UVic climate researchers take us from projection to prediction

Climatologists study historical climate change using observational data, which only became more complete and globally consistent after the 1950s. As a result, we don’t fully understand the early climate. Also, of course, the future remains unknown. Climate models help simulate both past and future conditions but they come with large uncertainties, which make it harder to know how what we’ll actually see and feel in the future differs from the past — and how best to prepare for changes in events like floods, wildfires and droughts. 

“Using mathematics and statistics, we can combine observations of past changes with simulations of the past and future to reduce uncertainty about the unknown,” says Francis Zwiers, scientist emeritus at the Pacific Climate Impacts Consortium (PCIC). Seeing the potential of this approach, Tong Li — a post-doctoral fellow and lead author of a study recently published in Science Advances — joined forces with Zwiers and Xuebin Zhang, director of PCIC, to tackle the challenge. 

In their article, the research team shows that their newly implemented method can significantly reduce uncertainty in future global warming projections. As a result, these “constrained projections” should be viewed not just as hypothetical possibilities, but as skillful predictions with a clear understanding of where that skill comes from. 

In parallel with looking into the future, Li’s innovations also improve how these techniques are applied to the past.  

“I wanted to explore how we can go back in time — to periods when human activities like fossil fuel use did not strongly affect the climate. This helps us better understand past climate change, including the part that can be attributed to human activity,” says Li.  

Thus, the team can now tell policymakers how much different regions of Canada have warmed since the 1850s. Zwiers explains that systematic collection of weather observations across Canada began as the railways developed from late in the nineteenth century but that there isn’t enough data for all of Canada before 1948 to combine with climate models. The methods the team use allow them to improve estimates over the whole period starting in 1850 to the end of the present century. Once the team finishes considering Canada, it will then also begin to consider different regions across the world, including many where observed data are severely lacking.  

“As a high-latitude country, Canada contains many different climate regions, and experiences amplified warming in its northern areas — making it an especially valuable case for climate research,” Li says.  

Li’s next research challenge sounds a lot like an existential question.  

“How can we travel back in time and make use of climate models that look forward?” she asks. “How can we look into history and the future at the same time?” 

Her next paper, we predict, will tell us.