Book Description

The coastal old growth forest of northern southeast Alaska is ecologically unique to North America. These forests have developed under relatively short, cool, and extremely wet growing seasons. Cloud covered days are more common than clear days. Rainfall and temperature show highly variable pattern dependent upon proximity to mainland icefields and Pacific ocean, topography, and regional weather pattterns. Soil moisture is excessive and fire is absent. Wind is an important agenct causing change in the forest vegetation. Landslides and snow avalanches on steep mountain slopes and insects are minor, locally important agents of change. Most of these forests are in an old growth, climax condition. This is in sharp contrast to dry, fire influenced ecosystems where much of the forest is in young growth. Natural variation in this old growth forest was described and classified using field observations, principal components analysis, and stepwise discriminant analysis (SDA) of reconaisance level data from 875 stands. These stands were sampled over a 4 year period from 1981 through 1984. Stands were sampled in subalpine, streamside riparian, Pacific coastal oceanspray, and general upland forest zones. Stands varied from highly productive, single tree species dominated closed forests of either Sitka spruce (Picea sitchensis), western (Tsuga heterophylla) or mountain hemlock (Tsuga mertensiana) to unproductive, many species dominated open forests. Blueberry (Vaccinium spp.), devil's club (Oplopanax horridum), and a variety of fortis dominated the understory of the most productive stands. Blueberry, rusty menziesia (Menziesia ferruginea), copperbush (Cladothamnus pyrolaeflorus), skunk cabbage (Lysichitum americanum), deer cabbage (Fauria crista-galli), sedges, and bog or alpine tundra plants dominated the understory of unproductive forests. This variation has been captured and described by classifying the forest into 1 ecological type, 25 plant associations and 5 series [western hemlock, western hemlock-Alaska cedar (Chaemacyparis nootkatensis), mixed conifer, mountain hemlock, Sitka spruce]. Combined vegetation and environment variable SDA models provided the best overall prediction of plant association and series. Environment variable SDA models performed the worst and vegetation variable models produced moderate results. Understory vegetation was correctly predicted for 45 to 78% of the stands using both overstory and environment variables in the SDA model. Most stands were correctly classified to association when all variables were included in the model. Soil drainage, soil disturbance from flooding and erosion, and temperature appeared to be major factors affecting plant community composition and distribution. In the uplands, a soil drainage gradient was evident from the mixed conifer associations on the most poorly drained soils to the western hemlock associations on the best drained soils. A flooding disturbance gradient was evident in the riparian Sitka spruce associations from highly disturbed alder (Alnus spp.) to undisturbed blueberry associations. Gradients of growing season duration and soil drainage appeared important in high elevation mountain hemlock associations.