A Decade Of DesignSafe

Marking a decade of innovation, DesignSafe celebrates 10 years as the nation’s go-to cyberinfrastructure for natural hazards research, empowering groundbreaking research and building a safer future to strengthen communities and save lives.

DesignSafe’s computational resources, software, and expertise have empowered 10,000 scientists, engineers, students, and other researchers with tools for managing, analyzing, and sharing data on an expanding list of natural hazards including earthquakes, hurricanes, inland flooding, tornadoes, landslides, and wildfires.

As a cornerstone of the National Science Foundation (NSF)-funded Natural Hazards Engineering Research Infrastructure (NHERI), this platform extends far beyond technological innovation, playing a transformative role in advancing research practices and collaboration in the natural hazards field.

“NHERI, and DesignSafe in particular, plays a critical role for us to perform research to improve our understanding of the consequences of natural hazards,” said DesignSafe Principal Investigator Ellen Rathje, a professor in the Cockrell School of Engineering’s Maseeh Department of Civil, Architectural and Environmental Engineering at The University of Texas at Austin.

The engine behind DesignSafe is TACC, enabling researchers to perform advanced computational research, manage massive datasets, and collaborate across institutions.

“TACC is a partner that puts the researcher first, and they realize that it’s important to build a bridge to the natural hazards research community,” Rathje said. “I credit TACC and its leadership for helping DesignSafe have a successful first decade.”

Improved Building Safety After Florida Hurricanes

Rathje points to a powerful example of DesignSafe’s impact: field reconnaissance images from the devastating 2017–2018 hurricanes in Florida, including Hurricanes Irma and Michael, were used to identify critical vulnerabilities in building performance, leading to improvements in safety and resilience in future building design.

“These structures were elevated because of the water hazard but doing so enhanced the potential for wind damage,” Rathje said. 

Building on the hurricane field observations, NHERI researchers conducted controlled wind tunnel experiments at the Florida International University Wall of Wind facility, capable of generating Category 5 hurricane-force winds. Scaled models of both elevated and non-elevated buildings were tested, and the results confirmed the field observations: elevated structures experienced greater wind-related damage. These experiments provided critical validation of the observed trade-offs between flood protection and wind vulnerability.

“As a direct result of these findings, recent updates to the building codes now include wind loading force coefficients associated with elevated structures such that buildings built in the future will be designed better to withstand elevated wind loads,” Rathje said.

Expands Capabilities to Support Wildfires

Wildfire is the newest natural hazard incorporated into DesignSafe’s portfolio. 

TACC is a partner that puts the researcher first, and they realize that it’s important to build a bridge to the natural hazards research community.

Researchers began publishing data associated with wildfires in DesignSafe after the 2018 Camp Fire in Paradise, California, and since that time the interest in wildfires has continued to grow, culminating in 2025 with reconnaissance efforts capturing critical imagery from the Palisades wildfire of Los Angeles. 

“This marked DesignSafe’s first major wildfire research response following the NSF’s directive to expand its focus to include this increasingly devastating hazard,” Rathje said. “Researchers are going into the field to understand the evolution of wildfires and the damage to the infrastructure.”

Researcher Erica Fischer of Oregon State University has been actively engaged with DesignSafe and the NHERI RAPID facility to study wildfire impacts on civil infrastructure since the 2018 Camp Fire. An associate professor of Civil and Construction Engineering, Fischer has contributed multiple comprehensive wildfire datasets to DesignSafe, featuring Google Street View imagery, drone images, and high-resolution 3D LIDAR scans. 

“Numerical simulations and machine learning require significant amounts of data for validation and training,” Fischer states. “While wide-spread fire impacts on communities have been occurring in the U.S. for over a century, it is only in the last decade that we are systematically collecting post-fire data on civil infrastructure. Without this data, across all hazards, we are unable to validate our community impact simulations and are limited in training our machine learning models.”

To explore the application of AI models trained in wildfires to other hazards and vice versa, the team used the years of data stored within DesignSafe. 

“We dove deep into DesignSafe’s extensive data archives to gather drone image from other NHERI researchers,” Fischer said. “Our goal was to know whether our algorithm could also be adapted to hurricanes, earthquakes, and a broad range of other natural hazards.”

[DesignSafe] allows us full transparency of all the science that we are doing — it is quite powerful.

“The most important thing for science is that it be transparent,” she added. “DesignSafe allows sharing of data, numerical simulation models, reports, and milestones in ways that are citable. It allows us full transparency of all the science that we are doing — it is quite powerful.”

Bridging AI and Civil Engineering

DesignSafe is increasingly integrating AI and machine learning to enhance its capabilities and impact. This includes using AI for tasks like building recognition, damage assessment from images, and predicting wind pressure coefficients. Additionally, DesignSafe provides resources for researchers to utilize AI/ML in their own projects, such as through Jupyter notebooks, interactive data analytics, and access to high performance computing. 

Beyond enabling scientists to mine publicly available data for AI-driven scientific discovery, DesignSafe is also pioneering the development of AI-powered chatbots to enhance its interface, making the platform more intuitive and accessible for researchers.

“The AI chatbot will come back with a clear, textual, description of exactly what you are asking, whether it is ‘What is all the LIDAR data that’s available at a location?’ Or ‘How do I run one million simulations with the data on a high performance computer,’” Rathje said.

A notable example of AI innovation at DesignSafe is the NSF-funded Chishiki-AI project, launched in 2023 and led by AI and civil engineering experts at UT Austin in collaboration with Cornell University. Named after the Japanese word for ‘knowledge,’ Chishiki-AI aims to accelerate the integration of AI into civil engineering, advancing research, education, and practical applications in hazard resilience.

“As AI continues to evolve, we see an urgent need to educate civil engineers and equip cyberinfrastructure professionals to handle the new challenges which AI and civil engineering brings together,” said Krishna Kumar, principal investigator of Chishiki-AI,  an assistant civil engineering and affiliate faculty member professor at UT’s Oden Institute for Computational Engineering and Sciences.

For example, engineers must understand the limitations of AI, such as convolutional neural networks trained to recognize structures in environments with bright lighting or minimal graffiti, which may underperform in more variable, real-world situations. Conversely, cyberinfrastructure professionals need deeper insight into the unique demands of civil engineering, including how to deploy large-scale AI models in post-disaster scenarios where power and connectivity are limited.

“Civil engineering presents unique challenges, not only in the context of natural hazards addressed by DesignSafe, but also in emerging areas like autonomous construction,” Kumar said. “As we push the field forward, we are embracing an AI-accelerated engineering paradigm that redefines how we meet society’s evolving infrastructure needs.”

Looking ahead, Rathje envisions a deeper, more transformative integration of AI with DesignSafe data, unlocking new possibilities for research and innovation.

“We want to make the AI and machine learning training more seamless,” she said. “One way we’re working to make it easier is developing models that can automatically apply tags to images upon upload with damage information.” 

In addition, AI-enabled pre-curation of data shows strong potential to streamline the initial stages of data organization, helping DesignSafe users jump-start the upload process with greater efficiency and confidence.