Note: This piece, which I penned alongside Carla Roesch, PhD, Colleen Golja, PhD, and Catherine Ember, all of whom I work with at ARC, originally ran on ARC’s Substack here. You should sign up for that substack as well.

Most people, if asked how global warming works, will give you a version of the same answer: greenhouse gases trap heat that would otherwise escape to space. That answer is correct. But it is also only half of the equation.

The rate at which the planet warms is set by Earth’s Energy Imbalance (EEI): the difference between the sunlight Earth absorbs and the thermal radiation it releases back to space.

Greenhouse gases act on the outgoing side, reducing the amount of heat that escapes. How much incoming sunlight gets absorbed, is a far messier question, governed largely by Earth’s albedo, or how reflective its surface is. Bright surfaces like ice, snow and clouds reflect sunlight back into space preventing it from warming the planet. As these surfaces have shrunk in recent decades, the planet has become less reflective, absorbing more solar energy and increasing EEI. Satellites such as NASA’s CERES instruments have recorded this change: a steady rise in absorbed radiation since 2001 coupled with an observed 0.6% decrease in Earth’s reflectivity.

That shift looks small on paper, but its consequences are not. By some estimates, the (direct) warming contribution from this loss of reflectivity is now contributing about as much to EEI (and thus the rate of warming) as all CO2 emissions since 1750.

Taken together, rising greenhouse gas emissions and Earth’s declining reflectivity have roughly doubled to tripled EEI since 2000 (depending on whether we look at 10 year averages, or more recent numbers), and we’re now seeing the effects on temperature, with clear evidence that the rate of warming in the last 10 years was double previous decades. Cutting greenhouse gas emissions remains, rightly, the central focus of climate policy and is a non-negotiable priority. But the loss of reflectivity of the planet’s surface is starting to draw focus on its own (see for example this recent piece in the Economist). We don’t yet know if the rate of warming will sustain, continue to accelerate, or revert back to the long-term average. But, so far, the EEI trends show no sign of slowing, which is cause for serious concern.

Identifying where Earth’s albedo is falling is relatively straightforward using satellite data. Losses of ice and snow, along with other surface changes such as urban development and deforestation, account for roughly 27% of the global decline in reflectivity. The majority of the remaining share is due to reduced cloud cover worldwide.

We have a decent understanding of what drives the surface changes. Warming temperatures and human developments are well understood. What drives the reduction in clouds is much more complicated. Two mechanisms appear to be dominant:

Aerosols include anything from smog to dust to sea salt spray; one that matters most for Earth’s energy balance is sulfate aerosols. These are formed from gases emitted from fossil fuel combustion, industrial processes, as well as naturally through volcanic eruptions. Sulfates are reflective in their own right (they scatter incoming sunlight), but more importantly are extremely hydrophilic, which means they are great CCNs. As a result, anthropogenic aerosols have been masking a meaningful share of global warming for decades, reducing current temperatures by an estimated 0.5°C. In other words, aerosols increase Earth’s reflectivity, decreasing EEI; without them, the imbalance would be even larger than it already is.

That said, sulfates and other aerosols are extremely harmful to human health, which has driven a sustained global push to cut them, primarily through reductions in SO₂ emissions. Cleaner fuel use and wider adoption of retrofittable scrubbing technology at industrial facilities and power plants have cut global SO₂ emissions by ~50%+ since 1970. Perhaps the most consequential recent move came in 2020, when the International Maritime Organization introduced new limits on sulfur content in shipping fuel, abruptly cutting international shipping-related SO₂ emissions by ~80%. This change alone drove a large reflectivity loss; its warming-equivalent increase is estimated at up to three years of global CO₂ emissions (at current emissions rates). One regulation, made unilaterally, likely raised global temperatures by up to ~0.17°C.

Recent studies show that the climate models that most accurately reproduce the observed albedo trends exhibit higher climate sensitivities (the amount of warming the planet will experience under a doubling of CO₂). If these models are more accurate in this regard, i.e., if climate sensitivity is higher than many other models assume, it could suggest that future warming may be more extreme than the ~2.9°C, which, as noted recently by climate scientist Zeke Hausfather, is the median temperature estimate for the year 2100 under the recently updated “medium” scenario (which attempts to forecast future warming trajectories and assumes ("Current emissions policies frozen at 2025 levels").

For climate mitigation and adaptation, it is imperative that we reduce these uncertainties to narrow the ranges of possible climate futures as much as possible so we can advance efforts across many other disciplines and in other domains in more targeted fashions.

That open question motivated development of CLARO, the CLouds, Aerosol, and Reflectivity Observatory (CLARO). CLARO is not a climate intervention program. It is a measurement, modeling, and observation platform designed to resolve the uncertainty around aerosols, clouds, reflectivity, and Earth's energy (im)balance, which, some argue, is the largest single source of uncertainty with respect to climate sensitivity. We’ll share more details on the program itself in a future newsletter later this fall – stay tuned, and in the interim, you can explore more here.

ICYMI the latest cool news out of SIlicon Valley, and because I’m probably not getting to a news roundup-type article anytime soon, I wanted so shout ot Network Ocean, which builds data centers in the ocean with solar panels on to (full disclosure, I’m a proud investor). The company itself noted, “Last week [two weeks ago now], NetworkOcean made history as the 1st to run a customer AI workload from the ocean, powered by solar. We went from design to deployed in the water in 53 days.”

There’s a lot to prove in the data centers at sea space, but exciting stuff notheless. Hats off always to people who take the proverbial plunge vs. arm-chair experting. More here.