Skip to main content
U.S. flag

An official website of the United States government

Publication Date
11 November 2022

STITCHES: A Comprehensive Solution to Emulating Climate Model Output

Subtitle
Building new climate outcome scenarios by stitching together pieces of existing ones.
Print / PDF
Powerpoint Slide
Image
|
Science

Climate models are a main source of future climate information, but they are extremely expensive and slow to run. Studies concerned with the effects of climate change on human and natural systems often need climate information for emissions trajectories not run by climate models. Previously, emulators have attempted to create climate information for these trajectories. However, they usually fall short of the information demands of increasingly sophisticated impact models. STITCHES, a new emulator, works by stitching together building blocks of actual climate model experiment outcomes to create new trajectories. By construction, STITCHES output has the same complexity and richness of climate model output to meet the needs of impact modelers. 

Impact

Producing “cheap,” comprehensive, and coherent future climate information allows researchers to study a wider range of socioeconomic futures and their consequences in a consistent modeling framework. It also allows feedbacks from climate impacts to the socioeconomic processes to be efficiently quantified and explored. That STITCHES can complement climate modeling may relieve pressure on the climate modeling community to run many different scenarios, potentially allowing experimental resources to focus on different aspects of future climate uncertainty.

Summary

STITCHES relies on the fact that many climate variables behave in sync with warming and cooling global temperatures (GSAT). Therefore, STITCHES first focuses on matching subsequent segments of a GSAT time series from a new scenario (which can be produced by simple, computationally cheap climate models) to segments of GSAT trajectories from existing experiments of complex and computationally expensive climate models. After identifying the building blocks (i.e., experiment-time-window combinations) and sequence that make up the target trajectory, the emulator can stitch together many types of climate model output from the same experiment-time-window combinations from the complex climate models.

Point of Contact
Marshall Wise
Institution(s)
Pacific Northwest National Laboratory
Funding Program Area(s)
Publication