alpine space - man & environment, vol. 7: Global Change and Sustainable Development in Mountain Regions © 2009 iup • innsbruck university press, ISBN 978-3-902571-97-7 Optimizing a Monitoring Network for Assessing Ambient Air Quality in the Athabasca Oil Sands Region of Alberta, Canada Witold Frączek 1) , Andrzej Bytnerowicz 2) and Allan Legge 3) 1) ESRI, Application Prototype Lab, Redlands, California, USA 2) US Forest Service, Paciic Southwest Research Station, Riverside, California, USA 3) Biosphere Solutions, Calgary, Alberta, Canada To ensure a high level of conidence in the results of any geostatistical interpo- lation, it is very important to have an adequate number of well distributed air quality sampling stations in a monitoring network. What is the adequate number of sampling stations and what is the best approach to optimize their distribution? Could GIS with a special emphasis on geostatistics help to answer these questions? 1. Introduction to geostatistics Geostatistics is a discipline of science which applies statistical methods for spatial interpolation. Even though geostatistics was developed independently from geographic information systems (GIS), today it has become an integral part of GIS. The research performed by meteorologists, geologists, foresters, and other scientists can beneit from applying GIS aided by geostatistics. Geostatistics is applicable when the studied phenomena are the regionalized vari- ables, which fall between random and deterministic variables. Geographic distribu- tion of the regionalized variables cannot be mathematically described as determin- istic; yet the distribution of intensity of those phenomena is not random. Most of the natural phenomena that take place in the atmosphere, seawater or soil meet the criteria of this category. Distribution of air temperature, salinity of oceans, soil moisture, and ore deposits concentration in a geologic layer are all examples of regionalized variables. Crop yield prediction and the distribution of air pollutants might also be a subject of geostatistical analysis even if those are not representative of the natural phenomena. Since we cannot observe the world exhaustively, we must sample. The ultimate criterion for sampling is to obtain an adequate representation of the phenomenon under study. Spatial sampling is an important problem in environmental studies