Species and Stand Management Options for Wood Production from Small Grower Plantations in Central Vietnam

Abstract

Acacia hybrid (Acacia mangium Willd. × A. auriculiformis A. Cunn. Ex Benth.) dominates plantation wood production in central Vietnam. Dependence on a single species may increase biological risks. The potential of eucalypt as an alternative was examined by comparing the growth and survival of acacia hybrid and eucalypt hybrid (Eucalyptus urophylla S.T. Blake × E. pellita F. Muell.) clones in Quang Tri province at three planting densities (1333, 1667 and 2222 trees ha⁻¹). The experiment was planted on an eroded shallow soil common in the region. At age 5 years, survival of acacia (74%) was higher than that of eucalypt (67%), a consequence of high mortality from wind damage for one eucalypt clone. Eucalypt was taller by about 2 m, but stem diameters of acacia and eucalypt were very similar. For both taxa, diameter decreased significantly as planting density increased. Across planting densities, mean standing volume was 107 and 108 m³ ha⁻¹ for acacia and eucalyptus, respectively. Linear regressions of stocking at 5 years on volume accounted for over half of the variance in acacia and eucalypt plot volumes, demonstrating the strong effect of stocking on yield. There were similarly strong effects of stocking on stem diameter. Acacia hybrid plantations of nearby small growers had stockings at age 5 years that averaged over 2500 stems ha⁻¹. Growers planted at higher densities and allowed their trees to multi-stem. Their standing volumes at age 5 years ranged from 83 to 102 m³ ha⁻¹. Understanding how to reduce tree mortality would assist growers to choose planting densities and stand management that optimise growth, log diameter classes and net returns.

1. Introduction

Vietnam has the largest acacia plantation estate of any nation worldwide [1,2], and acacia plantations dominate Vietnam’s wood supply. They support regional livelihoods through wood sales, local processing and chip export [3,4]. Eucalypt plantations also contribute to commercial wood production, but at a smaller scale. Nationally, Vietnam’s acacia and eucalypt plantation areas reached 2.2 and 0.3 Mha, respectively, by 2022 [2] and are increasing. Over half of the total area is managed by small growers owning 5 hectares or less [1,5].

In central Vietnam, the first small grower plantations established in the late 1980s with support from the United Nations World Food Program used Acacia auriculiformis, A. mangium and Eucalyptus camaldulensis Dehnh. [6]. Plantations of E. camaldulensis were severely damaged by fungal leaf pathogens such as Cylindrocladium quinqueseptatum Boedijn and Reitsma [7], and by the mid-1990s, growers had abandoned this species. Since the early 2000s, most growers in central Vietnam have planted selected clones of acacia hybrid (A. mangium × A. auriculiformis), which are productive throughout lowland Vietnam [8]. Acacia mangium and some eucalypt hybrid clones are also planted, but on a much smaller scale. Other timber tree species are little planted by small growers, who favour the fast growth and profitable early wood harvests obtainable from acacia hybrid [9].

A previous study [10] on a shallow, stony and eroded soil, typical of local hill sites in Quang Tri province, central Vietnam, found that with conservation of slash and litter during the inter-rotation period, avoidance of ploughing and burning and regular but moderate weed control, productivity of acacia hybrid was maintained over two successive rotations. In the second rotation at age 7.6 years, mean annual increment (MAI) was 20 m³ ha⁻¹, and 46% of this wood volume was classified as sawlog (a mid-point log diameter > 12 cm, the local market criterion). This experiment was planted at 1428 trees ha⁻¹, a lower planting density than used by small growers. Application of 10 g P tree⁻¹ at planting gave a modest improvement in early growth; higher rates gave no additional benefit [10]. A third rotation on this site, using the same set of clones, the same planting density and similar site and stand management, had an MAI at age 3.7 years, averaged across five fertiliser and weed control treatments, of 22 m³ ha⁻¹ (Forest Science Centre of North Central Vietnam, unpublished data, 2022). Shortly afterwards, this experiment was lost through the construction of a highway.

Attacks by pests and diseases and wind damage caused by frequent typhoons [11] are threats to wood production in the region. In Indonesia, outbreaks in the stem canker Ceratocystis Ellis & Halst. and Ganoderma Karst. root rot led to the abandonment of 600,000 ha of A. mangium plantations and compromised national wood supply [12]. This major setback is being managed by replacing A. mangium with E. pellita and its hybrids, which are more resistant to these diseases [13]. Ceratocystis is present in Vietnam [14] and while it has not yet had a notable impact on wood production, it poses a serious threat. Even if acacia hybrid plantations remain productive, the deployment of two or more alternative planting taxa, both spatially across local plantation landscapes and in successive rotations on individual planting blocks, might help to reduce the build-up of pests and diseases [12]. It is therefore prudent to explore alternative planting species, and the associated silvicultural techniques that they will require.

The majority of plantations in central Vietnam are established at high stockings of 2000 or more stems ha⁻¹ and are harvested on short rotation cycles of about 5 years. They yield small-diameter (<12 cm over bark) logs that are primarily converted to woodchips for export or the local manufacturing of medium-density fibreboard [1,2,15]. There has been little research published in Vietnam, or elsewhere, on the effect of planting density on wood volume and log diameter outcomes for tropical acacia species. In Brazil, a study evaluating 18 eucalypt clones across 27 tropical and subtropical sites [16] found that as planting density increased from 450 to 4000 stems ha⁻¹, wood production per hectare in the third year of growth increased markedly, but most of this increase occurred as planting density increased from 450 to 2000 stems ha⁻¹, with minimal further gains at higher densities at most of the sites and for most of the clones. It is therefore important to investigate optimum planting densities for both acacia and eucalypt plantations for central Vietnam, especially because most growers use high planting densities [10] and allow trees to multi-stem, resulting in high stocking [5].

The experiment reported here is a comparative study of the performance of acacia and eucalypt hybrid clones on a shallow soil, common in the region, which tested whether newly developed, disease-tolerant eucalypt hybrid clones provide an option for wood production. The effects of initial planting density on standing volume and log diameter were investigated. To complement the experimental study, we also surveyed stand establishment and management practices and the production of nearby plantations of acacia hybrid managed by small growers to assess their current growth rates and consider the applications of research results for growers.

2. Materials and Methods

2.1. Site

The experiment was located on gently sloping land within the Cam Lo field station of the VAFS Forest Science Centre for North Central Vietnam near Dong Ha, in Quang Tri Province, central Vietnam (16°28′ N, 106°5′ E, 50 masl). Dong Ha meteorological station, 10 km to the southeast of the experiment, has a mean annual temperature of 26 °C and mean annual rainfall of 3080 mm. Mean monthly temperature ranges from 20 °C in January to 30 °C in June. The wettest months (August- to January-inclusive) receive over 75% of the total annual rainfall, but the remaining months of the year all receive over 70 mm on average. Most years have a dry season (consecutive months with <40 mm) of 1 or 2 months.

The original natural forest was destroyed by war and logging over the latter decades of the twentieth century. This caused significant soil erosion and degradation, and by the 1980s, the local area was occupied by thickets of low, woody scrub. There were two successive plantation rotations at the experimental site. In the first, part of the site was planted to Eucalyptus sp. and the balance to A. crassicarpa A. Cunn. Ex Benth. In the second, the whole site was planted with acacia hybrid, which was harvested in July 2018. Merchantable wood and some firewood were removed, while fine branches (<1 cm diameter), foliage and litter were left on-site and were not burned.

The general description of the local soil is yellow to brown ferralitic grey soils [17], corresponding to Ferric acrisols/Skeletic acrisols in the FAO system. The soil profile features (Figure 1a) and soil properties were in general similar to those described for a previous experiment [10]. Here, soils had a somewhat lower stone content and less degradation of the A horizon. The top 0–10 cm of the profile was a grey–brown, sandy clay texture containing small angular lateritic stones. The textural components clay, silt, sand and coarse sand were 24, 38, 22 and 16%, respectively. In the next 10–30 cm horizon, the soil was a brown clay with a higher proportion of decaying reddish–brown laterite fragments of irregular shape, up to 2 cm in size. Below 30 cm, to a depth of 50 cm or more, the soil was a yellow–brown clay, highly compact with no structure, and embedded with fragments of pale brown sedimentary parent rock.

2.2. Experimental Design

A split plot design was used, with planting density treatments in the main plots and acacia and eucalypt sub-plots within each main plot. There were four replicates, laid out contiguously in a 2 × 2 arrangement. A spacing of 3 m between tree rows and spacings of 1.5, 2.0 or 2.5 m between trees within rows gave planting densities of 1333, 1667 and 2222 stems ha⁻¹. For these three planting densities, the species net plots had 80, 104 and 128 trees (8 rows × 10 trees, 8 rows × 13 trees and 8 rows × 16 trees), respectively (

Table 1. Dimensions of species plots and sequences of clones within net plots.

	Planting Density (Trees ha−1)
	1332	1667	2222
Spacing (m × m)	3 × 2.5	3 × 2	3 × 1.5
Number of trees in gross plot	10 × 12	10 × 15	10 × 20
Number of trees in net plot (rows × trees)	8 × 10	8 × 13	8 × 16
Dimensions of gross plot (m × m)	30 × 30	30 × 30	30 × 30
Dimensions of net plot (m × m)	24 × 25	24 × 26	24 × 24
Area of net plot (m2)	600	624	576
Sequence of clones (acacia)	(BV) 16, 10, 33, 75, 16, 10, 32, 75
Sequence of clones (eucalyptus)	(UP) 54, 35, 72, 223, 54, 35, 72, 223

). The species net plots were surrounded by one or more buffer rows of the same clones at the same spacing, to give gross plot dimensions of 30 m × 30 m for species plots at each planting density (Table 1). Allocations of planting density treatments to main plots and of acacia or eucalypt to the species plots were randomised.

The experiment used four acacia hybrid clones (A. mangium × auriculiformis BV10, BV16, BV73 and BV75) and four eucalypt hybrid clones (E. urophylla × pellita UP35, UP 54, UP72 and UP223), all developed by the Vietnamese Academy of Forest Science. These acacia and eucalypt clones were each represented by two separate line plots in the rows within their respective species net plots. To simplify planting, the sequences of individual clones allocated to the eight line plots within the species net plots (Table 1) were kept the same across all stocking treatments and replicates.

2.3. Establishment and Management

Planting positions were dug using a tracked excavator, creating 40 × 40 × 40 cm cultivated holes. Superphosphate (100 g per tree, equivalent to 7 g P per tree) was mixed into the soil in each planting hole before planting, which took place in December 2018. One month later, any seedlings that had died were re-planted using matching genetic stock. Weeds were controlled manually before planting and then twice annually to age 3 years. Trees were singled at age 6 months to remove competing stems, retaining one main stem. Side branches were pruned to a height of 1.5 m at age 2 years.

2.4. Assessment and Statistical Analysis

Tree height (Ht) and diameter at breast height (Dbh) of all live trees were measured 1.2, 2.2, 3 and 4 years after planting. Information from these annual measurements was used to examine the trend in mortality. For stand growth, we focussed on 5-year data. At this age, Dbh was measured directly, and Ht was predicted from Dbh. Sets of 36 acacia and eucalypt trees in the buffer rows, spanning the range of Dbh in the experiment, were felled and their heights measured. Dbh and Ht data of the felled sets of trees were used to develop the following Dbh–Ht relationships, which were applied to predict the Ht of all trees in the net plots:

• Acacia hybrid: Ht = 4.25 log Dbh + 3.3402 (r² = 0.91, n= 36)
• Eucalypt hybrid: Ht = 10.199 log Dbh − 9.204 (r² = 0.93, n= 36)

Stem volume over bark of each tree was calculated using the following equation:

Vol = ff × π × Ht × Dbh²/4 × 10,000

where ff = form factor, Vol = volume over bark in m³, Dbh = diameter at breast in cm and Ht = height in m. A form factor of 0.495, developed for young clonal acacia hybrid trees [18], was used here for both acacia and eucalypt. The total standing volume for each species plot was calculated by summing the volumes of all trees. Values for stocking, percentage survival, mean height and mean Dbh were calculated for species plots. Dbh, height and survival for each clone at the plot level were also calculated.

The significance of treatment effects was evaluated by a univariate fixed-effects analysis of variance, using the software package Genstat Release 22 (VSN International, Hemel Hempstead, United Kingdom). For each response variate, the set of species plot means was analysed with treatment specified as planting density × species and blocking structure as replicate/main plot/species plot. A separate analysis of variance was carried out to compare Dbh, height and survival of individual clones. This analysis specified treatment as planting density × clone and blocking structure as replicate/main plot. Survival was analysed for all five measurement times. Plots of residual versus fitted values were examined to confirm that the assumptions underlying the analysis of variance, that residuals should be independent of treatment identity and normally distributed around zero, were satisfied [19].

In view of the substantial and varying levels of mortality that were observed, we examined the relationships between stocking, standing volume and Dbh using plot level data. Linear regressions of standing volume on stocking and Dbh on stocking were calculated for each species.

2.5. Soil and Litter Sampling

The litter layer and the surface soil were sampled 5.4 years after planting. In the acacia and eucalypt plots established at the highest planting density (2222 stems ha⁻¹) in each replicate, samples were collected from six randomly located inter-row positions close to the plot centre. At each sampling point, all the litter within a 30 × 30 cm quadrat was collected, and then a 0–10 cm soil core, 6 cm in diameter, was taken. The litter and soil samples from each plot were bulked, bagged and transferred to the laboratory of the Forest Science Institute of South Vietnam for analysis. Litter samples were oven-dried and weighed, enabling the estimation of per-hectare litter mass. A representative sub-sample of litter was ground for analysis of C, N and P. Soil samples were thoroughly mixed and sieved to exclude fractions >2 mm. Subsamples were used for measuring pH (H₂O and KCl), organic C, Bray-1 extractable P and total N concentrations and soil texture, following procedures used previously [10,20]. The significance of differences in litter and soil properties under acacia and eucalyptus was compared using two-tailed t-tests.

2.6. Small Grower Plantations

To determine the productivity and stand conditions of acacia hybrid in small grower properties within about 5 km of the experimental sites and examine possible trends over time, we used inventories of four plantations carried out in 2015 [10] and conducted two further inventories.

In 2022, four plantations that were 4.9–5.0 years old were assessed. Within each plantation, four plots, each 15 m × 15 m, were located randomly. Height and Dbh of all live stems in the plots, including multiple stems, were measured. Individual stem volumes were calculated from Dbh and height as described above and summed to give plot volumes, and averages for each plantation were calculated.

Acacia hybrid typically develops multiple stems unless young saplings are singled during the first year, as was done in our experiments. In May 2024, ten small grower plantations of acacia hybrid, 4–5 years old, were assessed to understand the patterns of initial stocking, mortality and multi-stemming. In each plantation, a single randomly located plot covering 100 planted trees (10 rows × 10 planting positions) was measured, and the numbers of missing trees and surviving trees with single, double and triple stems were counted. From this data planting density, mortality and stocking (stems ha⁻¹) were calculated.

3. Results

3.1. Comparative Performance of Acacia and Eucalypt Hybrid

At age 5 years, the survival of acacia (mean of 73.9% across the three planting densities) was significantly (p < 0.01) higher than that of eucalypt (66.9%). Eucalypt was significantly (p < 0.001) taller than acacia, with a mean height across the three stockings of 16.0 m, compared with 13.9 m for acacia. The Dbh of acacia and eucalypt did not differ significantly, with overall means of 12.3 and 12.0 cm for the two species (

Table 2. Stocking, survival and growth at age 5 years.

	Acacia Hybrid					Eucalypt Hybrid
Planting density	Stock.	Surv.	Dbh	Ht	Vol.	Stock.	Surv.	Dbh	Ht	Vol.
trees ha−1	trees ha−1	%	cm	m	m3 ha−1	trees ha−1	%	cm	m	m3 ha−1
1332	1042	78.1	13	14.2	101	888	66.6	12.9	16.7	100
1667	1226	73.6	12.2	13.9	100	1086	65.1	12.1	16.2	104
2222	1558	70.1	11.7	13.7	119	1537	69.1	11.1	15.1	119
Mean	1275	73.9	12.3	13.9	107	1170	66.9	12	16	108
s.e.d. a (planting density)						100	6.0	0.27	0.13	9.9
s.e.d. (species)						40	1.9	0.16	0.08	4.2
s.e.d. (planting density × species)						111	6.5	0.33	0.16	11

^a Standard error of difference of treatment means.

).

Planting density had a significant effect on Dbh (p < 0.01) and height (p < 0.001). Dbh and height both decreased at higher stockings. Planting density did not significantly affect the survival percentage, but it significantly (p < 0.01) influenced stocking: planting densities of 1332, 1667 and 2222 stems ha⁻¹ resulted in stockings of 1042, 1226 and 1558 trees ha⁻¹ for acacia and 888, 1086 and 1537 trees ha⁻¹ for eucalypt (Table 2).

Averaged over the three planting densities, standing volume was 107 and 108 m³ ha⁻¹ for acacia and eucalyptus, respectively, equivalent to an MAI of 21–22 m³ ha⁻¹. At lower planting densities volumes were lower for both species, but the effect of planting density on volume was not statistically significant (

Table 3. Significance of treatment differences ^a from the analysis of variance.

	Stocking	Survival	Dbh	Height	Volume
Planting density	p < 0.01	n.s.	p < 0.01	p < 0.001	n.s.
Species	p < 0.05	p < 0.01	n.s.	p < 0.001	n.s.
Planting density × species	n.s.	n.s.	n.s.	p < 0.001	n.s.

^a n.s. = not significant.

). This was primarily a consequence of high variation in survival and consequently in volume.

The interaction between planting density and species was significant (p < 0.001) for height. Planting density had a stronger influence on height for eucalypt than for acacia. The interactions between planting density and species were not significant for Dbh, volume, stocking or survival (Table 3).

There were striking differences in the understory that had developed since the final weeding at age 3 years. The understory in the acacia was low, with a scattered shrub layer generally lower than 1 m, while in the eucalypt there was a much denser shrub understory exceeding 2 m in height (Figure 1b). There were clear differences in the structure of tree canopies. Eucalyptus trees (Figure 1c) had narrower, shallower and sparser crowns than those of acacia (Figure 1d). These understory and crown differences were consistent throughout the experiment.

The linear regressions of standing volume on stocking (Figure 2a) were significant (p < 0.001) and accounted for 58% and 59% of the variance for acacia and eucalypt, respectively. The corresponding relationships between Dbh and stocking (Figure 2b) were even stronger. Stem diameter decreased as stocking increased, the linear regressions accounting for 74% and 63% of the variance for acacia and eucalypt, respectively.

3.2. Differences among Clones

There were significant (p < 0.001) differences among the eight clones for height, Dbh and survival (

Table 4. Growth and survival of clones at age 5 years (mean of three planting densities).

		Dbh	Height	Survival
	Clone	cm	m	%
Acacia hybrid	BV10	12.2	13.9	80
	BV16	13.1	14.2	69
	BV73	12.2	13.9	72
	BV75	11.9	13.8	74
Eucalypt hybrid	UP223	12.3	16.2	76
	UP35	12.7	16.6	70
	UP54	11.3	15.4	79
	UP72	11.8	15.8	42
Critical difference (p = 0.05)		0.44	0.24	9.6
Significance of differences among clones		p < 0.001	p < 0.001	p < 0.001

), while interactions between clones and stockings were significant for Ht but not for Dbh or survival. At 5 years, the mean Dbh of clones ranged from 11.3 to 12.7 cm, and Ht from 13.8 to 16.2 m. However, eucalypt clone UP72 suffered high mortality, primarily as a result of typhoon damage, with only 42% survival at 5 years, compared to 69%–80% for the other clones (Table 4).

The mortality trends for the individual clones, pooled across the three planting densities, are shown in Figure 3. Tree death began soon after planting and survival of all clones had fallen below 87% at the first assessment at 1.3 years. Significant differences between clones had emerged by age 3 years.

3.3. Litter and Soil Characteristics

The mass of surface litter under acacia (4.4 t ha⁻¹) was slightly greater than that under eucalypt (3.6 t ha⁻¹) (

Table 5. Properties of litter and topsoil sampled at stand age 5.4 years.

	Acacia 1		Eucalyptus 1		Significance 2
Litter	Mean	s.e.	Mean	s.e
Dry mass (t ha−1)	4.4	0.01	3.6	0.4	n.s.
N (%)	1.17	0.11	1.03	0.09	n.s.
P (%)	0.064	0.01	0.051	0.01	n.s.
C (t ha−1)	1.8	0.1	1.3	0.1	p < 0.5
N (kg ha−1)	51.4	3.2	38.1	5.3	n.s.
P (kg ha−1)	4.8	0.4	3.4	0.5	n.s.
Soil (0–10 cm)
pH (H2O)	5.0	0.02	4.9	0.06	n.s.
pH (KCl)	4.3	0.04	4.2	0.02	n.s.
Organic C (%)	1.9	0.06	1.4	0.09	p < 0.01
N (%)	0.15	0.01	0.12	0.01	p < 0.01
Bray-1 P (ppm)	4.0	0.08	3.4	0.32	n.s.

¹ Mean of four plots, standard error of mean in brackets. ² Significance of difference between acacia and eucalyptus. n.s. = not significant.

). The concentrations of N and P in the acacia litter were likewise higher than those of eucalypt. As a consequence, the C, N and P contents per hectare of the acacia litter were each about one-third higher than those of eucalypt, the species difference for C content being significant. In the surface soil, the concentrations of organic carbon, total N and extractable P were higher under acacia than under eucalypt. The species differences for N and C were significant (Table 5). Soil pH was similar for both species.

3.4. Small Grower Acacia Hybrid Plantations

A typical small grower plantation is pictured in Figure 4, which shows the high stocking partly due to the proliferation of multi-stems. The results of the two inventories of growth, conducted seven years apart, are presented in

Table 6. Stand attributes and growth of small grower plantations.

			Planting Density	Stocking	Dbh	MAI	Mortality	Multi-Stem
Year	Sites		trees ha−1	stems ha−1	cm	m3 ha−1	%	%
	4	Mean	1825	1536	10.6	19	n.d. a	n.d.
		Max.	1900	1556	11.1	20
		Min.	1800	1511	10	17
2022	4	Mean	2750	2500	9.5	19	n.d.	n.d.
		Max.	3000	2756	10	21
		Min	2300	1989	8.8	17
2024	10	Mean	3062	2853	n.d.	n.d.	22	20
		Max.	4444	3644			28	28
		Min.	1667	1917			9	9
		s.e. b	302	179			2.2	1.9

^a n.d. not determined, ^b s.e. = standard error.

. The plantations assessed in 2022 were planted at densities of 2300 to 3000 trees ha⁻¹, much higher than those in 2015 (1800 to 1900 trees ha⁻¹). Accordingly, stocking at age ~5 years was higher in 2022 than in 2015 (2500 and 1536 stems ha⁻¹, respectively). This resulted in a lower Dbh (9.5 cm in 2022 and 10.6 cm in 2015). Despite these difference in stand attributes, MAIs in 2015 and 2022 were about the same, within the range of 17–21 m³ ha^—1 and averaging 19 m³ ha⁻¹. Most growers surveyed in 2024 had planted at a high rate (a mean of 3062 trees ha⁻¹, standard error, 304). The mortality at age 4 years averaged 22% (range of 9 to 28%). The net loss of planted stock in these plantations was high (average 711 trees ha⁻¹, range from 170 to 1200 trees ha⁻¹). The percentage of multi-stemmed trees averaged 20%. Because multi-stemming tended to compensate for mortality, stocking had fallen a little below planting density, averaging 2853 stems ha⁻¹.

4. Discussion

Since its introduction in the early 2000s, acacia hybrid has enabled growers to increase rates of wood production relative to those of the species it replaced, primarily A. mangium, A. auriculiformis, E. camaldulensis and E. urophylla [6,9]. In central Vietnam, the productivity of acacia hybrid plantations varies greatly. Based on the inventory of 220 small growers in four provinces, four site quality classes with a mean MAI of 5-year rotations ranging from 14.2 m³ ha⁻¹ (the lowest site quality class) to 28.8 m³ ha⁻¹ (highest class) were described [9]. In another study, the MAI of 5-year-old plantations at six sites in Thua Thien Hue province averaged 28.7 m³ ha⁻¹ [21]. Our surveys, carried out in 2015 and 2022, found that MAI of small grower plantations in Quang Tri province averaged 19 m³ ha⁻¹ for 5-year rotations (Table 6). Even over a distance of about 200 m down a 10% slope within an experiment, a downslope increase in growth of 38% was observed [10]. Several acacia hybrid clones that display very similar growth rates in comparative trials [8] are planted throughout central Vietnam, so genetic differences are unlikely to explain such spatial variation. Differences in the physiological quality of clonal planting stock, arising from differing nursery management, might affect the growth rate [22]. But high variation in site factors, common in war-damaged landscapes including the degree of soil erosion, soil profile depth, stoniness, texture and fertility, undoubtedly contribute to the reported high variation in productivity. Management regimes also differ between growers. Each grower wishes to make the best use of their land, and therefore, the question is: what are the options—in terms of genetics choices and site and stand management—for growers to sustain and increase wood yields and net returns over successive rotations?

4.1. Species Choices and Genetic Improvement

To help manage bio-physical risk, it is necessary to identify alternative species that could substitute for acacia hybrid should disease, pests or other unforeseen threats impact its viability. The results presented here show that selected clones of the recently developed eucalypt hybrid combination E. urophylla × pellita may yield rates of wood production comparable to those from acacia hybrid on shallow hill soils. The sets of four acacia and eucalypt clones yielded similar volumes at age 5 years, at each of the three planting densities. The MAI of approximately 24 m³ ha⁻¹ at the highest planting density of 2222 stems ha⁻¹ was about 20% greater than those achieved by nearby small growers (Table 6). The comparison of standing volumes between species is not precise. The form factor developed for acacia hybrid [18] was used for both acacia and eucalypt. A specific form factor for these eucalypt clones needs to be determined. Bark thickness, which may vary between taxa, also affects the yield of usable wood.

Acacia hybrid plantations in Vietnam are typically established using about three clones supplied by the forest nursery and planted in mixture, although some plant a single clone. Planting clone mixtures has been practiced by many, but not all growers in Vietnam, for more than two decades. In Brazil [23], there were increases in wood yields when eucalypt clones were planted in mixture in comparison to planting them in monoclonal blocks, under both experimental conditions (an average 9.8% increase) and commercial conditions (an average 6.3% increase), and most individual clones produced more wood in mixture than in monoculture.

All the acacia and eucalypt clones tested in our experiments grew well given the limitations of the soil and site. While the acacia hybrid clones have been well-tested and planted commercially throughout Vietnam [8,24], the eucalypt hybrid clones are less well-tested. The substantial mortality of eucalypt clone UP72, largely due to wind damage at age 3–4 years, reduced overall eucalypt volume. Clone UP72 was found to be similarly susceptible to wind damage in a trial at Ba Vi, Hanoi Province, northern Vietnam (N.D. Kien, pers. comm. 2024), so replacing it in clone mixtures with another clone displaying better survival would be advisable.

The breeding of new acacia and eucalypt clones can offer some increase in production [25]. In trials at Cam Lo and elsewhere in Vietnam, a new triploid acacia hybrid clone had significantly higher Dbh and wood volume than the acacia hybrid clones used in the current experiment [24]. Selecting acacia and eucalypt clones that are tolerant to Ceratocystis and other serious diseases [26,27] can help to reduce ecological risks and widen planting options.

4.2. Site Management

Under eucalypt stands, a more vigorous understory re-grew after weed control was discontinued from age 3 years (Figure 1b). The reason may have been that the narrower and sparser crowns of eucalypts (Figure 1c) are less effective in shading and suppressing understory growth than those of acacias. We were not able to assess if understory competition had any effect on stand growth; this will require attention if eucalypts are planted widely [14]. The results showed that while vegetation management is critical in the first 2–3 years after planting, beyond this stage it is possible to foster a biodiverse understory in productive single-species plantations. A long-term study in south Vietnam [28] showed that in a third rotation, A. auriculiformis plantation, following the cessation of herbicide application two years after planting, a diverse understory could be restored with no adverse effects on stand attributes including LAI and growth. Applied research on the dynamics of plantation development and understory (weed) management is an area warranting significant attention in Vietnam [5,11].

Acacia mangium plantations fix nitrogen in the soil symbiotically [29]. There were significant gains in soil N with stand age in an A. auriculiformis plantation in South Vietnam [20]. Our experiment was established on a site which was previously planted to acacia hybrid, and hence, the eucalypt hybrid may have benefitted from the legacy effect of the acacia rotation, as has been found in Sumatra [13]. This source of N supply would decline over successive eucalypt rotations, since eucalypts do not fix N. In Sumatra [13] and more extensively in Brazil [30], eucalypts also respond to potassium fertiliser. The results in this study indicate that litter mass and the associated levels of C, N and P are somewhat higher under acacia than under eucalypt (Table 5), a factor in favor of acacia, especially under the low input management adopted by small growers. On a soil closely comparable to that of this experiment, the application of 10 g of P per tree at planting appeared sufficient to maintain production over three successive acacia hybrid rotations ([10], Forest Science Centre of North Central Vietnam, unpublished data, 2022), but that may not be the case for eucalypts. For E. pellita in Sumatra, significant additional wood volume was obtained from higher rates of P than those applied in our study [31]. Eucalypts may also require the application of N fertiliser over time.

The critical importance of conserving site resources (organic matter, nutrient capital and overall productive capacity of soils) is now well-established as an essential prerequisite for sustainable plantation forestry, especially in short rotations in the tropics [32] and specifically in Vietnam [5,28] and Indonesia [13]. Our field observations suggest that most small growers now retain slash and litter on site and refrain from burning and repeated ploughing, which were common practices in the past.

4.3. Planting Density, Stand Management and Survival

In short-rotation forestry in this region, planting at a spacing of 3 m × 2 m (1667 stems ha⁻¹) may be appropriate. This will result in over 1000 stems ha⁻¹ at age 5 years, assuming typical current levels of mortality. It is prudent to maintain stocking through the rotation at or above 1000 stems ha⁻¹ (Figure 2a). If stocking falls below this, the standing volume at 5 years, the age at which most local small growers harvest, will be reduced below the site potential. Similar findings emerged from a broad-scale assessment of acacia and eucalypt plantation productivity across Southeast Asia [33] and from a recent experimental study in Sumatra, Indonesia [13]. In our experiment, trees were singled in the first year, so very few trees developed multiple stems. The majority of small growers do not single their trees, allowing many to multi-stem ([5], Figure 3) and employ high planting densities (Table 5). The resulting very high stockings do not result in increased volume production but lead to a mean Dbh seldom exceeding 10 cm at the typical harvest age. The numbers of trees lost over 5-year rotations on small grower plantations are high, averaging 711 trees ha⁻¹ in the plantations surveyed in 2024.

If growers reduced planting densities from the current levels of about 3000 stems ha⁻¹, establishment costs could be lowered through reductions in the number of planting holes to be dug and cultivated, the number of seedlings purchased and in fertiliser and tending costs. This would not reduce wood yields but would increase stem diameter classes, thus potentially improving net returns. A detailed financial analysis is beyond the scope of this study, but we estimate, based on current costs, that establishment costs would be reduced by VND 5–7.5 M (USD 200–300) ha⁻¹ if stockings were reduced from 3000 to 2000 stems ha⁻¹.

The present management followed by most small growers will hinder any production of saw logs from short rotations (Figure 2b). Stand thinning and extending the rotation length would produce larger diameter logs [34,35] and are promoted, for example, through government policy, extension training and certification schemes [1,36]. We note that on local soils, planting densities for acacia hybrid at around 1400–1600 stems ha⁻¹ can yield about 50% sawlog by volume by age 7–8 years without thinning, assuming current mortality levels of 20%–30% and provided that effective singling is implemented [10].

In Quang Tri province, prices paid in 2024 for acacia hybrid logs, at the sawmill gate, were VND 1.1 M, 1.2 M and 1.3 M per green tonne for logs with midpoint diameters less than 12 cm, 12–15 cm and 15–20 cm, respectively (L.X. Toan, unpublished data, 2024). These prices provide no incentive for growers to grow sawlogs. Elsewhere in Vietnam, for example, in the southern province of Bin Duong where the sawmilling industry is prominent, larger-diameter acacia logs fetch higher unit prices [37].

Tropical eucalypts, unlike acacia, do not multi-stem from planting. They do so if coppiced after harvest. Most eucalypt growers in Southeast Asia opt for a second coppice rotation, rather than replanting [33]. This gives the option of regulating stocking (i.e., stem numbers) in the coppiced rotation. Since acacia hybrid does not coppice, it must be replanted after each rotation.

4.4. Future Research Needs

Acacia hybrid is the predominant planting species in central Vietnam, as evidenced by the ongoing increase in plantation area and grower numbers. However, research and the wider adoption of best management practices have not kept pace with this expansion. The limited survey results reported here suggest no productivity improvement in small grower plantations over time. Despite government policies advocating sawlog production [36], there has been little uptake of appropriate silviculture in the region. We have shown here and previously [10] that the management practices employed by government centres and by nearby growers are very different. For example, the planting densities employed by most smallholders are very high, and they allow multi-stemming. This does not increase wood production and results in small-diameter stems. The reasoning underlying these practices is not well-understood. It may be that very high stockings are seen as a way to suppress weed growth. Mortality is a major constraint for good management and production. Yet, the causes of tree mortality in the region, and how mortality might be reduced, are not known. Likewise, it is not known whether rankings of individual clones for growth and survival determined in experiments remain the same under contrasting smallholder management regimes. Much refinement is required to improve weed management and fertiliser use.

Most research in Vietnam is carried out on government-run regional centres and researchers tend to count on the extension of the results through occasional field days and extension literature [36]. To date, very little research has been conducted within Vietnam’s small grower plantations and by directly engaging growers, despite their pivotal role in national wood supply [5]. The importance of on-farm adaptive research involving networks of farmers for developing and scaling up improved methods in agroforestry has been described and advocated [38]. In the context of this paper, a detailed adaptive on-farm research approach specifically targeted for small grower forestry in the tropics, especially in Vietnam, has been proposed [1,11]. At its core is a network of simple on-farm experiments. It will be rewarding to direct a proportion of future research and application effort along those lines, focussed on on-farm experimentation and demonstrations.

5. Conclusions

On a shallow hill soil in central Vietnam, newly developed clones of the interspecific hybrid E. urophylla × pellita achieved growth at 5 years comparable to that of acacia hybrid (A. mangium × auriculiformis). Eucalyptus hybrid may offer an alternative planting option to acacia hybrid, which currently dominates plantations in the region. Planting density significantly influenced stem diameter but not survival in both species. Stocking at 5 years had a strong influence on standing volume and stem diameter in both species, confirming the importance of survival in determining wood yield and log characteristics. Small growers favour high planting densities and do not single their young trees. This does not increase harvestable wood volume, but results in very high stocking and small log diameters at harvest. To improve productivity, value and net returns, efforts should be directed to partnering with small growers in on-farm research and demonstrations.