By Heather Ah San and Andrew Hard
Models are prototypes of our world; they are abstractions of reality. They can manifest themselves as architectural replicas for construction projects, clay mock-ups for automotive design, or even as the paper airplanes that litter the hallways of elementary schools.
When scientists encounter a problem too big to test, they use models. Ayn Shlisky, a regional forest planner at the USDA Forest Service uses models dealing with vegetation growth and development. Models, she says, are especially useful for experiments where full-scale tests are unrealistic.
“You want to try and mimic how the real world works, so that you can either test an assumption of test a hypothesis that you wouldn’t be able to do in the real world,” said Shlisky. “You can do things with models that represent something that may happen over hundreds or thousands of years.”
Since the beginning of humanity, people have been using creative ways to tell stories about the natural world. Hieroglyphs on cave walls, spoken-word narratives, hand-written epics, and emotive folk songs are just some ways that people have told stories in humanity’s short history. In the digital age, another form of storytelling is done with models.
Models can reach into the past and predict the future. Vegetation models use data collected from forests and catalog research to predict how certain landscapes—from murky waters to old growth forests—will look and function in the coming years. Before scientists can make those models though, they have to collect and analyze the data.
Geographic Information Systems: Viewing The World Through A Different Lens
Geographic Information Systems are, in the simplest sense, tools that collect and analyze geographical data, which can then be used to make models. It merges mapmaking, statistical analysis, and database technology into one, cohesive computer program that creates multi-layered visual representations of data. This form of data-visualization has the ability to take abstract statistics, correlations, and trends and give it all a face.
In 1855, an English physicist named John Snow plotted points on a map of London to visually represent cases of cholera that were popping up all around the city. Snow’s studies of the distribution of these outbreaks led to the discovery of a contaminated water pump that was infecting London residents with the fatal disease. Although his use of data plotting was antique by today’s standards, his analysis of statistical and geographical data from an aerial perspective has led many to credit John Snow with using the first form of GIS.
In a more modern sense, the origins of GIS can be traced to Canada. The Canadian Federal Department of Forestry and Rural Development developed a program called the Canada Geographic Information System, which was used to collect, analyze, and visualize “land capability” data for rural sections of Canada. By mapping and displaying information about Canadian soils, agriculture, recreation, wildlife, waterfowl, forestry, and land use visually, Canada created the world’s first true operational GIS system.
Today, the uses of this system are nearly infinite. Kathy Stroud, the Maps/GIS Librarian at the University of Oregon in Eugene, Ore., explained the diversity of the program.
“Natural resource land management was the focus of the first GIS, but social scientists are starting to see the power of looking at it, or the health industry, which is something that’s been growing over the past ten years: where do the homeless live? Do lower income people have access to healthcare clinics?” she said.
However, the main uses of Geographic Information Systems usually center on environmental issues like endangered animals or plants.
“The modeling aspect is going to be huge with the climate change. It’s such a huge area of concern and it affects all of us,” Stroud explained. “Climate change can affect the timber industry and what plants are growing, how vulnerable you are to fire, and if it affects the rainfall patterns. It’s also been used in disaster management.”
GIS was mostly used by government agencies up until the 1990s when ESRI, a software development and services company, came out with desktop software that was much simpler to use. One piece of that software was called ArcView, an entry-level desktop version of GIS.
“In the GIS world, ArcView has made GIS more accessible to people who don’t know how to work on other platforms,” said Ayn Shlisky.
Despite GIS’ accessibility, it generally still has a distinct set of users.
“City planning departments [use it],” said Stroud. “The main number of users would probably be local and regional planners.”
Those planners also use another program, called the Vegetation Dynamics Development Tool, to make predictions on various pieces of land.
The Vegetation Dynamics Tool: Building Models
Simon Bisrat is a modeling analyst working with Portland State University in conjunction with the Forest Service. The state and transition models he uses look at the current state of a particular vegetation type, and the transition it may go through. The state of a vegetation type is defined by age range, cover type (type of vegetation), canopy layers, percent canopy coverage, and size class. The transitions, which could be anything from human activity to fire to insect attack, are represented in arrows. The model uses the boxes and arrows to guide the user to making a prediction on how a particular landscape will function, given these inputs.
“VDDT does simulations across time,” he said. “If you don’t have models that simulate across time, what is the alternative? The bottom line will be: if you don’t have any other very practical and accurate techniques to predict the futures, you depend on models.”
In short, the GIS team collects the data, Bisrat’s team builds and runs models, and analysts make predictions.
The Integrated Landscape Assessment Project
One current example in which both vegetation modeling and GIS are used is in the Integrated Landscape Assessment Project, a two-year Portland-based project that builds models of current and future vegetation landscapes in four states—Arizona, New Mexico, Oregon and Washington. It then uses these models to inform decision-making for land managers, planners, and analysts at regional, state and on-the-ground levels.
The project was funded by the American Recovery and Reinvestment Act and has employed 50 people, including graduate students and recent graduates in Portland.
Miles Hemstrom, who has been working with landscape assessments and modeling for about 15 years, leads the team.
As a forest ecologist in the early ‘80s, Hemstrom saw what happened to forest communities with changes in the timber harvest and realized the downside to looking at the land from one perspective. He later asked himself what would happen if he looked at public land from multiple perspectives, using the economics, ecology and disturbances such as fire and logging as lenses.
“(The) project is personally a culmination of long time effort to pull together how to think about whole land,” he said.
When he received funding from the recovery act two years ago, it gave him the chance to push this vision ahead. One main concern for Hemstrom was to make the models and projections readable and usable for those who work and manage land.
“Most of what we’re doing is not new, “ Hemstrom said.
The landscape assessment, Hemstrom said, is merely building on already existing geographical models by updating information and making predictions for future vegetation and potential risks to the vegetation.
“What our team is trying together is a what to think about how those changes are happening across big areas of land, and what we might do to either get better results… have a more sustainable impact, produce green jobs,” Hemstrom said.
The analysts based at Portland State University use geographical information and analysis information to create the vegetation models. Those models have predictive value, but may not be entirely accurate, says Bisrat.
“The beauty of VDDT is if you build models and do a prediction, once it reaches say a forest manager, it can tell this is the input, this is the output,” he said “There are boxes so if you see it with a forest manager who doesn’t have a good modeling knowledge or background he can understand.”
Shlisky also sees the development tool as the perfect fit for the project.
“[The visualization of the models] really mimics the way that ecosystems, forests and rangelands operate ecologically,” she said. “They include attributes or parameters about ‘What’s the rate of vegetation growth? What’s the frequency of disturbance, whether it’s fire or insects or disease? What happens if you go in and harvest a few trees? What happens if climate change has a particular effect?’”
However, Hemstrom said for the purpose of this project VDDT works perfectly when covering large areas of land, and thus the data is easier to translate for users.
Data for practical applications
Using GIS to visually communicate this “story” behind a landscape to become accessible and relevant not just to decision-makers and land managers, but, eventually, to everyone.
The projected completion date for the Integrated Landscape Assessment Project project is early 2012. The use of GIS has been a huge game changer for how managers and the public view the natural land, manage impacts and understand the consequences of management actions.
“If we can help people think about what it looks like, and what their experience has been, and how that forest can change—it might blow down, it might burn up—we can build a set of stories in people’s minds and what they care about,” said Miles Hemstrom.
Shilsky is excited about having the models to use for forest planning. “I think it’s going to significantly bump up our capacity for making scientifically sound decisions.”
“I’m looking forward to having that capacity and those resources available to us, as well as I envision contributing back to the development of those models because, through time, we’re going to learn more about the way systems function.”