WildfireDB contains over 17 million data points that capture how fires have spread in the contiguous United States over the last decade 

A team at UC Riverside led by computer science assistant professor Ahmed Eldawy is collaborating with researchers at Stanford University and Vanderbilt University to develop a dataset that uses data science to study the spread of wildfires. The dataset can be used to simulate the spread of wildfires to help firefighters plan emergency responses and conduct evacuation. It can also help simulate how fires might spread soon under the effects of deforestation and climate change, and aid risk assessment and planning of new infrastructure development.

The open-source dataset, named WildfireDB, contains over 17 million data points that capture how fires have spread in the contiguous United States over the last decade. The dataset can be used to train machine learning models to predict the spread of wildfires.

“One of the biggest challenges is to have a detailed and curated dataset that can be used by machine learning algorithms,” said Eldawy. “WildfireDB is the first comprehensive and open-source dataset that relates historical fire data with relevant covariates such as weather, vegetation, and topography.”

First responders depend on understanding and predicting how a wildfire spreads to save lives and property and to stop the fire from spreading. They need to figure out the best way to allocate limited resources across large areas. Traditionally, fire spread is modeled by tools that use physics-based modeling. This method could be improved with the addition of more variables, but until now, there was no comprehensive, open-source data source that combines fire occurrences with geospatial features such as mountains, rivers, towns, fuel levels, vegetation, and weather.

Eldawy, along with UCR doctoral student Samriddhi Singla and undergraduate researcher Vinayak Gajjewar, utilized a novel system called Raptor, which was developed at UCR to process high-resolution satellite data such as vegetation and weather. Using Raptor, they combined historical wildfires with other geospatial features, such as weather, topography, and vegetation, to build a dataset at a scale that included most of the United States.

WildfireDB has mapped historical fire data in the contiguous United States between 2012 to 2017 with spatial and temporal resolutions that allow researchers to home in on the daily behavior of fire in regions as small as 375-meter square polygons. Each fire occurrence includes the type of vegetation, fuel type, and topography. The dataset does not include Alaska or Hawaii.

To use the dataset, researchers or firefighters can select information relevant to their situation from WildfireDB and train machine learning models that can model the spread of wildfires. These trained models can then be used by firefighters or researchers to predict the spread of wildfires in real-time. 

“Predicting the spread of wildfire in real-time will allow firefighters to allocate resources accordingly and minimize loss of life and property,” said Singla, the paper’s first author. 

The paper, “WildfireDB: an open-source dataset connecting wildfire spread with relevant determinants,” will be presented at the 35th Conference on Neural Information Processing Systems (NeurIPS 2021) Track on Datasets and Benchmarks and is available here. A visualization of the dataset is available here. Eldawy, Singla, and Gajjewar were joined in the research by Ayan Mukhopadhyay, Michael Wilbur, and Abhishek Dubey at Vanderbilt University; and Tina Diao, Mykel Kochenderfer, and Ross Shachter at Stanford University.

Exoplanets are experiencing a stratospheric rise. In the three decades since the first confirmed planet orbiting another star, scientists have cataloged more than 4,000 of them. As the list grows, so too does the desire to find Earth-like exoplanets — and to determine whether they could be life-sustaining oases like our own globe. 

The coming decades should see the launch of new missions that can gather ever-larger amounts of data about exoplanets. Anticipating these future endeavors, a team at the University of Washington and the University of Bern has computationally simulated more than 200,000 hypothetical Earth-like worlds — planets that have the same size, mass, atmospheric composition, and geography as modern Earth — all in orbit of stars like our sun. Their goal was to model what types of environments astronomers can expect to find on real Earth-like exoplanets.

As they report in a paper accepted to the Planetary Science Journal and submitted Dec. 6 to the preprint site arXiv, on these simulated exoplanets, one common feature of present-day Earth was often lacking: partial ice coverage.

“We essentially simulated Earth’s climate on worlds around different types of stars, and we find that in 90% of cases with liquid water on the surface, there are no ice sheets, like polar caps,” said co-author Rory Barnes, a UW professor of astronomy and scientist with the UW’s Virtual Planetary Laboratory. “When ice is present, we see that ice belts — permanent ice along the equator — are actually more likely than ice caps.”

The findings shed light on the complex interplay between liquid water and ice on Earth-like worlds, according to lead author Caitlyn Wilhelm, who led the study as an undergraduate student in the UW Department of Astronomy. 

“Looking at ice coverage on an Earth-like planet can tell you a lot about whether it’s habitable,” said Wilhelm, who is now a research scientist with the Virtual Planetary Laboratory. “We wanted to understand all the parameters — the shape of the orbit, the axial tilt, the type of star — that affect whether you have ice on the surface, and if so, where.”

The team used a 1-D energy balance model, which computationally imitates the energy flow between a planet’s equator and poles, to simulate the climates on thousands of hypothetical exoplanets in various orbital configurations around F-, G- or K-type stars. These classes of stars, which include our G-type sun, are promising candidates for hosting life-friendly worlds in their habitable zones, also known as the “Goldilocks” zone. F-type stars are a bit hotter and larger than our sun; K-type stars are slightly cooler and smaller.

In their simulations, the orbits of the exoplanets ranged from circular to pronounced oval. The team also considered axial tilts ranging from 0 to 90 degrees. Earth’s axial tilt is a moderate 23.5 degrees. A planet with a 90-degree tilt would “sit on its side” and experience extreme seasonal variations in climate, much like the planet Uranus.

According to the simulations, which encompassed a 1-million-year timespan on each world, Earth-like worlds showed climates ranging from planet-wide “snowball” climates — with ice present at all latitudes — to a steaming “moist greenhouse,” which is probably similar to Venus’ climate before a runaway greenhouse effect made its surface hot enough to melt lead. But even though most environments in the simulations fell somewhere between those extremes, partial surface ice was present on only about 10% of hypothetical, habitable exoplanets.

The model included natural variations over time in each world’s axial tilt and orbit, which in part explains the general lack of ice on habitable exoplanets, according to co-author Russell Deitrick, a postdoctoral scientist at the University of Bern and researcher with the Virtual Planetary Laboratory. 

“Orbits and axial tilts are always changing,” said Deitrick. “On Earth, these variations are called Milankovitch cycles and are very small in amplitude. But for exoplanets, these changes can be quite large, which can eliminate ice or trigger ‘snowball’ states.”

When partial ice was present, its distribution varied by a star. Around F-type stars, polar ice caps — like what Earth sports currently — were found about three times more often than ice belts, whereas ice belts occurred twice as often as caps for planets around G- and K-type stars. Ice belts were also more common on worlds with extreme axial tilts, likely because seasonal extremes keep the polar climates more volatile than equatorial regions, according to Wilhelm.

The team’s findings of ice on these simulated Earth-like worlds should help in the search for potentially habitable worlds by showing astronomers what they can expect to find, especially regarding ice distribution and the types of climates.

“Surface ice is very reflective, and can shape how an exoplanet ‘looks’ through our instruments,” said Wilhelm. “Whether or not ice is present can also shape how a climate will change over the long term, whether it goes to an extreme — like a ‘snowball Earth’ or a runaway greenhouse — or something more moderate.”

Ice alone, or its absence, does not determine habitability, though.

“Habitability encompasses a lot of moving parts, not just the presence or absence of ice,” said Wilhelm.

Life on Earth has survived snowball periods, as well as hundreds of millions of ice-free years, according to Barnes.

“Our own planet has seen some of these extremes in its own history,” said Barnes. “We hope this study lays the groundwork for upcoming missions to look for habitable signatures in exoplanet atmospheres — and to even image exoplanets directly — by showing what’s possible, what’s common, and what’s rare.”

Rachel Mellman, a recent UW graduate in astronomy, is a co-author of the paper. The research was funded by NASA through grants to the Virtual Planetary Laboratory.