WITH EACH PASSING SEASON, it becomes increasingly clear that environmental issues
will pose one of the greatest challenges we face in the 21st century, and perhaps beyond. Somehow
we must find ways not only to ameliorate the damage caused by industrialization, mass consumption,
loss of habitat and species diversity, and climate change, while simultaneously creating new technologies
that make sustainable economic development possible. Some experts predict if we do not find solutions
soon, we may find ourselves in an irreversible environmental crisis within a few decades.
When she began her work in environmental forecasting a quarter century ago, civil engineering professor Efi Foufoula-Georgiou, co-founder of the National Center for Earth-surface Dynamics (NCED) at the St. Anthony Falls Laboratory, and the current Joseph T. and Rose S. Ling Chair in Environmental Engineering, focused on one highly variable natural feature of river systems—precipitation.
Efi Foufoula-Georgiou, professor of civil engineering, examines data in non-traditional ways to create a conceptual
framework useful for predicting the effects of precipitation.
In those days, the most common way to measure precipitation was to use rain gauges. Then came weather radar systems, which gave much more information
about precipitation across time and space. And finally satellites began beaming back images of broad swaths of the earth’s atmosphere. Each additional
measurement tool not only added a different layer of data, but operated at widely different scales, from on-site data gathered from a rain gauge measuring
precipitation within only a few square feet, to data gathered by satellites measuring up to several square kilometers.
Potentially, data from a variety of sources operating
at different scales is a recipe for more accurate description and, ultimately, prediction. The challenge,
said Foufoula-Georgiou, is how to make sense of all that data to create a conceptual framework useful
for predicting the effects of precipitation.
“If you visit any natural environment, there are going to be a lot of variables from one point to another
point and over the course of time,” she said. “Those variables will govern many things, like the ecology and biology of a river valley and how pollutants
and sediment are flushed into streams.”
If one is “willing to look at data in non-traditional ways,” it may become possible to see what she calls “hidden structures” that can form the basis for sound environmental forecasting even of highly complex natural systems.
Models such as the one shown above are used to study some of the processes that shape the world’s highest
mountains. Using lab experiments, field work, innovative data analysis techniques, and mathematical modeling, professor Foufoula-Georgiou’s team works to model and forecast natural and human-induced effects on the environment
that lead to changes in precipitation
rates, erosion, climate, pollutant transport, floods, and shoreline migration. Ultimately,
environmental
forecasting is at the heart of effective resource management and policy decision-making.
“The tool for that is mathematical ‘renormalization,’
looking at things in different scales through multi-scale analysis,” she explains. “Despite the huge variability of what we see, if you renormalize the data, so you know what parameters to use, you can see structure behind variability. That makes it possible to go from one scale to another in a simple way and to use this information to validate predictions
based upon mathematical modeling.”
In short, by extrapolating an understanding of the hidden structure of a natural environment, Foufoula-
Georgiou’s research team can make predictions about the effects of precipitation and other variables even at sites from which measurements are not available. It also makes it possible to use measurements
from smaller systems (such as those gathered
from the University’s new Outdoor StreamLab) to predict changes that will occur in larger systems, or to make measuraments from larger systems (like global precipitation rates taken from satellite sensors)
relevant to predicting changes in smaller, more local systems.
“The complexity in all natural processes is mind-boggling,” she said. “Yet there is hope with the right lab experimentation and with new data becoming available in more locations, it will be possible to understand
these variables in a way that we can translate
knowledge from one place to another and be able to predict patterns of variability in places that are not currently monitored—and may never be monitored.”
