Background
The authors write that "atmospheric downward longwave radiation at the surface (Ld) quantifies the atmospheric greenhouse effect," which makes its evaluation one of the primary objectives of those who seek to divine its future impact on the planet.

What was done
According to Ma et al., the modeled Ld of 44 general circulation models (GCMs) that participated in the CMIP5 program were compared with land-based observations acquired at 47 sites of the Baseline Surface Radiation Network (BSRN), 51 sites of the Coordinated Energy and water cycle Observations Project (CEOP), 34 sites from the AmeriFlux network over the United States, 18 sites from the AsiaFlux network in Asia, and 12 sites where buoy-based measurements were made over tropical oceans over the period 1992-2005.

What was learned
The three researchers determined, among a number of other things, that (1) "GCMs in CMIP5 could not accurately simulate the diurnal variation of surface temperature and water vapor," and that (2) "CMIP5 GCMs are still poor in producing monthly anomalies of Ld," which is likely due to (3) "the GCM's poor performance in simulating seasonal variation of clouds and monthly anomalies of air temperature and water vapor."

What it means
In light of these several shortcomings, we have to ask ourselves the proverbial question about the integrity of the world's most up-to-date climate models: are we there yet? Apparently not; for however close the models may seem to be getting to replicating reality, they still fall significantly short of the mark.