3, ± SE). The open circles show on a monthly basis the fraction of time in which the [CO2] measured at the center of the ring and at the top of the canopy deviated by < 20% from the preset target [CO2] (550 ppm).

After 3 yr of growth at elevated [CO2], poplar trees were coppiced, and above-ground biomass was harvested. During the second rotation, several new shoots resprouted from the coppiced main stem and rooting systems of the first rotation cycle. Fertilization was applied only during the second rotation, since initial soil N content (between 7.7 and 10.4 µg [inline image and inline image] g−1 soil; Hoosbeek et al., 2004) was very high at the onset of the experiment. In the fertilized treatments (half of each plot), a total amount corresponding to 212 kg N (N-P2O5-K2O: 20-6-6) ha−1 during the first year and 290 kg N (ammonium nitrate: 34-0-0) ha−1 during the second and third years of the second rotation were supplied. An overview of the course of the POP/EUROFACE experiment is provided in Fig. 1c. For further plantation details, see Scarascia-Mugnozza et al. (2006).

Synthesis of results

A compilation of observed treatment effects in the POP/EUROFACE project, from the establishment of the plantation in 1999 until the final harvest in 2004, is given in Table 1. In general, three poplar species responded similarly to elevated [CO2] and fertilization, and therefore an average response across the three poplar species is shown. For more details on how different species responded, we refer to published studies as shown at the end of Table 1. Results in Table 1 are organized by leaf, tree and stand levels. Not all parameters were continuously measured throughout the two rotations for all three species, and empty white cells indicate no measurements.

Table 1. Summary of the average responses of three Populus species to elevated [CO2] (550 ppm), nitrogen (N) fertilization or CO2× N over the duration of the 6 yr POP/EUROFACE experimentThumbnail image of

Growth dynamics and productivity

Poplar growth was highly responsive to elevated [CO2] and at the end of the second rotation (i.e. after 6 yr of CO2 treatment), total woody biomass increased by 23%, within the range previously reported for trees (Ceulemans & Mousseau, 1994; Wullschleger et al., 1997; Ainsworth & Long, 2005). During the first year after planting and after coppicing, relative and absolute growth rates of poplar stems were strongly stimulated by elevated [CO2] (Table 1). In addition, after coppicing, poplars produced significantly more and thicker shoots in elevated [CO2] than in ambient [CO2] (Liberloo et al., 2004). This initial developmental enhancement increased stem basal area by 68 and 13% during the first growing season of the first and second rotations, respectively (Fig. 3a). Trees maintained larger basal area in elevated [CO2] during the remainder of each rotation, making the increase in basal area rather than in height the main factor contributing to larger yield in elevated [CO2] (Table 1). During the first rotation, above-ground biomass yielded 18–21 Mg ha−1 yr−1 in ambient and elevated [CO2], respectively (Calfapietra et al., 2003a). Coppicing the trees increased above-ground production and by the end of the second rotation, above-ground biomass production was 27 and 31 Mg ha−1 yr−1 in ambient and elevated [CO2], respectively (Liberloo et al., 2006).

Figure 3.

Average end of season stem basal area (a) and average yearly maximum leaf area index (LAImax) (b) of three poplar species (± SE) growing for 6 yr in elevated (closed circles) and ambient [CO2] (open circles). Significant CO2 effects are indicated with an asterisk (P ≤ 0.05). Data represent continuous measurements of trees growing in permanent growth plots and are adapted from Calfapietra et al. (2003b), Liberloo et al. (2004, 2005); unpublished (a) and from Gielen et al. (2001, 2003), and Liberloo et al. (2006) (b).

Initially, the development of poplar trees was enhanced during the first growing season of both rotations, but this stimulation in growth rate by elevated [CO2] disappeared during the second and third years when analyzing trees of similar size rather than of same age (Table 1; Calfapietra et al., 2003b; Liberloo et al., 2005). This initial growth enhancement corresponds to the findings for various herbaceous and tree species (Bazzaz et al., 1993; Norby et al., 1995; Poorter et al., 1996) and is in line with the findings of Centritto et al. (1999), who stated that elevated [CO2] speeds up early development, after which larger trees in elevated [CO2] ontogenetically reduce their growth rate (Jarvis & Jarvis, 1964). In addition, the ontogenetic decline after the first growing season was enhanced by an increased effect of competition. By the end of the first rotation, the stimulating CO2 effect was limited to only competitively advantaged trees (Calfapietra et al., 2003b). During the second rotation, a significant decline in the number of shoots per tree limited CO2 growth enhancement (Liberloo et al., 2004).

Similar to our findings, a sustained increase in basal area growth was found for pine trees growing in FACE for 8 yr (Moore et al., 2006). The average 18.7% enhancement of basal area increment (BAI) was also stronger in dominant trees. The faster growth of elevated [CO2] trees required more nitrogen from the soil (Finzi et al., 2002), such that suppressed trees could be more competitively disadvantaged in elevated [CO2] compared with ambient [CO2]. A similar response was found for five different aspen clones growing for 4 yr at AspenFACE (McDonald et al., 2002). Competition increased in the elevated [CO2] aspen stands; consequently, the size difference between large and small trees increased in elevated [CO2] and the CO2 response was greater for competitively advantaged individuals. However, at AspenFACE, the growth enhancement in elevated [CO2] was much weaker compared with POP/EUROFACE and Duke FACE. Aspen trees showed a significant increase in volume index only during the second and third years of fumigation (Isebrands et al., 2001). Results from the ORNL experiment show a positive enhancement of wood increment only during the first year of the FACE experiment; thereafter the increased carbon assimilation was allocated below ground to increased production of fine roots (Norby et al., 2002, 2004).

In POP/EUROFACE, elevated [CO2] significantly increased below-ground biomass production (Calfapietra et al., 2003a; Lukac et al., 2003). Total standing root biomass (i.e. coarse + fine roots) was enhanced by 47–76% in elevated [CO2] during the first rotation (Lukac et al., 2003) and poplars allocated 13% more of their standing root biomass into deeper soil horizons (Table 1) (Lukac et al., 2003). In addition, fine root turnover increased in elevated [CO2], although significantly only for P. alba and P. nigra (Lukac et al., 2003). This increase was explained by a decrease in life span of individual roots (Lukac et al., 2003), as was similarly found for aspen clones growing in elevated [CO2] (Kubiske et al., 1998). However, fine root turnover rate in a sweetgum stand declined when growing for 9 yr in elevated [CO2], attributed to both an increased root diameter and a deeper root proliferation (Iversen et al., 2008). A 6 yr minirhizotron study at Duke FACE revealed no significant CO2 effect on fine root turnover, although survival analysis of individual roots showed fine roots had shorter life spans in elevated [CO2] (Pritchard et al., 2008).

After coppicing the poplars, allocation of extra C taken up in elevated [CO2] shifted towards above-ground C pools. Coppicing had a negative impact on both coarse and fine root biomass and CO2 stimulation (Fig. 4a,b; Lukac et al., 2003). Fine root biomass declined shortly after coppicing in all three species (Fig. 4a), as large fine root biomass was no longer needed to support the smaller above-ground structures. Therefore, a temporal increase of fine root litter into the soil may have occurred immediately after coppicing. In addition, the stimulating CO2 effect on fine root biomass disappeared during the second rotation (Fig. 4a, Table 1). This suggests that water and nutrient acquisition capacity of the root systems before coppicing was far greater than the plant's requirements after coppicing. After coppicing, coarse root biomass continued to increase, although at a slower rate and with a slightly smaller positive CO2 effect on average coarse root biomass (Fig. 4b). Fine roots, because of their high turnover rate and high N and carbohydrate content, are a relatively large source of labile C and N in the soil (Iversen et al., 2008). Fine root turnover is therefore considered as one of the major pathways of new carbon entering soil C pools. These findings contrast with results from noncoppiced forests, such as the sweetgum stand (Norby et al., 2004), where fine root production of sweetgum trees almost doubled in elevated [CO2], contributing to a sustained increase in NPP. A strong above-ground biomass response in the first year was replaced by an equally strong response of fine root productivity from the second year onwards. Similarly, fine root production of aspen stands was higher when grown for 2 yr in elevated [CO2] (King et al., 2001). However, fine roots of pine trees at Duke FACE were less responsive to elevated [CO2], and the increase in NPP was mainly caused by the above-ground stem increment (Hamilton et al., 2002). The higher soil N availability at the POP/EUROFACE site compared with the other FACE sites might have decreased allocation towards fine roots as N requirements could have been more readily fulfilled.

Figure 4.

Average yearly maximum fine root standing biomass (a) and average yearly coarse root standing biomass (b) of three poplar species (± SE) growing for six consecutive years in the unfertilized, ambient (open circles) and elevated (closed circles) [CO2] plots. Significant CO2 effects are indicated with an asterisk (P ≤ 0.05). Data are adapted from Lukac et al. (2003); unpublished.

Leaf area index

Biomass yield is closely linked to leaf area, in particular in wood-producing crops (Hinckley et al., 1992). During the first year after planting, when poplars were still in their exponential growth phase, the LAI was significantly stimulated by elevated [CO2] (Fig. 3b) (Gielen et al., 2001). Poplar canopies were taller and contained more sylleptic branches (Gielen et al., 2002), carrying more and larger leaves (Ferris et al., 2001) (Table 1). However, the LAI stimulation disappeared after canopy closure, possibly as a result of increased leaf turnover caused by enhanced shading and competition in elevated [CO2] (Gielen et al., 2001). Results from leaf litter production confirmed the loss of LAI stimulation after canopy closure (Cotrufo et al., 2005).

In contrast with the first rotation, a larger LAI was maintained in elevated [CO2] until the end of the second growing season, despite early canopy closure already in the first growing season (Liberloo et al., 2005). The CO2-enriched poplars produced more shoots per tree and allocated more biomass to branches (Liberloo et al., 2006). However, by the end of the second rotation, this CO2 stimulation of LAI had disappeared.

Results on the LAI response from different elevated [CO2] experiments vary. In the POP/EUROFACE experiment, trees increased their LUE in elevated [CO2] across the two rotation cycles (Calfapietra et al., 2003a; Liberloo et al., 2006). Coppicing the poplars might have improved their efficiency of using the available space and light in the canopy by changing their growth form. This could explain why the LAI enhancement in elevated [CO2] was sustained longer during the second growing cycle. However, no structural changes in the canopy were found in an expanding sweetgum forest (Norby et al., 2003), and at Duke FACE, only small effects on LAI were found in the initial years of the experiment (DeLucia et al., 2002). However, exposure of aspen and birch trees in the aspen FACE experiment increased LAI, ascribed to the larger trees in elevated [CO2] (Karnosky et al., 2003), but trees at that time were still young and the canopy had not closed yet. Exposure of a scrub oak forest to elevated [CO2] did increase LAI, but this was before canopy closed (Hymus et al., 2003). Therefore, it seems that for a closed canopy with high LAI, the observed increase in NPP under elevated [CO2] can mainly be attributed to an increase in LUE, without an increase in LAI (DeLucia et al., 2005; Norby et al., 2005).

Consistent photosynthetic stimulation

Growing poplars for a period of 6 yr in elevated [CO2] increased Asat (light-saturated photosynthesis at growth [CO2]) by 31% on average across three species (Fig. 5; Table 1). During the entire first rotation, elevated [CO2] significantly stimulated photosynthesis (Bernacchi et al., 2003), despite the significant decreases in photosynthetic capacity in P. nigra (Vcmax, maximum carboxylation rate of Rubisco; and Jmax, maximum electron transport rate) and P. alba (Jmax). The gross primary production in the first rotation, estimated with a radiation transfer and energy balance model (Wittig et al., 2005), was significantly increased up to 251% in elevated [CO2] (Table 1). After canopy closure, this stimulation declined but still remained significant. Results obtained during the second rotation were, however, more contrasting. A study of Bernacchi et al. (2003) showed a large down-regulation of Vcmax and Jmax, especially for P. nigra, during the first year after coppicing, such that Asat of P. nigra and P. × euramericana was no longer increased under elevated [CO2]. However, two studies after 5 and 6 yr of growth in elevated [CO2] at the POP/EUROFACE experiment could not indicate any clear photosynthetic down-regulation (Calfapietra et al., 2005; Liberloo et al., 2007) and at the end of the second rotation cycle, Asat was stimulated by 49% averaged across three species (Liberloo et al., 2007).

Figure 5.

Mean response (% change) of photosynthesis and leaf photosynthetic parameters for three poplar species (± SE) growing during 6 yr in elevated [CO2]. Jmax, maximum rate of electron transport; Vcmax, maximum Rubisco carboxylation rate; gs, stomatal conductance; Asat, light saturated photosynthesis at growth [CO2]; SLA, specific leaf area; Na, nitrogen content on leaf area basis; Nm, nitrogen content on leaf mass basis; starch, starch content on leaf area basis. References to the individual yearly measurements are cited in Table 1.

It has been demonstrated that increased soluble sugars under elevated [CO2] promote the occurrence of down-regulation by decreasing the amounts of Rubisco protein through a hexokinase sensory system (Moore et al., 1998, 1999; Rogers et al., 1998; Long et al., 2004; Ainsworth & Rogers, 2007). However, in the POP/EUROFACE experiment, the concentration of soluble carbohydrates in both young and fully expanded leaves did not increase. By contrast, Davey et al. (2006) found large increases in starch concentrations (a mean increase of 85%; Fig. 5) but no changes in the amounts of Calvin cycle proteins, or in the starch synthetic enzyme ADP-glucose pyrophosphorylase (AGPase). The authors suggested that the observed reduction in photosynthetic capacity in the study of Bernacchi et al. (2003) was caused by short-term feedback responses occurring at the enzyme activity level, rather than by a change at the photosynthetic protein content level. The observed reductions in photosynthetic capacity were small: on average across 6 yr and the three species, Jmax decreased with 2% and Vcmax with 8% in elevated [CO2] (Fig. 5).

The sustained photosynthetic enhancement may in part be explained by the lack of a foliar nutrient deficiency. Nutrient limitations induced by rapid growth at elevated [CO2] can limit photosynthesis by reducing the amounts of Calvin cycle photosynthesis enzymes (Long & Drake, 1992). Leaf nitrogen content (expressed on a mass basis; Nm) was relatively high (on average 30 mg g−1) and significantly decreased by 8% in elevated [CO2] across 6 yr (Fig. 5, Table 1; Gielen et al., 2003; Tricker et al., 2004; Calfapietra et al., 2005, 2007; Liberloo et al., 2007). This consistent reduction in Nm matched the steady decrease of SLA (specific leaf area; cm2 g−1) in elevated [CO2] (mostly in the upper canopies, Fig. 5; Ferris et al., 2001; Gielen et al., 2001, 2003; Tricker et al., 2004; Calfapietra et al., 2005; Liberloo et al., 2007). However, leaf nitrogen content expressed on a leaf area basis (Na) generally did not change, indicating the decrease in Nm was a dilution effect caused, at least partly, by increased amounts of starch (Fig. 5).

Similarly, at the ORNL FACE site, down-regulation did not occur and photosynthetic enhancement was maintained for 3 yr (Sholtis et al., 2004), explained by the lack of a foliar nutrient deficiency. By contrast, photosynthesis of pine tree needles acclimated to elevated [CO2] at the Duke FACE site (Rogers & Ellsworth, 2002), although only for the 1-yr-old needles. A selective down-regulation of Rubisco following higher soluble carbohydrate concentrations caused this response, while leaf N concentrations were not deficient.

Increased photosynthesis at elevated [CO2] is often accompanied by a decrease in stomatal conductance (gs) (Medlyn et al., 2001). In the POP/EUROFACE experiment, gs was reduced 8% (Fig. 5) across 6 yr, but effects were variable. Initially, when canopies were still open, gs decreased because of a reduced stomatal density (Tricker et al., 2005). But when canopies gradually closed, stomatal closure dominated the negative stomatal response to elevated [CO2]. This significantly improved water use efficiency at the leaf level (Tricker et al., 2005). The improvement did not, however, translate to the stand level, as whole-tree sap flow was increased by elevated [CO2], by 12 and 23% before and after coppicing, respectively (Tricker et al., in press).

These findings suggest that there is a progression in stomatal sensitivity to elevated [CO2]. Hydrological CO2 responses can strongly depend on the scale of temporal and spatial observation (Wullschleger et al., 2002), but the response of gs may also vary considerably with other environmental factors, such as water availability (Wall et al., 2001; Gunderson et al., 2002; Ainsworth & Rogers, 2007). However, results of a 4 yr study on sweetgum at Duke FACE revealed a consistent and sustained reduction of gs across a variety of environmental conditions (Herrick et al., 2004).

Nitrogen availability

Fertilization increased leaf nitrogen content expressed on both a leaf area (Na) (Marinari et al., 2006) and a leaf mass (Nm) basis (Calfapietra et al., 2005; Marinari et al., 2006) initially, but also no effect of fertilization on leaf N content was found (Liberloo et al., 2007). Leaves of P. × euramericana were generally larger under fertilization, increasing their SLA (Calfapietra et al., 2005). Fertilization slightly affected the structure of poplar trees: fertilized trees produced more leaves in the top of the canopy (Liberloo et al., 2007). Below ground, fertilization decreased biomass of coarse roots of all species during the second year of the fertilization (M. Lukac, unpublished), and decreased stump biomass of P. alba by the end of the second rotation (Liberloo et al., 2006).

Despite these small effects of fertilization on leaf N content and size, tree structure and below-ground productivity, soil N availability in this SRC was sufficient to meet the needs for a sustained larger total mass production during 6 yr of growth in elevated [CO2].

How did these poplars avoid nutrient limitation when they consistently produced larger amounts of biomass in elevated [CO2]? Poplars were planted on former agricultural soil initially characterized by a high nutrient availability (for an overview of initial soil analysis, see Hoosbeek et al., 2004; Liberloo et al., 2006). Apart from this high background concentration of nutrients in the soil, increased fine root production, and thus soil exploration, by poplars in elevated [CO2] could also have alleviated N demand. However, this was only true during the first rotation, as coppicing the trees shifted below-ground C allocation from fine roots towards coarse roots and decreased the CO2 stimulation on fine roots (Fig. 4a,b). Furthermore, only P. alba significantly augmented the vertical exploration of the soil by allocating more roots to deeper soil horizons (Table 1; Lukac et al., 2003).

An additional pathway by which plants can increase soil exploration for N is through mycorrhizal symbiosis. Mycorrhizal fungi can help to sustain a positive response of NPP by increasing the plants’ nutrient uptake. During the first rotation, elevated [CO2] increased arbuscular mycorrhizal (AM) and ectomycorrhizal (ECM) infection of poplar roots by, on average, 23 and 24%, respectively, but the effect was species-specific (Table 1; Lukac et al., 2003). This response is slightly stronger than the 15% average increase in mycorrhizal abundance reported for temperate trees in elevated [CO2] in a meta-analysis by Treseder (2004). In the Duke FACE site, ECM significantly increased by 14% in elevated [CO2], whereas no consistent increase in AM was found (Garcia et al., 2008). However, Parrent & Vilgalys (2007) did not detect any CO2 effect on the extramatrical mycelia in the Duke FACE site; the increased investment in mycorrhizas was possibly preferentially invested into root-associated structures and not into the extramatrical hyphae.

In addition, poplar trees avoided nutrient limitation during the first rotation through increased NUE (Calfapietra et al., 2007). Increased NUE enhanced the amount of C fixed per unit of N taken up from the soil, decreasing the amount of N necessary to maintain high rates of NPP in elevated [CO2] (Finzi et al., 2002). Consequently, N concentration of most of the tree tissues decreased in elevated [CO2] (Cotrufo et al., 2005; Calfapietra et al., 2007). There were some indications that increased NUE was sustained over the second rotation, at least for P. × euramericana, which showed increased photosynthetic NUE and Vcmax/Na ratio at the end of the second rotation (Liberloo et al., 2007).

Our findings contrast with results from similar FACE experiments. While trees in POP/EUROFACE maintained high rates of NPP partly attributed to higher NUE, Finzi et al. (2007) showed that trees in the Duke Forest, ORNL and Aspen FACE site increased their N uptake by increased C allocation below ground. At the ORNL site, a deeper root proliferation of sweetgum trees supported the increased NPP (Norby & Iversen, 2006). At Duke FACE, Finzi et al. (2006) reported that 6 yr of growth in elevated [CO2] did not affect NUE on a plant tissue basis, but did widen the C-N ratio of the entire ecosystem. Although trees accumulated more N in the ecosystem through time, the initiation of PNL was precluded by an increase in C : N ratio of the entire ecosystem. Given that declining rates of mineralization were measured at this site, it is unclear how long this may delay the onset of PNL (Finzi et al., 2006).

C storage

In line with the source-sink balance theory of Stitt (1991) and in agreement with the hypothesis that the amount of sucrose produced by plants regulates the balance between carbon uptake and its consumption in sinks (Farrar, 1996; Miglietta et al., 1998), poplars were able to avoid photosynthetic down-regulation as a result of their capacity to accommodate extra carbon fixed in elevated [CO2]. Initially, poplar growth responded very positively to elevated [CO2], resulting in larger trees carrying more leaves and supporting larger rooting systems. However, coppicing the poplar trees partly shifted the extra carbon taken up in elevated [CO2] towards above-ground pools.

We determined the amount of carbon that was stored in the various above- and below-ground C pools at the end of each rotation in the unfertilized, ambient and elevated [CO2] plots (Fig. 6), following the methods reported in Gielen et al. (2005). Elevated [CO2] significantly increased total plant C, by 24% and 21% after the first and second rotations, respectively (Fig. 6). By the end of each rotation, C pools in stems, branches, stumps and coarse roots of elevated [CO2] grown poplars were significantly larger (Fig. 6). Although poplars increased LAI, this positive effect decreased throughout each rotation. At the end of the first rotation, elevated [CO2] did not affect the leaf litter C pool (Fig. 6; Gielen et al., 2005). However, the forest litter C pool was larger (though not significantly larger in the unfertilized treatments) in elevated [CO2] at the end of the second rotation cycle (Fig. 6; Hoosbeek & Scarascia-Mugnozza, 2008). Elevated [CO2] stimulated the fine root C pool more than the other pools during the first rotation (+77%), but this positive effect on fine roots strongly declined after coppicing (+25%) and was no longer significant in the second rotation. Coppicing did not affect relative CO2 enhancement of above-ground woody biomass or stump biomass (Fig. 6). The mean total amount of C (total plant C plus soil C) standing in this poplar SRC at the end of the second rotation was 7535 and 8043 g C m−2 in ambient and elevated [CO2] treatments, respectively. Across two rotation cycles, up to 6784 g C m−2 (or 64%) in elevated and 5671 g C m−2 (or 59%) in ambient [CO2] of the total forest carbon was stored in the above-ground harvestable biomass (i.e. the sum of above-ground biomass production of two harvests). The total plant biomass pools gradually exceeded the soil C pools (down to 0.2 m) (Fig. 6).

Figure 6.

Overview of above- and below-ground carbon pools at the end of the first and second rotations in the unfertilized, ambient and elevated [CO2] plots. Significant CO2 effects are indicated with an asterisk (P ≤ 0.05). Mean values for three poplars species ± SE are reported. References to the individual measurements are cited in Table 1. At the end of the first rotation, litter C pool was defined as the sum of litter C produced in 2001 and the C remaining on the forest floor after decomposition of the litter from previous years. Litter C at the end of the second rotation was directly measured by sampling the forest floor.

It has been shown that there is a strong coupling between carbon uptake by the canopy and below-ground metabolic processes (Steinmann et al., 2004). During the first rotation, elevated [CO2] enhanced soil C input via increased fine root biomass and turnover (Table 1, Figs 4a, 6). However, coppicing strongly reduced C allocation into fine roots, and at the end of the second rotation standing fine root biomass was no longer significantly stimulated (Figs 4a, 6). These findings are in contrast with findings from the ORNL sweetgum forest, where C in fine roots made up one-third of the NPP and the flux of carbon into the soil nearly doubled in elevated [CO2] because of this stimulated fine root production (Iversen et al., 2008).

An important pathway through which carbon entered the soil at the POP/EUROFACE was the mycorrhizal mycelium (Godbold et al., 2006), which was responsible for 62% of C input into SOM during the first rotation. Carbon allocation to mycorrhizas was not measured during the second rotation. It is possible that, despite the smaller fine root systems, the mycorrhizal mycelium could still have acted as a major carbohydrate sink in elevated [CO2], thus helping to avoid photosynthetic down-regulation and PNL.

As elevated [CO2] increased C storage in above- and below-ground biomass, it was expected that carbon storage in the elevated [CO2] soil would increase through an increased C input from litter, roots and mycorrhizas. Initial C content of the soil at the inception of the POP/EUROFACE experiment in the unfertilized plots was 1.08% (± 0.02) and 0.91% (± 0.12) in ambient and elevated [CO2], respectively (Hoosbeek et al., 2006a), and indicated no pre-treatment differences in soil C or soil N. However, at the end of the first rotation, total soil carbon content had increased significantly more in the ambient [CO2] plots than in the elevated [CO2] plots (Table 1, Fig. 7; Hoosbeek et al., 2004), even though the amount of newly stored C was significantly higher in the elevated [CO2] plots (Table 1). In addition, soil N content had decreased in elevated [CO2] as significantly more N was lost from the soil at 0–10 cm depth (Table 1; Hoosbeek et al., 2004). The authors hypothesized that the opposite effects of elevated [CO2] on total soil C and N, and Cnew during the first rotation were caused by a priming effect of the newly incorporated litter that increased decomposition of older SOM. In the elevated [CO2] plots, the soil contained more labile C (Hoosbeek et al., 2006a), which stimulated the microbial biomass and activity (Table 1; Moscatelli et al., 2005a,b).

Figure 7.

Change of top soil C (0–20 cm) in unfertilized, ambient (open circles) or elevated (closed circles) [CO2] plots compared with the start of the POP/EUROFACE experiment (1999). Significant changes in top soil C (P ≤ 0.05) are indicated with an asterisk.

By contrast, during the second rotation the priming effect had run its course (Hoosbeek et al., 2006a), as C content of the soil (0–20 cm) increased more in the elevated [CO2] plots (334 g C m−2) compared with the ambient [CO2] plots (238 g C m−2) (Fig. 7; Hoosbeek & Scarascia-Mugnozza, 2008). During the second rotation, microbial biomass N, expressed as a fraction of total N, significantly increased in elevated [CO2] (Table 1; Lagomarsino et al., 2008). Elevated [CO2] enhanced the immobilization of ammonium-N into microbial biomass (Lagomarsino et al., 2008). This decreased soil N availability through which the priming effect ceased (Hoosbeek & Scarascia-Mugnozza, 2008). The loss of older C during the first rotation and the increased accumulation of new C during the second rotation under FACE resulted in roughly equal increases of soil C under ambient CO2 and FACE during the 6 yr experiment. However, physical fractionation revealed that elevated [CO2] increased the stabilization of C in the mineral soil (Hoosbeek & Scarascia-Mugnozza, 2008).

In the sweetgum ORNL site, C input into soil organic matter (SOM) significantly increased in the top 5 cm of soil after 5 yr of elevated [CO2] exposure that increased root production (Jastrow et al., 2005). However, at the Duke FACE site, an increased C input to the forest floor from litter of pine trees was interpreted as a transient response, as fast turnover rates of the litter appeared to constrain the potential size of this carbon sink (Schlesinger & Lichter, 2001). Still, van Groenigen et al. (2006) reported in their recent meta-analysis that soils have the potential to increase their C storage by 5.4% in elevated [CO2]. But they also showed that this stimulation may be limited to sites where N is abundant and supplemented at rates > 30 kg ha−1 yr−1. The high N availability at the POP/EUROFACE site might therefore have facilitated the increase in soil C that arose after the priming effect ceased.

Conclusions

Tellingly, the choice of species may strongly affect the C sequestration capacity of a forest ecosystem (Hoosbeek et al., 2006b; Hyvönen et al., 2007). The three poplar species planted at the POP/EUROFACE experiment performed equally well when growing in an abundance of nutrients, sunlight and water, and produced very large amounts of woody biomass for potential use as biofuel. In addition, poplar yield of the three species was equally stimulated in elevated [CO2], by, on average, 20%, and coppicing the trees did not affect this increase in yield. On the contrary, coppicing did strongly increase above-ground productivity by 66%, suggesting that poplar wood from a SRC might be an excellent choice as a biofuel. The rotation length of fast-growing trees such as poplars is typically 3–5 yr as coppicing reinvigorates the trees, thus accelerating growth towards a theoretical maximum (Sennerby-Forsse, 1995). The predicted increase of productivity under future elevated [CO2] might reduce the optimal rotation length of a poplar SRC as increased competition for light and space would decrease the yield through self-thinning.

Coppicing the trees temporarily enhanced the dieback of fine roots and, as trees produced significantly more wood after coppicing, C allocation partly shifted towards above-ground pools and away from the soil. It is not likely that the fine root biomass would continue to decline further; it is more probable it would oscillate in sync with the coppice cycles. More long-term research beyond two rotation cycles is therefore needed to study the further development of the below-ground system under coppice management. Soils hold the best promise for long-term C storage as a result of the stabilization of SOM. Despite the fact that the increase of soil C in elevated [CO2] was not significant during the 6 yr experiment, physical fractionation revealed that the amount of stabilized SOM increased in elevated [CO2] (Hoosbeek & Scarascia-Mugnozza, 2008).

Although fertilization usually has a positive effect on soil C accumulation (Johnson & Curtis, 2001) and enhances forest productivity (Tamm, 1991), in our experiment fertilization did not interact with either C storage or biomass production. High initial soil nutrient status of this former agricultural land probably precluded any fertilization effect.

When considering the effects of forest management on C storage, one should also think about the life cycle of the removed biomass. Biofuels as an energy source reduce CO2 emissions by fossil fuel replacement (Eriksson et al., 2007). Using wood biomass as a biofuel does not remove carbon from the atmosphere in the long term, unless coupled with post-burn carbon capture and storage technology. It could be argued that, with regard to our findings of increased biomass production in high [CO2], the potential of SRC for C sequestration in wood will increase in the future. This, however, has to be hedged with the marginal increase of energy consumption needed to manage, harvest and process larger poplar biomass.

Acknowledgements

We wish to thank the reviewers for their useful comments and suggestions in rewriting this manuscript. In addition, we want to thank David Tissue for language revisions. We thank the entire POP/EUROFACE consortium for their collaboration during the course of the project. We want to thank all our Italian colleagues (DISAFRI, Viterbo, Italy) for their continuous help and logistic support. In particular, we would like to thank A. Zaldei for continued technical support to the experiment. Funding was provided by the EC Fifth Framework Program, Environment and Climate RTD Program, research contract EVR1-CT-2002-40027 (EUROFACE) and by the Centre of Excellence ‘Forest and Climate’ of the Italian Ministry of University and Research (MIUR). ML is a postdoctoral fellow on the Fund for Scientific Research-Flanders (F.W.O. Flanders).

Ancillary