Interactive Visual Analytics Promotes Pattern Discovery in 50-year Record of Cone Production in Upper-slope Conifers of the Pacific Northwest

Although cyclical cone production is a salient feature of many conifer life histories, relatively little is known about the temporal patterns of cone production in high-elevation conifer forests and factors that link conifer population dynamics to climate dynamics. Since 1959, J. Franklin and others have collected data on cone production of upper slope conifers in the Cascade Range of Oregon and Washington. Cone counts were obtained for nine species in 37 plots spanning 800 km from northern Washington to southern Oregon in ten National Forests. The data set has been difficult to analyze because it is large (45,704 observations), it contains many trees (934 distinct trees of 9 species), and trees died/were added during the study period.

We develop a web-based visual-analysis tool that facilitates the collaborative visual inspection of distribution patterns in large multi-species data sets with a focus on temporal trends. The tool is useful in the early stages of data exploration and collaboration, prior to further statistical analysis. In this poster, we present case studies of ecologists using this tool to explore the cone production data set. In particular, ecologists use this tool to overview the dataset, examine and compare subsets of data in detail, and detect distribution patterns, temporal trends, and exceptions. Moreover, ecologists can asynchronously share views/ hypotheses with others for further analyses. The project is a collaboration between ecologists and computer scientists at Andrews LTER.