Advisory group turns to computer model for Cook Inlet spill control

Predictions helping prevention

Posted: Sunday, December 03, 2000

Predicting the future was once considered the domain of fortune tellers hunched over crystal balls. But now scientists do it using the technique called computer modeling.

Thursday afternoon, the Cook Inlet Regional Citizens Advisory Council held a workshop presenting its model for predicting possible oil spills in Cook Inlet.

The Kenai-based public watchdog group, established to oversee oil spill prevention and response, is making the software available free to anyone via its Web site.

"We are trying to make it available so people can test it," said Susan Saupe, the council's science coordinator.

The council members also hope awareness of the project will encourage the collection of more information on winds and currents to improve the model's accuracy, she said.

Anyone using the model can click on a computer screen map to mark the location of an oil spill and add in conditions such as the time and wind conditions. The program then tracks the simulated spilled oil as it moves fast-forward across the water or onto shore in response to currents, tides, winds and turbulence.

The model program is designed so that anyone with minimal computer experience can use it and so that it can be loaded into a laptop and taken out to a spill or training site.

The program is called the Cook Inlet Oil Spill Model.

A team of specialists has been developing it over the past five years. The team leader is Bryan Pearce, a computer modeler who runs Pearce Engineering and teaches at the University of Maine. CIRCAC has contracted about $80,000 to $90,000 for the project through 2001.

The model has several uses, according to the workshop speakers.

The computer visuals condense a huge amount of numerical data and complex mathematical interrelationships into a simple video display. It makes it easier to explain what is happening.

"You can transmit a lot of information just by looking at the screen," Pearce said.

The oil spill model divides the inlet into 40,000 boxes, calculating wind, tides, currents and resulting oil slick movements for each segment and converting the results into images. The rapid advances in computer speed and capacity have improved modeling capabilities even since CIRCAC began the project.

"The laptops are more powerful than what we called super computers 15 years ago," Pearce said.

John Whitney, from the National Oceanic and Atmospheric Administration, said computer modeling has developed so much over the past five years that models are specialized into ones for response and ones for planning.

Response models predict in a real spill where the oil will go next, so responders can get workers and gear to crucial areas for protection and cleanup. Such a model must allow input of "course corrections" as observers monitoring a moving oil slick send in new information.

Planning models help industry, agencies and organizations target prevention and response preparations. For example, NOAA has developed a Trajectory Analysis Planner, which predicts where spills are mostly likely or most risky. It uses computer power to run thousands of simulations and statistically calculate odds of oiling. The information allows planners to evaluate the incidents they are most likely to encounter.

NOAA already has applied the analysis to San Francisco Bay and other sites in California and Hawaii.

Saupe said CIRCAC is using that approach to plan its environmental monitoring. It can focus sampling on beaches that are mostly likely to be hit by spills. The information also can help response managers make informed advance decisions about using dispersants.

CIRCAC plans to develop a risk map of inlet shorelines using calculations similar to but on a smaller scale than NOAA's, she said.

Making oil spill computer models more accurate is a major concern, the presenters said.

Pearce and Whitney recommended skepticism about computer models.

The CIOSM is a bargain model, designed for use by laymen. It could be more complex, more accurate -- but harder to use, Pearce stressed.

"In the end, we've always opted to keep it as simple as possible," he said.

Cook Inlet is very difficult to model, the scientists said.

For example, the tides and currents are extreme, and the rips in the inlet's center include the strongest downwelling areas in the world. The downwelling is not yet factored into the models.

Despite the challenges, the improving computer technology and increasing sophistication of physical oceanography give hope that the modeling approach will become more useful in years to come. CIRCAC intends to continuously improve CIOSM in the years ahead, Saupe said.

"You can only improve your models by collecting real data," she said. "There really needs to be some money spent in Cook Inlet."

Subscribe to Peninsula Clarion

Trending this week:


© 2018. All Rights Reserved. | Contact Us