6 Comments
User's avatar
Dave Woolcock's avatar

So - is this what they have been doing with the temperature record? (Truncated normalisation)

Expand full comment
John Dee's avatar

It's a distinct possibility because otherwise they'd have to junk a large number of stations whose records got going after 1961. What we want is a whistleblower!

Expand full comment
Dave Woolcock's avatar

And I do like worked examples to make your point - really clear

Expand full comment
John Dee's avatar

Cheers Dave, I wanted to work this through with rainfall pretty thoroughly. The same applies to temperature but folk get heated about that (couldn't resist). Coming up will be a similar series using the UK34 data whereby I flip to looking at tmax, tmin and tmean.

Expand full comment
Joel W. Hay, PhD's avatar

I think you should look at first differences; that nets out the time invariance of the rain, climate, location and atmospherics. Also in your covid time series.

Expand full comment
John Dee's avatar

That's all planned for when I get stuck into ARIMA time series analysis!

Expand full comment