Evaluation of the NASA SRB Surface Fluxes as a Function of Meteorological State

Accurate determination of the surface radiation budget is a critical component of understanding the global energy and water cycles and climate change. The only way to provide surface radiation budget estimates that are consistent and global is via satellite-based products; such estimates are now available convering several decades. In order to make use of these products in climate studies, it is imperative that we understand and minimize their uncertainties. While general estimates of uncertainty can be made using surface observations at multiple locations around the globe, more focused investigation is required to identify the specific conditions under which errors in the satellite products occur.

In this study, we evaluate the NASA/GEWEX Surface Radiation Budget (SRB) radiative flux estimates as a function of meteorological state. We have constructed a classification of large-scale meteorological states at the ARM Climate Research Facility’s Southern Great Plains (SGP) and Darwin sites using ECMWF ERA-interim data. Typical SGP states include various stages of frontal passages during winter and summer convective conditions. Typical Darwin states include dry season, transition season and monsoon conditions. Time series of these states were assembled for the period 1979-2010 at the SGP and 1979-2012 at Darwin. We also obtained ground-based solar and longwave irradiance measurements at these locations. Simultaneous ground-based measurements and satellite estimates are composited based on the meteorological state classifications. We then carry out a statistical comparison of the satellite-derived and measured irradiances based on the meteorological state populations and draw conclusions regarding how well or poorly the satellite retrieval algorithms perform. For example, the smallest differences in downwelling solar irradiance at the SGP site occur when cold fronts are in the vicinity, while the largest differences in the longwave downward irradiance occur for cold conditions dominated by northerly winds. This analysis provides insights into conditions under which the largest errors occur and what might be reponsible for those errors. This will allow us to develop improved algorithms for the creation of surface-based estimates.