A hierarchical approach for scaling forest inventory and fuels data from local to landscape scales in the Davis Mountains, Texas, USA

Seasonal variability of photosynthetic characteristics influences growth of eight tropical tree species at two sites with contrasting precipitation in Panama

“…Regression trees were pruned using the complexity parameter, which measures the heterogeneity within the subgroups of the regression tree. The complexity parameters selected to prune the regression trees (a) minimized the crossvalidated error rate and (b) reduced the number of splits of the final trees (Poulos et al, 2007;Park et al, 2010). R 2 was calculated for both trees to assess the proportion of model variance explained.…”

Fire models and methods to map fuel types: The role of remote sensing

“…A few years later, Falkowski et al (2005) implemented a different integrated method by combining digital analysis of ASTER imagery with gradient modelling methods for mapping fuel layers for fire behaviour modelling with FARSITE. More recently, Poulos et al (2007) used field data, remotely sensed data and landscape metrics derived from digital elevation models to develop predictive vegetation and fuel maps in Texas. They developed a hierarchical approach for scaling fuel maps from local to landscape scales.…”

Developing custom fire behavior fuel models from ecologically complex fuel structures for upper Atlantic Coastal Plain forests

“…Because cluster analysis had been successfully used to model fuels with reasonable precision over large areas (e.g., Poulos et al, 2007), it was logical to assume that fire behavior could also be handled in a similar manner. This study shows the validity of that assumption, as we were able to reduce ecologically variable and complex surface fuels to a tractable set of 7 FBFMs calibrated to FCCS point data for wind and fuel moisture conditions of the area.…”

“…Progress has been made over the last several decades in using statistical modeling to predict and explain the distribution of surface fuels with varying degrees of success. Several studies have successfully used hierarchical approaches involving cluster analysis and regression trees to model fuels with reasonable precision and limited bias over large areas (Keane et al, 2001;Rollins et al, 2004;Reich et al, 2004;Poulos et al, 2007). More recently Pierce et al (2009) evaluated several methods including gradient nearest neighbor, linear models, regression trees and several geostatistical methods to map fuels in western Washington, Oregon and California.…”