The second dust bowl cometh for America, supercomputer warns
Droughts, flash floods the future for the Midwest ... probably
In the 2004 disaster flick The Day After Tomorrow, the world is subsumed by extreme weather as a violent climate collapse encases the Northern Hemisphere in ice and snow in a matter of days.
However, a more likely future may be that of the dust bowl droughts depicted in Christopher Nolan's film Interstellar, released 10 years later.
A climate model developed by researchers at the Department of Energy's Argonne National Labs, projects prolonged droughts across much of the US which will be followed by brief but devastating floods. But these events won't happen overnight.
Instead, they're forecast to place with increasing frequency over the next 50 years. But even by the middle of the century — just a short 27 years from now — simulations suggest that large portions of the Midwest will be in a state of persistent drought, and the American West isn't looking much better off, despite recent rainstorms that have raised hopes of more lush times ahead.
The model forecasts climate patterns in blocks down to 12 square kilometers — about 4.6 square miles for those who don't speak metric — and was simulated using supercomputers at the National Energy Science Computing Center at Lawrence Berkeley National Lab and at the Argonne Leadership Computing Facility.
The work is intended to help lawmakers make informed decisions when making policy around near-term droughts and floods. However, the work has only just begun, and the team is now working to improve the model's resolution to four square kilometers (1.5 square miles) and use machine-learning techniques to identify both short-term and long-term problems.
"Now we're looking to understand long-term drought better," said Rao Kotamarthi, who leads Argonne's climate and earth system science department, in a DoE blog post this week.
- Biden: I want standard EV chargers made in America by 2024 – get on it
- Atos tossed €20m to build Max Planck Society supercomputer
- Used EV car batteries find new life storing solar power in California
- Power grid worries force Amazon to run Oregon datacenters using fuel cells
Even still, the group has limited the long-term trajectory of these forecasts to 50 years, due to the high degree of inherent uncertainty at play. "These are projections. They're not predictions," Argonne's Brandi Gamelin said in the post.
In what should come as a surprise to no one, the climate is an incredibly complex system with countless variables that need to be accounted for. This is one of the reasons why high-resolution, 10-day weather forecasts are so hard to get right. Though, we've been told those should be getting better any day now.
When it comes to predicting drought conditions, there are more than 50 metrics, including temperature, precipitation, evapotranspiration that have to be accounted for. To get around that Argonne researchers developed a new measure called the Standard Vapor Pressure Deficit drought index (SVDI), which is calculated independently of precipitation.
According to Gamelin, many equate declining precipitation with drought, but this isn't always a good indicator. SVDI allows researchers to measure evaporative demand. The higher it is the more moisture that's drawn out of vegetation and soil.
Another benefit is vapor pressure deficit is relatively easy to model, according to Kotamarthi.
However, extreme drought isn't the only thing Argonne researchers' models forecast. They also predict brief but intense periods of precipitation — a characteristic of many drought prone areas — leading to extensive flooding.
According to researchers, the American Midwest could bear the brunt of these extreme weather events as the climate continues to shift. While precipitation might sound like a reprieve from drought conditions, the researchers note that as the soil dries out, it becomes hydrophobic, causing it to repel water. They note that similar phenomena have been observed with wildfires in California.
Ultimately, scientists hope that improved models will give policymakers something to think about when approaching climate issues. ®