Home > Technology + Innovation

Regional Climate Modeling

Regional Climate Modeling

The Department of Energy’s Argonne National Laboratory is providing actionable information to companies, agencies, cities, and utilities.

By JD Amick

As the global effects of climate change become more and more apparent, the impacts of extreme climate and weather events are increasingly taking their toll on infrastructure systems within the United States. The Department of Energy’s Argonne National Laboratory is using their leading-edge climate modeling and infrastructure science capabilities to provide actionable information to companies, agencies, cities, and utilities to help safeguard their infrastructure in the face of these challenges for decades to come.

While global climate modeling data is readily available, Argonne is using its unique combination of supercomputing capabilities and climate modeling and infrastructure science expertise to use that data and turn it into higher resolution, local scale climate impact models.

“Argonne takes this a step further,” explained Tom Wall, Senior Infrastructure and Preparedness Analysist and leader of Argonne’s Climate Impact Data and Decision Support effort. “Our team at Argonne takes these raw climate variables and turns them into climate impact data tailored to the specific needs of a particular company or agency.”

The first element to the development of this climate impact data is to “downscale” the global models that display data on a scale of 100-150 km blocks. As Atmospheric and Earth Scientist Jiali Wang describes it, “There are two ways to take low-resolution global and regional climate models and localize them to a neighborhood level: statistical downscaling and dynamic downscaling.”

Statistical downscaling is used to form statistical relationships between global/regional climate models and local observations to “back-cast” historical climate trends of a particular area in order to infer projections about that area’s future climate.

While this method of downscaling does not require much relative computing power, and is therefore widely available, it has significant shortcomings. “It functions on the assumption that these statistical relationships will remain constant over time,” Wang elaborated. “But this isn’t the case. In the face of climate change, we are seeing these relationships and distributions change over time.”

At Argonne, the climate modeling team uses dynamic downscaling to produce high-resolution climate models on a scale of 12 km blocks. Dynamic downscaling refers to using atmospheric physics- and chemistry-based models to dynamically extrapolate future regional climate, in order to produce high-resolution, “neighborhood” scale projections.

This is a massive amount of data, however, and it is processed to create hourly climate projections for the next several decades. One year of these hourly projections would take more than 4,000 hours of computation time on a standard computer. Argonne’s supercomputing capabilities allow the climate modeling team to process this data on a large scale at incredible rates, finishing more than 50 million total core hours of computing in mere months. “We are using these capabilities to explore further downscaling to 4 km blocks to improve the model’s performance, reduce bias, and boost our overall confidence in future projections,” Wang explained.

Argonne researchers achieved this using a combination of modeling systems such as the Weather Research and Forecasting (WRF) model and the WRF-Hydrological (WRF-Hydro®) system for modeling regional climate at the 12 km scale and inland flooding due to precipitation down to even the 200 m scale. In conjunction with the Advanced Circulation Model (ADCIRC) for modeling coastal flooding due to storm surge down to the 50 m scale, the research team at Argonne achieved this detailed, high-resolution “neighborhood” modeling.

Argonne recently collaborated with AT&T, using these climate resiliency science capabilities to analyze and address issues with AT&T’s network infrastructure in the Southeastern United States, a region that has been hit with some of the largest hurricanes and other extreme weather events in recent history.

According to a CNBC article, such natural disasters have cost AT&T $847 million since 2016. As part of this collaboration, AT&T paid Argonne $375,000 to create a Climate Change Analysis Tool to model the impacts of coastal and inland flooding, as well as wind for the coming 30 years in Florida, Georgia, North Carolina, and South Carolina.

Moved by strong concern for repeated losses from similar natural disasters in the Southeast, AT&T worked with Argonne to identify this region as the target for developing the climate impact model. Throughout this process, Argonne helped determine what data would be useful for AT&T in terms of safeguarding their infrastructure.

“We needed to provide more than just information on precipitation values,” explained Eugene Yan, Principal Scientist in Earth Science. “We had to bridge the gap between these complex models and these industries by showing how climate change would directly impact their network infrastructure.”

The team used the dynamically downscaled climate models to illustrate the impact of extreme climate change in the future. In order to do this, the team used generalized extreme value distributions to identify what “extreme” weather might look like in the coming decades and with what frequency it might occur.

This information directly informed an analysis of Intensity Duration Frequency (IDF) curves for coastal and inland flooding in this region, a fundamental tool used when planning infrastructure. “As a critical piece for civil and structural engineers to incorporate into their designs,” Yan explained, “IDF curves offer information on probability and potential danger of flooding due to hurricanes and heavy rainfall.”Using the Climate Change Analysis Tool, Argonne researchers were able to improve the analysis of IDF curves in this region to more accurately inform engineering decisions. The Argonne team used non-stationary, time dependent parameters in conjunction with historical data to create what Yan called the “next generation IDF curves.” Yan further explained that these impact models are novel in that they not only predict what future extreme climate events will look like, but also quantify the uncertainty around these predictions that is inherent when combining climate projections from numerous global models.

Not only would the impacts of flooding on buildings and infrastructure vary dramatically from block to block, but so too would the impacts of wind levels. This modeling for future extreme wind levels can greatly inform wind loading needs for buildings in the future.

Argonne is currently working with AT&T and other parties to expand the Climate Change Analysis Tool to encompass the impacts of forest fires and other extreme climate effects and weather events.

“This work goes beyond the physical structures themselves and potential economic losses,” Wall elaborated. “At the core of any engineering work is a commitment to public safety.” Communication services, utilities infrastructure, physical infrastructure, are all critical to safety and relief efforts during natural disasters and weather emergencies. “The Climate Change Analysis Tool is crucial for preparing and safeguarding this infrastructure and these services in the decades to come.”

Organizations interested in learning how they might tap into Argonne’s expertise, facilities, and tools to unlock technical challenges and seize opportunities—including tools to help safeguard their infrastructure from climate events—should contact partners@anl.gov

JD Amick is a freelance writer based in Chicago.