High-performance computing helps researchers predict flooding for Iowa communities

Iowa Flooding and Professor Krajewski

Rivers are everywhere in Iowa. A network of 3 million water pathways negotiates the state’s landscape, with a half million of them carrying water at any given time.

Many of those waterways come in the form of streams that are small, unnamed, and harmless … most of the time. But when the conditions are right, they can go from tame to troublesome in a hurry.

More than 1,600 communities across the state are at risk for flooding, and researchers at the University of Iowa-based Iowa Flood Center are using high-performance computing (HPC) to improve flood monitoring and prediction. With complex mathematical algorithms, the HPC cluster produces projections every 15 minutes of what river levels in each town will be for the next five days.

“Creating a flood forecasting system for the state of Iowa requires an immense number of computations to predict for every locality in Iowa that’s near a river,” says Ricardo Mantilla, an assistant research engineer at the Flood Center. “You can’t stop it from happening, but you want to know it’s coming as early as possible.”

The Flood Center’s prediction models mimic the aggregation of water, taking into account a multitude of factors: rainfall data from federal radar reports, water levels from instruments in the field, soil type data from the Department of Agriculture, information about land use and vegetation and where crops are in the growth cycle, and a lot of unknowns that also influence the results.

For example, researchers have a pretty good indication of how much it will rain, but when the soil is already wet it holds less water. They don’t have data on saturation, so they don’t know how much of the rain will infiltrate to end up in the rivers – 10%, 20%, 50%?

“Because of what’s happening in the environment, there are a lot of ‘what if’ scenarios,” says Scott Small, who earned his doctorate in Applied Mathematical and Computational Sciences at the UI and is now a post-doctoral research scholar at the Flood Center. “Due to the uncertainties, we are testing four different models at a time.”

Contributing to the challenge of predicting river levels are the size of the state and the need to produce timely forecasts as a situation unfolds. Iowa consists of over 56,000 square miles of land, and new rainfall data comes through every five minutes. HPC is necessary to make calculations for such an immense area at a pace fast enough to keep up with rapidly changing circumstances.

“There is no operational system in the world as detailed as ours,” says Flood Center Director Witold Krajewski, a professor of civil and environmental engineering at the UI. “It’s fair to say that if not for the HPC cluster, we wouldn’t be doing the work we are doing.”

The National Weather Service (NWS) issues flood predictions every six hours for 100 Iowa communities, and human forecasters are part of that process. The Flood Center aims to fill in the gaps for the remaining 1,500 towns that are subject to flooding, and to provide more frequent forecasts.

Currently the NWS forecasts are more accurate than the computer-generated ones, but the Flood Center is constantly evaluating and refining its models to improve them, with the goal of eventually delivering computer-generated predictions that are comparable in accuracy to human experts. The center issued its first predictions in May of 2013; prior to that, the focus was on research and development of the models, and on creating a process to assess the accuracy of the predictions.

The predictions are accessible through the Iowa Flood Information System at Iowafloodcenter.org, which provides Iowans access to flood inundation maps, real-time flood information, flood forecasts, and interactive visualizations. There is an overview of flood warnings for the state, and users can click locations on maps or search for a specific city, street address, business, or landmark to view data for a specific watershed. The inundation maps provide detailed data on water levels during floods and allow community leaders or residents to visualize potential flood conditions for their area.

The severe flooding of the Iowa and Cedar Rivers in 2008 was the catalyst for the Flood Center. A team of UI researchers pooled their expertise to collaborate on flood-related initiatives, and within a year of the flood, they had attracted $500,000 in funding from the National Science Foundation. In 2009, Iowa legislators established the new flood center and allocated $1.3 million to support it.

A portion of that funding was invested in the university’s first high-performance computing cluster, Helium, which is administered through Information Technology Services. The UI is now in the process of building its second supercomputer, Neon, which is slated to come online in early 2014.

“We always know that if the situation calls for it – if it’s an emergency and we think we can do our job better with more computation power – we can get more,” Krajewski says. “We’re thankful for that. It’s a good feeling to know that we have the resources to substantially increase our computation capabilities if necessary.”