AN ANALYSIS OF RECENT GLACIER AREA CHNAGES IN THE NORTHERN CASCADE RANGE, USA

Inventories of glaciers whose sizes have been measured over time are commonly used to evaluate climate change signals. Most of these inventories have been of a small number of glaciers with uncertain statistical sampling properties, and even with large glacier inventories, there is typically an unknown threshold of detectable change. Using 2006 satellite imagery, we have developed an inventory of glacier areas for 742 glaciers in the Cascades Range of Washington, USA, all of which had previously been measured from 1958 aerial imagery. For the modern inventory, we completed repeated measurements of many glaciers using satellite images from other years, analyzing both the intrinsic errors to the remote sensing process and operator errors. We determined that interannual snowcover variations are the largest source of error. We have concluded that a slight majority of the glaciers shows no statistically detectable change over the half-century. While the great majority of the remaining glaciers are shrinking, a few are growing. Statistical analyses assisted by overlaying glacier boundaries with a digital elevation model failed to show any geographic or physiological pattern that could account for which glaciers would grow and which should shrink. Bootstrapping analyses were used to reinforce the statistical validity of these measurements and also to determine that a sample size of around 40 glaciers is the minimum needed to reliably understand glacier change if the 742 were treated as a population.