The correction to the HadCRUT4 global temperature series as described by Cowtan and Way [1] and applied to the CSALT model is evaluated by the following fit:

Fig 1 : The top panel shows the original HadCRUT4 CSALT fit and the bottom panel is the CW Hybrid correction. No filtering was applied to the data. Note the narrowing of the gap in recent years

The previous post predicted that the recent diverging decrease in warming was an artifact of not taking measurements of the rapidly warming Arctic region, and since energy is balanced over the earth a decrease was to be expected.  But once estimates of the Arctic and other regional warming was taken into account, this divergence would close up.

From the residual (Figure 2) one can see the variance shrink as expected.. Recall that the residual is flat over the entire span starting in 1880, as the CSALT components including the growing ln(CO2) factor, perform an excellent job of modeling the temperature profile.

Fig 2 : The CW Hybrid fit residual is flatter as indicated in the bottom panel

The filtered fit (Figure 3) makes the shrinking of the divergence more apparent.

Fig 3 : Applying a 12-month filter the narrowing of the gap be becomes clearer

The CO2 correlation in Figure 4 tightens as the CW Hybrid correction is applied.

Fig 4 : The correlation between temperature anomaly and ln(CO2) sensitivity is slightly stronger pushing the TCR from 1.95C to above 2C.

The CW Hybrid correction is subtle but necessary.

The point of these corrections, no matter how small, is that they are needed to reduce the epistemic (i.e. systemic) uncertainty of the scientific modeling. The fact that our measurements are not always in sync with our models is a challenge that we can always work on to correct.

So we applaud that C&W correct for the lack of Arctic and other region data. And John Kennedy corrects for the biases introduced by changing measurement procedures [2]. These are good things.

Related Posts

1. CSALT and SST corrections
2. Detailed Analysis of CSALT Model
3. CSALT model
4. Climate Variability and Inferring Global Warming
5. CSALT Ju-Jutsu

References