Evaluation of the Tropical TOA Flux Diurnal Cycle in MERRA and ERA-Interim Retrospective Analyses

Abstract

Reanalysis model output is extensively used in atmospheric research and must be rigorously and continuously evaluated to understand the strengths and weaknesses. This paper evaluates the tropical top-of-atmosphere (TOA) flux diurnal cycle in NASA’s Modern-Era Retrospective Analysis for Research and Applications (MERRA) and the ECMWF Interim Re-Analysis (ERA-Interim) against Clouds and the Earth’s Radiant Energy System (CERES) synoptic edition 3A (SYN Ed3A) TOA flux data. MERRA and ERA-Interim are able to reproduce large-scale features of the diurnal cycle, including land–ocean contrast. MERRA and ERA-Interim, however, fail to reproduce many regional features of the climatological annual diurnal cycle. The TOA flux diurnal cycle errors in regions dominated by convective diurnal cycles are 5–10 times larger than in nonconvective regions. These errors in the TOA radiative flux diurnal cycle are primarily attributed to errors in the cloud diurnal evolution and specifically the failure to reproduce diurnally forced propagating convection. The largest diurnal cycle errors are found in ocean convective regions (e.g., Indian and equatorial Pacific Oceans); the observed longwave cloud forcing (LWCF) diurnal evolution in several oceanic convective regions shows two peaks: an afternoon and a near midnight peak; however, the reanalysis models produce a single midnight peak. The outgoing longwave radiation (OLR) diurnal cycle over tropical land is 20%–30% too weak in both reanalyses. The small diurnal cycle errors in marine stratocumulus regions are a result of two common misrepresentations in MERRA and ERA-Interim: 1) the dissipation of marine stratocumulus clouds from morning to afternoon is too slow and 2) the cloud diurnal cycle is too weak. Overall, the intermodel differences in the representation of the TOA flux diurnal cycle are smaller than the differences between reanalysis models and observations.