This website's content is no longer actively maintained, but the material has been kept on-line for historical purposes.
The page may contain broken links or outdated information, and parts may not function in current web browsers.

Please note that ISCCP data processing is now performed at NOAA/NCEI.
Please visit the NOAA/NCEI website for access to ISCCP H data products and other up-to-date information.

Special Version of ISCCP DX Dataset From ScaRaB satellites

Determination of Top-of-Atmosphere Longwave Radiative Fluxes:
A Comparison Between Two Approaches Using ScaRaB Data


The TOA outgoing longwave radiation (OLR) can be directly estimated from satellite measurements of broadband radiances, such as the Nimbus-7 ERB, ERBE, ScaRaB, and the ongoing CERES. Alternatively, OLR can also be indirectly inferred from narrowband radiances, either by a regression-based narrowband-to-broadband conversion technique or by a detailed radiative transfer calculation with the cloud and surface radiative properties retrieved from narrowband radiances and other correlative atmospheric and surface datasets as input to the radiative transfer model.

Comparison between the model calculated OLR, using ISCCP retrieved cloud properties and other atmospheric and surface datasets, and the ERBE OLR shows that while the two fluxes agree reasonably well in general, systematic differences exist in specific locations suggesting particular local surface, atmospheric and cloud conditions are responsible for the disagreement between these two products. It is imperative to identify the sources of these discrepancies, because these systematic differences, though relatively small in magnitude, are certainly significant when compared with the radiation budget changes that appear interannually or that are being considered in relation to possible climate changes induced by changes in atmospheric composition and, consequently, cannot be ignored in evaluating the effects of cloud-radiation interactions on the climate. Identifying the sources of discrepancies will also help to identify the problems associated with the two approaches. Unfortunately, examining and isolating the sources of disagreement is complicated by the effects of different temporal and spatial sampling of the ISCCP and ERBE measurements.

Data and Method

The existence of both narrowband and broadband channels on the ScaRaB radiometer provides a unique opportunity to compare the narrowband-based ISCCP and broadband-based ERBE approaches using coincident, collocated, and coangular radiance measurements and flux estimates.

The following diagram schematically shows the key steps involved in the present study.

(1) ScaRaB narrowband radiances (VIS and IRW) are first used to retrieve the cloud and surface radiative properties using the ISCCP algorithm. The narrowband-retrieved quantities are then used as input (along with other ancillary datasets) to the radiative transfer model to calculate the broadband radiances, which will be compared with the simultaneously measured ScaRaB broadband radiances. Any discrepancies in the calculated and measured broadband radiances point to the uncertainties in the input data, satellite retrieval method, and radiative transfer model itself.

(2) For the pixels where the calculated and observed radiances agree, the model-calculated fluxes are compared to the fluxes converted from the measured broadband radiances using the ERBE ADMs, as well as the ADMs parameterized by Stubenrauch et al. (1993). Thus, we can isolate and examine in detail the effects of angle model differences between the two ADMs and the radiative transfer modelís treatment of the angle integration and quantify the effects that clouds have on the anisotropy of the radiation field.

ScaRaB Flow Chart (JPG)

Major Conclusions and Discussion

For the clear sky, the model underestimates the observed TOA LW radiances on the global average with the bias correlated to the column precipitable water amount. Although this bias can be substantially reduced by changing the water vapor profiles at lower latitudes, the exact source of the bias still requires study. A similar but smaller bias is also observed for the low clouds, primarily due to the clear sky effect since the TOA longwave radiation is largely determined by the atmosphere above when the clouds are low-level. For high clouds, depending on the optical thickness, the model-simulated LW radiances can be either less than or greater than the observed ones. When the clouds are optically very thin, the model underestimates the TOA LW radiance partly as a result of the fact that these clouds tend to be placed too high during the visible adjustment step of the ISCCP retrieval. In general, the modeled TOA LW radiances agree better with observations for the nadir-viewing pixels than for the pixels viewed from limb, since the excessively large limb-viewing pixels are more likely to be heterogeneous or partially cloud-filled.

Compared with the radiative transfer model, the limb-darkening in the ERBE angular models appears to be too weak for the optically thin clouds, but too strong for the optically thick clouds, and the difference between the ERBE ADMs and the angle treatment of the radiative transfer model increases with cloud height. Limitations in the radiative transfer model calculations, however, prevent definitive judgement on the ERBE angular models for the following reasons: 1) The radiative transfer modelís angle treatment is sensitive to the uncertainties in the cloud vertical structure (e.g., multi-layer clouds); and 2) the ERBE ADMs are generated using the complete range of observed cloud conditions, not just clear and overcast, whereas the plane-parallel radiative transfer model used in this study inevitably assumes optical homogeneity on the underlying pixels, which tends to underestimate the limb-darkening for the broken three-dimensional cloud fields. Nevertheless, this study clearly demonstrates that obtaining more accurate instantaneous longwave flux estimations from ERBE approach would require additional cloud classes based on cloud height and optical thickness as is planned for the advanced CERES analysis.


Stubenrauch, C., J.-P. Duvel, and R. S. Kandel, 1993: Determination of longwave anisotropic emission factors from combined broad- and narrowband radiance measurements. J. Appl. Meteor., 32, 848-856.

Further Information

Please also see:

Chen, T., and W.B. Rossow, 2002: Determination of top-of-atmosphere longwave radiation fluxes: A comparison between two approaches using ScaRaB data. J. Geophys. Res., 107, no. D8, 4070.

Other related studies that use ISCCP-ScaRaB DX data:

tubenrauch, C.J., V. Briand, and W.B. Rossow, 2002: The role of clear-sky identification in the study of cloud radiative effects: Combined analysis from ISCCP and the Scanner of Radiation Budget. J. Appl. Meteorol., 41, 396-412.

On-Line ISCCP Datasets | Global Radiation Budget
Related Projects | What's New