Researchers Improve Science of Predicting Catastrophes

By Don Comis
October 22, 2004

Agricultural Research Service scientists in New Mexico and cooperators have joined forces to improve the science of predicting catastrophes, ranging from forest fires to desertification and global warming.

By pooling their research, the scientists have constructed a theoretical mathematical framework supported by data as a first step toward developing the tools--including computer models—and designing the experiments needed to forecast and avert catastrophes that can begin with a single tree or shrub, person or event.

A paper on the research appears in the current issue of theProceedings of the National Academy of Sciences.

ARS scientists Debra Peters, Brandon Bestelmeyer and Kris Havstad cooperated on the research with Roger Pielke, Sr., a Colorado State University atmospheric scientist; Craig Allan, a U.S. Geological Survey fire ecologist; and Stuart Munson-McGee, a New Mexico State University chemical engineer.

Peters, an ecologist at the ARS Jornada Experimental Range near Las Cruces, N.M., began the inquiry by wondering about similar dynamics between desertification and wildfires. She and Bestelmeyer, also an ecologist, and research leader Kris Havstad—both at the Jornada—and cooperators now theorize that there are common elements between catastrophes that involve propagating events—like disease epidemics. Such events occur in four stages, with thresholds between each one. At each stage, both the pace of events and the dominating processes or forces change. When events cross the threshold after the third stage, they can become irreversible.

For example, in the first stage of desertification, a shrub invades a grass patch, creating bare spaces by taking water from grasses. The shrub later drops seeds and spreads, crossing a threshold into stage two.

While overgrazing by cattle can push land over the threshold from stage two to three, larger forces—often weather, such as wind and drought—become the dominant drivers after stage three.

This framework helps researchers and decision makers avoid surprises by taking into account different dominant processes and dynamics between stages.

ARS is the U.S. Department of Agriculture's chief scientific research agency.