Growth and wood basic density of acacia hybrid clones at three locations in Vietnam

Growth and wood basic density of acacia hybrid clonesat three locations in Vietnam

Le Dinh Kha • Chris E. Harwood • Nguyen Duc Kien •

Brian S. Baltunis • Nguyen Dinh Hai • Ha Huy Thinh

Received: 11 August 2010 / Accepted: 30 March 2011 / Published online: 10 April 2011� Springer Science+Business Media B.V. 2011

Abstract Field trials testing a total of 27 clones of the interspecific hybrid Acaciamangium 9 A. auriculiformis and seedling controls of the parental species were estab-

lished at Ba Vi and Yen Thanh in the north of Vietnam and Long Thanh in the south. At

both Ba Vi and Yen Thanh there were significant (P \ 0.001) differences in height and

diameter at breast height (DBH) among 22 tested clones at 4 years. At Long Thanh, twelve

hybrid clones did not differ significantly in DBH at age 3 years, but did (P \ 0.001) at age

5 years. At the two northern sites the acacia hybrid clones had significantly greater DBH

than control seedlots of the parental species. At Long Thanh, DBH of the hybrid clones and

A. mangium was similar, with a genetically improved seedlot of A. mangium displaying the

best DBH. Mean wood basic density at breast height of the acacia hybrid clones was

539 kg m-3 at Yen Thanh at age 8 years, and 473 kg m-3 at Long Thanh at age 5 years;

density for A. mangium at Long Thanh was only slightly lower than the hybrid clones at

461 kg m-3. Linear regressions of Pilodyn penetration (PP) at breast height on wood basic

density explained 60% of the variance in density of treatments (clones and control seed-

lots) at Yen Thanh and 36% at Long Thanh. There were significant differences between

hybrid clones in PP at all three trial sites. Clonal DBH performance was not strongly

correlated across the three trial sites; Pearson correlations of clone mean DBH between

pairs of sites ranged from -0.47 to 0.20. Clonal rankings for PP were more stable, with

Pearson correlations between pairs of sites ranging from r = 0.71 to 0.78.

Keywords Acacia mangium 9 auriculiformis hybrid �Clone-by-environment interaction � Pilodyn

L. Dinh Kha � N. D. Kien (&) � N. D. Hai � H. H. ThinhResearch Centre for Forest Tree Improvement, Forest Science Institute of Vietnam,Tu Liem, Hanoi, Vietname-mail: [email protected]

C. E. HarwoodCSIRO Ecosystem Sciences, Private Bag 12, Hobart 7001, Australia

B. S. BaltunisCSIRO Plant Industry, GPO Box 1600, Canberra, ACT 2601, Australia

123

New Forests (2012) 43:13–29DOI 10.1007/s11056-011-9263-y

Introduction

The development of fast-growing selected clones of the natural interspecific hybrid

combination A. mangium Willd. 9 A. auriculiformis A. Cunn. ex Benth. (hereafter

referred to as acacia hybrid) is a major contribution to plantation forestry and the rural

economy in Vietnam (Kha 2009). By the end of 2009, the area of acacia hybrid plantations

was estimated at 232,000 ha by Vietnam’s Ministry of Agriculture and Rural Development

(MARD, unpublished data), making acacia hybrid the most important plantation tree

variety in Vietnam. Wood of acacia hybrid has been found to be suitable for export

woodchips, kraft pulp production in Vietnamese pulp mills, and also for sawing to produce

furniture components (Kha 2001). Acacia hybrid is also planted on a large scale for

pulpwood production in Karnataka State, India (Amanulla et al. 2004) and plantations are

under development in Malaysia and other SE Asian countries (Mohd. Ghazali et al. 2007).

Pure-species plantations of A. mangium and A. auriculiformis have developed as

important plantation species in Vietnam, following seed introductions from natural prov-

enances in Papua New Guinea and north Queensland in the second half of the twentieth

Century (Nghia and Kha 1998; Hai et al. 2008a). Commencing in 1992, acacia hybrid

clones were developed from occasional hybrid individuals identified in young A. mangiumplantations in Vietnam (Kha 2001). To date, most acacia hybrid genotypes selected for

clonal testing appear to be first-generation A. mangium 9 auriculiformis hybrids, pre-

sumably arising from open-pollination by A. auriculiformis fathers growing near the A.mangium mother trees that supplied the seed used to establish the plantations. Many of the

acacia hybrid genotypes identified in the A. mangium plantations are not acceptable for

production forestry, because they are unsatisfactory with respect to one or more important

traits such as ease of rooting, growth, branch size, stem form or pest and disease sus-

ceptibility (Kha 2001). Phenotypically superior candidate acacia hybrid individuals are

coppiced, brought into clonal propagation via rooted stem cuttings, followed by clonal

testing of those clones that are easy to propagate. Candidates under clonal testing are

captured in micropropagation to enable mass-production of clonal hedge plants. Clone

trials at multiple locations are established to rank the clones for survival, growth, stem

form and branching, resistance to pests and diseases and wood properties (Kha 2001).

Well-tested, outstanding individual hybrid clones are then approved by MARD for use in

production plantations.

From the first set of clonal tests in the 1990s, six hybrid clones with high productivity,

health and stem quality (clone numbers BV10, 16 and 32, and TB 3, 6 and 12) were

approved by MARD for commercial use in the year 2000. These selected clones were

shown to significantly out-perform pure-species seedlots of A. mangium and A. auriculi-formis in field trials at trial locations in lowland northern, central and southern Vietnam

(Kha 2001). On a productive site at Bau Bang, in Binh Duong province in southern

Vietnam, with intensive cultivation, productivity of the best acacia hybrid clones in a

5-year rotation at close spacing (3 m 9 2 m) has exceeded 30 m3 ha-1 year-1 (Kha et al.

2005). At Ba Vi in northern Vietnam on shallow and infertile soils and with a cool and dry

winter season, productivity of acacia hybrid can reach 15 m3 ha-1 year-1 while that of

A. mangium on these sites is only 9 m3 ha-1 year-1 (Dao 2003).

Ideally, a large clonal plantation estate needs to be supported by a program of breeding

and clonal development and testing to supply many improved production clones. This

provides a safeguard against potential susceptibility to emerging pests and diseases, pro-

vides clones adapted to the range of target planting environments, as well as yielding

ongoing improvement in growth performance and wood properties (Burdon and Aimers-

14 New Forests (2012) 43:13–29

123

Halliday 2006). The Forest Science Institute of Vietnam (FSIV) and other research

agencies in Vietnam have therefore developed additional production clones. By 2008,

MARD had approved a total of nineteen acacia hybrid clones for use in operational

plantations. However, only six (BV10, 16, 32 and 33, and TB6 and 12) are preferred by

plantation growers (FSIV, unpublished data).

When selecting clones for deployment, the relative weightings to be placed on traits that

are important for the target production systems must be considered. Surveys of enterprises

involved in acacia hybrid growing and processing in Vietnam (PH Hai and CE Harwood,

unpublished data) showed that smallholder growers, who manage about half of the acacia

plantations in Vietnam, typically sell their crops to buyers based on stand volume. Export

woodchip mills purchase logs by fresh weight and sell woodchips based on dry weight of

the chips they sell, giving them an interest in wood density and log moisture content as

well as volume. Pulp mills value the percentage of pulp yield obtainable per unit of dry

wood, in addition to basic density and volume (Greaves et al. 1997). Profitability of

sawmilling and subsequent processing of acacia wood in Vietnam is affected not only by

log size and straightness, but by the extent of knot-related defects, wood stiffness and

hardness, and the proportion of more durable heartwood, which affect the value of sawn

boards produced. It is therefore apparent that different interest groups would differ in the

relative weightings they give to different traits, when ranking a set of candidate clones.

However, increased volume and higher basic density at harvest are desirable traits for all

stakeholders, as they increase the volume of acacia wood and its value for all applications.

Harvest ages for acacia hybrid plantations range from about 5 years, for pulpwood

rotations on the most productive sites in Vietnam, to about 12 years for plantations on less

productive sites or plantations that aim to produce a high percentage of logs of sufficient

size for sawmilling. An assessment age of 3–5 years for clone trials therefore corresponds

to approximately half rotation age for commercial plantations in Vietnam, depending on

the region and the production system. Selection traits (Greaves et al. 1997) measured at this

age which can be expected to correlate closely with harvest-age stand volume include

survival, stem diameter at breast height (DBH) and total tree height. Wood basic density of

breast-height disks or cores or pilodyn penetration (PP) at breast height taken at around

half rotation-age usually correlate well with whole-tree basic density at harvest age for

temperate eucalypt species (Greaves et al. 2003), although these relationships have been

less well-studied for tropical acacia species (Hai et al. 2008b).

The extent of clone-by-environment interaction in Vietnam has recently been explored

for A. auriculiformis. In field trials of a total of 130 clones of this species at three sites in

northern central and southern Vietnam, Hai et al. (2008a) found low genotypic correlations

for growth traits between the sites, particularly between the southern site and the two other

sites. This result indicated that clone-by-environment interaction for growth was important

and would need to be taken into account in developing a clonal deployment strategy for

this species.

This paper reports the performance of a total of 27 selected acacia hybrid clones from

FSIV’s clonal testing programs, in clonal tests at three locations in northern and southern

Vietnam. Pure-species seedling controls of A. mangium and A. auriculiformis, raised from

seedlots of known and appropriate origin, were included as control treatments in the trials.

The acacia hybrid clones under test were selected by FSIV from a much larger number of

candidate acacia hybrid genotypes identified as natural hybrids in young A. mangiumplantations, the best of which were captured for initial clonal testing. We examine the

following questions:

New Forests (2012) 43:13–29 15

123

• Are there significant differences among the acacia hybrid clones under test, for the

selection traits of height, DBH and wood density as indicted by PP?

• How reliable is PP for predicting wood basic density at breast height?

• Are there strong phenotypic correlations among different selection traits at the clone-

mean level, for example between DBH and PP?

• To what extent do the rankings of the tested acacia hybrid clones for DBH and PP vary

across different regions of Vietnam, indicating clone-by-environment interaction?

• How do the acacia hybrid clones perform, compared to the pure-species controls?

Implications of the results obtained from the trials are considered for ongoing breeding,

clonal selection and deployment strategies for acacia hybrid clonal forestry in Vietnam.

Materials and methods

Trial locations, and clones and seedlots tested

The three test sites were located in lowland regions of northern and southern Vietnam

(Table 1). They represented site types widely available for acacia hybrid plantations in

Vietnam. The two trial sites in northern Vietnam were on sloping land with relatively

shallow soils, with some site degradation through loss of topsoil during previous land use.

The site in the south represents a more productive soil type, and a climate more conducive

to year-round growth of tropical acacias.

The trials at BaVi and Yen Thanh used randomised complete block designs with three

replicates, and ten-tree line treatment plots, while the trial at Long Thanh used three

Table 1 Details of the three trial sites

Ba Vi Yen Thanh Long Thanh

Latitude (N) 21�070 18�590 10�470

Longitude (E) 105�260 105�270 106�590

Altitude (m) 60 44 35

Soil Ferralitic clay loamwith heavy laterization

Ferralitic clay loam Sandy alluvium

Mean annualrainfall (mm)

1,680 1,620 2,070

Mean annualtemperature (�C)

23.2 24.3 26.1

Site preparation Machine-ripped Hand cultivated Ploughed

Planting time Sept 2000 Nov 2000 July 2005

Fertilizer (per tree) 3 kg cattle manure? 0.3 kg superphosphateper tree

3 kg cattle manure ?0.3 kg superphosphate? 0.2 kg NPK per tree

500 g organicbio-fertilizer ?100 g NPK per tree

Design

No. of replicates 3 3 3

No. of trees per plot 10 10 49

Spacing (m) 3 9 2 4 9 2.5 4 9 2.5

No. of clones tested 23 23 12

16 New Forests (2012) 43:13–29

123

replicates of 49-tree (7 9 7) square plots. Most of the replicate boundaries at the three sites

were surrounded by external perimeter rows of acacia hybrid clones or other acacia species

at the trial spacing, although this was not universally the case. Planting pits, of size

40 9 40 9 40 cm, were cultivated by hand, and fertilizer was placed in the partly-filled pit

which was then back-filling with soil prior to planting. Fertilizer prescriptions varied

among the trials, following local operational practice. Trials were hand-weeded twice per

year for at least the first 2 years.

A total of 27 acacia hybrid clones were tested (Table 2). These clones had been selected

by FSIV as outstanding in initial clone screening trials. All of the clones were identified as

natural hybrids within plantations or nursery seedlots raised from A. mangium seed sources

and are most probably the offspring of A. mangium mothers and A. auriculiformis fathers.

At Ba Vi and Yen Thanh, 23 clones were tested, 22 of these being in common to the two

sites, while at Long Thanh only twelve clones were tested, three of these being tested only

at this site. Clones named BV were selected from A. mangium plantations at Ba Vi in

northern Vietnam whereas clone series TB and T were selected from A. mangium plan-

tations at Dong Nai and Binh Duong provinces in southern Vietnam.

For controls, one pure-species seedlot of each of A. mangium and A. auriculiformis was

included at Ba Vi and Yen Thanh, while three different seedlots of A. mangium were

included at Long Thanh. The A. auriculiformis control comprised a bulk seedlot from the

Coen River, Queensland provenance, which has proved to be an outstanding A. auriculi-formis provenance in trials in Vietnam (Hai et al. 2008b). Of the A. mangium control

seedlots, two were from unpedigreed selectively thinned Seed Production Areas (SPAs) in

Vietnam. The Pongaki SPA at Ba Vi, northern Vietnam was developed from a 2 ha stand

planted with the Pongaki (Papua New Guinea, PNG) provenance of A. mangium, selec-

tively thinned during stand development from initial stand density of about 1,000 stem-

s ha-1 to retain about 200 superior trees per hectare. The Dong Ha SPA in central Vietnam

was developed in a similar way from a 2 ha A. mangium stand established using seed from

many superior provenances from PNG and far north Queensland, together with seed

from first generation seedling seed orchards of these provenances developed in Australia.

The Dong Nai seedlot of A. mangium was a bulk collection from unimproved plantations of

A. mangium of unknown provenance origin from Dong Nai province, southern Vietnam.

Table 2 Acacia hybrid clones and seedling controls tested at the three trial sites

Location Clones and seedlots tested

Ba Vi AauriCRa, AmanPb, BV5, BV10, BV16, BV29, BV32, BV33, BV71, BV72, BV74,BV75, CQ58, CQ62, T1, T3, T4, T5, T6, T7, T8, TB12, TB3, TB6, TB39

Yen Thanh AauriCRa, AmanPb, BV5, BV10, BV16, BV29, BV32, BV33, BV71, BV72, BV73,BV74, BV75, CQ58, T1, T3, T4, T5, T6, T7, T8, TB12, TB3, TB6, TB9

Long Thanh AmanPb, AmanDNc, AmanDHd, BV10, BV15, BV16, BV33, BV71, BV72, BV73,BV74, BV75

a A. auriculiformis Coen River provenanceb A. mangium Pongaki seed production areac A. mangium Dong Nai land raced A. mangium Dong Ha seed production area

New Forests (2012) 43:13–29 17

123

Measurement and assessment

Tree height (HT) was measured to the nearest 0.1 m using height poles at ages 3 and

4 years, and estimated using a Vertex hypsometer at later ages. Diameter at breast height

(1.3 m) over bark (DBH) was measured to the nearest mm using diameter tapes. Pilodyn

penetration (PP) was assessed at all three sites. Approximately 20, 15 and 34 trees per

treatment (i.e. clone or seedlot) were assessed for PP at Ba Vi, Yen Thanh and Long

Thanh, respectively. After removal of a bark window, two pilodyn shots, 90o to one

another, were taken from each tree at breast height. The mean value for PP for each tree

was calculated for subsequent analysis. A number of trees (three trees per treatment at Yen

Thanh, and six trees per treatment at Long Thanh) that had been sampled for PP were

subsequently felled and breast-height wood disks were obtained for basic density deter-

mination using the standard method OM-258 of TAPPI (1998).

At Long Thanh, stem volumes over bark of 89 individual trees (17 A. mangium trees and

72 acacia hybrid trees) at age 5 years were calculated by summing frustum volumes

determined from diameter measurements made every 2 m up the stem to small-end

diameter of 6 cm. Individual tree volume was predicted from the following regression:

VOL ¼ 0:0013� DBH1:8953

where VOL is stem volume over bark to small-end diameter of 6 cm in m3, and DBH is

diameter at breast height over bark in cm. The regression accounted for 82% of the

variance in stem volume. Volume of each tree surviving to age 5 years was predicted using

the regression, and plot volumes were calculated by summing individual tree volumes of

all surviving trees in the plot, and expressed on a per-hectare basis.

Statistical analyses

Linear model (1) was used to analyse the data sets at the individual-tree level, in order to

test the significance of treatment effects and estimate the performance of treatments.

yijk ¼ lþ REPi þ TREATMENTj þ REP:PLOTij þ eijk ð1Þ

where yijk is the observed phenotypic measurement, l is the overall mean, REPi is the fixed

effect of replicate, TREATMENTj is the fixed effect of clone or seedlot treatment,

REP:PLOTij is the random plot effect (10-tree row plots at Ba Vi and Yen Thanh; and

7 9 7 square plots at Long Thanh), and eijk is the residual error. The significance of fixed

effects was tested with Wald-type F-statistics (Kenward and Roger 1997) in ASReml

(Gilmour et al. 2006).

The trait of VOL (wood volume per hectare) was analysed at the plot level using a

simpler fixed-effects model (2):

yijk ¼ lþ REPi þ TREATMENTj þ eij ð2Þ

where yij is the observed plot volume, l is the overall mean, REPi is the fixed effect of

replicate, TREATMENTj is the fixed effect of clone or seedlot treatment and eij is the

residual error.

For traits where the treatment effect was significant, Best Linear Unbiased Estimators

(BLUE) were estimated for each clone and seedling control: these are equivalent to least-

squares means. This analysis was then repeated with the pure-species control treatments

excluded, to test the significance of differences among the hybrid clones.

18 New Forests (2012) 43:13–29

123

Pearson correlations among the BLUE clone estimates were calculated for each paired-

trait comparison within each trial as an approximate indication of the genotypic correlation

between traits (numbers of clones tested were too low for genotypic correlations to be

realistically estimated). Additionally, across-trial Pearson correlations were estimated in

the same way to indicate the between-site genotypic correlations for each trait. The data for

the pure-species seedling control treatments were excluded when these Pearson correla-

tions were estimated, because in most cases pure-species values differed substantially from

those of the hybrids and would have dominated the correlation estimates, whereas interest

in the correlations centred on the hybrid clones.

Clonal mean repeatabilities (H2�C) for the assessed traits at each site were estimated from

a mixed effects model in which clone and plot within replicate were treated as random and

replicate as fixed.

H2�C ¼

r2clone

r2clone þ

r2rep:plot

r þ r2e

rt

;

where r = the number or replications and t = the harmonic mean number of ramets per

clone per replicate. Pure-species control seedlots were excluded from these calculations of

clonal repeatability.

Mean values for wood basic density were calculated for each treatment at the Yen

Thanh and Long Thanh sites, and the relationship between breast-height disk wood basic

density and PP was examined by calculating the linear regressions of density on PP, at the

treatment (clone or seedlot) mean level.

Results

Survival and growth

Survival was good in all three trials, with 4-year site mean survival of 82.7% and 84.9% at

Ba Vi and Yen Thanh, respectively, and 3-year site mean survival of 88.3% at Long Thanh

(Table 3). Survival had declined to 68.7% at Ba Vi at age 9 years, 77.5% at Yen Thanh at

8 years, and 83.5% at Long Thanh at 5 years. The decline in survival at the first two sites

was partly due to line plots of A. auriculiformis and A. mangium and some of the slowest-

growing hybrid clones being over-topped by adjacent rows of faster-growing hybrid

clones, leading to suppressed growth and poorer survival.

There were significant (P \ 0.001) differences among treatments for HT and DBH at all

three trial sites except for the 3-year measure of these traits at Long Thanh (Table 4). Eight

acacia hybrid clones and one control seedlot of A. mangium were tested at all three trial

sites (Table 2). Early height growth of these eight acacia hybrid clones was fastest at Long

Thanh, averaging 4.7 m per year up to age 3 years, and slowest at Ba Vi, where annual

height increment up to age 4 years of these clones was only 3.0 m, while the annual

increment at Yen Thanh was 3.5 m up to age 4 years. The Pongaki SPA seedlot of

A. mangium had slightly slower early height growth at all three sites, relative to the mean

of these eight hybrid clones. This difference was most pronounced at Ba Vi, where

A. mangium was 26% below the average of the clones at age 4 years, and least pronounced

at Long Thanh where it was only 9% below the average at age 3 years. At the two northern

sites, A. auriculiformis was significantly inferior in early height growth to the acacia hybrid

clones and to A. mangium (Fig. 1). Treatments differed significantly (P \ 0.01) in their

New Forests (2012) 43:13–29 19

123

per-hectare volume at age 5 years at Long Thanh (Table 4). Estimated volume mean

annual increment (MAI) for the acacia hybrid clones at age 5 years averaged 34 m3 ha-1,

with clone BV10 having the highest MAI at 39 m3 ha-1, and clone BV75 the lowest at

30 m3 ha-1 (Table 5). Corresponding MAI estimates for A. mangium seedlots were 31, 38

and 44 m3 ha-1 for the Dong Nai, Pongaki and Dong Ha seed sources respectively.

Clonal mean repeatabilities were 0.59 or higher for HT at 4 years and DBH at age

4 years at the two northern sites and for DBH at Long Thanh at age 5 years (Table 6),

although they were low and non-significant for HT and DBH at 3 years at Long Thanh.

Age-age Pearson correlations for DBH (Table 7) were high at all three sites; 0.96 at Ba Vi,

0.92 at Yen Thanh and 0.88 at Long Thanh, indicating rankings of clones obtained at the

earlier ages (4, 4 and 3 years) were reliable indicators of rankings at the later ages of 9, 8

and 5 years, respectively. Age-age correlations for HT at Ba Vi and Yen Thanh were

similarly high.

Pilodyn penetration and wood basic density

There were significant differences in pilodyn penetration among acacia hybrid clones at all

three sites (P \ 0.001 at Ba Vi and Yen Thanh, P \ 0.01 at Long Thanh, Table 4), and

clonal mean repeatabilities for PP were high at all three sites, ranging from 0.70 to 0.79

(Table 6). Pearson correlations between DBH at age 4 or 5 years and PP ranged from

-0.09 at Yen Thanh to 0.21 at Ba Vi, indicating that clonal values for DBH at age 4 or

5 years (the most reliable indicator of growth) and wood basic density as indicated by PP

were not closely related (Table 7).

Table 3 Mean values for survival, height (HT) and DBH, and pilodyn penetration at the three trial sites

Age Ba Vi Yen Thanh Long Thanh

Survival (%) 3 – – 88.3

4 82.7 84.9

5 – – 83.5

8 – 77.5 –

9 68.7 – –

HT (m) 3 – – 13.3

4 10.7 13.5 –

8 – 18.2 –

9 14.6 – –

DBH (cm) 3 – – 10.8

4 9.2 10.1 –

5 – – 14.2

8 – 15.2 –

9 14.8 – –

Pilodyn (mm) 5 – – 12.0

8 – 9.8 –

9 8.8 – –

Basic density(kg m-3) 5 – – 474

8 539 –

20 New Forests (2012) 43:13–29

123

Wood basic density at breast height across all treatments averaged 539 kg m-3 at

8 years at Yen Thanh, and 474 kg m-3 at 5 years at Long Thanh (Table 3). Only three

trees of A. mangium were assessed for density at Yen Thanh, so the density of this species

could not be realistically estimated there. The linear regressions of treatment (clone or

seedlot) mean basic wood density at breast height on treatment mean PP at both Yen Thanh

and Long Thanh were significant (P \ 0.001), but accounted for only 60 and 36%,

respectively of the variance in treatment density (Fig. 2).

Table 4 Significance of differ-ences among acacia hybrid clonetreatments (pure species controlsexcluded)

n.s., not significant

* 0.01 \ P \ 0.05;** 0.001 \ P \ 0.01;*** P \ 0.001

Trait Treatment differences

(a) Ba Vi, ages 4 and 9 years

DBH-4 (cm) ***

HT-4 (m) ***

DBH-9 (cm) ***

Height-9 (m) ***

Pilodyn-9 ***

(b) Yen Thanh, ages 4 and 8 years

DBH-4 (cm) ***

HT-4 (m) ***

DBH-8 (cm) ***

Height-8 (m) ***

Pilodyn-8 ***

(c) Long Thanh, ages 3 and 5 years

DBH-3 (cm) n.s.

HT-3 (m) n.s.

DBH-5 (cm) ***

Vol per ha (m3) **

Pilodyn-5 **

0

1

2

3

4

5

6

Ba Vi Yen Thanh Long Thanh

An

nu

al i

ncr

emen

t (m

)

hybrid clones

mangium

auriculiformis

Fig. 1 Mean annual height increment for the first 4 years (at Ba Vi and Yen Thanh) or the first 3 years (atLong Thanh) for eight acacia hybrid clones in common to the three trials, A. mangium (Pongaki SPA) andA. auriculiformis (Coen River provenance). Bars above A. mangium show critical difference (P = 0.05)between height increments of individual treatments in each trial

New Forests (2012) 43:13–29 21

123

Performance and ranking of clones and controls at different trial sites across Vietnam

At the two northern sites, Ba Vi and Yen Thanh, the pure species controls were clearly

inferior to all or most of the acacia hybrid clones for four-year DBH (Fig. 3). However, at

Long Thanh the A. mangium seedlot from the Dong Ha SPA displayed the largest DBH

Table 5 Estimates of meanannual increment of wood vol-ume over bark to small end stemdiameter of 6 cm at Long Thanh,at age 5 years

Taxon Seed source or clone Mean annual volumeincrement (m3 ha-1)

A. mangium Dong Nai 31

SPA Pongaki 38

SPA Dong Ha 44

Acacia hybrid BV10 39

BV15 28

BV16 35

BV33 33

BV71 34

BV72 31

BV73 34

BV74 34

BV75 30

TB11 33

TB12 34

TB15 36

Standard error ofdifference of means

3

Table 6 Clonal mean

repeatabilities (H2�C)

Trait H2�C

(a) Ba Vi ages 4 and 9 years

DBH-4 0.95 ± 0.02

HT-4 0.96 ± 0.02

Pilodyn-9 0.79 ± 0.07

(b) Yen Thanh ages 4 and 8 years

DBH-4 0.59 ± 0.15

HT-4 0.63 ± 0.14

Pilodyn-8 0.78 ± 0.08

(c) Long Thanh ages 3 and 5 years

DBH-3 0.07 ± 0.47

HT-3 –

DBH-5 0.59 ± 0.20

Pilodyn-5 0.70 ± 0.15

22 New Forests (2012) 43:13–29

123

(Fig. 3) and wood volume per hectare (Table 5) at age 5 years, exceeding all the hybrid

clones, and had low PP, indicating relatively high basic density (Fig. 3). At this site, the

Dong Ha SPA seedlot of A. mangium had significantly greater DBH than did the Dong Nai

and Pongaki SPA seedlots of this species (Fig. 3).

Clones BV10, BV 32, BV 33 and BV75 had DBH above the site mean for acacia hybrid

clones, and PP below the site mean (indicating higher than average basic density) at both

Ba Vi and Yen Thanh. These clones were selected as superior clones, indicated by solid

squares in Fig. 3. Three of these four clones (BV10, BV33 and BV75) were also tested at

Long Thanh (Fig. 3c), where all three were below-average for PP, but clone BV75 was

below-average for DBH and wood volume per hectare. Thus, clones BV10 and BV33 rank

among the fastest-growing acacia hybrid clones at all three trial sites and appear to have

average or above-average wood density.

Site–site correlations in PP rankings for the acacia hybrid clones (r = 0.71–0.78,

Table 8) were relatively strong, and clearly better than the correlations between sites in

4–5 year DBH (-0.54 to 0.33) and 3–4 year HT (-0.44 to 0.63). The site–site correlations

between Ba Vi and Long Thanh are poorly estimated, being based on only seven clones in

common to the two trials, similarly the correlations between Yen Thanh and Long Thanh is

based on only eight clones in common. At Ba Vi and Yen Thanh, the A. auriculiformiscontrol seedlot had relatively large PP, compared to most acacia hybrid clones, indicating

lower wood basic density (Fig. 3).

Table 7 Pearson correlations between predicted clonal values for different traits of acacia hybrid clonesat each site

D4 H4 D9 H9

(a) Ba Vi

H4 0.96

D9 0.96 0.86

H9 0.97 0.94 0.94

P9 0.21 0.17 0.20 0.16

D4 H4 D8 H8

(b) Yen Thanh

H4 0.82

D8 0.92 0.56

H8 0.72 0.75 0.71

P8 -0.09 -0.02 0.12 -0.11

D3 H3 D5

(d) Long Thanh

H3 0.40

D5 0.88 0.49

P5 0.17 -0.31 0.08

D#, DBH at measurement age; H#, height at measurement age; P#, pilodyn penetration at measurement age

New Forests (2012) 43:13–29 23

123

Discussion

Growth

The trend of faster early growth of both acacia hybrid and A. mangium with decreasing

latitude (from 21�070 at Ba Vi to about 10�470 at Long Thanh) is commonly observed in

trials and operational plantations of A. auriculiformis, A. mangium and acacia hybrid in

Vietnam and is consistent with the same trend for A. mangium evident in provenance trials

in many tropical countries (Harwood and Williams 1991). While the volume estimates for

Long Thanh are approximate only, being developed from a DBH-volume regression that

accounted for only 81% of variance in stem volume of 89 sampled trees, they confirm the

high wood production potential of acacia hybrid clones and genetically improved seedlots

of A. mangium on productive sites in southern Vietnam, as previously reported for acacia

hybrid by Kha et al. (2005) and Kha (2009). The higher rates of fertilizer application used

at the two northern sites were clearly less important in determining stand growth rates than

were the inherent site differences in soil and climate.

R2 = 0.60

400

450

500

550

600

7 8 9 10 11 12 13

Pilodyn penetration (mm)

Bas

ic d

ensi

ty (

kg/c

u m

)

(a) Yen Thanh

R2 = 0.36

420

440

460

480

500

520

9 10 11 12 13 14

pilodyn penetration (mm)

bas

ic d

ensi

ty (

kg/c

u m

)

(b) Long Thanh

Fig. 2 Pilodyn penetration as a predictor of wood basic density of clone and seedlot treatments at a YenThanh and b Long Thanh. The percentage of variance explained by the linear regression of treatment meanbasic density on treatment mean pilodyn penetration is shown for each site

24 New Forests (2012) 43:13–29

123

7

8

9

10

11

12

4 6 8 10 12

DBH at 4 years (cm)

PP

(m

m)

at 9

yea

rs hybrid

hybrid (bestclones)

auriculiformis

mangium

clone BV10

(a) Ba Vi

7

8

9

10

11

12

5 10 15

DBH at 4 years (cm)

PP

(m

m)

at 8

yea

rs

hybrid

hybrid (best clones)

auriculiformis

mangium

Clone BV10

(b) Yen Thanh

10

11

12

13

14

12 13 14 15 16 17 18

DBH at 5 years (cm)

PP

(m

m)

at 5

yea

rs

hybrid

hybrid (bestclones)mangium

A. mangium Dong Ha SPA

clone BV10

A. mangiumDong Nai

(c) Long Thanh

Fig. 3 Treatment means forDBH and pilodyn penetration(PP) at the three trial sites.Horizontal and vertical bars atlower left hand side of graphsshow critical differences(P = 0.05) between treatmentsfor DBH and PP, respectively.Hybrid clones shown in bold areclones with above-average DBHand below average PP at both BaVi and Yen Thanh; only three ofthese four clones were tested atLong Thanh

New Forests (2012) 43:13–29 25

123

The Dong Ha SPA seed source had the largest DBH of the three A. mangium seedlots tested

at Long Thanh. This was expected given that it was developed from a genetic base comprising

superior seed orchard progenies and the best natural provenances, and had been selectively

thinned, retaining the best 25% of trees for seed production. This seedlot displayed the largest

DBH of all treatments at Long Thanh, and there was no clear DBH advantage of the acacia

hybrid clones over the other two A. mangium seed sources tested there. Stem form and bark

thickness, as well as DBH, affect under-bark wood volume, and there may be systematic

differences in these traits between A. mangium and the acacia hybrid, the latter tending to have

thinner bark, so the slight growth advantage of genetically improved A. mangium over acacia

hybrid at Long Thanh suggested by its larger DBH should be treated with caution. None-

theless, the result at Long Thanh suggests that the acacia hybrid clones under test do not

outgrow the best seed sources of A. mangium in southern Vietnam.

The acacia hybrid clones clearly out-performed the Pongaki SPA seed source of A.mangium and the Coen River natural provenance of A. auriculiformis at the two northern

sites. The majority of the clones had four-year HT and DBH that exceeded that of the pure

species by more than the critical difference for significance (P = 0.05). This result concurs

with the findings of earlier trials comparing the first hybrid clones and pure-species controls in

northern Vietnam reported by Kha (2001). Because competition effects between adjacent line

plots become more important with increasing age, treatment rankings for HT, DBH and

survival at the 8- and 9-year measures are considered to be less reliable than those obtained

from the 4-year measures at the Ba Vi and Yen Thanh trials, particularly at Ba Vi where initial

spacing was closer at 3 9 2 m than in the other trials which used 4 9 2.5 m. Inter-plot

competition effects tend to accentuate early-age treatment rankings for growth (White et al.

2007), so age-age correlations for HT and DBH must also be interpreted with caution.

Wood basic density

At Long Thanh, mean density of seventeen A. mangium trees sampled was slightly lower

than the average of 72 acacia hybrid trees tested there (461, compared to 476 kg m-3).

Table 8 Across-trial correlations indicating stability of clonal performance across sites for diameter atbreast height (DBH) height (HT) and pilodyn penetration (PP), with measurement age in years shown foreach trait

Ba Vi (DBH4) Yen Thanh (DBH4)

(a) DBH

Yen Thanh (DBH4) 0.33

Long Thanh (DBH5) -0.54 0.20

Ba Vi (HT4) Yen Thanh (HT4)

(b) HT

Yen Thanh (HT4) 0.63

Long Thanh (HT3) -0.44 0.28

Ba Vi (PP9) Yen Thanh (PP8)

(e) Pilodyn

Yen Thanh (PP8) 0.71

Long Thanh (PP5) 0.78 0.76

26 New Forests (2012) 43:13–29

123

The lower density at Long Thanh, compared to Yen Thanh, may result in part from

environmental differences between the two sites, with much faster growth at Long

Thanh, but another contributing factor is the tendency for density to increase with age.

Wood basic density of acacia hybrid at Ba Vi was shown to increase with age up to

8 years by Kim et al. (2008), and in A. mangium from 6 to 12 years in Sumatra,

Indonesia (Siagian et al. 1999). Increase in density with age for A. auriculiformis can

also be inferred from results of a study in a 5�-year-old clonal trial of this species in

southern Vietnam (Hai et al. 2009). The mean breast-height density of 40 A. auriculi-formis clones was 520 kg m-3 for heartwood and 560 kg m-3 for sapwood, laid down at

a later stage in the growth of the trees. The overall density at breast height of

540 kg m-3 for 5�-year-old A. auriculiformis clones reported by Hai et al. (2009) was

substantially higher than the breast-height mean for twelve acacia hybrid clones at

practically the same age reported here, 476 kg m-3.

Pilodyn penetration for the A. auriculiformis control seedlot was above the average for

acacia hybrid clones at Ba Vi and Yen Thanh (Fig. 3), suggesting lower wood basic

density. This unexpected result may be due in part to the suppression of diameter growth of

A. auriculiformis in the last four years of stand growth at Ba Vi and Yen Thanh (annual

radial increment from age 4 years onwards being only 0.2–0.3 cm). The pilodyn pin

(10-11 mm mean PP at age 8 or 9 years) would have penetrated a higher proportion of

lower-density A. auriculiformis wood laid down at an earlier age, relative to that of hybrid

clones, which displayed much faster radial growth from year 4 onwards. Wood basic

density of eight-year-old trees of six of the acacia hybrid clones, growing in a separate trial

at Ba Vi, increased from pith to bark by as much as 200 kg m-3 from the centre of the stem

to the outermost wood (Kim et al. 2008).

Treatment means for basic density used in the density-PP regressions were not pre-

cisely estimated, with only 3 trees per treatment sampled at Yen Thanh and 6 trees per

treatment at Long Thanh. This imprecision may have contributed to the substantial

unexplained variance in the pilodyn—basic density regressions evident in Fig. 2. More

systematic studies are needed to determine the number of ramets per clone, using either

direct measurement of density on wood disks or cores, or a regression of density on PP,

that must be sampled, and the relationship between breast-height and whole-tree basic

density, to attain a desired level of precision in the estimation of clonal wood basic

density (Raymond and MacDonald 1998; Kim et al. 2008). Nevertheless, the regressions

demonstrate that PP provides an approximate guide to the ranking of acacia hybrid

clones for wood basic density, with greater mean PP generally indicating clones of lower

basic density.

Implications for breeding and clonal deployment

The relatively good DBH growth of almost all of the acacia hybrid clones under test is

evident from Fig. 3. Hybrid clone means for DBH at 4 or 5 years span a range of only

2–3 cm at each site, with the exception of two clones (CQ58 and CQ62) which grew

very slowly at Ba Vi. Pre-screening of a much larger initial set of hybrid individuals

during candidate selection and initial screening trials conducted by FSIV has eliminated

many inferior hybrid genotypes which would have performed poorly, if they had been

included in the clone trials reported here. For example, the seven clones in the series

BV5-BV33 tested here were the best-performing clones from a total of 60 promising

two-year-old acacia hybrid individuals identified in A. mangium plantations near Ba Vi

(Kha 2001). Similarly, the five clones in the series BV71-BV75 tested here were chosen

New Forests (2012) 43:13–29 27

123

from approximately 100 acacia hybrid seedlings that were identified as hybrids in the

nursery at Ba Vi from progeny of a bulk seedlot of A. mangium, and grown in a field

trial to age 2 years, prior to selection and cloning of the most promising hybrid phe-

notypes. Wider ranges in clonal DBH performance were reported in three clonal tests of

A. auriculiformis candidate clones that had been less intensively pre-screened (Hai et al.

2008a).

The lack of any obvious relationship between DBH and pilodyn at the clone level

evident in the three trials, as indicated by the low Pearson correlations between DBH and

PP, suggest that it is quite feasible to breed and select hybrid clones displaying both rapid

growth and high wood basic density in a scaled-up breeding program. However, the

numbers of acacia hybrid clones that have been well-tested for both traits are small, and it

is possible that a negative genetic correlation might emerge in a large breeding and testing

program, as has been found for Pinus radiata D. Don. (Wu et al. 2008). Substantial genetic

variation in growth and wood basic density of A. auriculiformis has been demonstrated in

progeny trials (Hai et al. 2008b) and clone trials (Hai et al. 2008a), and significant genetic

difference in growth rate among three Vietnamese seed sources of A. mangium has been

demonstrated at Long Thanh in the present study. Controlled crossing between A. auric-uliformis and A. mangium parent trees that have high genetic rankings for both growth and

wood basic density or setting up hybridizing seed orchards that incorporate a number of

selected superior genotypes of the two parent species to produce natural F1 hybrid off-

spring via open pollination are two possible ways to effectively produce a range of new

hybrid genotypes that can be tested to identify clones with both superior vigour and basic

density.

Clones BV10 and BV33 have remained in use in Vietnam for over 15 years, since first

tested in 1992, without obvious loss of vigour, relative to the newer hybrid clones such as

BV75 that were identified 5–10 years later. Management of these first production clones in

tissue culture by regular re-capture from vigorously growing sapling ramets has prevented

any major loss of vigour through maturation. This ability to maintain clones of acacia

hybrid in a vigorous state for over a decade without detectable loss of rooting vigour and

growth potential is an important advantage for clonal forestry with acacia hybrids, enabling

thorough testing of individual clones prior to their mass deployment for a prolonged

period, and contrasts with ageing and loss of vigour observed in clones of many coniferous

tree species (Burdon and Aimers-Halliday 2006).

Inter-site correlations of clonal rankings for DBH and PP suggest that wood density

rankings of clones are likely to be more stable across contrasting sites than are growth

rankings. A similar conclusion was reached by Kien et al. (2010), who studied early growth

and wood basic density of a larger set of clones of E. camaldulensis at three sites in

southern, central and northern Vietnam. Although the number of acacia hybrid clones

tested remains small, it appears that a deployment strategy should be regionally based, with

some clones deployed to particular regions rather than across the entire country, to better

exploit clone-by-environment interaction for growth. Nevertheless, some clones such as

BV 10 and BV33 tested here display good and stable growth above-average wood density

across all three trials (Fig. 3).

Acknowledgments We thank staff of the Forest Science Institute of Vietnam, in particular the Researchcentre for Forest Tree Improvement, for assistance with trial establishment, management and assessment.The Australian Centre for International Agricultural Research supported the study. Drs Chris Beadle, RodGriffin and Garth Nikles provided helpful comments on an earlier draft of this paper.

28 New Forests (2012) 43:13–29

123

References

Amanulla BKM, Jayakumar MN, Torvi RK (2004) Growth and productivity of acacia hybrids on degradedforest lands and other wastelands in Western Ghats region of Karnataka. Ind For 130(5):537–550

Burdon RD, Aimers-Halliday J (2006) Managing risk in clonal forestry. CAB reviews: perspectives inagriculture, Veterinary Science. Nutr Nat Resour 1(35)

Dao DN (2003) Continuing evaluation of growth and soil improvement ability of acacia hybrid in com-parison with its parental species from age 5 years (in Vietnamese). MSc Thesis, Forestry University ofVietnam

Gilmour AR, Gogel BJ, Cullis BR, Welham SJ, Thompson R (2006) ASReml User Guide Release 2.0. VSNInternational Ltd, Hemel Hempstead

Greaves BL, Borralho NMG, Raymond CA (1997) Breeding objective for plantation eucalypts grown forproduction of kraft pulp. For Sci 43(4):465–472

Greaves BL, Borralho NMG, Raymond CA (2003) Early selection in eucalypt breeding in Australia—optimum selection age to minimise the total cost of kraft pulp production. New For 25(3):201–210

Hai PH, Jansson G, Harwood C, Hannrup B, Thinh HH (2008a) Genetic variation in growth, stemstraightness and branch thickness in clonal trials of Acacia auriculiformis at three contrasting sites inVietnam. For Ecol Manage 255(1):156–167

Hai PH, Jansson G, Harwood C, Hannrup B, Thinh HH, Pinyopusarerk K (2008b) Genetic variation in woodbasic density and knot index and their relationship with growth traits for Acacia auriculiformisA. Cunn ex Benth in Northern Vietnam. N Z J For Sci 38(1):176–193

Hai PH, Jansson G, Hannrup B, Harwood C, Thinh HH (2009) Use of wood shrinkage characteristics inbreeding of fast-grown Acacia auriculiformis a. Cunn. ex Benth in Vietnam. Ann For Sci 66(6):611–619

Harwood CE, Williams ER (1991) A review of provenance variation in growth of Acacia mangium.In: Carron LT, Aken KM (eds) Breeding technologies for tropical acacias. Australian Centre forInternational Agricultural Research, pp 22–30

Kenward MG, Roger JH (1997) Small sample inference for fixed effects from restricted maximum likeli-hood. Biometrics 53(3):983–997

Kha LD (2001) Studies on the use of natural hybrids between Acacia mangium and Acacia auriculiformis inVietnam. Agriculture Publishing House, Hanoi

Kha LD (2009) Development of Acacia mangium x Acacia auriculiformis clones in Vietnam. La lettre del’ATIBT, Dec. 2009. Association Technique Internationale des Bois Tropicaux, France, pp 18–21

Kha LD, Hai ND, Thinh HH, Tuan PV, Thinh LB (2005) Clonal tests of natural hybrids at Ba Vi, Yen Thanhand Bau Bang (report for registration new varieties). Forest Science Institute of Vietnam

Kien ND, Jansson G, Harwood C, Almqvist C (2010) Clonal variation and genotype by environmentinteractions in growth and wood density in Eucalyptus camaldulensis at three contrasting sites inVietnam. Silvae Genet 59(1):17–28

Kim NT, Ochiishi M, Matsumura J, Oda K (2008) Variation in wood properties of six natural acacia hybridclones in northern vietnam. J Wood Sci 54(6):436–442

Mohd. Ghazali H, Wan Rasidah K, Rosazlin A, Rozita A, Rosdi K, Ab. Rasip AG (2007) Growth of acaciahybrid (A. mangium * A. auriculiformis) on coastal sandy soil. J Trop For Sci 19 (4):243–244

Nghia NH, Kha LD (1998) Selection of Acacia species and provenances for planting in Vietnam. In:Turnbull JW, Crompton HR, Pinyopusarerk K (eds) Recent developments in acacia planting. ACIARproceedings no. 82. Australian Centre for International Agricultural Research, Canberra, pp 130–135

Raymond CA, MacDonald AC (1998) Where to shoot your pilodyn: Within tree variation in basic density inplantation Eucalyptus globulus and E. nitens in Tasmania. New For 15 (3):205–221

Siagian RM, Darmawan S, Saepuloh (1999) Chemical composition of Acacia mangium Willd at several agesharvested from first rotation growth. Buletin Penelitian Hasil Hutan 17(1):57–66

White TL, Adams WT, Neale DN (2007) Forest genetics. CABI IntWu HX, Ivkovic M, Gapare WJ, Matheson AC, Baltunis BS, Powell MB, McRae TA (2008) Breeding for

wood quality and profit in Pinus radiata: a review of genetic parameter estimates and implications forbreeding and deployment. N Z J For Sci 38:56–87

New Forests (2012) 43:13–29 29

123