Desenvolvimento de modelos alométricos para estimar biomassa e carbono de mudas de espécies arbóreas, em áreas atingidas por tempestades de vento em Manaus (Am)

Abstract

A data set with 504 trees belonging to the natural regeneration (from 30 cm in height and < 5 cm dbh), was used to test 10 allometryc models, five linear and five non linear, to estimate total dry biomass of standing seedlings and saplings. The data was collected destructively, near Manaus (AM), Central Amazonia, on a site covered by dense terra-firme moist forest on plateaus dominated by yellow oxisols, affected by wind storms in 2005. The micro sites created by the wind gust made possible the access to individuals that came to the area after the disturbance. The data set was divided in different size categories. The hole data set (4 cat), RN 1 (30-50 cm in height), RN 2 (50-100 in height), RN 3 (100-200 cm in height), RN 4 (from 200 cm in height and less than 5 cm dbh) and 38 individuals of Goupia glabra Aubl. (Cupiuba) belonging to the 4 cat. For the whole data set 6 models estimated precisely the dry biomass, for RN 1 non models, RN 2 eight models, RN 3 six models, RN 4 one model, and two for cupiuba, being the difference between observed and estimated mean dry weight below 5%. The use of the whole data set produces more consistent results than the use of the different size categories. For 4 cat the use of non linear model with a single independent variable, root collar diameter (Dc), presents results as consistent and precise as the models that include seedling and sapling total height in the model. The individuals belonging to RN 1 and RN 2 was not divided in biomass components, RN 3 was divided in roots and shoot, RN 4 in roots, steams and foliage. The mean carbon content for the different size categories and different biomass components are: RN 1 and RN 2 (45,25% ± 0,07), RN 3 roots (41,84% ± 0,10), shoot (47% ± 0,05), RN 4 roots (46,15% ± 0,11), steam (46,05% ± 0,06) and foliage (47,37% ± 0,08). The weighted means are: RN 3 (46,05% ± 0,06), RN 4 (46,15% ± 0,07) and 4 cat (46,04% ± 0,07). The use of a single equation to estimate total dry weight to represent the different size categories, applying the covariance analysis with dummy variables was not possible. The regression equation of the model six to represent the whole data set can be used to follow carbon dynamics for natural regeneration seedlings and saplings, as well as to estimate carbon sequestration in the beginnings of reforestation/afforestation activities.