Improvements to Broadband Shortwave and Longwave Top of Atmosphere Fluxes Derived from Geostationary Satellites

Monitoring of top-of-atmosphere (TOA) broadband (BB) shortwave (SW) and longwave (LW) fluxes is important for many applications, including cloud-radiative interactions, as well as evaluating climate change. Geostationary satellites provide coverage over large areas of the world and thus can be valuable to this endeavor, however, most are equipped with only narrowband (NB) instruments. Narrowband-to-broadband (NB-BB) flux conversions have been employed operationally within the NASA Langley SatCORPS (Satellite Cloud Observations and Radiative Property retrieval System) cloud and radiation property dataset in order to derive these climatically relevant parameters at pixel-level resolution. However, these conversions have historically been based on a simple land or ocean domain-specific formulation, based on matching coincident geostationary NB with CERES (Clouds and the Earth's Radiant Energy System) BB fluxes from approximately twice-daily Terra or Aqua overpasses. Limitations in this approach include an incomplete sampling over the full solar zenith angle range, failure to capture scene type-based differences, and usage of only one NB channel (0.65 µm for SW, 11 µm for LW).

To improve the accuracy of the derived TOA fluxes and their diurnal cycle, a number of improvements are made to the SatCORPS geostationary NB-BB fit procedure. Both LW and SW NB-BB fits are derived based on more specific IGBP (International Geosphere-Biosphere Programme) scene types.  Longwave fits will expand upon the operational use of 11-µm channel data by also utilizing channels sensitive to water vapor. The shortwave process employs a theoretical radiative transfer model that is not be limited by sampling issues at high solar zenith angles. These methods are initially applied to GOES-13 covering the ARM (Atmospheric Radiation Measurement) Mobile Facility GO-AMAZON (Green Ocean-Amazon) 2014 campaign, then applied to other satellites and domains.  Validation is provided by comparisons with Fu-Liou modeled fluxes, as well as the hourly one-degree resolution gridded CERES Edition 4 SYN1DEG product.