This article was originally posted in March 2019. It has been re-posted here following a server issue in which the original post was accidentally removed. To learn more about the most recent work from Dr. Tapio Schneider’s research group, some of which uses machine learning to improve climate modeling, visit his research group’s website.
In Spring 2018, I visited the Kavli Institute for Theoretical Physics (KITP) in Santa Barbara, California for a conference on Frontiers in Oceanic, Atmospheric, and Cryospheric Boundary Layers. For five days, oceanographers, atmospheric scientists, and climate researchers convened to discuss the many ways in which planetary boundary layers – horizontal slabs of air and water constrained by interfaces like the ocean surface – influence our weather and climate. I went there to present a poster for my own work on the effects of surface waves on wind. In addition, I spent some time enjoying the beautiful spring weather before wildfires ravaged nearby forests that summer amid unseasonably warm and dry conditions.
One of the invited speakers – Dr. Tapio Schneider of the Climate Dynamics Group at California Institute of Technology – discussed how global climate models (GCMs) struggle to accurately predict climate without representing more complex cloud physics – namely, cloud formation processes that are intricately linked to chaotic air movement that occurs within atmospheric boundary layers. A big challenge for climate researchers is to figure out how to tackle the low spatial resolution climate models typically use. Within a network of interlinked grid cells, complex, multi-scale motions in the atmosphere are typically averaged into one value over tens of kilometers (typically 60 by 90 miles).
Considering clouds and climate
In this coarse simulation grid, boundary layer processes like wave growth and cloud formation are too small in scale to simulate with much complexity; instead, climate models account for them by assuming that their large-scale impact can be averaged over large stretches of space using empirical data. With current climate models struggling to recreate past climates, researchers like Dr. Schneider question whether important details about the complexity of small-scale physics are being overlooked.
Dr. Schneider and co-authors sought to address this issue in a study that made headlines. By using a model called large eddy simulation (LES) that resolves the spatial scales of chaotic motion in clouds, they inferred how the atmosphere might respond to future conditions in which humans have done little to curb carbon dioxide emissions. They simulated the cloud dynamics and solar energy balance over a patch of ocean in the subtropics, where layers of low-lying, highly reflective stratocumulus clouds are known to form amid persevering wind patterns.
It’s all about balance
Energy from the sun takes a number of paths when it enters our atmosphere. Some of it travels through the entire layer and warms the Earth’s surface, while some is reflected back to space by bright surfaces like clouds and ice. At the same time, gases such as water vapor (moisture) and carbon dioxide in the atmosphere easily absorb the sun’s energy in the form of longwave (or infrared) radiation – this is what is known as the greenhouse effect (see Fig. 2).
All of these different pathways amount to net zero planetary heat gain if the atmosphere’s overall temperature isn’t changing – but it is. The idea that warming could become worse, and even beyond our control, has troubled many climate researchers who note that ice and clouds – both reflective surfaces that maintain the Earth’s balance of solar energy – are expected to be reduced in warmer conditions.
The results of this simulation suggest that stratocumulus clouds – which are known to deflect the sun’s energy over a significant portion of the world’s oceans – could begin to break up and eventually disappear altogether as carbon dioxide concentrations rise. Today, those concentrations are around 400 parts per million (ppm) on average. What happens when they rise to 800 ppm, or even 1600 ppm – which could be possible in the next 100 to 200 years if we do not reduce carbon emissions?
In addition to 4°C of warming from these emissions, the study suggests that the loss of stratocumulus clouds could add an additional 8°C to average temperatures in the atmosphere. This degree of warming has been contextualized by researchers as a condition similar to when crocodiles and palm trees inhabited the Arctic over 50 million years ago. In this simulation, clouds did not return to Earth until carbon dioxide concentrations dropped below 300 ppm – a much lower level than even our current conditions.
The cause of a cloudless world
According to the study, a world without clouds could be partially caused by a more opaque (carbon-filled) upper atmosphere above the height where these clouds typically form. More carbon dioxide in the upper atmosphere means more solar radiation is absorbed at these heights, bringing its temperature closer to that of the warmer lower atmosphere. This warming weakens convection – the chaotic mixing driven by temperature differences that connects the upper atmosphere to the sea surface by vertical movement of air. With weaker convection, the atmosphere has a weaker connection to its moisture supply at the sea surface. At the same time, a warmer environment would also result in more evaporation that provides more moisture to the lower atmosphere. However, because of weak convection, this extra moisture would not reach the upper atmosphere to form clouds; instead, it could draw in warm, dry air from nearby that would further destroy clouds.
The study notes some uncertainties on larger scales; connecting small-scale phenomena like clouds to weather and climate patterns is not an easy task. Regardless, these findings highlight the significance of clouds in the climate system, and a key shortcoming in our current GCMs. Perhaps the little things matter after all.
I’m a PhD student at the University of Rhode Island’s Graduate School of Oceanography. I use a small-scale computer model to study how physical features like surface waves at the air-sea interface produce friction for the wind that can limit momentum, energy, gas, and heat exchange between the ocean and atmosphere. In the future, I hope to learn more about the role waves play in different parts of the world as weather and climate patterns evolve. Also, I love to write.