How to Use the Right Tools to Design for Climate Change

Legal liability is looming

The American Institute of Architects (AIA) updated its code of ethics in 2018 to say their members “should … anticipate extreme weather events” and support their clients in responding to the rapidly rising potential for such events. Later that year, a group of attorneys raised a red flag about grievously outdated weather data, on which design relies heavily. The Conservation Law Foundation warned that the practice of using historical weather files from the far-flung past could come with legal risks.

In 2022, a group within AIA’s Strategic Council agreed, saying designing for weather patterns of the past may indicate a “failure to act reasonably in the face of ascertainable climate risk.” In other words, the legal standard of care may have changed.

Two years later, the practice of designing for the past has not. But are architects to blame?

Data are catching up; codes and standards are not

The weather data referenced by our standard procedures, codes, ASHRAE 90.1, and other standards still do not anticipate how conditions will shift with climate change. In most cases, these data and practices don’t even reflect the most recent past.

Luckily, the U.S Department of Energy (DOE) has promised to finally make future-looking weather files available to the design and construction industry by the end of 2024. These files will supposedly check every box. They will:

• Predict future weather under multiple climate scenarios and for many locations
• Be distributed for free by a trusted source
• Live in a file format that mechanical, electrical, and plumbing (MEP) designers and engineers actually use

This is exactly what the climate advocates have been calling for. But is this problem solved?

Not entirely—or at least not yet. And it could be years before consistent guidance and standards emerge. Because until codes and standards reference future weather files, design professionals have to make decisions on their own about how and when to reference these files. And that leaves the legal standard of care in limbo.

But the ethical standard of care is not. With plenty of high-quality data available on how future weather is likely to play out under different climate scenarios, doing our best to design for climate change’s new normal is a moral imperative.

But how are people navigating all that? Here, we explore the generally accepted—and more fringe—use cases for future weather files. 

Use case 1: improve the accuracy of modeling tools

Weather files are the data source underlying almost all of the tools we use to model how a proposed building design will perform. They are used by multiple disciplines, according to a recent report from HGA, using dozens of software programs, including:

• Climate Studio (daylighting analysis, glare analysis, and façade radiation analysis)
• Design Builder (energy sensitivity studies and conceptual HVAC design)
• WUFI (hygrothermal analysis)
• Coolvent (bulk airflow analysis and computational fluid dynamics)

These programs, and many more, are used to predict and understand a building’s performance over its lifetime. Presumably, then, a future weather file, rather than one based on weather patterns from 20 or 30 years ago, would more accurately represent the conditions a building will actually experience. Many software tools already allow a user to upload their own weather files, and quite a few professionals already choose to use weather files from more recent history than is typical for the profession. 

For example, the most widely referenced weather file for energy modeling is the TMY3. These “typical meteorological year” files are based on a hypothetical model. Researchers from the National Renewable Energy Lab (NREL) looked at weather conditions over multiple years for a given location and created an hourly dataset to represent “normal” weather for that area. However, these files are built from weather data between the years of 1991 and 2005, missing data points from the years that brought us Hurricane Sandy, two decades of California wildfires and a few atmospheric rivers, numerous recent “heat domes,” and the 2021 Texas ice storms.

TMY3 (the 3 indicates the third iteration of this file type) is the default within most energy simulation engines, including Energy Plus.  However, many leading professionals already run future scenarios by substituting weather files from Climate.ONEBuilding.org, according to Victor Braciszewski, P.E., of SmithGroup. The TMYx files (the “x” distinguishes this file types from the ‘official’ NREL versions) provided on that site were created in the same manner as the TMY3 and by some of the same researchers, but they use meteorological data from 2007 through 2021.

Energy codes permit this. Both ASHRAE 90.1 and the International Code Council (ICC) energy codes prescribe only the characteristics of allowable weather file datasets—not which datasets must be used, according to Braciszewski. So picking a more recent dataset from a reputable source isn’t frowned upon. But it also isn’t explicitly supported. That means the modeler must take some additional steps of their own volition, without consistent standards or guidance to lean on. (Some states have required modelers to use specific weather files that the state created from a more recent time horizon, as is the case in California with the CALMAC weather files. But it isn't the norm.)

This leeway in the code also allows for using future weather files for energy modeling. Leading professionals focused on climate adaptation already purchase future weather files (fTMYs—where the “f” stands for future) from Weathershift, a service provided by Arup and Argos Analytics. And, the DOE project is intended to make fTMYs available for free, backed by a rigorous and transparent development process. These DOE files will arguably be better than Weathershift because they will be “downscaled” from global climate models to specific geographical locations in a way that takes highly localized, topography-dependent weather variability into account. (This is because they will use a combination of dynamic and statistical downscaling. The HGA report referenced earlier highlights the differences.)  

Using multiple fTMY files for energy modeling could increase confidence that a given building will meet the energy code or its operational carbon goal despite expected changes in weather under different climate scenarios. This process could also help a user make further tweaks if the performance falls short under different scenarios. Running multiple scenarios is key here because we don’t yet have certainty about what future weather will bring. The Intergovernmental Panel on Climate Change (IPCC) has established many possible futures to capture what might happen if the global economy decarbonizes rapidly or not at all. These are called ‍‍Representative Concentration Pathways (RCP). So when using fTMY files, design teams typically take multiple scenarios into account.

“We should not be using one set of climate data,” according to Nathan Kegel, vice president of Integrated Environmental Solutions. “We have to be able to understand a wide range of outcomes.” Pulling different fTMYs for different climate scenarios helps develop that understanding, he says.

Used properly, fTMY files could also help users futureproof in areas beyond energy consumption. For example, when making decisions about:

• shading for peak loads
• glazing optimization
• managing moisture in the building envelope
• natural ventilation strategies

In all these cases, the tools and method of doing the analysis won’t change: the background assumptions about weather data will hopefully just more accurately reflect what a future building will experience. The experts that BuildingGreen spoke with generally thought that using future weather files for performance modeling was a good idea. Who could resist having a more accurate model?

Using future weather files would catch on more quickly, added Braciszewski, if the energy code required them and if modeling tools like Energy Plus made it easy to reference them. Then enforcement might become an issue. Kegel noted that even where more recent historical files are mandated, code reviewers don't always check that the right files are used.

Use case 2: stress test for extreme events

Some of the future weather files being developed promise to enable new kinds of analysis—notably, stress testing for extreme events by seeing how a building would perform during an extreme weather event. While a TMY file is meant to represent a “typical” year, explicitly downplaying outliers, there’s another type of weather file called the Extreme Meteorological Year (XMY) that attempts to capture the impact of extreme weather events. To date, there has not been a widely accepted method for creating an XMY file. However, part of the DOE work is for Sandia National Lab to develop a tool called the Multi-scenario Extreme Weather Simulator (MEWS). 

So far, this tool has generated weather files to show the shifts in intensity and frequency of 10- and 50-year heat waves. Next, the same algorithm will be applied to hurricanes and extreme precipitation. 

A logical use of these files would be to stress test a given design to understand how it might handle an extreme event. Will the roof drainage system be overloaded? Will the mechanical systems be able to keep temperatures within the desired setpoints? Will natural ventilation still work if temperatures don’t drop as much at night? The tricky part is deciding what one should do with the results.

“There is a difference between understanding how your building is going to perform and looking at whether your building is going to freeze on the coldest day of the year,” Braciszewski told BuildingGreen. “That is a different set of liability.”

For those system-sizing questions, MEP engineers have typically referenced the Climatic Design Conditions in the ASHRAE Fundamentals Handbook. This section prescribes the climate conditions an HVAC system should be designed to accommodate. But are those conditions currently based on future weather files? No, they are not. Most are based on calculations that consider the 26-year period from 1994 to 2019 (ASHRAE Fundamentals 2021). 

Nevertheless, ASHRAE sets the legal standard of care for the industry.

“If my heat pump isn’t sufficient, I better be able to point to the methodology that I used to size it,” says Braciszewski. Engineers know that running an ASHRAE calculation with a design-day value will hold up as a reasonable degree of care. Most projects haven’t even used the Multi-scenario Extreme Weather Simulator. And it could be a long while before this question is brought before a court.

And yet there are also consequences to sizing systems based on weather data that we know are outdated.

Kim Shinn, P.E., of TLC Engineering Solutions has begun using fTMY files from Weathershift. Shinn is finding that even without capturing the extreme weather events, the precipitation values in those files indicate that the plumbing engineers need to be designing for more roof drainage than they otherwise would. That might be because the International Plumbing Code’s (IPC) rainfall rate intensity map is subject to a vote by its membership, so there have been few “wholesale changes” to the map, according to Shinn. IPC also allows plumbing designers to use the National Weather Service’s Precipitation Frequency Data Server (PFDS), which “generally reflects higher rainfall intensities,” says Shinn, but those data are still historical.

Weathershift also indicates designers should be making different decisions about the infrastructure for mechanical equipment. Under RCP 8.5, the “worst-case” pathway where emissions continue to rise until 2100, (see Demystifying Climate Change Lingo: From Scope 3 to SSP), mechanical equipment may need to be sized differently by 2080. Although that time horizon is beyond the lifespan of the equipment Shinn is designing today, the results indicate to him that they should be future-proofing current designs for future weather. That means allowing more space for bigger equipment and distribution systems to be installed in the future, for example, by increasing the size of:

• Main duct runs
• Vertical piping shafts
• Mechanical runs

These spaces in a building are not easily upgraded once a building is built.

Shinn presents this information to owners and lets them make their own decisions about whether to accept the risk or upsize certain systems at an increased cost. He also ensures that when he does use fTMY files for sizing, his recommendations are more conservative than ASHRAE’s current guidance. This way, he can argue he “exceeds the standard of care.”

Some engineers have held back from using Weathershift, partly because it is a product sold by a competitor engineering firm, and partly because it is not an open-source tool with full transparency around the methodology. But even when DOE makes fTMYs freely available, many engineers may not be comfortable choosing which greenhouse gas scenarios to model for clients.

“You have to understand that most mechanical engineers are used to following traditional methods. They may not be comfortable with picking which global climate policies are more likely,” says Braciszewski. He’s currently exploring the impact of future weather data, but “it is going to be tough to change the status quo.”

To really change the paradigm, ASHRAE will need to change its Climatic Design Information chapter, and other standards and guidelines, to incorporate the future-weather-file set that will be provided by DOE. This hurdle was forecast by DOE, which has stated that “success is getting building codes, ratings, and standards to require using future weather as part of their design criteria.”

Are we kicking the ball down the road (again)?

While future weather files like the ones DOE is developing are “twenty years overdue,” according to Kegel, he wonders whether we’ll still be left with an underlying problem once they’re available.

“It is really important that we [the AEC industry] have access to weather data that can be kept up to date without relying on a grant proposal,” Kegel argues. Adoption of future weather data is likely to be higher because the files will be free, but the moment they are released they will be outdated, he says.

Kegel thinks artificial intelligence might be used to keep updating the files as future weather models are released. Otherwise, he said, we’ll be seeing the same calls for an update in a few years.

More on designing for future weather

The Future of Designing with Future Weather Data

Future Climate and Professional Liability: AIA Weighs In

Future Heat Waves Will Kill Power and AC—and People