Using the reanalysis, we can calculate also the mean of the downward radiation flux. Of course, it is the reanalysis, with a coarse resolution and its own limitations (especially discontinuities due to inclusion of varying kinds of data), but I think it is quite good nonetheless, if we look at the values after 1981 (start of the satellite era). So, the downward radiation flux, or the infrared emitted by the atmosphere toward Earth surface (one funny thing when you have some knowledge in physics, you know you are a being of light and that you are shining
). Data are for the whole above 69°N (radiation data are along a Gauss grid, I take values up to 69.5217°N, cell index 11), averaged according to surface (hoping I make no mistakes), for each day from 1st January, 1981 :
It is looking quite like the graph of the temperature according to the DNMI, which is not unexpected by the way. We are tempted to say, a warmer atmosphere radiates more energy, so what is the point?
Actually there is more in this. The mean level of radiation for Earth is around 5 - 6 km, so around 500 hPa, take or given a couple of hPa. Relationship between temperature and radiation is not linear, so we can calculate the flux of a black-body with our old friend, the equation of a black-body :
Flux = 5.67e-8 * T^4
The grid for the temperatures are not the same as the one for the radiation data, but I didn't care about this. The precision of the reanalysis is probably far worse than the small difference induced by averaging over a slightly different area. So I used the region above 67.5°N for the temperatures, calculate the radiation flux from the temperature and compare with DWLR, and averaged for the three months from November to January (February is still ongoing un 2017). The year of reference is that of January (ie., NDJ 2017 is the average of ND 2016 and J 2017) :
Values are roughly the same, so yeah 500 hPa is a good level. But correlation is not looking good actually... I look to other levels, but the correlation is not significantly better. As a side-note, the temperatures at 500 hPa :
So, what if we try with the precipitable water?
It looks way better...
So I detrend the series, to compare the cross-correlation. Anomalies of the downward radiation flux, explained by the anomalies of the black-body emmision :
Anomalies of the downward radiation flux, explained by the anomalies of the precipitable water :
Anomalies of the anomalies of the precipitable water, explained by the temperature at 500 hPa :
I will not try an argument about chickens and eggs. It is of course difficult to disentangled all the mechanisms ongoing. But at least the increasing of water vapor, linked to warming of the temperatures but also to the decrease of Arctic sea ice, is increasing downward radiations.
The major point is that a warming of 20°C or 30°C is not impossible at surface is thus not impossible. With global warming, the "thin" -a 2 km thick and 20°C inversion is massive for an inversion in the absolute, but compared to the whole atmosphere this it is not so thick nor so cold- the "thin" layer of permanent inversion is set to be destroyed, with only marginal warming above. Usually there is around 5 to 10°C between surface and 850 hPa. Even a 7-8°C lapse rate with a 850 hPa layer around 250K would imply a mean surface temperature a bit below 260K, around -15°C, barely enough cold for sea ice. This graph shows the warming of the Arctic layers :
The surface 1000 hPa is warming fast and is now warmer than the 850 hPa for the first time since 1981 (and probably since many millenniums...). And the strength of the inversion (or of the now non-inversion) taken as the difference between the 850 hPa and 1000 hPa temperatures :
I will post the spreadsheet with the data a bit latter
P.S. : THe spreadsheet www.climatvisu.fr/Neven_ASIF/dlwr_out_2.ods