Visualizing Healthy Forests
DG Researchers Explore Ways of Developing New Software for Monitoring Ecosystems
By Karen Heyman
For the DGRC

Eco-Informatics
	• Researcher profile: Judith B. Cushing • Project profile:   Spatial Data Infrastructure for Ecological Research • Project home page • Resource: LTERnet

Many DG researchers work on projects that involve forms of database integration and data mining, (see accompanying article), but where does the data for data mining come from? In the field of forest ecology, it comes from researchers who go into forests and actually hand-count the trees, since most remote sensing programs, like satellite imaging, do not yet offer good enough resolution to answer many ecologists' questions.

"If they take a picture from the air, they don't know in detail the structure of the tree. They need to know not only the width and height of the crown, but what's inside," says computer scientist and DG researcher Judith Cushing of The Evergreen State College, "We even know of a researcher who climbed the trees and measured the location and size of every branch along with how much foliage was on each, and another who measured epiphyte coverage on each branch."

But if that seems a time-consuming way to go about things, it's even worse when the ecologists get back down on the ground, at least from the point of view of computer scientists. Back in their offices, the ecologists, limited in their resources and untrained in sophisticated computer modeling, enter their data into Excel spreadsheets and other off-the-shelf software not customized for their specialty. Worse, none of the commercial programs can be integrated together, so the analytical opportunities provided by data mining are even more out of reach than upper branches. "Right now, it's a tremendous amount of work for the researchers to analyze their data," says Cushing, "What inspired our work was seeing how hard it is for ecologists to document their datasets."

Cushing and her group set out to create software that could answer all those needs. They first found an even more fundamental challenge: Forest ecology does not have a universally defined set of data structures. To a layperson, a tree is simply a trunk with branches and leaves. But to specialists, trees and their forests are spatial-temporal datasets, filled with statistics about everything from rates of carbon reabsorption to air turbulence.

In order to even hope for automation, Cushing and her group first had to create categories of data that would be relevant to the research community. They did it by pouring over forty years of journal articles, 500 in all. By analyzing the literature, they were able to derive common structural representations and thereby create a template of data structures, which researchers could use as is or customize to build databases using Cushing's group's software that combines those templates into viable database designs. Essentially, they designed a framework, with categories to cover nearly every consideration in the field - but that would also give power users the ability to write their own custom add-ons, if needed.

They then took it one step further and wrote a visualization tool. Visualization allows a different, more powerful kind of reading, explains Cushing. On an Excel spreadsheet, or a histogram, you can show data in two dimensions - you might show how many trees are infected with a parasite, and you could show the infestation relative to the age of the trees or the species of tree. But with a graphic that represents each tree as a silhouette, you can see not only how many trees are infected, but where in the forest the tree is and sometimes much more detail like how old the tree is or what part of the tree is infected. "One researcher looked at our visualization and saw instantly that one species of tree was acting as a natural barrier to the infection's spread," says Cushing.

Latest DG News

• dg.o 2006 Convenes May 21-24, 2006   • dg.o 2006 Early Registration Ends April 10th! • dg.o 2006 Issues CFP - Tutorials • dg.o 2006 Issues CFP - Workshops • dg.o 2006 features Workshops on:    eRulemaking    GeoInformatics • dg.o 2006 features Tutorial on:    •Social Network Analysis • New DG Team Pursues eRulemaking • IEEE ISI2006 Convenes May 22-24, 2006 • eChallenges e-2006 Issues CFP • DG Research Helps Predict Urban Growth • Swapping Secrets of the Double Helix • UK and DO-Wire Launch e-Gov Best Practices wiki • DG Team Develops "Virtual Agora" for e-Gov • Mapping for Times of Crisis • Exploring Detection of Crisis Hotspots • Report: Mass eMail Campaigns Harmful • Scenario-Based Designs for Stat Studies • US, EU Explore Info Integration • DG Team Develops Digital Interpreter • DG Study Gives Teeth to FBI • Research Smooths Road for Small Businesses • DG Researchers Parsing in Tongues • e-Gov Journal Issus Call for Articles • See all news stories Contribute to dgOnline

Cushing's research is an example of a growing field dubbed "eco-informatics." Like biology's new genomic darling, bioinformatics, it is a specific application of informatics, the study of information, to a particular discipline. For practitioners, the challenges are how to integrate and then mine enormous, disparate datasets, often created at different times in different programming languages, and frequently with undefined terms, or worse different terms, for the same concept.

Her work is well-known and welcome to researchers at the Long Term Ecological Research Network (LTER), established by the NSF. "The problems we work on at LTER sites need to be studied over different time scales and in diverse places," says Nicole Kaplan, a Colorado-based LTER data manager in the shortgrass steppe. "The topics we study - like wildlife management, or dynamics of plague in prairie dog metapopulations, will impact government decisions regarding land use. There are even longer-term questions, like regional climate changes, that need to be answered over decades. Yet we've never had a data integration tool like this before; it's all been PIs ad hoc forcing information together from diverse files."

"As a computer scientist, I look at their work and say if we knew earlier the characteristics of the datasets we would be able to do more for you, we could help you do a lot to manage your data," says Cushing. Still, it's not easy on her side either. She estimates it may take up to five years to fully develop the software tools, though they put preliminary prototypes into the hands of scientists as they are developed. The ecologists are eagerly awaiting them, "Judy's tools will be able to provide a visual summary of data over space and time that helps PIs begin to analyze data in ways that let them see significant trends," says Kaplan, "In addition, if they see their data in a new way, they may be able to ask new questions and start new investigations."

In the meantime, there'll still be plenty of ecologists climbing trees.