Productivity and Biomass Properties of Poplar Clones Managed in Short-Rotation Culture as a Potential Fuelwood Source in Georgia

: Georgian forests are very valuable natural resources, but due to the lack of affordable alternatives to ﬁrewood, people are forced to use forest resources illegally and unsustainably. The aim of this study was to determine the productivity and biomass properties of four poplar clones from Aigeiros and Tacamahaca and one control clone, considering their wood and bark characteristics and their proportion in the stems. Short-rotation woody crops with these clones represent a potential source of commercial fuelwood production in Georgia as an alternative to natural forests. These tree characteristics were evaluated after three years of growth. The survival of the clones was generally high. No signiﬁcant differences in biomass production (dry matter, DM) were found among the four clones tested (DM of approximately 4 Mg ha − 1 yr − 1 ), while the control clone achieved signiﬁcantly lower values for DM. The biomass speciﬁc density was exceptionally high, at 481–588 kg m − 3 , which was a result of the high proportion of bark mass in the stem (23.3–37.7%), with a density almost twice that of wood. On the other hand, the tested clones had a very high ash content in the biomass (2.6–4.5%), which negatively affected their energy potential expressed as a lower heating value (17,642–17,849 J g − 1 ). Our preliminary results indicated that both the quantity and quality of biomass are important factors to justify the investment in an intensive poplar culture. The four clones should be further considered for commercial biomass production and tested at different sites in Georgia to evaluate the genotype-by-environment interactions and identify the site conditions required to justify such an investment.


Introduction
Forests in Georgia occupy about 40% of the country's territory [1]. They are exceptionally valuable natural resources at both the regional and global level [1,2]. The forest ecosystems surviving in the mountain massifs of the Georgian Caucasus are the last untouched ("virgin") forests in the temperate climate zone of the Earth [3]. As much as 95-98% of Georgian forests are of natural origin [1], and almost all of them are still naturally regenerated [2]. The species composition, structural diversity, and other characteristics of Georgian forests shape their rich biodiversity, with approximately 400 occurring tree and in breeding programs [26]. In general, most poplar breeding programs aim to develop new genotypes to achieve higher production, greater tolerance to pests and diseases, and specific wood quality [25,[27][28][29]. When considering biomass as an energy feedstock, important energy-related characteristics of interest include a favorable wood-to-bark ratio, high specific gravity (density) [30,31], and heating value [32][33][34], as well as a low ash/residue content [33] and porosity [35]. In general, despite the fact that some studies have shown that the heating value of bark is comparable to that of wood [36], in SRWC biomass, the proportion of bark mass to total harvested biomass is nevertheless considered negative. Since most minerals taken up by trees are accumulated in the bark, the ash mass fraction in the whole tree is predominantly associated with the bark fraction [37]. Therefore, the energy analysis of tree stems should be carried out considering the properties of wood and bark and their proportion in stems. These properties are influenced by several factors, including stem diameter, changes in cambium with age, genetic control, and environmental factors [37]. Previous studies [34,[36][37][38][39][40][41] on the clonal variation of many of these traits in hybrid poplars and other species of the Salicaceae family (willow) have reported a considerable variation in the bark-to-wood ratio and ash mass fraction, a moderate variation in specific density, and little variation in heating value.
Superior poplar genotypes from Aigeiros (e.g., Populus nigra, Populus deltoides) and Tacamahaca (e.g., Populus trichocarpa, Populus maximowiczii, Populus balsamifera) sections are the most commonly planted in Europe and potentially the most promising in terms of biomass production for energy purposes in Georgia. Therefore, the aim of our preliminary study was to determine the productivity and biomass characteristics considering the wood and bark properties and their proportion in the stems of five poplar clones from the Aigeiros and Tacamahaca sections as a potential source of commercial fuelwood production in Georgia.

Experimental Location, Soil, and Climatic Condition
The experiment was established in eastern Georgia in the Mtskheta Municipality (41 • 55 01.75 N 44 • 46 18.37 E, 586 m a.s.l.) in the village Jighaura, on the right bank of the Tezami River, in the spring of 2017 ( Figure 1). The long-term annual mean temperature is 11.89 • C, while the annual precipitation is 632.45 mm. The annual water balance (precipitation minus potential evapotranspiration) of the study site amounts to −410.81 mm. The detailed bioclimatic characteristics of the study site are presented in Table 1 [42]. Table 1. Bioclimatic characteristics of the study site in Jighaura (eastern Georgia) according to the EuMedClim database [42].

Climatic Variable (Unit) Value
Annual mean temperature ( • C) 11.89 Mean diurnal temperature range ( • C) 11.02 Maximal temperature of the warmest month ( • C) 29 The soil type of the experimental plot belongs to Fluvisols according to the World Reference Base for Soil Resources (IUSS Working Group WRB, 2015), which was developed based on river deposits and correlated with the national classification as an alluvial soil azonal type [43]. Fluvisols generally have low production and are used as cropland or hay meadows [43]. A soil survey conducted prior to the establishment of the experiment revealed that the soil was characterized by a very low organic matter content (less than 2% in the humic horizon), moderate alkalinity (pH 8), silty clay texture, and high stoniness index. The soil was poor in available phosphorus, rich in potassium, and had a high calcium carbonate content that increased with depth [44]. In winter 2017, the soil was plowed to a depth of 30-40 cm and superficially flattened before planting.

Planting Material and Study Design
The trial was established in early spring in 2017 using 22 cm long-dormant hardwood cuttings that had previously been soaked in water for 24 h to promote rooting.
The genetic material consisted of five clones. Three clones, 89.M.061, 89.M.004, and Kocabey (TR-77/10), were obtained from the Poplar and Fast Growing Forest Trees Research Institute, PFGFTRI in Izmit, Turkey. AF8 is a commercial clone suitable for short-rotation plantations produced by Alasia Franco Vivai in Italy (12030 Cavallermaggiore (CN)), while Populus pyramidalis Roz (syn. Populus nigra "Italica", P. italica Moench., P. fastigiata Desf.) was used as a control. We continue to refer to the name P. pyramidalis for consistency. The species was imported from Italy in the last century and showed good adaptability in Georgia. Detailed information about the genetic material used in this study is presented in Table 2.
Twenty cuttings from each poplar clone were used for planting. The cuttings were planted in a completely randomized design with a spacing of 2.0 m between rows and 0.4 m within rows, resulting in a density of 12,500 cuttings per hectare. No mineral fertilizers were used, while an optimal irrigation regime was applied to mitigate the water deficit.
Manual weeding was carried out as required during the growing seasons.

Measurement of Tree Characteristics
After three years of growth, the height and diameter at breast height (DBH; measured at a height of 1.3 m) of all trees were measured. DBH was measured using a digital caliper (Mitutoyo, type CD -15DC, UK), while tree height was measured using a calibrated pole. The survival rate was determined by the number of living trees. Aboveground biomass (dry mass, DM) production was calculated using an allometric function commonly used in tree biomass and volume modeling [45]: log 10 (DM)= β 0 + β 1 × log 10 (DBH) + β 2 log 10 (H), where DBH is the diameter at breast height; H is the height; and β 0 , β 1 , and β 2 are the parameters of the function (β 0 = −7.31111, β 1 = −0.402705, β 2 = 4.33891, R 2 = 0.958). The parameters of the function were fitted based on the aboveground biomass of sample trees taken from each genotype. The sample trees were selected for each poplar clone according to their average height and diameter in the experiment [46]. Immediately after felling, the sample trees were weighted. The two sample discs representing the given clone (discs cut at breast height, 1.3 m) were taken to the laboratory to assess the biomass properties [46,47].

Biomass Properties
All the properties of stem biomass were analyzed in the Laboratory of Production Technology and Quality Assessment of Biofuels certified by the Polish Center for Accreditation PCA (accreditation number AB1585). Analyses were conducted according to standardized procedures, the required number of replicates, and the required repeatability of the results. Since the parameters of the drying process have a significant impact on the specific density and porosity of the dried material [48], the stem samples of all clones were dried under strictly the same conditions in a laboratory dryer (SLW 115, Pol-Eko, WodzisławŚląski, Poland) at 105 • C until a constant mass was obtained. Then, the samples were divided into subsamples consisting of bark and wood of a given poplar clone. Based on the mass of bark and wood, the mass fraction (%) of wood Sw and bark Sb in the steam sample was determined. Specific density (SD), as an important physical parameter characterizing the ratio of the mass of a substance to its volume (volume including internal pores), was measured based on a displacement method [33] using a quasi-fluid pycnometer (GeoPyc 1360, Micromeritics Instrument Corp., Norcross, GA, USA). The detailed procedure for SD determination was described in our previous work [33]. Since the solid wood and bark samples consist of the cell walls and cell cavities, which contain air and small amounts of gum and other substances, we also determined their absolute density (AD). AD refers to the density of the material without internal pores. Measurements of AD were taken using helium gas for displacement with a gas pycnometer (AccuPyc II 1340 Micromeritics Instrument Corp., Norcross, GA, USA). Based on SD and AD, the porosity p of the samples was calculated according to the equation: For the proximate analysis of wood and bark, the samples were prepared according to the EN ISO 14 780:2017 standard [49]. The samples (of wood and bark) were ground to a particle size of <1 mm in a laboratory mill (PX-MFC 90D Polymix, Kinematika, Lucerne, Switzerland). Subsequently, the determination of the ash mass fraction A d was performed according to EN ISO 18122:2015 [50] in the muffle furnace (FCF 7SM Czylok, Jastrzebie-Zdrój, Poland) and the heating value according to EN ISO 18125:2017 [51] using a calorimeter (IKA C6000, IKA-Werke, Staufen, Germany). The heating value, also called the calorific value, of the biomass was defined by the higher heating value (HHV) and lower heating value (LHV). The HHV is essentially the energy content on a dry basis and indicates the maximum amount of heat that a material can produce when burnt completely [52]. Unlike HHV, the LHV takes into account energy losses, such as the energy used to vaporize water [52]. The difference between these values is practically constant when converted to the dry state and depends mainly on the hydrogen content in the fuel [53]. Therefore, we report both values while in the statistical analysis, and in the following, we refer only to the LHV since this value is more applicable in practice.
The relative self-ignition temperature (TSI) of the samples was determined according to the guidelines of Commission Regulation (EC) No. 440/2008 [54]. The TSI determination was carried out in the temperature-programmed laboratory oven (FCF 2R/TZ-BM Czylok, Jastrzębie-Zdrój, Poland). The test was performed until the sample temperature reached 400 • C, then the TSI was calculated. Details of the procedure are described elsewhere [31]. The synthetic information about the methods, standards, and equipment used for the assessment of biomass quality parameters followed by the repeatability and errors of the measurements are presented in Table 3. Table 3. Standards followed for biomass quality parameters and analytical equipment used with information on measurement accuracy.

Statistical Analysis
To test for variation among the examined clones in terms of mean growth and biomass traits, linear models were implemented. The mathematical model for the single-factor completely randomized design was applied for growth traits (DBH, H, DM), which was appropriate with the study layout: where x ij is the dependent variable, µ is the mean, α i is the effect of clone i, and ε ij is the random error. Meanwhile, linear models with the interaction effect were applied to assess differences in the biomass properties (SD, A d , LHV, p). In the latter analysis, we considered the effect of clone, material (bark, wood), and interaction between the clone and material: where x ijk is the dependent variable, µ is the mean, α i is the effect of clone i, β j is the effect of material j, (αβ) ij is the effect of interaction between clone i and material j, and ε ijk is the random error. Generally, the analyses for biomass properties were performed separately for wood and bark characteristics; however, since it was unlikely that the feedstock would be de-barked in commercial operations, we also calculated the average values for stem characteristics as a weighted average of the wood and bark content in the stem.
When the ANOVA indicated a significant effect (p ≤ 0.05), post hoc comparisons between clones, materials, or interactions between clone and material were performed using a least significant difference (LSD) test.
Pearson correlation coefficients were used to test the relationships between the measured characteristics, while principal component analysis (PCA) was used to further explore the variation and determine patterns in the dataset. In the PCA, the multidimensional space was reduced to eigenvectors (the principal components) corresponding to the eigenvalues that explained the largest part of the variation in the dataset. In this analysis, the eigenvectors were calculated based on the constructed correlation matrix. All the statistical analyses were performed with Statistica (data analysis software system), version 13 [55].

Survival and Growth Parameters of Clones
The survival rate of the poplar clones was generally high. Clone 89.M.061 had the lowest survival rate among the tested genotypes (90%), while two clones, Kocabey and AF8, recorded a 100% survival rate at the age of three. The average diameters of the various genotypes were statistically significantly different (p < 0.05). Clone 89.M.061 had the highest mean DBH value (28.4 mm), which was significantly greater than the DBH values of all the other clones (p < 0.05 in the LSD test, Figure 2a). In contrast, the control clone of P. pyramidalis had the lowest DBH, at 11.6 mm, which was almost 2.5-fold smaller than for the thickest clone.
The heights of trees of various clones were not as diverse as in the case of DBH values. The four tested clones differed significantly in height (298.6-320.1 cm) from P. pyramidalis, which produced the shortest trees (200.3 cm; p < 0.05 in the LSD test, Figure 2b).
The dry biomass, estimated on the basis of the allometric equation, showed significant differences among clones. Biomass production data were calculated on an area basis, taking into account the survival rate, and did not show statistically different results for the three clones: 89.M.004, AF8, and Kocabey. The 89.M.061 had a slightly lower biomass production but did not considerably differ from that of Kocabey. The largest difference in DM productivity was exhibited between the tested clones and control (p < 0.05 in the LSD test, Figure 2c). The mean annual increment (MAI) of dry mass reflected the results for DM per ha and ranged between 2.96 (89.M.061) and 4.14 Mg ha −1 yr −1 (89.M.004). In contrast, the MAI produced by the P. pyramidalis achieved a value of 0.78 Mg ha −1 yr −1 (Figure 2d).

Biomass Properties
All the biomass properties were generally influenced by the wood-to-bark ratio in the stem. The proportion of bark in the stem of the different clones varied from 23.3% in Kocabey to 37.7% in P. pyramidalis (Table 4). Overall, the specific density of stem biomass (wood and bark) was high and ranged from 481 kg m −3 (89.M.061) to 588 kg m −3 (AF8). There were significant differences (p < 0.05) between the wood and bark SD (Figure 3). The wood density was 422 kg m −3 , while the bark density was almost twice as high, at 817 kg m −3 . Therefore, the bark content resulted in higher mean density values in the clones with a greater ratio of bark to wood in the stem, e.g., P. pyramidalis. These results were in contrast to porosity, which was lowest for the biomass of clone AF8 and highest for that of 89.M.061. The mean porosity obtained for wood was 71.1%, and for bark it was 43.7%. The obtained results for porosity were significantly different between these two components of the stem. The ash mass fraction in the stem biomass was generally high and ranged from 2.6% to 4.5%. The lowest ash mass fraction was observed in the biomass of AF8 and gradually increased with the biomass of the Turkish clones until it reached the level of P. pyramidalis, which had the highest ash mass fraction. Similarly to density, the ash mass fraction was considerably higher in the bark (9.5%) than in the wood (1.1%).   The highest HHV among all the tested clones was achieved by the clone AF8 (19,103 J g −1 ), while the lowest was achieved by P. pyramidalis (18,922 J g −1 ). The LHV results were in agreement with the values obtained for the HHV of the clones and ranged from 17,642 J g −1 (P. pyramidalis) to 17,849 J g −1 (AF8). In general, wood was characterized by a significantly higher LHV (18,205 J g −1 ) than bark (16,508 J g −1 ). The overall mean values for clones are presented in Table 4, while the detailed data on all the biomass characteristics obtained with the LSD test are shown in Figure 3.

Relationship between Traits
The Pearson correlation coefficients showed that the tree height (H) was strongly and positively correlated with DBH, while strong negative correlations were observed between H and bark share (Sb) or ash mass fraction (A d ). Similarly, DBH was negatively correlated with Sb and specific density (SD). The correlation matrix also showed that the trees with a high bark mass fraction in the stem had a higher ash mass fraction, which also negatively affected the heating value (LHV; HHV was excluded from the analysis as explained in Materials and Methods). Table 5 summarizes the correlation coefficients between the studied traits. Wood mass fraction and porosity were removed from the analysis because they are the converse of bark mass fraction and specific density, respectively. The results of the PCA are shown in Table 6 and Figure 4. The eigenvalues calculated, based on the correlation matrix, showed the dominance of the first principal component, which explained 52.52% (eigenvalue = 4.2) of the variance and was mainly associated with H, Sb, A d , and LHV, while the second PC was mainly associated with DBH, S, TSI, and SD, and explained another 37.63% (eigenvalue = 3.0) of the variance. Overall, the first two components explained a large part of the variance, amounting to 90.15% The locations of the clones in the PCA ordination space showed very different patterns in terms of growth and biomass properties ( Figure 1). Generally, however, clones AF8 and in laser extend Kocabey demonstrated a higher survival rate, TSI, and heating value, with a lower ash mass fraction. The 89.M.004 clone demonstrated high DBH and H but a lower SD. Clone 89.M.061 had, desirable, a high DBH but also a higher ash mass fraction and lower SD than the other clones. The reference clone P. pyramidalis was distinguished by showing the lowest productivity traits (DBH and H) and the highest share of bark in the stem among all the selected clones, as well as a greater ash mass fraction and specific density.

Discussion
High biomass production in SRWC plantations depends on an optimum combination of climatic and soil physical conditions, the availability of moisture during the growing season, the availability of nutrients, the aeration of the soil and suitable species, and clone selection [12]. Thus, the highest yields of 20-25 Mg ha −1 yr −1 are expected to be achieved under suitable climatic conditions, with additional optimization of soil fertility and moisture through fertilization and irrigation. Geographically, most records of the highest productivity of poplars in SRWCs are reported from southern Europe [56][57][58][59] and more southern locations in North America [60][61][62]. The results of several studies also highlight the notable role of clone selection by showing large differences in productivity between clones under the same culture, which can range from less than 1 to more than 25 Mg ha −1 yr −1 [63,64].
This study showed a relatively uniform level of biomass production among the tested poplar clones at the study site in eastern Georgia. The only exception was the control, P. pyramidalis, which differed significantly from all other clones in terms of biometric traits and achieved the lowest growth parameters. During the three-year period of our study, the highest DBH was achieved by clone 89.M.061, while the highest trees were produced by 89.M.004. The highest biomass production (DM) was achieved by the two clones, 89.M.004 followed by AF8. However, the MAI for these clones was not high and amounted to approximately 4 Mg ha −1 yr −1 . The following two clones, Kocabay and 89.M.061, were characterized by an MAI at a level of about 3 Mg ha −1 yr −1 . The lowest MAI was produced by the control clone P. pyramidalis (below 1 Mg ha −1 yr −1 ).
Previous studies in the Black Sea region-with similar climatic conditions to those in our study site-showed the high potential (promising results) of all Turkish clones investigated in this study. They were all characterized by significantly higher biomass production than the control clones at the experimental sites established in central Anatolia, south-east Anatolia, and east Anatolia [20]. The black poplar clone Kocabey, native to the region, was already registered by the National Poplar Commission (Turkey), while 89.M.004 achieved an MAI of 62 m 3 ha −1 yr −1 in first-stage clone trials in the Black Sea region [20]. The MAI produced by 89.M.004 was more than twice that of the control clone and can be roughly converted to an MAI of DM of approximately 24 Mg ha −1 yr −1 . Similarly, the AF8 clone has shown a high biomass production in many trials and in commercial plantations in the Mediterranean region, with an MAI of about 19-24 Mg ha −1 yr −1 [25,32,65]. The low productivity of the investigated clones in the current study may, therefore, suggest unsuitable growth conditions.
Optimal growth conditions for hybrid poplar cultivation have been studied extensively in Italy [65]. According to these studies, the optimal annual average temperature may be in a wide range from 8.5 to 17 • C, with summer precipitation of 100-150 mm. These ranges of climatic parameters indicate that our study site falls within limits for optimal climatic conditions. Regarding soil properties, a suitable soil pH for short-rotation poplar cultivation should fall within the range of 5.5 to 7.5. Soils with either a highly acidic or alkaline pH or poorly or excessively drained; infertile; stony; shallow; or degraded by erosion, compaction, or salinization reduce the productivity required to achieve satisfactory results in cultivation [12]. Land suitability analyses for SRWC have defined the threshold of potential annual biomass productivity as more than 8 Mg ha −1 yr −1 for suitable land [32,66,67]. The soil conditions at the study site, with a pH of 8, a low organic matter content, and high stoniness, exceeded suitable soil conditions, resulting in the low productivity of the clones, which was far too low for sustainable biomass production on a commercial scale. In order to obtain reliable results for the productivity of the currently tested clones, we propose that further experimental phases should focus on establishing multiple trials under different site conditions. Further experimental phases would certainly limit the pool of potentially useful sites for poplar cultivation in Georgia but would help to achieve the high productivity required to justify an investment in intensive culture. As a supplementary to the above, we suggest the implementation of soil fertilization.
Regarding the biomass properties, they showed, as expected, a considerable variation between clones in terms of bark-to-wood ratio and ash mass fraction, but also quite a large variation in terms of specific density and porosity and a low variation in the lower heating value.
Specific density is an important parameter in assessing the quality of feedstocks for bioenergy production. High densities give a higher energy output on a volume basis and help to reduce transportation costs [37]. The specific density values for the poplar clones in the experiment were exceptionally high (481-588 kg m −3 ), much higher than the wood densities of the 3-year-old poplar clones in the study of Tharakan et al. (330-370 kg m −3 ) [37] or the 10-year-old poplar clones in continental Canada (325-366 kg m −3 ) [29].
From a biomass production perspective, selecting clones that combine a high wood density with a high volume of growth maximizes the production and energy output on an area basis [37]. Our preliminary results showed that the best combination of productivity and density was achieved for the clone AF8, but all other clones tested in our study were also characterized by a high wood and bark density.
The exceptionally high values of specific densities for biomass raw materials obtained in our study resulted directly from the fact that all the clones were characterized by a relatively high bark mass fraction in the stem, ranging from 23.3% to 37.7%, with a density almost twice that of wood.
In general, the bark mass fraction results for the sample trees, which reached DBH values of 11-28 mm in our study, were comparable to results obtained in other studies on poplar clones from similar DBH ranges [68]. Bark mass fraction is known to decrease with increasing DBH and age [34,36,68]. As described by Guidi et al. [68] for 2-year-old shoots of Populus deltoids L., the bark mass fraction gradually decreased from 35.8% to 7.1% in the 10-90 mm diameter range. The stabilization of bark mass fraction (less than 8%) was visible when shoots reached 60 mm in diameter [68]. In the present study, we also found a clear negative correlation between bark mass fraction and growth traits DBH and H, suggesting that rotation length should be extended if the stabilization of the wood-to-bark ratio is deemed desirable.
Although bark was characterized by a very high SD in our study, it should be emphasized that the bark mass fraction in biomass tends to be considered a negative characteristic of the energy feedstock. Since most of the minerals taken up by trees are accumulated in the bark, the ash content in the whole tree is highly dependent on the bark mass fraction [33]. Consequently, a high ash mass fraction in the biomass feedstock negatively affects the heating value.
The heating value of the biomass as a bioenergy feedstock of the investigated clones was comparable to publications on hybrid poplars [32,33,69], furthermore to the heating value of other tree species [70], and suggested a similar or even better energetic potential compared to waste biomass (bio-waste) from the agri-food industry [8,51,[70][71][72]. As expected, a high variation in the heating value of the biomass of the poplar clones was not detected. The difference in the heating value mainly arose from the different share and composition of bark and wood.
The lower heating value of the tested clones for wood was in the range of 18,100-18,300 J g −1 and was higher than the LHV determined for bark: 16,000-17,000 J g −1 .
Overall, when considering woody biomass as a bioenergy feedstock according to EN ISO 17225-1:2014 [73], typical LHV values of bark and wood from various tree species are comparable for bark (19,(0)(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)200 J g −1 ) and for wood (18,,100 J g −1 ). However, the range of variation in these values is much larger for bark (17,100-21,300 J g −1 ) than for wood (18,,800 J g −1 ) [73]. The bark contains less cellulose and more lignin and extractives than wood [34,74]. For this reason, the quantities of these components relative to the extractives and ash mass fraction of the biomass have an impact on the combustion temperatures and energy potential of the feedstocks [34]. As our results showed, the energy potential of poplar bark decreased with increasing ash mass fraction (r = −0.72). The ash content of the bark was notably high, ranging from 7.6% (clone AF8) to 11.4% (clone 89.M.061). According to EN ISO 17225-1:2014 [73], the typical ash content of bark should fall within the range of 0.8-5%, which is consistent with several empirical studies conducted on different tree species [36,37,[75][76][77], although higher values of ash were also found [46,75,78]. Nevertheless, the exceptionally high ash content obtained in the studied poplar clones prompted us to consider the specific factors that influenced these results. It seems likely that this phenomenon can be explained in part by the environmental conditions of the study area. The dry, windy climate may cause mineral dust to be deposited on the bark surface, which is not adequately removed by the low rainfall. As a result, the high ash mass fraction in the bark reduces the energy potential of whole stems. Anderson and Zsuffa [79], in turn, draw attention to the possible presence of a genotype-environment interaction with respect to nutrient availability.
Despite the differences in the LHV and ash content between wood and bark in the present study, the LHVs of whole stems for different clones were comparable and ranged from 17,600 to 17,800 J g −1 . The values obtained were lower than the typical value for poplar of 18,400 J g −1 according to EN ISO 17225-1:2014. This result was undoubtedly influenced by the high bark content in the biomass and the high ash content in the bark.
In this context, it should be noted that bark fuel contains silicon oxide (SiO 2 ) and alkali metals such as potassium (K), sodium (Na), calcium (Ca), and magnesium (Mg), which determine the behavior of ash during combustion. Low melting temperatures can form very "sticky" silicate melts that occur during the combustion process and form deposits on the burner grate [36,75]. Bark can form slag at 850 • C, while the actual combustion process takes place at a temperature of approximately 1100 • C [80]. Therefore, due to the different chemical properties of bark and wood, the bark mass fraction should not be ignored when assessing the overall fuel quality of crops.
Regarding the clones tested, the Kocabey clone had the lowest bark-to-wood ratio, while the most desirable overall biomass properties were obtained by the AF8 clone, which had not only a desirable high wood and bark density but also lower ash content and a higher heating value than all the other clones. Similar to the productivity results, the control clone was characterized by inferior biomass feedstock properties compared to the remaining clones.

Conclusions
In this study, the potential biomass yield and biomass energy properties of poplar clones in SRWC under the specific climate and environmental conditions of eastern Georgia were recognized for the first time to contribute to solving the energy deficit in the country. Our preliminary results indicated that both the quantity and quality of biomass are important factors in justifying the investment in an intensive poplar culture. None of the clones tested in this study site had all the desired quantitative and qualitative characteristics, although all four clones tested performed better than the control clone. The productivity of the tested clones at the current stage of the experiment was far too low to justify the investment in intensive culture, while the biomass properties of the tested clones as an energy feedstock were generally slightly below the normalized values. The tested clones had desirable, exceptionally high densities on the one hand and exceptionally high ash content in the biomass, on the other hand, which negatively affected the energy potential of these clones. In addition, we showed that a high bark mass fraction in the stems had a negative effect on the quality of the biomass feedstock for energy purposes. Bark content can be controlled by appropriate clone selection and silvicultural means. To improve biomass as an energy feedstock, we, therefore, propose an extension of the rotation length, which should favor larger stem dimensions and, thus, a lower and more stable bark mass fraction in the stems.
We also pointed out a number of potential difficulties that Georgia will face in preparing to invest in biomass energy plantations. Future research should focus on the currently tested clones, perhaps enriched with commercially available clones that perform well under similar climatic conditions, and test them at multiple sites in Georgia to further identify the potential for genotype-by-environment interactions, which will enable the selection of stable clones for operational use. In addition, it would be beneficial to conduct experiments to analyze the effects of fertilization and irrigation on the growth of poplar clones at different test sites.

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

Acknowledgments:
We would like to thank Omar Kiguradze (Georgian Technical University), Koba Chiburdanidze (Agricultural University of Georgia), and Fabiano Bertaina (Biopoplar S.R.L.) for their technical support during the experiment. We also thank Piotr Mroczek from the Forest Research Institute (Poland) for preparing a map.

Conflicts of Interest:
The authors declare no conflict of interest. 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.