3.1. Intergeneric, interspecific and intraspecific wood density analysis
Among the different part of the wood, significance difference in wood mean density between the two genera is observed only on heartwood, as evaluated by Kruskall Wallis test (p < 00001). The mean HWD of Dalbergia and Diospyros genera are respectively 0.961 ± 0.142 g∙cm-3 and 1.070 ± 0.165 g∙cm-3. This difference could be attributed to a higher content of extractives in the heartwood of Diospyros. However, some authors such as Zanne et al. (2009) found the opposite results, with higher wood density for Dalbergia compared to Diospyros genus. As summarize in table 2 for both genera, Malagasy species generally have denser wood compared to species of the same genus growing in other countries. The wood of Malagasy Dalbergia and Diospyros species is also denser than that of most native species. The average wood density of Dalbergia is similar to Stephanostegia capuronii Mark (WD = 0.975 g∙cm-3) and Scolopia madagascariensis Si (WD = 0.950 g∙cm-3), while the average wood density of Diospyros is similar to Neobeguea mahafaliensis Leroy (WD = 1.040 g∙cm-3) and Cedrelopsis grevei Baillon (WD = 1.000 g∙cm-3) (Blaser et al. 1993).
Between-species, wood mean density varies significantly either based on CWD, SPW or HWD for Dalbergia and Diospyros species. The wood mean density variation is however higher based on HWD, with coefficients of variation of 13.92% and 16.33% for Dalbergia and Diospyros species respectively. Mean HWD for Dalbergia species ranges from 0.655 ± 0.070 g∙cm-3 (Dalbergia baronii Baker) to 1.299 g∙cm-3 (Dalbergia rakotovaoi) while it ranges from 0.740 ± 0.072 g∙cm-3 (Diospyros subtrinervis H. Perrier) to 1.273 ± 0.047 g∙cm-3 (Diospyros malandy H.N. Rakouth, Randrianaivo, G.E. Schatz & Lowry) for Diospyros species. Intraspecific density variation ranges (between trees) are closely similar for both genera, with a Coefficient of Variation of 1.51% (Dalbergia madagascariensis) up to 26.68% (Dalbergia ambongoensis) for Dalbergia species, and 1.94% (Diospyros crassifolia) up to 22.51% (Diospyros sp.) for Diospyros species.
The Coefficient of Variation of the HWD seems to be the best proxy for defining wood density-based groups. HWD values show that the 35 Dalbergia (n = 150) species can be classified in eleven groups, as shown in Table 1 and Fig. 4. Among the different Dalbergia groups that exhibit significantly distinct HWD ranges, Dalbergia rakotovaoi and Dalbergia baronii stand out as distinct classes. The former has the highest HWD of 1.299 g∙cm-3, while the latter has the lowest HWD of 0.655 ± 0.07 g∙cm-3. For the Diospyros species (n = 72), seven groups can be observed. Diospyros malandy has the highest HWD, while Diospyros subtrinervis has the lowest. This interspecific density variation can be attributed to the difference in anatomical characteristics between the species which affect its wood density. The comparison of our results to the published anatomical data of Malagasy Dalbergia and Diospyros species (Sandratriniaina et al. 2021; Ramanantsialonina et al. 2022) allow to understand that species with a larger tangential vessel diameter tend to have lower heartwood density. For example, this is the case with Dalbergia baronii (HWD = 0.655 g∙cm-3) and Dalbergia davidii (HWD = 1.172 g∙cm-3), whose wood is among the lighter and denser, respectively, based on HWD values, and they have mean TVD of 136 µm and 64 µm (Ramanantsialonina et al. 2022). In contrast, Diospyros malandy (HWD = 1.273 g∙cm-3) and Diospyros lewisiae (HWD = 0.780 g∙cm-3), which also differ significantly in terms of heartwood density values, have mean TVD of 43 µm and 73 µm, respectively (Sandratriniaina et al. 2021). This is supported by Fichtler et al. (2012), who assessed, across various tropical species, the relationships between wood anatomical variables and tree growth site conditions. They also found a negative correlation between vessel diameter and wood density (Fichtler et al. 2012). This can be explained by the fact that wood density is generally influenced by the proportion of solid material (cellulose, lignin, etc.) relative to the volume of empty spaces (vessels, cavities, etc.). The presence of larger vessels thus contributes more significantly to creating void spaces within the wood, resulting in a reduction of its density. Other factors for which data are not available, and whose effect on wood density could not be assessed for Malagasy Dalbergia and Diospyros, may also interact with these wood anatomical factors to determine wood density. These factors include for example genetic factors such as tree growth rate (Hietz et al. 2013), or environmental variables such as soil type.
Table 1 Mean values (± standard deviation) of wood density at 12% moisture content according to wood section (whole core, heartwood, sapwood) for the 38 Dalbergia and 29 Diospyros.
Botanical name (Dalbergia, n= 165)
|
n
|
CWD
|
SWD
|
HWD
|
Class
|
Dalbergia rakotovaoi
|
1
|
1.111
|
1.092
|
|
1.299
|
|
a
|
Dalbergia davidii Bosser & R. Rabev.
|
2
|
1.145
|
± 0.005
|
1.091
|
|
1.172
|
± 0.043
|
ab
|
Dalbergia chermezonii R. Vig.
|
2
|
1.132
|
± 0.039
|
1.091
|
± 0.040
|
1.154
|
± 0.054
|
abc
|
Dalbergia humbertii R. Vig.
|
5
|
1.090
|
± 0.076
|
1.006
|
± 0.019
|
1.140
|
± 0.111
|
abc
|
Dalbergia bemarivensis Phillipson & N. Wilding
|
8
|
1.119
|
± 0.059
|
1.062
|
± 0.083
|
1.138
|
± 0.105
|
abc
|
Dalbergia suaresensis Baill.
|
5
|
1.039
|
± 0.059
|
1.015
|
± 0.046
|
1.074
|
± 0.061
|
abc
|
Dalbergia lemurica Bosser & R. Rabev.
|
3
|
1.051
|
± 0.018
|
0.905
|
± 0.118
|
1.070
|
± 0.039
|
abcd
|
Dalbergia trichocarpa Baker
|
3
|
1.009
|
± 0.045
|
0.975
|
± 0.048
|
1.053
|
± 0.040
|
abcde
|
Dalbergia purpurascens Baill.
|
12
|
1.027
|
± 0.043
|
1.022
|
± 0.052
|
1.048
|
± 0.062
|
abcde
|
Dalbergia neoperrieri Bosser & R. Rabev.
|
1
|
0.977
|
|
0.857
|
|
1.040
|
|
abcdef
|
Dalbergia pseudobaronii R. Vig.
|
5
|
0.994
|
± 0.015
|
0.970
|
± 0.030
|
1.037
|
± 0.041
|
abcdef
|
Dalbergia obtusa Lecompte
|
7
|
0.994
|
± 0.057
|
0.952
|
± 0.044
|
1.018
|
± 0.070
|
abcdef
|
Dalbergia abrahamii Bosser & R. Rabev.
|
7
|
0.961
|
± 0.053
|
0.913
|
± 0.038
|
1.001
|
± 0.072
|
abcdef
|
Dalbergia tricolor Drake
|
6
|
0.947
|
± 0.061
|
0.925
|
± 0.057
|
0.993
|
± 0.125
|
abcdef
|
Dalbergia emirnensis Bosser & R. Rabev.
|
14
|
0.994
|
± 0.054
|
1.019
|
± 0.044
|
0.989
|
± 0.065
|
abcdef
|
Dalbergia aff. greveana Baill.
|
1
|
0.945
|
|
0.898
|
|
0.986
|
|
abcdef
|
Dalbergia chlorocarpa R. Vig.
|
1
|
0.964
|
|
0.940
|
|
0.977
|
|
abcdef
|
Dalbergia greveana Baill.
|
5
|
0.957
|
± 0.086
|
0.941
|
± 0.081
|
0.972
|
± 0.122
|
abcdef
|
Dalbergia densicoma Baill.
|
6
|
0.931
|
± 0.101
|
0.928
|
± 0.100
|
0.946
|
± 0.142
|
abcdef
|
Dalbergia aff. purpurascens Baill.
|
1
|
0.935
|
|
0.909
|
|
0.946
|
|
abcdef
|
Dalbergia occidentalis
|
4
|
0.932
|
± 0.057
|
0.914
|
± 0.054
|
0.942
|
± 0.084
|
abcdef
|
Dalbergia ambongoensis Baill.
|
2
|
0.912
|
± 0.133
|
0.916
|
± 0.115
|
0.933
|
± 0.249
|
abcdef
|
Dalbergia bathiei R. Vig.
|
5
|
0.845
|
± 0.052
|
0.819
|
± 0.039
|
0.911
|
± 0.070
|
abcdef
|
Dalbergia glaucocarpa Bosser & R. Rabev.
|
5
|
0.900
|
± 0.069
|
0.903
|
± 0.068
|
0.880
|
± 0.096
|
bcdef
|
Dalbergia antsirananae Phillipson, Crameri & N. Wilding
|
7
|
0.888
|
± 0.052
|
0.895
|
± 0.052
|
0.867
|
± 0.047
|
bcdef
|
Dalbergia leandrii
|
2
|
0.866
|
± 0.003
|
0.864
|
± 0.010
|
0.867
|
± 000
|
cdef
|
Dalbergia viguieri Bosser & R. Rabev.
|
4
|
0.912
|
± 0.069
|
0.906
|
± 0.078
|
0.862
|
± 0.027
|
cdef
|
Dalbergia madagascariensis (Baker) Bosser & R. Rabev.
|
4
|
0.884
|
± 0.050
|
0.891
|
± 0.066
|
0.858
|
± 0.013
|
cdef
|
Dalbergia normandii Bosser & R. Rabev.
|
6
|
0.842
|
± 0.119
|
0.828
|
± 0.119
|
0.841
|
± 0.215
|
cdef
|
Dalbergia cloiselii Drake
|
1
|
0.820
|
|
0.787
|
|
0.839
|
|
cdef
|
Dalbergia monticola Bosser & R. Rabev.
|
12
|
0.856
|
± 0.045
|
0.875
|
± 0.062
|
0.838
|
± 0.063
|
def
|
Dalbergia orientalis Bosser. & R. Rabev.
|
5
|
0.786
|
± 0.056
|
0.791
|
± 0.075
|
0.830
|
± 0.168
|
def
|
Dalbergia aff. chapelieri Baill.
|
4
|
0.815
|
± 0.110
|
0.842
|
± 0.100
|
0.781
|
± 0.116
|
ef
|
Dalbergia L. f.
|
1
|
0.755
|
|
0.790
|
|
0.721
|
|
ef
|
Dalbergia baronii Baker
|
5
|
0.743
|
± 0.082
|
0.757
|
± 0.075
|
0.655
|
± 0.070
|
f
|
Dalbergia chapelieri Baill.
|
1
|
0.711
|
|
0.711
|
|
|
|
-
|
Dalbergia manongarivensis Bosser & R. Rabev.
|
1
|
0.828
|
|
0.828
|
|
|
|
-
|
Dalbergia rajeryi
|
1
|
1.128
|
|
1.128
|
|
|
|
-
|
Botanical name (Diospyros, n= 132)
|
n
|
CWD
|
SWD
|
HWD
|
Class
|
Diospyros malandy H.N. Rakouth, Randrianaivo, G.E. Schatz & Lowry
|
4
|
1.018
|
± 0.044
|
0.976
|
± 0.042
|
1.273
|
± 0.047
|
a
|
Diospyros crassifolia A.G. Linan, G.E. Schatz & Lowry
|
5
|
1.089
|
± 0.037
|
1.006
|
± 0.038
|
1.232
|
± 0.024
|
ab
|
Diospyros torquata H. Perrier
|
3
|
1.008
|
± 0.030
|
0.983
|
± 0.046
|
1.225
|
± 0.095
|
abc
|
Diospyros obscurinerva A.G. Linan, G.E. Schatz & Lowry
|
5
|
1.087
|
± 0.049
|
1.024
|
± 0.059
|
1.211
|
± 0.039
|
abc
|
Diospyros analamerensis H. Perrier
|
5
|
1.054
|
± 0.022
|
1.012
|
± 0.037
|
1.180
|
± 0.023
|
abc
|
Diospyros chitoniophora Capuron ex A.G. Linan, G.E. Schatz & Lowry
|
6
|
1.044
|
± 0.032
|
0.988
|
± 0.016
|
1.166
|
± 0.073
|
abc
|
Diospyros humbertiana H. Perrier
|
4
|
1.123
|
± 0.021
|
1.035
|
± 0.124
|
1.152
|
± 0.054
|
abc
|
Diospyros platycalyx Hiern
|
5
|
0.998
|
± 0.041
|
0.927
|
± 0.042
|
1.149
|
± 0.046
|
abc
|
Diospyros clusiifolia (Hiern) G.E. Schatz & Lowry
|
6
|
0.935
|
± 0.053
|
0.911
|
± 0.074
|
1.059
|
± 0.098
|
bcd
|
Diospyros toxicaria Hiern
|
7
|
0.937
|
± 0.033
|
0.912
|
± 0.050
|
1.031
|
± 0.042
|
cd
|
Diospyros ferrea (WillDiospyros) Bakh.
|
4
|
1.032
|
± 0.033
|
1.030
|
± 0.041
|
1.020
|
± 0.023
|
cde
|
Diospyros littoralis Capuron ex G.E. Schatz & Lowry
|
5
|
1.006
|
± 0.025
|
1.004
|
± 0.026
|
1.000
|
|
cde
|
Diospyros pubiramulis A.G. Linan, G.E. Schatz & Lowry
|
1
|
0.961
|
|
0.956
|
|
0.978
|
|
cde
|
Diospyros occlusa H. Perrier
|
6
|
0.874
|
± 0.084
|
0.836
|
± 0.035
|
0.973
|
± 0.162
|
cde
|
Diospyros tropophylla (H. Perrier) G.E. Schatz & Lowry
|
5
|
0.946
|
± 0.030
|
0.946
|
± 0.031
|
0.952
|
|
cde
|
Diospyros haplostylis Boivin ex Hiern
|
4
|
0.937
|
± 0.088
|
0.917
|
± 0.117
|
0.911
|
|
cde
|
Diospyros rubripetiolata G.E. Schatz & Lowry
|
4
|
0.917
|
± 0.018
|
0.924
|
± 0.011
|
0.894
|
|
cde
|
Diospyros baronii (H. Perrier) G.E. Schatz & Lowry
|
5
|
0.767
|
± 0.059
|
0.732
|
± 0.068
|
0.875
|
± 0.103
|
de
|
Diospyros sp. L.
|
4
|
0.929
|
± 0.081
|
0.927
|
± 0.084
|
0.866
|
± 0.195
|
de
|
Diospyros bardotiae H.N. Rakouth, G.E. Schatz & Lowry
|
5
|
0.864
|
± 0.033
|
0.864
|
± 0.033
|
0.860
|
|
de
|
Diospyros ultima G.E. Schatz & Lowry
|
4
|
0.816
|
± 0.065
|
0.811
|
± 0.073
|
0.835
|
|
de
|
Diospyros ramisonii G.E. Schatz & Lowry
|
1
|
0.861
|
|
0.861
|
|
0.796
|
|
de
|
Diospyros randrianasoloi G.E. Schatz, Lowry & Mas
|
3
|
0.695
|
± 0.090
|
0.750
|
± 0.183
|
0.792
|
|
de
|
Diospyros lewisiae Mas, G.E. Schatz & Lowry
|
6
|
0.866
|
± 0.066
|
0.873
|
± 0.062
|
0.780
|
|
de
|
Diospyros subtrinervis H. Perrier
|
5
|
0.751
|
± 0.058
|
0.758
|
± 0.063
|
0.740
|
±0.072
|
e
|
Diospyros squamosa Bojer
|
6
|
1.006
|
± 0.040
|
1.006
|
± 0.040
|
|
|
-
|
Diospyros brevipedicellata G.E. Schatz, Lowry & Mas
|
5
|
0.818
|
± 0.037
|
0.817
|
± 0.037
|
|
|
-
|
Diospyros gracilipes Hiern
|
5
|
0.843
|
± 0.085
|
0.843
|
± 0.085
|
|
|
-
|
Diospyros sakalavarum H. Perrier
|
4
|
0.795
|
± 0.070
|
0.795
|
± 0.070
|
|
|
-
|
CWD: mean density of the whole core, SWD: mean density of the sapwood, HWD: mean density of the heartwood, n: number of cores (trees) per species. Species groups defined by different letters have significantly different HWD intervals at a threshold of p > 0.05
This study significantly enriches the density database for Malagasy Dalbergia and Diospyros species as wood density was previously measured for only seven species of Dalbergia and two species of Diospyros (Richter and Dallwitz 2019; Louppe Dominique 2008; Gerard et al. 2016; Lisan n.d). Among publised density data found in the literature, only those that provided density at 12% moisture content and infradensity, which were subsequently converted into density at 12% moisture content according to the equation defined by Vieilledent et al. (2018), were compared with the results of our study. Some authors provide one average wood density value for trees belonging to several Dalbergia spp or Diospyros spp species (Rakotovao et al. 2012).
Regarding the interspecific variation of wood density, our results show that density variation is more pronounced in the heartwood, allowing the distinction of 11 Dalbergia and 7 Diospyros groups with significantly different density value ranges.
Table 2 Comparison of the average wood density of Dalbergia and Diospyros species in the literature.
Genus
|
n
|
Average wood density (g∙cm-3)
|
Countrys
|
References
|
Dalbergia
|
3
|
0.915 ± 0.02
|
India
|
Zanne et al. 2019
|
1
|
0.990
|
Madagascar
|
Rakotovao et al. 2012
|
2
|
0.857 ± 0.08
|
India
|
Reyes et al. 1992
|
5
|
0.837 ± 0.148
|
Madagascar
Africa
|
Louppe et al. 2008
|
1
|
0.850
|
Madagascar
|
Lisan, n.d
|
3
|
0906 ± 0.079
|
Madagascar
|
Cooke et al. 2008
|
spp
|
0.995
|
Madagascar
|
Rakotovao et al. 2012
|
38
|
0.961 ± 0.14
|
Madagascar
|
Current research
|
Diospyros
|
14
|
0.838 ± 0.01
|
India
|
Zanne et al. 2019
|
10
|
0.857 ± 0.08
|
Tropical Asia
Tropical Africa
Tropical America
|
Reyes et al. 1992
|
spp
|
0.990
|
Madagascar
|
Rakotovao et al. 2012
|
1
|
1.000
|
Madagascar
|
Louppe et al. 2008
|
29
|
1.070 ± 0.160
|
Madagascar
|
Current research
|
n : number of species, spp : several species whose list and number are not determined
3.2. Radial wood density variation
Kruskall Wallis test shows that there is no statistically significant difference in density values along the radial direction of the wood for the Dalbergia genus (p = 0.19). Mean density values of the core segments are nearly constant between the pith and the seventh segment (rho = -0,011), with an average value ranging from 0.982 ± 0.119 g∙cm-3 to 1.014 ± 0.100 g∙cm-3. Density then decreases in the three last segments, with an average value of 0.941 ± 0.105 g∙cm-3 to 0.982 ± 0.172 g∙cm-3 (Fig. 5). The wood density for Diospyros genus however varies significantly (p < 0.001) from the pith to the bark. The mean density values for the core segments range from 0.852 ± 0.104 g∙cm-3 to 1.010 ± 0.158 g∙cm-3. Table 3 shows the results of Spearman correlation test which reveals a decrease in wood density from the pith towards the bark, but the radial density trend is relatively weak (rho = -0,27).
Analysis of density variance at the species level shows a significant difference in radial wood density within the tree (p < 0.05) for Dalbergia abrahamii Bosser & R. Rabev., Diospyros chitoniophora Capuron ex A.G. Linan, G.E. Schatz & Lowry and Diospyros platycalyx Hiern (Fig. 6 and Fig. 7). Furthermore, table 3 indicate that these species exhibit a strong negative correlation (r or rho > 0.5) between wood density and the radial distance from the pith. The wood density for Dalbergia bathiei R. Vig., Diospyros crassifolia A.G. Linan, G.E. Schatz & Lowry and Diospyros occlusa H. Perrier decreases as one moves from pith towards the bark. However, the difference in density along the radial direction is not statistically significant (p < 0.5). This decrease is likely associated with the accumulation of extractives in the inner wood during the heartwood formation (Lehnebach et al. 2019). The accumulation of extractives can effectively raise the wood density, resulting in the heartwood being denser than the sapwood. Some researchers have suggested that the shade tolerance and physiological strategy of species during ontogeny also influence the radial density pattern of wood (Woodcock 2015). However, very little data is available in existing literature regarding the shade tolerance of Dalbergia and Diospyros species from Madagascar. The available information, nevertheless, suggests that there may be both shade-tolerant and light-demanding species within the same genus of Dalbergia or Diospyros (Blaser et al. 1993; Cooke et al. 2008; Razafimamonjy 2011; CITES 2016). The decreasing trend of radial wood density from the pith to the bark might be a characteristic feature of shade-tolerant species, as they generally produce dense wood during the juvenile stage (Hietz et al. 2013; Plourde et al. 2014) to limit growth and enhance survival in the understory of forests, providing protection against the risks of mechanical injury and pathogen attacks (Alvarez-Clare and Kitajima, 2007). They subsequently produce lighter wood in the adult stage to promote further growth when they reach higher levels in the canopy (Woodcock 2015). The prevalence of the decreasing trend in the radial density profile, observed in 10 out of the 17 analyzed species, is consistent with the literature which confirms that species from the Dalbergia and Diospyros genera are generally semi-shade tolerant (CITES 2016). Conversely, the increasing trend of radial wood density variation can be observed in light-demanding species that produce low-density wood during the juvenile stage to promote growth, and then denser wood in the outer part of the trunk during the adult stage (Woodcock 2015).
Six of the studied species show a U-shaped or inverted U-shaped radial density trend. The factors underlying this trend are not well-known in the literature, but according to some authors, such a trend is particularly found in trees in humid tropical forests (Williamson et al. 2012). This is consistent with four species analyzed species in our study (Dalbergia glaucocarpa , Dalbergia suaresensis, Diospyros ultima, and Diospyros ferrea), which display a similar trend and are found in forests located in humid or subhumid bioclimatic zones. However, the findings of this study also show that this trend can be observed in species inhabiting dry tropical forests (Dalbergia lemurica and Dalbergia emirnensis).
Table 3 Effect of radial distance from the pith to the bark on HWD and its trend along the radial direction of the wood
Genus
|
Botanical name
|
n
|
p-value
|
r
|
rho
|
Dalbergia
|
Dalbergia abrahamii Bosser & R. Rabev.
|
3
|
0.033
|
-0.64
|
|
Dalbergia bathiei R. Vig.
|
2
|
0.310
|
|
-0.59
|
Dalbergia bemarivensis Phillipson & N. Wilding
|
2
|
1
|
|
0.02
|
Dalbergia emirnensis var. decaryi Bosser & R. Rabev.
|
6
|
< 0.010
|
|
0.23
|
Dalbergia glaucocarpa Bosser & R. Rabev.
|
3
|
0.820
|
|
-0.06
|
Dalbergia lemurica Bosser & R. Rabev.
|
2
|
0.560
|
|
-0.25
|
Dalbergia obtusa Lecomte
|
5
|
0.210
|
-0.43
|
|
Dalbergia suaresensis Baill.
|
3
|
0.410
|
-0.31
|
|
Diospyros
|
Diospyros analamerensis H. Perrier
|
3
|
< 0.010
|
-0.07
|
|
Diospyros bardotiae H.N. Rakouth. G.E. Schatz & Lowry
|
4
|
0.850
|
0.20
|
|
Diospyros chitoniophora Capuron ex A.G. Linan. G.E. Schatz & Lowry
|
4
|
< 0.001
|
|
-0.86
|
Diospyros clusiifolia (Hiern) G.E. Schatz & Lowry
|
2
|
0.0819
|
-0.34
|
|
Diospyros crassifolia A.G. Linan. G.E. Schatz & Lowry
|
2
|
0.055
|
-0.92
|
|
Diospyros ferrea (Willd.) Bakh.
|
3
|
0.103
|
0.36
|
|
Diospyros occlusa H. Perrier
|
2
|
0.182
|
|
-0.71
|
Diospyros platycalyx Hiern
|
3
|
< 0.0001
|
-0.82
|
|
Diospyros ultima G.E. Schatz & Lowry
|
3
|
0.570
|
-0.33
|
|
n: number of cores, P-value: significance of Anova or Kruskall Wallis test, r and rho: Pearson and Spearman correlation coefficients, (r/ rho > 0.5 means strong correlation). Values in bold indicate a significant variation in radial density value and a strong correlation between density and radial distance from pith to bark.
3.3. Influence of growth site climatic conditions and tree diameter on average wood density
The results of correlation tests between wood density and climatic parameters are given in table 4. For both Dalbergia and Diospyros species, wood density increases with increasing temperature at the tree's growing site. This positive correlation is weak for SWD and moderate for HWD. On the other hand, an increase in precipitation corresponds to a decrease in wood density, although this correlation is weak, except for HWD in Diospyros species.
As shown in table 5, the forest stands location sites has therefore a significant influence on the CWD, SWD, and HWD, for both Dalbergia (p = 0.002; p = 0.011; p = 0.020) and Diospyros (p = 0.001; p = 0.03; p = 0.004). For the two genera, trees in subarid and dry bioclimates have higher values of CWD, SWD, and HWD compared to trees growing in subhumid and humid bioclimates. However, average CWD for the Diospyros genus is higher in humid bioclimates, where there is more precipitation, than in subhumid bioclimates. These results are in line with previous research by Vaughan et al. (2019) who observed an increase in wood density during drought conditions. Similarly, Ibanez et al. (2016) found that aridity conditions characterized by low precipitation and high temperatures promote the formation of high-density wood. In hardwood species, these results can be attributed to the reduction in average vessel width to limit maximum xylem hydraulic conductivity in response to water scarcity and high temperature (Hacke et al. 2016; Fajardo and Piper 2022). The variation in vessel diameter serves as a crucial characteristic for assessing the tree's adaptation to water availability (Hacke et al. 2016) and the decrease in average vessel width is an ecophysiological response of the tree to resist embolism (Mendez-Alonzo et al. 2012; Hacke et al. 2016). Additionally, Thomas et al (2007) affirmed that higher temperatures promote increased wood density by reducing vessel and fiber lumen area while increasing fiber wall thickness.
Regarding the influence of tree diameter on wood density, the Dbh of sampled trees varied widely between species and collection sites, ranging from 1.5 to 80 cm. The correlation between average core density and the diameter of the trees is very weak, whether for CWD (rho = - 0.08), SWD (rho = -0.06), or HWD (rho = -0.11). This can be attributed to the influence of multiple factors on tree diameter, including tree age, tree growth rate, environmental growth conditions, etc. (Ramananantoandro et al. 2016; Dey et al. 2017; Nabais et al. 2018; Mevanarivo et al. 2020). Furthermore, wood density can be influenced by other factors such as wood chemical composition (Vermaas 1975; Singleton et al. 2003) and anatomical structure (Pritzkow et al. 2014). All these factors can interact in complex ways, resulting in varying effects on wood density. The results of this study are consistent with those of Ramananantoandro et al. (2016) who worked on trees from the humid forests of Madagascar, as well as Flore et al. (2021) who studied trees from the lower stratum of the semi-deciduous forest in Cameroon. Both studies observed no significant correlation between wood diameter and its average density. However, Tenouewa et al. (2022) reported contradictory results for Acacia auriculiformis wood, where density increased with increasing tree diameter. This suggests that the influence of tree diameter on wood density may vary among different species. Limited data regarding the relationship between tree diameter and wood density can be found in the existing literature. Therefore, a comparative approach with other wood genera or species could provide additional information to more accurately assess this relationship.
Table 4 Values of rho correlations between climatic parameters (Temperature and precipitation) and average wood density (CWD, SWD, and HWD).
|
Dalbergia
|
Diospyros
|
|
CWD
|
SWD
|
HWD
|
CWD
|
SWD
|
HWD
|
Temperature
|
0.46
|
0.33
|
0.48
|
0.35
|
0.25
|
0.49
|
Precipitation
|
-0.32
|
-0.36
|
-0.22
|
-0.30
|
-0.17
|
-0.49
|
Table 5 Mean value ± standard deviation of CWD, SWD and CWD for species belonging to the genus Dalbergia and Diospyros according to bioclimatic zones.
|
|
Bioclimatic zone
|
Genus
|
Density (g∙cm-3)
|
Subarid
|
Dry
|
Subhumid
|
Humid
|
Dalbergia
|
CWD
|
1013.6 ± 59.2 (a)
|
995.6 ± 57.6 (a)
|
911.3 ±117.8 (b)
|
829.7 ± 58.5 (c)
|
SWD
|
1011.9 ± 63.4 (a)
|
963.4 ± 79.5 (a)
|
898.1 ± 103.3 (b)
|
835.6 ± 89.9 (c)
|
HWD
|
1022.3 ± 94.0 (a)
|
1012.6 ± 116.8 (a)
|
911.6 ± 141.1 (b)
|
827.7 ± 125.1 (b)
|
Diospyros
|
CWD
|
1123.1 ± 129.0 (a)
|
984.9 ± 86.3 (a)
|
869.5 ± 44.6 (b)
|
888.6 ± 52.3 (a)
|
SWD
|
1131.4 ± 123.0 (a)
|
994.7 ± 84.4 (a)
|
886.0 ± 47.8 (b)
|
873.0 ± 42.4 (b)
|
HWD
|
1121.0 ± 68.6 (a)
|
1016.0 ± 4 6.5 (a)
|
924.1 ± 25.0 (b)
|
869.7 ± 18.1 (b)
|
Different letters indicate groups with significantly different average density values
3.4. Operational application of the results
The updated density database of Malagasy Dalbergia and Diospyros wood can be harvested to manage sustainably forest ressources and enhance the valorization of wood properties. It can be firstly used to estimate the aboveground biomass (AGB) and carbon content for malagasy Dalbergia and Diospyros trees throughout allometrics models (Chave et al. 2005). Several authors have used wood density in the development of allometric equations for estimating AGB in dense rainforests (Vieilledent et al. 2012; Ramananantoandro et al. 2015) and secondary forests of Madagascar (Randrianasolo et al. 2019). Njana et al. (2016) worked on three species from Tanzanian mangrove forests (Avicennia marina (Forssk.) Vierh, Sonneratia alba J. Smith, and Rhizophora mucronata Lam.) and found that, in addition to dendrometric parameters, the accuracy of multispecies allometric models improves when wood density is considered. An accurate knowledge of AGB can significantly contribute to the sustainable utilization of forest resources as a decision tool regarding timber harvesting, conservation, restoration, and forest regeneration. Furthermore, the quantification of AGB provides a better understanding of a forest's capacity to sequester carbon from the atmosphere (Chave et al. 2005; Ramananantoandro et al. 2015). This holds paramount importance in the context of climate change, as forests play a crucial role in reducing atmospheric carbon dioxide (CO2) levels.
In the context of wood utilization, understanding interspecific density variation allows to better valorize wood properties by choosing the most suitable species for the intended use. In the furniture industry, for example, Dalbergia and Diospyros are highly sought after, even though wood density value is unknown for the majority of species in these two genera. The availability of such knowledge contributes however to improve the aesthetics, durability, and quality of furnitures, as density is correlated with various mechanical properties (Larjavaara and Muller-Landau 2011) and wood durability (Chave et al. 2009; Larjaavara and Muller-Landau 2010). Denser woods are inherently harder, more resistant to impacts and wear, less susceptible to deformation, and better equipped to withstand changes in humidity and temperature (Cabral et al. 2022). These makes wood density as an important criterion for selecting wood species for the production of furniture subject to heavy use.