The Boreal Forest of Interior Alaska: Patterns, Scales, and Climate Change 770 views

According to a variety of field observations, most forest types of the boreal forest in Interior Alaska can be found at unique elevation ranges and topographic slopes and aspects. My analysis of spatial interactions among fire, vegetation type, and topography at 1km resolution suggests that these spatial patterns are still represented at this scale. In order to understand drivers of vegetation type distribution and change, a hierarchical logistic regression model was developed. The model indicates that the distinction between tundra versus forest is driven by elevation, precipitation, and south to north aspect. The separation between deciduous forest versus spruce forest is driven by fire interval and elevation. The identification of black versus white spruce uses fire interval and elevation as the main drivers. The model was validated in Interior Alaska and Northwest Canada where it could predict vegetation with good accuracy. The logistic regression model could also be used to distinguish bog vegetation from all other vegetation types and improved in predictive ability when actual fire history was included in model development. The model was then used to identify vegetation response to environmental change by imposing changes in temperature, precipitation, and fire interval. Black spruce remains the dominant vegetation type under all scenarios expanding most under warming coupled with increasing fire interval. White spruce is clearly limited by moisture once average growing season temperatures exceed 2°C. Deciduous forests expand their range the most when decreasing fire interval, warming, and increasing precipitation are combined. Tundra is replaced by forest under warming but expands under precipitation ii increase. Model predictions agree with current knowledge of the response of vegetation types to climate change. The response of vegetation types to environmental changes is not linear when two changes are imposed simultaneously. The last chapter explores the compatibility and accuracy of currently existing classifications for Interior Alaska and the effect of scale. Overall agreement among the classifications is very low; low kappa values indicate that much of the agreement among the classifications can be attributed to random chance. The resolution of the vegetation classifications affects the representation of vegetation types: the major vegetation types eliminate the less abundant types with increasing coarseness. Note: Abstract extracted from PDF text

PHD (Doctor of Philosophy)

English

All rights reserved (no additional license for public reuse)

2003-05-01