Shocking to me, at least.
You may recall from my previous post that I came up with a simple-minded way of calculating yearly anomalies, that is by subtracting the mean of all data from a given location from each of the yearly means.
Hansen describes a much more complicated method of calculating the yearly anomalies. I won't go into it at this point, but I was concerned that my simple approach might make a significant difference.
And so I compared the two. That is, anomalies calculated using Hansen's method with my simple method. And here is the result:
The black plot is dset1 with anomalies calculated by Hansen's method. The red dashed plot is dset1 with anomalies calculated using my method. If you have trouble seeing the red plot it's because it's right on top of the black one.
Here's a plot of the difference between the two:
So the difference is very small. Once again Hansen's complicated approach appears to be unnecessary.