By Daniel Dunaief
Minghua Zhang spent a sabbatical year in China trying to improve climate models, which included analyzing errors of current models.
A professor at the School of Marine and Atmospheric Sciences at Stony Brook University, Zhang focused on the Southern Great Plains of the United States. He explored how the current models did not accurately simulate convection, which created a warm and dry bias.
In convection, heat and moisture get carried upward. The models account for summer rainfall but do not calculate the organizational structure of the convective systems, which led them to simulate insufficient precipitation.
By adding in the new information, Zhang predicts in recent research published in the scientific journal, Nature Communications, that the projected warming in the region would be 20 percent less than previously thought. Precipitation, meanwhile, would be about the same as it is today, instead of the drying that was previously anticipated.
“The resolution of the models is not high enough to predict the change of the convection with a high degree of precision,” Zhang explained in an email.
He suggested that using 10 times the specificity of model calculations, he expects a clearer picture of the likely climate by the end of the 21st century. This is like looking through binoculars at a nondescript moving shape in the distance. By adding focal power to the lens, the image can become sharper and clearer.
The climate is a big picture view of trends over the course of many years. That is distinct from weather, which involves day-to-day variations and which meteorologists describe each morning and evening with colorful images of cold and warm fronts on local maps.
“You have many things you can’t see and now you have better binoculars,” he suggested. “This tiny thing in the binoculars can make a bigger impact. What we see is that these [variables] actually matter.”
Zhang suggested that a climate model that better accounts for summer rainfall still includes higher temperatures in this sensitive region. “The warming is going to be there and will be significant,” he said. If carbon dioxide emissions continue at their current rate, the warming will still be about five degrees by the end of the century, he suggested. That, he predicted, will still have a significant impact on agriculture.
Edmund Chang, a professor at the School of Marine and Atmospheric Sciences at Stony Brook who was not involved in this study, said this research addresses “a specific bias of the model that needs to be taken care of.” He added that researchers know that the “models are not perfect” and suggested that the “scientific and climate modeling community is trying to refine and improve” these tools.
Chang agreed that the refinement “doesn’t change the fact that we still project a large increase in the temperatures over the central United States.”
The Southern Great Plains region has some unique elements that make climate predictions challenging. It has considerable organized convection, which increases the occurrence of tornadoes. There’s also a large coupling between the soil moisture and the clouds, which means that whatever happens on the land has feedback for the atmosphere.
Zhang said his research focus is on climate simulation modeling. He knew the models had problems simulating convective events, which is why he started exploring this specific region. “The way the models are constructed, the granules are not small enough,” he said.
Chang expected that this work would “spur more research on trying to understand this mechanism. Model developers will need to try to find out how to improve the physical model.”
Zhang has been working for the last two years with scientists from Tsinghua University in China, which included his time on sabbatical. “When you are on sabbatical, you have more time to really think about problems,” he said.
Chang added that sabbaticals can provide some time to focus on specific scientific questions. During a typical semester that includes administrative responsibilities and teaching, professors “are very busy,” He said. “We really don’t have an extended period of time to focus on one project. The sabbatical gives us a chance to focus.”
Zhang hopes this study “motivates people to think about how to improve their models in describing” other climate systems.
One of the many challenges scientists like Zhang face in developing these climate models is that their computers are still not powerful enough to resolve elements like clouds, which not only dot the landscape and provide shade during the summer but also send the sun’s energy back into space.
The system he’s studying is “chaotic by nature,” which makes accounting for elements that change regularly challenging. He suggested that these studies were akin to the butterfly effect. Scientists have suggested that someone who went back in time and committed a seemingly trivial act, like killing a single butterfly, might return to his familiar time and surroundings to discover profound changes.
While that’s an exaggeration, that’s still the kind of system he said researchers are confronting as they try to account for, and weigh, climate defining factors. That’s why he’s looking for statistical, or probabilistic, predictions that are averaged over time periods.
The United States, China and the European Union are all pursuing more powerful computers for these kinds of applications, Zhang said.
Zhang, who is the editor-in-chief of the Journal of Geophysical Research: Atmospheres, has been involved in an advisory capacity with the United States Department of Energy in developing these models. A
s for this specific effort, Zhang said he was pleased that the paper pointed out a research direction to refine models for climate in this area. “What we see is that these things [including convection] actually matter,” he said. “That’s the main contribution of this paper.”