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

Abstract

This study combined hierarchical cluster analysis and classification and regression tree algorithms to quantify vegetation and fuel characteristics and to generate spatially explicit vegetation and fuels maps for forest and fire management in the Davis Mountains of west Texas, USA. We used field data, landscape metrics derived from digital elevation models, and spectral information from remotely sensed imagery to (1) determine recent changes in forest stand structure in relation to historical fire exclusion, (2) quantify the effects of fire exclusion on fuel accumulation patterns, and (3) develop predictive vegetation and fuels maps for our study area. Four vegetation types were identified by cluster analysis including: mesic woodlands, pinyon pine forests, alligator juniper forests, and gray oak forests. Vegetation types varied by elevation, landform type, potential relative radiation (PRR), and spectral signature. Age data suggested that the majority of pines in the Davis Mountains established near 1920, just after the widespread 1916 fire and favorable climatic conditions in 1919. Three fuel types were identified that also varied by elevation, landform type, PRR and spectral characteristics, although the importance of these variables in distinguishing fuel types differed from the environmental variables that discerned the vegetation types. Forest stand densities and fuel accumulations were high in the Davis Mountains, which was probably the result of fire exclusion from grazing activities beginning in the early 1900s. Results from this study will be used to implement forest and fire management activities directed toward ecosystem restoration and maintenance.