Stand Age Affects Biomass Allocation and Allometric Models for Biomass Estimation: A Case Study of Two Eucalypts Hybrids

Abstract

We studied the effects of stand age on the allocation of biomass and allometric relationships among component biomass in five stands ages (1, 3, 5, 7, and 8 years old) of two eucalypts hybrids, including Eucalyptus urophylla × E. grandis and E. urophylla × E. tereticornis, in the Leizhou Peninsula, China. The stem, bark, branch, leaf, and root biomass from 60 destructively harvested trees were quantified. Allometric models were applied to examine the relationship between the tree component biomass and predictor variable (diameter at breast height, D, and height, H). Stand age was introduced into the allometric models to explore the effect of stand age on biomass estimation. The results showed the following: (1) Stand age significantly affected the distribution of biomass in each component. The proportion of stem biomass to total tree biomass increased with stand age, the proportions of bark, branch, and leaf biomass to total tree biomass decreased with stand age, and the proportion of root biomass to total tree biomass first decreased and then increased with stand age. (2) There were close allometric relationships between biomass (i.e., the components biomass, aboveground biomass, and total biomass per tree) and diameter at breast height (D), height (H), the product of diameter at breast height and tree height (DH), and the product of the square of the diameter at breast height and tree height (D²H). The allometric relationship between biomass and measurement parameters (D, H, DH, D²H) could be applied to the biomass assessment of eucalypts plantation. (3) Allometric equations that included stand age as a complementary variable significantly improved the fit and enhanced the accuracy of biomass estimates. The optimal independent variable for the biomass prediction model varied according to each organ. These results indicate that stand age has an important influence on biomass allocation. Allometric equations considering stand age could improve the accuracy of carbon sequestration estimates in plantations.

1. Introduction

At present, global climate change, such as global warming, extreme weather happening frequently, glacier retreating, and global sea levels rising, is becoming more and more serious [1], and the mitigation of and adaptation to global climate change should not be delayed. Forest ecosystems are an important terrestrial carbon sink, with a high carbon storage capacity, globally accumulating about 2.4 ± 0.4 Pg C per year [2], and they play a critical role in mitigating global climate change and balancing the ecosystem carbon cycle. The area of plantations in China is 7954.28 × 10⁴ ha, accounting for 36.45% of the national forest area, and the storage capacity of plantations is (33.88 million m³), accounting for 19.29% of the national forest storage capacity [3]. Plantations have a high potential for carbon sink, absorbing and fixing CO₂ through photosynthesis conversion to the form of biomass in plants, which helps to reduce greenhouse gas emissions by increasing carbon sinks and plays an important role in improving the ecological environment and mitigating climate change [4].

Biomass is the amount of organic matter actually living in a unit area at a given point in time [5]. Tree biomass is an important indicator for evaluating forest carbon stocks [6]. The accumulation of biomass not only reflects tree growth and adaptation to the environment, but also the productivity and carbon sequestration capacity of a forest ecosystem, and its role as a carbon sink [7,8]. Biomass is influenced by biotic and abiotic factors, such as plant genetic characteristics, interspecific competition, stand age, precipitation, soil conditions, altitude, and annual average temperature [9,10,11,12]. For example, Zhang et al. [13] found that altitude had an important effect on above–belowground biomass distribution, and the altitude gradient pattern of forest above–belowground biomass allocation distribution was determined by climate factors. Dai et al. [14] revealed that the efficiency of plant photosynthesis decreases because of a reduction in temperature, which would lead to a reduction in carbohydrate production and potentially less growth investment in aboveground parts, so as to influence biomass distribution. Accurate biomass estimation of trees is important for quantifying forest carbon storage and for understanding the carbon cycle of a forest ecosystem in the context of the global carbon cycle. Estimating the biomass of forests may rely on different techniques at different scales, from remote sensing at the regional level to direct weighing at the local level [15,16]. Because direct tree biomass measurements are destructive and costly, this method is not suitable for estimating large-scale forest biomass [15]. Allometric equation, a statistical model that predicts tree biomass (stem, bark, branch, leaf, and root) from easily measured variables, has been widely used in biomass estimation [17,18,19,20]. The diameter at breast height (D), tree height (H), cross-sectional area at breast height, crown morphology (crown width and shape), and wood density are common independent variables in allometric models [7,21,22,23]. In previous research, allometric models have been shown to provide accurate and reliable estimates of biomass in forest ecosystems [24,25,26,27]. For example, McNicol et al. [24] and Peng et al. [25] used a power function model with D²H as the predictive variable to estimate aboveground biomass. Yang et al. [26] found a good model fit using D and H as predictive variables (W = c D^a H^b) to estimate the total biomass of Quercus acutissima in middle-aged forests and using only D as a predictive variable (W = c D^a) to estimate the total biomass of Q. acutissima in young and mature forests. And Yang et al. [26] revealed the parameters of the allometric equation’s dependence on the stand age.

Stand age provides a clear indication of cumulative biomass. Many studies have emphasized the importance of stand age for the assessment of forest biomass and carbon sink potential [28,29,30,31]. Stand age also has a huge impact on tree size, shape, and traits, alters biomass allocation among tree components, and consequently affects allometric relationships [19,26]. For example, in early periods of growth, more biomass is allocated to stems for height growth and to branches and leaves for canopy expansion in order to increase competitive advantage for light resources over their neighbors [19]. Peichl et al. [32] and Chen et al. [33] reported differences in tree biomass between different stand ages. Some studies proved that the influence of stand age was an important determinant of forest biomass [16,31,34]. Thus, there is a need to develop age-specific allometric equations for rapid and reliable estimation of tree component biomass at different developmental stages. Although previous studies have explored stand age-based biomass modeling, stand age-driven biomass modeling also varies by forest type and tree species because of species-specific physiological growth traits [35]. In addition, different environmental conditions, e.g., biogeographic region, climatic conditions, soil type, and stand category, may lead to differences in stand age-specific biomass modeling [10,34,36]. Therefore, the development of stand age-dependent biomass models is important for practical applications to reveal biomass changes in fast-growing plantation forests (e.g., eucalypts).

Eucalypts is an important species for afforestation in southern China due to its fast growth, short rotation period, and excellent material. Eucalypts plantations in China exceed 5.6 million ha, accounting for 2.57% of the country’s forest area (218.22 million ha) and 7.04% of the country’s plantation forest (79.54 million ha) [3,37]. The area of eucalypts plantation in China is second only to India and Brazil [37]. Eucalypts has the function of carbon sequestration. Accurately quantifying the biomass of eucalypts plantations is critical for understanding carbon stocks in forest ecosystems and mitigating global climate change. A great deal of research has been conducted on allometric models of eucalypts biomass [38,39], but the information about allometric models considering the effects of stand age is limited. Therefore, the objectives of this study were to (1) analyze the effect of stand age on biomass distribution among tree components (i.e., stems, bark, branch, leaf, and root); (2) examine if the predictive capacity of the allometric equations that include stand age as a complementary variable is improved.

2. Material and Methods

2.1. Study Site

This study was carried out in the Leizhou Peninsula (109°57′~110°8′ E, 20°55′~21°8′ N) in Zhanjiang, Guangdong Province, China. The study site was within one of the major production areas of eucalypts. The study area had a typical oceanic monsoon climate, with an annual mean temperature of 23.1 °C. The annual relative humidity was 80.4%, the mean annual precipitation was 1711.6 mm, and there was an annual mean of 2003.6 h of sunshine. The elevation was approximately 50 m, and the soil type was Rhodi-Udic Ferralosols. The asexual lines E. urophylla × E. grandis 3229 and E. urophylla × E. tereticornis LH1 were used, and there were minimal differences in the terrain or management measures.

2.2. Sample Selection and Biomass Measurement

During 2023, five E. urophylla × E. grandis and E. urophylla × E. tereticornis stands (1, 3, 5, 7, and 8 years old) in the Leizhou Peninsula were selected. Three 20 m × 20 m sample plots were established randomly within each stand, each plot was separated by 500 m, and the diameter at breast height (D, 1.3 m) and tree height (H) was measured (

Table 1. Characteristics E. urophylla × E. grandis and E. urophylla × E. tereticornis stands. (mean ± standard deviations).

Species	Stand Age (Year)	Regeneration Method	Stand Density (Trees/ha)	Average Tree Height (m)	Average DBH (cm)
E. urophylla × E. grandis	1	seedling	1200	5.550 ± 0.235	4.300 ± 0.245
	3	seedling	1400	12.467 ± 0.480	8.550 ± 0.164
	5	seedling	1250	19.167 ± 0.258	13.278 ± 0.441
	7	seedling	1150	20.483 ± 0.479	13.85 0 ± 0.197
	8	seedling	1100	21.717 ± 0.471	15.033 ± 0.186
E. urophylla × E. tereticornis	1	seedling	1350	5.750 ± 0.418	6.017 ± 0.426
	3	seedling	1400	11.342 ± 0.111	8.117 ± 0.075
	5	seedling	1300	18.550 ± 0.565	11.797 ± 0.259
	7	seedling	1000	21.983 ± 0.911	15.000 ± 0.379
	8	seedling	1000	22.450 ± 0.616	15.875 ± 0.395

). According to the measurement data, two average trees (i.e., trees with the same diameter at breast height (D) and tree height as the average diameter and average tree height of the stand) in each sample plot were harvested. That is, six average trees were harvested per stand for destructive sampling. In total, 60 average trees were sampled. After felling, D and H were recorded. The trees were cut using a hand saw at 1.3 m, 2 m, 4 m, 6 m … 20 m, and 22 m of the tree height, and subsequently at 2 m intervals up to the apex. Bark was removed from the sampled stem. The fresh mass of the stem, bark, branches, and leaves was measured in the field. After the above measurements, the root systems of all sampled trees were excavated with a hook machine, soil clinging to the roots was removed, and the fresh mass of the roots was measured. Approximately 60 g subsamples of each component were selected and brought to the laboratory. The subsamples were oven-dried at 85 °C until they reached a constant weight. The dry mass (i.e., biomass) of all components was estimated according to the ratio of the fresh mass and dry mass of the subsamples.

2.3. Biomass Modeling

According to the methodology of McNicol et al. [24], allometric relationships between biomass components and D, H were often expressed using the following power function equation:

y_i = a₁ (x)^b1

(1)

where y_i is the biomass (kg) of the component (i.e., stem, bark, branch, leaf, root, aboveground) or the total biomass of a single tree; x is D, H, DH or D²H; and a₁ and b₁ are equation parameters.

All data were logarithmically (base 10) transformed to ensure a linear relationship between variables and to meet the requirements of homogeneity [19,24]. Therefore, we established the following equation [24]:

lg (y_i) = a₂ + b lg (x)

(2)

where y_i is the biomass (kg) of component (i.e., stem, bark, branch, leaf, root, aboveground) or the total biomass of a single tree; x is D, H, DH or D²H; and a₂ and b are equation parameters.

According to Xiang et al. [19], we introduced stand age as a complementary variable into the allometric equation to determine whether the equation was improved. The individual effect of stand age and the interactions between stand age and other predictor variables (i.e., D, H, DH, or D²H) were considered. Then, we reconstructed the stand age-specific allometric equations by introducing the stand age factor based on Equation (2):

lg (y_i) = a₂ + b lg(x) + c (age) + d (age) lg (x)

(3)

where y_i is the biomass (kg) of a tree component (i.e., stem, bark, branch, leaf, root, aboveground) or the total biomass of a single tree; x is D, H, DH or D²H; age is the stand age; and a₂, b, c, and d are equation parameters.

2.4. Data Analysis

The coefficient of determination (R²), total relative error (TRE), residual sum of squares (RSS), root mean squared error (RMSE), and Akaike information criterion (AIC) were used to assess the performance of all models:

$$TRE={\displaystyle \frac{{\displaystyle \sum \left(y_i-{\widehat{y}}_i\right)}}{{\displaystyle \sum {\widehat{y}}_i}}}\times 100$$

(4)

$$RSS={\displaystyle \sum _{i=1}^n{\left(y_i-{\widehat{y}}_i\right)}^2}$$

(5)

$$RMSE=\sqrt{{\displaystyle \frac{{\displaystyle \sum _1^n{\left(y_i-{\widehat{y}}_i\right)}^2}}{n}}}$$

(6)

$$AIC=2k+n\ln \left({\displaystyle \frac{RSS}{n}}\right)$$

(7)

where y_i is the observed biomass value, ${\widehat{y}}_i$ is the predicted biomass value obtained by the allometric models, n is the number of samples, and k is the number of parameters in the model.

3. Results

3.1. Biomass Allocation Among Tree Components

The stem of E. urophylla × E. grandis accounted for the largest proportion of biomass, (42.17%–74% of total tree biomass) and increased with stand age. Root biomass represented 14.18%–25.84% of total tree biomass and first decreased with increasing stand age and then increased with increasing stand age. The proportions of bark, branch, and leaf biomass decreased with stand age. Crown biomass (branch + leaf) was higher in 1 year and 3 years plantations and represented 22.89% and 17.45%, respectively, and crown biomass decreased with stand age. The partitioning of biomass components in E. urophylla × E. tereticornis was equivalent to that in E. urophylla × E. grandis. Stem biomass accounted for 55%–73.68% of total tree biomass in the five stands and increased with stand age. Root biomass accounted for 11.90%–22.28% of the total tree biomass and first decreased as the stand age increased, then increased with stand age increased. Stem and root biomass accounted for 71.93%–90.69%, while crown biomass accounted for 4.38%–16.75%, with a higher contribution occurring in 1 year and 3 years (Figure 1).

3.2. Allometry Among Tree Biomass Components

Biomass was scaled allometrically in eucalypts. The stem, bark, roots, aboveground biomass, and total biomass were closely related to diameter at breast height, with an R² > 0.917. Branch and crown biomass had R² values = 0.794 and 0.733, respectively, and leaf biomass had an R² = 0.390 (Figure 2). The stem, bark, aboveground biomass, and total biomass were closely related to tree height, with an R² > 0.968 (Figure 3). Branch, crown, and root biomass had R² values = 0.865, 0.833, and 0.703, respectively, and leaf biomass had an R² = 0.534 (Figure 3). There were close allometric relationships among the components’ biomasses. The R² between root and stem biomass was 0.786, the R² between crown and stem biomass was 0.675, and the R² between root and crown biomass was 0.675 (Figure 4). Our results demonstrate that an allometric equation was suitable for estimating the component biomass of eucalypts, and this relationship could be used to estimate biomass when diameter at breast height, tree height, or the biomass of one component is known.

3.3. Effect of Stand Age on Biomass Components

Stand age had a significant effect on the component biomass (including stem, bark, branch, leaf, root, aboveground, and crown) and total biomass of eucalypts (p < 0.05, Tables S1 and S2). However, tree variety had no significant effect on the component biomass or total biomass of eucalypts, except for stem and leaf (Tables S1 and S2). This study focused only on the effect of stand age on eucalypts biomass.

Introducing stand age, which was used to predict component (including stem, bark, branch, leaf, root, aboveground, and crown) biomass and total biomass, improved the fit of the allometric model. The R² of the stand age-specific allometric model was higher than that of a model that only considered D or H; in particular branch, leaf and root estimates were significantly improved. The R² values of stand age-specific allometric models based on D that were used to predict branch, leaf, and root biomass improved by 9.45%, 31.03%, and 4.69%, respectively (

Table 2. Allometric equations for tree biomass components and fitting statistics with stand age.

Independent Variable	Dependent Variable	a	b	c	d	R2	p	RSS	TRE	RMSE	AIC
D	Stem biomass	−1.490	2.595	0.216	−0.143	0.986	0.000	2409.382	0.917	6.337	229.567
H		−1.033	1.969	−0.066	0.081	0.988	0.000	2628.201	1.369	6.618	234.783
DH		−1.401	1.283	−0.075	0.035	0.995	0.000	1766.375	−0.303	5.426	210.940
D2H		−1.486	0.893	−0.014	0.008	0.994	0.000	1578.636	−0.889	5.129	204.198
D	Bark biomass	−2.146	2.503	0.412	−0.349	0.970	0.000	39.750	2.130	0.814	−16.705
H		−1.818	2.119	0.045	−0.054	0.974	0.000	20.893	1.089	0.590	−55.295
DH		−2.129	1.295	0.107	−0.059	0.983	0.000	26.621	−0.375	0.666	−40.760
D2H		−2.185	0.883	0.185	−0.060	0.981	0.000	29.322	3.100	0.699	−34.960
D	Branch biomass	−1.124	0.806	0.737	−0.523	0.869	0.000	83.809	2.813	1.182	28.052
H		−1.468	1.302	0.559	−0.373	0.897	0.000	65.885	2.383	1.048	13.614
DH		−1.385	0.595	0.641	−0.222	0.883	0.000	71.773	1.895	1.094	18.750
D2H		−1.315	0.362	0.680	−0.159	0.878	0.000	76.784	4.648	1.131	22.799
D	Leaf biomass	−0.941	0.025	0.845	−0.605	0.511	0.000	15.468	9.001	0.508	−73.333
H		−1.743	1.445	0.288	−0.227	0.549	0.000	12.080	7.468	0.449	−88.167
DH		−1.526	0.518	0.545	−0.201	0.502	0.000	13.348	9.924	0.472	−82.178
D2H		−1.360	0.245	0.666	−0.161	0.496	0.000	14.106	8.879	0.485	−78.864
D	Root biomass	−0.543	0.880	0.054	0.049	0.960	0.000	452.262	1.159	2.745	129.195
H		−0.412	0.728	−0.094	0.146	0.967	0.000	492.147	3.277	2.864	134.266
DH		−0.553	0.482	−0.102	0.076	0.967	0.000	434.119	1.824	2.690	126.738
D2H		−0.578	0.329	−0.067	0.043	0.965	0.000	434.716	2.447	2.692	126.821
D	Aboveground biomass	−1.045	2.189	0.310	−0.212	0.981	0.000	3423.626	1.035	7.554	250.647
H		−0.767	1.846	0.005	0.025	0.992	0.000	2313.416	−0.681	6.209	227.128
DH		−1.045	1.146	0.031	−0.004	0.994	0.000	1338.961	−1.481	4.724	194.318
D2H		−1.095	0.782	0.097	−0.019	0.991	0.000	1400.379	−1.073	4.831	197.009
D	Total biomass	−0.736	1.941	0.264	−0.165	0.981	0.000	5349.070	1.002	9.442	277.420
H		−0.491	1.648	−0.021	0.052	0.993	0.000	3592.137	−0.432	7.738	253.529
DH		−0.743	1.026	0.002	0.012	0.993	0.000	2171.750	−0.885	6.016	223.337
D2H		−0.787	0.699	0.065	−0.007	0.990	0.000	2768.508	−2.397	6.793	237.903

Notes: D is the diameter at breast height, H is the tree height, DH is the product of diameter at breast height and tree height, D²H is the product of square of the diameter at breast height and tree height, R² is the coefficient of determination, RSS is the residual sum of squares, TRE is total relative error, RMSE is the root mean squared error, AIC is Akaike information criterion, and a, b, c, and d are model parameters.

and Figure 2). The R² values of stand age-specific allometric models based on H that were used to predict branch, leaf, and root biomass improved by 3.70%, 2.81%, and 37.56%, respectively (Table 2 and Figure 3). The RSS, RMSE, and AIC of stand age-specific allometric models based on H that predicted bark, branch, and leaf biomass were lower than those of other models. The RSS, TRE, RMSE, and AIC of the stand age-specific allometric model based on DH that predicted root and aboveground biomass were lower than those of other models. The RSS, RMSE, and AIC of the stand age-specific allometric model based on DH that predicted stem and total biomass were lower than those based on D or H (Table 2). The RSS, TRE, RMSE, and AIC of the stand age-specific allometric model based on D²H that predicted stem biomass were lower than those of other models, while the RSS, RMSE, and AIC of the models that predicted root, aboveground, and total biomass were lower than those based on D or H.

The ratios among biomass components were significantly different across stand age (

Table 3. Component biomass ratios in different stand age classes and species classes. (mean ± standard deviations).

Factors		Roots/Stem Ratio	Crown/Stem Ratio	Roots/Crown Ratio	Belowground/Aboveground Ratio
Stand age classes
1 year		0.511 ± 0.138 a	0.372 ± 0.218 a	1.688 ± 0.643 b	0.319 ± 0.055 a
3 years		0.301 ± 0.066 b	0.311 ± 0.073 a	0.987 ± 0.219 c	0.199 ± 0.040 c
5 years		0.185 ± 0.029 c	0.098 ± 0.022 b	1.963 ± 0.469 b	0.150 ± 0.023 d
7 years		0.218 ± 0.039 c	0.069 ± 0.013 b	3.281 ± 0.885 a	0.191 ± 0.036 c
8 years		0.310 ± 0.055 b	0.099 ± 0.018 b	3.161 ± 0.389 a	0.262 ± 0.043 b
Species classes
EG	1 year	0.620 ± 0.108 a	0.555 ± 0.150 a	1.175 ± 0.304 b	0.351 ± 0.058 a
ET		0.403 ± 0.042 b	0.189 ± 0.036 b	2.200 ± 0.434 a	0.287 ± 0.031 b
EG	3 years	0.293 ± 0.08	0.318 ± 0.104	0.952 ± 0.253	0.194 ± 0.047
ET		0.309 ± 0.055	0.305 ± 0.028	1.023 ± 0.196	0.205 ± 0.035
EG	5 years	0.201 ± 0.03 a	0.093 ± 0.025	2.233 ± 0.480 a	0.166 ± 0.022 a
ET		0.168 ± 0.016 b	0.102 ± 0.019	1.694 ± 0.282 b	0.135 ± 0.012 b
EG	7 years	0.203 ± 0.032	0.078 ± 0.010 a	2.643 ± 0.48 b	0.176 ± 0.028
ET		0.232 ± 0.043	0.06 ± 0.008 b	3.919 ± 0.719 a	0.206 ± 0.039
EG	8 years	0.341 ± 0.038 a	0.106 ± 0.018	3.261 ± 0.379	0.285 ± 0.028
ET		0.278 ± 0.054 b	0.091 ± 0.016	3.060 ± 0.407	0.239 ± 0.045

Note: different lowercase letters indicate significant differences (p < 0.05). ET is E. urophylla × E. tereticornis, EG is E. urophylla × E. grandis.

). The ratios of roots to stem biomass were 0.511, 0.301, 0.185, 0.218, and 0.310 across the stand age, and these ratios first decreased and then increased with stand age. The ratios of crown to stem biomass were 0.372, 0.311, 0.098, 0.069, and 0.099, and first decreased and then stabilized with stand age. The ratios of root to crown biomass were 1.688, 0.987, 1.963, 3.281, and 3.161, and first decreased and then increased with stand age. The ratio of belowground to aboveground biomass first decreased and then increased with stand age (Figure S1). Overall, the ratios between biomass components varied greatly in young stands (1 year and 3 years) and differed significantly from those in nearly mature stands (7 years and 8 years, Table 3). Therefore, large errors in the estimation of component biomass may occur if an allometric model does not consider eucalypts stand age.

3.4. Biomass Prediction Comparisons Among Tree Components

Except for branches and leaves, the biomass values predicted by the stand age-specific allometric models for stem, bark, root, aboveground, and total tree biomass were comparable with the observed biomass values (Figure 5). These results indicate that stand the age-specific allometric models based on D, H, DH, and D²H were able to efficiently estimate stem, bark, root, aboveground, and total tree biomass with a high degree of precision. The R² values for observed biomass (bark, branch, and leaf) versus predicted biomass (using H) from the stand age-specific allometric model were 0.945, 0.753, and 0.375, respectively, which were better than those from models based on D, DH, or D²H (Figure 5). For bark, branches and leaves, the H-based stand age-specific allometric model was more precise, and the RSS, RMSE, and AIC values using H as a predictor variable were smaller than those from other models (Table 2). The R² values for observed stem biomass values predicted biomass (using D²H) from the stand age-specific allometric model to be 0.979, and that value was higher than those from the other allometric models (Figure 5). For stem, the stand age-specific allometric model using D²H as a predictor variable was better than models based on D, H, or DH, and the TRE, RSS, RMSE, and AIC of the based D²H stand age-specific allometric model were smaller than those from other allometric models (Table 2). For roots, aboveground biomass, and total tree biomass, the values predicted by the DH- and D²H-based stand age-specific allometric models were close to the observed values, with R² values of 0.920 and 0.921 (root biomass), 0.986 and 0.985 (aboveground biomass), and 0.984 and 0.982 (total biomass), which were higher than those from the models using D or H as predictor variables (Figure 5). These results were consistent with the RSS, RMSE, and AIC values from the DH- and D²H-based stand age-specific allometric models, which were smaller than those from other models (Table 2).

4. Discussion

4.1. Biomass Allocation Characteristics

Biomass production is a primary function of forest ecosystems and is influenced by numerous processes: roots capture nutrients from the soil; stems and branches provide mechanical support and conduct water and nutrients; and leaves fix carbon [40]. Balancing the allocation of biomass to different organs enables trees to perform necessary physiological activities and achieve normal growth, thus reflecting their resource use efficiency and environmental adaptability [41,42]. The effect of stand age on biomass allocation is a result of tree growth and the balance of resource allocation among biomass components in a way that maintains the physiological activities and functions of each organ [22]. Our results show that except for 1 year E. urophylla × E. grandis (42.17%) and E. urophylla × E. tereticornis (55.43%) and 3 years E. urophylla × E. grandis (56.24%) and E. urophylla × E. tereticornis (55%), the proportions of stem biomass to total tree biomass in E. urophylla × E. grandis (65.18%–74%) and E. urophylla × E. tereticornis (69.62%–73.68%) (Figure 1) were higher than the average stem biomass ratio for forest ecosystems in China (58.20%) [43]. The proportions of stem biomass to total tree biomass in E. urophylla × E. grandis and E. urophylla × E. tereticornis from all stand ages were higher than the biomass of 1 year–7 years E. urophylla × E. grandis (24.91%–66.78%) measured by Zhang et al. [44] in Guangdong, but lower than the biomass of 1 year–8 years E. urophylla × E. grandis (51.07%–98.48%) measured by Fu et al. [45] in Guangxi. In this study, the proportion of leaf biomass to total tree biomass of eucalypts (0.91%–10.01%) at all stand ages was smaller than that measured by Zheng et al. [39] (2.37%–23.63%) and Fu et al. [45] (2.17%–21.01%).

Stand age significantly affected biomass allocation in each organ of eucalypts. We observed an increase in the proportion of stem biomass to total tree biomass and a decrease in the proportion of bark, branch, and leaf biomass to total tree biomass as stand age increased (Figure 1), which was consistent with other studies [19,26]. Changes in biomass allocation with age may be explained by the strategies that trees employ during stand development [19]. Stem biomass is an important component of tree biomass, which represents a net accumulation process with stand age [26]. In order to occupy the upper canopy space and obtain more lights, the branches and leaves of young eucalypts would undergo rapid expansion, which is conducive to photosynthesis and transpiration and crucial for the survival of young eucalypts [32,46]. The proportions of branch and leaf biomass to total tree biomass in eucalypts were higher in young stands, demonstrating that more carbohydrates may be allocated to branches and leaves to help eucalypts grow rapidly [22]. Trees will adjust their biomass allocation strategy according to competitive demands at different stages [26]. Therefore, changes in biomass distribution during stand development must be considered to obtain accurate biomass predictions.

Roots play an important role in material cycling and energy flow in forest ecosystems [47]. Roots not only absorb and transport water and nutrients but also input organic matter into the soil, which maintains soil fertility, and fix carbon [23,47]. In this study, the proportions of root biomass to total tree biomass in eucalypts of all stand ages were 11.90%–25.84% (Figure 1), which indicates the importance of root biomass. A well-developed root system improves tree competition for water and nutrients, which is conducive to growth [26]. We also found that the proportion of root biomass to total tree biomass first decreased and then increased as stand age increased (Figure 1), which is not consistent with previous studies [45,47]. The proportion of root biomass to total tree biomass decreased as stand age increased in Fu et al. [45], while an increase with stand age was reported by Han et al. [47], and an initial increase and then a decrease with stand age was found by Zhang et al. [44]. However, Chen et al. [33] showed that the proportion of root biomass to total tree biomass gradually stabilized as stand age increased. These differences may result from different tree species and study sites. Root biomass is also related to soil moisture, nutrients, stand structure, and community composition [47]. In addition, the root biomass may be estimated based on the relationship between root biomass and the biomass of other components [48]. In this study, root biomass was closely related to stem biomass and crown biomass (R² = 0.786 and 0.675, respectively) (Figure 4), which was similar to the results of other studies [49,50]. The ratio of belowground to aboveground biomass of eucalypts ranged from 0.150 to 0.319 (Table S2), which was close to that of other forest types in China (0.233) [51]. The ratio of aboveground biomass to root biomass helps us estimate root biomass from the more easily measured aboveground biomass and also reflects investment in aboveground organs versus roots in response to environmental change [23,52].

4.2. Analysis of Biomass Allometric Equations

Stand age had a significant effect on component biomass in eucalypts. Other studies have also reported an effect of stand age on the relationship between biomass and measurement factors [19,26]. Hong et al. [53] showed that the introduction of a stand age factor improved the fit and the prediction accuracy of their biomass model. Similarly, Wang et al. [50] reported that there were differences in aboveground and belowground biomass allocation by Abies nephrolepis at different growth stages, and a biomass estimation without considering stand age could be slightly biased. Our results show that stand age-specific allometric equations for component biomass in eucalypts improved the estimates of biomass, especially in branch and leaf biomass equations (Figure 2 and Figure 3, Table 2), which was similar to the results of previous studies [32,54]. These are the results of changes in tree growth and a combination of stand age and tree social status [19]. Therefore, we should consider the effect of stand age on biomass to improve the accuracy of biomass estimation. Similar findings have been applied at other species, such as Chinese fir (Cunninghamia lanceolata), Mongolian pine (Pinus sylvestris L. var. mongolica Litv.), and White pine (Pinus strobus) [4,19,32].

Using D, H, or combinations of D and H as predictor variables in allometric equations is a simple and effective method for estimating tree biomass. In this study, we used E. urophylla × E. grandis and E. urophylla × E. tereticornis from different-aged stands and constructed and screened an optimal biomass model, which was reliable for estimating the biomass of eucalypts in the Leizhou Peninsula of Guangdong Province (Table 2). Stand age-specific allometric models for stem, bark, root, aboveground, and total biomass per tree based on D, H, DH, or D²H explained more than 96.5% of variation in the data (Table 2). Stand age-specific allometric models for branch biomass based on D, H, DH, or D²H explained more than 86.9% of variation in the data but stand age-specific allometric models for leaf biomass explained only 49.6%–54.9% of variation in the data (Table 2). Except for leaf biomass, the stand age-specific allometric models for other components (stem, bark, branch, root, aboveground, and total biomass) based on D, H, DH, or D²H had high explanatory power and provided good estimates of biomass (Table 2). Similar trends in leaf biomass have been reported in other studies [20,55] and may be due to the felling process, which causes a loss of leaves that results in a large error. Leaf biomass is strongly influenced by canopy traits and tree growth because of canopy competition [56]. In addition, leaf biomass is influenced by light, moisture, temperature, soil nutrients, and interspecific competition [20]. The fit of the leaf biomass model was lower than that of other biomass components (Table 2). However, in general, stand age-specific allometric models fit better (Figure 2 and Figure 3; Table 2).

The D and H, which are accessible and accurate measurements, are widely used in allometric models [21]. Our study confirmed a strong correlation between component biomass and D in eucalypts, and a stronger correlation between branch and leaf biomass and H (Table 2). It is possible that branches and leaves were more responsive to competition and the environment [57], and that competition for canopy dominance may have resulted in a greater allocation of photosynthate to height than to radial growth [46]. H was added into the biomass allometric model as the second parameter to improve its accuracy [24,25,58]. Our results showed that stem biomass predicted by the stand age-specific allometric model based on D²H fitted well and had high explanatory power. The R² value for this model was higher than that of other models, and the RSS, TRE, RMSE, and AIC were lower than those of other models (Table 2), demonstrating that D²H was strongly intrinsically linked to stem biomass. The stem biomass predicted with D²H was comparable to the observed data (Figure 5). The DH and D²H were reliable predictors of root, aboveground, and total tree biomass, and those models had higher R² values and lower RSS, TRE, RMSE, and AIC values (Table 2), indicating that DH and D²H were both strongly intrinsically related to root, aboveground, and total tree biomass. Root, aboveground, and total tree biomass predicted with DH or D²H were comparable to the observed data (Figure 5), indicating that DH and D²H improved the accuracy of the models. This may be because DH and D²H captured the variation in tree structure and avoided the problem of multicollinearity between D and H as separate variables in an allometric model [36,59]. These models are suitable for predicting the component biomass of eucalypts in Leizhou Peninsula, Guangdong province, China. However, when using these models to predict the biomass of eucalypts plantations in other place of China, they should preferably be calibrated to improve the precision and accuracy of biomass estimation because biomass is affected by light, moisture, temperature, soil nutrients, altitude, etc. [13,20,57], which affect the relationship between component biomass and measuring factor (i.e., D and H).

5. Conclusions

There was no significant effect of different eucalypts varieties on biomass except for stem and leaf biomass (Figure 1). Stand age significantly affected the allocation of biomass in eucalypts. With increasing stand age, the proportion of stem biomass to total tree biomass increased, the proportion of bark, branch, and leaf biomass to total tree biomass decreased, and the proportion of root biomass to total tree biomass first decreased and then increased (Figure 1).

This study confirms that organ biomass, aboveground biomass, total biomass, and D, H, DH, and D²H were allometrically consistent and can be used to accurately estimate the biomass of eucalypts plantation (Figure 2 and Figure 3, Table 2).

Stand age-specific allometric equations that include stand age as a complementary predictor variable, significantly improved the prediction effect and the accuracy of biomass assessment (Table 2). The optimal independent variables in the biomass prediction models differed according to each organ. Bark, branch, and leaf biomass were best predicted with H as the independent variable, stem biomass was best predicted with D²H as the independent variable, and root, aboveground, and total biomass were best predicted with DH or D²H (Table 2 and Figure 5). Therefore, stand age-specific allometric equations are recommended to improve the estimates of component biomass in eucalypts plantations.