Difference in longwave radiation between top and bottom of atmosphere

I downloaded data from 1979 to today of NCEP reanalysis-2 of upward longwave radiation from this page. Look for the files that are about 30MB in size.

This first graph is from nominal top of atmosphere. The orange line is a moving average I'm experimenting with in Excel2013:

This next graph is of the same quantity but from earth's surface:

Now here's the second graph minus the first:

The difference between the IR at the surface and the top of atmosphere (TOA) increases a for a few years, then levels off for about 15 years.  But overall not much trend.

Both quantities rise a bit over the measurement period, but surface flux rises slightly faster than TOA flux.  Let's call this difference between surface upward longwave radiation flux (ULRF) and top of atmosphere ULRF the "greenhouse dividend" even though it could be caused by any means.

The greenhouse dividend sort of increases for a while, but it stalls about 15 years ago. If it really was a greenhouse dividend it should more closely match the Keeling Curve (below), but doesn't.

Using the outgoing longwave radiation (OLR) file (9.53MB file "olr.mon.mean.nc") from here, here's global OLR:

It should be decreasing if CO2 was causing heat-trapping, but it isn't.

Now let's compare the greenhouse dividend to OLR and it sort of matches.

Here's the above graph but with particular focus on the moving averages I applied in Excel to help us compare:

The upward flux greenhouse dividend should positively correlate with CO2 (indicating more heat trapping) and the OLR should negatively correlate. But neither correspond to the Keeling Curve:

Three graphs above, notice how the OLR data (blue line) starts in 1974, then there's a gap of one year, then OLR continues where upward radiation starts in 1979. I guess there was a NOAA satellite that measured OLR that operated from 1974 to 1978. Then a gap year in 1978 before NOAA satellites resumed measuring again.

The upward radiation flux data starts 1 Jan 1979 and appears to correspond to the satellite era. I'm not yet sure how this data relates to the longer running similar data that goes back to 1948 I graphed a couple of posts ago.

But none of the data I've found so far corresponds to CO2 or "heat trapping".  To quote John Daly: still waiting for greenhouse...