> "What is it you think is worthwhile investigating next?"
I think the significant issues are with the occasional historical event that impacts the ENSO measure but has little to do with the underlying oscillations. As I pursued the detailed fitting, these events became more and more evident.
The ones that I can name are the following, going back in time:
1. The Pinatubo volcanic eruption of 1991
2. The El Chichon eruption of 1982
3. The Agung eruption of 1963/1964
4. The Pacific climate shift of 1976/1977
5. A hot period during WWII from 1941 to 1945 that is coincidental with limited data or miscalibration as the military took over measurements
6. A cool period during WWI from 1917 to 1918 that may have similar problems
7. The Ksudach eruption of 1907 , *"Major eruptions in the Kamchatka have a disproportional effect on global climate, due to the low troposphere elevation at these latitudes, and the ease of dispersal of ash, aerosol, and gas."* This one shows up in my CSALT model of global temperatures even though it is not as large a volcano, VEI=5, as the VEI=6 volcanoes above.
The basic issue is still trying to remove all the extraneous noise so we can see the fundamental signal of interest.