Note to Reporters: A photo of Dan Cooley is available with this release at news.colostate.edu
Scientists can’t predict when disasters like the massive Colorado flooding in September or recent wildfires will occur, but a new study hopes to assess risks that could alert officials of the potential for disasters in specific areas.
The $4.9 million, five-year study, sponsored by the National Science Foundation and the Department of Energy, will be led by Dan Cooley, associate professor in the Department of Statistics at Colorado State University. He will work with scientists from the University of California-Berkeley, the University of North Carolina-Chapel Hill, and Lawrence Berkeley National Labs.
Cooley hopes to use statistical models to help determine the risk of extreme events like flooding, wildfires and droughts in certain areas. He said data will be mined from satellite images, climate models and other sources to assess the risk of natural disasters.
“These events take place because of the right combination of events, such as heat, wind and lack of moisture,” he said, referring to the large number of highly destructive Colorado wildfires over the past two years. “This is all about risk assessment, not weather prediction. We hope to improve, from a statistical point of view, our ability to determine risks associated with extreme events.”
Cooley said Colorado’s recent floods, which killed eight and caused hundreds of millions of dollars in damage across the state, were far different from the 1976 Big Thompson Flood near Loveland, which claimed 143 lives and caused more than $30 million in damage. The recent floods were spread over several days and an area that included hundreds of square miles. The Big Thompson Flood was contained to a relatively small area.
“When assessing risks we want to look at both types of events and create a model that approximately tracks the true risk of each type of event occurring,” he said. “The recent floods in Colorado were quite different from anything on record because they were so widespread. The Big Thompson flood and the 1997 Spring Creek flood in Fort Collins flood (that killed five and severely damaged the CSU campus) were much more intense, concentrated events. The difference shows the need to understand the spatial and temporal aspects of these events.”
Complicating the process, Cooley said, is our ever-changing climate, which complicates the way past data is analyzed. As a result, “100-year events” may occur more or less frequently than researchers have come to believe.
“We will develop probability-based measures of risk that take account of the fact that extreme weather events cannot be predicted with certainty,” he said. “Was the recent flooding in Boulder caused by climate change? That can’t be answered with a simple ‘yes’ or ‘no,’ but we can assess the probability of that event taking place in a changed climate.”
Another aspect of the study will be working with social scientists to determine the potential impact of extreme weather events on people across a broad range of socio-economic circumstances.