Produtividade de Eucalyptus saligna com base nas propriedades físico-hídricas do solo e parametrização do modelo ecofisiológico 3-PG

Abstract

This study was developed in Serra do Sudeste and Encosta do Sudeste physiographic regions of Rio Grande do Sul state. The objective was to identify the physical and hydraulic soil properties related to forest yield sites in two distinct areas, including physical-structural properties and soil available water capacity (AW) of Argissolo Vermelho-Amarelo Distrófico típico, Argissolo Vermelho-Amarelo Distrófico latossólico (ARG) and Neossolo Quartzarênico Órtico típico (NEO), as well as perform parameterization of the Physiological Principles to Predicting Growth model (3-PG). In the two sites with high (ARG) and low (NEO) productive potential capacity, the growth and productivity of stands characterization was based on data from forest inventories and from destructive sampling and cube by Smalian method of 30 trees of medium quadratic mean diameter (dg). Soil properties were determined in undisturbed and disturbed soil samples, taken from 0.00-0.10, 0.10-0.20, 0.20-0.40, 0.40-0.60, 0.60-0.80, 0.80-1.00 e 1.00-1.20 m soil depths, in 12 sampling points. Eucalyptus yield was different in the ARG (from clay loam to clay texture) and NEO (sandy texture) areas, characterized by mean values of 373.9 m³ ha-1 and 272.3 m³ ha-1, for 89-months-old stands. In the ARG, AW ranged from 0.100 up to 0.304 cm3 cm-3. The highest soil bulk density (BD; 1.64 Mg m-3) was observed in the 0.00-0.10 and 0.10-0.20 m soil layers. Multiple regression analysis with BD and macroporosity (Mac) of the 0.40-0.60 m soil layer had r2 ~ 0.22 in variance estimation of mean annual volume increment (MAIV). Soil air permeability (Kar) was the main soil property related to productivity in ARG stands. In the NEO (sandy texture) areas, AW ranged from 0.025 up to 0.04 cm3 cm-3 and BD had the highest values (1.60 Mg m-3) in the 0.80-1.00 m soil layer. The BD determined from surface up to 0.60 m explained 14% of MAIV variance and 27% of the total height variance (h). Multiple regression that included BD and Kar of the surface layers explained around 23% of mean annual stem biomass increment (MAISB) variance. In the parameterization of the model 3-PG we obtained estimates characterized by r2 higher than 0.9 for the wood volume with bark (V; m3 ha-1), wood biomass (BL; Mg ha-3) and total height (h), and equal to 0.78 for the diameter at breast high (dbh; cm). The statistical efficiency described by the Nash-Sutcliffe indicator had values higher than 0.8, while the PBIAS indicator showed a tendency of underestimation of less than 12%. In the validation process of the model, performed in unused portions in the parameterization, the results were similar to those obtained in the efficiency evaluation, except for the variable dbh, which was underestimated at the lowest ages. The results showed that the soils with high water availability have their physical quality, based on productivity, determined by the properties that represent the flow of solutes and gases, while coarse soils, highly drainable, have their productive capacity inversely related to BD. The 3-PG model was efficient and accurate for the estimation of E. saligna productivity in the observed climatic conditions, with a higher sensitivity for coarse textured soils.