GIS-based modeling of pinyon-juniper woodland structure in the great basin

Abstract

Pinyon-juniper woodlands are a dominant vegetation type across the Intermountain West, covering an estimated 30 million ha. Increases in density, cover, and aerial extent over recent decades have been associated with altered fire regimes, tree mortality, and reduced understory productivity. In our study we sought to determine how environmental factors influence woodland structure and to exploit such abiotic influences to improve our ability to model woodland structure from digital aerial photography. We developed and applied a methodology for automated tree canopy cover mapping using an object-oriented classifier. The resulting classification was used to scale up stand inventory measurements (0.1 ha) across 11 mountain ranges to a 3,250 km2 area. Predictive models were compared with field-measured values from a validation sample for tree canopy cover, basal area, height, and stand density index. Linear mixed-effects regression models demonstrated consistent relationships between woodland structure attributes and environmental variables. More productive sites supported greater tree densities, whereas less productive sites supported fewer, larger trees. Relatively weak relationships for certain variables (e.g., tree height) suggest that woodland structure-environment relationships are partially obscured by natural disturbances and land use history. © 2009 by the Society of American Foresters.