*Article* **Color and Chemical Composition of Timber Woods (***Daniellia oliveri***,** *Isoberlinia doka***,** *Khaya senegalensis***, and** *Pterocarpus erinaceus***) from Different Locations in Southern Mali**

**Mohamed Traoré 1,2 and Antonio Martínez Cortizas 1,3,\***


**Abstract:** Wood characteristics and properties are related to various factors connected to the biochemical processes that occur in the tree during wood formation, but also, to the interactions with the environmental conditions at the tree growing location. In addition to climatic factors, several investigations drew attention to the significance of the influence of other environmental parameters at the tree growing location. In this perspective, this work aimed to characterize the variation in color and chemical composition of timber wood from different locations in southern Mali, of trees growing under the same climatic conditions. To do so, a total of 68 grounded wood samples, from 4 timber wood species (*Daniellia oliveri*, *Isoberlinia doka*, *Khaya senegalensis*, and *Pterocarpus erinaceus*), were analyzed using CIELab color space and FTIR-ATR. Overall, the results indicated that the variation in wood color and chemical properties can be related to the local environmental conditions. *Pterocarpus erinaceus* presented significant differences between samples from the three areas according to the highest number of variables (color parameters, molecular composition determined by FTIR-ATR spectroscopy, and FTIR-ATR ratios). *Daniellia oliveri* and *Khaya senegalensis*, however, showed significant differences between areas of provenance for a lower number of variables. *Isoberlinia doka*, for its part, showed no significant differences and seems to be less sensitive to environmental factors. Furthermore, the results revealed that important differences exist between wood samples from Kati and Kéniéba.

**Keywords:** tropical timber; color parameters; mid-infrared spectroscopy; FTIR ratios; environmental factors; West African Sahel

#### **1. Introduction**

Wood characteristics and properties (e.g., aesthetical, physical, and chemical) are important parameters that make it valuable for a given end-use [1]. In construction, wood density and other mechanical properties, as well as its natural durability, are among the most considered wood characteristics to face biotic and abiotic threats, whereas, in the art industry, the required characteristics are related to its easy manipulation for carving or a certain specific tune when referring to musical instruments [2–4]. It is well known that wood characteristics and properties are related to various factors connected to the biochemical processes that occur in the tree during wood formation, but also, to the interactions with the environmental conditions at the tree growing location [5]. Over the last decade, numerous studies have been carried out on a large variety of wood species, in temperate regions and in tropical regions, to unravel the biogeochemical processes that are behind the composition of this complex lignocellulosic material [6–8]. Despite

**Citation:** Traoré, M.; Martínez Cortizas, A. Color and Chemical Composition of Timber Woods (*Daniellia oliveri*, *Isoberlinia doka*, *Khaya senegalensis*, and *Pterocarpus erinaceus*) from Different Locations in Southern Mali. *Forests* **2023**, *14*, 767. https://doi.org/10.3390/f14040767

Academic Editors: Vicelina Sousa, Helena Pereira, Teresa Quilhó and Isabel Miranda

Received: 4 March 2023 Revised: 4 April 2023 Accepted: 6 April 2023 Published: 8 April 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

the great advances in the field, more work is needed, especially taking advantage of the multidisciplinary aspect of wood research, to obtain a better understanding of the endogenous and exogenous factors controlling wood characteristics and properties.

The effects of the environmental conditions on wood structure and composition come out with various parameters related to the climate, as well as to physical, chemical, and biological processes at the tree growing location [9,10]. Variations of these environmental factors can be related to parameters that reflect such changes in wood structure and composition [11,12]. For instance, tree-ring width is one of the most used proxies in wood research to provide a general overview of tree growth [13,14]. The analysis of anatomical features and the structure of growth rings (earlywood width, latewood width, total ring width) provides insights into the main growth-limiting climatic factors (e.g., temperature, precipitation, drought, etc.), and enables the retrieval of information about the modification of structural features by soil moisture content, depth of groundwater table, slope, etc. [6,15]. Likewise, wood chemical composition (molecular, elemental, and isotopic) constitutes a potential means for the characterization of wood and for the understanding of the effect that the variability in environmental and climatic conditions has on the quality and properties of wood [16,17]. To our understanding, the development of these promising approaches still requires further research work in order to provide both a theoretical basis (i.e., an understanding of the factors behind wood composition and its characteristics) and a wellgrounded methodological approach for wood characterization and fingerprinting (i.e., a means to determine the specific features concerning a given factor).

An increasing number of studies are conducted to develop methodological approaches, using accessible and handy analytical techniques. This approach has been important to obtain a better understanding of wood physical and chemical characteristics in relation to environmental influences. To give an example, CIELab color space parameters have been used to study the link between wood color variation and various factors, including genetic, climatic, and edaphic [18–20]. In a like manner, the various applications of infrared spectroscopy revealed the great potential of this technique to investigate the link between wood molecular composition and tree species, as well as the wood provenance [21–24]. Additionally, data from these techniques can be used in parallel with other parameters related to structure and composition to foster the multiproxy potentially associated with wood research disciplines [20,25,26].

The main objective of the present research is to assess the variation in the characteristics (physical and chemical) of timber wood from three different areas of provenance, in a limited geographical area, under the same climatic conditions. To do so, we have applied CIELab color space and FTIR-ATR spectroscopy to wood samples of four timber wood species (*Daniellia oliveri*, *Isoberlinia doka*, *Khaya senegalensis*, and *Pterocarpus erinaceus*) from different locations (Kéniéba, Kita, and Sibi) in southern Mali. Data analysis has been performed using uni- and multivariate statistical methods (one-way ANOVA and principal component analysis).

#### **2. Materials and Methods**

#### *2.1. Origin of the Wood Materials and Wood Species*

The wood materials used for this research belong to 11 wood cross-sections obtained from individual planks of commercialized timbers, from three different areas in Mali: Kéniéba (12◦5004300 N, 11◦1400800 W), Kita (13◦1 04100 N, 9◦3005 00 W), and Sibi (12◦2203900 N, 8 ◦1903500 W) (Figure 1). All three areas are located, under the same climatic conditions, in the Sudano–Guinean climatic zone with average annual rainfall ranging between 800 and 1200 mm, and ferralitic soils predominate [4,27]. The consideration of this geographical delimitation is related to the main purpose of this study, which intends to draw attention to other factors characterizing the local environmental conditions (e.g., location of origin) rather than climatic factors that seem to be the most studied factors driving changes in wood properties.

wood properties.

to other factors characterizing the local environmental conditions (e.g., location of origin) rather than climatic factors that seem to be the most studied factors driving changes in

**Figure 1.** Map showing the area of origin of the studied woods. **Figure 1.** Map showing the area of origin of the studied woods.

The 11 wood cross-sections correspond to four timber wood species: *Daniellia oliveri*, *Isoberlinia doka*, *Khaya senegalensis*, and *Pterocarpus erinaceus*, which are indigenous African tree species with a geographic distribution ranging from West to East Africa [28–30]. These species are associated with high natural resistance to the less favorable climatic conditions of the tropical arid Sahelian and semiarid Sudanian zones in the West African Sahel [31,32]. Further, the selected wood species are among the most used timber wood with vital importance for local economic development [33,34]. According to the IUCN Red List, *D. oliveri* and *I. doka* are registered as the least concerned species, whereas *K. senegalensis* and *P. erinaceus* are, respectively, indicated as vulnerable and threatened species [35–38]. Moreover, these species are being affected by widespread forest degradation because of The 11 wood cross-sections correspond to four timber wood species: *Daniellia oliveri*, *Isoberlinia doka*, *Khaya senegalensis*, and *Pterocarpus erinaceus*, which are indigenous African tree species with a geographic distribution ranging from West to East Africa [28–30]. These species are associated with high natural resistance to the less favorable climatic conditions of the tropical arid Sahelian and semiarid Sudanian zones in the West African Sahel [31,32]. Further, the selected wood species are among the most used timber wood with vital importance for local economic development [33,34]. According to the IUCN Red List, *D. oliveri* and *I. doka* are registered as the least concerned species, whereas *K. senegalensis* and *P. erinaceus* are, respectively, indicated as vulnerable and threatened species [35–38]. Moreover, these species are being affected by widespread forest degradation because of increasing demand for forest products and excessive exploitation combined with the changing environmental conditions without appropriate regeneration measures [33,39].

increasing demand for forest products and excessive exploitation combined with the changing environmental conditions without appropriate regeneration measures [33,39].

divided into two identical parts along the radial direction. The first part was prepared following standard dendrochronological methods to improve the visibility of the growth

*2.2. Sample Preparation* 

#### *2.2. Sample Preparation*

*Forests* **2023**, *14*, x FOR PEER REVIEW 4 of 14

Once at the laboratory, for different analytical purposes, each cross-section was divided into two identical parts along the radial direction. The first part was prepared following standard dendrochronological methods to improve the visibility of the growth rings and related macroscopic and microscopic features. The wood surface was manually polished with fine sandpaper (grain sizes from P600 to P1200, Federation of European Producers of Abrasives). This first part enabled us to obtain images of transverse sections, using a Leica Emspira 3 digital microscope (Leica Microsystems GmbH, Wetzlar, Germany), as presented in Figure 2. As for the second part, in the first step, it was divided along the radial direction into small pieces of 3 to 4 cm. Thereafter, a portion (10–15 g) of each wood piece was milled to a very fine powder and homogenized using a Retsch mixer mill MM 301 (Retsch GmbH, Haan, Germany), with final fineness <5 µm, and oven dried at 30 ◦C for two weeks before the analysis using CIELab color space and infrared spectroscopy techniques. It is worth mentioning that the main reason for milling the wood to very fine powder is related to the later application of the X-ray fluorescence techniques to determine the elemental concentrations, not to the color measurement nor the infrared spectroscopy technique. rings and related macroscopic and microscopic features. The wood surface was manually polished with fine sandpaper (grain sizes from P600 to P1200, Federation of European Producers of Abrasives). This first part enabled us to obtain images of transverse sections, using a Leica Emspira 3 digital microscope (Leica Microsystems GmbH, Germany), as presented in Figure 2. As for the second part, in the first step, it was divided along the radial direction into small pieces of 3 to 4 cm. Thereafter, a portion (10–15 g) of each wood piece was milled to a very fine powder and homogenized using a Retsch mixer mill MM 301 (Retsch GmbH, Germany), with final fineness < 5 µm, and oven dried at 30°C for two weeks before the analysis using CIELab color space and infrared spectroscopy techniques. It is worth mentioning that the main reason for milling the wood to very fine powder is related to the later application of the X-ray fluorescence techniques to determine the elemental concentrations, not to the color measurement nor the infrared spectroscopy technique.

**Figure 2.** Transverse sections images of the studied wood. DO for *Daniellia oliveri*, ID for *Isoberlinia doka*, KS for *Khaya senegalensis*, and PE for *Pterocarpus erinaceus*. Z1, Z2, and Z3 refer to Kita, Sibi, and Kéniéba, respectively. **Figure 2.** Transverse sections images of the studied wood. DO for *Daniellia oliveri*, ID for *Isoberlinia doka*, KS for *Khaya senegalensis*, and PE for *Pterocarpus erinaceus*. Z1, Z2, and Z3 refer to Kita, Sibi, and Kéniéba, respectively.

In total, 68 grounded wood samples (see Table 1) were analyzed with two commonly used analytical techniques for wood color and chemical characterization, under room

Minolta Inc., Japan). The CIELab color space provides numerical values for color coordinates: lightness (L\*), green–red component (a\*), blue-yellow component (b\*),

*2.3. Analytical Techniques* 

#### *2.3. Analytical Techniques*

In total, 68 grounded wood samples (see Table 1) were analyzed with two commonly used analytical techniques for wood color and chemical characterization, under room standard temperature and humidity conditions. Color parameters were measured in the CIELab color space using a Konica-Minolta CR-5 Chroma Meter for solids (Konica Minolta Inc., Japan). The CIELab color space provides numerical values for color coordinates: lightness (L\*), green–red component (a\*), blue-yellow component (b\*), chroma (C\*), and hue (h◦ ). The infrared spectra were recorded, using an Agilent Cary 630 FTIR Spectrometer (Agilent Technologies Inc., USA), by Fourier-transform infrared spectroscopy in attenuated total reflectance mode (FTIR-ATR). Spectra were obtained in the mid-infrared region (MIR, 4000–400 cm−<sup>1</sup> ) by performing 100 scans at a resolution of 4 cm−<sup>1</sup> . Further analytical details are provided elsewhere [40,41].

**Table 1.** Total number of samples (per species and area) analyzed using CIELab and FTIR-ATR.


#### *2.4. Data Analysis*

For the color data, data analysis addressed the color coordinates including lightness (L\*), redness (a\*), yellowness (b\*), chroma (C\*), and hue (h◦ ). For the FTIR data, in the first place, we focused on a selected number of infrared bands on the basis of peak identification using the R package *andurinha* [42]. Then, we applied principal component analysis (PCA) using varimax rotation to the selected MIR bands, on the correlation mode. The extracted PCA components were further used, instead of individual FTIR bands. Finally, we calculated the following FTIR ratios: lateral order index (LOI, 1425/895), guaiacyl/syringyl ratios (1265/1230 and 1510/1595), lignin/carbohydrate ratios (1510/1460, 1510/1375, and 1510/895). These FTIR ratios mostly help to determine changes in the molecular structure of polysaccharides and lignin [43,44].

Furthermore, to evaluate whether there were significant differences between wood samples from the three areas (Kéniéba, Kita, and Sibi), for each wood species, we applied one-way analysis of variance (ANOVA) using as variables the CIELab color parameters, the extracted PCA components, and the calculated FTIR ratios.

The statistical tests, one-way ANOVA and principal component analysis, were performed using SPSS 27.

#### **3. Results and Discussion**

#### *3.1. Variations in Wood Color Parameters*

The color data indicated that the influence of the provenance of the wood sample can depend on the tree species. As presented in Table 2, no significant differences were observed between samples from the three areas where the timbers were collected, for *Daniellia oliveri* and *Isoberlinia doka*. However, for *Khaya senegalensis* and *Pterocarpus erinaceus*, there were significant differences between samples from the three areas. It appears that woods from Kita are highly different from woods of the two other areas (Sibi and Kéniéba). As for *K. senegalensis*, the highest yellowness (b\*) values were associated with wood from Kita, the lowest values for Kéniéba, and intermediate values for wood from Sibi (Figure 3). Also, the highest hue (h◦ ) values were related to wood from Kita, and wood from Sibi and Kéniéba similarly indicated the lowest h◦ values. Regarding *P. erinaceus*, woods from Kita were characterized by the highest values of lightness (L\*) and redness (a\*), and the lowest values were obtained for woods from Sibi and Kéniéba, which appear to be analogous. The h◦ values for *P. erinaceus* were found to be the opposite of what has been pointed out by the wood samples of *K. senegalensis*.


**Table 2.** Summary of the one-way ANOVA using the CIELab color parameters.

(1) significant differences at *p* < 0.01; and (2) significant differences at *p* < 0.05. (1) significant differences at *p* < 0.01; and (2) significant differences at *p* < 0.05.

**Figure 3.** Bar chart of the mean values of color parameters showing significant differences between samples from the three areas; (**a**) for *K. senegalensis* (*n* = 8 for Kita, *n* = 8 for Sibi and *n* = 8 for Kéniéba), (**b**) for *P. erinaceus* (*n* = 6 for Kita, *n* = 5 for Sibi and *n* = 6 for Kéniéba); (the labels a, ab, and b refer to group classification in ascending order of the Student–Newman–Keuls post-hoc test). **Figure 3.** Bar chart of the mean values of color parameters showing significant differences between samples from the three areas; (**a**) for *K. senegalensis* (*n* = 8 for Kita, *n* = 8 for Sibi and *n* = 8 for Kéniéba), (**b**) for *P. erinaceus* (*n* = 6 for Kita, *n* = 5 for Sibi and *n* = 6 for Kéniéba); (the labels a, ab, and b refer to group classification in ascending order of the Student–Newman–Keuls post-hoc test).

These differences observed using the wood color parameters may be related to the variation in wood contents in relation to the tree growing locations. As mentioned in various studies, wood color is connected to its chemical contents, especially the extractive compounds [45,46]. Therefore, the color characteristics of wood can result from processes linked to genetic effects but also the effects of environmental conditions at the tree These differences observed using the wood color parameters may be related to the variation in wood contents in relation to the tree growing locations. As mentioned in various studies, wood color is connected to its chemical contents, especially the extractive compounds [45,46]. Therefore, the color characteristics of wood can result from processes

growing location [20,47]. L\* and a\* are usually connected to the extractives and phenols, while b\* is usually related to lignin and polysaccharides [46,48]. The latter may suggest

*senegalensis* may be due to processes that control the chemical content in the wood cell walls (polysaccharides and lignins) and, on the other hand, the significant differences observed between the wood samples of *P. erinaceus* may be due to variation in the wood extractive contents. Similar results have been found for wood from trees growing in

different environmental conditions, including different soil properties [19,49,50].

linked to genetic effects but also the effects of environmental conditions at the tree growing location [20,47]. L\* and a\* are usually connected to the extractives and phenols, while b\* is usually related to lignin and polysaccharides [46,48]. The latter may suggest that, on the one hand, the significant differences observed for the wood samples of *K. senegalensis* may be due to processes that control the chemical content in the wood cell walls (polysaccharides and lignins) and, on the other hand, the significant differences observed between the wood samples of *P. erinaceus* may be due to variation in the wood extractive contents. Similar results have been found for wood from trees growing in different environmental conditions, including different soil properties [19,49,50].

#### *3.2. Variations in Wood FTIR Absorption Bands*

The principal component analysis extracted six components (with 26, 25, 14, 13, 10, and 7% of explained variance, respectively) that explained about 95% of the total variance in the dataset. All the selected bands showed a communality value above 0.85. The first component (PC1) is characterized by large positive loadings for bands related to vibrations in aromatic rings and in carboxylic groups in lignin (near 1623, 1606, and 1592 cm−<sup>1</sup> ) and wagging in crystalline cellulose (at 1318 cm−<sup>1</sup> ); and large negative loadings for bands associated with carbonyl vibration in ester and acetyl groups in hemicellulose (near 1735, 1720 and 1703 cm−<sup>1</sup> ) (Table 3). The second component, PC2, shows positive loadings for vibrations related to the molecular structure in guaiacyl and syringyl lignin moieties (1510, 1264, 1245, 1228, and 822 cm−<sup>1</sup> ), C-H deformation in lignin and carbohydrates (1461 cm−<sup>1</sup> ), C-H out of plane in cellulose (895 cm−<sup>1</sup> ), and negative loadings for O-H bond vibration of absorbed water on cellulose (near 3323 and 1655 cm−<sup>1</sup> ) (Table 3). Regarding the third and fourth components (PC3 and PC4), they are characterized by large positive loadings for methyl and methylene stretching of aliphatic structures in cellulose and in extractive compounds (2935, 2922, and 2853 cm−<sup>1</sup> ) and C-O stretching in polysaccharides (1055, 1031, 1019, and 986 cm−<sup>1</sup> ), respectively (Table 3). The fifth component, PC5, is characterized by positive loadings for methoxyl and aromatic skeletal vibrations in lignin and in extractive compounds (near 1560, 1543, and 1422 cm−<sup>1</sup> ). The last extracted component, PC6, presents positive loadings for bands associated with C-O-C stretching in cellulose (1124 and 1107 cm−<sup>1</sup> ) (Table 3). For more insight, the average relative absorption spectra of the wood species with the associated assignment for the absorption bands are presented in Figure 4.


**Table 3.** List of FTIR bands (with loading values) in relation to the extracted PCA components.

The one-way ANOVA applied to the PCA component scores enabled us to identify the components that show differences between wood samples from the three geographical areas. A summary of this statistical analysis is provided in Table 4. It appears evident that the significance of differences, based on FTIR data, between the sources of the studied samples depends on the wood species. For instance, no significant differences were observed for *I. doka*, whereas significant differences were shown by *D. oliveri* for PC4, by *K. senegalensis* for PC3 and PC4, and by *P. erinaceus* for PC5 and PC6. Overall, samples from Kita and Kéniéba showed high component scores (Figure 5). For *D. oliveri*, the samples from Kéniéba

were associated with positive scores of PC4, whereas samples from Kita and Sibi were similarly associated with negative scores of PC4. As for *K. senegalensis*, woods from Kita were associated with negative scores of PC3 and positive scores of PC4, whereas woods from Kéniéba were characterized by positive scores of PC3 and negative scores of PC4. For both PC3 and PC4, intermediate values of component scores were related to wood from Sibi, and regarding *P. erinaceus*, the woods from Kita were associated with negative scores of PC5 and PC6, whereas woods from Sibi and Kéniéba were similarly associated with positive scores of PC5 and PC6. *Forests* **2023**, *14*, x FOR PEER REVIEW 8 of 14

**Figure 4.** Average FTIR absorption spectra of the studied wood species, indicating the related bond vibrations for FTIR absorption bands. **Figure 4.** Average FTIR absorption spectra of the studied wood species, indicating the related bond vibrations for FTIR absorption bands.

The one-way ANOVA applied to the PCA component scores enabled us to identify


**Table 4.** Summary of the one-way ANOVA using component scores of PCA components.

**Wood Species PCA Component df F Sig.** 

PC1 2 0.21 0.81 PC2 2 0.41 0.67 PC3 2 0.42 0.66 PC4 2 6.99 0.01 (1) PC5 2 0.44 0.65 PC6 2 3.03 0.08

PC1 1 1.54 0.25 PC2 1 0.01 0.92 PC3 1 0.15 0.71 PC4 1 0.00 0.99 PC5 1 2.93 0.13 PC6 1 0.35 0.57

**Table 4.** Summary of the one-way ANOVA using component scores of PCA components.

*Daniellia oliveri*

*Isoberlinia doka* 


**Table 4.** *Cont.*

(1) significant differences at *p* < 0.01; and (2) significant differences at *p* < 0.05. (1) significant differences at *p* < 0.01; and (2) significant differences at *p* < 0.05.

**Figure 5.** Bar chart of the mean values of component scores of PCA factors showing significant differences between samples from the three areas; (**a**) for *D. oliveri* (*n* = 5 for Kita, *n* = 6 for Sibi and *n* = 6 for Kéniéba), (**b**) for *K. senegalensis* (*n* = 8 for Kita, *n* = 8 for Sibi and *n* = 8 for Kéniéba), and (**c**) for *P. erinaceus* (*n* = 6 for Kita, *n* = 5 for Sibi and *n* = 6 for Kéniéba); (the labels a, ab, and b refer to group classification in ascending order of the Student–Newman–Keuls post-hoc test). **Figure 5.** Bar chart of the mean values of component scores of PCA factors showing significant differences between samples from the three areas; (**a**) for *D. oliveri* (*n* = 5 for Kita, *n* = 6 for Sibi and *n* = 6 for Kéniéba), (**b**) for *K. senegalensis* (*n* = 8 for Kita, *n* = 8 for Sibi and *n* = 8 for Kéniéba), and (**c**) for *P. erinaceus* (*n* = 6 for Kita, *n* = 5 for Sibi and *n* = 6 for Kéniéba); (the labels a, ab, and b refer to group classification in ascending order of the Student–Newman–Keuls post-hoc test).

From the results described above, the FTIR signal related to the main macromolecules (cellulose, hemicellulose, and lignin) and extractive contents can be explored to evaluate the relation between the variations in wood characteristics and provenance of the studied timber wood. Again, it seems that the wood of the studied tree species was not equally

glycosidic bonds (in polysaccharides) and aliphatic structures (in polysaccharides and extractives) are the most indicative of the identified chemical variations in relation to the origin of the wood samples for *D. oliveri* and *K. senegalensis*. In fact, the variation in

From the results described above, the FTIR signal related to the main macromolecules (cellulose, hemicellulose, and lignin) and extractive contents can be explored to evaluate the relation between the variations in wood characteristics and provenance of the studied timber wood. Again, it seems that the wood of the studied tree species was not equally affected by the different local environmental conditions. These results suggest that glycosidic bonds (in polysaccharides) and aliphatic structures (in polysaccharides and extractives) are the most indicative of the identified chemical variations in relation to the origin of the wood samples for *D. oliveri* and *K. senegalensis*. In fact, the variation in polysaccharide contents in wood can also be linked to the degree of disturbance related to the environmental conditions at the tree growing location [17,51]. Thus, we could conclude that the local environmental conditions at Kéniéba and Kita have a significant influence on the polysaccharide contents of wood of *D. oliveri* and *K. senegalensis*, respectively. This result is also in agreement with what has been described earlier for the color data, for *K. senegalensis*, about the significant differences showed by the yellowness (b\*) values, which can be related to the polysaccharide contents [46,48]. Nevertheless, we should also draw attention to the fact that biochemical responses to environmental factors may vary depending on the wood species [51]. Furthermore, for *P. erinaceus*, it seems that extractive compounds with aromatic structures are also indicative of changes that differentiate between wood from different provenance areas. Previous studies on this wood species, in Togo and in Mali (West Africa), revealed that its extractive contents depend on the soil properties at the tree growing location [4,52]. Finally, these findings highlight that wood chemistry and biochemical processes of wood formation are also related to environmental factors other than climate [9,53].

#### *3.3. Variations in Wood FTIR Ratios*

The one-way ANOVA applied to the FTIR ratios dataset revealed the possibility to differentiate the wood samples of the three studied areas. The summary of these results is provided in Table 5. No significant differences were found for samples of *I. doka* and *K. senegalensis*, whereas significant differences were found for *D. oliveri* based on the guaiacyl/syringyl (GS2, 1510/1595) and lignin/carbohydrate (LC1, 1510/1460; LC2, 1510/1375; and LC3, 1510/895) ratios, and for *P. erinaceus* for all the considered ratios (Table 5). Again, as illustrated in Figure 6, samples from Kita and Kéniéba stand out by the average values of the FTIR ratios. For *D. oliveri*, samples from the areas of Kita and Sibi appeared to be very much alike with positive values of the FTIR ratios, whereas samples from the area of Kéniéba were associated with negative values of the FTIR ratios. Regarding *P. erinaceus*, all FTIR ratios except for the guaiacyl/syringyl ratio (GS1, 1265/1230) showed significantly lower values for Kita, as compared to Sibi and Kéniéba, which presented similar values. In contrast, the 1265/1230 ratio values were significantly higher for wood from Kita than for the wood from Sibi and Kéniéba.


**Table 5.** Summary of the one-way ANOVA using FTIR ratios.


**Table 5.** *Cont.*

(1) significant differences at *p* < 0.01; and (2) significant differences at *p* < 0.05. (1) significant differences at *p* < 0.01; and (2) significant differences at *p* < 0.05.

**Figure 6.** Bar chart of the mean values of FTIR ratios showing significant differences between samples from the three areas; (**a**) for *D. oliveri* (*n* = 5 for Kita, *n* = 6 for Sibi and *n* = 6 for Kéniéba), (**b**) for *P. erinaceus* (*n* = 6 for Kita, *n* = 5 for Sibi and *n* = 6 for Kéniéba); (the labels a, ab, and b refer to group classification in ascending order of the Student–Newman–Keuls post-hoc test). **Figure 6.** Bar chart of the mean values of FTIR ratios showing significant differences between samples from the three areas; (**a**) for *D. oliveri* (*n* = 5 for Kita, *n* = 6 for Sibi and *n* = 6 for Kéniéba), (**b**) for *P. erinaceus* (*n* = 6 for Kita, *n* = 5 for Sibi and *n* = 6 for Kéniéba); (the labels a, ab, and b refer to group classification in ascending order of the Student–Newman–Keuls post-hoc test).

as the lateral order index (LOI), is an empirical crystallinity index that is associated with the number of crystalline structures of cellulose [43,54]. The guaiacy/syringyl ratios (1265/1230 and 1510/1595) inform about the molecular structure of the lignin contents and the relative number of methoxyl groups in guaiacyl and syringyl nuclei [55,56]. The lignin/carbohydrate ratios (1510/1460, 1510/1375, and 1510/895) are indicative of the relative proportion of lignin and polysaccharide compounds [44,57]. Thus, according to the results of the FTIR ratios, it can be deduced that, for *D. oliveri*, the wood samples from Kéniéba are characterized by a higher proportion of syringyl than guaiacyl groups in the molecular structure of lignin compounds and the highest proportion of polysaccharide compounds. As for *P. erinaceus*, the wood samples from Kita appear to be associated with the less ordered crystalline structure in cellulose, the highest proportion of methoxyl groups in the syringyl nuclei of lignin compounds, and the lowest proportion of polysaccharide compounds. Moreover, the known correlation between the lignin/carbohydrate ratio suggests that *P. erinaceus* wood from Kita may have lower mechanical strength in comparison to *P. erinaceus* wood from Sibi and Kéniéba [58,59].

The results of the FTIR ratios for *D. oliveri* and *P. erinaceus* suggest that they can be used to assess the changes in the molecular structure of the wood cell walls macromolecules related to provenance. Each of the calculated FTIR ratios is somehow

The results of the FTIR ratios for *D. oliveri* and *P. erinaceus* suggest that they can be used to assess the changes in the molecular structure of the wood cell walls macromolecules related to provenance. Each of the calculated FTIR ratios is somehow connected to specific chemical characteristics. For instance, the ratio 1425/895, also known as the lateral order index (LOI), is an empirical crystallinity index that is associated with the number of crystalline structures of cellulose [43,54]. The guaiacy/syringyl ratios (1265/1230 and 1510/1595) inform about the molecular structure of the lignin contents and the relative number of methoxyl groups in guaiacyl and syringyl nuclei [55,56]. The lignin/carbohydrate ratios (1510/1460, 1510/1375, and 1510/895) are indicative of the relative proportion of lignin and polysaccharide compounds [44,57]. Thus, according to the results of the FTIR ratios, it can be deduced that, for *D. oliveri*, the wood samples from Kéniéba are characterized by a higher proportion of syringyl than guaiacyl groups in the molecular structure of lignin compounds and the highest proportion of polysaccharide compounds. As for *P. erinaceus*, the wood samples from Kita appear to be associated with the less ordered crystalline structure in cellulose, the highest proportion of methoxyl groups in the syringyl nuclei of lignin compounds, and the lowest proportion of polysaccharide compounds. Moreover, the known correlation between the lignin/carbohydrate ratio suggests that *P. erinaceus* wood from Kita may have lower mechanical strength in comparison to *P. erinaceus* wood from Sibi and Kéniéba [58,59].

#### **4. Conclusions**

The main conclusion of the study is that the variation in wood color and chemical properties can be related to the local environmental conditions at the tree growing locations. Since the three studied areas are under the Sudano–Guinean climatic zone, the observed variations in wood properties are most likely connected to the local soil properties and/or landforms. Additionally, it was clear that the studied tree species were not equally affected by the different local environmental conditions. *P. erinaceus* appears to be the most sensitive species, as it presented significant differences between samples from the three areas for the largest number of variables (color parameters, PCA components, and FTIR-ATR ratios) used in this research. *D. oliveri* and *K. senegalensis*, however, showed significant differences between sample provenance for a lower number of variables. *I. doka*, for its part, showed no significant differences. Furthermore, the results revealed that important differences exist between wood samples from Kati and Kéniéba; these differences are related to color parameters (redness (a\*) and hue (h◦ )), as well as a number of FTIR signals related to chemical content in the wood cell walls (polysaccharides and lignins).

Finally, from this work, we believe that other factors characterizing the local environmental conditions (e.g., soil properties) rather than climatic factors can drive significant changes in wood properties. Certainly, further research could help to determine specific links between wood properties (physical and chemical) and environmental factors (e.g., soil factors) that are not necessarily climatic.

**Author Contributions:** Conceptualization, M.T. and A.M.C.; methodology, M.T. and A.M.C.; investigation, M.T. and A.M.C.; resources, M.T. and A.M.C.; data curation, M.T. and A.M.C.; writing original draft preparation, M.T.; writing—review and editing, M.T. and A.M.C.; supervision, A.M.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research is funded by Grupos de Referencia Competitiva (ED431C 2021/32) by Xunta de Galicia. It was developed within the framework of the visiting fellowship program for African researchers (Programa BECAS ÁFRICA-MED 2021-2022) of the Spanish Agency for International Development Cooperation (AECID). MT is currently funded by Plan Galego I2C Modalidade A (ED481B-2022-017).

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** The authors would like to thank Bakary Traoré (Katibougou IPR/IFRA, Mali) who provided the wood samples. Also, the authors would like to thank the editors and reviewers for their time and insightful suggestions that helped to improve our manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest. And the funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

## **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
