Open Source Software Library Helps Users Quantify Seasonal Patterns in Noisy Environmental Data
The Landscape Dynamics Assessment Tool (LanDAT), is generating interest among GIS professionals, academics, and planners who have attended workshops focused on describing the analytical advantages for tracking vegetation change using LanDAT’s online map viewer. The LanDAT system’s analytical techniques are generalizable, allowing workshop attendees to apply these techniques to their own detailed localized or multiple-variable analyses. LanDAT’s Eastern Threat Center developers have responded by creating a free, easy to obtain and use library in the software language, R, which is widely used in the ecological and environmental sciences. Now, anyone with a desktop computer can reproduce the same core vegetation-based datasets that are part of LanDAT, or produce new derived measures from different environmental variables such as temperature and precipitation. The software library calculates and returns intuitive representations of users’ input data through a process called polar transformation that substitutes key measures of vegetation dynamics such as length of season and seasonality in favor of the noisy multitude of values in the input data. These polar measures are simple, easy to map, and facilitate the analysis of uncommon vegetation change and patterns, including multi-year cycles and shifts in timing and seasonal intensity. The advantage of the LanDAT analysis technique is the subject of a paper submitted to the International Conference on Data Mining where LanDAT developers plan to present it to a larger community of researchers interested in vegetation dynamics as well as other kinds of data mining and pattern analysis.
Pictured: A LanDAT map shows seasonal intensity across the Upper Midwest. The inset time series plot shows an example output from the LanDAT R library for temperature and precipitation. Map image from www.landat.org and graph image by Bjorn Brooks, U.S. Forest Service.
External Partners/Collaborators: Oak Ridge National Laboratory; University of Wisconsin-Madison