Abstract: With the recent advancement of geospatial techniques (e.g., remote sensing, Geographical Information Systems, and GPS), geodatasets have grown dramatically in size and number and become more widely distributed. This provides ecologists unprecedented opportunities to explore ecological problems at larger spatial scale than before. At the same time, the wealth of data demands improvements in geospatial techniques in order to fully explore the current data capacity. In this dissertation, I sought to refine, combine, and develop new geospatial techniques that when applied to a real-world ecological problem, generated new and more comprehensive understanding of the ecological system examined. First, I refined several commonly used spatial analytical techniques such as paired quadrat variance (PQV) and Ripley's K functions. I clarified some misinterpretations of PQV methods, and proposed a GIS approach in correcting the edge-effect problem associated with the Ripley's K function in irregular-shaped study areas. Second, I combined several methods to better understand spatial patterns: (1) I combined Ripley's K and semivariance to study point patterns, and (2) I combined PQV, two term local quadrat variance, new local variance, and their three-term counterparts to study transect data. Third, I developed a new environmental niche model to model potential niche using presence-only data. Fourth, I developed a hybrid classifier which integrated an object-based and a knowledge-based classification method in mapping dead trees from high spatial resolution images. Finally, these geospatial methods were applied to analyze and model the spread of a new forest disease 'sudden oak death' in California at the landscape and regional scales. At the landscape level, I found that topographic factors were the most influential factors in controlling the presence of dead trees, followed by foliar hosts of the disease. At the regional scale, I found that the majority of disease risk would occur in coastal areas of California. In summary, this study demonstrated that the new, combined and refined geospatial tools used here helped gain a better understanding of a forest disease process at multiple scales.