Estimativa de peso de copas de Pinus caribaea Morelet variedade hondurensis Barr. et Golf. e Pinus oocarpa Schiede, através de parâmetros dendrométricos
Resumo
Resumo: O presente trabalho de pesquisa teve como principais objetivos quantificar os resíduos vegetais deixados no solo após a operação de desbaste, através de equações estimativas obtidas pela análise de regressão.
Foram utilizadas 30 árvores de Pinus caribaea hondurensis e 30 de Pinus oocarpa com 7,5 anos de idade, provenientes de povoamentos de propriedade da Reflorestadora Sacramento RESA Ltda., Sacramento, Minas Gerais.
Os modelos de equações testados, através do procedimento "Forward", para avaliar o peso de copas e seus componentes para ambas as espécies foram:
Log P = a + b Log X1
Log P = a + b Log X1 + c Log X2
Log P = a + b Log DAP
onde a, b e c são os coeficientes da equação e X1 e X2 são respectivamente a primeira e a segunda variáveis escolhidas pelo processo seletivo.
Testaram-se também, para as equações estimadoras, em função do DAP, as condicionantes básicas de regressão, a fim de verificar a homogeneidade de variância, normalidade dos dados e independência dos resíduos.
Na maior parte das vezes, o DAP foi eleito como a variável com melhor poder estimativo de pesos e, nos casos em que outra foi a variável escolhida, a capacidade estimada da equação em função do DAP apresentou valores bastante semelhantes. Para se estimar o peso total das copas, as melhores equações foram:
1) Pinus caribaea hondurensis
Log y = 2,237425382 + 1,6099561 Log DAP (R2 =0,87)
2) Pinus oocarpa
Log Y = 2,054712709 + 1,77965996 Log DAP (R2 = 0,90)
Finalmente, são mostradas tabelas de peso total de copas para ambas as espécies, baseadas nas equações propostas. Abstract: The main objective of this research was to quantify the residue originated from thinning operation in pine plantations through the estimative of the crown weight of felled trees.
Thirty trees of each species, hondurensis and Pinus oocarpa, 7,5 years old, from stands owned by "Reflorestadora Sacramento - RESA Ltda.", in Sacramento County, Minas Gerais State, Brazil, were used in the research.
A "Forward" regression program was used to identify the variables with higher partial correlation coefficients in relation to the parameters to be estimated. The mathematical models used to estimate the crown and its components weights were:
Log P = a + b Log X1
Log P = a + b Log X1 + cLog X2
Log P = a + b Log DBH
where,
P = weight in grams
a, b, c = coefficients of the equations
X1 = the variable with higher correlation coefficient
X2= the variable with the second higher coefficient.
For the equations based on the DBH it was tested the homogeneity of the variance, the normality of the data, and the residual independence. All tests resulted positive.
DBH was the variable with higher correlation coefficient in relation to total crown weight and most of the crown components. Even when it was not the first choice, DBH was able to estimate the weights with a precision almost as good as the highly correlated variables.
For estimating total crown weight, the recommended equations were:
1) Pinus caribaea hondurensis:
Log y = 2,237425382 + 1,6099561 Log DAP (R2 = 0,87)
2) Pinus oocarpa:
Log y = 2,054712709 +1,77965996 Log DAP (R2 = 0,90)
Finally. tables showing the total crown weight for both species were built, based on the proposed equations.
Collections
- Dissertações [482]