dc.contributor.advisor | Dalla Corte, Ana Paula, 1980- | pt_BR |
dc.contributor.other | Sanquetta, Carlos Roberto, 1964- | pt_BR |
dc.contributor.other | Behling, Alexandre, 1987- | pt_BR |
dc.contributor.other | Universidade Federal do Paraná. Setor de Ciências Agrárias. Programa de Pós-Graduação em Engenharia Florestal | pt_BR |
dc.creator | Stolle, Lorena | pt_BR |
dc.date.accessioned | 2022-11-17T11:47:53Z | |
dc.date.available | 2022-11-17T11:47:53Z | |
dc.date.issued | 2021 | pt_BR |
dc.identifier.uri | https://hdl.handle.net/1884/75078 | |
dc.description | Orientadora: Profa. Dra. Ana Paula Dalla Corte | pt_BR |
dc.description | Coorientadores: Prof. Dr. Carlos Roberto Sanquetta, Prof. Dr. Alexandre Behling | pt_BR |
dc.description | Tese (doutorado) - Universidade Federal do Paraná, Setor de Ciências Agrárias, Programa de Pós-Graduação em Engenharia Florestal. Defesa : Curitiba, 19/08/2021 | pt_BR |
dc.description | Inclui referências | pt_BR |
dc.description | Área de concentração: Manejo Florestal | pt_BR |
dc.description.abstract | Resumo: O volume total de uma população florestal é estimado por meio de técnicas de amostragem utilizadas nos inventários florestais, em que o volume estimado na área da unidade amostral é extrapolado para a área do talhão. Neste trabalho, o objetivo foi avaliar e comparar dois métodos para a estimativa do volume total em florestas plantadas: 1) método tradicional de inventários florestais (inferência por área); e 2) método de árvore individual, no qual a estimativa é realizada pelo produto do volume médio individual (vmi) pelo número de árvores detectadas remotamente nos talhões (inferência por indivíduo). A pesquisa foi conduzida em três talhões florestais: Pinus taeda (16 anos), Pinus taeda (7 anos) e Eucalyptus urograndis (5 anos). Em cada talhão, foi realizado o censo para as variáveis: diâmetro a 1,30 m e altura total. Os volumes das árvores foram estimados pelo modelo de Schumacher-Hall, ao passo que a soma dos seus volumes individuais foi considerada como o volume total paramétrico. Em seguida, foi executada a etapa da estimativa do volume total da população. No método 1 (por área), foram analisados dois procedimentos de amostragem (aleatória simples e sistemática). No método 2 (por número de árvores), o vmi utilizado foi obtido de duas maneiras: (a) média do volume individual de cem árvores selecionadas aleatoriamente, reamostradas pelo método Bootstrap e, (b) pela amostragem sistemática. O número de árvores (N) foi obtido por um algoritmo de contagem automática aplicado em imagens obtidas por -uma aeronave remotamente pilotada (ARP). Adicionalmente, no método 2, foram simuladas diferentes taxas de detecção de indivíduos (90% a 110%), enquanto para o método 1, foram simuladas diferenças (±10%) na área do talhão. Os resultados dos dois métodos de estimativa de volume total da população foram comparados ao volume total paramétrico pela diferença absoluta, diferença relativa e pela abrangência dos intervalos de confiança obtidos por cada método. O método 1, tradicional de inventário florestal (inferência por área), apresentou as menores diferenças em relação ao volume do censo. O processo de amostragem sistemático produziu os melhores resultados para os talhões 1 e 3, enquanto o processo aleatório foi o melhor para o talhão 1. O método 2 (abordagem por número de árvores) apresentou o volume total paramétrico dentro do intervalo de confiança (95% de probabilidade) ao se utilizar o vmi da amostragem aleatória de cem árvores e o número de árvores detectadas automaticamente, com diferença relativa de no máximo 4,7%. A variação em área e em número de árvores permitiu verificar que o acerto em 100% da área ou em 100% da do número de árvores não garante o menor erro de estimativa para o total, pois ela também depende da acurácia da estimativa do volume médio individual. O método 2 (abordagem por número de árvores) pode ser utilizado como uma alternativa ao inventário florestal tradicional. | pt_BR |
dc.description.abstract | Abstract: The total volume of a forest population is estimated through sampling techniques used in forest inventories, in which the estimated volume in the sample unit area is extrapolated to the stand area. The aim of this work was to evaluate and compare two methods for estimating the total volume in planted forests: 1) traditional method of forest inventories (inference by area); and 2) individual tree method, in which the estimate is performed by the product of the mean individual volume by the number of trees remotely detected in the stands (inference by number of trees). The study was conducted in three forest stands: Pinus taeda (16 years of age), Pinus taeda (7 years of age) and Eucalyptus urograndis (5 years of age). A census was carried out in each stand for the variables: diameter at 1.30 m (d) and total height. Tree volumes were estimated by the Schumacher-Hall's model and the sum of their individual volumes was considered as the total parametric volume. Then, the step for estimating the total population volume was performed. Thus, two sampling procedures (simple random and systematic) were analyzed in method 1 (by area). Next, the mean individual volume used in method 2 (number of trees) was obtained in two ways: (a) mean individual volume of one hundred randomly selected trees, resampled by the Bootstrap method; and (b) by systematic sampling. The number of trees (N) was obtained by an automatic counting algorithm applied to images obtained by RPA. Additionally, different detection rates of individuals (90% to 110%) were simulated in method 2, while differences (±10%) in the stand area were simulated for method 1. The results of the two methods for estimating total population volume were compared to the total parametric volume by the absolute difference, relative difference and by the amplitude of the 95% confidence intervals obtained by each method. Method 1, traditional forest inventory (inference by area), showed the smallest differences in relation to the volume of the census. The systematic sampling process produced the best results for plots 1 and 3, and the random process for plot 1. Method 2 (number of trees) presented the total parametric volume included in the confidence interval when using the mean individual volume of random sampling of one hundred trees and the number of trees automatically detected, with a relative difference of at most 4.7%. The variation in area and number of trees enabled us to verify that the correctness in 100% of the area or in 100% of the number of trees does not guarantee the smallest estimate error for the total, as it depends on the accuracy of the estimate of the average volume individual. Method 2 (number of trees) can be used as an alternative to the traditional forest inventory. | pt_BR |
dc.format.extent | 1 recurso online : PDF. | pt_BR |
dc.format.mimetype | application/pdf | pt_BR |
dc.language | Português | pt_BR |
dc.subject | Levantamentos florestais | pt_BR |
dc.subject | Arvores - Medição | pt_BR |
dc.subject | Arvores - Amostragem | pt_BR |
dc.subject | Recursos Florestais e Engenharia Florestal | pt_BR |
dc.title | Inventário florestal : inferência por área ou por número de árvores? | pt_BR |
dc.type | Tese Digital | pt_BR |