search

lawinslow / AncientIce

1 stars 0 forks source link

Comment 16: Intervals on CO2 before-after #27

Closed rBatt closed 9 years ago

rBatt commented 9 years ago
  1. Comment: Figure 4 A: I would have expected a wider confidence interval on the 1581--1681 CO2 regression coefficient, if only because the CO2 signal was much weaker then. Any commentary on why the uncertainty is actually wider for the 1897--1997 interval?

This is mostly a "just check or talk about it" type comment, but I know another comment pointed out that the time series we're comparing aren't necessarily stationary, so you can get mirage correlations. So I'll have to compare the intervals the reviewer is mentioning after the new analysis

rBatt commented 9 years ago

I think the CI's for suwa ice date and CO2 are wider for the second half because the first half is just a flat line. The supplement figure in f69ad2367ca99a95998b6a180cc0fc432c68f8eb makes this pretty clear (page 2 of the pdf, last set of scatter plots on that page)