Nghiên cứu khả năng tích lũy sinh khối ở hệ sinh thái rừng thứ sinh lá rộng thường xanh tại kon hà nừng, gia lai TT

1

FOREST SCIENCE INSTITUTE OF VIET NAM

TRAN HOANG QUY

RESEARCH ON BIOMASS ACCUMULATION CAPACITY IN SECONDARY EVERGREEN
BROADLEAF FOREST ECOSYSTEMS IN KON HA NUNG, GIA LAI PROVINCE

SPECIALITY: FOREST MEASUREMENT AND PLANNING
CODE: 9.62.02.08

SUMMARY OF FORESTRY DISSERTATION

HA NOI, 2020
1


2

The work had been completed at Forest Science Institute of Vienam

Scientific Suppervisor:
1. GS.TSKH. NGUYEN NGOC LUNG

Reviewer 1:
Reviewer 2:
Reviewer 3:

The dissertation had been defended at thesis assessment Council of Forest Science Institute of Vietnm,
04 Duc Thang Street, North Tu Liem district, Hanoi city.

Time …., date…. month…. year 2020

For more information of the thesis can be searched at:
Library: Forest Science Institute of Vietnam
National library of Vietnam


3

INTRODUCTION
1. The urgency of the research
The accurate estimation of annual biomass increase of tropical forests
were urgently to minimizing the inadequate of net carbon stock. It was to
be realized an inconsistence of former and current research relevant to the
estimation of biomass and absorbability and emission of carbon of forest
ecosystems. The difference may be because inadequate data base and
various methods used to estimate biomass and carbon of forests. Former
research often based on direct measurement in small sample plots leading
to higher biomass estimate. Recent research, however, was based on forest
inventory data and supply biomass data at national level or regional level.
An adequate method to estimate forest biomass was urgently needed to
minimize the inadequate in carbon monitoring. There were many
systematical and wholly researches on biomass and carbon accumulation
for planted and natural forests that had been done in Vietnam which could
be oriented for further research. However, there were some limits: (i) Only
determined biomass at the time of samples collection; (ii Used methods
might destroy the research objects (cutting trees, digging soil to collect
samples); (iii) Dead biomass accumulated annually had been not
considered; (iv) Most research overlooked fine roots production dynamics
(ø<2mm). For scientific background addition of estimating biomass of

evergreen broadleaf forest ecosystems in Kon Ha Nung- Gia Lai, it is
necessary to implement a systematically research on biomass
accumulation capacity and to enhancing degree of accuracy estimating
biomass accumulated, especially underground biomass of forest
ecosystems. Therefore, the project: “research on biomass accumulation
capacity in secondary evergreen broadleaf forest ecosystems in Kon Ha
Nung, Gia Lai province” had been chosen for PhD thesis.
2. Research objective
Overall: to supplement scientific background on biomass and carbon
estimation of evergreen broadleaf forest ecosystems in Kon Ha Nung, Gia
Lai province.
Specific:
To determine biomass accumulation capacity of evergreen broadleaf
forest ecosystems to make background for quantifying environmental
services of the forests; and
To enhance the degree of accuracy in determining belowground
biomass of evergreen broadleaf forest ecosystems.


4

3. Research object
The thesis research on two forest status of the secondary evergreen
broadleaf forest in Kon Ha Nung, Gia Lai. The two forest status chosen for
the research were: (i) Little influenced forests (RiBTĐ); (ii) Restoration
forests (RPH).
4. Research scope:
The experimental forest in Kon Ha Nung belonging to Tropical
Research Centre, Vietnamese Academy of Forest Sciences.
5. The significance of the thesis

Scientific significance: The thesis has supplemented scientific basis
on determining biomass of evergreen broadleaf forests, especially
underground biomass of evergreen broadleaf forest ecosystems in Kon Ha
Nung.
Practical significance:
The thesis has determined biomass
accumulation capacity of the evergreen broadleaf forest ecosystems in Kon
Ha Nung, Gia Lai. Besides, the thesis also contributed to enhance the
degree of accuracy for estimation of underground biomass for future
scientific works.
6. New contribution of the thesis
(1) Applying new method to determine fine roots biomass of
evergreen broadleaf forest ecosystem in Ko Ha Nung, Gia Lai
(2) Determining biomass increase of evergreen broadleaf forest in
Kon Ha Nung, Gia Lai.
7. Thesis content
The thesis had 117 pages, with 30 tables and 29 figures. There are
109 references. The content of the thesis includes:
Introduction: 4 pages
Chapter 1. Overview of the research issues: 25 pages
Chapter 2. Research contents and methods: 21 pages
Chapter 3. Results and Discussion: 63 pages
Conclusions and Recommendations: 2 pages.
CHAPTER 1. OVERWIEW OF RESEARCH ISSUES
1.1. Some basic principles
1.2. In the world
1.2.1. Research on accumulation of aboveground biomass (AGB)
1.2.2. Research on accumulation of belowground biomass (BGB
1.2.3. Research on relationships between biomass and stand
parameters

5

1.3. In Vietnam
1.3.1. Research on accumulation of aboveground biomass
1.3.2. Research on accumulation of belowground biomass
1.4. Discussion
Research results in and out Vietnam are important scientific
background to orient further research on forest biomass and carbon.
However, there are some limits of the formerly research: (i) Many
researches only determined biomass at the research time; (ii) They used
research methods destroying research objects by cutting ang digging
sample trees to have biomass, these methods gained exact data but
destroyed research objects and were very expensive, difficult to verify and
to inherit for further research like increase, structure, biomass
accumulation potentials. (iii) Biomass part accumulated belowground
were little delt, especially the estimation of fine root production in many
researches in Vietnam.
Therefore, determination of net ecosystem production (NEP) of
forest ecosystems in Vietnam, especially fine root production has to be
interested more in the next time in order to evaluate practical
environmental value of forest ecosystems for regulation CO2 in
atmosphere contributing decreasing green-house effects and climate
change.
Starting from these practical points, the project “Research on
biomass accumulation capacity of secondary evergreen broadleaf forest
ecosystems in Kon Ha Nung, Gia Lai” contributing to overcome above
limitations in biomass research in Vietnam.
CHAPTER 2. RESEARCH CONTENTS AND METHODOLOGY

2.1. Research contents
(1) Additional research on silvicultural characteristics of secondary
evergreen broadleaf forests
(2) Biomass accumulation capacity of evergreen broadleaf forest
ecosystems (AGB production capacity; BGB production capacity (coarse
Ø>2mm and fine roots Ø<2mm); and total biomass accumulated annually)
(3) Enhancing the degree of accuracy of determination method of BGB
(include coarse and fine roots).
2.2. Research methods
2.2.1. Research methods for content 1
Data collection


6

* The thesis inherited 10 permanent sample plots established in 2004 by
the project: “Research to apply advanced technological progress and
measurements to manage sustainably forests in central highland” (Tran
Van Con et al., 2006) and continued up to 2015 by the research project:
“research on silvicultural characteristics of some main natural forest
ecosystems in Vietnam (Tran Van Con et al., 2010; 2015). The thesis only
inherited data collected in year 2012 to compared with data collected in
2017 by the thesis in 10 sample plots. The 10 1-ha sample plots belonging
to two forest status: (i) Little influenced forests (RiBTĐ) which had been
harvested with lower intensity and had been very good habilitated, it
remain a structure nearby native forests, and belong to very rich forest
status according circulation 34 of MARD (Forest volume >300 m 3/ha) that
were sample plots: 2, 6, 8 and 9; (ii) Restoration forests were forests which
had been harvested with higher intensity and had good restoration and
become to the status medium or rich forest status according circulation 34

of MARD (forest volume around 200 m3/ha) that were sample plots: 1, 3,
4, 5, 7 and 10.
* The thesis additionally established 6 experimental plots of 900 m 2 each
in two forest status. The experimental layout method was according the
research project: “Applying advanced methods in evaluating biomass
accumulation potential of some main forest ecosystem in Vietnam” (done
3/2014-3/2016 by Tran Van Do et al., 2016.
Data analysis
+ Species composition is percentage of species participating the forest
association determining in1/10 of importance value of the species IV.
+ The diversity Shannon-Wiener (H’) index calculated by formula: H’=∑(Ni)(lnNi) with i=1,2,…,s.
+ Species mixed ratio calculated by S/N (in which S is the number of
species and N is the number of individuals in the sample plot)
2.2.2. Research methods for content 2
Experimental layout
In each sample plot for two forest status, experimental layout was the
same to ensure the objective accuracy of gained results. At each site
selected to collect data, establish a plot with area of 900 m 2 (30 m x 30 m).
These plots were divided into 25 subplots with an area 36 m 2 (6 m x 6 m)
like figure 2.2.
6m
6m


7

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22
23
30m

24

25

30m

6m

-

-

-

-

Figure 2.2. Experimental layout and collecting data in sample plots
(30m x 30m)

Note:
In 4 subplots (8, 12, 14, 18), fine roots were collected by soil, in one
subplot there were 18 soil cores, a total 72 soil cores were collected. 3
cores in each plot were collected periodically 4 month/time x 4 plots = 12
core/4 months/time.
In 4 subplots (7, 9, 17, 19), fine root development was determined by
burying plate box (1 box/plot) with the size (long 29,7 cm, wide 21 cm)
according to scan machine A4. Data collected by link with a computer to
scan images, periodically one month/time from 1/2016-1/2017.
Determine respiration of soil microorganism were conducted in 4 subplots
(3, 11, 15, 23). Besides, decomposition bag was used to determine
respiration of soil microorganism had been also arranged in the same four
subplots.
Data collection
For trees with diameter at breast height ≥ 6cm: using dendrometer to
measure diameter increase of trees (5 trees per species).
For the litter fall: establish 12 subplots sized 1mx1m to collect litter
falls of the forests to determine raw biomass and then bring to labor to dry
by temperature 105oC to determine dry biomass.
Method to determine root biomass:
• For coarse roots:
Three 1-m x 1-m x 1-m soil blocks (1m 3 soil volume) were used to
measure the coarse root diameter in July 2015 and July 2016.
In July 2015, the soil in each block was carefully removed using shovels
and trowels. The diameter of each root segment was measured at three


8

-

•

points: two ends on the wall of soil block and at the middle. The points for
measuring coarse root diameter were marked with red ink and numbered.
After measurements were completed, the soil was filled in to cover all the
roots. In July 2016, the soil was removed again, and coarse root diameters
were remeasured at the marked points and recorded. The procedure was
conducted carefully to minimally affect the roots and to ensure that the
coarse roots continued to grow naturally.
To establish allometric relationship between biomass and diameter and
length of root, coarse roots were excavated and measured in four 1-m × 1m × 1-m soil blocks (1-m3 soil volume). Two blocks were sampled in June
(rainy season), and two others were sampled in December (dry season) in
2016. All coarse roots were collected, cut into segments, and measured for
fresh mass (M in grams), length (L in centimeters), and diameters (Ø in
centimeters). Coarse roots with a diameter of 1 cm were cut with scissors
and the larger roots were cut with a handsaw. Fresh root segments were
sampled and transferred to the laboratory to determine dry mass by drying
to constant mass in an 80°C oven (Tran Van Do et al., 2018).
For fine roots: using CIE (Continuous Inflow Estimate) of Osawa and
Aizawa (2012) and modified by Tran Van Do et al. (2015) to estimated
fine root production.
Method to estimate decomposition death fine root:
For evaluating the decomposition ratio of dead fine roots, RIWP
sheets (root-impermeable, water-permeable) were used… The RIWP sheet
has a pore size of approximately 6 μm and blocks the ingrowth of almost
all fine roots; however, fine soil particles, rainwater, mycorrhizal hyphae,
and other microorganisms can penetrate through the sheet.
Put 1,2 - 1,7 g dry roots in the bag, each bag has own sign. Before
burying to the field all sheets were soaked in ordinary water in room

temperature for 24 hours to ensure that water content in fine roots inside
bags was the same that the field. The bags were buried at a depth of 20cm
which is the zone most fine roots distribute in the soil. 40 litter bags were
buried at 8 points of the sample plot. Litter bags were collected periodic
with the same time to collect soil cores, each time collected 5 bags to
determine decomposition ratio in each period correlatively. After
collection, the remaining fine roots inside the litter bags were separated
from soil particles by washing and sieving and then oven-dried to constant
mass for calculating the decomposition ratio.
Selection allometric equation between biomass and diameter D1,3


9

- Applied the method proposed by Kettering et al., (2001), in which
allometric equation between individual biomass and breast height diameter
had been used. Equation B = aD1,3b, in which B is biomass, D1,3 is diameter
at breast height and a and b are parameters.
- Using biomass data of 36 sample trees to determine biometric
relationship equation to estimate AGB of individual trees of evergreen
broadleaf forest.
Methods to determine respiration of soil micro-organism
Two different methods were used to compare and determine suitable
levels of each method for research respiration of soil micro-organism in
Vietnam forests.
(i) Determine respiration of soil micro-organism by using tight bunch
used gas exchange machine.
(ii) Determine respiration of micro-organism by using litter
decomposition bags
Data were collected periodic 4 months/time for living, death roots, death

biomass aboveground (Litter falls).
Data analysis
• Net primary production (NPP)
NPP had been calculated by formula 2.2:
(2.2)
In which: ΔM is AGB increase;
ΔCr: Increase of coarse roots;
Lf: Biomass of litter falls;
Fp: Biomass of fine roots.
(i) AGB increase (ΔM) was determined by the formula 2.3:
(2.3)
In which: Mi= biomass at first measurement;
Mj = biomass at the second measurement.
(ii) Biomass increase of coarse roots (ΔCr):
Diameter and length of coarse roots collected at time ti and tj (Δt;
tj ≥ ti) were used to estimate biomass of coarse roots per unit of surface
area during ∆t, assuming that soil block with surface area A (m 2) were
sampled. Allometric relationship of coarse root biomass to root diameter
and length were determined from a separate sample of the coarse roots that
were excavated, collected, cut into segments measured for length L (cm)
and diameter Ø (cm) and then oven-dried to constant mass Cr (g). The


10

allometric relationship between root segment dry mass (Cr) and Ø2·L was
established using the following equation:
Cr =a (φ2.L)b
(2.4)
In which a and b are parameters to be estimated.

Equation 2.4 can be expressed as a lineare relationship on double
logarithmic coordinated as shown in equation 2.5:
logCr = loga + blog(φ2.L)
(2.5)
The parameters a and b were estimated based on observed data for L, φ
and Cr from coarse roots in sample soil blocks. The mass of coarse roots
per unit surface area (A) can be calculated from equation 2.5 if the coarse
root diameter and length data from the soil block of A surface area are
known. Next, if the diameter and length of coarse roots in a soil block of a
known surface area (e.g., m2) are measured at times ti and tj (Δt; t j ≥ t i),
coarse root production (ΔCr) during Δt can be estimated using
equation 2.6:
ΔCr = Crj - Cri
(2.6)
where Cri and Crj are estimated from Equation 2.4 with Li (length of
coarse root at time ti), Øi (diameter of coarse root at time ti), Lj (length of
coarse root at time tj), and Øj (diameter of coarse root at time tj). Through
this allometric conversion, the mass of coarse roots (e.g., g m−2) can be
estimated at times ti and tj, respectively.
Biomass of litter falls (Lf) was determined by formula 2.7
(2.7)
In which: Vi is litter mass of each subplot to collect litter falls;
Si is plot area.
Fine root biomass (Fp) was detemined by formula 2.8:
(2.8)
In which: Bi and Bj are live fine root biomass in the same unit area at the
measured time (ti) and at the second measurement (tj); Ni and Nj are fine
root death biomass in the same unit area at the measured time ti and tj; γij
is decomposition ratio of ine roots in the time ∆t (tj-ti). Bi, Bj, Ni, Nj were
determined by soil cores as described above.

(iv) Allometric relationship establishment for coarse roots:
The root segments were classified into five groups based on diameter:
Ø = 0,2-1,0; >1,0-2,0; >2,0-3,0; >3,0-5,0 cm and >5,0 cm. Five
corresponding relationships were fit using regression to estimate
parameters a and b shown in Equation 2.5, the coefficient of determination
(R2 ), and the bias. The deviation of the predicted vs. measured masses of


11

all root segments was calculated for each relationship using the equation
[error (%) = 100 × (observed mass − predicted mass) / observed mass]
(Chave et al. 2005) [41], and this is called the error of estimation. For
validation the dataset of root segments was randomly split into model
development (80%) and validation (20%). This data-splitting and modelcalibration process was repeated 30 times. Validation statistics were
calculated for all five groups of root diameters and a combined group with
all diameters. The validation statistics included percentage bias, root mean
square percentage error (RMSPE), and mean absolute percent error
(MAPE). Validation statistics were computed for each realization of
randomly selected data and then averaged over the 30 realizations.
Percentage bias =
(2.9)
RMSPE =
(2.10)
MAPE =
(2.11)
where R is the number of realizations (30); nr is the number of root
segments per realization r; and Ori and Pri are the observed and predicted
root mass for the ith root segment in realization r, respectively. After
validating, final estimates of model parameters (a and b of Equation 1) and

their standard errors were computed using the entire, combined dataset
(Bao Huy et al. 2016).
CHAPTER 3. RESULTS AND DISCUSSION
3.1. Supplemental research on silvicultural characteristics of
secondary evergreen broadleaf forests
3.1.1. Stand structure characteristics
Natural evergreen broadleaf forests in the research region were
synthesized from 10 permanent sample plots in table 3.1 indicates that, for
trees with diameter D1,3≥10cm, the number of tree species vary from 69
to 94 species and the species mixed ratio from 1/5 to 1/8 (meaning each 5
up to 8 individuals there one species). Shannon-Wiener (H’) varied little
(3,75-4,03) between sample plots telling vegetation structure in the
research region was homogeneous. Number of dominant species (IV≥5%)
take part in species composition formula varies from 2-9 species. Species
composition structure deals with the combination and the participation
levels of all plant components in the vegetational association. Species
composition is an important indicator telling number of species and
percentages of one species or a group of any species in forest stand.


12

Species composition is also an indicator for biodiversity, stability and
sustainability of the forest ecosystems.
Species composition structure changed according to time and space.
On the same ecological region, although there are homogenous conditions
of climate, weather, soils... however at a determinized research site, there
are differences in species composition due to different in topography or
history of the establishment and development ò the forests. Therefore, in
overall area, the number of dominant species had been changed. This study

analyzed species composition by 1-ha sample plots to have a general view
over species composition by sample plots in particular and in evergreen
broadleaf forests in general.
Generalization about characters of species composition structure in
each sample plots in research area expressing in measurement in year 2012
and the measurement after 5 years (2017). Species composition formula
was written with important index IV≥0,5, species with IV< 0,5 were added
to other species group. In the 2012-measurement was shown that in 1-ha
sample plot species number with D1,3 ≥10 cm varied from 69 to 92 species,
in which only 1-5 species had IV≥0,5 and were written in composition
formula. Data of species number had shown that there was changed in
species number in 2017- measurement compared with 2012-measurement
proving changing in species number and basal area of each species.
Forests in the research region were natural forests which were little
influenced and had a stable structure, however, there are some changes due
to effect of dynamic process of increment, mortality and diameter class
variation.
3.1.2. Stand diameter increase
Species with sample number more than 100 individuals which have
two times measured and can be calculated diameter increase, these
diameter increases varied from 0,202 cm/year (Kháo hoa thưa) 603
cm/year (Giổi nhung); maximal observed diameter varied from 22 cm
(Kháo hoa thưa) from 140 cm (Giổi nhung). Average diameter increase of
all measured species have different ages neutralizing different growth
stages with slow or fast growing rate in large sample number conditions.
3.2. Biomass accumulation capacity of evergreen broadleaf forests
3.2.1. AGB accumulation capacity
3.2.1.1. Selection of allometric model for estimating AGB
To estimate AGB, many researchers have been used allometric
equations between biomass of individual tree and tree measured

13

undestroyed sample trees. Most used allometric equation was power
equation in form B = aD1,3b which B is biomass, D1,3 is diameter at breast
height and a and b were estimated parameters varying dependent on sites.
This variation was the source of main estimated error by using relationship
equations undistinguished particular sites. However. Gathering B and D 1,3
for each particular site is impossible due to destroying research objects.
Method to select parameter a and b undestroying research objects had been
proposed by Kettering et al., (2001). Parameter a and b can be estimated
by relationship between H and D1,3 specifical for each site: H = kD 1,3c and
so b = c + 2. Parameter a can be estimated from average wood density
(WD) of the site with a = WD*r in which r is a relatively stable coefficient.
From dataset of 36 sample trees in Kon Ha Nung.
The thesis had been determined the relationship between height and
diameter, that was: H = 2,732*D0,57 (R=0,992). So that b = 2+c = 2 + 0,57
= 2,57. Parameter a=WD*r. The relationship equation will became B =
r*WD*D1,3c+2 (Quirine M Ketterings et al., 2001) [67]. Using data
collected in Kon Ha Nung there are r = 0,151 and average wood density is
0,5, so that, parameter a of the equation is a = 0,151*0,5 = 0,0755. The
allometric equation to be finded is:
B = 0,0755*D1,32,57 (R = 0,995)
The thesis had used data collected from 36 sample trees to validate 8
relationship equations often used in order to find one which was most
suitable (which had minimal error).
The absolute value of minimal error is 8,4% and the maximal is 79,8% and
the average error is -36,8%. Equation 2 had 7 positive error and 29
negative error, average error is -8,9%, minimal absolute error is 0,1% and

maximal absolute error is 43,8%. Equation 3 was similar equation 1, there
is no positive error had systematical errors and estimated higher biomass.
Equation 4 had 3 positive error and 33 negative error estimated higher
biomass with average was 53,5%, absolute value of minimal error is
17,5% and maximal is 111,7%. Equation 5 had average error of -5,8%,
minimal absolute error is 0,5% and maximal is 16,3%. Equation 6 had
maximal average error (-105%), minimal absolute value was 31% and
maximal was 163,6%. Equation 7 had average error of -23,3%, the value
of absolute minimal error was 1% and maximal was 52,4%. Equation 8
had positive and negative error equal (17 and 19) average error is -0,8%,
minimal absolute value of error is 0,1 and maximal is 31,2.


14

So that, equation 8: B=0,0755*D2,57 had been chosen to estimate
AGB.
3.2.1.2. AGB estimated from 10 permanent sample plots
a) Living biomass
Calculation results from 4 permanent sample plots of minorinfluenced forests were accumulated in table 3.7a.
Table 3.7. AGB and AGB increase in 10 permanent sample plots
Table 3.7a. AGB and AGB increment of 4 permanent samle plots in minorinfluenced forests
N (trees/ha)

Dtb (cm)

2

G (m /ha)

M (tons/ha)

Plot
2012

2017

2012

2

558

550

6

432

8

2012

2017

25.6

26.7 40.86

465

26.1

26.1

529

541

24.8

9

557

563

23.3

TB

519

STD

2012

2017

43.4

354.73

374.7

3.99

33.6 36.87

301.76

334.7

6.59

25.5 39.18 42.87

363.71 385.22

4.30

24.4

34.7 38.55

301.34 322.27

4.19

530 24.95 25.66 38.25 40.42

330.39 366.72

4.77

59.54 44.10

1.22

2017

ΔM
(t/ha/n
)

0.98

3.18

3.21

33.50

33.81

1.22

Table 3.7a showed that. AGB of permanent sample plots of minorinfluenced forests varied from 301.34 to 363.71 tons/ha. averaging 330.39
± 33.50 tons/ha in 2012 and increased to 322.27 - 385.22 tons/ha.
averaging 366.72 ± 33.81 tons/ha in 2017; thus AGB increment of minorinfluenced forests varied from 3.99 to 6.59 tons/ha/year. averaging 4.77 ±

1.22 tons/ha/year.
Table 3.7b. Dry BGB and BGB increment of 6 permanent sample plots in
restoration forests


15

N (trees/ha)

Dtb (cm)

2

G (m /ha)

M (tons/ha)

Plot
2012

2017

2012

2017

2012

2017

1

397

410

24.5

24.7

24.7

25.5

196.54 197.79

0.25

3

434

453

24.2

25.0 26.28

29.4

202.46 238.85

7.28

4

483

510

22.5

23.1 25.89 28.68

198.3 241.32

8.60

5

648

640

21.6

22.6

241.66 274.68

6.60

7

626

643

23.1

23.9 34.41

38.3

259.87 296.68

7.36

10

616

614

22.5

23.2 32.83

35.7

254.55 288.94

6.88

TB

534

545 23.07 23.75 29.37

32

225.56 256.38

6.16

STD

109 101.3

1.11

0.95

32.11 34.42

4.20

4.88

2012

ΔM
(t/ha/n
)

29.65

2017

37.38

2.98

Calculation results from 6 permanent sample plots of restoration
forests were systhezed in table 3.7b shown that AGB varied from 196.54
to 259.97 tons/ha in year 2012. meet an average of 225.56±29.65 tons/ha
increasing from 197.79 to 296.68 tons/ha with average of 256.38±37.38
tons/ha in year 2017. so that AGB-icrease of restoration forests varied
from 0.25 to 8.60 tons/ha/year. with an average of 6.16±2.98 tons/ha/year.
Biomass increase in restoration forests is higher than in little influenced
forests but varied in a wider range. These were agreed with biological
rules in forests.
b) Biomass of litter falls
Table 3.8 were litter falls (VRR) had been synthezed from 4
permanent sample plots of status little influenced forest and 6 permanent
sample plots of status of restoration forest. From this table was shown that
litter fall bimass in little influenced forest varied from 6.40 to 7.74
tons/ha/year. meet an average of 7.28±0.40 tons/ha/year; For restoration
forest. Litter fall biomass varied from 6.39 to 12.19 tons/ha/year. averaged

of 9.19±2.34 tons/ha/year. higher than that in little influenced forest but
had wide variation saying that the restoration had been not yet stable as the
little influenced forests.


16

Table 3.8. Litter fall bioomass in evegreen broadleaf forests
Little influenced forests
PLOT
2
6
8
9
Mean
Standar
d error

VRR in
Subplot
(g/m2/day)
1.89
2.05
1.92
2.12
2.00
0.11

Restoration forests
Convert to

ha
tons/ha/year
)
6.90
7.48
7.01
7.74
7.28
0.400

PLOT
1
3
4
5
7
10
Mean
Standar
d error

VRR in
subplot
(g/m2/day)

Convert to
ha
tons/ha/year)

1.75

2.34
3.34
3.25
2.08

6.39
8.54
12.19
11.86
7.59

2.34
2.52

8.54
9.19

0.64

2.34

3.2.1.3. AGB calculated from Experiments
a) Litter fall biomass (death biomass)
Litter fall biomass of 6 experimental plots from two forest status was
synthesized in table 3.9 shown that litter fall biomass in restoration forest
was higher than that in little influence forest. Litter fall mass had an
average of 1.64 g/m2/day (equivalent with 6.02 tons/ha/year) for the little
influenced forest and 2.26 g/m2/day (equivalent with 8.25 tons/ha/year) for
the restoration forest. Litter fall mass here were lower than in permanent
sample plots. but these differences were statistical not significant.

Table 3.9. Litter fall biomass in two forest status
Little influenced forest

Restoration forest

Plot
(g/m2/day)

(tons/ha/year)

(g/m2/day)

(tons/ha/year)

1

1.75

638.75

2.01

733.65

2

1.42

518.3

2.42

883.3

3

1.74

635.1

2.36

861.4

Average

1.64

598.6

2.26

824.9

Standard error 0.19

69.35

0.22

80.3

b) AGB-increase
Based on data of diameter increase measured by dendrometer. AGBincrease were estimated by equation 2.4 as the form:
ΔM = ΔM= a*Dj^b - a*Di^b


17

In which parameter a = 0.0755 and b = 2.57.
In 3 sample plots of little influenced forest. there were 90 trees of 18
species measured and estimated an average biomass-increase of 6.35±6.08
kg/tree. the tree had maximal biomas-increase of 22.44 kg and the minimu
of 1.08 kg.
For the restoration forest. there were 120 trees of 24 species (5
trees/species) and estimated an average biomass-increase of 7.87±7.75
kg/tree with tree had minimal biomass-increase was 0.99 and maximal of
29.24 kg (see table 3.10b). AGB-increase in little influenced forest was
4.31±0.17 tons/ha/year and in restoration forest was 6.91±0.32
tons/ha/year. higher than that in little influenced forest. Biomass increase
in restoration forests varied wide than that in little influenced forests.
Larger diameter classes played higher role for biomass increase for
both forest status. In the little influenced forests. all trees with diameter
D1.3 ≤ 30 cm contributed of 14 % of total AGB and 30.8 % of biomassincrease (figure 3.3). While in restoration forest. all trees with D 1.3 ≤ 30 cm
contributed of 13.3 % of total AGB and 10.7 % AGB-increase (figure 3.4).
Figure 3.3. Biomass and biomass-increase in little influenced forests in
Kon Ha Nung
Figure 3.4. Biomass and biomass-increase in restoration forests in
Kon Ha Nung
c) Total AGB

The total AGB of evergreen broadleaf forests in Kon Ha Nung had
been synthesized in table 3.11 shown that. AGB-increase varied from 10.3
to 15.2 tons/ha/year. maximal in restoration forest and minimal in little
influenced forest. In Which. the percentage of live biomass account from
40.6 to 52.3% and litter fall biomass (death biomass) account from 47.7 to
59.4%.
Table 3.11. AGB-increase of forests in Kon Ha Nung
Litter fall biomass
Little influenced forests
AGB (tons/ha/year) 6.02 ± 0.66
Carbon
3.01
(tons/ha/year)
Percentage (%)
47.7
Restoration forests
AGB (tons/ha/year) 8.25 ± 0.88

Live biomass

Total AGB

4.31 ± 0.17

10.33 ± 0.83

2.20

516

52.30
6.91 ± 0.32

15.16 ± 1.20


18

Carbon
(tons/ha/year)
Percentage (%)

4.12

3.30

55.70

43.30

7.60

These results shown that AGB-increase in little influenced forests
was (10.33 tons/ha/year) and less than that in restoration forests (15.16
tons/ha/year).
3.2.2. BGB accumulation capacity
3.2.2.1. Coarse root increase
Measured data were used to estimate coarse root masses in 3
experimental time from June 2015 to June 2016 based on relationship
equation between root production and diameter and length after the

formula (2.4): Cr=a*(φ2*L). The difference in the root masses between 4
June 2015 and 10 June 2016 was coarse root production for 371 days on a
soil surface area of 1 m2. Differences were calculated separately for the 3
experimental soil blocks. The mean and standard error were calculated for
each soil plot with surface area 1 m 2. These values were then converted to
production units g/m2/day and tons/ha/year (see table 3.14).
Table 3.14. Coarse root production increase between two measurements
Soil block
Biomass (g/m2)
Biomass
(1m2
production in
sureface
4.6.2015
10.6.2016
371 days g/m2
area
1
268.1
631.4
363.3

0.98

2

542.3

880.4

338.1

0.91

3

339.9

701.4

361.5

0.97

Mean
Standard
error

383.43

737.73

354.3

0.95

116.1

104.85

11.48

0.03

Biomassincrease
(g/m2/day)

Allometric relationship between coarse root production and diameter
and lenght
Figure 3.7. Regression between the logarithm of Ø2L and that of root mass
(g). each dot corresponds to a root segment.
Root mass is a function of root diameter and length. Regression
analyze results had determinated relationship equation of: y=0.913x+0.064
(figure 3.7). in the form of equation (2.6) in the methodology. The
equation had high correlation coefficient (R2 = 0.975). Replace the
parameters of equation 2.6 in equation 2.5 we have:


19

Cr = 1.1588x(Ø2×L)0.913
Estimated results of coarse root production had been synthezed in table
3.14. Coarse root production divided into diameter classes had been
susthezed in figure 3.8a shown that higher biomass production was
observed in smaller coarse roots. coarse root in the 0.2-1.0 cm diameter
class produced 0.42 g/m2/day. contributing to 42.9% of the total coarse
root production (figure 3.8b). Production was 0.27; 0.17; 0.06 and 0.03
g/m2/day for the 1.0-2.0; 2.0-3.0; 3.0-5.0 and >5.0 diameter classes.
respectively. The contribution to total coarse root production was 29; 19; 6
and 3% for the 1.0–2.0; 2.0–3.0; 3.0–5.0 and >5 cm diameter classes.

respectively.
The total coarse root production measured in research region in Kon Ha
Nung was 0.95 ±0.19 g/m2/day equivalent of 3.5±0.68 tons/ha/year (table
3.15).
Table 3.15. Coarse root production estimating for forests in Kon Ha Nung
Diameter
class
(cm)

Coarse root production
(g/m2/day)
Mean
Standard error

Coarse root production
(tons/ha/year)
Mean
Standard error

0.2-1.0

0.42

0.05

1.6

0.18

1.0-2.0

0.27

0.01

1

0.04

2.0-3.0

0.17

0.0006

0.6

0.02

3.0-5.0

0.06

0.003

0.2

0.01

>5.0

0.03

0.0002

0.1

0.01

All roots

0.95

0.19

3.5

0.68

a
b
2
Figure 3.8a. Coarse root production increase (g/m /day); b. Contribution
of diameter classes (%)
3.3.2. 2. BGB-increase for fine roots
a. Distribution of fine roots in soil depth
Fine roots distributed mainly in the soil depth 0-20 cm for both forest
status little influenced forest (figure 3.9a) and restoration forest (figure
3.9b). The more deep in the soil the little fine roots contributed. In the little
influced forest. there were 50.9 % of fine roots distributed in depth 0-20

cm. while in restoration forest were 53.8 %. Fine roots live. die and
decompose in little influenced and restoration forests had no differences.


20

Fine roots had decomposed of 0.006 g/m2/day in little influenced and
0.011 g/m2/day in restoration forest. Fine roots dead in the both forest
status were the same. While fine roots produced in little influenced forest
was 0.20 g/m2/day and in restoration forest was 0.18 g/m2/day.
a
b
Figure 3.9a. Fine root distribution in soil depth of little influenced forest;
b. Fine root distribution in soil depth of restoration forest.
Figure 3.9 presented live fine root distribution thể (figure 3.9a) and
death fine root (figure 3.9b) in the depth of soil. From figure 3.9a shown
that on research forests. live fine roots can contribute in depth of 1 m.
However. more than 50% of live fine roots distributed in 0-20 cm. here is
the soil layer with high humus matters. nutrients. and water. Therefore.
live fine roots were concentrated to take up them to feeed trees. These
leads to the distribution of death roots in this layer.
b. Decomposition rate of fine roots
Decomposition rates of fine roots in litter bags buried in different
locations were different. variation range of decompositon rates was from
0.39 to 1 (nearly entirely decomposed). These shown that micro
enviroment in baried location affected decomposion rates.
Litter bags baried in high humid sites with rich humus and micro
organism acted strongly had high decomposion rate. After barying 10
months. fine roots had decomposion rate of an average 0.54. Like that.
after ca. 20 months death fine roots will be composed entirely to give back

nutrients to the soil. Decomposion rate of fine roots dependend mainly on
climate conditions. where high humid. high temperature and high rain falls
promoting microorganism action then the decomposion rate increased
Decomposition rate of fine roots with diameter (ϕ ≤ 1 mm) was hihgher
than that of which with diameter (1 < ϕ ≤ 2 mm). In the time from March
to June. decompostion rate of fine roots was 0.284. while that of larger fine
roots was 0.213 (Tran Van Do và cộng sự. 2015).
c. Fine roor production
Fine root live, die and decompose in little influenced forest (figure
3.10a) and in restoration forest (figure 3.10b) were no difference. Fine
roots were decomposed of 0.006 g/m2/day in little influenced forest and
0.011 g/m2/day in restoration forest. Death fine roots were the same in both
forest status. While total fine root production in little influenced was 0.20
g/m2/day and in restoration forest was 0.18 g/m 2/day. Total fine root


21

production in little influenced forest was 0.73±0.28 and in restoration
forest was 0.66±0.25 tons/ha/year.
Besides, fine root biomass depends on forest objects. research region.
forest age. climate conditions. soils. Like that. research on fine root
production had been conducted for each forest. climate and soil zone
differently. From that. the understand about fine root production. the role
of fine root in forest ecosystems and carbon cycle in forest ecosystems
sufficient and perfect.
Total biomass produced below ground was synthezed in table 3.18 shown
that BGB of minor-influenced forest was 4.24±0.96 tons/ha/year (in which
coarse root biomass account of 85.1%) and in restoration forest was
4.16±0.93 tons/ha/year with 84.5% of coarse root.

a
b
Figure 3.10a. Fine root decomposed (d). die (m). and live (p) in little
influenced forest; b. Fine root decomposed (d). die (m) and live (p) in
restoration forest.
Table 3.18. BGB of experiment forest
Fine root (tons/ha/year)
Coarse root (tons/ha/year)
Total (tons/ha/year)

Little influenced forest
0.73±0.28
3.5±0.68
4.24±0.96

Restoration forest
0.66±0.25
3.5±0.68
4.16±0.93

3.2.2.3. Respiration of soil micro-organism
Results had differences due to measurement time and forest status
(figure 3.11). From figure 3.11 had seen. that respiration of microorganism in both forest status in August (rainy season) was higher than in
June (dry season). Respiration in restoration forest was higher than that in
little influenced forest.
Figure 3.11. Respiration of soil micro organism in forests of Kon Ha Nung
Respiration of soil micro organism in little influenced forest was
lower than that in restoration at any time (figure 3.11). The mean value of
respiration of soil micro organism was 3±0.2 g/m 2/day (equivalent with 1.5
g carbon/m2/day) for little influenced forest and 3.4 g/m 2/day (equivalent

with 1.7 g carbon/m2/day) for restoration forest (table 3.19).
Table 3.19. Average respiration of micro organism in forest ò Kon Ha
Nung
2

Biomass (g/m /day)

Little influenced forest
3.0±0.2

Restoration forest
3.4±0.3


22

Carbon (g/m2/day)

1.5±0.1

1.7±0.15

So that, average respiration of soil micro organism in forests of Ko
Ha Nung was 10.95±0.73 tons/ha/year dry biomass in little influenced
forest lower than that in restoration forest 12.41±1.10 tons/ha/year.
3.2.3. Total biomass accumulated annually
Table 3.20. Total biomass- increase of research forests
AGB
Little influenced forest
Biomass

10.33±0.83
(tons/ha/year)
Caron
5.2
(tons/ha/year)
Restoration forest
Biomass
15.16±1.2
(tons/ha/year)
Carbon
7.6
(tons/ha/year)

BGB

Respiration

Total

4.24±0.96

10.95±0.73

3.62±1.06

2.12

5.45

1.8

4.16±0.93

12.41±1.10

7.02±1.03

0.99

6.2

3.5

Results synthezed in table 3.20 shown that biomass increase annually
in little influenced forest was 3.62±1.06 tons/ha/year lower than that in
restoration with 7.02±1.03 tons/ha/year.
Biomass accumulation capacity of natural forests in north west
region was 7.82 tons/ha/year for little influenced forest (IIIB) was 13.36
tons/ha/year for restoration forest. Biomass accumulation capacity in north
east region was 6.8 tons/ha/year for little influenced forest was 7.6
tons/ha/year for restoration forest (Trần Văn Đô et al.. 2016) .
Like that, biomass calculation capacity of evergreen broadleaf forests
were different in ecological zones and depended strongly on forest status.
General trend was minor-influenced forests accumulated lower biomass
than restoration forests in all ecological zones. These are corresponded
with natural rules.
3.3. Enhance the accurate of root biomass estimating methods
Belowground net primary production (BNPP) in forests included fine
root production and coarse root production. It is not easy to estimate
BNPP. because roots are small. numerous and underground. Destructive

sampling by excavating entire root systems of individual sampled trees can
miss up to 30% of coarse roots. This is a costly. labor-intensive method.
and it may be prohibited in context of forest ban. The thesis applied a new
developed method to estimate coarse root production without destructive
sampling of trees. The method is developed and applied to an evergreen


23

broadleaf forest in Vietnam (Tran Van Do và cộng sự. 2018) [103].
Measurements of diameter and length of coarse roots at times tj and ti (Δt;
tj ≥ ti) are used to estimate root mass (Cr) from root length and diameter
squared (Ø2·L). The results indicated that the combined error of estimating
mass of coarse roots using such relationship was a 3.4 % overestimate. The
coarse root production in this study was 0.99 g/m2/day. It is concluded that
the present method to estimate coarse root production using allometry
betwen Cr and Ø2·L is relatively easy and applicable to any forests where
small blocks of soil can be excavated to measure and remeasure coarse
roots over time.
3.3.1. Enhance the accurate of coarse root biomass estimating method
3.3.1.1. Allometry establishment
A total of 268 coarse root segments were collected in the 31-m 3 soil
blocks (1 m × 1 m × 1 m). There were 4 root segments in the >5 cm
diameter class. 6 root segments in the 3.0–5.0 cm diameter class. 11 root
segments in the 2.0–3.0 cm diameter class. and 95 root segments in the
0.2–1.0 cm diameter class (Table 3.21). The total length of the coarse root
segments collected in the 3 soil blocks was 16.590 cm. and the total dry
mass was 16.257 g (Table 3.21). The increase in diameter of the coarse
roots was size-dependent with smaller roots on average increasing more in
diameter than larger roots. The diameters of the larger coarse root classes

increased less (Table 3.21).
Coarse roots had diamter of 0.2 - 1.0 cm contained up to 60% water.
whereas roots >5 cm in diameter contained 34% water. The range of
moisture content was greater in the smalle diameter classes. The validation
statstics for different root diameter classes and a combined diameter class
are shown in table 3.22.
The calibrated model for roots >5.0 cm in diameter had the lowest
percentage bias of –2.3. Bias increased progressively for the 0.2–1.0 cm
diameter class (–4.1). the 2.0–3.0 cm diameter class (–5.1). the 1.0–2.0 cm
diameter class (–8.0). and the 3.0–5.0 cm diameter class (–13.3). The
general model for the mixed diameter class of all root segments had –6.6
percent bias. The allometric relationship based on all coarse root segments
combined had a correlation coefficient of determination (R 2 ) of 0.84
(Table 3.20). and it overestimated total coarse root biomass by 18.8%
(Table 3.22). The combined estimation errors were smaller when the five
coarse root diameter classes were applied individually with their
corresponding allometric relationships (Table 3.22).


24

Table 3.21. Length, diameter increase, dry mass and moisture of coarse
roots
Diameter
class (cm)
0.2 - 1.0
1.0 - 2.0
2.0 - 3.0
3.0 - 5.0
>5.0

Total

Diameter
increase
(mean±SE)
1.34 ±
0.12
1.06 ± 0.11
0.95 ± 0.08
0.88 ± 0.06
0.63 ± 0.07

Range of
water
content
(%)

Number of
root
segments

Range of
root length
(cm)

Range of
dry root
mass (g)

152

5 - 125

2 - 129

48 - 60

95
11
6
4
268

3 - 98
7 - 89
4 - 78
27 - 48
16.590

2 - 288
8 - 418
44 - 876
400 - 1.188
16.257

43 - 50
40 – 46
36 - 42
32 - 34

Table 3.22. Validation for models of different diameter classes and mixed
diameter
Diameter class
(cm)
0.2-1.0
1.0-2.0
2.0-3.0
3.0-5.0
3.0-5.0
3.0-5.0

Percentage bias

RMSPE

MAPE

-4.1
-8.0
-5.1
-13.3
-13.3
-13.3

31.7
36.3
7.8
35.6
35.6
35.6

19.8
19.7
13.6
29.0
29.0
29.0

Table 3.23. Estimated parameters of relationship equations for diameter
classes
Diameter
class
(cm)
0.2-1.0
1.0-2.0
2.0-3.0
3.0-5.0
>5.0
All roots

R2

Estmated
parameters

a
0.68
1.8289
0.76
0.5568

0.88
0.2056
0.87
2.7416
0.67
2.1717
0.9758 0.9131

b
0.6104
1.0085
1.1820
0.7213
0.8147
0.9758

Standard error
a
0.0781
0.0902
0.1337
0.0861
0.1501
0.0504

b
0.0653
0.0672
0.0540
0.0394

0.1088
0.0424

Estmation error
of coarse root
biomass (%)
11.6
3.9
0.3
5.2
0.6
18.8

The relationship for coarse roots with diameter >5 cm had a
correlation coefficient of 0.67 and overestimated 0.6 percent of the root
mass. The relationship for coarse roots in the 3.0–5.0 cm diameter class
had a correlation coefficient of 0.87 and overestimated the root mass by
5.2 percent. The relationship for coarse roots in the 2.0–3.0 cm diameter
class had a correlation coefficient of 0.88 and underestimated root mass by


25

0.3 percent. The relationship for coarse roots in the 1.0–2.0 cm diameter
class had a correlation coefficient of 0.76 and overestimated root mass by
3.9 percent. The relationship for coarse roots in the 0.2–1.0 cm diameter
class overestimated by 11.6 percent. (Table 3.23). The combined error for
estimating coarse root mass using these five relationships was a 3.4%
overestimate compared with an overestimate of 18.8 percent for the single
combined model. Root mass is a function of root diameter and length.

From collected number of root segments. root length and diameter and
mass (table 3.20). it was established relationship equation between survey
information for all coarse roots in form of linear equation as
y=0.913x+0.064 (figure 3.8). equation (2.6) in methodology. The equation
had a correlation coefficient of (R 2 = 0.975) and can be used to estimate
coarse root biomass without collecting root data by traditional destructive
method. Replace the parameters of equation 2.6 in equation 2.5 we have:
Cr = 1.1588x(Ø2×L)0.913
Research results shown that coarse root biomass had of 18.8%
overestimated. so that the accurate of relationship equation was high and
can be used to estimate coarse root biomass of the forests.
3.3.1.2. Improvement of coarse root estimation method
The method used in this thesis was modified to be simple and easy to
apply with a minimum of equipment and without destructive sampling of
trees. Trees are not excavated; rather. a sample of coarse roots is collected
to establish the allometric relationship of root mass to diameter and length.
Repeated measurements over time of root diameter and length are used to
estimate coarse root production for Δt. It is easier to collect all coarse roots
(Ø > 2 mm) in sampling soil blocks using the present method than to
collect all coarse roots of sampled trees using previously published
methods. which may miss 30 percent of all coarse roots when sampling
sampled trees (Ogino 1977, Niiyama et al. 2010).
In this thesis, we divided the coarse root segments into five diameter
classes (Ø = 0.2–1.0. >1.0–2.0. >2.0–3.0. >3.0–5.0 cm. and >5.0 cm) and
established five corresponding allometric relationships, which reduced
composite estimation and resulted in a combined overestimate error of
3.4%. Meanwhile, if we combined all root segments (Ø ≥ 2 mm) and
established one allometric relationship for all diameter classes, the coarse
root biomass was overestimated by 18.8 percent.
3.3.2. Improvement of fine root estimation method