Top-of-atmosphere and surface radiation budget from broadband radiometers
Earth's energy imbalance, currently around 0.9 W/m² by CERES analyses, is the root signal of long-term climate change. Broadband radiometers on Terra, Aqua, NOAA-20 and Meteosat measure it, but a ~2 W/m² absolute calibration uncertainty means the trend matters more than any single epoch value.
Sensors
- CERES (Clouds and the Earth's Radiant Energy System) on Terra and Aqua: Twin broadband scanning radiometers measuring shortwave (0.3–5 µm) and longwave (8–12 µm total-channel) radiance. Terra crosses the equator at 10:30 local time; Aqua at 13:30. Footprint at nadir is roughly 20 km, degrading to ~40 km at scan edges. Absolute calibration uncertainty is approximately 1 W/m² for shortwave and 0.5 W/m² for longwave, with a combined flux uncertainty near 2 W/m² after angular distribution model (ADM) corrections. Both instruments have operated continuously since 2000 and 2002 respectively, providing the longest consistent TOA flux record available.
- CERES on NOAA-20 (JPSS-1): Launched in 2017, this instrument extends the CERES time series into the JPSS era with a 13:30 equatorial crossing time similar to Aqua. Cross-calibration with the heritage instruments is ongoing and critical: inter-satellite offsets of even 0.3 W/m² can alias into apparent decadal trends. NOAA-20's CERES also supports near-real-time energy budget monitoring with latency of roughly one to two days for Level-2 products.
- GERB (Geostationary Earth Radiation Budget) on Meteosat Second Generation: GERB provides broadband shortwave and total-channel radiance from geostationary orbit at 15-minute temporal sampling, with a native footprint of roughly 50 km at sub-satellite point. The high cadence resolves the diurnal cycle of reflected solar and emitted thermal radiation that polar-orbiting CERES samples only twice per day. GERB and CERES are routinely cross-validated; GERB's absolute accuracy is somewhat lower than CERES but its temporal density is irreplaceable for regional energy budget closure studies.
- ERBE (Earth Radiation Budget Experiment, heritage archive): The predecessor scanning and non-scanning radiometers on ERBS, NOAA-9 and NOAA-10 operated from 1984 to 1999. The archive is the only source of pre-CERES broadband flux data and is used to extend decadal trend analyses, though inter-mission calibration offsets limit quantitative comparisons. Useful primarily for establishing climatological baselines rather than precise trend attribution.
What a 0.9 W/m² imbalance actually means to measure
The Earth absorbs roughly 240 W/m² of solar radiation averaged over the globe and emits nearly the same amount as longwave thermal radiation. The difference, currently estimated at about 0.9 W/m² by published CERES analyses, is the net energy accumulating in the climate system. That number sounds small. Spread across the entire surface area of the Earth, it is equivalent to detonating several Hiroshima-scale devices every second, continuously. The signal is real and consequential.
The measurement problem is that 0.9 W/m² sits well inside the absolute calibration uncertainty of the instruments designed to detect it. CERES shortwave channels carry roughly 1 W/m² absolute uncertainty; longwave channels around 0.5 W/m². Combined, the uncertainty envelope on instantaneous global-mean flux is close to 2 W/m². The imbalance is therefore not directly observable from a single satellite epoch. It is inferred from multi-year trend analysis, anchored by independent in-situ ocean heat content measurements from the Argo float network. Satellite data and ocean data must agree before the figure is trusted.
How CERES converts a radiance reading into a flux
A radiometer in orbit measures radiance: energy arriving from one direction. The climate system needs flux: energy integrated over the entire hemisphere above a surface. Bridging that gap requires Angular Distribution Models (ADMs), which describe how different scene types (ocean, desert, cloud, ice) scatter and emit radiation in different directions. CERES ADMs were built from millions of coincident multi-angle observations and are scene-classified using co-registered MODIS imagery on the same Terra and Aqua platforms.
The ADM correction is not small. For a broken cloud field viewed at a large scan angle, the radiance-to-flux conversion factor can differ by 20 percent or more from the clear-ocean factor. Errors in scene classification propagate directly into flux errors, which is why the CERES instrument is always operated alongside a high-resolution imager. The published Edition 4 CERES data products incorporate revised ADMs and improved cloud retrievals; earlier editions should not be mixed with current ones in trend analyses without documented inter-edition adjustments.
Temporal sampling is a separate source of uncertainty. CERES on any single polar platform samples a given location twice per day. Diurnal cycles in cloud cover and surface temperature mean that those two snapshots may not represent the daily mean accurately. GERB on Meteosat, sampling every 15 minutes, is the primary tool for characterising and correcting this diurnal aliasing over Europe and Africa.
What the surface radiation budget adds, and why it is harder
The top-of-atmosphere (TOA) budget is the cleanest measurement because there is nothing above the satellite to complicate it. The surface radiation budget requires knowing how much of the incoming solar and downwelling longwave radiation actually reaches the ground after the atmosphere has absorbed and scattered it. That requires a full atmospheric column model: aerosol optical depth, cloud properties, water vapour, and ozone all modify the surface flux relative to the TOA flux.
CERES produces surface radiation budget estimates by combining its TOA measurements with radiative transfer calculations constrained by ancillary atmospheric profiles. These are model-mediated products, not direct observations. The published uncertainty on surface shortwave flux is around 5 W/m² regionally, rising to 10 W/m² or more in persistently cloudy regions where cloud optical depth retrievals are least reliable. Users who need surface radiation for agricultural, solar-energy or ecological applications should treat these as climatological inputs rather than point-specific ground truth.
The decadal trend problem: calibration drift versus real change
Detecting a change in Earth's energy imbalance over a decade requires that the instruments themselves have not drifted by more than the signal being sought. CERES uses onboard calibration sources and vicarious calibration against stable desert targets, but demonstrating sub-0.3 W/m² stability over twenty years is genuinely difficult. The CERES Science Team addresses this through a combination of internal consistency checks, cross-calibration between instruments on different platforms, and comparison against the independent ocean heat content record.
The practical implication for any analysis of decadal radiation budget trends: the relative change from year to year within a single instrument's record is more reliable than the absolute value at any epoch. A step change in apparent flux that coincides with a satellite transition or an ADM edition update should be treated with suspicion until the calibration team has published an assessment. This is not a flaw unique to CERES; it applies to every long-term climate data record assembled from successive satellite generations.
Regional applications: where the numbers become actionable
Global-mean energy balance is a climate-science quantity. Regional radiation budget anomalies, by contrast, have direct operational uses. A persistent positive shortwave anomaly over a semi-arid region may indicate reduced cloud cover or decreased aerosol loading, both of which affect surface temperature and evapotranspiration. Negative longwave anomalies at TOA over the Arctic in winter signal reduced sea-ice extent through the well-documented ice-albedo feedback mechanism.
CERES regional monthly composites at 1-degree resolution are the standard input for evaluating climate model radiation schemes. Discrepancies between modelled and observed TOA shortwave cloud radiative effect, which can reach 5 to 10 W/m² regionally in the tropics, remain one of the primary diagnostics for cloud parameterisation errors in general circulation models. Governments commissioning climate-risk assessments and national adaptation plans increasingly request these regional diagnostics as part of their evidence base. Satellize can compile and contextualise CERES regional products for specific national domains, as it does for agricultural analytics in the Kingdom of Tonga crop-estimation programme, where surface energy balance feeds into crop-stress modelling.
For solar-energy developers, CERES-derived surface shortwave climatologies provide a satellite-based cross-check on ground station records and reanalysis products. The 5 W/m² regional uncertainty is acceptable for site screening at the feasibility stage; bankable yield assessments still require ground-based pyrheliometry.
Honest limits, and what comes next
The 2 W/m² absolute calibration uncertainty is the governing constraint on this entire field. It cannot be engineered away with better algorithms; it requires a more stable onboard reference or an independent absolute standard in orbit. NASA's CLARREO Pathfinder instrument, designed to demonstrate spectrally resolved absolute calibration at the 0.3 W/m² level, flew a demonstration on the International Space Station. Whether a full CLARREO mission proceeds depends on funding decisions not yet finalised at the time of writing.
Cloud cover imposes a different kind of limit. CERES retrieves cloud properties from co-registered MODIS or VIIRS imagery, but optically thick convective systems saturate the visible channels and introduce retrieval ambiguity in the longwave. The polar night removes shortwave observations entirely for months at a time over the Arctic and Antarctic, where the surface energy budget is most sensitive to change.
Despite these constraints, the CERES record is the most complete and internally consistent broadband radiation budget dataset in existence. For climate-treaty verification, national climate assessments, model evaluation and solar-resource characterisation, it remains the reference against which all other approaches are measured.
Typical figures
| Nadir footprint (CERES) | ~20 km at nadir, ~40 km at scan edge |
| Nadir footprint (GERB) | ~50 km at sub-satellite point |
| Temporal revisit (CERES polar) | 2 overpasses per day per platform; ~4 per day with Terra + Aqua combined |
| Temporal revisit (GERB geostationary) | 15-minute cadence |
| Spectral channels | Shortwave 0.3–5 µm; longwave (total minus shortwave); window channel 8–12 µm on some configurations |
| Absolute calibration uncertainty (flux) | ~1 W/m² shortwave, ~0.5 W/m² longwave; combined ~2 W/m² after ADM correction |
| Surface flux regional uncertainty | ~5 W/m² shortwave in clear sky; up to ~10 W/m² in persistently cloudy regions |
| Standard gridded product resolution | 1-degree monthly means (CERES EBAF Edition 4); hourly 1-degree for SYN1deg product |
| Archive depth | CERES: March 2000 (Terra) to present; ERBE heritage: 1984–1999 |
| Level-2 product latency | 1–2 days for near-real-time; ~3 months for science-quality Edition products |
Analytics Satellize can run
| Regional TOA energy budget anomaly report | Differencing of CERES EBAF Edition 4 monthly 1-degree composites against a user-defined climatological baseline period, with cloud radiative effect decomposition | Annual PDF report with gridded anomaly maps and time-series charts for a defined national or regional domain |
| Surface shortwave climatology for solar-resource screening | CERES SYN1deg hourly surface downwelling shortwave, aggregated to monthly and annual means with percentile distributions; cross-checked against ERA5 reanalysis | GIS layer (GeoTIFF) and summary statistics table per candidate site polygon |
| Cloud radiative effect trend analysis | Separation of all-sky and clear-sky TOA fluxes from CERES EBAF to isolate shortwave and longwave cloud radiative effect; Mann-Kendall trend test applied to monthly anomaly series | Trend report with significance levels and breakpoint analysis flagging instrument-transition artefacts |
| Climate model radiation bias assessment | Comparison of CMIP6 model TOA shortwave and longwave fluxes against CERES EBAF reference, following published CERES Science Team evaluation protocols | Bias maps and Taylor diagram summary for specified model ensemble, delivered as a technical annex for national adaptation plan submissions |
| Diurnal radiation cycle characterisation | Fusion of GERB 15-minute geostationary observations with CERES polar snapshots using published diurnal correction algorithms to reconstruct full daily flux cycle | Monthly diurnal composite dataset for Europe/Africa domains, in NetCDF with accompanying methodology note |
| Energy imbalance context briefing | Synthesis of CERES global-mean net flux estimates with published Argo ocean heat content data, following the approach documented in Loeb et al. published analyses, to frame observed imbalance within calibration uncertainty bounds | Four-page executive briefing for non-specialist government audiences, updated annually |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.