Climate change is increasingly impacting U.S. solar energy generation, causing inaccuracies in pre-construction energy models, according to a new study conducted by climate and energy analytics company Sunairio.

Traditional, industry-standard solar production risk assessment approaches are based on decades-old methods that use historical data and which omit the growing influence of climate trends. As a result, pre-construction energy models over-predict solar production from new projects by up to 5–12%, according to the study. This impact is strongest in the Midwest and Southeast, followed by the East Coast — regions where climate change is causing decreased irradiance through cloudier weather patterns and greater precipitation.

While other recent studies have identified and attempted to diagnose solar asset underperformance in the U.S., the Sunairio study is the first to quantify the influence of climate change on this problem.

“We can’t expect to create accurate solar production estimates if we’re using historical weather data not adjusted for the growing influence of climate change. Those estimates get worse and worse over time, as climate change causes greater and greater deviations from historical norms,” said Rob Cirincione, CEO of Sunairio. “This is the same problem plaguing flood maps. Future weather risks are being estimated from historical data, but the climate is changing so quickly that modeling based on historical analysis can drastically underestimate the frequency of extreme events and impact of climate trends going forward.”

To conduct the analysis, Sunairio selected a representative sample of 100 actual utility-scale solar sites across the U.S. and calculated forward-looking, 15-year production estimates for each — taking into account current and expected climate-induced changes in local weather patterns informed by the company’s novel climate simulation techniques.

Sunairio compared these findings to two industry-standard methods — typical meteorological year (TMY) analysis and historical time series analysis — neither of which adjust historical weather data for climate trends. Sunairio found that TMY methods, which have changed little since their introduction in 1978, create a sunnier-than-typical year and thus overly rosy solar generation forecasts.

Across all U.S. sites studied, Sunairio found an average production gap of 2% currently, increasing to 5% by 2034. On an individual site basis, the difference between using conventional backward-looking analysis and a climate-change-aware simulation can be as high as 5% currently and 12% by 2034. For solar project owners, missing energy production translates to much lower project returns. For a project that borrows 75% of capital costs, a 5% revenue drop will cause a 20% drop in equity returns.

The current misalignment in solar production estimates compared to actual generation contributes to decreased confidence from renewable energy investors at a time when stronger investment is needed to reach the U.S. goal of transitioning to a carbon-pollution-free power sector by 2035. More accurate solar production risk assessment can help solar deals meet pro-forma targets and increase investor confidence while improving system planning to further bolster grid reliability.