Thesis (M.S.)--West Virginia University, 2005. / Title from document title page. Document formatted into pages; contains iv, 81 p. : ill. (some col.). Includes abstract. Includes bibliographical references (p. 32).

A raster-based geographic information system (GIS), in which a 100% inventory of certain measurements can be stored and manipulated, can provide inexpensive, low variance timber stand volume estimates. Timber volume can be correlated to such topographic variables as slope, aspect, slope position, slope shape, and elevation which can be produced through the manipulation of a digital elevation model. More precise estimates can be achieved through a procedure which utilizes the 100% sample of this information as opposed to the more traditional method of random sampling. Through the use of such a procedure, the potential for more cost effective estimations is available. The use of this same procedure for periodic re-inventory estimations could also provide a more precise estimate. The model-based approach, which utilizes all the information provided by the GIS, produced a more precise estimate of stand volume than did simple random sampling. Statification of the area into two basic cover types, mixed pine/hardwood and upland hardwood, did not provide a more precise estimate of the total stand volume. However, the model-based approach did provide better individual group estimates than random sampling; relative efficiency was 1.380 for mixed pine/hardwood and 1.121 for upland hardwood. In terms of 'cost effectiveness, the percent reduction of the initial sample plots placed in the field to achieve the lower variance would be 29.7 % for mixed pine/hardwood and 14.8 % for upland hardwoods. For reinventory, a more precise estimate was achieved through the use of the model-based approach as opposed to an estimate derived from probability proportional to the volume estimates computed from the previously derived volume equations. The relative efficiency of the model-based estimate as compared to the probability proportional to size sample, was 1.097. / M.S.

Thesis (Ph. D.)--Oregon State University, 2009. / Printout. Includes bibliographical references (leaves 120-129). Also available on the World Wide Web.

This thesis contributes to the understanding of two of the most intriguing questions that forest ecologists have faced over recent decades: 1) how high diversity is maintained in species-rich ecosystems; and 2) what is the role of spatio-temporal environmental variation in structuring forest communities. The aims of the research were to ascertain how species composition varies both spatially and temporally and how changes in the vegetation can be understood in the context of species coexistence theories (niche versus neutral). A group of 38 sympatric species, including 9 species of Quercus, on which little ecological research has been undertaken, were used in this study. The data used in this project include eleven years of periodic remeasurements of permanent plots established in high-altitude oak forests in Mexico. Adult, sapling and seedling trees were studied as well as their environmental surrounding. Spatial and temporal variations in forest composition were analysed using multivariate statistical approaches. The results show that there are discrete communities in these mixed oak forests that correspond to specific environments. At a broad scale the study area can be classified into two floristic zones, a mesic zone characterised by associations that include Quercus candicans, Q. laurina and Q. castanea and; a xeric zone dominated by Q. crassipes. However of a finer scale of analysis important variation in composition was associated with different life stages of the trees, with adult trees showing much stronger environmental associations than seedlings and saplings. Successional pathways and rates vary at relatively fine scales. This may be as a result of dominance alternation between dominant canopy species. Micro-niche zonation processes caused by a high degree of environmental heterogeneity combined with individual species traits explain the coexistence of phylogenetically similar sympatric Quercus species. A hierarchy of processes, each acting at a different spatial and temporal scale, determines species diversity and coexistence. The overall findings support the idea that niche differentiation rather than chance events such as dispersal limitation, are more important in permitting species coexistence.

Orientador: Célia Regina Lopes Zimback / Coorientador: Diego Augusto de Camps Moraes / Banca: Luis Gustavo Frediani Lessa / Banca: Osmar Delmanto Junior / Resumo: A elevada produtividade do eucalipto ocasiona a necessidade do domínio de técnicas e ferramentas precisas acerca dos povoamentos florestais, como o inventário florestal, procedimento fundamental para o monitoramento da produção que permite o conhecimento do máximo potencial das florestas, além disso, é essencial para o planejamento de atividades de corte, colheita e suprimento de madeira. As técnicas convencionais para o levantamento da produção florestal utilizam-se de métodos tradicionais de estatística (estatística clássica), considerando apenas a adoção de valores médios para a tomada de decisões, não explorando as correlações espaciais que possam existir entre as parcelas amostrais. Então, um tipo de interpolador geoestatístico que permite definir a estrutura de dependência espacial dos dados é a krigagem, que juntamente com a metodologia de amostras virtuais, pode-se tornar uma alternativa viável para obtenção de um variograma com modelagem aceitável. Desta forma, o objetivo geral deste estudo foi empregar a metodologia de amostras virtuais no planejamento da malha amostral do inventário florestal, a fim de possibilitar a determinação de um número de parcelas que possam ser estabelecidas na área para obtenção de produtividade, sem que haja perda de precisão ou aumento de custos. A área de estudo abrange uma área plantada de 287,66 ha, pertencente a empresa Eucatex S/A, localizada em Itatinga, no estado de São Paulo. Os dados dendrométricos de volume (m³.ha-1) foram obti... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The high productivity of eucalyptus trees makes it necessary to master accurate techniques and tools for forest, such as the forest inventory, a fundamental procedure for monitoring production that allows the knowledge of the maximum potential of forests, and is essential for planning of harvesting and supply of timber. Conventional techniques for surveying forest production use traditional statistical methods (classical statistics), considering only the adoption of average values for decision making, not exploring the spatial correlations that may exist between the sample plots. So, a kind of geostatistical interpolator that allows defining the structure of spatial dependence of data is kriging, which together with the methodology of virtual samples, can become a viable alternative to obtain a variogram with acceptable modeling. In this way, the aim of this study was to use the methodology of virtual samples in the planning of the sampling of the forest inventory, in order to allow the determination of a number of parcels that can be established in the area to obtain productivity, without losing accuracy or increasing costs. The study area covers a planted area of 287.66 ha, belonging to Eucatex S / A, located in Itatinga, in the state of São Paulo. The volume dendrometric data (m³.ha-1) were obtained through inventories carried out by the company and were estimated in 32 permanent rectangular plots for continuous forest inventories (CFI) and 32 permanent rectangular plots plus 66 temporary plots for the pre-cut inventory (PCI). The following steps were performed with the CFI data for spatialization of the original data in the ArcGIS program: application of an univariate statistic on the variable volume. ha-1; generation of the experimental variogram; variogram adjustment; validation of the theoretical model; application of kriging; generation of productivity maps year by year. Subsequently ... / Mestre

The potential of airborne LiDAR technology to quantify forest structure within eucalypt forests has been evaluated with a focus on the understorey stratum. To achieve this, three studies have been undertaken using multiple (4) LiDAR datasets acquired over three test areas located in Wedding Bells State Forest, Coffs Harbour, Australia. Initially, the effects of sensor configuration were evaluated using field measurements collected from three structurally and topographically differing field plots (40 x 90 m areas). Results indicated that canopy height profiles derived from LiDAR data at the plot scale were largely unaffected by a change in platform altitude from 1000 to 3000 m (p &gt 0.05). In addition, the derivation of individual tree attributes was found to be highly sensitive to the density of LiDAR observations whilst higher platform altitudes showed an increased proportion of single returns over forested areas. In the second study, an innovative field based approach was developed to sample the structure of the understorey (horizontally and vertically) for LiDAR validation purposes. Using two separate LiDAR datasets, this research confirmed that mean understorey height and understorey cover can be effectively mapped in areas of low to medium canopy cover whilst no significant relationship (p &gt 0.05) was identified between field and LiDAR estimates of maximum understorey height. In the third study, an optimised LiDAR beam interception model was developed and validated, and then applied to assess the interaction of extrinsic and intrinsic factors of the LiDAR survey. This demonstrated that the probability of beam interception through the forest canopy can be affected by factors both intrinsic (e.g. crown cover) and extrinsic (e.g. scan angle) to the structure of the canopy. Overall, the results of this research indicate that optimising the sensor configuration is important to the derivation of particular forest structural attributes and significantly, there is potential for LiDAR technology to provide quantitative and spatially detailed estimates of key understorey attributes such as mean height and cover.

Thesis (Ph. D.)--University of Helsinki. / Includes bibliographical references (p. 86-93).

Thesis (doctoral)--Swedish University of Agricultural Sciences, 2001. / Appendix includes reprints of five papers and manuscripts, four co-authored with others. Includes bibliographical references. Also partially issued electronically via World Wide Web in PDF format; online version lacks appendix

A forest inventory was made by a two-man team in the Emory oak (Quercus emoryi) woodlands, near the Huachuca Mountains, in southeastern Arizona. Two plot sizes, 1/10th, and 1/25th hectare, and three basal area factors, 2, 4 and 6 (square meters per hectare) were used. Tree tally time was recorded, but the travelling time between plots was not recorded. Total number of trees, total basal area, and total volume of trees per hectare were measured, summarized, and analyzed. Significant differences were noted between plot sampling and point sampling. Results suggested that with plot sampling, 1/25th hectare plot was more efficient than 1/10th hectare plot for all measures of forest densities. With point sampling, basal area factor 6 had the highest relative sampling efficiency in terms of trees per hectare. Basal area factor 2 had the highest relative sampling efficiency in terms of basal area and volume per hectare. From this information, more efficient forest inventories of the Emory oak woodlands can be designed.

Thesis (M.S.)--University of Wisconsin--Stevens Point, 2007. / Includes bibliographical references (leaves 72-86).