Spelling suggestions: "subject:"1inear approximation method"" "subject:"alinear approximation method""
1 |
Proje??o diam?trica com base em dados observados antes e ap?s o desbaste em povoamentos de eucaliptoLacerda, Talles Hudson Souza 16 February 2017 (has links)
?rea de concentra??o: Manejo florestal e silvicultura. / Submitted by Jos? Henrique Henrique (jose.neves@ufvjm.edu.br) on 2017-06-09T22:52:32Z
No. of bitstreams: 2
talles_hudson_souza_lacerda.pdf: 1852089 bytes, checksum: 5f25d81aee4d02d93913bfc83196ecb3 (MD5)
license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Rodrigo Martins Cruz (rodrigo.cruz@ufvjm.edu.br) on 2017-06-14T19:22:36Z (GMT) No. of bitstreams: 2
talles_hudson_souza_lacerda.pdf: 1852089 bytes, checksum: 5f25d81aee4d02d93913bfc83196ecb3 (MD5)
license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-06-14T19:22:36Z (GMT). No. of bitstreams: 2
talles_hudson_souza_lacerda.pdf: 1852089 bytes, checksum: 5f25d81aee4d02d93913bfc83196ecb3 (MD5)
license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5)
Previous issue date: 2017 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior (CAPES) / O objetivo desse trabalho foi avaliar, do ponto de vista estat?stico e biol?gico, simula??es
realizadas por dois modelos de distribui??o diam?trica, ajustados pelos m?todos de
aproxima??o linear e m?xima verossimilhan?a, em planta??es de eucalipto submetidos a
desbaste. Os dados foram provenientes de um povoamento h?brido de Eucalyptus grandis x
Eucalyptus urophylla, sob regime de desbaste, localizado no nordeste da Bahia, vinculados ?
empresa BAHIA SPECIALTY CELLULOSE. Os dados utilizados neste estudo foram obtidos
nas idades 27, 40, 50, 61, 76, 87, 101, 112, 122, 137, 147, 158 e 165 meses. Esse povoamento
foi submetido a tratamentos de remo??o seletiva de 20%, 35% e 50%, nas idades 58 e 142
meses. Utilizou-se dois modelos de distribui??o diam?trica, empregando bases de dados
observadas aos 27 meses (antes do primeiro desbaste), aos 61 meses (ap?s o primeiro
desbaste) e aos 147 meses (ap?s o segundo desbaste). Por meio dos modelos gerou-se tr?s
sistemas, os quais se diferiram no m?todo de ajuste da fun??o Weibull. No sistema 1 os
par?metros da fun??o Weibull foram ajustados pelo m?todo de aproxima??o linear. No
sistema 2 e no sistema 3, os par?metros foram ajustados pelo m?todo da m?xima
verossimilhan?a. As proje??es realizadas pelos sistemas foram confrontadas com as
distribui??es diam?tricas observadas, por meio do teste de ader?ncia Kolmogorov-Smirnov a
1% de signific?ncia, e pelo teste F de Graybill, com n?vel de signific?ncia de 5%. Os tr?s
sistemas proporcionaram distribui??es diam?tricas projetadas estatisticamente semelhantes ?s
observadas, antes e ap?s o desbastes. O sistema 2 apresentou um maior percentual de
proje??es n?o significativas para os dois testes estat?sticos empregados. As simula??es
realizadas pelos modelos apresentaram realismo estat?stico e tend?ncia do crescimento da
distribui??o de di?metros para diferentes porcentagens de desbaste. Houve maior efici?ncia
dos modelos ao se utilizar distribui??es diam?tricas observadas em idades imediatamente
antes do desbaste. As proje??es das distribui??es diam?tricas, empregando-se como base
inicial as distribui??es observadas antes do primeiro desbaste e imediatamente ap?s os
desbastes (simula??es 1, 2 e 3), foram mais precisas do que as proje??es obtidas quando
foram utilizadas somente as distribui??es diam?tricas observadas antes do primeiro desbaste
como base inicial para as proje??es e, em seguida, simulados os desbastes nas idades previstas
e, por ?ltimo, realizadas as proje??es empregando-se a distribui??o estimada remanescente do
desbaste como base inicial para projetar as distribui??es para idades subsequentes (simula??es
4, 5 e 6). / Disserta??o (Mestrado) ? Programa de P?s-Gradua??o em Ci?ncia Florestal, Universidade Federal dos Vales do Jequitinhonha e Mucuri, 2017. / The objective of the study was evaluated from the statistical and biological point of view,
simulations performed by two models of diametric distribution, adjusted by linear
approximation and maximum likelihood methods, in eucalyptus plantations submitted to
thinning. The data were found in a hybrid settlement of Eucalyptus grandis x Eucalyptus
urophylla, under thinning regime, located in the northeast region of Bahia, linked to the
company BAHIA ESPECIALIDADE CELULOSE. The data used in this study 27, 40, 50, 61,
76, 87, 101, 112, 122, 137, 147, 158 and 165 months. This population was submitted to
treatments of selective removal of 20%, 35% and 50%, in the ages 58 and 142 months. Two
diametric distribution models were used, using data bases observed at 27 months (before the
first thinning), at 61 months (after the first thinning) and at 147 months (after the second
thinning). By means of the models three systems were generated, the channels did not differ
any method of adjustment of the Weibull function. No system 1 of the Weibull function
parameters were adjusted by the linear approximation method. In system 2 and in system 3
the parameters were adjusted by the maximum likelihood method. As the projections
performed by the systems were compared with the observed diametric distributions, using the
Kolmogorov-Smirnov test at 1% significance, by the Graybill F test, with a significance level
of 5%. The three systems provided by the statistically projected diametric distributions for
observations, before and after the deviations. System 2 presents a higher percentage of nonsignificant
projections for the two statistical tests used. As simulations of model execution
demonstrated statistical realism and tendency of growth of the distribution of diameters for
different percentages of thinning. There was greater efficiency of the models of use of
diametric distributions observed in ages before thinning. As the projections of the diametric
distributions, using as an initial basis as distributions observed before the first thinning and
after the slabs (simulations 1, 2 and 3), were more accurate than the projections obtained
when only diametric distributions observed before the first Thinning as the initial basis for the
projections and then simulated the lagging at the predicted ages and finally performed as
projections using an estimated remnant distribution of the thinning as the initial basis for
designing as distributions for subsequent ages (simulations 4, 5 and 6).
|
Page generated in 0.1181 seconds