Ajuste de modelos para la estimación de biomasa y carbono en plantaciones de Eucalyptus dunnii

The commercial plantations of Eucalyptus dunnii has increased in surface through all the national territory, which was promoted by the raise in the demand of the raw material for pulp production. These plantations are characterized by high primary productivity, accumulating biomass and promoting carbon sequester. Nonetheless, the direct quantification of biomass and carbon stocks presents limitations due to the high operational effort and time needed. The adjustment of predictive models is an efficient form for obtaining relevant information for the sustainable management of forest plantations. This study aims to obtain models for estimating biomass and to quantify the carbon stock, in populations of E. dunnii at ages of 3 and 7 years old on soils CONEAT 2. Based on these characteristics 4 stratums were defined, in which 48 forest inventory plots were measured, distributed in equal number among the stratums. The data obtained was used for selecting 58 trees for destructive evaluation, sectioning the aerial biomass in stem (SB), thick branches (TcB), thin branches (TnB), leaves (LB) and total (TB). Initially, independent nonlinear regression models were fitted for fractions and total biomass, considering both ages separately. The models selected were finally adjusted in simultaneous for obtaining additive models for the fractions through the process of nonlinear seemingly unrelated regression (NSUR). The quantification of carbon stocks from populations were realized using the aerial biomass estimated, and carbon concentrations obtained with individual fractions samples. The results presented the following order of models fit quality for the biomass variables: TB > SB > TcB > LB > TnB. The models of SB and TB presented high predictive capacity, with R² adjusted superior to 95%. The models of TcB, TnB and LB presented higher variability in the estimates compared to those of SB and TB, nonetheless, the obtained values of R² adjusted were over 70%. Regarding the forest productivity, the values verified were similar among the stratums of 3 years old plantations and presented greater difference at those of 7 years, what can be associated to the site quality and competition among trees. The total biomass stock for populations of 3 and 7 years old were 65.93 y 119.51 Mg ha^-1, respectively. Respect to the fractions of TcB, TnB and LB, the population of 3 years presented superior absolute and relative values respect to those of 7 years, indicating a raise in stem biomass assignation and crown reduction at older ages. The results obtained show an efficient and precise estimation of forest biomass, and the productivity levels verified indicate a high potential for carbon sequestration with E. dunnii forest plantations.