Description

In the chy montaue forests of western North America frequent disturbances of variable types and intensities are a source of fine-scale heterogeneity in forest structure. Structural variability retlects complex spatial and temporal interactions between plant growth (e.g., establishment, dispersal, and competition), environmental gradients (e.g., aspect, soil type, or water availability), different types of natural disturbances (e.g., wildfires, windthrow, and insect outbreaks), and various levels of human intervention. New initiatives that attempt to model forest landscape management based on the spatial and temporal dynamics of natural disturbances require detailed maps and field data for a range of parameters across large areas. Traditional two-stage forest inventory based on medium-scale aerial photography and stratified field sampling is not well suited for this task due to limits in photo resolution and the complex requirements for statistical rigor of field sampling. On the other hand, new procedures based on digital remote sensing data and automated image analysis are still limited in their analytical capabilities and general applicability. Consequently, alternative approaches for exploiting existing technologies and refining sampling procedures remain of interest to resource managers and field staff. In this paper I present a novel three-stage approach for mapping the variability of forest stand conditions across mountainous landscapes. It aims to increase the accuracy and efficiency of traditional airphoto-based mapping by combining supplementary high-resolution aerial photography with a sampling strategy developed for regional-scale vegetation inventory The first stage involves the acquisition and interpretation of 1:2,000-scale small-format aerial photography. Even in steep terrain, existing helicopter-mounted camera boom systems are capable of acquiring continuous strips of stereo images. Helicopter flight lines are purposively [purposely] located to follow significant topographic or environmental gradients, a sampling design first proposed by Gillison (1985). In the second stage, field plots are systematically placed along these transects to provide empirical data of stand structure and disturbance dynamics, as well as reference data for the measurements on stereo pairs. In the third stage, field data and 70-mm airphotos are used in combination to calibrate and verify the classification key for mapping forest conditions on standard 1:15,000-scale aerial photography across the landscape of interest. I have applied this method in four study areas in the Kamloops Forest Region of southern interior British Columbia, each one covering between 2000 and 5000 hectares. Dry montane forests appear particularly suited for this sampling design as they are influenced by a mix of recurrent, often patchy disturbances of variable severity. Also, stands tend to have less dense and uniform crown closure than moister or cooler forest types. To fully capture fine-scale heterogeneity in vertical and horizontal forest structure, I expanded the structural vegetation classification of O*Hara (1996) to 12 stand structure classes. The resulting maps of forest conditions will form the basis for a quantitative spatial analysis of the influence of topographic gradients on stand structure and natural disturbances. © University of Idaho 2000. Abstract reproduced by permission.