Above- and below-ground biomass relationships of Leucaena leucocephala (Lam.) de Wit in different plant stands

The above and below–ground biomass (AGB and BGB) relationship is often used to assess the impact of biotic and abiotic effects on the growth and development of individual plants. The AGB and BGB relationship of the same tree species in different habitats can change significantly because of environmental stress. To investigate how the tree size, the biomass allocation and BGB/AGB ratio of Leucaena leucocephala (Lam.) de Wit varied according to spacing and mixed plant patterns in a valley–type savanna of southwest China, we examined the growth of L. leucocephala, and sampled 23 individuals for biomass measurement in each of four treatments (close/wide spacing of Leucaena leucocephala monocultures, mixed plantation of Leucaena leucocephala and Eucalyptus camaldulensis, and mixed plantation of Leucaena leucocephala and Eucalyptus citriodora), and then determined the regression relationships between AGB and BGB of L. leucocephala in different plant stands. Our results indicated that mixed planting significantly reduced all growth metrics for the tree sizes of L. leucocephala and increased the value of BGB/AGB. Changing plant spacing in monocultures had a significant impact on AGB and TB (Total Biomass) of L. leucocephala, but it had no significant effect on the other metrics. Within mixed plant schemes, L. leucocephala significantly reduced the biomass allocation to leaves and small roots and increased the allocation to coarse root biomass. There were no significant differences in tree size and biomass allocation of L. leucocephala between different spacing regimes in monocultures or between different mixtures in mixed plant stands. The correlation between BGB and AGB of L. leucocephala in all plant stands was consistent with the model of allometric growth, and AGB can be used to accurately estimate BGB. Interestingly, the correlations were not exactly the same. BGB and AGB in monoculture showed isometric growth, and their values in mixed plant stands showed allometric growth. BGB also increased faster than AGB. The findings indicated that L. leucocephala allocated more biomass to the root system when it was planted with Eucalyptus.


Site description and trial establishment
The study was carried out in the artificial vegetation restoration area of Desert Ecosystem Station in Yuanmou County of Yunnan Province, State Forestry Administration of China (25˚40' N, 101˚52' E), and the study area was approximately 30 km south of the Jinsha River. The site was situated in a relatively low elevation valley-type savanna (approximately 1, 200 m a.s.l.), where it was extremely hot and dry due to the influence of the monsoon, mountainous terrain, and canyons. The year is divided into a wet and dry season. Average annual rainfall  is 584 mm. The largest rainfall year (891 mm) occurred in 1998, and the driest year (471 mm) occurred in 2010. Rainfall distribution tends to be wet season dominant (June-October) with only 8% of the annual rainfall falling in the dry season (November-May of the following year) when evaporative demand is highest. The average annual potential evapotranspiration (1991-2015) is 3, 848 mm. Thus, the trees grow in an environment where the atmospheric water demand is on average 6.6 times the rainfall. The long-term average annual temperature is 22˚C. The lowest average monthly temperature was 15˚C (December), with an extremely low temperature of -1˚C. The highest average monthly temperature was 27˚C (May), with an extremely high temperature of 42˚C. The annual accumulated temperature (�10˚C) was 8, 003˚C (The meteorological data in this section were from the weather station of Yuanmou County). The main type of soils is classified as Ferralic Arenosols according to the FAO Taxonomy [29][30]. There is shallow topsoil, and the subsoil layers are deep, compact, and have a high level of gravel (mass fraction > 35%) with poor water retention. Thus, the trees grow in an environment where there is a high degree of desertification and severe soil erosion, with ravines on the soil surface.
The experimental forests were planted in 1998. In this paper, the experimental design consisted of four treatments (two spacing and two mixed plant treatments). The spacing treatments included 100% L. leucocephala at close spacing (1.5 m×3.0 m) and 100% L. leucocephala at wide spacing (3.0 m×3.0 m). Mixed plant treatments included 50% L. leucocephala + 50% E. camaldulensis at 3.0 m×3.0 m spacing which we will refer to as "M. W. camaldulensis" and 50% L. leucocephala + 50% E. citriodora at 3.0 m×3.0 m spacing which we will refer to as "M. W. citriodora" The average diameters at breast height (DBH) of E. camaldulensis and E. citriodora were 19.8±1.9 cm (Mean±SE) and 17.4±0.9 cm, the average heights were 11.2±1.0 m and 13.6±0.9 m, and the average crown diameters were 5.1±0.2 m and 4.7±0.1 m, respectively. We also used the parameters of average DBH, average height, and average crown diameter to describe the effect of the plant stand on growth of L. leucocephala.

Biomass sampling
Sampling for biomass of L. leucocephala was carried out in the winter of 2015 when the trees were 17 years old. As a preliminary step, we randomly selected 100 trees in each of the four treatments and collected the data of each tree including the diameter at breast height (DBH), the height, and the crown diameter. The crown diameter was calculated according to the crown's orthogonal projection on the ground. All of these data were used to verify the growth characteristics and the DBH class distribution of the plants in the four treatments. According to the DBH class distribution of the trees, 23 trees with different DBH in each of the four treatments were selected to sample biomass. For example, the DBH range of L. leucocephala in close spacing treatment ranged from 6.9 to 15.7 cm, and the DBH could be divided into 5 class levels. These classes included 6 cm<DBH�8 cm, 8 cm<DBH�10 cm, 10 cm<DBH�12 cm, 12 cm<DBH�14 cm, and 14 cm<DBH�16 cm, and 4-5 trees were selected from each DBH class level.

Above-ground biomass sampling
Each sample tree of L. leucocephala was cut at ground level (leaving 10 cm of stump) and the crown (including branches and foliage) was removed from the stem. The aboveground biomass (AGB) was separated into stem biomass, branch biomass, and foliage biomass. The stem at the top was cut at 3 cm diameter and the part smaller than 3 cm was included in the branches portion. Foliage was removed from the branches by hand, and both were weighed fresh in the field using a field dynamometer and the values were recorded. Subsamples of branches and foliage were labeled, bagged, and then transported to the laboratory of Desert Ecosystem Research Station in Yuanmou County of Yunnan Province, State Forestry Administration. The subsamples were then dried at 85˚C to a constant weight for determination of water content. The dry weight for branches and foliage was then calculated. The stem was divided into the base (ground to 1.3 m), middle (1.3 m to halfway from top), and top. Diameters, lengths, and fresh weight were recorded for each stem section. Disks from the stem were taken at ground level, breast height, and top. Fresh weight was determined for each component and all samples were oven dried to a constant weight at 85˚C. Weights were recorded to calculate dry weights of tree parts.

Below-ground biomass sampling
Belowground biomass of L. leucocephala was determined by the stratified excavation method. For each tree selected for sampling, the scope of the root system excavation was set as a circle with a radius of 1.0 m and the base of the trunk as the center. The sample tree roots were harvested to a depth of 100 cm by manual excavation to collect total belowground biomass, and the root samples were weighed for each 20 cm of depth. Thus, the vertical distribution of below-ground biomass was divided into 5 layers, i. e., 0-20, 20-40, 40-60, 60-80, 80-100 (unit: cm). Due to the time involved in sampling and the much smaller number of roots encountered, excavation of the 50-100 cm depth was reduced to half of the area used for the 0-50 cm excavation. All roots from these areas regardless of their excavated source were included in the measurement of root biomass. Thus, the part of the root system of the sample tree that grew outside of the sampled area was not sampled, but any roots that grew into the soil from other trees were included. In any event, it was impossible to excavate and recover the entire root system for each sample tree because of root intermingling, the enormous effort required for excavation, root separation, and the destruction of plots that occurred during the excavation. It should be pointed out that the belowground biomass might not be sampled fully (in a 1 m radius around trees); and the BGB as well as BGB/AGB were likely underestimated.
The belowground biomass (BGB) was separated into the vertical taproot (the primary root) and the lateral root. The vertical taproot was excavated to a diameter (or diameters in cases where it split) of 2 cm, and the part smaller than 2 cm was included as the lateral root. According to the size of root diameter (RD), the lateral root was separated into small root (RD � 1.0 cm), medium root (1.0 cm < RD � 2.0 cm), and coarse root (RD > 2.0 cm). All root systems were extracted from the soil and separated from the soil particles and sediments by gentle shaking and removal by hand. The fresh weight was measured in the field. The root samples were oven dried to constant weight at 85˚C and then used to calculate total oven-dry weights of root system parts.

Data processing and statistical analyses
We analyzed the spatial distribution of vertical belowground biomass according to the root system distribution model (RSDM) proposed by Gale [31]: where, Y is the ratio of the sum of biomass from the topsoil (0 cm) to the depth d (cm) of the overall belowground biomass and β is the extinction coefficient. A larger β value indicates a greater proportion of deep root biomass; while a smaller β-value indicates that a greater proportion of the root system is present in the surface soil or in the lower level of the crown.
The regression between above-and below-ground biomass was developed using the equation: y = b . x a . Taking the logarithm (base 10) of each side of the equation, we performed a linear conversion and obtained the following equation: log (y) = log (b) + a log (x), where, x and y represent the below-and aboveground biomass, respectively. Log (b) represents the intercept of the regression equation between above-and belowground biomass, and a represents the slope of the regression equation, which is the relative growth index. The relative growth index suggests isometric growth if there is no significant difference between 'a' and '1.0,' and it suggests allometric growth when 'a' is significantly larger or smaller than '1.0' [32]. Standardized major axis estimation (SMA) was used to estimate the slope parameters and the intercept [33] using the SMATR Version 2.0 software (http://www.bio.mq.edu.au/ecology/SMATR). The Sigmaplot 11.0 software was used for plotting graphs. Differences in tree sizes (n = 100), the biomass, and BGB/AGB ratio (n = 23) of L. leucocephala between treatments were tested using an analysis of variance (ANOVA) and Tukey's test. Rejection level was set when α < 0.05 in all analyses. Analyses were conducted using SPSS 16.0 (IBM; Armonk, NY).

Tree sizes
There was a significant difference in L. leucocephala growth among different plant stands (Table 1). L. leucocephala trees in monoculture (close spacing and wide spacing) were significantly taller and had significantly larger DBH and crown diameter than plants in mixed stands (M. W. camaldulensis, M. W. citriodora) (p<0.05). However, changing the plant spacing and mixture of tree species had no significant effect on growth.

The biomass distribution and BGB/AGB ratio
Higher AGB, BGB, and TB as well as lower BGB/AGB ratios of L. leucocephala were observed in monoculture as compared to mixed plant stands (Table 2). AGB, BGB, TB, and BGB/AGB ratios showed a significant response to plant stands (monoculture or mixed plant stands) (p<0.05). Increasing the plant spacing significantly increased AGB and TB of L. leucocephala (p<0.05), and it had no significant effect on the biomass of BGB and BGB/AGB (p>0.05). Changing the mixture of tree species had no significant effect on the biomass distribution or BGB/AGB ratio of L. leucocephala (p>0.05).

Distribution characteristics of aboveground biomass
The AGB of L. leucocephala was mainly concentrated in the stem, with a lowest overall proportion of AGB of 55.9% (in Wide spacing of the monoculture) and a highest proportion of 68.1% (in M. W. citriodora of the mixed plant stands), and no significant difference was found in stem biomass between different plant stands (p>0.05) (Fig 1). Higher branch and foliage biomasses were observed in monoculture than in mixed plant stands. Changing the plant spacing and mixture of tree species had no significant effect on the distribution characteristics of AGB.

Distribution characteristics of belowground biomass
The BGB of L. leucocephala was mainly concentrated in the primary root, with the lowest overall proportion of BGB at 60.4% (in Wide spacing of the monoculture) and the highest proportion at 63.2% (in M. W. citriodora of the mixed plant stands) (Fig 2). No significant difference was found in the primary root biomass between different plant stands (p>0.05). Significantly higher coarse root biomass and lower small root biomass were observed in mixed plant stands when compared to monoculture (p<0.05). Changing the plant spacing and mixture of tree species had no significant effect on the distribution characteristics of BGB. The BGB of L. leucocephala decreased gradually with depth, and it was mostly concentrated in the soil surface between 0-40 cm (Fig 3). Statistical comparison of means showed that BGB within the top 0-40 cm of the surface soil in monoculture (74.30% in Close spacing; 80.79% in Wide spacing) was significantly higher than that in the mixed plant stands (55.66% in M. W. camaldulensis; 60.59% in M. W. citriodora) (p<0.05).

The relationship between above-and belowground biomass
LogBGB of L. leucocephala in different plant stands was closely correlated with LogAGB, and a strong linear relationship was found between them (Fig 4).  Table 3). Among these, the fitting slopes of L. leucocephala grown in monoculture (close spacing and wide spacing) showed no significant

Discussion
In the present study, almost all metrics of L. leucocephala tree size were greater in monoculture (Tables 1 and 2). All treatments showed a significant response to planting stand (monoculture or mixed plant stands) (p<0.05), reflecting increased competition for L. leucocephala in the mixed plant stands. However, changing the plant spacing and mixture of tree species had no significant effect on most growth metrics and BGB/AGB ratio. This indicated that mixed plantations with Eucalyptus significantly influenced tree sizes and biomass distribution of L. leucocephala,but spacing did not affect these characteristics.
The results also suggested that interspecific competition was higher than intraspecific competition for L. leucocephala. Similarly, mixtures that contain nitrogen-fixing species, such as 'E. globules + Acacia mearnsii,' 'E. saligna + Facaltaria moluccana,' and 'Eucalyptus grandis + Acacia mangium,' were found to have significant effects on plantation yields [5,[21][22]. However, these studies showed that the tree sizes of nitrogen-fixing trees increased in the mixed plant stands, which contrasted with the results of the present study. The differences in results can be explained by the different age of plant stands and the different growth environment. In the studies by Laclau et al. [5] and Bauhus et al. [21], investigations were respectively carried out 30 months and 6.5 years after planting, when there was less intense competition in the plantation. With continued tree growth, competition within the forest would become increasingly intense, and the situation might change significantly. On the other hand, Binkley et al. [22] established their experiments at 480 m elevation on the northeast coast of Hawaii, where the average annual rainfall was approximately 4000 mm. Under such conditions, nitrogen fixing trees might have adapted to use the higher water supply and soil fertility via developing stronger stems and root systems, and the complementarities between nitrogen fixing trees and non-nitrogen fixing trees were therefore greater than their competition for resources in the mixed plantations.
Furthermore, the non-nitrogen fixing trees of Eucalyptus in the mixed plantations was a fast growing plant with growth rates that routinely exceeded 35 m 3 ha -1 year -1 [34][35]. It could also be seen from our survey data that the Eucalyptus had larger DBH values (2.0-2.2 times), heights (1.3-1.7 times), and crown diameters than L. leucocephala in the mixed plantations (see the Materials and methods section and Table 1). However, there is an evident negative impact of Eucalyptus on soil water availability for other species [36][37][38]. When Eucalyptus is grown, the land of mixed plantations would become more water-impoverished. As a result, the competition for water between Eucalyptus and L. leucocephala is more intense, and the growth of L. leucocephala can be restrained.
The root system is the most important organ of plant nutrition. Reduction in tree sizes of L. leucocephala in the mixed planting stands might be determined by the structure and interactions of the root systems. In the present study, results suggested that the top 0-40 cm of the surface soil was very important for L. leucocephala for absorbing nutrients and water as well as the exchange of materials with the external environment. Furthermore, BGB within the top camaldulensis represent 50% L. leucocephala + 50% E. camaldulensis at 3.0 m×3.0 m spacing; M. W. citriodora represent 50% L. leucocephala + 50% E. citriodora at 3.0 m×3.0 m spacing. Different letters between bars indicate significant differences (p<0.05). https://doi.org/10.1371/journal.pone.0207059.g001 Above-and below-ground biomass relationships of Leucaena leucocephala 0-40 cm of the surface soil in monoculture was significantly higher than that in the mixed plant stands (p<0.05) (Fig 3). The results of RSDM suggested that mixtures had lower β-values than monoculture. This showed that L. leucocephala in the mixed plant stands had deeper root systems than in monoculture. Simultaneously, when planted with Eucalyptus, L. leucocephala significantly reduced the biomass allocation to leaves and small roots as well as increased coarse root biomass (p<0.05) (Figs 2 and 3). This indicated that more competition was found in the roots of L. leucocephala in the mixed plant stands and the coarse root competition was Above-and below-ground biomass relationships of Leucaena leucocephala particularly strong. It suggested that the water absorption of L. leucocephala in mixed plant stands was more restricted than in monoculture. This might be related to the suppression of L. leucocephala growth by tall Eucalyptus and the higher water consumption characteristics of Eucalyptus [37][38]. Tang et al. [39] indicated that there were no significant interactions between the roots of L. leucocephala and Eucalyptus 10 years after establishment of the mixed plant stands in the valley-type savanna. This was due to the fact that rocky soil inhibited root growth. However, the situation might be changed after 17 years of trees growth. This was similar to some studies in other regions [40][41]. As plants grow, fierce competition between the  Isometric allocation hypotheses and allometric allocation hypotheses for the relationship between AGB and BGB of the individual plants have been widely evaluated based on field observations. Many studies on the biomass allocation of trees grown in good conditions have supported the existence of isometric growth between AGB and BGB [3,14,42]. However, in harsh ecosystems, the allometric allocation hypothesis between AGB and BGB has often been confirmed. Plants would allocate more biomass to competing organs, thereby increasing the absorption and utilization of limited resources [10,[43][44]. In the present study, the relationship between AGB and BGB of L. leucocephala in monoculture with different spacing exhibited isometric growth, which supported the isometric allocation hypothesis. The mixed plant stands showed allometric growth, which supported the allometric allocation hypothesis. Furthermore, BGB increased faster than AGB (Table 3). This showed that the competition for available water content in the mixed plant stands was more intense. This was consistent with the results of the previous analysis on tree size of L. leucocephala in different planting lands. With increasing of water stress, plants would allocate more biomass to the root system and less to the AGB [8][9][10]. Therefore, biomass allocation in monoculture and mixed plant stands might support the different allocation hypothesis. Although the allocation relationship between BGB and AGB is not equivalent to the BGB/AGB ratio, they become comparable in different plant stands. In the present study, values of BGB/AGB ratio in mixed plant stands were significantly increased (Table 2). Generally, BGB to AGB ratios have a tendency to increase in drier, harsher conditions [8,10]. These results suggested that there was a significant change in the proportion of AGB and BGB allocation for L. leucocephala when planted with Eucalyptus, and more biomass was allocated to the root system. In addition, these values were higher than the IPCC [45] default value of 0.25 for hardwood species and thus might significantly increase the estimate of BGB of L. leucocephala in the dry and hot environment, especially for mixed plantations.

Conclusion
Our results suggested that mixed plantations with Eucalyptus significantly decreased almost all growth metrics of L. leucocephala, including DBH, height, average crown diameter, AGB, BGB and TB, while it significantly increased BGB/AGB ratios. Increasing the plant spacing only significantly increased AGB and TB, and had no significant effect on the other metrics. Compared with in monoculture, L. leucocephala in mixed plantations with Eucalyptus significantly reduced the biomass allocation to leaves and small roots, and the plants increased the allocation to coarse root biomass. This study also showed that the correlation between AGB and BGB of L. leucocephala in all plant stands was consistent with the model of allometric growth. However, the correlation was not exactly the same. It showed isometric growth in monoculture, and allometric growth in mixed plant stands.