GROWTH AND ENERGY QUALITY OF EUCALYPTUS WOOD IN DIFFERENT CROP-LIVESTOCK-FOREST SPATIAL ARRANGEMENTS

The aim of the present study was to assess the growth and energetic properties of genetic material of eucalyptus grown in several integrated crop-livestock system spatial arrangements in Barra-do-Garças City (MT). The genetic material of 76-month old Eucalyptus spp from eight spatial arrangements were assessed according to the following properties: diameter, height and volume growth, mortality rate, dry matter yield, sapwood to heartwood ratio, form factor, bark volume, basic density, superior calorific value, ash content, volatile matter, fixed carbon and trunk energy. Data analysis was performed by adjusting growth function, test for model identity and parameters equality and ANOVA tests to Scott-Knott’s mean clustering test. Eucalyptus grandis x E. urophylla clones from recorded the highest energy efficiency due to their high yield volume, which was decisive to determine estimated energy available per tree and cropland area. Growth, stem shape and basic density varied among, and within, the same genetic materials, depending on their respective spatial arrangement. However, solid biomass properties (high calorific value, ash content, fixed carbon and volatile matter) remained stable.


INTRODUCTION
The afforestation of conventional cropland, known as integrated crop-livestock-forest systems (CLFS) or agroforestry systems, is one of the techniques embraced by the National Plan for Brazilian Low Carbon Emissions in Agriculture (ABC Plan). These systems are part of the commitment to reduce greenhouse gas emissions assumed by Brazil at the 15th Conference of the Parties (COP 15) held in Copenhagen, back in 2009. The ABC Plan aims at reducing greenhouse gas emissions in agriculture, in order to improve natural resource management and adapt agriculture to climate change (MAPA 2018).
In recent years, there has been a considerable increase in the CLFS adoption in the country, totaling 11.5 million hectares, and Mato Grosso, with 1.5 million hectares, represents 13% of the national total (EMBRAPA 2018). However, in Mato Grosso, crop-livestock integration covers 89% of this area, with the forestcomponent being present in only 11% of the area occupied (GIL et al., 2015).
Yet, there are economic, technical and cultural barriers to the integration of trees into conventional agriculture. Forest integration requires a great deal of capital, time, labor, training and market knowledge. Moreover, it is a risky strategy that may trigger insecurity, as the integration of unsuitable tree species into cropping systems and/or spatial arrangements most likely lead to trading issues and to decreased crop yield (DIAS-FILHO;FERREIRA, 2008).
However, increasing the area planted with forest species in CLFS would be desirable due to the greater land use efficiency, environmental benefits provided by trees and mutual and complementary benefits, since FLORESTA, Curitiba, PR, v. 50, n. 4, p. 1707-1716, out/dez 2020. Tonini, H. et.al. ISSN eletrônico 1982 10.5380/rf.v50 i4. 59650 1708 agriculture and livestock can cover the negative cash flow that occurs during the forest investment maturity period (PORFÍRIO-DA-SILVA, 2014) Employment of biomass-energy alternatives to grain drying, slaughterhouses and ceramics industries in Mato Grosso State has increased firewood yield by 194% in the last ten years. As a result of it, eucalyptus plantations increased from 37.392 ha in 2006 to 212.815 ha in 2015, thus becomingthe forest genus with the largest planted area (FAMATO 2013).
Further scientific research is needed to help farmers from Mato Grosso State decide exactly what eucalyptus species or clones to plant, and how to plant them, for bioenergy production. Nevertheless, properties rather than growth and bioenergy productionsuch as basic density, calorific value, dry matter yield and biomass should be assessed, as they can be altered by the genetic material, age, area and intercropping. Thus, The aim of the present study was to assess the growth and energetic properties of genetic material of eucalyptus grown in several integrated crop-livestock system designs in Barra-do-Garças City (MT)..

MATERIALS AND METHODS
Data was collected in the crop fields of Embrapa in partnership with Agropecuária Fazenda Brasil, in Barra-do-Garças City (MT), at coordinates 14° 59'25.34"S and 52°16'21.05". The city is located in Southeastern Mato Grosso State (Figure 1), in a transition area between the Amazon and Cerrado biomes (MARACAHIPES et al., 2011). Tropical Savanna (Aw) is the local climate, with two well-defined seasons: dry (May to September) and rainy (October to April) (SOUZA et al., 2013). Historically, total annual rainfall in Barra-do-Garças City ranged from 1200 to 2000 mm (SOUZA et al., 2013). Data from INMET (2018) indicated that it ranged from 958.2mm to 1907.8 mm during the study's experiment; furthermore, 2013 and 2017 recorded the highest and lowest rainfall rates, respectively. The highest rainfall was recorded from October to March of 2018 (91.9% of the total), with six dry months, which record rainfall lower than 60 mm, on average. The predominant soil in the city is the Dystrophic Red-Yellow Latosolmedium-textured, well drained and flat (MARACAHIPES et al., 2011).
The CLFS was established over a total area of 100 hectares in December 2010. However, the assessed systems evaluated totaled 49.6 hectares (Table 1). Soybean (Glycine max L.) crops were grown together with trees in the first five years; then, brachiaria grass (Brachiaria ruziziensis Satrf.) was introduced to the system. Eucalyptus seedlings were planted after land demarcation, there was crop desiccation and post-emergent herbicide application to soybeans. Pre-sowing fertilization of forests species was performed by applying 70 kg ha -1 triple superphosphate (approximately 28 kg ha -1 P2O5) and 100 g hole -1 NPK fertilizer (6:30:6) to each planting hole. Three months later, top-dressing was performed by manually applying 12:0:24 NPK added with 1.15% Boron, at dose of 110 g plant -1 , to the crown cover area.
Five eucalyptus treatments were assessed: two Eucalyptus grandis x Eucalyptus urophylla clones (CL1 and CL2; trade names: H13 and GG100, respectively); two hybrids Eucalyptus urophylla x Eucalyptus camaldulensis (URC) and Eucalyptus grandis x Eucalyptus camaldulensis (GRC) and Eucalyptus camaldulensis (CAM), Acacia mangium (AMG) and Tectona grandis (TEC) from seedlings planted in eight spatial arrangements (Table 1). Test sample selection was based on tolerance to drought and on local nursery availability.
The spatial arrangements A5 to A8 grew mixed crops (Table 1): A5 consisted of Eucalyptus camaldulensis and Acacia mangium at 1:1 ratio (50% eucalyptus and 50% acacia); A6 and A8 had two Eucalyptus grandis x Eucalyptus urophylla clones (CL1 and CL2); A7 had CL1 and Tectona grandis (teak). FLORESTA, Curitiba, PR, v. 50, n. 4, p. 1707-1716  Data on diameter at breast height (DBH) growth (cm), height (m), individual volume (m 3 ) and mortality (%) was collected from each arrangement. DBH and height growth data was collected from 18 permanent sample plots, in the center of two middle rows to avoid edge effects. The minimum number of 30 plants per sample was grown in spatial arrangements with plots ranging from 114 to 360 m 2 . Plots were measured four times: At 31, 51, 63 and 76-old months.
Individual volume, bark volume (%), form factor, heartwood to sapwood ratio (H:S) and biomass properties were obtained from thirty trees; three of them were selected per arrangement and species. Trees in the middle rows were selected under the following criteria: quadratic mean DBH, good plant health status,good crown form and exposure to sunlight (North and South in A2, A3 and A4, North, South and central in A1 and A5).
Sample trees were felled at the age of 76 months. Mean individual trunk volume was calculated by the Smalian's formula. Tree mortality was obtained from counting failures and dead trees in all rows. Data was converted as follows: √ 100 . Basic density (g/cm -3 ), heartwood and sapwood ratio (H:S) and bark volume percentage were determined from approximately 2.5 cm-thick discs from cross-sections extracted at base, DBH and at 25%, 50%, 70% and 90% of the total height. Basic density was determined through mean volume-weighted, based on Trugilho (2009), under standard NBR 11941-02. Calorific value was determined using a bomb calorimeter (standard NBR 8633/84); quick analysis was performed in muffle furnace (standard NBR 8112/83).
Heartwood:sapwood ratio was obtained by measuring four radii in each cross section with the aid of a millimeter-graded ruler, based on Pereira et al. (2013). Stem taper was calculated through artificial form factor and stem volume (%), based on Finger (1992). The individual dry mass was obtained by multiplying the trunk volume by the basic density and the energy content per tree (KW.h), multiplying the dry mass by the superior calorific power, according to Santos et al. (2011).
Diameter and height growth curves were calculated using Chapman-Richards growth functions optimized by the Levenberg-Marquardt algorithm. Variables were expressed as follows: Wherein: Y, dependent variable (diameter at breast height in cm, or total height in m); A, asymptote; K, growth coefficient; R, shape coefficient; T, age (in months).
Good fit was evaluated based on the coefficient of determination, standard error of estimates (%) and residual standard error. Significance of estimated coefficients was tested through the asymptotic 95% confidence interval. Coefficients were considered insignificant if the interval did not include zero.
After model fitting, the likelihood-ratio chi-squared test was performed to assess the equality of coefficients in growth curves (REGAZZI; SILVA, 2004). The hypotheses were: The decision rule for this test is that H0 at 1% significance level should be rejected if ϰ 2 calculated ≥ ϰ 2 α(υ), wherein υ = degrees of freedom (ρΩρω) Wherein: N, number of observations; Nl, Neperian logarithm; SSR Ω, total sum of squares; RSS ω, residual sum of squares; ρΩ , number of estimated parameters in the full model; ρω, number of estimated parameters in the reduced model.
The resulting dataset was subjected to Kolmogorov-Smirnov Test for Normality and Levene's Homogeneity of Variance Test. Correlations among genetic materials and spatial arrangements were calculated by the growth curve model and by ANOVA, based on a completely randomized design applied to ten treatments (eucalyptus genetic materials and spatial arrangements), with three replications (trees). Means among treatments were compared through Scott-Knott's test, and means among controls (clones x hybrids) were compared through t-test,both at 95% confidence interval.

RESULTS
A common equation could not fit the CAM species' DHB data, since H0 5 should not be accepted (Table  2 and Figure 2 A) with restriction in the parametric space for all parameters of the model. H0 5 was not significant for height growth, thus it was calculated through a common equation (Table 2 and Figure 2 B). H0 5 was rejected for both variables of CL2, thus it proved the effect of spatial arrangements on diameter at breast height (DHB) growth and height (Table 2), as shown by the growth curve model (Figures 2C and 2D).
DBH growth curves between CL1's A6 and A7 was not significant according to H0 5 and data of A1, A6 and A7 can be grouped for height comparison purposes. H0 4 indicated that A8 differed from the others and had parametric space for equality between parameters A and R restricted. Identifiability analysis results were graphically portrayed through the growth curve model (Figures 2E and 2F). Table 2. Likelihood-ratio chi-squared test (ϰ 2 ) performed to assess coefficient equality between diameter (DBH) and height (H). Tabela 2. Teste de razão de verossimilhança com aproximação pelo qui-quadrado (ϰ 2 ) para avaliar a identidade de parâmetros para diâmetro (DAP) e altura (H).  Figure 2. Growth curve of diameter at breast height (DBH) and height of Eucalyptus camaldulensis (A and B), clone 2 (C and D) and clone 1 (E and F). Identity test analysis provided equations and estimated parameters for all the assessed genetic materials and spatial arrangements (Table and Figure 3). Statistical adjustment for R 2 and Syx% ranged from 0.30 to 0.77 and from 10.5% to 27.4% for DBH, and from 0.28 to 0.81 and 9.8% to 22.6% for height. All coefficients were significant at 5% level and clones presented better fit statistics.    Mean annual increment in diameter ranged from 2.8 to 4.3 cm at 76-month old trees; CL1 and CL2 (A6 and A7) had the highest rates (Table 4 and Figure 3A), whereas CAM (A4 and A5) had the lowest ones (Table 4 and Figure 3 B). Mean annual increment in total height ranged from 2.6 to 4.3 m; CL1 and CL2 (A6, A7 and A8) had the highest rates, whereas CAM (A4 and A5) had the lowest ones (Table 4 and Figures 3 C and D).
Total individual volume and form factor have varied depending on the genetic material and on the spatial arrangements (Table 4). Individual volume ranged from 0.220 to 0.565 m 3 ; CL1 and CL2 (A6 and A7) presented the highest rates, whereas CAM (A4 and A5) presented the lowest ones. Form factor ranged from 0.35 to 0.49 in the two groups: (I) GRC, CL1 and CL2 in A3, A6 and A8 (0.35 to 0.39, 0.37, on average); (II) CL1, URC and CAM in A1, A2, A4, A5, A7 and A8 (0.41 to 0.49, 0.44, on average).

DISCUSSION
Overall, clones grown in large vital growth space recorded the highest rates of individual volume and diameter growth. CL1 (A6 and A7) and CL2 (teak) recorded similar rates due to CL1's little competition for sunlight against the slow-growing teak (CL2), as identified by the growth curve model (Table 3) and by ANOVA (Table 5). Height difference between 76-months old CL1 and CL2 was 11.6 m.
There was no statistical difference between the survival rates of clones and seedling materials (T(8) = 0.17; p = 0.43). Therefore, the high efficiency rate of cloned plants can be associated with high growth and trees homogeneity rates, since these plants recorded the lowest diameter and height coefficients of variation at the age of 76 months (24.08% and 17.5% for seedlings; 10.9% and 9.02% for clones , respectively).
Mean mortality rate (30.8%) and its variance (4.3 to 53%) were higher than those of studies by Franchini et al. (2014) and Nieri et al. (2018), who found rates ranging from 0 to 23% in different Eucalyptus and Corymbia plant species grown in crop-livestock-forest systems. This variable proves the establishment potential / adaptability of plants to the current ecological status of croplands (MACEDO et al., 2010), mainly, to limited climatic conditions, such as low humidity and 6 month long dry seasons observed in this region Spatial arrangements had little effect on plant height growth, as no significant difference was detected through the following analyses: Growth curve of CL1 among A1, A6 and A7 (Table 2 and Figure 3C); ANOVA of CL1 and CL2 among A6, A7 and A8. Therefore, the gains in volumetric increase were due to the greater expansion in basal area in the spatial arrangements with the largest growth vital space available to plants..
The aforementioned effect was highlighted by CL1 in spatial arrangements 1 to 8: A1 presented lower initial density and smaller lgrowth vital space than A8 (Table 1); it also grew thinner, presented lower trees, with lower individual volume and dry matter. This finding corroborates the study by Henskens et al. (2001), who found that volume and yield are directly affected by density and tree spatial patterns. The authors have demonstrated that isolated trees are more efficient in converting sunlight into biomass through photosynthesis, because of their large leaf area, intercepts more radiant energy.
CAM wood growth and properties (except for basic density) were not affected by the spatial arrangements. This finding suggests that there was little competition between CAM and Acacia mangium. Accordingly, Oliveira et al. (2015) found that A. mangium is oftentimes quickly suppressed by Eucalyptus when these species are grown together. CAM (18.07%) and CL1, in A6, have recorded higher bark volume (%), which varied among spatial arrangements. High bark volume (%) is unsuitable for bioenergy and coal generation, since its mineral content can generate greater amounts of ash in the burning process. This ash is corrosive and accumulates in the furnace, thus, it forces intermittent pauses during the process for machine cleaning (ANDRADE et al., 2013).
Basic density (0.47 to 0.61 g/ cm -3 ) recorded values below those recommended for direct burning (VALE et al., 2002). Authors suggest that the ideal mean ratio between semi-hardwood and hardwood must range from 0.65 to 0.80 g/cm -3 . Basic density varied among different species and within the same species in plant spatial arrangements For CL1 and CL2, the increase in the number of trees per unit area and the reduction of growth vital space reduced the diameter growth and basic density. This finding corroborates studies by Sereghetti et al. (2015) and Rocha et al. (2016), who found low basic density in small spacings of Eucalyptus grandis x Eucalyptus urophylla and Eucalyptus grandis x Eucalyptus camaldulensis clones, respectively. For CAM, there was no statistically significant