|
MODELAGEM DE MATERIAL COMBUSTÍVEL EM PLANTAÇÕES DE Pinus taeda L. e Eucalyptus dunnii Maiden Resumo: Um inventário foi conduzido em Três Barras, Santa Catarina, para quantificar e modelar material combustível superficial vivo e morto de plantações de Pinus taeda e Eucalyptus dunnii. De forma sistemática foram estabelecidas 20 parcelas em oito idades de plantio para cada espécie, sendo determinados a carga de combustível, o DAP, a altura dominante, o diâmetro dominante e a área basal em todas as plantações, além da altura da vegetação baixa do eucalipto e a espessura da liteira do pinus. DAP, hdom, idade e espessura da camada de acículas foram as variáveis que melhor se relacionaram com a carga de combustível das plantações de P. taeda e, idade, área basal e hdom com a carga de E. dunnii. Vinte e um modelos foram ajustados através dos métodos Stepwise e "Todas as Regressões Possíveis", sendo nove para P. taeda e doze para E. dunnii. Além de selecionados pela distribuição dos resíduos, R2 e erro padrão da estimativa, critérios de ordem prática foram também considerados na escolha dos modelos. O melhor modelo para estimar acículas foi ajustado com base na espessura "S" da liteira de P. taeda, enquanto que para liteira de E. dunnii foi utilizada a idade "I" na formulação do modelo , ambos com R2 acima de 0,85. Para carga de combustível morto de Eucalyptus dunnii, de até 2,5 cm de espessura, o modelo não linear foi o que proporcionou o melhor ajuste, utilizando a idade das plantações como variável explicativa e, para P. taeda a maior precisão foi obtida com usando a espessura da liteira, ambos com R2 acima de 0,85. Para determinação da carga de combustível total das duas espécies, os melhores modelos obtidos foram para E. dunnii e para P. taeda, ambos com R2 abaixo de 0,71, resultado que pode ter sido influenciado pela carga de combustível vivo e resíduos de exploração do povoamento anterior, observadas nas idades mais jovens. Ficou constatado que um ajuste para peso de combustível vivo do sub-bosque de E. dunnii não proporciona resultados satisfatórios se, no modelo, além da altura da vegetação baixa, uma outra variável como idade ou área basal não for incluída. Sete gráficos para estimativa indireta de combustíveis e um "Calibrador de Carga" de acículas de P. taeda são apresentados. Abstract: A surface füel loading inventory was conducted in Três Barras/Santa Catarma-Brazil to appraise dead woody füel and understory live füel withm a l. 8 m strata over mineral soil. Twenty 1.0 m piots were systematically allocated m eight difíerent Eucatyptus dunnii and Pinus ïaeda stands ages, totalmg 320 piots. Total füel load, dbh, dominant dbh, dominant height and basal área at ali piots, as well as litter depth of Pinus ïaeda and understory low vegetation height of Eucalypíus dunnii were measured. Analysis showed that for Pinus taeda good correlations were obtamed between d.b.h, age, litter depth and íüel loading, whereas for Eucatyptus dunnii good correlations were obtained between füel loading and age and füel loading and basal área. The dominant height showed aiso high correlation with the füel accumulated up to timelag 10 for the two species. Nine füel modeis were fitted to evaluate needies weight and others downed füeis weight o f Pinus taeda and twelve to estirnate dead füel loading and live füeis weight for Eucályptus dunnii understory by the Stepwise and Ali Possibles Regressions methods. The equations were selected by coefficient of determination (R 2 ), standard error of estirnate (Syx), and aiso by other practical critena like facility of measurement and propensity to error m the vanables obtainment. The best model to predict pine needies loading was Wa = 2.356794.5' where "S" is the depth litter, while for Eucalypíus dunnii litter füel loading was i^=1.697347+0,001005./ 2 , ali with R 2 above 0.85. For dead füel with timelag 10, the non-lmear model W^ = [/'•^"'yexp^O.OinoS./)] was the best for Eucatyptus dunnii, that hás the age as mdependent variable, while for Pinus taeda was used the litter depth (S) as mdependent variable lü obtain the followmg model: iyio=1.889811+1.569117.5'+0.145218.5' 2 . For both species the R 2 was higher than 0.85. No good fít was obtamed for total füel loading, probably because the live füel and siash remaming for past clearcutting had contributed strongly in the young ages. In both cases the best coefficients of determination were between 0.67 to 0.70. It was observed that if age or basal área associated with understory high, were not included, no good accuracy could be obtained to predict Eucályptus dunnú live füel loading. |