Modelagem de distribuição de espécies com dados de manejo de precisão em florestas tropicais no leste do Acre
The studies, structured into three chapters, aim to develop techniques for improving the management of forest resources using inventory data with areas of precision technical managed in eastern Acre state through the application of predictive models of global distribution species of wood interest. T...
| Autor principal: | Figueiredo, Symone Maria de Melo |
|---|---|
| Grau: | Tese |
| Idioma: | por |
| Publicado em: |
Instituto Nacional de Pesquisas da Amazônia – INPA
2020
|
| Assuntos: | |
| Acesso em linha: |
https://repositorio.inpa.gov.br/handle/1/4956 http://lattes.cnpq.br/6095147404214692 |
| Resumo: |
|---|
|
The studies, structured into three chapters, aim to develop techniques for improving the management of forest resources using inventory data with areas of precision technical managed in eastern Acre state through the application of predictive models of global distribution species of wood interest. To estimate the distribution of the species, it was used the maximum entropy method (Maxent) using environmental variables: elevation, slope, aspect, normalized difference vegetation index (NDVI) and height above the nearest drainage (HAND). It was also tested the inclusion of biological variables, represented by the density of individuals of forest species, as predictors in the models. In the first chapter, the objective was to evaluate if the use of forest inventories can improve the estimation of the probability of occurrence, identify the limits of potential distribution and habitat preference of a group of timber species. The results showed that compared to a random distribution, the method Maxent reached an accuracy of 86%, on average, as predicted geographical distribution of species. The elevation and the NDVI were the most important variables. Ceiba pentandra (samaúma), Castilla ulei (caucho) and Hura crepitans (assacu) is more likely to occur in nearby water course areas. Clarisia racemosa (guariúba), Amburana acreana (cerejeira), Aspidosperma macrocarpon (pereiro), Apuleia leiocarpa (cumaru-cetim), Aspidosperma parvifolium (amarelão) and Astronium lecointei (aroeira) may also occur in upland forest with well drained soils. In the second chapter, the aim was to evaluate the effect of spatial scale of forest inventories as occurrences of data source applied to interpolation of the potential distribution models of species. In modeling, the occurrence data were divided into four geographical regions and several sampling schemes were tested. The use of occurrence data in only one geographic region with natural environmental characteristics, increased both overfitting of models to input data such as errors of omission. The sampling scheme in diagonal and the use of lower threshold values influenced the improvement of the performance of the models. The forest inventories can be used to predict areas with high probability of species, since they are located in forest management plans that represent the environmental range of the projection area of the models. In the third and final chapter, data logging occurrence of forest species were grouped by classes with an interval of 20 cm in diameter at breast height (DBH). The objective was to examine whether there is a relationship between the diameter size classes and the distribution in space, including as predictors, environmental and biological variables. Six predictor variables were selected by species, by the method of all possible regressions. The models had an average of good performance (AUC = 0.7; failure rate = 8.8%), but were significantly influenced by the sample size, due to the limited size of the study area restricted to management plans. The elevation and the NDVI were the most important environmental predictors and the density of A. acreana and C. racemosa stood out among the biological variables, by the Jacknife test. The contribution of biological variables in the models shows the need of expanding the studies about the interaction between species. The trees of A. lecointei, C. racemosa and C. pentandra with DBH ≥ 100 cm are more likely to occur in environment with higher elevations of the land, but it was not observed a significant relation between the size of the potential distribution area and the diameter class. The modeling approaches used in this study have potential application to other tropical species poorly studied, and the
results can help to improve the management of the species that are commercially exploited. |