Researchers at the Hong Kong University of Science and Technology (HKUST) developed an improved method and found that tropical cloud feedback may have amplified the greenhouse effect by 71% more than earlier estimates [6, 25]. Their results, published in Nature Communications, point to low marine boundary-layer clouds as a particularly potent amplifier . The study also ruled out the possibility that tropical low clouds could have a cooling effect that offsets warming
.
The strength of low-cloud feedback depends on where SSTs warm. A "warmer-get-higher" framework shows that spatial patterns in SST increase drive changes in tropical cloud fraction and height, which in turn govern the feedback magnitude .
The observed warming pattern over recent decades — with stronger warming in the western Pacific and weaker warming in the eastern Pacific — has historically moderated the low-cloud feedback somewhat [1, 8]. However, climate models project that pattern will shift toward greater eastern Pacific warming by late century, likely strengthening the feedback .
Updated emergent constraints using the latest data give a transient climate response (TCR) of about 1.81 K (very likely range 1.28–2.33 K), a modest increase over pre-2019 estimates . Some analyses suggest climate sensitivity (warming per CO₂ doubling) could be closer to 4.5°C, rather than the commonly assumed 2.5–3°C range
.
The observed decline in low-cloud albedo has contributed to the record-breaking Earth energy imbalance, which reached all-time highs in 2023 [9, 12]. Low-cloud reduction explains approximately 70% of the recent energy imbalance trend .
A sharp drop in sulfur aerosol emissions from global shipping after 2020, following International Maritime Organization regulations, has further reduced the reflectivity of low clouds over the Pacific and Atlantic, adding to the warming . This "cleaner air" effect is partly responsible for the surge in global temperatures in 2023–2024
.
A July 2026 study found that tropical low-cloud cover actually increased by about 0.18% per decade from 1979 to 2025, suggesting the historical cloud response has partially muted warming over the past ~47 years . This increase was mainly in stratocumulus clouds, which are especially effective at reflecting sunlight, producing a cooling effect estimated at about −0.79 W m⁻² K⁻¹
. However, climate models consistently show this protective effect will reverse and weaken under continued warming, with low-cloud feedback turning more strongly positive in the future
.
The years 2023 and 2024 were the warmest on record, with 2024 being the first calendar year more than 1.5°C above pre-industrial levels [11, 14]. The past 11 years have been the 11 warmest on record .
The evidence strongly converges on tropical low-cloud loss as a major amplifier of global warming. Satellite data already show a detectable decline in low-cloud amount and albedo, and observational constraints indicate the feedback is at the high end of model ranges. This raises the central estimate of climate sensitivity and suggests future warming may be faster than the average CMIP6 model projection — unless deep and rapid emissions reductions are achieved. The main remaining uncertainties are the exact magnitude of the feedback and how shifting SST warming patterns will modulate it over coming decades.