GoKawiilReflected SW radiation from CERES and its decomposition
Products and record length
Here we obtain monthly, 1° all-sky TOA SW fluxes between January 2001 and December 2025 (inclusive) from the CERES EBAF Edition 4.2.1 product15,16. These data are derived from measurements of SW radiance between 0.3 and 5.0 μm from CERES instruments onboard the Terra, Aqua and NOAA-20 satellites14,15,16 and these records have been synced in the EBAF product. To decompose broadband SW reflection at TOA into components including CRE, clear-sky surface and atmospheric reflections, we obtain TOA clear-sky (for the total region) SW fluxes and calculated fluxes at the surface also from the EBAF Edition 4.2.1 product. Furthermore, to enable a cloud type decomposition of the CRE, we obtain the Flux By Cloud Type (FBCT) Edition 4A product17 between January 2023 and December 2022 (inclusive). There is an apparent offset in TOA SW fluxes between those measured from NOAA-20 and those measured from Terra+Aqua, preventing us from using the FBCT product beyond 2022. The FBCT product integrates information from the Moderate Resolution Imaging Spectroradiometer (MODIS) to partition CERES observed fluxes into seven cloud effective pressure (P eff ) bins and six cloud optical depth (τ cld ) bins. We further condense the P eff –τ cld into nine bins, representing nine cloud types: cumulus (Cu), stratocumulus (Sc) and stratus (St) for low-level clouds capped at 680 hPa; altocumulus (Ac), altostratus (As) and nimbostratus (Ns) for mid-level clouds between 680 hPa and 440 hPa; and cirrus (Ci), cirrostratus (Cs) and cumulonimbus (Cb) for high-level clouds above 440 hPa. The separation of τ cld for a given cloud level is set at 3.6 and 23, following ref. 55 and illustrated in the inset of Fig. 2a.
It worth noting that the EBAF and FBCT products are produced with different CERES data streams15. Specifically, in the process of producing monthly mean fluxes, the diurnal filling in the EBAF product incorporates geostationary data to capture a complete diurnal cycle56, whereas the FBCT product assumes a diurnally invariant cloud field fixed at the time of the CERES overpasses. This is particularly relevant here because TOA SW fluxes are known to exhibit diurnal variability owing to cloud evolution, with deep convective and stratocumulus clouds showing distinctly different diurnal cycles57. Moreover, the EBAF and FBCT products are also not internally synced in terms of the total TOA radiation, as EBAF is also adjusted within observational uncertainty to match observed ocean heat uptake58. As a result, EBAF all-sky global mean SW reflection is about 3 W m−2 greater than that from FBCT. Given these differences in processing streams, results may therefore appear different between the two products. In fact, the climatological E–W symmetry appears robustly at the 27° E meridian based on the EBAF record, whereas it appears robustly at the 28° E meridian based on the FBCT product. Thus, whenever the FBCT product is analysed, the 28° E meridian is used to separate the EH and the WH. The interannual variability in E–W symmetry inferred from the two products tracks closely (not shown), supporting the robustness of the results presented here.
In a nutshell, the EBAF product is used for the majority of the study for its longer record and more complete diurnal filling, and we use its latest available record from January 2001 to December 2025. The FBCT product from January 2003 to December 2022 is used when cloud type decomposition is involved.
Cloud radiative effect and its decomposition into cloud type
The cloud contribution to the total all-sky TOA SW reflection is defined as the CRE, calculated as the difference between all-sky reflection (R) and clear-sky reflection (for the total region; R clr ). The CRE is then further decomposed into contributions from each cloud type, calculated as the cloud fraction (f cldtyp ) weighted flux difference between upwelling overcast TOA flux for that cloud type (\({F}_{{\rm{cldtyp}}}^{\uparrow }\)) and the upwelling clear-sky flux (\({F}_{{\rm{clr}}}^{\uparrow }\)):
$${R}_{{\rm{cldtyp}}}={f}_{{\rm{cldtyp}}}({F}_{{\rm{cldtyp}}}^{\uparrow }-{F}_{{\rm{clr}}}^{\uparrow })$$
Although the CRE is not independent of clear-sky fluxes, the simplicity of its calculation offers a straightforward way to interpret cloud changes themselves.
Clear-sky atmospheric and surface decomposition
... continue reading