Subtle Differences from Regular ISCCP DX Data
Product
ISCCP-ScaRaB DX dataset is constructed to be as close as possible to the regular ISCCP DX dataset. However, due to the nature of ScaRaB instrument, ISCCP-ScaRaB DX dataset slightly differs from the regular ISCCP DX dataset in the following three aspects:
Although the regular ISCCP DX native (NAT) dataset
has a nominal resolution of 30 km 30 km sampled image pixels), the
retrieved cloud (and surface) properties are actually radiatively weighted
averages over the small-scale (~5 km) variations present in each pixel.
The cloud and surface retrievals in the special ISCCP-ScaRaB DX dataset,
however, represent radiatively weighted averages over a larger spatial
area (~60 km at nadir) as a result of the coarse resolution of ScaRaB
scanner.
In the polar regions of regular ISCCP DX dataset,
near-infrared (NIR: ~3.7 um) radiances are used to augment detection and
radiative analysis of clouds over snow and ice surfaces. Because the NIR
channel is not available on the ScaRaB scanner, cloud detection and
retrieval are perfomed using VIS and IR channels only (over the entire
globe). As a result, the three NIR-related variables are not meaningful in
this dataset.
In the regular ISCCP DX dataset, three bytes are
used to store radiances for up to three extra wavelengths, if available.
In the special ISCCP-ScaRaB DX dataset, the TOA broadband SW and LW fluxes
(estimated by the ERBE methodology from broadband radiances, measured by
the ScaRaB instrument and co-located with the narrowband radiances used for
the ISCCP analysis) are packed into these three bytes. These two flux
values can be extracted using the sample read program.