Russ Frith : CE 699


Introduction

Snow load and snow depth geographical variables cannot be measured at all parts of space; thus, research involving those variables generally requires the use of interpolation techniques for the studies. Observations are taken at points and spatial interpolation is used to obtain full spatial coverage of Alaska's geographic extent. Spatially accurate estimates for those variables requires investigating the relation between snowfall accumulation and secondary data such as elevation data, wind effects, and sea effects. Incorporating those influences on the spatial estimates provides more accurate estimates than approaches based on one parameter like snowfall accumulation. This study treats the mapping of annual average snow depth for Alaska for sparce point data using co-kriging and geographic weighted regression. By using spatial relationships between meteorlogical observations and variables derived from digital elevation modeling, optimum spatial distributions of mean annual snow fall are aimed to be defined. Alaska's unique Arctic climate and topography introduce snow fall mechanisms and snow depth distributions which vary widely across the state. This study presents recommendations for determining snow loads on roof tops in Alaska. A three part procedure is proposed which includes establshing a design concept for dealing with roof snow loads, considering snow conditions at a construction site, and considering snow conditions on the roof itself.