ENSO and other variability in 116 and 118

[islasimpson]

Here are some CVDP diagnostics of runs 116 and 118, compared to run 112 and a couple of segments of 104 from @phillips-ad.

(1) 116 and 118 are both improved in terms of the dry slot issue...

(2) ENSO teleconnections to the North Pacific look a bit better in 116 than they were in 112. 118 looks a bit bad though. Maybe too short to say anything for sure about this.

(3) In the plot above there are also indications that the structure of ENSO is too narrow in the latitudinal direction, particularly in 118. Not sure if this would be related to all the MLD issues @gustavo-marques?

(4) 116 and 118 both seem to be lacking in tropical Pacific SST variability as you can see in the SST standard deviation

and the bias in this standard deviation from obs

and in the Nino3.4 monthly variance

So, in terms of how 116 and 118 are doing, in tropical Pacific and related variability it's mixed. Dry slot issue is definitely better, but a lack of ENSO variance and perhaps an ENSO structure that's too latitudinally narrow are both features to look out for moving forward.

@gustavo-marques, @fobryan3, @dlawrenncar, @adamrher, @PeterHjortLauritzen, @swrneale, @cecilehannay, @phillips-ad

[islasimpson]

@megandevlan

[megandevlan]

My hunch is that 118 especially is still a little too short, but the ENSO variances are super low compared to other simulations (using years 2-end of runs):
.

But pretty surprisingly, the westward extent of the strongest correlation between SST anomalies and nino 3.4 index is improving a little bit (at least for the 0.5 contour). That said, the negative correlations are really strongly reduced too, which seems less good.