One of the rare places in New York City where you can get a good view of clouds is the Central Park reservoir. Looking north from its edge, a gap between the buildings is wide enough to see them roll in from the harbour. This is where climate scientist Kara Lamb suggests we meet for a bit of cloud watching.
When we do, the sky is crowded with fluffy cumulus beneath a ceiling of altostratus – one, I venture, is very like a whale. But Lamb, who studies clouds at the city’s Columbia University, sees something less whimsical. “Clouds are fascinating because they are cool to look at,” she says. “But I think about them more from a climate perspective.” That means making sense of how the sunlight they reflect and the heat they trap below influences Earth’s temperature.
Advertisement
What casual cloud watchers may not know is that determining how this balance will change in a warming world makes clouds the biggest unknown in predicting future climate change. Will the world warm by a manageable 1.5°C or a hellish 4.5°C, given a doubling of carbon dioxide from pre-industrial levels? Our poor grasp of clouds is the biggest culprit regarding this uncertainty.
But researchers are making progress. Lamb, for one, is focused on ice crystals in clouds, which play a surprisingly large role in their climate impact. Others use cloud chambers, with plans for one tall enough to produce rain. New satellites that can see inside storms and artificial intelligences that can simulate realistic atmospheres are also helping. Of course, we may not like the answers we get.
People have long puzzled over clouds. “Indeed, can anyone understand the spreading of the clouds and the thunderings of His pavilion?” asks the Bible’s Book of Job. Walking with around the reservoir, she recalls rewatching Carl Sagan address the US Congress in 1985 to explain the basic science of climate change. At one point he mentioned uncertainties about future warming due to clouds. She is struck that not much has changed. “It’s really the same fundamental scientific understanding he was expressing 40 years ago,” she says.
We do know the features of clouds that matter most for Earth’s energy balance. Top of the list are their total area, altitude and optical depth, which is a measure of how much light they block. For instance, low, bright clouds have a strong cooling effect because they reflect lots of sunlight without trapping much of the thermal radiation, or warmth, rising from Earth’s surface. Conversely, high clouds like wispy cirrus can have a warming effect because they trap heat below them without reflecting much sunlight back into space.
Clouds and climate change
We also know that while clouds contribute to cooling overall, climate change will alter their characteristics in ways that will reduce that cooling. We just don’t know by how much. “It’s a rather serious situation that we’re in,” says at the US National Oceanic and Atmospheric Administration in Colorado. “Clouds are not well-understood enough for us to understand just how much they will be offsetting that climate change.”
The global climate models we use to predict the climate decades from now rely on a combination of equations that draw on physics and statistics to represent the changing atmosphere. But these “general circulation models” are relatively coarse, and even when run on the most powerful supercomputers they are unable to resolve the small-scale processes that affect clouds. “It’s like trying to study cells without a microscope,” says at the National Center for Atmospheric Research in Colorado.
To make do, climate modellers use dozens of parameters based on observations of clouds to represent their “overall statistical effects”, says Lamb. “We can model maybe a single cloud or a field of clouds with relatively high accuracy, but if you want to take that and put it into a climate model, you have to simplify it a lot more.” That simplification adds a large degree of uncertainty to our best estimates of how clouds will respond to rising temperatures.
Just how much uncertainty was made clear in a landmark study by at the University of New South Wales in Australia and his colleagues. They in clouds caused by global warming that will, in turn, influence climate change. These cloud feedbacks include less cloud cover overall, a shift of mid-latitude clouds towards the poles and more water in icy clouds (see “The cloud forecast”, below). By their estimate, the total effect of the feedbacks ranged from no impact on the amount of warming to a substantial increase.
The other major source of uncertainty alongside these cloud feedbacks is the role of particles suspended in the atmosphere called aerosols. From dust and the soot emitted by power plants to floating fungi, aerosols serve as seeds on which water droplets or ice crystals can form and create a cloud. “Clouds have a very, very difficult time forming in the absence of these particles,” says Feingold. The size, type and number of particles affect the brightness of clouds by changing the number of droplets or crystals in them.
Understanding aerosols
Current climate models can represent how aerosols affect the brightness of clouds fairly well, but the models don’t do a great job of capturing the particles’ knock-on effects, such as changes to the lifetime of a cloud, says at the University of California, San Diego. Better understanding becomes particularly important as we reduce air pollution from burning dirty fuels, which decreases the aerosols in the atmosphere. Since 2020, for example, shipping has been forced to clean up its act. Yet estimates of how much warming this will cause vary widely. This is concerning given the current geoengineering debates about spraying aerosols into the sky to reduce warming, says Watson-Parris. “If we can’t even measure the effect of reducing the emissions of the entire globe’s shipping fleet by 80 per cent four years on, what hope do we have of managing an intervention?”
Even seemingly small details can have outsized significance. The current focus of Lamb’s research, for instance, is how ice crystals grow in high-flying cirrus clouds. Crystal size matters. “It is one of the most important, sensitive factors in the model,” says Lamb, who has been called the “ice-cloud puzzle master” for her work. Depending on how ice crystals in cirrus clouds are modelled, the predictions show a huge range of warming equivalent to doubling the amount of CO2 in the atmosphere, she says. Yet there is no clear understanding of how aerosols and other factors affect the formation of this ice high in the sky. (Unfortunately, Lamb and I didn’t see any cirrus during our walk in the park; the upper atmosphere was obscured by clouds.)
Nevertheless, researchers are making headway in their efforts to model clouds with greater precision. As a result, cloud-related uncertainty in the major Intergovernmental Panel on Climate Change (IPCC) reports on climate change decreased by around 50 per cent between 2014 and 2023. “The needle is gradually moving,” says Watson-Parris. While that is good news for cloud science, it may not be good news for the climate: between the two IPCC reports, the projected warming effect due to cloud feedbacks increased by about 20 per cent. Still, the magnitude of this effect remains unclear, says Watson-Parris, with projections ranging from negligible to, in the most extreme emissions scenarios, .
The diversity of clouds is such that narrowing this range of possible futures won’t come from a single breakthrough, but from many researchers coming at the problem from different directions, says Feingold. For instance, two recent studies addressed uncertainties around how huge storm clouds called tropical anvil clouds will respond to warming. These develop when water and ice rising in a cloud hits the boundary between the troposphere and the stratosphere above it and spread out over an area much larger than the storm itself, resulting in a distinctive anvil shape. These formations can have a cooling or a warming effect depending on their area, altitude and thickness.
Earlier studies suggested global warming would boost the cooling power of such clouds by reducing the thinner areas of the anvil that produce a heating effect. But , at the University of Washington in Seattle and his colleagues found that the thick part of the anvil, which has a cooling effect, would decline even more. This would tip the balance towards a net warming effect. The second study, based on satellite measurements of tropical anvil clouds, discovered that they are as temperatures rise.
Other researchers are turning to cloud chambers to better understand cloud formation at a much finer level. “We have a machine in our lab that allows us to create a well-controlled, small piece of atmosphere,” says at Michigan Technological University, which hosts the only research cloud chamber in the US. “The big models are capturing all the details on a large scale. We have this box where we know almost everything… it’s reductionism at its finest.”
The chamber enables researchers to scrutinise how the distribution of water throughout a cloud is affected by different factors such as aerosols, turbulence and temperature, information that is crucial to understanding its brightness. This can come down to tracking how colliding droplets interact in painstaking detail: “the millions of coalescences that lead to a raindrop”, says Shaw. To better understand such interactions, he is part of a consortium called the (ACDC-2) collaboration. The group is designing a bigger cloud chamber that, at 10 metres high, will be tall enough for water vapour to form drizzle. “We’re not only going to have aerosols in clouds,” says Shaw, who thinks of the chamber somewhat like a particle accelerator for cloud science. “We’re going to have drizzle. So, we’re going to have collisions.”
Yet more researchers are seeking to reduce uncertainty by developing a new class of high-resolution climate models able to represent the atmosphere on a 1-kilometre grid. “Global cloud-resolving models are something I’m really excited about,” says Sokol. “Then it’s a realistic planet.” The drawback, however, is that these models require so much computing power they can only make near-term predictions or simulate limited areas, and they don’t yet include ocean, ice or biosphere dynamics. Here, machine-learning algorithms trained on those cloud-resolving models could play a supporting role. One such model, recently released by Google researchers, can accurately represent clouds using far less computing power than physics-based models. However, Watson-Parris says it is important to keep as much physics involved as possible to ensure projections are reliable.
The promise of EarthCARE
Improvements to climate models will also come from better observations of real clouds. One important source of this is a satellite launched in May called the Earth Cloud, Aerosol and Radiation Explorer (EarthCARE). A joint mission of the European Space Agency and the Japanese Aerospace Exploration Agency, it is designed to answer questions about the sun-reflecting and heat-trapping effects of clouds and aerosols. It will provide the highest-resolution view of this yet, says at the European Centre for Medium Range Weather Forecasting.
EarthCARE promises greater detail than other set-ups by combining multiple instruments on a single satellite to enable simultaneous measurements of clouds, aerosols and the incoming and outgoing radiation that influences warming. EarthCARE is also equipped with lidar, a laser-based detector, that can measure the shape and type of aerosols, and Doppler radar that can measure the speed of falling particles inside clouds. In June, the radar , showing a cross-section of rain and ice falling out of a storm. The satellite is expected to begin delivering complete data in 2025, which will be used to test how well climate models are representing clouds, and to make improvements to them.
“EarthCARE is not going to give us all the answers,” says Forbes. “But I think it will give us a step change in terms of how we represent the clouds.” With all these efforts to solve the cloud conundrum, perhaps the fog is finally lifting.
The cloud forecast
Ongoing global warming will change clouds and those changes will, in turn, influence how much the planet heats up. These are known as cloud feedbacks. They come in many different forms, but three in particular are expected to increase warming.
Loss of clouds at sea
Marine boundary-layer clouds – low-level formations found over much of the ocean – can cover more than a third of Earth's surface at once and have a substantial cooling effect by reflecting sunlight. However, warmer temperatures will lead to drier air above the ocean, which is expected to reduce the area covered by these clouds, resulting in the most significant of the warming cloud feedbacks.
High clouds get higher
Clouds rise when they are warmer than the surrounding atmosphere, and hence less dense than it. In a warmer world, the altitude at which they stop rising tends to increase. This is a problem because higher clouds have a stronger warming effect. Overall, this is expected to result in a moderate increase in warming.
Fewer clouds over land
Hotter temperatures on land mean less water vapour in the air. This leads to fewer clouds at all levels, but especially those at low altitudes, which reflect lots of sunlight back into space. This feedback is expected to have a small warming effect.
James Dinneen is an environment reporter for 91av based inNew York
