Sensitivity of Longwave Fluxes to Clouds and Meteorology: Establishing Uncertainties for GEWEX SRB Longwave Release 4

The International Satellite Cloud Climatology Project (ISCCP) has released an updated cloud and surface properties data product called the H-series (Young et al. 2018). This climate data record has made improvements to the past data products (D-Series, for example) by having access to the higher spatial resolution data available in the ISCCP B1U data sets. Because of the higher spatial resolution (10km vs. 30km) of the input data, data gridded to 1° x 1° now has a much greater number of pixels and thus a much finer granularity in cloud fractions than with the D-series.

The NASA Global Energy and Water Exchanges (GEWEX) Surface Radiation Budget (SRB) project is embarking on the creation of a new release of surface and top-of-atmosphere fluxes using the ISCCP H-series cloud properties. This data set overlaps CERES, providing comparable comparisons of fluxes, but also predates CERES which allows for longer term assessments of fluxes. Besides the higher spatial resolution, the ISCCP project has recalibrated and improved cloud retrievals. Additionally, the input meteorology (temperature and specific humidity profiles) for ISCCP H series has been completely revised and released as the ISCCP nnHIRS. The nnHIRS is used for the H series cloud properties such as cloud top temperature and pressure.

For this study, we evaluate the effect of the updated cloud properties on the longwave fluxes. In particular, the fluxes are analyzed by the results of the cloud overlap scheme that was developed and used in the GEWEX SRB Release 3.1 Longwave data set. That data set featured a maximum overlap within a high, middle, and low cloud pressure bounds. A random overlap between those high, middle and low layers was then used to estimate cloud fractions obscured by higher cloud overcast conditions. This scheme is reviewed and compared to new updated schemes adapting more recent treatments. The cloud properties within high, middle and low cloud classes are also analyzed to see if there are significant changes from the original D-series and to CERES and CALIPSO/CloudSat observations in a few limited cases. We also evaluate the effect using this nnHIRS data set within the longwave algorithm against using another reanalysis assimilation, such as the NASA/GMAO’s MERRA-2 re-analysis. We assess the resulting fluxes by running the algorithm for selected periods and comparing to surface based flux measurements and to other radiative flux data sets such as CERES.