A unique model/data comparison capability is made possible by the unique viewing geometry from NASA’s DSCOVR Mission Lissajous orbital location around the Lagrangian L1 point. The key point of this unique location is the large orbital inclination relative to the perpendicular of the Sun-Earth line-of-sight. This circumstance enables periodic Sun-Earth-Satellite phase angle shifts ranging from 2-degrees to 12-degrees with repeating ∼3-month periodicity. At such extreme phase angles, backscattered radiation for spherical cloud-top particles is strongly phase angle dependent, but not for irregular-shaped ice particles. Also key, are the near-hourly high-resolution EPIC images that have been converted to radiative solar fluxes by extensive use of ancillary satellite data and CERES-based ADMs. These EPIC-derived SW fluxes, integrated over the Earth’s sunlit hemisphere, constitute the EPIC Composite dataset of 1-day resolution global-scale reflected SW fluxes, which have been shown to agree well with CERES reflected SW fluxes. Using the EPIC data as a template, the DSCOVR satellite ephemeris enables aggregation of climate GCM run-time output over the sunlit hemisphere with the same viewing geometry as EPIC. Generating the GCM-equivalent global-scale SW flux dataset, together with the EPIC data, forms the basis for a new paradigm in model/data comparisons. The key advantages of this DSCOVR-style approach are the (1) identical space-time sampling with identical viewing geometry and complex, but identical averaging over the diurnal cycle between observations and climate GCM output data, (2) preservation of short-period variability at 1-day resolution due to the Earth’s rotation, and (3) self-consistent weather noise suppression by identical averaging over the sunlit hemisphere. Early examples of the EPIC data variability drew concerns from colleagues worried that the variability in the EPIC data might be modeling noise. There is no other way to resolve this concern but to find another data source that shows the same degree of variability. Definitive comparisons to NISTAR measurements presented in this study unequivocally confirm that the global EPIC-derived variability is indeed real, and not a data artifact.
Lacis et al. (Wed,) studied this question.