Back to content.
I realized that my previous post I was not clear enough about the science to be convincing.
So let me try again, this time pointing out exactly what is wrong with the conclusions of Ding et al 2017.
First of all, what does Ding et al claim.
I think it was Steven who correctly pointed out that there are TWO parts to Ding et al's claims of influence of "internal variability" on Arctic sea ice :
(1) about 60% of the September Arctic sea-ice extent decline since 1979 is caused by trends in summertime atmospheric circulation.
(2) about 70% of those atmospheric circulation trends is caused by internal variability (and hence the remaining 30% is anthropogenic).
Now the problem I have with the paper are very clearly in point (1) :
It started with Ding et al claim in the abstract of that point :
Here, we present evidence that trends in summertime atmospheric circulation may have contributed as much as 60% to the September sea-ice extent decline since 1979.
In my previous post, I noted that Ding et al uses "GL-Z200" (geopotential height at 200 mb, over Greenland) as their metric for "atmospheric circulation" over the Arctic.
That immediately means that the conclusion should have read something like this :
Here, we present evidence that trends in summertime geopotential height may have contributed as much as 60% to the September sea-ice extent decline since 1979.
Now, this should raise some eyebrows here and there, since geopotential height is clearly directly correlated with temperature. If the temperature goes up, so does geopotential height.
In fact, the ideal gas law (PV=nRT) DICTATES that geopotential height goes up with temperature.
Which is sustained in the ERA data set, where Ding et al itself reports that there is a R=0.94 correlation between geopotential height and temperature.
So AGW, which is certainly a part (if not all) of the trend in temperature over the Arctic is certainly also a part (if not all) of the trend in geopotential height.
And thus Ding et al 2017 can no longer claim that the Z200 trend (which they call "atmospheric circulation trend") is a source of "internal variability".
But that's not all.
The even bigger problem with the paper is in which they determined the influence of "atmospheric circulation" (oops, I mean geopotential height) on Arctic sea ice decline.
If you want to investigate that influence with a sea ice/climate model, you would run the model with the ERA climate parameters, and compare it with a run that has geopotential height adjusted (subtract the geopotential height anomalies from the ERA data).
But that is not what Ding et al did.
They adjusted ALL variables (geopotential height, temperature, moisture content, downwelling LW radiation etc), essentially making the 'climate' constant.
Specifically, this is described here in the paper :
Exp-6: Same as Exp-5 except that the atmospheric forcing is modified to excise the forcing associated with the trends in the Greenland circulation pattern. To remove the circulations trend from the observat ions, we first construct the thirty-six-year seasonal (JJA) averaged time series of the Z200 index over Greenland, Z200 GL (GL-Z200 in Fig. 1c). We then linearly regress a key variable B
against this time series to obtain spatial pattern β(x,y) of the variable associated with the Greenland circulation index. Specifically, for the variable B we have
B(x, y, t) = β(x, y) × Z200 GL (t) (1)
where B represents a forcing field (for example, 10 m zonal wind, DLR, temperature, and so on), x and y indicate the location, t indicates time (JJA), Z200 GL is the Greenland Z200 index (GL-Z200 in Fig. 1c), and β is the regression coefficient. In the second step, the seasonal mean anomalous value of each forcing field is subtracted from the observed daily (or 6-hourly) forcing data during the summer—rendering a modified forcing that does not include variability or trends in variables that are asso ciated with Z200 GL. In the nine non-summer months, the forcing is exactly the same as that used in the Exp-5 control experiment. Given a strong correlation between circulation and surface winds, temperature, specific humidity, sea-level pressure, and downwelling long wave radiation in the Arctic, variability and trends in these six variables that are associated with Z200 GL are processed and removed from the forcing. The initial states of ocean, sea ice and atmosphere in Exp-5 and Exp-6 are exactly the same.
I highlighted in bold the section that describes that indeed they adjusted ALL variables to the extent that they linearly associate with Z200 GL. And because the correlation between all these variables and Z200 is very high (see figure 1c in the paper), they effectively eliminated ALL trends in ALL variables from their Exp-6 run, which means they effectively made the climate constant in that run.
Since the climate did not change in that run, they obtained a much lower ice melt rate.
Essentially the conclusion should thus have been :
Here, we present evidence that trends in summertime climate change may have contributed as much as 60% to the September sea-ice extent decline since 1979.
Which means that (this experiment 5 versus 6 suggests that) :
60% of Arctic sea ice reduction is caused by summer-time climate change, while 40% is caused by climate change over the remaining 9 months.
Which is an interesting conclusion, but as you can see it has NOTHING to do with ANY sort of "natural variability", so Ding et al cannot use it in their variability conclusion either.
You can still choose not to believe that this paper is seriously flawed, but you can no longer claim that you did not know.