Nutrient accumulation in aboveground biomass of planted tropical trees: a meta-analysis

Abstract

Efficient nutrient use is essential for biomass production by tropical trees growing in infertile soils. Accumulation of nitrogen (N) and phosphorus (P) in the aboveground biomass of four groups of tree stands [Acacia, Eucalyptus, N₂-fixing trees excluding Acacia, and other non-N₂-fixing broadleaved (ONNFB) trees] were investigated using meta-analyses of a range of biomass data to test the hypothesis that fast-growing Acacia and Eucalyptus trees accumulate fewer nutrients. Data for 83 tropical tree stands were selected from the literature. Standardized major axis regressions were applied between the log₁₀-transformed biomass and N or P accumulation. Nutrient use efficiency was compared with aboveground biomass and topsoil conditions. The slope of the regression between aboveground biomass and N accumulation for Eucalyptus was significantly smaller than the slopes for the N₂-fixing trees (excluding Acacia) and the ONNFB trees. N use efficiency of Eucalyptus increased with biomass more than that of N₂-fixing trees (excluding Acacia) and the ONNFB trees, because their stems and twigs tended to accumulate less N than in the other groups as biomass increased. The regressions between aboveground biomass and P accumulation had a common slope, and the intercepts of Acacia and Eucalyptus were significantly lower than that of ONNFB trees. P use efficiency of Acacia was consistently higher than that of the ONNFB trees. P use efficiency is more affected by other factors like soil conditions than is N use efficiency, and the differences in the tree groups significantly affect the use efficiency of both nutrients. These results explained some aspects of the general suitability of Acacia and Eucalyptus species for tropical plantations on infertile soils.

Key words:

• biomass production
• nitrogen
• nutrient use efficiency
• phosphorus
• tropical forest plantation

INTRODUCTION

Tropical forest plantations are increasingly important because of their ability to enhance biomass/carbon resources in an area (FAO 2006) and improve degraded soil organic matter after disturbances (Lugo 1997; Garay et al. 2004; Macedo et al. 2008; Wang et al. 2010). Fast-growing trees, including several species of Eucalyptus and Acacia, are often chosen for such purposes, because these trees can produce a profitable yield of more than 15 m³ ha⁻¹ year⁻¹ of biomass for small log production (Cossalter and Pye-Smith 2003). Weathered soils produced by long-term pedogenesis, which are usually infertile, have developed across wide areas of the humid tropics. In addition, harvesting operations and removal of tree biomass can cause serious nutrient deficits in forest soils (Nykvist 2000; Laclau et al. 2010; Dovey 2012). Wood biomass production is primarily regulated by nutrients, rather than water, in wet tropical areas (Jordan 1985). Species with the lowest nutrient demands may therefore be the best biomass producers in such soils.

Fast-growing tree species are generally considered to accumulate soil nutrients faster than slow-growing ones when multiple rotations are implemented (Cossalter and Pye-Smith 2003). However, whether nutrient accumulation per unit biomass differs between fast- and slow-growing trees is not known. Previous studies compared nutrient use efficiencies among several fast-growing trees (Kumar et al. 1998; Hiremath et al. 2002), some Acacia species (Kimaro et al. 2007) and some Eucalyptus species (Laclau et al. 2000), but differences in nutrient use efficiency among Acacia, Eucalyptus and other slower-growing trees are undocumented. Some fast-growing tree species have strategies that allow them to grow on degraded soils. In a nursery experiment, Acacia mangium Willd. had enhanced phosphorus (P) use efficiency under P-limited conditions (Ribet and Drevon 1996). Some Acacia species selectively translocate a large proportion of P (relative to nitrogen, N) away from their leaves before senescence (He et al. 2011; Inagaki et al. 2011). Some Eucalyptus plantations experience higher primary productivity with a low N investment (Binkley et al. 2004). Therefore, some Acacia and Eucalyptus trees may use nutrients more efficiently than other species.

Both N and P are major nutrients for tree growth (Chapin et al. 2011). However, they differ in their sources (Chapin et al. 2011), site availability (McGroddy et al. 2004; Chapin et al. 2011), metabolic forms (Matzek and Vitousek 2009) and degrees of translocation in plant organs (Laclau et al. 2001; Hättenschwiler et al. 2008; Inagaki et al. 2011). Therefore, they may not accumulate in aboveground biomass in the same way. Recent meta-analyses have clarified the relationship between leaf nutrient concentrations and some environmental and genetic factors (e.g. McGroddy et al. 2004; Townsend et al. 2007; Fyllas et al. 2009). A study of leaf analyses across the Amazon Basin revealed that leaf N is influenced mainly by genetic factors (family/genus/species), while leaf P is more strongly affected by the environment (Fyllas et al. 2009). Thus, the differing degree of structural dependency on these nutrients should be taken into account when assessing genetic differences in these levels in trees growing under different environmental conditions. In addition, the factors regulating N and P levels in plant organs other than leaves are poorly known (Kerkhoff et al. 2006). Understanding the accumulation of nutrients and total aboveground biomass in plants, including woody parts, will require additional research.

Trees increase the ratio of structural organs (woody parts) to metabolic organs (leaves) during their growth (Mori et al. 2010; Dovey 2012). When nutrient accumulation in aboveground biomass is considered, shift in investment between metabolic and structural organs must be considered. To evaluate this shift, forest stands ranging from juvenile to mature are needed for analysis. In contrast to leaf analyses, measuring the aboveground biomass and accumulated nutrients of entire trees is difficult because it requires destructive measurements. Until recently, very limited information on the aboveground biomass of tropical trees has been available. The results of a few international research projects, such as “Site management and productivity in tropical plantation forests” conducted by the Center for International Forestry Research (CIFOR), have improved the body of knowledge and, recently, a large amount of biomass and nutrient content data has become available (Nambiar et al. 2004; Nambiar 2008). Most of these studies have included soil parameters that help to elucidate the nitrogen use efficiency (NUE) and phosphorus use efficiency (PUE) characteristics of species growing on various levels of soil fertility.

We investigated the accumulation and allocation of N and P in the aboveground biomass of four tree groups [Acacia trees, Eucalyptus trees, N₂-fixing trees excluding Acacia, and other non-N₂-fixing broadleaved (ONNFB) trees], across a range of biomasses, using a meta-analysis. The aim of this study was to test the hypothesis that fast-growing Acacia and Eucalyptus tree species use smaller amounts of nutrients per unit of aboveground biomass for their growth than other tree species. Effects of topsoil conditions on N and P accumulation in tree biomass are also discussed.

MATERIALS AND METHODS

Sources of data on aboveground biomass and nutrient accumulation in tropical timber trees

Aboveground biomass and nutrient data were derived from reports on tropical plantation trees. The majority of the data was taken from a CIFOR report (Yamada et al. 2004), while other information was gleaned from the literature (see Appendix 1). Data for conifers were omitted because only two datasets were available. We compiled data for 83 tropical tree stands: 20 stands of Acacia species, 21 stands of N₂-fixing species excluding Acacia (hereafter simply called “N₂-fixing species”), 26 stands of Eucalyptus species, and 16 stands of ONNFB species. Acacia and Eucalyptus are two large genera of fast-growing trees and are planted throughout much of the tropics and in some temperate areas (Cossalter and Pye-Smith 2003). Most of the N₂-fixing trees sampled belong to the Fabaceae, with the exception of Casuarina equisetifolia L. (Casuarinaceae). The ONNFB trees sampled grew in India and Central America, since the available data were limited to these areas.

The mean annual increments (MAI) of biomass (dry weight basis) of Acacia and Eucalyptus were 16.3 and 15.7 Mg ha⁻¹ year⁻¹, respectively (Fig. 1). These MAI exceeded 15 m³ ha⁻¹ year⁻¹, given the wood densities of 0.55 g cm⁻³ for Acacia and 0.64 g cm⁻³ for Eucalyptus (Ogata et al. 2008). However, the MAI of ONNFB trees was significantly smaller than those of Acacia and Eucalyptus, and the MAI of N₂-fixing trees was two-thirds of those of Acacia and Eucalyptus. In this study, we considered Acacia and Eucalyptus to be fast-growing trees, although a broader definition would encompass some rapidly growing trees in the N₂-fixing and ONNFB tree groups. Total aboveground biomass was partitioned into (1) leaf biomass and (2) stem + twig biomass; we analyzed nutrient accumulation in these structural units.

Figure 1 Mean annual increment (MAI) of aboveground biomass in four tree groups: Acacia, Eucalyptus, other non-nitrogen (N₂)-fixing broadleaved (ONNFB) trees, and N₂-fixing trees excluding Acacia. Bars indicate the standard error of the mean. Letters indicate significant differences among tree groups according to analysis of variance and post-hoc Tukey-Kramer honestly significant difference (HSD) tests (P < 0.05). Numbers in parentheses after the names of the tree groups indicate the numbers of tree stands.

Total N and available P concentrations in the surface soil layer, mostly at depths of 0–10 cm, were used. Although the extraction methods for available P differed among studies, we did not adjust the values because no method exists to account for a variety of soil conditions. The soil data from some A. mangium stands (Nykvist and Sim 2009) were provided as mass values (kg ha⁻¹) in the original literature, so we estimated the concentrations using assumed bulk density (1.2). Some of the data were supplied by N. Nykvist (pers. comm.).

Nitrogen and phosphorus use efficiency

In this study, we defined NUE and PUE as units of aboveground biomass per accumulated N and P, respectively (Kumar et al. 1998; Laulau et al. 2000; Kimaro et al. 2007). The original definition of nutrient use efficiency was net primary production per amount of nutrients taken up (Hirose 1975). We omitted belowground and litterfall components as these are not removed during harvesting. We targeted a range of tree ages when analyzing aboveground biomasses. Biomasses for the four tree groups were: (1) Acacia trees (6–326 Mg ha⁻¹), (2) N₂-fixing trees (18.2−121 Mg ha⁻¹), (3) Eucalyptus trees (5.9−324 Mg ha⁻¹) and (4) ONNFB trees (2.5–191 Mg ha⁻¹) (see Appendix 1).

Statistical analyses

We applied standardized major axis (SMA) regression (Sokal and Rohlf 2011) to examine relationships between the log₁₀-transformed biomass (Mg ha⁻¹) and accumulated N or P (kg ha⁻¹). All biomass data were categorized into one of the four tree groups listed above. The differences in slopes and intercepts of the SMA regressions were tested between these four tree groups using Sidak’s adjusted pair-wise test, depending on the significance of the null hypothesis that the slopes were equal. Analysis of covariance (ANCOVA) was applied to NUE and PUE with the factors of aboveground biomass and the tree group. Differences in slopes were tested with Holm’s adjusted pairwise test, and the differences in intercepts were tested using the Tukey-Kramer honestly significant difference (HSD) test. Log₁₀-transformed NUE and PUE were compared by analysis of variance (ANOVA) among species represented by more than four stands. A post-hoc Tukey-Kramer HSD test was performed if the ANOVA was significant. All statistical analyses were performed using R 2.15.3 (R Core Team 2013) with the SMATR 3.3 package (Warton et al. 2012).

RESULTS

Relationship between aboveground biomass and N and P accumulation

To evaluate the biomass allocation and nutrient accumulation in an entire tree and in each structural unit at different growth stages, we compared total aboveground biomass, leaf biomass and stem + twig biomass to N and P accumulation. All the regressions comparing biomass to N and P accumulation across the four tree groups were significant except those between total aboveground biomass as well as stem + twig biomass and P accumulation for ONNFB trees (P = 0.33 and 0.62, respectively; Table 1). Although almost all of the regression slopes were similar among tree groups, the slopes of the regressions between total aboveground biomass and N accumulation and between stem + twig biomass and N accumulation were significantly different (P < 0.001). The slope for the regression of total aboveground biomass and N accumulation of Eucalyptus (0.615; 95% confidential interval, 0.535–0.706) was significantly lower (P < 0.05) than 1 and those for the N₂-fixing and ONNFB trees (0.848–1.027). The slope of the regression of total aboveground biomass and N accumulation of Acacia was also lower than 1 (0.751; 95% confidential interval, 0.638–0.884). Similarly, the slope of the regression of stem + twig biomass and N accumulation of Eucalyptus (0.485) was lower (P < 0.05) than those of the other tree groups (0.875–1.194), including the slope of Acacia. Since the slope of the regression between leaf biomass and N accumulation in Eucalyptus was not significantly different from those of the other tree groups, the declined slope between total aboveground biomass and N accumulation for Eucalyptus was caused by the lower N accumulation in stems + twigs with increasing biomass. The intercept of the regression between leaf biomass and N accumulation for Eucalyptus (1.143) was significantly smaller (P < 0.05) than those of other tree groups (1.307–1.445), and thus leaf N concentration of Eucalyptus was lower.

Table 1 Parameters of the standardized major axis (SMA) regressions for the relationships between total aboveground, leaf and stem + twig biomasses and nitrogen (N) and phosphorus (P) accumulation among four tree groups

	H0: slopes are equal	H0: no difference in intercept		Acacia (20)		N2-fixing trees ex. Acacia (21)		Eucalyptus (26)		Other non N2-fixing broadleaved trees (16)
Total aboveground biomass vs. N accumulation	0.001	< 0.001	Slope (a)	0.751	ab	0.848	a	0.615	b	1.027	a
			Intercept (b)	1.203		1.008		1.185		0.669
			R^2	0.89		0.88		0.89		0.82
			P-value	< 0.001		< 0.001		< 0.001		< 0.001
Total aboveground biomass vs. P accumulation	0.430	< 0.001	Slope (a)	0.790		0.790		0.790		0.790
			Intercept (b)	−0.266	b	−0.060	ab	−0.134	b	0.134	a
			R^2	0.63		0.33		0.50		0.07
			P-value	< 0.001		0.0061		< 0.001		0.3265
Leaf biomass vs. leaf N accumulation	0.394	< 0.001	Slope (a)	0.985		0.985		0.985		0.985
			Intercept (b)	1.419	a	1.445	ab	1.143	c	1.307	b
			R^2	0.97		0.90		0.84		0.77
			P-value	< 0.001		< 0.001		< 0.001		< 0.001
Leaf biomass vs. leaf P accumulation	0.088	< 0.001	Slope (a)	0.936		0.936		0.936		0.936
			Intercept (b)	−0.023	b	0.160	a	−0.020	b	0.161	a
			R^2	0.79		0.79		0.87		0.65
			P-value	< 0.001		< 0.001		< 0.001		< 0.001
Stem + twig biomass vs. stem + twig N accumulation	< 0.001	< 0.001	Slope (a)	0.944	a	0.875	a	0.485	b	1.194	a
			Intercept (b)	0.693		0.849		1.268		0.218
			R^2	0.92		0.81		0.88		0.71
			P-value	< 0.001		< 0.001		< 0.001		< 0.001
Stem + twig biomass vs. stem + twig P accumulation	0.226	< 0.001	Slope (a)	0.881		0.881		0.881		0.881
			Intercept (b)	−0.566	b	−0.317	ab	−0.361	ab	−0.088	a
			R^2	0.63		0.23		0.41		0.02
			P-value	< 0.001		0.0295		0.0014		0.6169

Slope (a) and intercept (b) indicate the parameters in the equation; log(y) = b + a log(x), where x is biomass and y is N or P accumulation in biomass. Letters indicate significant differences (P < 0.05) in slopes or differences in intercepts when a common slope was assumed. Numbers in parentheses after the names of the tree groups indicate the numbers of tree stands.

The common slope for the regressions between total aboveground biomass and P accumulation was 0.790 (95% confidential interval, 0.665–0.940) and significantly lower than 1. The intercepts for the regressions between total aboveground biomass and P accumulation for Acacia and Eucalyptus (−0.266 and −0.134, respectively) were significantly smaller than that for ONNFB trees (0.134) when the same slope was assumed. Correspondingly, the intercept of the regression between stem + twig biomass and P accumulation for Acacia (−0.566) was significantly lower than that for ONNFB trees (−0.088).

The coefficients of determination of the regressions were consistently higher in the relationships between total aboveground biomass and N accumulation (R² = 0.82–0.89) than the relationships between total aboveground biomass and P accumulation (R² = 0.07–0.63). Higher coefficients of determination were also observed between stem + twig biomass and N accumulation (R² = 0.71–0.92) than between stem + twig biomass and P accumulation (R² = 0.02–0.63). Those relationships between leaf biomass and N and P accumulations were not very different (R² = 0.77–0.97 and 0.65–0.87, respectively).

NUE and PUE for the four tree groups and each species

Results of the ANCOVA for the effects of total aboveground biomass and tree group on NUE were significant (adjusted R² = 0.71, F = 29.01, P < 0.001) for each factor (P < 0.001) and for their interaction (P = 0.0063). Aboveground biomass and NUE were positively correlated, and the least square mean slope was highest in Eucalyptus (Fig. 2). The slope for Eucalyptus was significantly higher than the slopes for the N₂-fixing and ONNFB trees (P < 0.05; adjusted pairwise test). The ANCOVA results for the same factors on PUE were also significant (adjusted R² = 0.43, F = 9.85, P < 0.001). The effects of aboveground biomass and tree group were significant (P < 0.001 and = 0.007, respectively), but their interaction was not (P = 0.48). Assuming a positive common slope, the intercepts were significantly larger in Acacia than in ONNFB trees (P < 0.05; Tukey-Kramer’s HSD test).

Figure 2 Relationships between total aboveground biomass and nitrogen use efficiency (NUE) and phosphorus use efficiency (PUE). The regression lines indicate the results of analysis of covariance for four tree groups: Acacia, Eucalyptus, other non-nitrogen (N₂)-fixing broadleaved (ONNFB) trees, and N₂-fixing trees excluding Acacia. Letters indicate significant differences in slopes for NUE due to the significant differences in the interaction, and in intercepts for PUE among tree groups assuming a common slope (P < 0.05). Numbers in parentheses after the names of the tree groups indicate the numbers of tree stands.

Average NUE of aboveground biomass for Eucalyptus grandis Hill ex Maiden was about 2.5-fold higher than those for ONNFB, N₂-fixing and Acacia trees (Fig. 3). The data for E. grandis stands represented more aboveground biomass (33.7–324.3 Mg ha⁻¹) than that for stands of Eucalyptus spp. (5.9–195.2 Mg ha⁻¹), whereas that for other Eucalyptus spp. represented stands of a broader age range, including 1-year-old stands in the metadata (see Appendix 1). Therefore, the NUE of E. grandis was high and was significantly greater than those of most of N₂-fixing trees and ONNFB trees (P < 0.05; Fig. 3). The average PUEs of A. mangium, E. grandis and Eucalyptus camaldulensis Dehnh. were 3.0–4.6-fold higher than those of ONNFB trees (P < 0.05).

Figure 3 Nitrogen use efficiency (NUE) and phosphorus use efficiency (PUE) for aboveground biomass of each of several species (n ≥ 4) and the remaining members of their tree groups. Letters indicate significant differences among species according to a post-hoc Tukey-Kramer honestly significant difference (HSD) test (P < 0.05). Numbers in parentheses indicate the numbers of plots. The top, bottom and middle lines in each box correspond to the 75th, 25th and 50th (median) percentiles, respectively. Whiskers extend from upper and lower quartiles ± 1.5 × (interquartile ranges). Asterisks indicate the arithmetical means.

Relationship between topsoil conditions and NUE and PUE

Because the metadata included various soil conditions (see Appendix 1), the relationships between total N and NUE and between available P and PUE were compared in the topsoils (Fig. 4). The regression between the total N and NUE was not significant (P > 0.05), whereas the SMA regression between available P and log₁₀-transformed PUE was significant (P = 0.019; R² = 0.098), although the P extraction methods varied.

Figure 4 Relationship between total nitrogen (N) in the topsoil and nitrogen use efficiency (NUE, left) and between available phosphorus (P) in the topsoil and phosphorus use efficiency (PUE, right).

DISCUSSION

Nutrient accumulation and allocation in Acacia and Eucalyptus

Acacia accumulated less P than did ONNFB trees. Larger Eucalyptus stands had higher NUE, and some Eucalyptus species tended to have higher PUE than did ONNFB trees. These results support the hypothesis that fast-growing Acacia and Eucalyptus tree species use fewer nutrients per unit of aboveground biomass, with the exception of N in Acacia.

P accumulation per unit biomass in Acacia and Eucalyptus tended to be small throughout the biomass range. Agricultural leguminous plants generally require a relatively large amount of P, with the exception of some tropical trees (Vitousek et al. 2002). For Acacia, the results corresponded with those of a pot experiment in which high PUE occurred under P-limited conditions (Ribet and Drevon 1996). Some Acacia species, including A. mangium, can selectively translocate a large proportion (> 70% at maximum) of P before leaf senescence (He et al. 2011; Inagaki et al. 2011). The efficiency of P resorption in Eucalyptus has been reported to be more than 50% in many studies (Fife et al. 2008; Yuan and Chen 2009). Thus, P accumulation in aboveground biomass and P usage by both Acacia and Eucalyptus may be smaller than in other species, because P in litterfall is also small. In wet tropical regions, P tends to limit primary production more than N (Chapin et al. 2011). Such lower PUE would be an advantage under P-limited conditions. Acacia and Eucalyptus may be competitive with other species in terms of P accumulation.

The common slopes of the regressions between total aboveground biomass and P accumulation and between total aboveground biomass and PUE may indicate that the ratios of P accumulation in stem + twig biomass to P accumulation in leaf biomass are similar among all tree groups at each biomass size. PUE of all the tree groups increased with the total aboveground biomass increase, and the common slope of the regressions between total aboveground biomass and P accumulation was less than 1. These results suggested that P concentrations in total aboveground biomass of all the trees gradually decreases during growth, and the differences in regression intercepts indicated the differences in the P concentrations of total aboveground biomass of the tree groups. Tropical trees tend to have a smaller ratio of leaf biomass to total aboveground biomass than other trees; our meta-data indicated that this percentage for all meta-data was less than 36%, and the percentages for larger-biomass trees (> 100 Mg ha⁻¹) were only 2–12% (see Appendix 1). Therefore, P accumulation in stem + twig biomass became a large proportion of total P accumulations. In the case of Eucalyptus stands, P concentrations in stems were less than a half of those in leaves (Bouvet et al. 2013). The increase in woody structures, which contain less P, may enhance PUE during the growth periods of all the tree groups.

N accumulation in Eucalyptus was less than in the N₂-fixing and ONNFB trees as biomass increased. The NUE increased for Eucalyptus relative to these other tree groups as aboveground biomass became larger (Fig. 3). These results suggested that N concentrations in stems and twigs were low or decreased as the Eucalyptus trees grew. Binkley et al. (2004) noted that resource use efficiency (including NUE) increased with primary production in a Eucalyptus plantation dataset. A growth model of Eucalyptus also indicated that N concentrations in stem and twig biomass decreased as stem diameter increased, whereas stem P concentrations remained constant over time (Bouvet et al. 2013). Our findings supported those results. Our results also indicated that the slope of the regression between stem + twig biomass and N of Eucalyptus (0.485) was only about a half of the slopes of the other tree groups (0.875–1.194). Both NUE and PUE in Eucalyptus increased with aboveground biomass; the slope between stem + twig biomass versus their N accumulation (0.485) was distinctly lower than that between stem + twig biomass and P accumulation (0.881). Therefore, we hypothesized that Eucalyptus may selectively translocate a large proportion of N during wood senescence. While the ratio of woody parts to leaves increased as growth progressed, N concentration in woody cells may decrease. In addition, the differences in the regression slopes between total aboveground biomass and N accumulation may derive from differences in the degree of translocation of N from the woody structures in each tree group. However, more than 50% of P was translocated from stem-wood during senescence in Eucalyptus clone forests (Laclau et al. 2001; Sette et al. 2013), while less than 30% of N was translocated from stem-wood (Laclau et al. 2001). The translocation rates of N and P in twigs and in stems of other species are still poorly understood, and further studies are needed to understand why Eucalyptus forests had a lower slope.

Another possible reason for the differences in the slopes is differences in the ratios of N accumulation in stem + twig biomass to that in leaf biomass among the tree groups. The N concentration in leaves is more than 10-fold higher than that in wood in some forests (Wang et al. 1991; Nykvist et al. 1996; Stanley and Montagnini 1999; Bouvet et al. 2013), and these differences in N concentrations tend to exceed those in P concentrations. In contrast to P accumulation, some tree groups may have a higher ratio of N accumulation in leaves to N accumulation in stems + twigs, which may explain why the slopes between total aboveground biomass and N accumulation of the N₂-fixing and ONNFB trees were not significantly lower than 1.

Acacia and N₂-fixing trees contain large amounts of N. They harbor symbionts that are able to fix a considerable amount of N from atmospheric N₂; not all their accumulated N comes from available soil sources. In the case of A. mangium, the percentages of accumulated N from fixation have been estimated to be 24–65% (Galiana et al. 2002), 10–59% (Bouillet et al. 2008) and 54 (± 22)% (Mercado et al. 2011). Therefore, the amount of soil N used by N₂-fixing trees may be small even though they have high demands for N (Inagaki et al. 2009).

The smaller accumulations of N in Eucalyptus than in N₂-fixing and ONNFB trees and of P in Acacia and Eucalyptus than in ONNFB trees suggested that some Eucalyptus and Acacia trees used in tropical plantations can produce more biomass than can ONNFB trees per unit resource (N and P). Whether such efficient nutrient use was due to their larger MAI or to the properties of the Eucalyptus and Acacia species used in this study is difficult to determine. MAI is not a suitable parameter to use to answer this question because growth rates are not consistent among tree ages and species (West 2006). Our results were also limited to the species for which metadata were available (see Appendix 1). To better understand nutrient use in trees, detailed growth information for each species should be prepared. Additional studies concerning nutrient translocation in trees will also further our knowledge (Laclau et al. 2001; He et al. 2011; Inagaki et al. 2011; Sette et al. 2013).

N and P accumulations under various soil conditions

More than 80% of the N accumulation in total aboveground biomass was explained by the differences among the tree groups using the regressions, even though the metadata included various soil conditions. The total N in the topsoil did not influence NUE. In contrast, only 7–63% of the P accumulation in total aboveground biomass was explained by the differences among the tree groups, indicating that other factors influenced the P accumulation. Although the P extraction methods were different among the studies, the available P in the topsoil is likely to affect PUE.

Leaf and leaf-litter P of tropical trees tend to be more affected by soil conditions than are N levels. Fyllas et al. (2009) reported that leaf P is primarily regulated by environmental factors (47%), followed by genetic factors (23%), whereas leaf N is primarily regulated by genetic factors (ca. 40%). In comparisons of 45 tropical trees, the leaf-litter P concentration varied by a factor of 7, whereas the litter N concentration varied only by a factor of 3 (Hättenschwiler et al. 2008). The primary reason that P levels in plant structures are more influenced by environmental factors than are N levels may depend on the mobility of N and P in plants. In leaves, more than 50% of N is used to produce insoluble compounds such as photosynthetic enzymes (Rubisco) and proteins (Chapin et al. 2011). P in leaves seems to form more labile and mobile compounds; labile P, which consisted of >50% inorganic P, comprised most of the total P in the fresh litter horizon of temperate pine forests (Polglase et al. 1992). Thus, leaf P concentrations fluctuate more than leaf N ones, and the N:P ratio of leaf litter tends to be higher than that of fresh leaves in tropical forests with infertile soil because P resorption during leaf senescence increases under P-limited conditions (McGroddy et al. 2004; Hättenschwiler et al. 2008). Although the forms of N and P found in wood are poorly known, previous studies indicated that P concentrations in woody inner-stem cells produced during the previous year are considerably lower than those in woody cells produced during the current year in Eucalyptus (Laclau et al. 2001; Sette et al. 2013). Ostertag (2010) suggested that P tended to be stored after P fertilization to guard against possible P starvation, in Hawaiian forests. These differences in mobility of N and P may inform our results, i.e., PUE is affected by other factors more than by the tree groups, compared to NUE.

Although other factors including soil conditions may affect the nutrient use efficiency, particularly in PUE, the differences among the tree groups significantly affect NUE and PUE. The lower nutrient requirements for biomass production by Acacia and Eucalyptus can explain why they survive on degraded soils.

CONCLUSIONS

The results of this study indicated that some Eucalyptus trees can accumulate more aboveground biomass than N₂-fixing and ONNFB trees while incorporating less aboveground N. Furthermore, some Acacia and Eucalyptus trees can produce more aboveground biomass than ONNFB trees while using less P. While PUE is more affected by external factors, like soil conditions, than is NUE, differences among the tree groups significantly affected NUE and PUE. Although factors such as water availability and light capture should also be considered, our results explained some aspects of the general suitability of these tree groups for tropical plantations with infertile soils.

SUPPLEMENTARY MATERIAL

Supplementary material for this article is available online from http://dx.doi.org/10.1080/00380768.2014.929025

ACKNOWLEDGMENTS

We thank Drs. Hisashi Abe, Hidetoshi Shigenaga and Shigehiro Ishizuka at the Forestry and Forest Products Research Institute, Japan, and Dr. Shin Ugawa at Kagoshima University, Japan, for their valuable comments on this manuscript. We also thank Dr. Nils Nykvist in Rönninge, Sweden, for access to detailed soil data from the Mendolong site in Sabah, Malaysia. This study was financially supported by the Ministry of Agriculture, Forestry and Fisheries, Japan.